diff options
Diffstat (limited to 'capstone/suite/synctools/tablegen/ARM')
22 files changed, 38283 insertions, 0 deletions
diff --git a/capstone/suite/synctools/tablegen/ARM/ARM-digit.td b/capstone/suite/synctools/tablegen/ARM/ARM-digit.td new file mode 100644 index 000000000..d915f49be --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARM-digit.td @@ -0,0 +1,1098 @@ +//===-- ARM.td - Describe the ARM Target Machine -----------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// Target-independent interfaces which we are implementing +//===----------------------------------------------------------------------===// + +include "llvm/Target/Target.td" + +//===----------------------------------------------------------------------===// +// ARM Subtarget state. +// + +def ModeThumb : SubtargetFeature<"thumb-mode", "InThumbMode", + "true", "Thumb mode">; + +def ModeSoftFloat : SubtargetFeature<"soft-float","UseSoftFloat", + "true", "Use software floating " + "point features.">; + + +//===----------------------------------------------------------------------===// +// ARM Subtarget features. +// + +// Floating Point, HW Division and Neon Support +def FeatureVFP2 : SubtargetFeature<"vfp2", "HasVFPv2", "true", + "Enable VFP2 instructions">; + +def FeatureVFP3 : SubtargetFeature<"vfp3", "HasVFPv3", "true", + "Enable VFP3 instructions", + [FeatureVFP2]>; + +def FeatureNEON : SubtargetFeature<"neon", "HasNEON", "true", + "Enable NEON instructions", + [FeatureVFP3]>; + +def FeatureFP16 : SubtargetFeature<"fp16", "HasFP16", "true", + "Enable half-precision " + "floating point">; + +def FeatureVFP4 : SubtargetFeature<"vfp4", "HasVFPv4", "true", + "Enable VFP4 instructions", + [FeatureVFP3, FeatureFP16]>; + +def FeatureFPARMv8 : SubtargetFeature<"fp-armv8", "HasFPARMv8", + "true", "Enable ARMv8 FP", + [FeatureVFP4]>; + +def FeatureFullFP16 : SubtargetFeature<"fullfp16", "HasFullFP16", "true", + "Enable full half-precision " + "floating point", + [FeatureFPARMv8]>; + +def FeatureVFPOnlySP : SubtargetFeature<"fp-only-sp", "FPOnlySP", "true", + "Floating point unit supports " + "single precision only">; + +def FeatureD16 : SubtargetFeature<"d16", "HasD16", "true", + "Restrict FP to 16 double registers">; + +def FeatureHWDivThumb : SubtargetFeature<"hwdiv", + "HasHardwareDivideInThumb", "true", + "Enable divide instructions in Thumb">; + +def FeatureHWDivARM : SubtargetFeature<"hwdiv-arm", + "HasHardwareDivideInARM", "true", + "Enable divide instructions in ARM mode">; + +// Atomic Support +def FeatureDB : SubtargetFeature<"db", "HasDataBarrier", "true", + "Has data barrier (dmb/dsb) instructions">; + +def FeatureV7Clrex : SubtargetFeature<"v7clrex", "HasV7Clrex", "true", + "Has v7 clrex instruction">; + +def FeatureDFB : SubtargetFeature<"dfb", "HasFullDataBarrier", "true", + "Has full data barrier (dfb) instruction">; + +def FeatureAcquireRelease : SubtargetFeature<"acquire-release", + "HasAcquireRelease", "true", + "Has v8 acquire/release (lda/ldaex " + " etc) instructions">; + + +def FeatureSlowFPBrcc : SubtargetFeature<"slow-fp-brcc", "SlowFPBrcc", "true", + "FP compare + branch is slow">; + +def FeaturePerfMon : SubtargetFeature<"perfmon", "HasPerfMon", "true", + "Enable support for Performance " + "Monitor extensions">; + + +// TrustZone Security Extensions +def FeatureTrustZone : SubtargetFeature<"trustzone", "HasTrustZone", "true", + "Enable support for TrustZone " + "security extensions">; + +def Feature8MSecExt : SubtargetFeature<"8msecext", "Has8MSecExt", "true", + "Enable support for ARMv8-M " + "Security Extensions">; + +def FeatureSHA2 : SubtargetFeature<"sha2", "HasSHA2", "true", + "Enable SHA1 and SHA256 support", [FeatureNEON]>; + +def FeatureAES : SubtargetFeature<"aes", "HasAES", "true", + "Enable AES support", [FeatureNEON]>; + +def FeatureCrypto : SubtargetFeature<"crypto", "HasCrypto", "true", + "Enable support for " + "Cryptography extensions", + [FeatureNEON, FeatureSHA2, FeatureAES]>; + +def FeatureCRC : SubtargetFeature<"crc", "HasCRC", "true", + "Enable support for CRC instructions">; + +def FeatureDotProd : SubtargetFeature<"dotprod", "HasDotProd", "true", + "Enable support for dot product instructions", + [FeatureNEON]>; + +// Not to be confused with FeatureHasRetAddrStack (return address stack) +def FeatureRAS : SubtargetFeature<"ras", "HasRAS", "true", + "Enable Reliability, Availability " + "and Serviceability extensions">; + +// Fast computation of non-negative address offsets +def FeatureFPAO : SubtargetFeature<"fpao", "HasFPAO", "true", + "Enable fast computation of " + "positive address offsets">; + +// Fast execution of AES crypto operations +def FeatureFuseAES : SubtargetFeature<"fuse-aes", "HasFuseAES", "true", + "CPU fuses AES crypto operations">; + +// Fast execution of bottom and top halves of literal generation +def FeatureFuseLiterals : SubtargetFeature<"fuse-literals", "HasFuseLiterals", "true", + "CPU fuses literal generation operations">; + +// The way of reading thread pointer +def FeatureReadTp : SubtargetFeature<"read-tp-hard", "ReadTPHard", "true", + "Reading thread pointer from register">; + +// Cyclone can zero VFP registers in 0 cycles. +def FeatureZCZeroing : SubtargetFeature<"zcz", "HasZeroCycleZeroing", "true", + "Has zero-cycle zeroing instructions">; + +// Whether it is profitable to unpredicate certain instructions during if-conversion +def FeatureProfUnpredicate : SubtargetFeature<"prof-unpr", + "IsProfitableToUnpredicate", "true", + "Is profitable to unpredicate">; + +// Some targets (e.g. Swift) have microcoded VGETLNi32. +def FeatureSlowVGETLNi32 : SubtargetFeature<"slow-vgetlni32", + "HasSlowVGETLNi32", "true", + "Has slow VGETLNi32 - prefer VMOV">; + +// Some targets (e.g. Swift) have microcoded VDUP32. +def FeatureSlowVDUP32 : SubtargetFeature<"slow-vdup32", "HasSlowVDUP32", + "true", + "Has slow VDUP32 - prefer VMOV">; + +// Some targets (e.g. Cortex-A9) prefer VMOVSR to VMOVDRR even when using NEON +// for scalar FP, as this allows more effective execution domain optimization. +def FeaturePreferVMOVSR : SubtargetFeature<"prefer-vmovsr", "PreferVMOVSR", + "true", "Prefer VMOVSR">; + +// Swift has ISHST barriers compatible with Atomic Release semantics but weaker +// than ISH +def FeaturePrefISHSTBarrier : SubtargetFeature<"prefer-ishst", "PreferISHST", + "true", "Prefer ISHST barriers">; + +// Some targets (e.g. Cortex-A9) have muxed AGU and NEON/FPU. +def FeatureMuxedUnits : SubtargetFeature<"muxed-units", "HasMuxedUnits", + "true", + "Has muxed AGU and NEON/FPU">; + +// Whether VLDM/VSTM starting with odd register number need more microops +// than single VLDRS +def FeatureSlowOddRegister : SubtargetFeature<"slow-odd-reg", "SlowOddRegister", + "true", "VLDM/VSTM starting " + "with an odd register is slow">; + +// Some targets have a renaming dependency when loading into D subregisters. +def FeatureSlowLoadDSubreg : SubtargetFeature<"slow-load-D-subreg", + "SlowLoadDSubregister", "true", + "Loading into D subregs is slow">; + +// Some targets (e.g. Cortex-A15) never want VMOVS to be widened to VMOVD. +def FeatureDontWidenVMOVS : SubtargetFeature<"dont-widen-vmovs", + "DontWidenVMOVS", "true", + "Don't widen VMOVS to VMOVD">; + +// Some targets (e.g. Cortex-A15) prefer to avoid mixing operations on different +// VFP register widths. +def FeatureSplatVFPToNeon : SubtargetFeature<"splat-vfp-neon", + "SplatVFPToNeon", "true", + "Splat register from VFP to NEON", + [FeatureDontWidenVMOVS]>; + +// Whether or not it is profitable to expand VFP/NEON MLA/MLS instructions. +def FeatureExpandMLx : SubtargetFeature<"expand-fp-mlx", + "ExpandMLx", "true", + "Expand VFP/NEON MLA/MLS instructions">; + +// Some targets have special RAW hazards for VFP/NEON VMLA/VMLS. +def FeatureHasVMLxHazards : SubtargetFeature<"vmlx-hazards", "HasVMLxHazards", + "true", "Has VMLx hazards">; + +// Some targets (e.g. Cortex-A9) want to convert VMOVRS, VMOVSR and VMOVS from +// VFP to NEON, as an execution domain optimization. +def FeatureNEONForFPMovs : SubtargetFeature<"neon-fpmovs", + "UseNEONForFPMovs", "true", + "Convert VMOVSR, VMOVRS, " + "VMOVS to NEON">; + +// Some processors benefit from using NEON instructions for scalar +// single-precision FP operations. This affects instruction selection and should +// only be enabled if the handling of denormals is not important. +def FeatureNEONForFP : SubtargetFeature<"neonfp", + "UseNEONForSinglePrecisionFP", + "true", + "Use NEON for single precision FP">; + +// On some processors, VLDn instructions that access unaligned data take one +// extra cycle. Take that into account when computing operand latencies. +def FeatureCheckVLDnAlign : SubtargetFeature<"vldn-align", "CheckVLDnAlign", + "true", + "Check for VLDn unaligned access">; + +// Some processors have a nonpipelined VFP coprocessor. +def FeatureNonpipelinedVFP : SubtargetFeature<"nonpipelined-vfp", + "NonpipelinedVFP", "true", + "VFP instructions are not pipelined">; + +// Some processors have FP multiply-accumulate instructions that don't +// play nicely with other VFP / NEON instructions, and it's generally better +// to just not use them. +def FeatureHasSlowFPVMLx : SubtargetFeature<"slowfpvmlx", "SlowFPVMLx", "true", + "Disable VFP / NEON MAC instructions">; + +// Cortex-A8 / A9 Advanced SIMD has multiplier accumulator forwarding. +def FeatureVMLxForwarding : SubtargetFeature<"vmlx-forwarding", + "HasVMLxForwarding", "true", + "Has multiplier accumulator forwarding">; + +// Disable 32-bit to 16-bit narrowing for experimentation. +def FeaturePref32BitThumb : SubtargetFeature<"32bit", "Pref32BitThumb", "true", + "Prefer 32-bit Thumb instrs">; + +/// Some instructions update CPSR partially, which can add false dependency for +/// out-of-order implementation, e.g. Cortex-A9, unless each individual bit is +/// mapped to a separate physical register. Avoid partial CPSR update for these +/// processors. +def FeatureAvoidPartialCPSR : SubtargetFeature<"avoid-partial-cpsr", + "AvoidCPSRPartialUpdate", "true", + "Avoid CPSR partial update for OOO execution">; + +/// Disable +1 predication cost for instructions updating CPSR. +/// Enabled for Cortex-A57. +def FeatureCheapPredicableCPSR : SubtargetFeature<"cheap-predicable-cpsr", + "CheapPredicableCPSRDef", + "true", + "Disable +1 predication cost for instructions updating CPSR">; + +def FeatureAvoidMOVsShOp : SubtargetFeature<"avoid-movs-shop", + "AvoidMOVsShifterOperand", "true", + "Avoid movs instructions with " + "shifter operand">; + +// Some processors perform return stack prediction. CodeGen should avoid issue +// "normal" call instructions to callees which do not return. +def FeatureHasRetAddrStack : SubtargetFeature<"ret-addr-stack", + "HasRetAddrStack", "true", + "Has return address stack">; + +// Some processors have no branch predictor, which changes the expected cost of +// taking a branch which affects the choice of whether to use predicated +// instructions. +def FeatureHasNoBranchPredictor : SubtargetFeature<"no-branch-predictor", + "HasBranchPredictor", "false", + "Has no branch predictor">; + +/// DSP extension. +def FeatureDSP : SubtargetFeature<"dsp", "HasDSP", "true", + "Supports DSP instructions in " + "ARM and/or Thumb2">; + +// Multiprocessing extension. +def FeatureMP : SubtargetFeature<"mp", "HasMPExtension", "true", + "Supports Multiprocessing extension">; + +// Virtualization extension - requires HW divide (ARMv7-AR ARMARM - 4.4.8). +def FeatureVirtualization : SubtargetFeature<"virtualization", + "HasVirtualization", "true", + "Supports Virtualization extension", + [FeatureHWDivThumb, FeatureHWDivARM]>; + +// Special TRAP encoding for NaCl, which looks like a TRAP in Thumb too. +// See ARMInstrInfo.td for details. +def FeatureNaClTrap : SubtargetFeature<"nacl-trap", "UseNaClTrap", "true", + "NaCl trap">; + +def FeatureStrictAlign : SubtargetFeature<"strict-align", + "StrictAlign", "true", + "Disallow all unaligned memory " + "access">; + +def FeatureLongCalls : SubtargetFeature<"long-calls", "GenLongCalls", "true", + "Generate calls via indirect call " + "instructions">; + +def FeatureExecuteOnly : SubtargetFeature<"execute-only", + "GenExecuteOnly", "true", + "Enable the generation of " + "execute only code.">; + +def FeatureReserveR9 : SubtargetFeature<"reserve-r9", "ReserveR9", "true", + "Reserve R9, making it unavailable" + " as GPR">; + +def FeatureNoMovt : SubtargetFeature<"no-movt", "NoMovt", "true", + "Don't use movt/movw pairs for " + "32-bit imms">; + +def FeatureNoNegativeImmediates + : SubtargetFeature<"no-neg-immediates", + "NegativeImmediates", "false", + "Convert immediates and instructions " + "to their negated or complemented " + "equivalent when the immediate does " + "not fit in the encoding.">; + +// Use the MachineScheduler for instruction scheduling for the subtarget. +def FeatureUseMISched: SubtargetFeature<"use-misched", "UseMISched", "true", + "Use the MachineScheduler">; + +def FeatureNoPostRASched : SubtargetFeature<"disable-postra-scheduler", + "DisablePostRAScheduler", "true", + "Don't schedule again after register allocation">; + +// Enable use of alias analysis during code generation +def FeatureUseAA : SubtargetFeature<"use-aa", "UseAA", "true", + "Use alias analysis during codegen">; + +//===----------------------------------------------------------------------===// +// ARM architecture class +// + +// A-series ISA +def FeatureAClass : SubtargetFeature<"aclass", "ARMProcClass", "AClass", + "Is application profile ('A' series)">; + +// R-series ISA +def FeatureRClass : SubtargetFeature<"rclass", "ARMProcClass", "RClass", + "Is realtime profile ('R' series)">; + +// M-series ISA +def FeatureMClass : SubtargetFeature<"mclass", "ARMProcClass", "MClass", + "Is microcontroller profile ('M' series)">; + + +def FeatureThumb2 : SubtargetFeature<"thumb2", "HasThumb2", "true", + "Enable Thumb2 instructions">; + +def FeatureNoARM : SubtargetFeature<"noarm", "NoARM", "true", + "Does not support ARM mode execution">; + +//===----------------------------------------------------------------------===// +// ARM ISAa. +// + +def HasV4TOps : SubtargetFeature<"v4t", "HasV4TOps", "true", + "Support ARM v4T instructions">; + +def HasV5TOps : SubtargetFeature<"v5t", "HasV5TOps", "true", + "Support ARM v5T instructions", + [HasV4TOps]>; + +def HasV5TEOps : SubtargetFeature<"v5te", "HasV5TEOps", "true", + "Support ARM v5TE, v5TEj, and " + "v5TExp instructions", + [HasV5TOps]>; + +def HasV6Ops : SubtargetFeature<"v6", "HasV6Ops", "true", + "Support ARM v6 instructions", + [HasV5TEOps]>; + +def HasV6MOps : SubtargetFeature<"v6m", "HasV6MOps", "true", + "Support ARM v6M instructions", + [HasV6Ops]>; + +def HasV8MBaselineOps : SubtargetFeature<"v8m", "HasV8MBaselineOps", "true", + "Support ARM v8M Baseline instructions", + [HasV6MOps]>; + +def HasV6KOps : SubtargetFeature<"v6k", "HasV6KOps", "true", + "Support ARM v6k instructions", + [HasV6Ops]>; + +def HasV6T2Ops : SubtargetFeature<"v6t2", "HasV6T2Ops", "true", + "Support ARM v6t2 instructions", + [HasV8MBaselineOps, HasV6KOps, FeatureThumb2]>; + +def HasV7Ops : SubtargetFeature<"v7", "HasV7Ops", "true", + "Support ARM v7 instructions", + [HasV6T2Ops, FeaturePerfMon, + FeatureV7Clrex]>; + +def HasV8MMainlineOps : + SubtargetFeature<"v8m.main", "HasV8MMainlineOps", "true", + "Support ARM v8M Mainline instructions", + [HasV7Ops]>; + +def HasV8Ops : SubtargetFeature<"v8", "HasV8Ops", "true", + "Support ARM v8 instructions", + [HasV7Ops, FeatureAcquireRelease]>; + +def HasV8_1aOps : SubtargetFeature<"v8.1a", "HasV8_1aOps", "true", + "Support ARM v8.1a instructions", + [HasV8Ops]>; + +def HasV8_2aOps : SubtargetFeature<"v8.2a", "HasV8_2aOps", "true", + "Support ARM v8.2a instructions", + [HasV8_1aOps]>; + +def HasV8_3aOps : SubtargetFeature<"v8.3a", "HasV8_3aOps", "true", + "Support ARM v8.3a instructions", + [HasV8_2aOps]>; + +def HasV8_4aOps : SubtargetFeature<"v8.4a", "HasV8_4aOps", "true", + "Support ARM v8.4a instructions", + [HasV8_3aOps, FeatureDotProd]>; + +//===----------------------------------------------------------------------===// +// ARM Processor subtarget features. +// + +def ProcA5 : SubtargetFeature<"a5", "ARMProcFamily", "CortexA5", + "Cortex-A5 ARM processors", []>; +def ProcA7 : SubtargetFeature<"a7", "ARMProcFamily", "CortexA7", + "Cortex-A7 ARM processors", []>; +def ProcA8 : SubtargetFeature<"a8", "ARMProcFamily", "CortexA8", + "Cortex-A8 ARM processors", []>; +def ProcA9 : SubtargetFeature<"a9", "ARMProcFamily", "CortexA9", + "Cortex-A9 ARM processors", []>; +def ProcA12 : SubtargetFeature<"a12", "ARMProcFamily", "CortexA12", + "Cortex-A12 ARM processors", []>; +def ProcA15 : SubtargetFeature<"a15", "ARMProcFamily", "CortexA15", + "Cortex-A15 ARM processors", []>; +def ProcA17 : SubtargetFeature<"a17", "ARMProcFamily", "CortexA17", + "Cortex-A17 ARM processors", []>; +def ProcA32 : SubtargetFeature<"a32", "ARMProcFamily", "CortexA32", + "Cortex-A32 ARM processors", []>; +def ProcA35 : SubtargetFeature<"a35", "ARMProcFamily", "CortexA35", + "Cortex-A35 ARM processors", []>; +def ProcA53 : SubtargetFeature<"a53", "ARMProcFamily", "CortexA53", + "Cortex-A53 ARM processors", []>; +def ProcA55 : SubtargetFeature<"a55", "ARMProcFamily", "CortexA55", + "Cortex-A55 ARM processors", []>; +def ProcA57 : SubtargetFeature<"a57", "ARMProcFamily", "CortexA57", + "Cortex-A57 ARM processors", []>; +def ProcA72 : SubtargetFeature<"a72", "ARMProcFamily", "CortexA72", + "Cortex-A72 ARM processors", []>; +def ProcA73 : SubtargetFeature<"a73", "ARMProcFamily", "CortexA73", + "Cortex-A73 ARM processors", []>; +def ProcA75 : SubtargetFeature<"a75", "ARMProcFamily", "CortexA75", + "Cortex-A75 ARM processors", []>; + +def ProcKrait : SubtargetFeature<"krait", "ARMProcFamily", "Krait", + "Qualcomm Krait processors", []>; +def ProcKryo : SubtargetFeature<"kryo", "ARMProcFamily", "Kryo", + "Qualcomm Kryo processors", []>; +def ProcSwift : SubtargetFeature<"swift", "ARMProcFamily", "Swift", + "Swift ARM processors", []>; + +def ProcExynosM1 : SubtargetFeature<"exynosm1", "ARMProcFamily", "ExynosM1", + "Samsung Exynos-Mx processors", []>; + +def ProcR4 : SubtargetFeature<"r4", "ARMProcFamily", "CortexR4", + "Cortex-R4 ARM processors", []>; +def ProcR5 : SubtargetFeature<"r5", "ARMProcFamily", "CortexR5", + "Cortex-R5 ARM processors", []>; +def ProcR7 : SubtargetFeature<"r7", "ARMProcFamily", "CortexR7", + "Cortex-R7 ARM processors", []>; +def ProcR52 : SubtargetFeature<"r52", "ARMProcFamily", "CortexR52", + "Cortex-R52 ARM processors", []>; + +def ProcM3 : SubtargetFeature<"m3", "ARMProcFamily", "CortexM3", + "Cortex-M3 ARM processors", []>; + + +//===----------------------------------------------------------------------===// +// ARM Helper classes. +// + +class Architecture<string fname, string aname, list<SubtargetFeature> features> + : SubtargetFeature<fname, "ARMArch", aname, + !strconcat(aname, " architecture"), features>; + +class ProcNoItin<string Name, list<SubtargetFeature> Features> + : Processor<Name, NoItineraries, Features>; + + +//===----------------------------------------------------------------------===// +// ARM architectures +// + +def ARMv2 : Architecture<"armv2", "ARMv2", []>; + +def ARMv2a : Architecture<"armv2a", "ARMv2a", []>; + +def ARMv3 : Architecture<"armv3", "ARMv3", []>; + +def ARMv3m : Architecture<"armv3m", "ARMv3m", []>; + +def ARMv4 : Architecture<"armv4", "ARMv4", []>; + +def ARMv4t : Architecture<"armv4t", "ARMv4t", [HasV4TOps]>; + +def ARMv5t : Architecture<"armv5t", "ARMv5t", [HasV5TOps]>; + +def ARMv5te : Architecture<"armv5te", "ARMv5te", [HasV5TEOps]>; + +def ARMv5tej : Architecture<"armv5tej", "ARMv5tej", [HasV5TEOps]>; + +def ARMv6 : Architecture<"armv6", "ARMv6", [HasV6Ops, + FeatureDSP]>; + +def ARMv6t2 : Architecture<"armv6t2", "ARMv6t2", [HasV6T2Ops, + FeatureDSP]>; + +def ARMv6k : Architecture<"armv6k", "ARMv6k", [HasV6KOps]>; + +def ARMv6kz : Architecture<"armv6kz", "ARMv6kz", [HasV6KOps, + FeatureTrustZone]>; + +def ARMv6m : Architecture<"armv6-m", "ARMv6m", [HasV6MOps, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureMClass, + FeatureStrictAlign]>; + +def ARMv6sm : Architecture<"armv6s-m", "ARMv6sm", [HasV6MOps, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureMClass, + FeatureStrictAlign]>; + +def ARMv7a : Architecture<"armv7-a", "ARMv7a", [HasV7Ops, + FeatureNEON, + FeatureDB, + FeatureDSP, + FeatureAClass]>; + +def ARMv7ve : Architecture<"armv7ve", "ARMv7ve", [HasV7Ops, + FeatureNEON, + FeatureDB, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureAClass]>; + +def ARMv7r : Architecture<"armv7-r", "ARMv7r", [HasV7Ops, + FeatureDB, + FeatureDSP, + FeatureHWDivThumb, + FeatureRClass]>; + +def ARMv7m : Architecture<"armv7-m", "ARMv7m", [HasV7Ops, + FeatureThumb2, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureHWDivThumb, + FeatureMClass]>; + +def ARMv7em : Architecture<"armv7e-m", "ARMv7em", [HasV7Ops, + FeatureThumb2, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureHWDivThumb, + FeatureMClass, + FeatureDSP]>; + +def ARMv8a : Architecture<"armv8-a", "ARMv8a", [HasV8Ops, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC]>; + +def ARMv81a : Architecture<"armv8.1-a", "ARMv81a", [HasV8_1aOps, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC]>; + +def ARMv82a : Architecture<"armv8.2-a", "ARMv82a", [HasV8_2aOps, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC, + FeatureRAS]>; + +def ARMv83a : Architecture<"armv8.3-a", "ARMv83a", [HasV8_3aOps, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC, + FeatureRAS]>; + +def ARMv84a : Architecture<"armv8.4-a", "ARMv84a", [HasV8_4aOps, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC, + FeatureRAS, + FeatureDotProd]>; + +def ARMv8r : Architecture<"armv8-r", "ARMv8r", [HasV8Ops, + FeatureRClass, + FeatureDB, + FeatureDFB, + FeatureDSP, + FeatureCRC, + FeatureMP, + FeatureVirtualization, + FeatureFPARMv8, + FeatureNEON]>; + +def ARMv8mBaseline : Architecture<"armv8-m.base", "ARMv8mBaseline", + [HasV8MBaselineOps, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureHWDivThumb, + FeatureV7Clrex, + Feature8MSecExt, + FeatureAcquireRelease, + FeatureMClass, + FeatureStrictAlign]>; + +def ARMv8mMainline : Architecture<"armv8-m.main", "ARMv8mMainline", + [HasV8MMainlineOps, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureHWDivThumb, + Feature8MSecExt, + FeatureAcquireRelease, + FeatureMClass]>; + +// Aliases +def IWMMXT : Architecture<"iwmmxt", "ARMv5te", [ARMv5te]>; +def IWMMXT2 : Architecture<"iwmmxt2", "ARMv5te", [ARMv5te]>; +def XScale : Architecture<"xscale", "ARMv5te", [ARMv5te]>; +def ARMv6j : Architecture<"armv6j", "ARMv7a", [ARMv6]>; +def ARMv7k : Architecture<"armv7k", "ARMv7a", [ARMv7a]>; +def ARMv7s : Architecture<"armv7s", "ARMv7a", [ARMv7a]>; + + +//===----------------------------------------------------------------------===// +// ARM schedules. +//===----------------------------------------------------------------------===// +// +include "ARMSchedule.td" + +//===----------------------------------------------------------------------===// +// ARM processors +// + +// Dummy CPU, used to target architectures +def : ProcessorModel<"generic", CortexA8Model, []>; + +// FIXME: Several processors below are not using their own scheduler +// model, but one of similar/previous processor. These should be fixed. + +def : ProcNoItin<"arm8", [ARMv4]>; +def : ProcNoItin<"arm810", [ARMv4]>; +def : ProcNoItin<"strongarm", [ARMv4]>; +def : ProcNoItin<"strongarm110", [ARMv4]>; +def : ProcNoItin<"strongarm1100", [ARMv4]>; +def : ProcNoItin<"strongarm1110", [ARMv4]>; + +def : ProcNoItin<"arm7tdmi", [ARMv4t]>; +def : ProcNoItin<"arm7tdmi-s", [ARMv4t]>; +def : ProcNoItin<"arm710t", [ARMv4t]>; +def : ProcNoItin<"arm720t", [ARMv4t]>; +def : ProcNoItin<"arm9", [ARMv4t]>; +def : ProcNoItin<"arm9tdmi", [ARMv4t]>; +def : ProcNoItin<"arm920", [ARMv4t]>; +def : ProcNoItin<"arm920t", [ARMv4t]>; +def : ProcNoItin<"arm922t", [ARMv4t]>; +def : ProcNoItin<"arm940t", [ARMv4t]>; +def : ProcNoItin<"ep9312", [ARMv4t]>; + +def : ProcNoItin<"arm10tdmi", [ARMv5t]>; +def : ProcNoItin<"arm1020t", [ARMv5t]>; + +def : ProcNoItin<"arm9e", [ARMv5te]>; +def : ProcNoItin<"arm926ej-s", [ARMv5te]>; +def : ProcNoItin<"arm946e-s", [ARMv5te]>; +def : ProcNoItin<"arm966e-s", [ARMv5te]>; +def : ProcNoItin<"arm968e-s", [ARMv5te]>; +def : ProcNoItin<"arm10e", [ARMv5te]>; +def : ProcNoItin<"arm1020e", [ARMv5te]>; +def : ProcNoItin<"arm1022e", [ARMv5te]>; +def : ProcNoItin<"xscale", [ARMv5te]>; +def : ProcNoItin<"iwmmxt", [ARMv5te]>; + +def : Processor<"arm1136j-s", ARMV6Itineraries, [ARMv6]>; +def : Processor<"arm1136jf-s", ARMV6Itineraries, [ARMv6, + FeatureVFP2, + FeatureHasSlowFPVMLx]>; + +def : Processor<"cortex-m0", ARMV6Itineraries, [ARMv6m]>; +def : Processor<"cortex-m0plus", ARMV6Itineraries, [ARMv6m]>; +def : Processor<"cortex-m1", ARMV6Itineraries, [ARMv6m]>; +def : Processor<"sc000", ARMV6Itineraries, [ARMv6m]>; + +def : Processor<"arm1176j-s", ARMV6Itineraries, [ARMv6kz]>; +def : Processor<"arm1176jz-s", ARMV6Itineraries, [ARMv6kz]>; +def : Processor<"arm1176jzf-s", ARMV6Itineraries, [ARMv6kz, + FeatureVFP2, + FeatureHasSlowFPVMLx]>; + +def : Processor<"mpcorenovfp", ARMV6Itineraries, [ARMv6k]>; +def : Processor<"mpcore", ARMV6Itineraries, [ARMv6k, + FeatureVFP2, + FeatureHasSlowFPVMLx]>; + +def : Processor<"arm1156t2-s", ARMV6Itineraries, [ARMv6t2]>; +def : Processor<"arm1156t2f-s", ARMV6Itineraries, [ARMv6t2, + FeatureVFP2, + FeatureHasSlowFPVMLx]>; + +def : ProcessorModel<"cortex-a5", CortexA8Model, [ARMv7a, ProcA5, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureSlowFPBrcc, + FeatureHasSlowFPVMLx, + FeatureVMLxForwarding, + FeatureMP, + FeatureVFP4]>; + +def : ProcessorModel<"cortex-a7", CortexA8Model, [ARMv7a, ProcA7, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureSlowFPBrcc, + FeatureHasVMLxHazards, + FeatureHasSlowFPVMLx, + FeatureVMLxForwarding, + FeatureMP, + FeatureVFP4, + FeatureVirtualization]>; + +def : ProcessorModel<"cortex-a8", CortexA8Model, [ARMv7a, ProcA8, + FeatureHasRetAddrStack, + FeatureNonpipelinedVFP, + FeatureTrustZone, + FeatureSlowFPBrcc, + FeatureHasVMLxHazards, + FeatureHasSlowFPVMLx, + FeatureVMLxForwarding]>; + +def : ProcessorModel<"cortex-a9", CortexA9Model, [ARMv7a, ProcA9, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureHasVMLxHazards, + FeatureVMLxForwarding, + FeatureFP16, + FeatureAvoidPartialCPSR, + FeatureExpandMLx, + FeaturePreferVMOVSR, + FeatureMuxedUnits, + FeatureNEONForFPMovs, + FeatureCheckVLDnAlign, + FeatureMP]>; + +def : ProcessorModel<"cortex-a12", CortexA9Model, [ARMv7a, ProcA12, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureVMLxForwarding, + FeatureVFP4, + FeatureAvoidPartialCPSR, + FeatureVirtualization, + FeatureMP]>; + +def : ProcessorModel<"cortex-a15", CortexA9Model, [ARMv7a, ProcA15, + FeatureDontWidenVMOVS, + FeatureSplatVFPToNeon, + FeatureHasRetAddrStack, + FeatureMuxedUnits, + FeatureTrustZone, + FeatureVFP4, + FeatureMP, + FeatureCheckVLDnAlign, + FeatureAvoidPartialCPSR, + FeatureVirtualization]>; + +def : ProcessorModel<"cortex-a17", CortexA9Model, [ARMv7a, ProcA17, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureMP, + FeatureVMLxForwarding, + FeatureVFP4, + FeatureAvoidPartialCPSR, + FeatureVirtualization]>; + +// FIXME: krait has currently the same features as A9 plus VFP4 and HWDiv +def : ProcessorModel<"krait", CortexA9Model, [ARMv7a, ProcKrait, + FeatureHasRetAddrStack, + FeatureMuxedUnits, + FeatureCheckVLDnAlign, + FeatureVMLxForwarding, + FeatureFP16, + FeatureAvoidPartialCPSR, + FeatureVFP4, + FeatureHWDivThumb, + FeatureHWDivARM]>; + +def : ProcessorModel<"swift", SwiftModel, [ARMv7a, ProcSwift, + FeatureHasRetAddrStack, + FeatureNEONForFP, + FeatureVFP4, + FeatureMP, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureAvoidPartialCPSR, + FeatureAvoidMOVsShOp, + FeatureHasSlowFPVMLx, + FeatureHasVMLxHazards, + FeatureProfUnpredicate, + FeaturePrefISHSTBarrier, + FeatureSlowOddRegister, + FeatureSlowLoadDSubreg, + FeatureSlowVGETLNi32, + FeatureSlowVDUP32, + FeatureUseMISched, + FeatureNoPostRASched]>; + +def : ProcessorModel<"cortex-r4", CortexA8Model, [ARMv7r, ProcR4, + FeatureHasRetAddrStack, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-r4f", CortexA8Model, [ARMv7r, ProcR4, + FeatureHasRetAddrStack, + FeatureSlowFPBrcc, + FeatureHasSlowFPVMLx, + FeatureVFP3, + FeatureD16, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-r5", CortexA8Model, [ARMv7r, ProcR5, + FeatureHasRetAddrStack, + FeatureVFP3, + FeatureD16, + FeatureSlowFPBrcc, + FeatureHWDivARM, + FeatureHasSlowFPVMLx, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-r7", CortexA8Model, [ARMv7r, ProcR7, + FeatureHasRetAddrStack, + FeatureVFP3, + FeatureD16, + FeatureFP16, + FeatureMP, + FeatureSlowFPBrcc, + FeatureHWDivARM, + FeatureHasSlowFPVMLx, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-r8", CortexA8Model, [ARMv7r, + FeatureHasRetAddrStack, + FeatureVFP3, + FeatureD16, + FeatureFP16, + FeatureMP, + FeatureSlowFPBrcc, + FeatureHWDivARM, + FeatureHasSlowFPVMLx, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-m3", CortexM3Model, [ARMv7m, + ProcM3, + FeatureHasNoBranchPredictor]>; + +def : ProcessorModel<"sc300", CortexM3Model, [ARMv7m, + ProcM3, + FeatureHasNoBranchPredictor]>; + +def : ProcessorModel<"cortex-m4", CortexM3Model, [ARMv7em, + FeatureVFP4, + FeatureVFPOnlySP, + FeatureD16, + FeatureHasNoBranchPredictor]>; + +def : ProcNoItin<"cortex-m7", [ARMv7em, + FeatureFPARMv8, + FeatureD16]>; + +def : ProcNoItin<"cortex-m23", [ARMv8mBaseline, + FeatureNoMovt]>; + +def : ProcessorModel<"cortex-m33", CortexM3Model, [ARMv8mMainline, + FeatureDSP, + FeatureFPARMv8, + FeatureD16, + FeatureVFPOnlySP, + FeatureHasNoBranchPredictor]>; + +def : ProcNoItin<"cortex-a32", [ARMv8a, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"cortex-a35", [ARMv8a, ProcA35, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"cortex-a53", [ARMv8a, ProcA53, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC, + FeatureFPAO]>; + +def : ProcNoItin<"cortex-a55", [ARMv82a, ProcA55, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureDotProd]>; + +def : ProcessorModel<"cortex-a57", CortexA57Model, [ARMv8a, ProcA57, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC, + FeatureFPAO, + FeatureAvoidPartialCPSR, + FeatureCheapPredicableCPSR]>; + +def : ProcNoItin<"cortex-a72", [ARMv8a, ProcA72, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"cortex-a73", [ARMv8a, ProcA73, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"cortex-a75", [ARMv82a, ProcA75, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureDotProd]>; + +def : ProcessorModel<"cyclone", SwiftModel, [ARMv8a, ProcSwift, + FeatureHasRetAddrStack, + FeatureNEONForFP, + FeatureVFP4, + FeatureMP, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureAvoidPartialCPSR, + FeatureAvoidMOVsShOp, + FeatureHasSlowFPVMLx, + FeatureCrypto, + FeatureUseMISched, + FeatureZCZeroing, + FeatureNoPostRASched]>; + +def : ProcNoItin<"exynos-m1", [ARMv8a, ProcExynosM1, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"exynos-m2", [ARMv8a, ProcExynosM1, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"exynos-m3", [ARMv8a, ProcExynosM1, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"exynos-m4", [ARMv8a, ProcExynosM1, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"kryo", [ARMv8a, ProcKryo, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcessorModel<"cortex-r52", CortexR52Model, [ARMv8r, ProcR52, + FeatureUseMISched, + FeatureFPAO, + FeatureUseAA]>; + +//===----------------------------------------------------------------------===// +// Register File Description +//===----------------------------------------------------------------------===// + +include "ARMRegisterInfo-digit.td" +include "ARMRegisterBanks.td" +include "ARMCallingConv.td" + +//===----------------------------------------------------------------------===// +// Instruction Descriptions +//===----------------------------------------------------------------------===// + +include "ARMInstrInfo.td" +def ARMInstrInfo : InstrInfo; + +//===----------------------------------------------------------------------===// +// Declare the target which we are implementing +//===----------------------------------------------------------------------===// + +def ARMAsmWriter : AsmWriter { + string AsmWriterClassName = "InstPrinter"; + int PassSubtarget = 1; + int Variant = 0; + bit isMCAsmWriter = 1; +} + +def ARMAsmParser : AsmParser { + bit ReportMultipleNearMisses = 1; +} + +def ARMAsmParserVariant : AsmParserVariant { + int Variant = 0; + string Name = "ARM"; + string BreakCharacters = "."; +} + +def ARM : Target { + // Pull in Instruction Info. + let InstructionSet = ARMInstrInfo; + let AssemblyWriters = [ARMAsmWriter]; + let AssemblyParsers = [ARMAsmParser]; + let AssemblyParserVariants = [ARMAsmParserVariant]; + let AllowRegisterRenaming = 1; +} diff --git a/capstone/suite/synctools/tablegen/ARM/ARM.td b/capstone/suite/synctools/tablegen/ARM/ARM.td new file mode 100644 index 000000000..2e62a0790 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARM.td @@ -0,0 +1,1098 @@ +//===-- ARM.td - Describe the ARM Target Machine -----------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// Target-independent interfaces which we are implementing +//===----------------------------------------------------------------------===// + +include "llvm/Target/Target.td" + +//===----------------------------------------------------------------------===// +// ARM Subtarget state. +// + +def ModeThumb : SubtargetFeature<"thumb-mode", "InThumbMode", + "true", "Thumb mode">; + +def ModeSoftFloat : SubtargetFeature<"soft-float","UseSoftFloat", + "true", "Use software floating " + "point features.">; + + +//===----------------------------------------------------------------------===// +// ARM Subtarget features. +// + +// Floating Point, HW Division and Neon Support +def FeatureVFP2 : SubtargetFeature<"vfp2", "HasVFPv2", "true", + "Enable VFP2 instructions">; + +def FeatureVFP3 : SubtargetFeature<"vfp3", "HasVFPv3", "true", + "Enable VFP3 instructions", + [FeatureVFP2]>; + +def FeatureNEON : SubtargetFeature<"neon", "HasNEON", "true", + "Enable NEON instructions", + [FeatureVFP3]>; + +def FeatureFP16 : SubtargetFeature<"fp16", "HasFP16", "true", + "Enable half-precision " + "floating point">; + +def FeatureVFP4 : SubtargetFeature<"vfp4", "HasVFPv4", "true", + "Enable VFP4 instructions", + [FeatureVFP3, FeatureFP16]>; + +def FeatureFPARMv8 : SubtargetFeature<"fp-armv8", "HasFPARMv8", + "true", "Enable ARMv8 FP", + [FeatureVFP4]>; + +def FeatureFullFP16 : SubtargetFeature<"fullfp16", "HasFullFP16", "true", + "Enable full half-precision " + "floating point", + [FeatureFPARMv8]>; + +def FeatureVFPOnlySP : SubtargetFeature<"fp-only-sp", "FPOnlySP", "true", + "Floating point unit supports " + "single precision only">; + +def FeatureD16 : SubtargetFeature<"d16", "HasD16", "true", + "Restrict FP to 16 double registers">; + +def FeatureHWDivThumb : SubtargetFeature<"hwdiv", + "HasHardwareDivideInThumb", "true", + "Enable divide instructions in Thumb">; + +def FeatureHWDivARM : SubtargetFeature<"hwdiv-arm", + "HasHardwareDivideInARM", "true", + "Enable divide instructions in ARM mode">; + +// Atomic Support +def FeatureDB : SubtargetFeature<"db", "HasDataBarrier", "true", + "Has data barrier (dmb/dsb) instructions">; + +def FeatureV7Clrex : SubtargetFeature<"v7clrex", "HasV7Clrex", "true", + "Has v7 clrex instruction">; + +def FeatureDFB : SubtargetFeature<"dfb", "HasFullDataBarrier", "true", + "Has full data barrier (dfb) instruction">; + +def FeatureAcquireRelease : SubtargetFeature<"acquire-release", + "HasAcquireRelease", "true", + "Has v8 acquire/release (lda/ldaex " + " etc) instructions">; + + +def FeatureSlowFPBrcc : SubtargetFeature<"slow-fp-brcc", "SlowFPBrcc", "true", + "FP compare + branch is slow">; + +def FeaturePerfMon : SubtargetFeature<"perfmon", "HasPerfMon", "true", + "Enable support for Performance " + "Monitor extensions">; + + +// TrustZone Security Extensions +def FeatureTrustZone : SubtargetFeature<"trustzone", "HasTrustZone", "true", + "Enable support for TrustZone " + "security extensions">; + +def Feature8MSecExt : SubtargetFeature<"8msecext", "Has8MSecExt", "true", + "Enable support for ARMv8-M " + "Security Extensions">; + +def FeatureSHA2 : SubtargetFeature<"sha2", "HasSHA2", "true", + "Enable SHA1 and SHA256 support", [FeatureNEON]>; + +def FeatureAES : SubtargetFeature<"aes", "HasAES", "true", + "Enable AES support", [FeatureNEON]>; + +def FeatureCrypto : SubtargetFeature<"crypto", "HasCrypto", "true", + "Enable support for " + "Cryptography extensions", + [FeatureNEON, FeatureSHA2, FeatureAES]>; + +def FeatureCRC : SubtargetFeature<"crc", "HasCRC", "true", + "Enable support for CRC instructions">; + +def FeatureDotProd : SubtargetFeature<"dotprod", "HasDotProd", "true", + "Enable support for dot product instructions", + [FeatureNEON]>; + +// Not to be confused with FeatureHasRetAddrStack (return address stack) +def FeatureRAS : SubtargetFeature<"ras", "HasRAS", "true", + "Enable Reliability, Availability " + "and Serviceability extensions">; + +// Fast computation of non-negative address offsets +def FeatureFPAO : SubtargetFeature<"fpao", "HasFPAO", "true", + "Enable fast computation of " + "positive address offsets">; + +// Fast execution of AES crypto operations +def FeatureFuseAES : SubtargetFeature<"fuse-aes", "HasFuseAES", "true", + "CPU fuses AES crypto operations">; + +// Fast execution of bottom and top halves of literal generation +def FeatureFuseLiterals : SubtargetFeature<"fuse-literals", "HasFuseLiterals", "true", + "CPU fuses literal generation operations">; + +// The way of reading thread pointer +def FeatureReadTp : SubtargetFeature<"read-tp-hard", "ReadTPHard", "true", + "Reading thread pointer from register">; + +// Cyclone can zero VFP registers in 0 cycles. +def FeatureZCZeroing : SubtargetFeature<"zcz", "HasZeroCycleZeroing", "true", + "Has zero-cycle zeroing instructions">; + +// Whether it is profitable to unpredicate certain instructions during if-conversion +def FeatureProfUnpredicate : SubtargetFeature<"prof-unpr", + "IsProfitableToUnpredicate", "true", + "Is profitable to unpredicate">; + +// Some targets (e.g. Swift) have microcoded VGETLNi32. +def FeatureSlowVGETLNi32 : SubtargetFeature<"slow-vgetlni32", + "HasSlowVGETLNi32", "true", + "Has slow VGETLNi32 - prefer VMOV">; + +// Some targets (e.g. Swift) have microcoded VDUP32. +def FeatureSlowVDUP32 : SubtargetFeature<"slow-vdup32", "HasSlowVDUP32", + "true", + "Has slow VDUP32 - prefer VMOV">; + +// Some targets (e.g. Cortex-A9) prefer VMOVSR to VMOVDRR even when using NEON +// for scalar FP, as this allows more effective execution domain optimization. +def FeaturePreferVMOVSR : SubtargetFeature<"prefer-vmovsr", "PreferVMOVSR", + "true", "Prefer VMOVSR">; + +// Swift has ISHST barriers compatible with Atomic Release semantics but weaker +// than ISH +def FeaturePrefISHSTBarrier : SubtargetFeature<"prefer-ishst", "PreferISHST", + "true", "Prefer ISHST barriers">; + +// Some targets (e.g. Cortex-A9) have muxed AGU and NEON/FPU. +def FeatureMuxedUnits : SubtargetFeature<"muxed-units", "HasMuxedUnits", + "true", + "Has muxed AGU and NEON/FPU">; + +// Whether VLDM/VSTM starting with odd register number need more microops +// than single VLDRS +def FeatureSlowOddRegister : SubtargetFeature<"slow-odd-reg", "SlowOddRegister", + "true", "VLDM/VSTM starting " + "with an odd register is slow">; + +// Some targets have a renaming dependency when loading into D subregisters. +def FeatureSlowLoadDSubreg : SubtargetFeature<"slow-load-D-subreg", + "SlowLoadDSubregister", "true", + "Loading into D subregs is slow">; + +// Some targets (e.g. Cortex-A15) never want VMOVS to be widened to VMOVD. +def FeatureDontWidenVMOVS : SubtargetFeature<"dont-widen-vmovs", + "DontWidenVMOVS", "true", + "Don't widen VMOVS to VMOVD">; + +// Some targets (e.g. Cortex-A15) prefer to avoid mixing operations on different +// VFP register widths. +def FeatureSplatVFPToNeon : SubtargetFeature<"splat-vfp-neon", + "SplatVFPToNeon", "true", + "Splat register from VFP to NEON", + [FeatureDontWidenVMOVS]>; + +// Whether or not it is profitable to expand VFP/NEON MLA/MLS instructions. +def FeatureExpandMLx : SubtargetFeature<"expand-fp-mlx", + "ExpandMLx", "true", + "Expand VFP/NEON MLA/MLS instructions">; + +// Some targets have special RAW hazards for VFP/NEON VMLA/VMLS. +def FeatureHasVMLxHazards : SubtargetFeature<"vmlx-hazards", "HasVMLxHazards", + "true", "Has VMLx hazards">; + +// Some targets (e.g. Cortex-A9) want to convert VMOVRS, VMOVSR and VMOVS from +// VFP to NEON, as an execution domain optimization. +def FeatureNEONForFPMovs : SubtargetFeature<"neon-fpmovs", + "UseNEONForFPMovs", "true", + "Convert VMOVSR, VMOVRS, " + "VMOVS to NEON">; + +// Some processors benefit from using NEON instructions for scalar +// single-precision FP operations. This affects instruction selection and should +// only be enabled if the handling of denormals is not important. +def FeatureNEONForFP : SubtargetFeature<"neonfp", + "UseNEONForSinglePrecisionFP", + "true", + "Use NEON for single precision FP">; + +// On some processors, VLDn instructions that access unaligned data take one +// extra cycle. Take that into account when computing operand latencies. +def FeatureCheckVLDnAlign : SubtargetFeature<"vldn-align", "CheckVLDnAlign", + "true", + "Check for VLDn unaligned access">; + +// Some processors have a nonpipelined VFP coprocessor. +def FeatureNonpipelinedVFP : SubtargetFeature<"nonpipelined-vfp", + "NonpipelinedVFP", "true", + "VFP instructions are not pipelined">; + +// Some processors have FP multiply-accumulate instructions that don't +// play nicely with other VFP / NEON instructions, and it's generally better +// to just not use them. +def FeatureHasSlowFPVMLx : SubtargetFeature<"slowfpvmlx", "SlowFPVMLx", "true", + "Disable VFP / NEON MAC instructions">; + +// Cortex-A8 / A9 Advanced SIMD has multiplier accumulator forwarding. +def FeatureVMLxForwarding : SubtargetFeature<"vmlx-forwarding", + "HasVMLxForwarding", "true", + "Has multiplier accumulator forwarding">; + +// Disable 32-bit to 16-bit narrowing for experimentation. +def FeaturePref32BitThumb : SubtargetFeature<"32bit", "Pref32BitThumb", "true", + "Prefer 32-bit Thumb instrs">; + +/// Some instructions update CPSR partially, which can add false dependency for +/// out-of-order implementation, e.g. Cortex-A9, unless each individual bit is +/// mapped to a separate physical register. Avoid partial CPSR update for these +/// processors. +def FeatureAvoidPartialCPSR : SubtargetFeature<"avoid-partial-cpsr", + "AvoidCPSRPartialUpdate", "true", + "Avoid CPSR partial update for OOO execution">; + +/// Disable +1 predication cost for instructions updating CPSR. +/// Enabled for Cortex-A57. +def FeatureCheapPredicableCPSR : SubtargetFeature<"cheap-predicable-cpsr", + "CheapPredicableCPSRDef", + "true", + "Disable +1 predication cost for instructions updating CPSR">; + +def FeatureAvoidMOVsShOp : SubtargetFeature<"avoid-movs-shop", + "AvoidMOVsShifterOperand", "true", + "Avoid movs instructions with " + "shifter operand">; + +// Some processors perform return stack prediction. CodeGen should avoid issue +// "normal" call instructions to callees which do not return. +def FeatureHasRetAddrStack : SubtargetFeature<"ret-addr-stack", + "HasRetAddrStack", "true", + "Has return address stack">; + +// Some processors have no branch predictor, which changes the expected cost of +// taking a branch which affects the choice of whether to use predicated +// instructions. +def FeatureHasNoBranchPredictor : SubtargetFeature<"no-branch-predictor", + "HasBranchPredictor", "false", + "Has no branch predictor">; + +/// DSP extension. +def FeatureDSP : SubtargetFeature<"dsp", "HasDSP", "true", + "Supports DSP instructions in " + "ARM and/or Thumb2">; + +// Multiprocessing extension. +def FeatureMP : SubtargetFeature<"mp", "HasMPExtension", "true", + "Supports Multiprocessing extension">; + +// Virtualization extension - requires HW divide (ARMv7-AR ARMARM - 4.4.8). +def FeatureVirtualization : SubtargetFeature<"virtualization", + "HasVirtualization", "true", + "Supports Virtualization extension", + [FeatureHWDivThumb, FeatureHWDivARM]>; + +// Special TRAP encoding for NaCl, which looks like a TRAP in Thumb too. +// See ARMInstrInfo.td for details. +def FeatureNaClTrap : SubtargetFeature<"nacl-trap", "UseNaClTrap", "true", + "NaCl trap">; + +def FeatureStrictAlign : SubtargetFeature<"strict-align", + "StrictAlign", "true", + "Disallow all unaligned memory " + "access">; + +def FeatureLongCalls : SubtargetFeature<"long-calls", "GenLongCalls", "true", + "Generate calls via indirect call " + "instructions">; + +def FeatureExecuteOnly : SubtargetFeature<"execute-only", + "GenExecuteOnly", "true", + "Enable the generation of " + "execute only code.">; + +def FeatureReserveR9 : SubtargetFeature<"reserve-r9", "ReserveR9", "true", + "Reserve R9, making it unavailable" + " as GPR">; + +def FeatureNoMovt : SubtargetFeature<"no-movt", "NoMovt", "true", + "Don't use movt/movw pairs for " + "32-bit imms">; + +def FeatureNoNegativeImmediates + : SubtargetFeature<"no-neg-immediates", + "NegativeImmediates", "false", + "Convert immediates and instructions " + "to their negated or complemented " + "equivalent when the immediate does " + "not fit in the encoding.">; + +// Use the MachineScheduler for instruction scheduling for the subtarget. +def FeatureUseMISched: SubtargetFeature<"use-misched", "UseMISched", "true", + "Use the MachineScheduler">; + +def FeatureNoPostRASched : SubtargetFeature<"disable-postra-scheduler", + "DisablePostRAScheduler", "true", + "Don't schedule again after register allocation">; + +// Enable use of alias analysis during code generation +def FeatureUseAA : SubtargetFeature<"use-aa", "UseAA", "true", + "Use alias analysis during codegen">; + +//===----------------------------------------------------------------------===// +// ARM architecture class +// + +// A-series ISA +def FeatureAClass : SubtargetFeature<"aclass", "ARMProcClass", "AClass", + "Is application profile ('A' series)">; + +// R-series ISA +def FeatureRClass : SubtargetFeature<"rclass", "ARMProcClass", "RClass", + "Is realtime profile ('R' series)">; + +// M-series ISA +def FeatureMClass : SubtargetFeature<"mclass", "ARMProcClass", "MClass", + "Is microcontroller profile ('M' series)">; + + +def FeatureThumb2 : SubtargetFeature<"thumb2", "HasThumb2", "true", + "Enable Thumb2 instructions">; + +def FeatureNoARM : SubtargetFeature<"noarm", "NoARM", "true", + "Does not support ARM mode execution">; + +//===----------------------------------------------------------------------===// +// ARM ISAa. +// + +def HasV4TOps : SubtargetFeature<"v4t", "HasV4TOps", "true", + "Support ARM v4T instructions">; + +def HasV5TOps : SubtargetFeature<"v5t", "HasV5TOps", "true", + "Support ARM v5T instructions", + [HasV4TOps]>; + +def HasV5TEOps : SubtargetFeature<"v5te", "HasV5TEOps", "true", + "Support ARM v5TE, v5TEj, and " + "v5TExp instructions", + [HasV5TOps]>; + +def HasV6Ops : SubtargetFeature<"v6", "HasV6Ops", "true", + "Support ARM v6 instructions", + [HasV5TEOps]>; + +def HasV6MOps : SubtargetFeature<"v6m", "HasV6MOps", "true", + "Support ARM v6M instructions", + [HasV6Ops]>; + +def HasV8MBaselineOps : SubtargetFeature<"v8m", "HasV8MBaselineOps", "true", + "Support ARM v8M Baseline instructions", + [HasV6MOps]>; + +def HasV6KOps : SubtargetFeature<"v6k", "HasV6KOps", "true", + "Support ARM v6k instructions", + [HasV6Ops]>; + +def HasV6T2Ops : SubtargetFeature<"v6t2", "HasV6T2Ops", "true", + "Support ARM v6t2 instructions", + [HasV8MBaselineOps, HasV6KOps, FeatureThumb2]>; + +def HasV7Ops : SubtargetFeature<"v7", "HasV7Ops", "true", + "Support ARM v7 instructions", + [HasV6T2Ops, FeaturePerfMon, + FeatureV7Clrex]>; + +def HasV8MMainlineOps : + SubtargetFeature<"v8m.main", "HasV8MMainlineOps", "true", + "Support ARM v8M Mainline instructions", + [HasV7Ops]>; + +def HasV8Ops : SubtargetFeature<"v8", "HasV8Ops", "true", + "Support ARM v8 instructions", + [HasV7Ops, FeatureAcquireRelease]>; + +def HasV8_1aOps : SubtargetFeature<"v8.1a", "HasV8_1aOps", "true", + "Support ARM v8.1a instructions", + [HasV8Ops]>; + +def HasV8_2aOps : SubtargetFeature<"v8.2a", "HasV8_2aOps", "true", + "Support ARM v8.2a instructions", + [HasV8_1aOps]>; + +def HasV8_3aOps : SubtargetFeature<"v8.3a", "HasV8_3aOps", "true", + "Support ARM v8.3a instructions", + [HasV8_2aOps]>; + +def HasV8_4aOps : SubtargetFeature<"v8.4a", "HasV8_4aOps", "true", + "Support ARM v8.4a instructions", + [HasV8_3aOps, FeatureDotProd]>; + +//===----------------------------------------------------------------------===// +// ARM Processor subtarget features. +// + +def ProcA5 : SubtargetFeature<"a5", "ARMProcFamily", "CortexA5", + "Cortex-A5 ARM processors", []>; +def ProcA7 : SubtargetFeature<"a7", "ARMProcFamily", "CortexA7", + "Cortex-A7 ARM processors", []>; +def ProcA8 : SubtargetFeature<"a8", "ARMProcFamily", "CortexA8", + "Cortex-A8 ARM processors", []>; +def ProcA9 : SubtargetFeature<"a9", "ARMProcFamily", "CortexA9", + "Cortex-A9 ARM processors", []>; +def ProcA12 : SubtargetFeature<"a12", "ARMProcFamily", "CortexA12", + "Cortex-A12 ARM processors", []>; +def ProcA15 : SubtargetFeature<"a15", "ARMProcFamily", "CortexA15", + "Cortex-A15 ARM processors", []>; +def ProcA17 : SubtargetFeature<"a17", "ARMProcFamily", "CortexA17", + "Cortex-A17 ARM processors", []>; +def ProcA32 : SubtargetFeature<"a32", "ARMProcFamily", "CortexA32", + "Cortex-A32 ARM processors", []>; +def ProcA35 : SubtargetFeature<"a35", "ARMProcFamily", "CortexA35", + "Cortex-A35 ARM processors", []>; +def ProcA53 : SubtargetFeature<"a53", "ARMProcFamily", "CortexA53", + "Cortex-A53 ARM processors", []>; +def ProcA55 : SubtargetFeature<"a55", "ARMProcFamily", "CortexA55", + "Cortex-A55 ARM processors", []>; +def ProcA57 : SubtargetFeature<"a57", "ARMProcFamily", "CortexA57", + "Cortex-A57 ARM processors", []>; +def ProcA72 : SubtargetFeature<"a72", "ARMProcFamily", "CortexA72", + "Cortex-A72 ARM processors", []>; +def ProcA73 : SubtargetFeature<"a73", "ARMProcFamily", "CortexA73", + "Cortex-A73 ARM processors", []>; +def ProcA75 : SubtargetFeature<"a75", "ARMProcFamily", "CortexA75", + "Cortex-A75 ARM processors", []>; + +def ProcKrait : SubtargetFeature<"krait", "ARMProcFamily", "Krait", + "Qualcomm Krait processors", []>; +def ProcKryo : SubtargetFeature<"kryo", "ARMProcFamily", "Kryo", + "Qualcomm Kryo processors", []>; +def ProcSwift : SubtargetFeature<"swift", "ARMProcFamily", "Swift", + "Swift ARM processors", []>; + +def ProcExynosM1 : SubtargetFeature<"exynosm1", "ARMProcFamily", "ExynosM1", + "Samsung Exynos-Mx processors", []>; + +def ProcR4 : SubtargetFeature<"r4", "ARMProcFamily", "CortexR4", + "Cortex-R4 ARM processors", []>; +def ProcR5 : SubtargetFeature<"r5", "ARMProcFamily", "CortexR5", + "Cortex-R5 ARM processors", []>; +def ProcR7 : SubtargetFeature<"r7", "ARMProcFamily", "CortexR7", + "Cortex-R7 ARM processors", []>; +def ProcR52 : SubtargetFeature<"r52", "ARMProcFamily", "CortexR52", + "Cortex-R52 ARM processors", []>; + +def ProcM3 : SubtargetFeature<"m3", "ARMProcFamily", "CortexM3", + "Cortex-M3 ARM processors", []>; + + +//===----------------------------------------------------------------------===// +// ARM Helper classes. +// + +class Architecture<string fname, string aname, list<SubtargetFeature> features> + : SubtargetFeature<fname, "ARMArch", aname, + !strconcat(aname, " architecture"), features>; + +class ProcNoItin<string Name, list<SubtargetFeature> Features> + : Processor<Name, NoItineraries, Features>; + + +//===----------------------------------------------------------------------===// +// ARM architectures +// + +def ARMv2 : Architecture<"armv2", "ARMv2", []>; + +def ARMv2a : Architecture<"armv2a", "ARMv2a", []>; + +def ARMv3 : Architecture<"armv3", "ARMv3", []>; + +def ARMv3m : Architecture<"armv3m", "ARMv3m", []>; + +def ARMv4 : Architecture<"armv4", "ARMv4", []>; + +def ARMv4t : Architecture<"armv4t", "ARMv4t", [HasV4TOps]>; + +def ARMv5t : Architecture<"armv5t", "ARMv5t", [HasV5TOps]>; + +def ARMv5te : Architecture<"armv5te", "ARMv5te", [HasV5TEOps]>; + +def ARMv5tej : Architecture<"armv5tej", "ARMv5tej", [HasV5TEOps]>; + +def ARMv6 : Architecture<"armv6", "ARMv6", [HasV6Ops, + FeatureDSP]>; + +def ARMv6t2 : Architecture<"armv6t2", "ARMv6t2", [HasV6T2Ops, + FeatureDSP]>; + +def ARMv6k : Architecture<"armv6k", "ARMv6k", [HasV6KOps]>; + +def ARMv6kz : Architecture<"armv6kz", "ARMv6kz", [HasV6KOps, + FeatureTrustZone]>; + +def ARMv6m : Architecture<"armv6-m", "ARMv6m", [HasV6MOps, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureMClass, + FeatureStrictAlign]>; + +def ARMv6sm : Architecture<"armv6s-m", "ARMv6sm", [HasV6MOps, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureMClass, + FeatureStrictAlign]>; + +def ARMv7a : Architecture<"armv7-a", "ARMv7a", [HasV7Ops, + FeatureNEON, + FeatureDB, + FeatureDSP, + FeatureAClass]>; + +def ARMv7ve : Architecture<"armv7ve", "ARMv7ve", [HasV7Ops, + FeatureNEON, + FeatureDB, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureAClass]>; + +def ARMv7r : Architecture<"armv7-r", "ARMv7r", [HasV7Ops, + FeatureDB, + FeatureDSP, + FeatureHWDivThumb, + FeatureRClass]>; + +def ARMv7m : Architecture<"armv7-m", "ARMv7m", [HasV7Ops, + FeatureThumb2, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureHWDivThumb, + FeatureMClass]>; + +def ARMv7em : Architecture<"armv7e-m", "ARMv7em", [HasV7Ops, + FeatureThumb2, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureHWDivThumb, + FeatureMClass, + FeatureDSP]>; + +def ARMv8a : Architecture<"armv8-a", "ARMv8a", [HasV8Ops, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC]>; + +def ARMv81a : Architecture<"armv8.1-a", "ARMv81a", [HasV8_1aOps, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC]>; + +def ARMv82a : Architecture<"armv8.2-a", "ARMv82a", [HasV8_2aOps, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC, + FeatureRAS]>; + +def ARMv83a : Architecture<"armv8.3-a", "ARMv83a", [HasV8_3aOps, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC, + FeatureRAS]>; + +def ARMv84a : Architecture<"armv8.4-a", "ARMv84a", [HasV8_4aOps, + FeatureAClass, + FeatureDB, + FeatureFPARMv8, + FeatureNEON, + FeatureDSP, + FeatureTrustZone, + FeatureMP, + FeatureVirtualization, + FeatureCrypto, + FeatureCRC, + FeatureRAS, + FeatureDotProd]>; + +def ARMv8r : Architecture<"armv8-r", "ARMv8r", [HasV8Ops, + FeatureRClass, + FeatureDB, + FeatureDFB, + FeatureDSP, + FeatureCRC, + FeatureMP, + FeatureVirtualization, + FeatureFPARMv8, + FeatureNEON]>; + +def ARMv8mBaseline : Architecture<"armv8-m.base", "ARMv8mBaseline", + [HasV8MBaselineOps, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureHWDivThumb, + FeatureV7Clrex, + Feature8MSecExt, + FeatureAcquireRelease, + FeatureMClass, + FeatureStrictAlign]>; + +def ARMv8mMainline : Architecture<"armv8-m.main", "ARMv8mMainline", + [HasV8MMainlineOps, + FeatureNoARM, + ModeThumb, + FeatureDB, + FeatureHWDivThumb, + Feature8MSecExt, + FeatureAcquireRelease, + FeatureMClass]>; + +// Aliases +def IWMMXT : Architecture<"iwmmxt", "ARMv5te", [ARMv5te]>; +def IWMMXT2 : Architecture<"iwmmxt2", "ARMv5te", [ARMv5te]>; +def XScale : Architecture<"xscale", "ARMv5te", [ARMv5te]>; +def ARMv6j : Architecture<"armv6j", "ARMv7a", [ARMv6]>; +def ARMv7k : Architecture<"armv7k", "ARMv7a", [ARMv7a]>; +def ARMv7s : Architecture<"armv7s", "ARMv7a", [ARMv7a]>; + + +//===----------------------------------------------------------------------===// +// ARM schedules. +//===----------------------------------------------------------------------===// +// +include "ARMSchedule.td" + +//===----------------------------------------------------------------------===// +// ARM processors +// + +// Dummy CPU, used to target architectures +def : ProcessorModel<"generic", CortexA8Model, []>; + +// FIXME: Several processors below are not using their own scheduler +// model, but one of similar/previous processor. These should be fixed. + +def : ProcNoItin<"arm8", [ARMv4]>; +def : ProcNoItin<"arm810", [ARMv4]>; +def : ProcNoItin<"strongarm", [ARMv4]>; +def : ProcNoItin<"strongarm110", [ARMv4]>; +def : ProcNoItin<"strongarm1100", [ARMv4]>; +def : ProcNoItin<"strongarm1110", [ARMv4]>; + +def : ProcNoItin<"arm7tdmi", [ARMv4t]>; +def : ProcNoItin<"arm7tdmi-s", [ARMv4t]>; +def : ProcNoItin<"arm710t", [ARMv4t]>; +def : ProcNoItin<"arm720t", [ARMv4t]>; +def : ProcNoItin<"arm9", [ARMv4t]>; +def : ProcNoItin<"arm9tdmi", [ARMv4t]>; +def : ProcNoItin<"arm920", [ARMv4t]>; +def : ProcNoItin<"arm920t", [ARMv4t]>; +def : ProcNoItin<"arm922t", [ARMv4t]>; +def : ProcNoItin<"arm940t", [ARMv4t]>; +def : ProcNoItin<"ep9312", [ARMv4t]>; + +def : ProcNoItin<"arm10tdmi", [ARMv5t]>; +def : ProcNoItin<"arm1020t", [ARMv5t]>; + +def : ProcNoItin<"arm9e", [ARMv5te]>; +def : ProcNoItin<"arm926ej-s", [ARMv5te]>; +def : ProcNoItin<"arm946e-s", [ARMv5te]>; +def : ProcNoItin<"arm966e-s", [ARMv5te]>; +def : ProcNoItin<"arm968e-s", [ARMv5te]>; +def : ProcNoItin<"arm10e", [ARMv5te]>; +def : ProcNoItin<"arm1020e", [ARMv5te]>; +def : ProcNoItin<"arm1022e", [ARMv5te]>; +def : ProcNoItin<"xscale", [ARMv5te]>; +def : ProcNoItin<"iwmmxt", [ARMv5te]>; + +def : Processor<"arm1136j-s", ARMV6Itineraries, [ARMv6]>; +def : Processor<"arm1136jf-s", ARMV6Itineraries, [ARMv6, + FeatureVFP2, + FeatureHasSlowFPVMLx]>; + +def : Processor<"cortex-m0", ARMV6Itineraries, [ARMv6m]>; +def : Processor<"cortex-m0plus", ARMV6Itineraries, [ARMv6m]>; +def : Processor<"cortex-m1", ARMV6Itineraries, [ARMv6m]>; +def : Processor<"sc000", ARMV6Itineraries, [ARMv6m]>; + +def : Processor<"arm1176j-s", ARMV6Itineraries, [ARMv6kz]>; +def : Processor<"arm1176jz-s", ARMV6Itineraries, [ARMv6kz]>; +def : Processor<"arm1176jzf-s", ARMV6Itineraries, [ARMv6kz, + FeatureVFP2, + FeatureHasSlowFPVMLx]>; + +def : Processor<"mpcorenovfp", ARMV6Itineraries, [ARMv6k]>; +def : Processor<"mpcore", ARMV6Itineraries, [ARMv6k, + FeatureVFP2, + FeatureHasSlowFPVMLx]>; + +def : Processor<"arm1156t2-s", ARMV6Itineraries, [ARMv6t2]>; +def : Processor<"arm1156t2f-s", ARMV6Itineraries, [ARMv6t2, + FeatureVFP2, + FeatureHasSlowFPVMLx]>; + +def : ProcessorModel<"cortex-a5", CortexA8Model, [ARMv7a, ProcA5, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureSlowFPBrcc, + FeatureHasSlowFPVMLx, + FeatureVMLxForwarding, + FeatureMP, + FeatureVFP4]>; + +def : ProcessorModel<"cortex-a7", CortexA8Model, [ARMv7a, ProcA7, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureSlowFPBrcc, + FeatureHasVMLxHazards, + FeatureHasSlowFPVMLx, + FeatureVMLxForwarding, + FeatureMP, + FeatureVFP4, + FeatureVirtualization]>; + +def : ProcessorModel<"cortex-a8", CortexA8Model, [ARMv7a, ProcA8, + FeatureHasRetAddrStack, + FeatureNonpipelinedVFP, + FeatureTrustZone, + FeatureSlowFPBrcc, + FeatureHasVMLxHazards, + FeatureHasSlowFPVMLx, + FeatureVMLxForwarding]>; + +def : ProcessorModel<"cortex-a9", CortexA9Model, [ARMv7a, ProcA9, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureHasVMLxHazards, + FeatureVMLxForwarding, + FeatureFP16, + FeatureAvoidPartialCPSR, + FeatureExpandMLx, + FeaturePreferVMOVSR, + FeatureMuxedUnits, + FeatureNEONForFPMovs, + FeatureCheckVLDnAlign, + FeatureMP]>; + +def : ProcessorModel<"cortex-a12", CortexA9Model, [ARMv7a, ProcA12, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureVMLxForwarding, + FeatureVFP4, + FeatureAvoidPartialCPSR, + FeatureVirtualization, + FeatureMP]>; + +def : ProcessorModel<"cortex-a15", CortexA9Model, [ARMv7a, ProcA15, + FeatureDontWidenVMOVS, + FeatureSplatVFPToNeon, + FeatureHasRetAddrStack, + FeatureMuxedUnits, + FeatureTrustZone, + FeatureVFP4, + FeatureMP, + FeatureCheckVLDnAlign, + FeatureAvoidPartialCPSR, + FeatureVirtualization]>; + +def : ProcessorModel<"cortex-a17", CortexA9Model, [ARMv7a, ProcA17, + FeatureHasRetAddrStack, + FeatureTrustZone, + FeatureMP, + FeatureVMLxForwarding, + FeatureVFP4, + FeatureAvoidPartialCPSR, + FeatureVirtualization]>; + +// FIXME: krait has currently the same features as A9 plus VFP4 and HWDiv +def : ProcessorModel<"krait", CortexA9Model, [ARMv7a, ProcKrait, + FeatureHasRetAddrStack, + FeatureMuxedUnits, + FeatureCheckVLDnAlign, + FeatureVMLxForwarding, + FeatureFP16, + FeatureAvoidPartialCPSR, + FeatureVFP4, + FeatureHWDivThumb, + FeatureHWDivARM]>; + +def : ProcessorModel<"swift", SwiftModel, [ARMv7a, ProcSwift, + FeatureHasRetAddrStack, + FeatureNEONForFP, + FeatureVFP4, + FeatureMP, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureAvoidPartialCPSR, + FeatureAvoidMOVsShOp, + FeatureHasSlowFPVMLx, + FeatureHasVMLxHazards, + FeatureProfUnpredicate, + FeaturePrefISHSTBarrier, + FeatureSlowOddRegister, + FeatureSlowLoadDSubreg, + FeatureSlowVGETLNi32, + FeatureSlowVDUP32, + FeatureUseMISched, + FeatureNoPostRASched]>; + +def : ProcessorModel<"cortex-r4", CortexA8Model, [ARMv7r, ProcR4, + FeatureHasRetAddrStack, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-r4f", CortexA8Model, [ARMv7r, ProcR4, + FeatureHasRetAddrStack, + FeatureSlowFPBrcc, + FeatureHasSlowFPVMLx, + FeatureVFP3, + FeatureD16, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-r5", CortexA8Model, [ARMv7r, ProcR5, + FeatureHasRetAddrStack, + FeatureVFP3, + FeatureD16, + FeatureSlowFPBrcc, + FeatureHWDivARM, + FeatureHasSlowFPVMLx, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-r7", CortexA8Model, [ARMv7r, ProcR7, + FeatureHasRetAddrStack, + FeatureVFP3, + FeatureD16, + FeatureFP16, + FeatureMP, + FeatureSlowFPBrcc, + FeatureHWDivARM, + FeatureHasSlowFPVMLx, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-r8", CortexA8Model, [ARMv7r, + FeatureHasRetAddrStack, + FeatureVFP3, + FeatureD16, + FeatureFP16, + FeatureMP, + FeatureSlowFPBrcc, + FeatureHWDivARM, + FeatureHasSlowFPVMLx, + FeatureAvoidPartialCPSR]>; + +def : ProcessorModel<"cortex-m3", CortexM3Model, [ARMv7m, + ProcM3, + FeatureHasNoBranchPredictor]>; + +def : ProcessorModel<"sc300", CortexM3Model, [ARMv7m, + ProcM3, + FeatureHasNoBranchPredictor]>; + +def : ProcessorModel<"cortex-m4", CortexM3Model, [ARMv7em, + FeatureVFP4, + FeatureVFPOnlySP, + FeatureD16, + FeatureHasNoBranchPredictor]>; + +def : ProcNoItin<"cortex-m7", [ARMv7em, + FeatureFPARMv8, + FeatureD16]>; + +def : ProcNoItin<"cortex-m23", [ARMv8mBaseline, + FeatureNoMovt]>; + +def : ProcessorModel<"cortex-m33", CortexM3Model, [ARMv8mMainline, + FeatureDSP, + FeatureFPARMv8, + FeatureD16, + FeatureVFPOnlySP, + FeatureHasNoBranchPredictor]>; + +def : ProcNoItin<"cortex-a32", [ARMv8a, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"cortex-a35", [ARMv8a, ProcA35, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"cortex-a53", [ARMv8a, ProcA53, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC, + FeatureFPAO]>; + +def : ProcNoItin<"cortex-a55", [ARMv82a, ProcA55, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureDotProd]>; + +def : ProcessorModel<"cortex-a57", CortexA57Model, [ARMv8a, ProcA57, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC, + FeatureFPAO, + FeatureAvoidPartialCPSR, + FeatureCheapPredicableCPSR]>; + +def : ProcNoItin<"cortex-a72", [ARMv8a, ProcA72, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"cortex-a73", [ARMv8a, ProcA73, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"cortex-a75", [ARMv82a, ProcA75, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureDotProd]>; + +def : ProcessorModel<"cyclone", SwiftModel, [ARMv8a, ProcSwift, + FeatureHasRetAddrStack, + FeatureNEONForFP, + FeatureVFP4, + FeatureMP, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureAvoidPartialCPSR, + FeatureAvoidMOVsShOp, + FeatureHasSlowFPVMLx, + FeatureCrypto, + FeatureUseMISched, + FeatureZCZeroing, + FeatureNoPostRASched]>; + +def : ProcNoItin<"exynos-m1", [ARMv8a, ProcExynosM1, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"exynos-m2", [ARMv8a, ProcExynosM1, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"exynos-m3", [ARMv8a, ProcExynosM1, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"exynos-m4", [ARMv8a, ProcExynosM1, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcNoItin<"kryo", [ARMv8a, ProcKryo, + FeatureHWDivThumb, + FeatureHWDivARM, + FeatureCrypto, + FeatureCRC]>; + +def : ProcessorModel<"cortex-r52", CortexR52Model, [ARMv8r, ProcR52, + FeatureUseMISched, + FeatureFPAO, + FeatureUseAA]>; + +//===----------------------------------------------------------------------===// +// Register File Description +//===----------------------------------------------------------------------===// + +include "ARMRegisterInfo.td" +include "ARMRegisterBanks.td" +include "ARMCallingConv.td" + +//===----------------------------------------------------------------------===// +// Instruction Descriptions +//===----------------------------------------------------------------------===// + +include "ARMInstrInfo.td" +def ARMInstrInfo : InstrInfo; + +//===----------------------------------------------------------------------===// +// Declare the target which we are implementing +//===----------------------------------------------------------------------===// + +def ARMAsmWriter : AsmWriter { + string AsmWriterClassName = "InstPrinter"; + int PassSubtarget = 1; + int Variant = 0; + bit isMCAsmWriter = 1; +} + +def ARMAsmParser : AsmParser { + bit ReportMultipleNearMisses = 1; +} + +def ARMAsmParserVariant : AsmParserVariant { + int Variant = 0; + string Name = "ARM"; + string BreakCharacters = "."; +} + +def ARM : Target { + // Pull in Instruction Info. + let InstructionSet = ARMInstrInfo; + let AssemblyWriters = [ARMAsmWriter]; + let AssemblyParsers = [ARMAsmParser]; + let AssemblyParserVariants = [ARMAsmParserVariant]; + let AllowRegisterRenaming = 1; +} diff --git a/capstone/suite/synctools/tablegen/ARM/ARMCallingConv.td b/capstone/suite/synctools/tablegen/ARM/ARMCallingConv.td new file mode 100644 index 000000000..f173e423f --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMCallingConv.td @@ -0,0 +1,318 @@ +//===-- ARMCallingConv.td - Calling Conventions for ARM ----*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// This describes the calling conventions for ARM architecture. +//===----------------------------------------------------------------------===// + +/// CCIfAlign - Match of the original alignment of the arg +class CCIfAlign<string Align, CCAction A>: + CCIf<!strconcat("ArgFlags.getOrigAlign() == ", Align), A>; + +//===----------------------------------------------------------------------===// +// ARM APCS Calling Convention +//===----------------------------------------------------------------------===// +def CC_ARM_APCS : CallingConv<[ + + // Handles byval parameters. + CCIfByVal<CCPassByVal<4, 4>>, + + CCIfType<[i1, i8, i16], CCPromoteToType<i32>>, + + // Pass SwiftSelf in a callee saved register. + CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>, + + // A SwiftError is passed in R8. + CCIfSwiftError<CCIfType<[i32], CCAssignToReg<[R8]>>>, + + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, + + // f64 and v2f64 are passed in adjacent GPRs, possibly split onto the stack + CCIfType<[f64, v2f64], CCCustom<"CC_ARM_APCS_Custom_f64">>, + + CCIfType<[f32], CCBitConvertToType<i32>>, + CCIfType<[i32], CCAssignToReg<[R0, R1, R2, R3]>>, + + CCIfType<[i32], CCAssignToStack<4, 4>>, + CCIfType<[f64], CCAssignToStack<8, 4>>, + CCIfType<[v2f64], CCAssignToStack<16, 4>> +]>; + +def RetCC_ARM_APCS : CallingConv<[ + CCIfType<[i1, i8, i16], CCPromoteToType<i32>>, + CCIfType<[f32], CCBitConvertToType<i32>>, + + // Pass SwiftSelf in a callee saved register. + CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>, + + // A SwiftError is returned in R8. + CCIfSwiftError<CCIfType<[i32], CCAssignToReg<[R8]>>>, + + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, + + CCIfType<[f64, v2f64], CCCustom<"RetCC_ARM_APCS_Custom_f64">>, + + CCIfType<[i32], CCAssignToReg<[R0, R1, R2, R3]>>, + CCIfType<[i64], CCAssignToRegWithShadow<[R0, R2], [R1, R3]>> +]>; + +//===----------------------------------------------------------------------===// +// ARM APCS Calling Convention for FastCC (when VFP2 or later is available) +//===----------------------------------------------------------------------===// +def FastCC_ARM_APCS : CallingConv<[ + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, + + CCIfType<[v2f64], CCAssignToReg<[Q0, Q1, Q2, Q3]>>, + CCIfType<[f64], CCAssignToReg<[D0, D1, D2, D3, D4, D5, D6, D7]>>, + CCIfType<[f32], CCAssignToReg<[S0, S1, S2, S3, S4, S5, S6, S7, S8, + S9, S10, S11, S12, S13, S14, S15]>>, + + // CPRCs may be allocated to co-processor registers or the stack - they + // may never be allocated to core registers. + CCIfType<[f32], CCAssignToStackWithShadow<4, 4, [Q0, Q1, Q2, Q3]>>, + CCIfType<[f64], CCAssignToStackWithShadow<8, 4, [Q0, Q1, Q2, Q3]>>, + CCIfType<[v2f64], CCAssignToStackWithShadow<16, 4, [Q0, Q1, Q2, Q3]>>, + + CCDelegateTo<CC_ARM_APCS> +]>; + +def RetFastCC_ARM_APCS : CallingConv<[ + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, + + CCIfType<[v2f64], CCAssignToReg<[Q0, Q1, Q2, Q3]>>, + CCIfType<[f64], CCAssignToReg<[D0, D1, D2, D3, D4, D5, D6, D7]>>, + CCIfType<[f32], CCAssignToReg<[S0, S1, S2, S3, S4, S5, S6, S7, S8, + S9, S10, S11, S12, S13, S14, S15]>>, + CCDelegateTo<RetCC_ARM_APCS> +]>; + +//===----------------------------------------------------------------------===// +// ARM APCS Calling Convention for GHC +//===----------------------------------------------------------------------===// + +def CC_ARM_APCS_GHC : CallingConv<[ + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, + + CCIfType<[v2f64], CCAssignToReg<[Q4, Q5]>>, + CCIfType<[f64], CCAssignToReg<[D8, D9, D10, D11]>>, + CCIfType<[f32], CCAssignToReg<[S16, S17, S18, S19, S20, S21, S22, S23]>>, + + // Promote i8/i16 arguments to i32. + CCIfType<[i8, i16], CCPromoteToType<i32>>, + + // Pass in STG registers: Base, Sp, Hp, R1, R2, R3, R4, SpLim + CCIfType<[i32], CCAssignToReg<[R4, R5, R6, R7, R8, R9, R10, R11]>> +]>; + +//===----------------------------------------------------------------------===// +// ARM AAPCS (EABI) Calling Convention, common parts +//===----------------------------------------------------------------------===// + +def CC_ARM_AAPCS_Common : CallingConv<[ + + CCIfType<[i1, i8, i16], CCPromoteToType<i32>>, + + // i64/f64 is passed in even pairs of GPRs + // i64 is 8-aligned i32 here, so we may need to eat R1 as a pad register + // (and the same is true for f64 if VFP is not enabled) + CCIfType<[i32], CCIfAlign<"8", CCAssignToRegWithShadow<[R0, R2], [R0, R1]>>>, + CCIfType<[i32], CCIf<"ArgFlags.getOrigAlign() != 8", + CCAssignToReg<[R0, R1, R2, R3]>>>, + + CCIfType<[i32], CCIfAlign<"8", CCAssignToStackWithShadow<4, 8, [R0, R1, R2, R3]>>>, + CCIfType<[i32], CCAssignToStackWithShadow<4, 4, [R0, R1, R2, R3]>>, + CCIfType<[f32], CCAssignToStackWithShadow<4, 4, [Q0, Q1, Q2, Q3]>>, + CCIfType<[f64], CCAssignToStackWithShadow<8, 8, [Q0, Q1, Q2, Q3]>>, + CCIfType<[v2f64], CCIfAlign<"16", + CCAssignToStackWithShadow<16, 16, [Q0, Q1, Q2, Q3]>>>, + CCIfType<[v2f64], CCAssignToStackWithShadow<16, 8, [Q0, Q1, Q2, Q3]>> +]>; + +def RetCC_ARM_AAPCS_Common : CallingConv<[ + CCIfType<[i1, i8, i16], CCPromoteToType<i32>>, + CCIfType<[i32], CCAssignToReg<[R0, R1, R2, R3]>>, + CCIfType<[i64], CCAssignToRegWithShadow<[R0, R2], [R1, R3]>> +]>; + +//===----------------------------------------------------------------------===// +// ARM AAPCS (EABI) Calling Convention +//===----------------------------------------------------------------------===// + +def CC_ARM_AAPCS : CallingConv<[ + // Handles byval parameters. + CCIfByVal<CCPassByVal<4, 4>>, + + // The 'nest' parameter, if any, is passed in R12. + CCIfNest<CCAssignToReg<[R12]>>, + + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>, + + // Pass SwiftSelf in a callee saved register. + CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>, + + // A SwiftError is passed in R8. + CCIfSwiftError<CCIfType<[i32], CCAssignToReg<[R8]>>>, + + CCIfType<[f64, v2f64], CCCustom<"CC_ARM_AAPCS_Custom_f64">>, + CCIfType<[f32], CCBitConvertToType<i32>>, + CCDelegateTo<CC_ARM_AAPCS_Common> +]>; + +def RetCC_ARM_AAPCS : CallingConv<[ + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v8f16,v16i8, v4f32], CCBitConvertToType<v2f64>>, + + // Pass SwiftSelf in a callee saved register. + CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>, + + // A SwiftError is returned in R8. + CCIfSwiftError<CCIfType<[i32], CCAssignToReg<[R8]>>>, + + CCIfType<[f64, v2f64], CCCustom<"RetCC_ARM_AAPCS_Custom_f64">>, + CCIfType<[f32], CCBitConvertToType<i32>>, + + CCDelegateTo<RetCC_ARM_AAPCS_Common> +]>; + +//===----------------------------------------------------------------------===// +// ARM AAPCS-VFP (EABI) Calling Convention +// Also used for FastCC (when VFP2 or later is available) +//===----------------------------------------------------------------------===// + +def CC_ARM_AAPCS_VFP : CallingConv<[ + // Handles byval parameters. + CCIfByVal<CCPassByVal<4, 4>>, + + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>, + + // Pass SwiftSelf in a callee saved register. + CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>, + + // A SwiftError is passed in R8. + CCIfSwiftError<CCIfType<[i32], CCAssignToReg<[R8]>>>, + + // HFAs are passed in a contiguous block of registers, or on the stack + CCIfConsecutiveRegs<CCCustom<"CC_ARM_AAPCS_Custom_Aggregate">>, + + CCIfType<[v2f64], CCAssignToReg<[Q0, Q1, Q2, Q3]>>, + CCIfType<[f64], CCAssignToReg<[D0, D1, D2, D3, D4, D5, D6, D7]>>, + CCIfType<[f32], CCAssignToReg<[S0, S1, S2, S3, S4, S5, S6, S7, S8, + S9, S10, S11, S12, S13, S14, S15]>>, + CCDelegateTo<CC_ARM_AAPCS_Common> +]>; + +def RetCC_ARM_AAPCS_VFP : CallingConv<[ + // Handle all vector types as either f64 or v2f64. + CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>, + CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>, + + // Pass SwiftSelf in a callee saved register. + CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>, + + // A SwiftError is returned in R8. + CCIfSwiftError<CCIfType<[i32], CCAssignToReg<[R8]>>>, + + CCIfType<[v2f64], CCAssignToReg<[Q0, Q1, Q2, Q3]>>, + CCIfType<[f64], CCAssignToReg<[D0, D1, D2, D3, D4, D5, D6, D7]>>, + CCIfType<[f32], CCAssignToReg<[S0, S1, S2, S3, S4, S5, S6, S7, S8, + S9, S10, S11, S12, S13, S14, S15]>>, + CCDelegateTo<RetCC_ARM_AAPCS_Common> +]>; + +//===----------------------------------------------------------------------===// +// Callee-saved register lists. +//===----------------------------------------------------------------------===// + +def CSR_NoRegs : CalleeSavedRegs<(add)>; +def CSR_FPRegs : CalleeSavedRegs<(add (sequence "D%u", 0, 31))>; + +def CSR_AAPCS : CalleeSavedRegs<(add LR, R11, R10, R9, R8, R7, R6, R5, R4, + (sequence "D%u", 15, 8))>; + +// R8 is used to pass swifterror, remove it from CSR. +def CSR_AAPCS_SwiftError : CalleeSavedRegs<(sub CSR_AAPCS, R8)>; + +// The order of callee-saved registers needs to match the order we actually push +// them in FrameLowering, because this order is what's used by +// PrologEpilogInserter to allocate frame index slots. So when R7 is the frame +// pointer, we use this AAPCS alternative. +def CSR_AAPCS_SplitPush : CalleeSavedRegs<(add LR, R7, R6, R5, R4, + R11, R10, R9, R8, + (sequence "D%u", 15, 8))>; + +// R8 is used to pass swifterror, remove it from CSR. +def CSR_AAPCS_SplitPush_SwiftError : CalleeSavedRegs<(sub CSR_AAPCS_SplitPush, + R8)>; + +// Constructors and destructors return 'this' in the ARM C++ ABI; since 'this' +// and the pointer return value are both passed in R0 in these cases, this can +// be partially modelled by treating R0 as a callee-saved register +// Only the resulting RegMask is used; the SaveList is ignored +def CSR_AAPCS_ThisReturn : CalleeSavedRegs<(add LR, R11, R10, R9, R8, R7, R6, + R5, R4, (sequence "D%u", 15, 8), + R0)>; + +// iOS ABI deviates from ARM standard ABI. R9 is not a callee-saved register. +// Also save R7-R4 first to match the stack frame fixed spill areas. +def CSR_iOS : CalleeSavedRegs<(add LR, R7, R6, R5, R4, (sub CSR_AAPCS, R9))>; + +// R8 is used to pass swifterror, remove it from CSR. +def CSR_iOS_SwiftError : CalleeSavedRegs<(sub CSR_iOS, R8)>; + +def CSR_iOS_ThisReturn : CalleeSavedRegs<(add LR, R7, R6, R5, R4, + (sub CSR_AAPCS_ThisReturn, R9))>; + +def CSR_iOS_TLSCall + : CalleeSavedRegs<(add LR, SP, (sub(sequence "R%u", 12, 1), R9, R12), + (sequence "D%u", 31, 0))>; + +// C++ TLS access function saves all registers except SP. Try to match +// the order of CSRs in CSR_iOS. +def CSR_iOS_CXX_TLS : CalleeSavedRegs<(add CSR_iOS, (sequence "R%u", 12, 1), + (sequence "D%u", 31, 0))>; + +// CSRs that are handled by prologue, epilogue. +def CSR_iOS_CXX_TLS_PE : CalleeSavedRegs<(add LR, R12, R11, R7, R5, R4)>; + +// CSRs that are handled explicitly via copies. +def CSR_iOS_CXX_TLS_ViaCopy : CalleeSavedRegs<(sub CSR_iOS_CXX_TLS, + CSR_iOS_CXX_TLS_PE)>; + +// The "interrupt" attribute is used to generate code that is acceptable in +// exception-handlers of various kinds. It makes us use a different return +// instruction (handled elsewhere) and affects which registers we must return to +// our "caller" in the same state as we receive them. + +// For most interrupts, all registers except SP and LR are shared with +// user-space. We mark LR to be saved anyway, since this is what the ARM backend +// generally does rather than tracking its liveness as a normal register. +def CSR_GenericInt : CalleeSavedRegs<(add LR, (sequence "R%u", 12, 0))>; + +// The fast interrupt handlers have more private state and get their own copies +// of R8-R12, in addition to SP and LR. As before, mark LR for saving too. + +// FIXME: we mark R11 as callee-saved since it's often the frame-pointer, and +// current frame lowering expects to encounter it while processing callee-saved +// registers. +def CSR_FIQ : CalleeSavedRegs<(add LR, R11, (sequence "R%u", 7, 0))>; + + diff --git a/capstone/suite/synctools/tablegen/ARM/ARMInstrFormats.td b/capstone/suite/synctools/tablegen/ARM/ARMInstrFormats.td new file mode 100644 index 000000000..1d3b1414f --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMInstrFormats.td @@ -0,0 +1,2620 @@ +//===-- ARMInstrFormats.td - ARM Instruction Formats -------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// +// ARM Instruction Format Definitions. +// + +// Format specifies the encoding used by the instruction. This is part of the +// ad-hoc solution used to emit machine instruction encodings by our machine +// code emitter. +class Format<bits<6> val> { + bits<6> Value = val; +} + +def Pseudo : Format<0>; +def MulFrm : Format<1>; +def BrFrm : Format<2>; +def BrMiscFrm : Format<3>; + +def DPFrm : Format<4>; +def DPSoRegRegFrm : Format<5>; + +def LdFrm : Format<6>; +def StFrm : Format<7>; +def LdMiscFrm : Format<8>; +def StMiscFrm : Format<9>; +def LdStMulFrm : Format<10>; + +def LdStExFrm : Format<11>; + +def ArithMiscFrm : Format<12>; +def SatFrm : Format<13>; +def ExtFrm : Format<14>; + +def VFPUnaryFrm : Format<15>; +def VFPBinaryFrm : Format<16>; +def VFPConv1Frm : Format<17>; +def VFPConv2Frm : Format<18>; +def VFPConv3Frm : Format<19>; +def VFPConv4Frm : Format<20>; +def VFPConv5Frm : Format<21>; +def VFPLdStFrm : Format<22>; +def VFPLdStMulFrm : Format<23>; +def VFPMiscFrm : Format<24>; + +def ThumbFrm : Format<25>; +def MiscFrm : Format<26>; + +def NGetLnFrm : Format<27>; +def NSetLnFrm : Format<28>; +def NDupFrm : Format<29>; +def NLdStFrm : Format<30>; +def N1RegModImmFrm: Format<31>; +def N2RegFrm : Format<32>; +def NVCVTFrm : Format<33>; +def NVDupLnFrm : Format<34>; +def N2RegVShLFrm : Format<35>; +def N2RegVShRFrm : Format<36>; +def N3RegFrm : Format<37>; +def N3RegVShFrm : Format<38>; +def NVExtFrm : Format<39>; +def NVMulSLFrm : Format<40>; +def NVTBLFrm : Format<41>; +def DPSoRegImmFrm : Format<42>; +def N3RegCplxFrm : Format<43>; + +// Misc flags. + +// The instruction has an Rn register operand. +// UnaryDP - Indicates this is a unary data processing instruction, i.e. +// it doesn't have a Rn operand. +class UnaryDP { bit isUnaryDataProc = 1; } + +// Xform16Bit - Indicates this Thumb2 instruction may be transformed into +// a 16-bit Thumb instruction if certain conditions are met. +class Xform16Bit { bit canXformTo16Bit = 1; } + +//===----------------------------------------------------------------------===// +// ARM Instruction flags. These need to match ARMBaseInstrInfo.h. +// + +// FIXME: Once the JIT is MC-ized, these can go away. +// Addressing mode. +class AddrMode<bits<5> val> { + bits<5> Value = val; +} +def AddrModeNone : AddrMode<0>; +def AddrMode1 : AddrMode<1>; +def AddrMode2 : AddrMode<2>; +def AddrMode3 : AddrMode<3>; +def AddrMode4 : AddrMode<4>; +def AddrMode5 : AddrMode<5>; +def AddrMode6 : AddrMode<6>; +def AddrModeT1_1 : AddrMode<7>; +def AddrModeT1_2 : AddrMode<8>; +def AddrModeT1_4 : AddrMode<9>; +def AddrModeT1_s : AddrMode<10>; +def AddrModeT2_i12 : AddrMode<11>; +def AddrModeT2_i8 : AddrMode<12>; +def AddrModeT2_so : AddrMode<13>; +def AddrModeT2_pc : AddrMode<14>; +def AddrModeT2_i8s4 : AddrMode<15>; +def AddrMode_i12 : AddrMode<16>; +def AddrMode5FP16 : AddrMode<17>; +def AddrModeT2_ldrex : AddrMode<18>; + +// Load / store index mode. +class IndexMode<bits<2> val> { + bits<2> Value = val; +} +def IndexModeNone : IndexMode<0>; +def IndexModePre : IndexMode<1>; +def IndexModePost : IndexMode<2>; +def IndexModeUpd : IndexMode<3>; + +// Instruction execution domain. +class Domain<bits<3> val> { + bits<3> Value = val; +} +def GenericDomain : Domain<0>; +def VFPDomain : Domain<1>; // Instructions in VFP domain only +def NeonDomain : Domain<2>; // Instructions in Neon domain only +def VFPNeonDomain : Domain<3>; // Instructions in both VFP & Neon domains +def VFPNeonA8Domain : Domain<5>; // Instructions in VFP & Neon under A8 + +//===----------------------------------------------------------------------===// +// ARM special operands. +// + +// ARM imod and iflag operands, used only by the CPS instruction. +def imod_op : Operand<i32> { + let PrintMethod = "printCPSIMod"; +} + +def ProcIFlagsOperand : AsmOperandClass { + let Name = "ProcIFlags"; + let ParserMethod = "parseProcIFlagsOperand"; +} +def iflags_op : Operand<i32> { + let PrintMethod = "printCPSIFlag"; + let ParserMatchClass = ProcIFlagsOperand; +} + +// ARM Predicate operand. Default to 14 = always (AL). Second part is CC +// register whose default is 0 (no register). +def CondCodeOperand : AsmOperandClass { let Name = "CondCode"; } +def pred : PredicateOperand<OtherVT, (ops i32imm, i32imm), + (ops (i32 14), (i32 zero_reg))> { + let PrintMethod = "printPredicateOperand"; + let ParserMatchClass = CondCodeOperand; + let DecoderMethod = "DecodePredicateOperand"; +} + +// Selectable predicate operand for CMOV instructions. We can't use a normal +// predicate because the default values interfere with instruction selection. In +// all other respects it is identical though: pseudo-instruction expansion +// relies on the MachineOperands being compatible. +def cmovpred : Operand<i32>, PredicateOp, + ComplexPattern<i32, 2, "SelectCMOVPred"> { + let MIOperandInfo = (ops i32imm, i32imm); + let PrintMethod = "printPredicateOperand"; +} + +// Conditional code result for instructions whose 's' bit is set, e.g. subs. +def CCOutOperand : AsmOperandClass { let Name = "CCOut"; } +def cc_out : OptionalDefOperand<OtherVT, (ops CCR), (ops (i32 zero_reg))> { + let EncoderMethod = "getCCOutOpValue"; + let PrintMethod = "printSBitModifierOperand"; + let ParserMatchClass = CCOutOperand; + let DecoderMethod = "DecodeCCOutOperand"; +} + +// Same as cc_out except it defaults to setting CPSR. +def s_cc_out : OptionalDefOperand<OtherVT, (ops CCR), (ops (i32 CPSR))> { + let EncoderMethod = "getCCOutOpValue"; + let PrintMethod = "printSBitModifierOperand"; + let ParserMatchClass = CCOutOperand; + let DecoderMethod = "DecodeCCOutOperand"; +} + +// ARM special operands for disassembly only. +// +def SetEndAsmOperand : ImmAsmOperand<0,1> { + let Name = "SetEndImm"; + let ParserMethod = "parseSetEndImm"; +} +def setend_op : Operand<i32> { + let PrintMethod = "printSetendOperand"; + let ParserMatchClass = SetEndAsmOperand; +} + +def MSRMaskOperand : AsmOperandClass { + let Name = "MSRMask"; + let ParserMethod = "parseMSRMaskOperand"; +} +def msr_mask : Operand<i32> { + let PrintMethod = "printMSRMaskOperand"; + let DecoderMethod = "DecodeMSRMask"; + let ParserMatchClass = MSRMaskOperand; +} + +def BankedRegOperand : AsmOperandClass { + let Name = "BankedReg"; + let ParserMethod = "parseBankedRegOperand"; +} +def banked_reg : Operand<i32> { + let PrintMethod = "printBankedRegOperand"; + let DecoderMethod = "DecodeBankedReg"; + let ParserMatchClass = BankedRegOperand; +} + +// Shift Right Immediate - A shift right immediate is encoded differently from +// other shift immediates. The imm6 field is encoded like so: +// +// Offset Encoding +// 8 imm6<5:3> = '001', 8 - <imm> is encoded in imm6<2:0> +// 16 imm6<5:4> = '01', 16 - <imm> is encoded in imm6<3:0> +// 32 imm6<5> = '1', 32 - <imm> is encoded in imm6<4:0> +// 64 64 - <imm> is encoded in imm6<5:0> +def shr_imm8_asm_operand : ImmAsmOperand<1,8> { let Name = "ShrImm8"; } +def shr_imm8 : Operand<i32>, ImmLeaf<i32, [{ return Imm > 0 && Imm <= 8; }]> { + let EncoderMethod = "getShiftRight8Imm"; + let DecoderMethod = "DecodeShiftRight8Imm"; + let ParserMatchClass = shr_imm8_asm_operand; +} +def shr_imm16_asm_operand : ImmAsmOperand<1,16> { let Name = "ShrImm16"; } +def shr_imm16 : Operand<i32>, ImmLeaf<i32, [{ return Imm > 0 && Imm <= 16; }]> { + let EncoderMethod = "getShiftRight16Imm"; + let DecoderMethod = "DecodeShiftRight16Imm"; + let ParserMatchClass = shr_imm16_asm_operand; +} +def shr_imm32_asm_operand : ImmAsmOperand<1,32> { let Name = "ShrImm32"; } +def shr_imm32 : Operand<i32>, ImmLeaf<i32, [{ return Imm > 0 && Imm <= 32; }]> { + let EncoderMethod = "getShiftRight32Imm"; + let DecoderMethod = "DecodeShiftRight32Imm"; + let ParserMatchClass = shr_imm32_asm_operand; +} +def shr_imm64_asm_operand : ImmAsmOperand<1,64> { let Name = "ShrImm64"; } +def shr_imm64 : Operand<i32>, ImmLeaf<i32, [{ return Imm > 0 && Imm <= 64; }]> { + let EncoderMethod = "getShiftRight64Imm"; + let DecoderMethod = "DecodeShiftRight64Imm"; + let ParserMatchClass = shr_imm64_asm_operand; +} + + +// ARM Assembler operand for ldr Rd, =expression which generates an offset +// to a constant pool entry or a MOV depending on the value of expression +def const_pool_asm_operand : AsmOperandClass { let Name = "ConstPoolAsmImm"; } +def const_pool_asm_imm : Operand<i32> { + let ParserMatchClass = const_pool_asm_operand; +} + + +//===----------------------------------------------------------------------===// +// ARM Assembler alias templates. +// +// Note: When EmitPriority == 1, the alias will be used for printing +class ARMInstAlias<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, Requires<[IsARM]>; +class ARMInstSubst<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, + Requires<[IsARM,UseNegativeImmediates]>; +class tInstAlias<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, Requires<[IsThumb]>; +class tInstSubst<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, + Requires<[IsThumb,UseNegativeImmediates]>; +class t2InstAlias<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, Requires<[IsThumb2]>; +class t2InstSubst<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, + Requires<[IsThumb2,UseNegativeImmediates]>; +class VFP2InstAlias<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, Requires<[HasVFP2]>; +class VFP2DPInstAlias<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, Requires<[HasVFP2,HasDPVFP]>; +class VFP3InstAlias<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, Requires<[HasVFP3]>; +class NEONInstAlias<string Asm, dag Result, bit EmitPriority = 0> + : InstAlias<Asm, Result, EmitPriority>, Requires<[HasNEON]>; + + +class VFP2MnemonicAlias<string src, string dst> : MnemonicAlias<src, dst>, + Requires<[HasVFP2]>; +class NEONMnemonicAlias<string src, string dst> : MnemonicAlias<src, dst>, + Requires<[HasNEON]>; + +//===----------------------------------------------------------------------===// +// ARM Instruction templates. +// + + +class InstTemplate<AddrMode am, int sz, IndexMode im, + Format f, Domain d, string cstr, InstrItinClass itin> + : Instruction { + let Namespace = "ARM"; + + AddrMode AM = am; + int Size = sz; + IndexMode IM = im; + bits<2> IndexModeBits = IM.Value; + Format F = f; + bits<6> Form = F.Value; + Domain D = d; + bit isUnaryDataProc = 0; + bit canXformTo16Bit = 0; + // The instruction is a 16-bit flag setting Thumb instruction. Used + // by the parser to determine whether to require the 'S' suffix on the + // mnemonic (when not in an IT block) or preclude it (when in an IT block). + bit thumbArithFlagSetting = 0; + + // If this is a pseudo instruction, mark it isCodeGenOnly. + let isCodeGenOnly = !eq(!cast<string>(f), "Pseudo"); + + // The layout of TSFlags should be kept in sync with ARMBaseInfo.h. + let TSFlags{4-0} = AM.Value; + let TSFlags{6-5} = IndexModeBits; + let TSFlags{12-7} = Form; + let TSFlags{13} = isUnaryDataProc; + let TSFlags{14} = canXformTo16Bit; + let TSFlags{17-15} = D.Value; + let TSFlags{18} = thumbArithFlagSetting; + + let Constraints = cstr; + let Itinerary = itin; +} + +class Encoding { + field bits<32> Inst; + // Mask of bits that cause an encoding to be UNPREDICTABLE. + // If a bit is set, then if the corresponding bit in the + // target encoding differs from its value in the "Inst" field, + // the instruction is UNPREDICTABLE (SoftFail in abstract parlance). + field bits<32> Unpredictable = 0; + // SoftFail is the generic name for this field, but we alias it so + // as to make it more obvious what it means in ARM-land. + field bits<32> SoftFail = Unpredictable; +} + +class InstARM<AddrMode am, int sz, IndexMode im, + Format f, Domain d, string cstr, InstrItinClass itin> + : InstTemplate<am, sz, im, f, d, cstr, itin>, Encoding { + let DecoderNamespace = "ARM"; +} + +// This Encoding-less class is used by Thumb1 to specify the encoding bits later +// on by adding flavors to specific instructions. +class InstThumb<AddrMode am, int sz, IndexMode im, + Format f, Domain d, string cstr, InstrItinClass itin> + : InstTemplate<am, sz, im, f, d, cstr, itin> { + let DecoderNamespace = "Thumb"; +} + +// Pseudo-instructions for alternate assembly syntax (never used by codegen). +// These are aliases that require C++ handling to convert to the target +// instruction, while InstAliases can be handled directly by tblgen. +class AsmPseudoInst<string asm, dag iops, dag oops = (outs)> + : InstTemplate<AddrModeNone, 0, IndexModeNone, Pseudo, GenericDomain, + "", NoItinerary> { + let OutOperandList = oops; + let InOperandList = iops; + let Pattern = []; + let isCodeGenOnly = 0; // So we get asm matcher for it. + let AsmString = asm; + let isPseudo = 1; +} + +class ARMAsmPseudo<string asm, dag iops, dag oops = (outs)> + : AsmPseudoInst<asm, iops, oops>, Requires<[IsARM]>; +class tAsmPseudo<string asm, dag iops, dag oops = (outs)> + : AsmPseudoInst<asm, iops, oops>, Requires<[IsThumb]>; +class t2AsmPseudo<string asm, dag iops, dag oops = (outs)> + : AsmPseudoInst<asm, iops, oops>, Requires<[IsThumb2]>; +class VFP2AsmPseudo<string asm, dag iops, dag oops = (outs)> + : AsmPseudoInst<asm, iops, oops>, Requires<[HasVFP2]>; +class NEONAsmPseudo<string asm, dag iops, dag oops = (outs)> + : AsmPseudoInst<asm, iops, oops>, Requires<[HasNEON]>; + +// Pseudo instructions for the code generator. +class PseudoInst<dag oops, dag iops, InstrItinClass itin, list<dag> pattern> + : InstTemplate<AddrModeNone, 0, IndexModeNone, Pseudo, + GenericDomain, "", itin> { + let OutOperandList = oops; + let InOperandList = iops; + let Pattern = pattern; + let isCodeGenOnly = 1; + let isPseudo = 1; +} + +// PseudoInst that's ARM-mode only. +class ARMPseudoInst<dag oops, dag iops, int sz, InstrItinClass itin, + list<dag> pattern> + : PseudoInst<oops, iops, itin, pattern> { + let Size = sz; + list<Predicate> Predicates = [IsARM]; +} + +// PseudoInst that's Thumb-mode only. +class tPseudoInst<dag oops, dag iops, int sz, InstrItinClass itin, + list<dag> pattern> + : PseudoInst<oops, iops, itin, pattern> { + let Size = sz; + list<Predicate> Predicates = [IsThumb]; +} + +// PseudoInst that's in ARMv8-M baseline (Somewhere between Thumb and Thumb2) +class t2basePseudoInst<dag oops, dag iops, int sz, InstrItinClass itin, + list<dag> pattern> + : PseudoInst<oops, iops, itin, pattern> { + let Size = sz; + list<Predicate> Predicates = [IsThumb,HasV8MBaseline]; +} + +// PseudoInst that's Thumb2-mode only. +class t2PseudoInst<dag oops, dag iops, int sz, InstrItinClass itin, + list<dag> pattern> + : PseudoInst<oops, iops, itin, pattern> { + let Size = sz; + list<Predicate> Predicates = [IsThumb2]; +} + +class ARMPseudoExpand<dag oops, dag iops, int sz, + InstrItinClass itin, list<dag> pattern, + dag Result> + : ARMPseudoInst<oops, iops, sz, itin, pattern>, + PseudoInstExpansion<Result>; + +class tPseudoExpand<dag oops, dag iops, int sz, + InstrItinClass itin, list<dag> pattern, + dag Result> + : tPseudoInst<oops, iops, sz, itin, pattern>, + PseudoInstExpansion<Result>; + +class t2PseudoExpand<dag oops, dag iops, int sz, + InstrItinClass itin, list<dag> pattern, + dag Result> + : t2PseudoInst<oops, iops, sz, itin, pattern>, + PseudoInstExpansion<Result>; + +// Almost all ARM instructions are predicable. +class I<dag oops, dag iops, AddrMode am, int sz, + IndexMode im, Format f, InstrItinClass itin, + string opc, string asm, string cstr, + list<dag> pattern> + : InstARM<am, sz, im, f, GenericDomain, cstr, itin> { + bits<4> p; + let Inst{31-28} = p; + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", asm); + let Pattern = pattern; + list<Predicate> Predicates = [IsARM]; +} + +// A few are not predicable +class InoP<dag oops, dag iops, AddrMode am, int sz, + IndexMode im, Format f, InstrItinClass itin, + string opc, string asm, string cstr, + list<dag> pattern> + : InstARM<am, sz, im, f, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = iops; + let AsmString = !strconcat(opc, asm); + let Pattern = pattern; + let isPredicable = 0; + list<Predicate> Predicates = [IsARM]; +} + +// Same as I except it can optionally modify CPSR. Note it's modeled as an input +// operand since by default it's a zero register. It will become an implicit def +// once it's "flipped". +class sI<dag oops, dag iops, AddrMode am, int sz, + IndexMode im, Format f, InstrItinClass itin, + string opc, string asm, string cstr, + list<dag> pattern> + : InstARM<am, sz, im, f, GenericDomain, cstr, itin> { + bits<4> p; // Predicate operand + bits<1> s; // condition-code set flag ('1' if the insn should set the flags) + let Inst{31-28} = p; + let Inst{20} = s; + + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p, cc_out:$s)); + let AsmString = !strconcat(opc, "${s}${p}", asm); + let Pattern = pattern; + list<Predicate> Predicates = [IsARM]; +} + +// Special cases +class XI<dag oops, dag iops, AddrMode am, int sz, + IndexMode im, Format f, InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : InstARM<am, sz, im, f, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = iops; + let AsmString = asm; + let Pattern = pattern; + list<Predicate> Predicates = [IsARM]; +} + +class AI<dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, f, itin, + opc, asm, "", pattern>; +class AsI<dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : sI<oops, iops, AddrModeNone, 4, IndexModeNone, f, itin, + opc, asm, "", pattern>; +class AXI<dag oops, dag iops, Format f, InstrItinClass itin, + string asm, list<dag> pattern> + : XI<oops, iops, AddrModeNone, 4, IndexModeNone, f, itin, + asm, "", pattern>; +class AXIM<dag oops, dag iops, AddrMode am, Format f, InstrItinClass itin, + string asm, list<dag> pattern> + : XI<oops, iops, am, 4, IndexModeNone, f, itin, + asm, "", pattern>; +class AInoP<dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : InoP<oops, iops, AddrModeNone, 4, IndexModeNone, f, itin, + opc, asm, "", pattern>; + +// Ctrl flow instructions +class ABI<bits<4> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, BrFrm, itin, + opc, asm, "", pattern> { + let Inst{27-24} = opcod; +} +class ABXI<bits<4> opcod, dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : XI<oops, iops, AddrModeNone, 4, IndexModeNone, BrFrm, itin, + asm, "", pattern> { + let Inst{27-24} = opcod; +} + +// BR_JT instructions +class JTI<dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : XI<oops, iops, AddrModeNone, 0, IndexModeNone, BrMiscFrm, itin, + asm, "", pattern>; + +class AIldr_ex_or_acq<bits<2> opcod, bits<2> opcod2, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, LdStExFrm, itin, + opc, asm, "", pattern> { + bits<4> Rt; + bits<4> addr; + let Inst{27-23} = 0b00011; + let Inst{22-21} = opcod; + let Inst{20} = 1; + let Inst{19-16} = addr; + let Inst{15-12} = Rt; + let Inst{11-10} = 0b11; + let Inst{9-8} = opcod2; + let Inst{7-0} = 0b10011111; +} +class AIstr_ex_or_rel<bits<2> opcod, bits<2> opcod2, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, LdStExFrm, itin, + opc, asm, "", pattern> { + bits<4> Rt; + bits<4> addr; + let Inst{27-23} = 0b00011; + let Inst{22-21} = opcod; + let Inst{20} = 0; + let Inst{19-16} = addr; + let Inst{11-10} = 0b11; + let Inst{9-8} = opcod2; + let Inst{7-4} = 0b1001; + let Inst{3-0} = Rt; +} +// Atomic load/store instructions +class AIldrex<bits<2> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AIldr_ex_or_acq<opcod, 0b11, oops, iops, itin, opc, asm, pattern>; + +class AIstrex<bits<2> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AIstr_ex_or_rel<opcod, 0b11, oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + let Inst{15-12} = Rd; +} + +// Exclusive load/store instructions + +class AIldaex<bits<2> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AIldr_ex_or_acq<opcod, 0b10, oops, iops, itin, opc, asm, pattern>, + Requires<[IsARM, HasAcquireRelease, HasV7Clrex]>; + +class AIstlex<bits<2> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AIstr_ex_or_rel<opcod, 0b10, oops, iops, itin, opc, asm, pattern>, + Requires<[IsARM, HasAcquireRelease, HasV7Clrex]> { + bits<4> Rd; + let Inst{15-12} = Rd; +} + +class AIswp<bit b, dag oops, dag iops, string opc, list<dag> pattern> + : AI<oops, iops, MiscFrm, NoItinerary, opc, "\t$Rt, $Rt2, $addr", pattern> { + bits<4> Rt; + bits<4> Rt2; + bits<4> addr; + let Inst{27-23} = 0b00010; + let Inst{22} = b; + let Inst{21-20} = 0b00; + let Inst{19-16} = addr; + let Inst{15-12} = Rt; + let Inst{11-4} = 0b00001001; + let Inst{3-0} = Rt2; + + let Unpredictable{11-8} = 0b1111; + let DecoderMethod = "DecodeSwap"; +} +// Acquire/Release load/store instructions +class AIldracq<bits<2> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AIldr_ex_or_acq<opcod, 0b00, oops, iops, itin, opc, asm, pattern>, + Requires<[IsARM, HasAcquireRelease]>; + +class AIstrrel<bits<2> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AIstr_ex_or_rel<opcod, 0b00, oops, iops, itin, opc, asm, pattern>, + Requires<[IsARM, HasAcquireRelease]> { + let Inst{15-12} = 0b1111; +} + +// addrmode1 instructions +class AI1<bits<4> opcod, dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrMode1, 4, IndexModeNone, f, itin, + opc, asm, "", pattern> { + let Inst{24-21} = opcod; + let Inst{27-26} = 0b00; +} +class AsI1<bits<4> opcod, dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : sI<oops, iops, AddrMode1, 4, IndexModeNone, f, itin, + opc, asm, "", pattern> { + let Inst{24-21} = opcod; + let Inst{27-26} = 0b00; +} +class AXI1<bits<4> opcod, dag oops, dag iops, Format f, InstrItinClass itin, + string asm, list<dag> pattern> + : XI<oops, iops, AddrMode1, 4, IndexModeNone, f, itin, + asm, "", pattern> { + let Inst{24-21} = opcod; + let Inst{27-26} = 0b00; +} + +// loads + +// LDR/LDRB/STR/STRB/... +class AI2ldst<bits<3> op, bit isLd, bit isByte, dag oops, dag iops, AddrMode am, + Format f, InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : I<oops, iops, am, 4, IndexModeNone, f, itin, opc, asm, + "", pattern> { + let Inst{27-25} = op; + let Inst{24} = 1; // 24 == P + // 23 == U + let Inst{22} = isByte; + let Inst{21} = 0; // 21 == W + let Inst{20} = isLd; +} +// Indexed load/stores +class AI2ldstidx<bit isLd, bit isByte, bit isPre, dag oops, dag iops, + IndexMode im, Format f, InstrItinClass itin, string opc, + string asm, string cstr, list<dag> pattern> + : I<oops, iops, AddrMode2, 4, im, f, itin, + opc, asm, cstr, pattern> { + bits<4> Rt; + let Inst{27-26} = 0b01; + let Inst{24} = isPre; // P bit + let Inst{22} = isByte; // B bit + let Inst{21} = isPre; // W bit + let Inst{20} = isLd; // L bit + let Inst{15-12} = Rt; +} +class AI2stridx_reg<bit isByte, bit isPre, dag oops, dag iops, + IndexMode im, Format f, InstrItinClass itin, string opc, + string asm, string cstr, list<dag> pattern> + : AI2ldstidx<0, isByte, isPre, oops, iops, im, f, itin, opc, asm, cstr, + pattern> { + // AM2 store w/ two operands: (GPR, am2offset) + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> Rn; + let Inst{25} = 1; + let Inst{23} = offset{12}; + let Inst{19-16} = Rn; + let Inst{11-5} = offset{11-5}; + let Inst{4} = 0; + let Inst{3-0} = offset{3-0}; +} + +class AI2stridx_imm<bit isByte, bit isPre, dag oops, dag iops, + IndexMode im, Format f, InstrItinClass itin, string opc, + string asm, string cstr, list<dag> pattern> + : AI2ldstidx<0, isByte, isPre, oops, iops, im, f, itin, opc, asm, cstr, + pattern> { + // AM2 store w/ two operands: (GPR, am2offset) + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> Rn; + let Inst{25} = 0; + let Inst{23} = offset{12}; + let Inst{19-16} = Rn; + let Inst{11-0} = offset{11-0}; +} + + +// FIXME: Merge with the above class when addrmode2 gets used for STR, STRB +// but for now use this class for STRT and STRBT. +class AI2stridxT<bit isByte, bit isPre, dag oops, dag iops, + IndexMode im, Format f, InstrItinClass itin, string opc, + string asm, string cstr, list<dag> pattern> + : AI2ldstidx<0, isByte, isPre, oops, iops, im, f, itin, opc, asm, cstr, + pattern> { + // AM2 store w/ two operands: (GPR, am2offset) + // {17-14} Rn + // {13} 1 == Rm, 0 == imm12 + // {12} isAdd + // {11-0} imm12/Rm + bits<18> addr; + let Inst{25} = addr{13}; + let Inst{23} = addr{12}; + let Inst{19-16} = addr{17-14}; + let Inst{11-0} = addr{11-0}; +} + +// addrmode3 instructions +class AI3ld<bits<4> op, bit op20, dag oops, dag iops, Format f, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrMode3, 4, IndexModeNone, f, itin, + opc, asm, "", pattern> { + bits<14> addr; + bits<4> Rt; + let Inst{27-25} = 0b000; + let Inst{24} = 1; // P bit + let Inst{23} = addr{8}; // U bit + let Inst{22} = addr{13}; // 1 == imm8, 0 == Rm + let Inst{21} = 0; // W bit + let Inst{20} = op20; // L bit + let Inst{19-16} = addr{12-9}; // Rn + let Inst{15-12} = Rt; // Rt + let Inst{11-8} = addr{7-4}; // imm7_4/zero + let Inst{7-4} = op; + let Inst{3-0} = addr{3-0}; // imm3_0/Rm + + let DecoderMethod = "DecodeAddrMode3Instruction"; +} + +class AI3ldstidx<bits<4> op, bit op20, bit isPre, dag oops, dag iops, + IndexMode im, Format f, InstrItinClass itin, string opc, + string asm, string cstr, list<dag> pattern> + : I<oops, iops, AddrMode3, 4, im, f, itin, + opc, asm, cstr, pattern> { + bits<4> Rt; + let Inst{27-25} = 0b000; + let Inst{24} = isPre; // P bit + let Inst{21} = isPre; // W bit + let Inst{20} = op20; // L bit + let Inst{15-12} = Rt; // Rt + let Inst{7-4} = op; +} + +// FIXME: Merge with the above class when addrmode2 gets used for LDR, LDRB +// but for now use this class for LDRSBT, LDRHT, LDSHT. +class AI3ldstidxT<bits<4> op, bit isLoad, dag oops, dag iops, + IndexMode im, Format f, InstrItinClass itin, string opc, + string asm, string cstr, list<dag> pattern> + : I<oops, iops, AddrMode3, 4, im, f, itin, opc, asm, cstr, pattern> { + // {13} 1 == imm8, 0 == Rm + // {12-9} Rn + // {8} isAdd + // {7-4} imm7_4/zero + // {3-0} imm3_0/Rm + bits<4> addr; + bits<4> Rt; + let Inst{27-25} = 0b000; + let Inst{24} = 0; // P bit + let Inst{21} = 1; + let Inst{20} = isLoad; // L bit + let Inst{19-16} = addr; // Rn + let Inst{15-12} = Rt; // Rt + let Inst{7-4} = op; +} + +// stores +class AI3str<bits<4> op, dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrMode3, 4, IndexModeNone, f, itin, + opc, asm, "", pattern> { + bits<14> addr; + bits<4> Rt; + let Inst{27-25} = 0b000; + let Inst{24} = 1; // P bit + let Inst{23} = addr{8}; // U bit + let Inst{22} = addr{13}; // 1 == imm8, 0 == Rm + let Inst{21} = 0; // W bit + let Inst{20} = 0; // L bit + let Inst{19-16} = addr{12-9}; // Rn + let Inst{15-12} = Rt; // Rt + let Inst{11-8} = addr{7-4}; // imm7_4/zero + let Inst{7-4} = op; + let Inst{3-0} = addr{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; +} + +// addrmode4 instructions +class AXI4<dag oops, dag iops, IndexMode im, Format f, InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : XI<oops, iops, AddrMode4, 4, im, f, itin, asm, cstr, pattern> { + bits<4> p; + bits<16> regs; + bits<4> Rn; + let Inst{31-28} = p; + let Inst{27-25} = 0b100; + let Inst{22} = 0; // S bit + let Inst{19-16} = Rn; + let Inst{15-0} = regs; +} + +// Unsigned multiply, multiply-accumulate instructions. +class AMul1I<bits<7> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, MulFrm, itin, + opc, asm, "", pattern> { + let Inst{7-4} = 0b1001; + let Inst{20} = 0; // S bit + let Inst{27-21} = opcod; +} +class AsMul1I<bits<7> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : sI<oops, iops, AddrModeNone, 4, IndexModeNone, MulFrm, itin, + opc, asm, "", pattern> { + let Inst{7-4} = 0b1001; + let Inst{27-21} = opcod; +} + +// Most significant word multiply +class AMul2I<bits<7> opcod, bits<4> opc7_4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, MulFrm, itin, + opc, asm, "", pattern> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + let Inst{7-4} = opc7_4; + let Inst{20} = 1; + let Inst{27-21} = opcod; + let Inst{19-16} = Rd; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} +// MSW multiple w/ Ra operand +class AMul2Ia<bits<7> opcod, bits<4> opc7_4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : AMul2I<opcod, opc7_4, oops, iops, itin, opc, asm, pattern> { + bits<4> Ra; + let Inst{15-12} = Ra; +} + +// SMUL<x><y> / SMULW<y> / SMLA<x><y> / SMLAW<x><y> +class AMulxyIbase<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, MulFrm, itin, + opc, asm, "", pattern> { + bits<4> Rn; + bits<4> Rm; + let Inst{4} = 0; + let Inst{7} = 1; + let Inst{20} = 0; + let Inst{27-21} = opcod; + let Inst{6-5} = bit6_5; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} +class AMulxyI<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : AMulxyIbase<opcod, bit6_5, oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + let Inst{19-16} = Rd; +} + +// AMulxyI with Ra operand +class AMulxyIa<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : AMulxyI<opcod, bit6_5, oops, iops, itin, opc, asm, pattern> { + bits<4> Ra; + let Inst{15-12} = Ra; +} +// SMLAL* +class AMulxyI64<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : AMulxyIbase<opcod, bit6_5, oops, iops, itin, opc, asm, pattern> { + bits<4> RdLo; + bits<4> RdHi; + let Inst{19-16} = RdHi; + let Inst{15-12} = RdLo; +} + +// Extend instructions. +class AExtI<bits<8> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, ExtFrm, itin, + opc, asm, "", pattern> { + // All AExtI instructions have Rd and Rm register operands. + bits<4> Rd; + bits<4> Rm; + let Inst{15-12} = Rd; + let Inst{3-0} = Rm; + let Inst{7-4} = 0b0111; + let Inst{9-8} = 0b00; + let Inst{27-20} = opcod; + + let Unpredictable{9-8} = 0b11; +} + +// Misc Arithmetic instructions. +class AMiscA1I<bits<8> opcod, bits<4> opc7_4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, ArithMiscFrm, itin, + opc, asm, "", pattern> { + bits<4> Rd; + bits<4> Rm; + let Inst{27-20} = opcod; + let Inst{19-16} = 0b1111; + let Inst{15-12} = Rd; + let Inst{11-8} = 0b1111; + let Inst{7-4} = opc7_4; + let Inst{3-0} = Rm; +} + +// Division instructions. +class ADivA1I<bits<3> opcod, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, ArithMiscFrm, itin, + opc, asm, "", pattern> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + let Inst{27-23} = 0b01110; + let Inst{22-20} = opcod; + let Inst{19-16} = Rd; + let Inst{15-12} = 0b1111; + let Inst{11-8} = Rm; + let Inst{7-4} = 0b0001; + let Inst{3-0} = Rn; +} + +// PKH instructions +def PKHLSLAsmOperand : ImmAsmOperand<0,31> { + let Name = "PKHLSLImm"; + let ParserMethod = "parsePKHLSLImm"; +} +def pkh_lsl_amt: Operand<i32>, ImmLeaf<i32, [{ return Imm >= 0 && Imm < 32; }]>{ + let PrintMethod = "printPKHLSLShiftImm"; + let ParserMatchClass = PKHLSLAsmOperand; +} +def PKHASRAsmOperand : AsmOperandClass { + let Name = "PKHASRImm"; + let ParserMethod = "parsePKHASRImm"; +} +def pkh_asr_amt: Operand<i32>, ImmLeaf<i32, [{ return Imm > 0 && Imm <= 32; }]>{ + let PrintMethod = "printPKHASRShiftImm"; + let ParserMatchClass = PKHASRAsmOperand; +} + +class APKHI<bits<8> opcod, bit tb, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : I<oops, iops, AddrModeNone, 4, IndexModeNone, ArithMiscFrm, itin, + opc, asm, "", pattern> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + bits<5> sh; + let Inst{27-20} = opcod; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-7} = sh; + let Inst{6} = tb; + let Inst{5-4} = 0b01; + let Inst{3-0} = Rm; +} + +//===----------------------------------------------------------------------===// + +// ARMPat - Same as Pat<>, but requires that the compiler be in ARM mode. +class ARMPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsARM]; +} +class ARMV5TPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsARM, HasV5T]; +} +class ARMV5TEPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsARM, HasV5TE]; +} +// ARMV5MOPat - Same as ARMV5TEPat with UseMulOps. +class ARMV5MOPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsARM, HasV5TE, UseMulOps]; +} +class ARMV6Pat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsARM, HasV6]; +} +class VFPPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [HasVFP2]; +} +class VFPNoNEONPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [HasVFP2, DontUseNEONForFP]; +} +class Thumb2DSPPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsThumb2, HasDSP]; +} +class Thumb2DSPMulPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsThumb2, UseMulOps, HasDSP]; +} +class FP16Pat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [HasFP16]; +} +class FullFP16Pat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [HasFullFP16]; +} +//===----------------------------------------------------------------------===// +// Thumb Instruction Format Definitions. +// + +class ThumbI<dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, string asm, string cstr, list<dag> pattern> + : InstThumb<am, sz, IndexModeNone, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = iops; + let AsmString = asm; + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb]; +} + +// TI - Thumb instruction. +class TI<dag oops, dag iops, InstrItinClass itin, string asm, list<dag> pattern> + : ThumbI<oops, iops, AddrModeNone, 2, itin, asm, "", pattern>; + +// Two-address instructions +class TIt<dag oops, dag iops, InstrItinClass itin, string asm, + list<dag> pattern> + : ThumbI<oops, iops, AddrModeNone, 2, itin, asm, "$lhs = $dst", + pattern>; + +// tBL, tBX 32-bit instructions +class TIx2<bits<5> opcod1, bits<2> opcod2, bit opcod3, + dag oops, dag iops, InstrItinClass itin, string asm, + list<dag> pattern> + : ThumbI<oops, iops, AddrModeNone, 4, itin, asm, "", pattern>, + Encoding { + let Inst{31-27} = opcod1; + let Inst{15-14} = opcod2; + let Inst{12} = opcod3; +} + +// BR_JT instructions +class TJTI<dag oops, dag iops, InstrItinClass itin, string asm, + list<dag> pattern> + : ThumbI<oops, iops, AddrModeNone, 0, itin, asm, "", pattern>; + +// Thumb1 only +class Thumb1I<dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, string asm, string cstr, list<dag> pattern> + : InstThumb<am, sz, IndexModeNone, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = iops; + let AsmString = asm; + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb, IsThumb1Only]; +} + +class T1I<dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : Thumb1I<oops, iops, AddrModeNone, 2, itin, asm, "", pattern>; +class T1Ix2<dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : Thumb1I<oops, iops, AddrModeNone, 4, itin, asm, "", pattern>; + +// Two-address instructions +class T1It<dag oops, dag iops, InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : Thumb1I<oops, iops, AddrModeNone, 2, itin, + asm, cstr, pattern>; + +// Thumb1 instruction that can either be predicated or set CPSR. +class Thumb1sI<dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : InstThumb<am, sz, IndexModeNone, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = !con(oops, (outs s_cc_out:$s)); + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${s}${p}", asm); + let Pattern = pattern; + let thumbArithFlagSetting = 1; + list<Predicate> Predicates = [IsThumb, IsThumb1Only]; + let DecoderNamespace = "ThumbSBit"; +} + +class T1sI<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb1sI<oops, iops, AddrModeNone, 2, itin, opc, asm, "", pattern>; + +// Two-address instructions +class T1sIt<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb1sI<oops, iops, AddrModeNone, 2, itin, opc, asm, + "$Rn = $Rdn", pattern>; + +// Thumb1 instruction that can be predicated. +class Thumb1pI<dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : InstThumb<am, sz, IndexModeNone, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", asm); + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb, IsThumb1Only]; +} + +class T1pI<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb1pI<oops, iops, AddrModeNone, 2, itin, opc, asm, "", pattern>; + +// Two-address instructions +class T1pIt<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb1pI<oops, iops, AddrModeNone, 2, itin, opc, asm, + "$Rn = $Rdn", pattern>; + +class T1pIs<dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : Thumb1pI<oops, iops, AddrModeT1_s, 2, itin, opc, asm, "", pattern>; + +class Encoding16 : Encoding { + let Inst{31-16} = 0x0000; +} + +// A6.2 16-bit Thumb instruction encoding +class T1Encoding<bits<6> opcode> : Encoding16 { + let Inst{15-10} = opcode; +} + +// A6.2.1 Shift (immediate), add, subtract, move, and compare encoding. +class T1General<bits<5> opcode> : Encoding16 { + let Inst{15-14} = 0b00; + let Inst{13-9} = opcode; +} + +// A6.2.2 Data-processing encoding. +class T1DataProcessing<bits<4> opcode> : Encoding16 { + let Inst{15-10} = 0b010000; + let Inst{9-6} = opcode; +} + +// A6.2.3 Special data instructions and branch and exchange encoding. +class T1Special<bits<4> opcode> : Encoding16 { + let Inst{15-10} = 0b010001; + let Inst{9-6} = opcode; +} + +// A6.2.4 Load/store single data item encoding. +class T1LoadStore<bits<4> opA, bits<3> opB> : Encoding16 { + let Inst{15-12} = opA; + let Inst{11-9} = opB; +} +class T1LdStSP<bits<3> opB> : T1LoadStore<0b1001, opB>; // SP relative + +class T1BranchCond<bits<4> opcode> : Encoding16 { + let Inst{15-12} = opcode; +} + +// Helper classes to encode Thumb1 loads and stores. For immediates, the +// following bits are used for "opA" (see A6.2.4): +// +// 0b0110 => Immediate, 4 bytes +// 0b1000 => Immediate, 2 bytes +// 0b0111 => Immediate, 1 byte +class T1pILdStEncode<bits<3> opcode, dag oops, dag iops, AddrMode am, + InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : Thumb1pI<oops, iops, am, 2, itin, opc, asm, "", pattern>, + T1LoadStore<0b0101, opcode> { + bits<3> Rt; + bits<8> addr; + let Inst{8-6} = addr{5-3}; // Rm + let Inst{5-3} = addr{2-0}; // Rn + let Inst{2-0} = Rt; +} +class T1pILdStEncodeImm<bits<4> opA, bit opB, dag oops, dag iops, AddrMode am, + InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : Thumb1pI<oops, iops, am, 2, itin, opc, asm, "", pattern>, + T1LoadStore<opA, {opB,?,?}> { + bits<3> Rt; + bits<8> addr; + let Inst{10-6} = addr{7-3}; // imm5 + let Inst{5-3} = addr{2-0}; // Rn + let Inst{2-0} = Rt; +} + +// A6.2.5 Miscellaneous 16-bit instructions encoding. +class T1Misc<bits<7> opcode> : Encoding16 { + let Inst{15-12} = 0b1011; + let Inst{11-5} = opcode; +} + +// Thumb2I - Thumb2 instruction. Almost all Thumb2 instructions are predicable. +class Thumb2I<dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : InstARM<am, sz, IndexModeNone, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", asm); + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb2]; + let DecoderNamespace = "Thumb2"; +} + +// Same as Thumb2I except it can optionally modify CPSR. Note it's modeled as an +// input operand since by default it's a zero register. It will become an +// implicit def once it's "flipped". +// +// FIXME: This uses unified syntax so {s} comes before {p}. We should make it +// more consistent. +class Thumb2sI<dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : InstARM<am, sz, IndexModeNone, ThumbFrm, GenericDomain, cstr, itin> { + bits<1> s; // condition-code set flag ('1' if the insn should set the flags) + let Inst{20} = s; + + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p, cc_out:$s)); + let AsmString = !strconcat(opc, "${s}${p}", asm); + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb2]; + let DecoderNamespace = "Thumb2"; +} + +// Special cases +class Thumb2XI<dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : InstARM<am, sz, IndexModeNone, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = iops; + let AsmString = asm; + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb2]; + let DecoderNamespace = "Thumb2"; +} + +class ThumbXI<dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : InstARM<am, sz, IndexModeNone, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = iops; + let AsmString = asm; + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb, IsThumb1Only]; + let DecoderNamespace = "Thumb"; +} + +class T2I<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb2I<oops, iops, AddrModeNone, 4, itin, opc, asm, "", pattern>; +class T2Ii12<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb2I<oops, iops, AddrModeT2_i12, 4, itin, opc, asm, "",pattern>; +class T2Ii8<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb2I<oops, iops, AddrModeT2_i8, 4, itin, opc, asm, "", pattern>; +class T2Iso<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb2I<oops, iops, AddrModeT2_so, 4, itin, opc, asm, "", pattern>; +class T2Ipc<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb2I<oops, iops, AddrModeT2_pc, 4, itin, opc, asm, "", pattern>; +class T2Ii8s4<bit P, bit W, bit isLoad, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : Thumb2I<oops, iops, AddrModeT2_i8s4, 4, itin, opc, asm, cstr, + pattern> { + bits<4> Rt; + bits<4> Rt2; + bits<13> addr; + let Inst{31-25} = 0b1110100; + let Inst{24} = P; + let Inst{23} = addr{8}; + let Inst{22} = 1; + let Inst{21} = W; + let Inst{20} = isLoad; + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = Rt{3-0}; + let Inst{11-8} = Rt2{3-0}; + let Inst{7-0} = addr{7-0}; +} +class T2Ii8s4post<bit P, bit W, bit isLoad, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, string cstr, + list<dag> pattern> + : Thumb2I<oops, iops, AddrModeT2_i8s4, 4, itin, opc, asm, cstr, + pattern> { + bits<4> Rt; + bits<4> Rt2; + bits<4> addr; + bits<9> imm; + let Inst{31-25} = 0b1110100; + let Inst{24} = P; + let Inst{23} = imm{8}; + let Inst{22} = 1; + let Inst{21} = W; + let Inst{20} = isLoad; + let Inst{19-16} = addr; + let Inst{15-12} = Rt{3-0}; + let Inst{11-8} = Rt2{3-0}; + let Inst{7-0} = imm{7-0}; +} + +class T2sI<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : Thumb2sI<oops, iops, AddrModeNone, 4, itin, opc, asm, "", pattern>; + +class T2XI<dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : Thumb2XI<oops, iops, AddrModeNone, 4, itin, asm, "", pattern>; +class T2JTI<dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : Thumb2XI<oops, iops, AddrModeNone, 0, itin, asm, "", pattern>; + +// Move to/from coprocessor instructions +class T2Cop<bits<4> opc, dag oops, dag iops, string opcstr, string asm, + list<dag> pattern> + : T2I <oops, iops, NoItinerary, opcstr, asm, pattern>, Requires<[IsThumb2]> { + let Inst{31-28} = opc; +} + +// Two-address instructions +class T2XIt<dag oops, dag iops, InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : Thumb2XI<oops, iops, AddrModeNone, 4, itin, asm, cstr, pattern>; + +// T2Ipreldst - Thumb2 pre-indexed load / store instructions. +class T2Ipreldst<bit signed, bits<2> opcod, bit load, bit pre, + dag oops, dag iops, + AddrMode am, IndexMode im, InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : InstARM<am, 4, im, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", asm); + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb2]; + let DecoderNamespace = "Thumb2"; + + bits<4> Rt; + bits<13> addr; + let Inst{31-27} = 0b11111; + let Inst{26-25} = 0b00; + let Inst{24} = signed; + let Inst{23} = 0; + let Inst{22-21} = opcod; + let Inst{20} = load; + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = Rt{3-0}; + let Inst{11} = 1; + // (P, W) = (1, 1) Pre-indexed or (0, 1) Post-indexed + let Inst{10} = pre; // The P bit. + let Inst{9} = addr{8}; // Sign bit + let Inst{8} = 1; // The W bit. + let Inst{7-0} = addr{7-0}; + + let DecoderMethod = "DecodeT2LdStPre"; +} + +// T2Ipostldst - Thumb2 post-indexed load / store instructions. +class T2Ipostldst<bit signed, bits<2> opcod, bit load, bit pre, + dag oops, dag iops, + AddrMode am, IndexMode im, InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : InstARM<am, 4, im, ThumbFrm, GenericDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", asm); + let Pattern = pattern; + list<Predicate> Predicates = [IsThumb2]; + let DecoderNamespace = "Thumb2"; + + bits<4> Rt; + bits<4> Rn; + bits<9> offset; + let Inst{31-27} = 0b11111; + let Inst{26-25} = 0b00; + let Inst{24} = signed; + let Inst{23} = 0; + let Inst{22-21} = opcod; + let Inst{20} = load; + let Inst{19-16} = Rn; + let Inst{15-12} = Rt{3-0}; + let Inst{11} = 1; + // (P, W) = (1, 1) Pre-indexed or (0, 1) Post-indexed + let Inst{10} = pre; // The P bit. + let Inst{9} = offset{8}; // Sign bit + let Inst{8} = 1; // The W bit. + let Inst{7-0} = offset{7-0}; + + let DecoderMethod = "DecodeT2LdStPre"; +} + +// T1Pat - Same as Pat<>, but requires that the compiler be in Thumb1 mode. +class T1Pat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsThumb, IsThumb1Only]; +} + +// T2v6Pat - Same as Pat<>, but requires V6T2 Thumb2 mode. +class T2v6Pat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsThumb2, HasV6T2]; +} + +// T2Pat - Same as Pat<>, but requires that the compiler be in Thumb2 mode. +class T2Pat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [IsThumb2]; +} + +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// ARM VFP Instruction templates. +// + +// Almost all VFP instructions are predicable. +class VFPI<dag oops, dag iops, AddrMode am, int sz, + IndexMode im, Format f, InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : InstARM<am, sz, im, f, VFPDomain, cstr, itin> { + bits<4> p; + let Inst{31-28} = p; + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", asm); + let Pattern = pattern; + let PostEncoderMethod = "VFPThumb2PostEncoder"; + let DecoderNamespace = "VFP"; + list<Predicate> Predicates = [HasVFP2]; +} + +// Special cases +class VFPXI<dag oops, dag iops, AddrMode am, int sz, + IndexMode im, Format f, InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : InstARM<am, sz, im, f, VFPDomain, cstr, itin> { + bits<4> p; + let Inst{31-28} = p; + let OutOperandList = oops; + let InOperandList = iops; + let AsmString = asm; + let Pattern = pattern; + let PostEncoderMethod = "VFPThumb2PostEncoder"; + let DecoderNamespace = "VFP"; + list<Predicate> Predicates = [HasVFP2]; +} + +class VFPAI<dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : VFPI<oops, iops, AddrModeNone, 4, IndexModeNone, f, itin, + opc, asm, "", pattern> { + let PostEncoderMethod = "VFPThumb2PostEncoder"; +} + +// ARM VFP addrmode5 loads and stores +class ADI5<bits<4> opcod1, bits<2> opcod2, dag oops, dag iops, + InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : VFPI<oops, iops, AddrMode5, 4, IndexModeNone, + VFPLdStFrm, itin, opc, asm, "", pattern> { + // Instruction operands. + bits<5> Dd; + bits<13> addr; + + // Encode instruction operands. + let Inst{23} = addr{8}; // U (add = (U == '1')) + let Inst{22} = Dd{4}; + let Inst{19-16} = addr{12-9}; // Rn + let Inst{15-12} = Dd{3-0}; + let Inst{7-0} = addr{7-0}; // imm8 + + let Inst{27-24} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-9} = 0b101; + let Inst{8} = 1; // Double precision + + // Loads & stores operate on both NEON and VFP pipelines. + let D = VFPNeonDomain; +} + +class ASI5<bits<4> opcod1, bits<2> opcod2, dag oops, dag iops, + InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : VFPI<oops, iops, AddrMode5, 4, IndexModeNone, + VFPLdStFrm, itin, opc, asm, "", pattern> { + // Instruction operands. + bits<5> Sd; + bits<13> addr; + + // Encode instruction operands. + let Inst{23} = addr{8}; // U (add = (U == '1')) + let Inst{22} = Sd{0}; + let Inst{19-16} = addr{12-9}; // Rn + let Inst{15-12} = Sd{4-1}; + let Inst{7-0} = addr{7-0}; // imm8 + + let Inst{27-24} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-9} = 0b101; + let Inst{8} = 0; // Single precision + + // Loads & stores operate on both NEON and VFP pipelines. + let D = VFPNeonDomain; +} + +class AHI5<bits<4> opcod1, bits<2> opcod2, dag oops, dag iops, + InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : VFPI<oops, iops, AddrMode5FP16, 4, IndexModeNone, + VFPLdStFrm, itin, opc, asm, "", pattern> { + list<Predicate> Predicates = [HasFullFP16]; + + // Instruction operands. + bits<5> Sd; + bits<13> addr; + + // Encode instruction operands. + let Inst{23} = addr{8}; // U (add = (U == '1')) + let Inst{22} = Sd{0}; + let Inst{19-16} = addr{12-9}; // Rn + let Inst{15-12} = Sd{4-1}; + let Inst{7-0} = addr{7-0}; // imm8 + + let Inst{27-24} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-8} = 0b1001; // Half precision + + // Loads & stores operate on both NEON and VFP pipelines. + let D = VFPNeonDomain; +} + +// VFP Load / store multiple pseudo instructions. +class PseudoVFPLdStM<dag oops, dag iops, InstrItinClass itin, string cstr, + list<dag> pattern> + : InstARM<AddrMode4, 4, IndexModeNone, Pseudo, VFPNeonDomain, + cstr, itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let Pattern = pattern; + list<Predicate> Predicates = [HasVFP2]; +} + +// Load / store multiple + +// Unknown precision +class AXXI4<dag oops, dag iops, IndexMode im, + string asm, string cstr, list<dag> pattern> + : VFPXI<oops, iops, AddrMode4, 4, im, + VFPLdStFrm, NoItinerary, asm, cstr, pattern> { + // Instruction operands. + bits<4> Rn; + bits<13> regs; + + // Encode instruction operands. + let Inst{19-16} = Rn; + let Inst{22} = 0; + let Inst{15-12} = regs{11-8}; + let Inst{7-1} = regs{7-1}; + + let Inst{27-25} = 0b110; + let Inst{11-8} = 0b1011; + let Inst{0} = 1; +} + +// Double precision +class AXDI4<dag oops, dag iops, IndexMode im, InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : VFPXI<oops, iops, AddrMode4, 4, im, + VFPLdStMulFrm, itin, asm, cstr, pattern> { + // Instruction operands. + bits<4> Rn; + bits<13> regs; + + // Encode instruction operands. + let Inst{19-16} = Rn; + let Inst{22} = regs{12}; + let Inst{15-12} = regs{11-8}; + let Inst{7-1} = regs{7-1}; + + let Inst{27-25} = 0b110; + let Inst{11-9} = 0b101; + let Inst{8} = 1; // Double precision + let Inst{0} = 0; +} + +// Single Precision +class AXSI4<dag oops, dag iops, IndexMode im, InstrItinClass itin, + string asm, string cstr, list<dag> pattern> + : VFPXI<oops, iops, AddrMode4, 4, im, + VFPLdStMulFrm, itin, asm, cstr, pattern> { + // Instruction operands. + bits<4> Rn; + bits<13> regs; + + // Encode instruction operands. + let Inst{19-16} = Rn; + let Inst{22} = regs{8}; + let Inst{15-12} = regs{12-9}; + let Inst{7-0} = regs{7-0}; + + let Inst{27-25} = 0b110; + let Inst{11-9} = 0b101; + let Inst{8} = 0; // Single precision +} + +// Double precision, unary +class ADuI<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<2> opcod4, + bit opcod5, dag oops, dag iops, InstrItinClass itin, string opc, + string asm, list<dag> pattern> + : VFPAI<oops, iops, VFPUnaryFrm, itin, opc, asm, pattern> { + // Instruction operands. + bits<5> Dd; + bits<5> Dm; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{15-12} = Dd{3-0}; + let Inst{22} = Dd{4}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{19-16} = opcod3; + let Inst{11-9} = 0b101; + let Inst{8} = 1; // Double precision + let Inst{7-6} = opcod4; + let Inst{4} = opcod5; + + let Predicates = [HasVFP2, HasDPVFP]; +} + +// Double precision, unary, not-predicated +class ADuInp<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<2> opcod4, + bit opcod5, dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : VFPXI<oops, iops, AddrModeNone, 4, IndexModeNone, VFPUnaryFrm, itin, asm, "", pattern> { + // Instruction operands. + bits<5> Dd; + bits<5> Dm; + + let Inst{31-28} = 0b1111; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{15-12} = Dd{3-0}; + let Inst{22} = Dd{4}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{19-16} = opcod3; + let Inst{11-9} = 0b101; + let Inst{8} = 1; // Double precision + let Inst{7-6} = opcod4; + let Inst{4} = opcod5; +} + +// Double precision, binary +class ADbI<bits<5> opcod1, bits<2> opcod2, bit op6, bit op4, dag oops, + dag iops, InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : VFPAI<oops, iops, VFPBinaryFrm, itin, opc, asm, pattern> { + // Instruction operands. + bits<5> Dd; + bits<5> Dn; + bits<5> Dm; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{19-16} = Dn{3-0}; + let Inst{7} = Dn{4}; + let Inst{15-12} = Dd{3-0}; + let Inst{22} = Dd{4}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-9} = 0b101; + let Inst{8} = 1; // Double precision + let Inst{6} = op6; + let Inst{4} = op4; + + let Predicates = [HasVFP2, HasDPVFP]; +} + +// FP, binary, not predicated +class ADbInp<bits<5> opcod1, bits<2> opcod2, bit opcod3, dag oops, dag iops, + InstrItinClass itin, string asm, list<dag> pattern> + : VFPXI<oops, iops, AddrModeNone, 4, IndexModeNone, VFPBinaryFrm, itin, + asm, "", pattern> +{ + // Instruction operands. + bits<5> Dd; + bits<5> Dn; + bits<5> Dm; + + let Inst{31-28} = 0b1111; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{19-16} = Dn{3-0}; + let Inst{7} = Dn{4}; + let Inst{15-12} = Dd{3-0}; + let Inst{22} = Dd{4}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-9} = 0b101; + let Inst{8} = 1; // double precision + let Inst{6} = opcod3; + let Inst{4} = 0; + + let Predicates = [HasVFP2, HasDPVFP]; +} + +// Single precision, unary, predicated +class ASuI<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<2> opcod4, + bit opcod5, dag oops, dag iops, InstrItinClass itin, string opc, + string asm, list<dag> pattern> + : VFPAI<oops, iops, VFPUnaryFrm, itin, opc, asm, pattern> { + // Instruction operands. + bits<5> Sd; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{19-16} = opcod3; + let Inst{11-9} = 0b101; + let Inst{8} = 0; // Single precision + let Inst{7-6} = opcod4; + let Inst{4} = opcod5; +} + +// Single precision, unary, non-predicated +class ASuInp<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<2> opcod4, + bit opcod5, dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : VFPXI<oops, iops, AddrModeNone, 4, IndexModeNone, + VFPUnaryFrm, itin, asm, "", pattern> { + // Instruction operands. + bits<5> Sd; + bits<5> Sm; + + let Inst{31-28} = 0b1111; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{19-16} = opcod3; + let Inst{11-9} = 0b101; + let Inst{8} = 0; // Single precision + let Inst{7-6} = opcod4; + let Inst{4} = opcod5; +} + +// Single precision unary, if no NEON. Same as ASuI except not available if +// NEON is enabled. +class ASuIn<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<2> opcod4, + bit opcod5, dag oops, dag iops, InstrItinClass itin, string opc, + string asm, list<dag> pattern> + : ASuI<opcod1, opcod2, opcod3, opcod4, opcod5, oops, iops, itin, opc, asm, + pattern> { + list<Predicate> Predicates = [HasVFP2,DontUseNEONForFP]; +} + +// Single precision, binary +class ASbI<bits<5> opcod1, bits<2> opcod2, bit op6, bit op4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : VFPAI<oops, iops, VFPBinaryFrm, itin, opc, asm, pattern> { + // Instruction operands. + bits<5> Sd; + bits<5> Sn; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-9} = 0b101; + let Inst{8} = 0; // Single precision + let Inst{6} = op6; + let Inst{4} = op4; +} + +// Single precision, binary, not predicated +class ASbInp<bits<5> opcod1, bits<2> opcod2, bit opcod3, dag oops, dag iops, + InstrItinClass itin, string asm, list<dag> pattern> + : VFPXI<oops, iops, AddrModeNone, 4, IndexModeNone, + VFPBinaryFrm, itin, asm, "", pattern> +{ + // Instruction operands. + bits<5> Sd; + bits<5> Sn; + bits<5> Sm; + + let Inst{31-28} = 0b1111; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-9} = 0b101; + let Inst{8} = 0; // Single precision + let Inst{6} = opcod3; + let Inst{4} = 0; +} + +// Single precision binary, if no NEON. Same as ASbI except not available if +// NEON is enabled. +class ASbIn<bits<5> opcod1, bits<2> opcod2, bit op6, bit op4, dag oops, + dag iops, InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : ASbI<opcod1, opcod2, op6, op4, oops, iops, itin, opc, asm, pattern> { + list<Predicate> Predicates = [HasVFP2,DontUseNEONForFP]; + + // Instruction operands. + bits<5> Sd; + bits<5> Sn; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; +} + +// Half precision, unary, predicated +class AHuI<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<2> opcod4, + bit opcod5, dag oops, dag iops, InstrItinClass itin, string opc, + string asm, list<dag> pattern> + : VFPAI<oops, iops, VFPUnaryFrm, itin, opc, asm, pattern> { + list<Predicate> Predicates = [HasFullFP16]; + + // Instruction operands. + bits<5> Sd; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{19-16} = opcod3; + let Inst{11-8} = 0b1001; // Half precision + let Inst{7-6} = opcod4; + let Inst{4} = opcod5; +} + +// Half precision, unary, non-predicated +class AHuInp<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<2> opcod4, + bit opcod5, dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : VFPXI<oops, iops, AddrModeNone, 4, IndexModeNone, + VFPUnaryFrm, itin, asm, "", pattern> { + list<Predicate> Predicates = [HasFullFP16]; + + // Instruction operands. + bits<5> Sd; + bits<5> Sm; + + let Inst{31-28} = 0b1111; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{19-16} = opcod3; + let Inst{11-8} = 0b1001; // Half precision + let Inst{7-6} = opcod4; + let Inst{4} = opcod5; +} + +// Half precision, binary +class AHbI<bits<5> opcod1, bits<2> opcod2, bit op6, bit op4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : VFPAI<oops, iops, VFPBinaryFrm, itin, opc, asm, pattern> { + list<Predicate> Predicates = [HasFullFP16]; + + // Instruction operands. + bits<5> Sd; + bits<5> Sn; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-8} = 0b1001; // Half precision + let Inst{6} = op6; + let Inst{4} = op4; +} + +// Half precision, binary, not predicated +class AHbInp<bits<5> opcod1, bits<2> opcod2, bit opcod3, dag oops, dag iops, + InstrItinClass itin, string asm, list<dag> pattern> + : VFPXI<oops, iops, AddrModeNone, 4, IndexModeNone, + VFPBinaryFrm, itin, asm, "", pattern> { + list<Predicate> Predicates = [HasFullFP16]; + + // Instruction operands. + bits<5> Sd; + bits<5> Sn; + bits<5> Sm; + + let Inst{31-28} = 0b1111; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{11-8} = 0b1001; // Half precision + let Inst{6} = opcod3; + let Inst{4} = 0; +} + +// VFP conversion instructions +class AVConv1I<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<4> opcod4, + dag oops, dag iops, InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : VFPAI<oops, iops, VFPConv1Frm, itin, opc, asm, pattern> { + let Inst{27-23} = opcod1; + let Inst{21-20} = opcod2; + let Inst{19-16} = opcod3; + let Inst{11-8} = opcod4; + let Inst{6} = 1; + let Inst{4} = 0; +} + +// VFP conversion between floating-point and fixed-point +class AVConv1XI<bits<5> op1, bits<2> op2, bits<4> op3, bits<4> op4, bit op5, + dag oops, dag iops, InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : AVConv1I<op1, op2, op3, op4, oops, iops, itin, opc, asm, pattern> { + bits<5> fbits; + // size (fixed-point number): sx == 0 ? 16 : 32 + let Inst{7} = op5; // sx + let Inst{5} = fbits{0}; + let Inst{3-0} = fbits{4-1}; +} + +// VFP conversion instructions, if no NEON +class AVConv1In<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, bits<4> opcod4, + dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AVConv1I<opcod1, opcod2, opcod3, opcod4, oops, iops, itin, opc, asm, + pattern> { + list<Predicate> Predicates = [HasVFP2,DontUseNEONForFP]; +} + +class AVConvXI<bits<8> opcod1, bits<4> opcod2, dag oops, dag iops, Format f, + InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : VFPAI<oops, iops, f, itin, opc, asm, pattern> { + let Inst{27-20} = opcod1; + let Inst{11-8} = opcod2; + let Inst{4} = 1; +} + +class AVConv2I<bits<8> opcod1, bits<4> opcod2, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : AVConvXI<opcod1, opcod2, oops, iops, VFPConv2Frm, itin, opc, asm, pattern>; + +class AVConv3I<bits<8> opcod1, bits<4> opcod2, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : AVConvXI<opcod1, opcod2, oops, iops, VFPConv3Frm, itin, opc, asm, pattern>; + +class AVConv4I<bits<8> opcod1, bits<4> opcod2, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : AVConvXI<opcod1, opcod2, oops, iops, VFPConv4Frm, itin, opc, asm, pattern>; + +class AVConv5I<bits<8> opcod1, bits<4> opcod2, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : AVConvXI<opcod1, opcod2, oops, iops, VFPConv5Frm, itin, opc, asm, pattern>; + +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// ARM NEON Instruction templates. +// + +class NeonI<dag oops, dag iops, AddrMode am, IndexMode im, Format f, + InstrItinClass itin, string opc, string dt, string asm, string cstr, + list<dag> pattern> + : InstARM<am, 4, im, f, NeonDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", ".", dt, "\t", asm); + let Pattern = pattern; + list<Predicate> Predicates = [HasNEON]; + let DecoderNamespace = "NEON"; +} + +// Same as NeonI except it does not have a "data type" specifier. +class NeonXI<dag oops, dag iops, AddrMode am, IndexMode im, Format f, + InstrItinClass itin, string opc, string asm, string cstr, + list<dag> pattern> + : InstARM<am, 4, im, f, NeonDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", "\t", asm); + let Pattern = pattern; + list<Predicate> Predicates = [HasNEON]; + let DecoderNamespace = "NEON"; +} + +// Same as NeonI except it is not predicated +class NeonInp<dag oops, dag iops, AddrMode am, IndexMode im, Format f, + InstrItinClass itin, string opc, string dt, string asm, string cstr, + list<dag> pattern> + : InstARM<am, 4, im, f, NeonDomain, cstr, itin> { + let OutOperandList = oops; + let InOperandList = iops; + let AsmString = !strconcat(opc, ".", dt, "\t", asm); + let Pattern = pattern; + list<Predicate> Predicates = [HasNEON]; + let DecoderNamespace = "NEON"; + + let Inst{31-28} = 0b1111; +} + +class NLdSt<bit op23, bits<2> op21_20, bits<4> op11_8, bits<4> op7_4, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, string cstr, list<dag> pattern> + : NeonI<oops, iops, AddrMode6, IndexModeNone, NLdStFrm, itin, opc, dt, asm, + cstr, pattern> { + let Inst{31-24} = 0b11110100; + let Inst{23} = op23; + let Inst{21-20} = op21_20; + let Inst{11-8} = op11_8; + let Inst{7-4} = op7_4; + + let PostEncoderMethod = "NEONThumb2LoadStorePostEncoder"; + let DecoderNamespace = "NEONLoadStore"; + + bits<5> Vd; + bits<6> Rn; + bits<4> Rm; + + let Inst{22} = Vd{4}; + let Inst{15-12} = Vd{3-0}; + let Inst{19-16} = Rn{3-0}; + let Inst{3-0} = Rm{3-0}; +} + +class NLdStLn<bit op23, bits<2> op21_20, bits<4> op11_8, bits<4> op7_4, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, string cstr, list<dag> pattern> + : NLdSt<op23, op21_20, op11_8, op7_4, oops, iops, itin, opc, + dt, asm, cstr, pattern> { + bits<3> lane; +} + +class PseudoNLdSt<dag oops, dag iops, InstrItinClass itin, string cstr> + : InstARM<AddrMode6, 4, IndexModeNone, Pseudo, NeonDomain, cstr, + itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + list<Predicate> Predicates = [HasNEON]; +} + +class PseudoNeonI<dag oops, dag iops, InstrItinClass itin, string cstr, + list<dag> pattern> + : InstARM<AddrModeNone, 4, IndexModeNone, Pseudo, NeonDomain, cstr, + itin> { + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let Pattern = pattern; + list<Predicate> Predicates = [HasNEON]; +} + +class NDataI<dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string dt, string asm, string cstr, list<dag> pattern> + : NeonI<oops, iops, AddrModeNone, IndexModeNone, f, itin, opc, dt, asm, cstr, + pattern> { + let Inst{31-25} = 0b1111001; + let PostEncoderMethod = "NEONThumb2DataIPostEncoder"; + let DecoderNamespace = "NEONData"; +} + +class NDataXI<dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : NeonXI<oops, iops, AddrModeNone, IndexModeNone, f, itin, opc, asm, + cstr, pattern> { + let Inst{31-25} = 0b1111001; + let PostEncoderMethod = "NEONThumb2DataIPostEncoder"; + let DecoderNamespace = "NEONData"; +} + +// NEON "one register and a modified immediate" format. +class N1ModImm<bit op23, bits<3> op21_19, bits<4> op11_8, bit op7, bit op6, + bit op5, bit op4, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, string cstr, + list<dag> pattern> + : NDataI<oops, iops, N1RegModImmFrm, itin, opc, dt, asm, cstr, pattern> { + let Inst{23} = op23; + let Inst{21-19} = op21_19; + let Inst{11-8} = op11_8; + let Inst{7} = op7; + let Inst{6} = op6; + let Inst{5} = op5; + let Inst{4} = op4; + + // Instruction operands. + bits<5> Vd; + bits<13> SIMM; + + let Inst{15-12} = Vd{3-0}; + let Inst{22} = Vd{4}; + let Inst{24} = SIMM{7}; + let Inst{18-16} = SIMM{6-4}; + let Inst{3-0} = SIMM{3-0}; + let DecoderMethod = "DecodeNEONModImmInstruction"; +} + +// NEON 2 vector register format. +class N2V<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, bits<2> op17_16, + bits<5> op11_7, bit op6, bit op4, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, string cstr, list<dag> pattern> + : NDataI<oops, iops, N2RegFrm, itin, opc, dt, asm, cstr, pattern> { + let Inst{24-23} = op24_23; + let Inst{21-20} = op21_20; + let Inst{19-18} = op19_18; + let Inst{17-16} = op17_16; + let Inst{11-7} = op11_7; + let Inst{6} = op6; + let Inst{4} = op4; + + // Instruction operands. + bits<5> Vd; + bits<5> Vm; + + let Inst{15-12} = Vd{3-0}; + let Inst{22} = Vd{4}; + let Inst{3-0} = Vm{3-0}; + let Inst{5} = Vm{4}; +} + +// Same as N2V but not predicated. +class N2Vnp<bits<2> op19_18, bits<2> op17_16, bits<3> op10_8, bit op7, bit op6, + dag oops, dag iops, InstrItinClass itin, string OpcodeStr, + string Dt, list<dag> pattern> + : NeonInp<oops, iops, AddrModeNone, IndexModeNone, N2RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vm", "", pattern> { + bits<5> Vd; + bits<5> Vm; + + // Encode instruction operands + let Inst{22} = Vd{4}; + let Inst{15-12} = Vd{3-0}; + let Inst{5} = Vm{4}; + let Inst{3-0} = Vm{3-0}; + + // Encode constant bits + let Inst{27-23} = 0b00111; + let Inst{21-20} = 0b11; + let Inst{19-18} = op19_18; + let Inst{17-16} = op17_16; + let Inst{11} = 0; + let Inst{10-8} = op10_8; + let Inst{7} = op7; + let Inst{6} = op6; + let Inst{4} = 0; + + let DecoderNamespace = "NEON"; +} + +// Same as N2V except it doesn't have a datatype suffix. +class N2VX<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, bits<2> op17_16, + bits<5> op11_7, bit op6, bit op4, + dag oops, dag iops, InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : NDataXI<oops, iops, N2RegFrm, itin, opc, asm, cstr, pattern> { + let Inst{24-23} = op24_23; + let Inst{21-20} = op21_20; + let Inst{19-18} = op19_18; + let Inst{17-16} = op17_16; + let Inst{11-7} = op11_7; + let Inst{6} = op6; + let Inst{4} = op4; + + // Instruction operands. + bits<5> Vd; + bits<5> Vm; + + let Inst{15-12} = Vd{3-0}; + let Inst{22} = Vd{4}; + let Inst{3-0} = Vm{3-0}; + let Inst{5} = Vm{4}; +} + +// NEON 2 vector register with immediate. +class N2VImm<bit op24, bit op23, bits<4> op11_8, bit op7, bit op6, bit op4, + dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string dt, string asm, string cstr, list<dag> pattern> + : NDataI<oops, iops, f, itin, opc, dt, asm, cstr, pattern> { + let Inst{24} = op24; + let Inst{23} = op23; + let Inst{11-8} = op11_8; + let Inst{7} = op7; + let Inst{6} = op6; + let Inst{4} = op4; + + // Instruction operands. + bits<5> Vd; + bits<5> Vm; + bits<6> SIMM; + + let Inst{15-12} = Vd{3-0}; + let Inst{22} = Vd{4}; + let Inst{3-0} = Vm{3-0}; + let Inst{5} = Vm{4}; + let Inst{21-16} = SIMM{5-0}; +} + +// NEON 3 vector register format. + +class N3VCommon<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string dt, string asm, string cstr, + list<dag> pattern> + : NDataI<oops, iops, f, itin, opc, dt, asm, cstr, pattern> { + let Inst{24} = op24; + let Inst{23} = op23; + let Inst{21-20} = op21_20; + let Inst{11-8} = op11_8; + let Inst{6} = op6; + let Inst{4} = op4; +} + +class N3V<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op6, bit op4, + dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string dt, string asm, string cstr, list<dag> pattern> + : N3VCommon<op24, op23, op21_20, op11_8, op6, op4, + oops, iops, f, itin, opc, dt, asm, cstr, pattern> { + // Instruction operands. + bits<5> Vd; + bits<5> Vn; + bits<5> Vm; + + let Inst{15-12} = Vd{3-0}; + let Inst{22} = Vd{4}; + let Inst{19-16} = Vn{3-0}; + let Inst{7} = Vn{4}; + let Inst{3-0} = Vm{3-0}; + let Inst{5} = Vm{4}; +} + +class N3Vnp<bits<5> op27_23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, dag oops, dag iops,Format f, InstrItinClass itin, + string OpcodeStr, string Dt, list<dag> pattern> + : NeonInp<oops, iops, AddrModeNone, IndexModeNone, f, itin, OpcodeStr, + Dt, "$Vd, $Vn, $Vm", "", pattern> { + bits<5> Vd; + bits<5> Vn; + bits<5> Vm; + + // Encode instruction operands + let Inst{22} = Vd{4}; + let Inst{15-12} = Vd{3-0}; + let Inst{19-16} = Vn{3-0}; + let Inst{7} = Vn{4}; + let Inst{5} = Vm{4}; + let Inst{3-0} = Vm{3-0}; + + // Encode constant bits + let Inst{27-23} = op27_23; + let Inst{21-20} = op21_20; + let Inst{11-8} = op11_8; + let Inst{6} = op6; + let Inst{4} = op4; +} + +class N3VLane32<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string dt, string asm, string cstr, + list<dag> pattern> + : N3VCommon<op24, op23, op21_20, op11_8, op6, op4, + oops, iops, f, itin, opc, dt, asm, cstr, pattern> { + + // Instruction operands. + bits<5> Vd; + bits<5> Vn; + bits<5> Vm; + bit lane; + + let Inst{15-12} = Vd{3-0}; + let Inst{22} = Vd{4}; + let Inst{19-16} = Vn{3-0}; + let Inst{7} = Vn{4}; + let Inst{3-0} = Vm{3-0}; + let Inst{5} = lane; +} + +class N3VLane16<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string dt, string asm, string cstr, + list<dag> pattern> + : N3VCommon<op24, op23, op21_20, op11_8, op6, op4, + oops, iops, f, itin, opc, dt, asm, cstr, pattern> { + + // Instruction operands. + bits<5> Vd; + bits<5> Vn; + bits<5> Vm; + bits<2> lane; + + let Inst{15-12} = Vd{3-0}; + let Inst{22} = Vd{4}; + let Inst{19-16} = Vn{3-0}; + let Inst{7} = Vn{4}; + let Inst{2-0} = Vm{2-0}; + let Inst{5} = lane{1}; + let Inst{3} = lane{0}; +} + +// Same as N3V except it doesn't have a data type suffix. +class N3VX<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, + dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string asm, string cstr, list<dag> pattern> + : NDataXI<oops, iops, f, itin, opc, asm, cstr, pattern> { + let Inst{24} = op24; + let Inst{23} = op23; + let Inst{21-20} = op21_20; + let Inst{11-8} = op11_8; + let Inst{6} = op6; + let Inst{4} = op4; + + // Instruction operands. + bits<5> Vd; + bits<5> Vn; + bits<5> Vm; + + let Inst{15-12} = Vd{3-0}; + let Inst{22} = Vd{4}; + let Inst{19-16} = Vn{3-0}; + let Inst{7} = Vn{4}; + let Inst{3-0} = Vm{3-0}; + let Inst{5} = Vm{4}; +} + +// NEON VMOVs between scalar and core registers. +class NVLaneOp<bits<8> opcod1, bits<4> opcod2, bits<2> opcod3, + dag oops, dag iops, Format f, InstrItinClass itin, + string opc, string dt, string asm, list<dag> pattern> + : InstARM<AddrModeNone, 4, IndexModeNone, f, NeonDomain, + "", itin> { + let Inst{27-20} = opcod1; + let Inst{11-8} = opcod2; + let Inst{6-5} = opcod3; + let Inst{4} = 1; + // A8.6.303, A8.6.328, A8.6.329 + let Inst{3-0} = 0b0000; + + let OutOperandList = oops; + let InOperandList = !con(iops, (ins pred:$p)); + let AsmString = !strconcat(opc, "${p}", ".", dt, "\t", asm); + let Pattern = pattern; + list<Predicate> Predicates = [HasNEON]; + + let PostEncoderMethod = "NEONThumb2DupPostEncoder"; + let DecoderNamespace = "NEONDup"; + + bits<5> V; + bits<4> R; + bits<4> p; + bits<4> lane; + + let Inst{31-28} = p{3-0}; + let Inst{7} = V{4}; + let Inst{19-16} = V{3-0}; + let Inst{15-12} = R{3-0}; +} +class NVGetLane<bits<8> opcod1, bits<4> opcod2, bits<2> opcod3, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, list<dag> pattern> + : NVLaneOp<opcod1, opcod2, opcod3, oops, iops, NGetLnFrm, itin, + opc, dt, asm, pattern>; +class NVSetLane<bits<8> opcod1, bits<4> opcod2, bits<2> opcod3, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, list<dag> pattern> + : NVLaneOp<opcod1, opcod2, opcod3, oops, iops, NSetLnFrm, itin, + opc, dt, asm, pattern>; +class NVDup<bits<8> opcod1, bits<4> opcod2, bits<2> opcod3, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, list<dag> pattern> + : NVLaneOp<opcod1, opcod2, opcod3, oops, iops, NDupFrm, itin, + opc, dt, asm, pattern>; + +// Vector Duplicate Lane (from scalar to all elements) +class NVDupLane<bits<4> op19_16, bit op6, dag oops, dag iops, + InstrItinClass itin, string opc, string dt, string asm, + list<dag> pattern> + : NDataI<oops, iops, NVDupLnFrm, itin, opc, dt, asm, "", pattern> { + let Inst{24-23} = 0b11; + let Inst{21-20} = 0b11; + let Inst{19-16} = op19_16; + let Inst{11-7} = 0b11000; + let Inst{6} = op6; + let Inst{4} = 0; + + bits<5> Vd; + bits<5> Vm; + + let Inst{22} = Vd{4}; + let Inst{15-12} = Vd{3-0}; + let Inst{5} = Vm{4}; + let Inst{3-0} = Vm{3-0}; +} + +// NEONFPPat - Same as Pat<>, but requires that the compiler be using NEON +// for single-precision FP. +class NEONFPPat<dag pattern, dag result> : Pat<pattern, result> { + list<Predicate> Predicates = [HasNEON,UseNEONForFP]; +} + +// VFP/NEON Instruction aliases for type suffices. +// Note: When EmitPriority == 1, the alias will be used for printing +class VFPDataTypeInstAlias<string opc, string dt, string asm, dag Result, bit EmitPriority = 0> : + InstAlias<!strconcat(opc, dt, "\t", asm), Result, EmitPriority>, Requires<[HasVFP2]>; + +// Note: When EmitPriority == 1, the alias will be used for printing +multiclass VFPDTAnyInstAlias<string opc, string asm, dag Result, bit EmitPriority = 0> { + def : VFPDataTypeInstAlias<opc, ".8", asm, Result, EmitPriority>; + def : VFPDataTypeInstAlias<opc, ".16", asm, Result, EmitPriority>; + def : VFPDataTypeInstAlias<opc, ".32", asm, Result, EmitPriority>; + def : VFPDataTypeInstAlias<opc, ".64", asm, Result, EmitPriority>; +} + +// Note: When EmitPriority == 1, the alias will be used for printing +multiclass NEONDTAnyInstAlias<string opc, string asm, dag Result, bit EmitPriority = 0> { + let Predicates = [HasNEON] in { + def : VFPDataTypeInstAlias<opc, ".8", asm, Result, EmitPriority>; + def : VFPDataTypeInstAlias<opc, ".16", asm, Result, EmitPriority>; + def : VFPDataTypeInstAlias<opc, ".32", asm, Result, EmitPriority>; + def : VFPDataTypeInstAlias<opc, ".64", asm, Result, EmitPriority>; +} +} + +// The same alias classes using AsmPseudo instead, for the more complex +// stuff in NEON that InstAlias can't quite handle. +// Note that we can't use anonymous defm references here like we can +// above, as we care about the ultimate instruction enum names generated, unlike +// for instalias defs. +class NEONDataTypeAsmPseudoInst<string opc, string dt, string asm, dag iops> : + AsmPseudoInst<!strconcat(opc, dt, "\t", asm), iops>, Requires<[HasNEON]>; + +// Extension of NEON 3-vector data processing instructions in coprocessor 8 +// encoding space, introduced in ARMv8.3-A. +class N3VCP8<bits<2> op24_23, bits<2> op21_20, bit op6, bit op4, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, string cstr, list<dag> pattern> + : NeonInp<oops, iops, AddrModeNone, IndexModeNone, N3RegCplxFrm, itin, opc, + dt, asm, cstr, pattern> { + bits<5> Vd; + bits<5> Vn; + bits<5> Vm; + + let DecoderNamespace = "VFPV8"; + // These have the same encodings in ARM and Thumb2 + let PostEncoderMethod = ""; + + let Inst{31-25} = 0b1111110; + let Inst{24-23} = op24_23; + let Inst{22} = Vd{4}; + let Inst{21-20} = op21_20; + let Inst{19-16} = Vn{3-0}; + let Inst{15-12} = Vd{3-0}; + let Inst{11-8} = 0b1000; + let Inst{7} = Vn{4}; + let Inst{6} = op6; + let Inst{5} = Vm{4}; + let Inst{4} = op4; + let Inst{3-0} = Vm{3-0}; +} + +// Extension of NEON 2-vector-and-scalar data processing instructions in +// coprocessor 8 encoding space, introduced in ARMv8.3-A. +class N3VLaneCP8<bit op23, bits<2> op21_20, bit op6, bit op4, + dag oops, dag iops, InstrItinClass itin, + string opc, string dt, string asm, string cstr, list<dag> pattern> + : NeonInp<oops, iops, AddrModeNone, IndexModeNone, N3RegCplxFrm, itin, opc, + dt, asm, cstr, pattern> { + bits<5> Vd; + bits<5> Vn; + bits<5> Vm; + + let DecoderNamespace = "VFPV8"; + // These have the same encodings in ARM and Thumb2 + let PostEncoderMethod = ""; + + let Inst{31-24} = 0b11111110; + let Inst{23} = op23; + let Inst{22} = Vd{4}; + let Inst{21-20} = op21_20; + let Inst{19-16} = Vn{3-0}; + let Inst{15-12} = Vd{3-0}; + let Inst{11-8} = 0b1000; + let Inst{7} = Vn{4}; + let Inst{6} = op6; + // Bit 5 set by sub-classes + let Inst{4} = op4; + let Inst{3-0} = Vm{3-0}; +} + +// Operand types for complex instructions +class ComplexRotationOperand<int Angle, int Remainder, string Type, string Diag> + : AsmOperandClass { + let PredicateMethod = "isComplexRotation<" # Angle # ", " # Remainder # ">"; + let DiagnosticString = "complex rotation must be " # Diag; + let Name = "ComplexRotation" # Type; +} +def complexrotateop : Operand<i32> { + let ParserMatchClass = ComplexRotationOperand<90, 0, "Even", "0, 90, 180 or 270">; + let PrintMethod = "printComplexRotationOp<90, 0>"; +} +def complexrotateopodd : Operand<i32> { + let ParserMatchClass = ComplexRotationOperand<180, 90, "Odd", "90 or 270">; + let PrintMethod = "printComplexRotationOp<180, 90>"; +} + +// Data type suffix token aliases. Implements Table A7-3 in the ARM ARM. +def : TokenAlias<".s8", ".i8">; +def : TokenAlias<".u8", ".i8">; +def : TokenAlias<".s16", ".i16">; +def : TokenAlias<".u16", ".i16">; +def : TokenAlias<".s32", ".i32">; +def : TokenAlias<".u32", ".i32">; +def : TokenAlias<".s64", ".i64">; +def : TokenAlias<".u64", ".i64">; + +def : TokenAlias<".i8", ".8">; +def : TokenAlias<".i16", ".16">; +def : TokenAlias<".i32", ".32">; +def : TokenAlias<".i64", ".64">; + +def : TokenAlias<".p8", ".8">; +def : TokenAlias<".p16", ".16">; + +def : TokenAlias<".f32", ".32">; +def : TokenAlias<".f64", ".64">; +def : TokenAlias<".f", ".f32">; +def : TokenAlias<".d", ".f64">; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMInstrInfo.td b/capstone/suite/synctools/tablegen/ARM/ARMInstrInfo.td new file mode 100644 index 000000000..d4c342cee --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMInstrInfo.td @@ -0,0 +1,6167 @@ +//===- ARMInstrInfo.td - Target Description for ARM Target -*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file describes the ARM instructions in TableGen format. +// +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// ARM specific DAG Nodes. +// + +// Type profiles. +def SDT_ARMCallSeqStart : SDCallSeqStart<[ SDTCisVT<0, i32>, + SDTCisVT<1, i32> ]>; +def SDT_ARMCallSeqEnd : SDCallSeqEnd<[ SDTCisVT<0, i32>, SDTCisVT<1, i32> ]>; +def SDT_ARMStructByVal : SDTypeProfile<0, 4, + [SDTCisVT<0, i32>, SDTCisVT<1, i32>, + SDTCisVT<2, i32>, SDTCisVT<3, i32>]>; + +def SDT_ARMSaveCallPC : SDTypeProfile<0, 1, []>; + +def SDT_ARMcall : SDTypeProfile<0, -1, [SDTCisPtrTy<0>]>; + +def SDT_ARMCMov : SDTypeProfile<1, 3, + [SDTCisSameAs<0, 1>, SDTCisSameAs<0, 2>, + SDTCisVT<3, i32>]>; + +def SDT_ARMBrcond : SDTypeProfile<0, 2, + [SDTCisVT<0, OtherVT>, SDTCisVT<1, i32>]>; + +def SDT_ARMBrJT : SDTypeProfile<0, 2, + [SDTCisPtrTy<0>, SDTCisVT<1, i32>]>; + +def SDT_ARMBr2JT : SDTypeProfile<0, 3, + [SDTCisPtrTy<0>, SDTCisVT<1, i32>, + SDTCisVT<2, i32>]>; + +def SDT_ARMBCC_i64 : SDTypeProfile<0, 6, + [SDTCisVT<0, i32>, + SDTCisVT<1, i32>, SDTCisVT<2, i32>, + SDTCisVT<3, i32>, SDTCisVT<4, i32>, + SDTCisVT<5, OtherVT>]>; + +def SDT_ARMAnd : SDTypeProfile<1, 2, + [SDTCisVT<0, i32>, SDTCisVT<1, i32>, + SDTCisVT<2, i32>]>; + +def SDT_ARMCmp : SDTypeProfile<0, 2, [SDTCisSameAs<0, 1>]>; +def SDT_ARMFCmp : SDTypeProfile<0, 3, [SDTCisSameAs<0, 1>, + SDTCisVT<2, i32>]>; + +def SDT_ARMPICAdd : SDTypeProfile<1, 2, [SDTCisSameAs<0, 1>, + SDTCisPtrTy<1>, SDTCisVT<2, i32>]>; + +def SDT_ARMThreadPointer : SDTypeProfile<1, 0, [SDTCisPtrTy<0>]>; +def SDT_ARMEH_SJLJ_Setjmp : SDTypeProfile<1, 2, [SDTCisInt<0>, SDTCisPtrTy<1>, + SDTCisInt<2>]>; +def SDT_ARMEH_SJLJ_Longjmp: SDTypeProfile<0, 2, [SDTCisPtrTy<0>, SDTCisInt<1>]>; +def SDT_ARMEH_SJLJ_SetupDispatch: SDTypeProfile<0, 0, []>; + +def SDT_ARMMEMBARRIER : SDTypeProfile<0, 1, [SDTCisInt<0>]>; + +def SDT_ARMPREFETCH : SDTypeProfile<0, 3, [SDTCisPtrTy<0>, SDTCisSameAs<1, 2>, + SDTCisInt<1>]>; + +def SDT_ARMTCRET : SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>; + +def SDT_ARMBFI : SDTypeProfile<1, 3, [SDTCisVT<0, i32>, SDTCisVT<1, i32>, + SDTCisVT<2, i32>, SDTCisVT<3, i32>]>; + +def SDT_WIN__DBZCHK : SDTypeProfile<0, 1, [SDTCisVT<0, i32>]>; + +def SDT_ARMMEMCPY : SDTypeProfile<2, 3, [SDTCisVT<0, i32>, SDTCisVT<1, i32>, + SDTCisVT<2, i32>, SDTCisVT<3, i32>, + SDTCisVT<4, i32>]>; + +def SDTBinaryArithWithFlags : SDTypeProfile<2, 2, + [SDTCisSameAs<0, 2>, + SDTCisSameAs<0, 3>, + SDTCisInt<0>, SDTCisVT<1, i32>]>; + +// SDTBinaryArithWithFlagsInOut - RES1, CPSR = op LHS, RHS, CPSR +def SDTBinaryArithWithFlagsInOut : SDTypeProfile<2, 3, + [SDTCisSameAs<0, 2>, + SDTCisSameAs<0, 3>, + SDTCisInt<0>, + SDTCisVT<1, i32>, + SDTCisVT<4, i32>]>; + +def SDT_LongMac : SDTypeProfile<2, 4, [SDTCisVT<0, i32>, + SDTCisSameAs<0, 1>, + SDTCisSameAs<0, 2>, + SDTCisSameAs<0, 3>, + SDTCisSameAs<0, 4>, + SDTCisSameAs<0, 5>]>; + +def ARMSmlald : SDNode<"ARMISD::SMLALD", SDT_LongMac>; +def ARMSmlaldx : SDNode<"ARMISD::SMLALDX", SDT_LongMac>; +def ARMSmlsld : SDNode<"ARMISD::SMLSLD", SDT_LongMac>; +def ARMSmlsldx : SDNode<"ARMISD::SMLSLDX", SDT_LongMac>; + +def SDT_MulHSR : SDTypeProfile<1, 3, [SDTCisVT<0,i32>, + SDTCisSameAs<0, 1>, + SDTCisSameAs<0, 2>, + SDTCisSameAs<0, 3>]>; + +def ARMsmmlar : SDNode<"ARMISD::SMMLAR", SDT_MulHSR>; +def ARMsmmlsr : SDNode<"ARMISD::SMMLSR", SDT_MulHSR>; + +// Node definitions. +def ARMWrapper : SDNode<"ARMISD::Wrapper", SDTIntUnaryOp>; +def ARMWrapperPIC : SDNode<"ARMISD::WrapperPIC", SDTIntUnaryOp>; +def ARMWrapperJT : SDNode<"ARMISD::WrapperJT", SDTIntUnaryOp>; + +def ARMcallseq_start : SDNode<"ISD::CALLSEQ_START", SDT_ARMCallSeqStart, + [SDNPHasChain, SDNPSideEffect, SDNPOutGlue]>; +def ARMcallseq_end : SDNode<"ISD::CALLSEQ_END", SDT_ARMCallSeqEnd, + [SDNPHasChain, SDNPSideEffect, + SDNPOptInGlue, SDNPOutGlue]>; +def ARMcopystructbyval : SDNode<"ARMISD::COPY_STRUCT_BYVAL" , + SDT_ARMStructByVal, + [SDNPHasChain, SDNPInGlue, SDNPOutGlue, + SDNPMayStore, SDNPMayLoad]>; + +def ARMcall : SDNode<"ARMISD::CALL", SDT_ARMcall, + [SDNPHasChain, SDNPOptInGlue, SDNPOutGlue, + SDNPVariadic]>; +def ARMcall_pred : SDNode<"ARMISD::CALL_PRED", SDT_ARMcall, + [SDNPHasChain, SDNPOptInGlue, SDNPOutGlue, + SDNPVariadic]>; +def ARMcall_nolink : SDNode<"ARMISD::CALL_NOLINK", SDT_ARMcall, + [SDNPHasChain, SDNPOptInGlue, SDNPOutGlue, + SDNPVariadic]>; + +def ARMretflag : SDNode<"ARMISD::RET_FLAG", SDTNone, + [SDNPHasChain, SDNPOptInGlue, SDNPVariadic]>; +def ARMintretflag : SDNode<"ARMISD::INTRET_FLAG", SDT_ARMcall, + [SDNPHasChain, SDNPOptInGlue, SDNPVariadic]>; +def ARMcmov : SDNode<"ARMISD::CMOV", SDT_ARMCMov, + [SDNPInGlue]>; + +def ARMssatnoshift : SDNode<"ARMISD::SSAT", SDTIntSatNoShOp, []>; + +def ARMusatnoshift : SDNode<"ARMISD::USAT", SDTIntSatNoShOp, []>; + +def ARMbrcond : SDNode<"ARMISD::BRCOND", SDT_ARMBrcond, + [SDNPHasChain, SDNPInGlue, SDNPOutGlue]>; + +def ARMbrjt : SDNode<"ARMISD::BR_JT", SDT_ARMBrJT, + [SDNPHasChain]>; +def ARMbr2jt : SDNode<"ARMISD::BR2_JT", SDT_ARMBr2JT, + [SDNPHasChain]>; + +def ARMBcci64 : SDNode<"ARMISD::BCC_i64", SDT_ARMBCC_i64, + [SDNPHasChain]>; + +def ARMcmp : SDNode<"ARMISD::CMP", SDT_ARMCmp, + [SDNPOutGlue]>; + +def ARMcmn : SDNode<"ARMISD::CMN", SDT_ARMCmp, + [SDNPOutGlue]>; + +def ARMcmpZ : SDNode<"ARMISD::CMPZ", SDT_ARMCmp, + [SDNPOutGlue, SDNPCommutative]>; + +def ARMpic_add : SDNode<"ARMISD::PIC_ADD", SDT_ARMPICAdd>; + +def ARMsrl_flag : SDNode<"ARMISD::SRL_FLAG", SDTIntUnaryOp, [SDNPOutGlue]>; +def ARMsra_flag : SDNode<"ARMISD::SRA_FLAG", SDTIntUnaryOp, [SDNPOutGlue]>; +def ARMrrx : SDNode<"ARMISD::RRX" , SDTIntUnaryOp, [SDNPInGlue ]>; + +def ARMaddc : SDNode<"ARMISD::ADDC", SDTBinaryArithWithFlags, + [SDNPCommutative]>; +def ARMsubc : SDNode<"ARMISD::SUBC", SDTBinaryArithWithFlags>; +def ARMadde : SDNode<"ARMISD::ADDE", SDTBinaryArithWithFlagsInOut>; +def ARMsube : SDNode<"ARMISD::SUBE", SDTBinaryArithWithFlagsInOut>; + +def ARMthread_pointer: SDNode<"ARMISD::THREAD_POINTER", SDT_ARMThreadPointer>; +def ARMeh_sjlj_setjmp: SDNode<"ARMISD::EH_SJLJ_SETJMP", + SDT_ARMEH_SJLJ_Setjmp, + [SDNPHasChain, SDNPSideEffect]>; +def ARMeh_sjlj_longjmp: SDNode<"ARMISD::EH_SJLJ_LONGJMP", + SDT_ARMEH_SJLJ_Longjmp, + [SDNPHasChain, SDNPSideEffect]>; +def ARMeh_sjlj_setup_dispatch: SDNode<"ARMISD::EH_SJLJ_SETUP_DISPATCH", + SDT_ARMEH_SJLJ_SetupDispatch, + [SDNPHasChain, SDNPSideEffect]>; + +def ARMMemBarrierMCR : SDNode<"ARMISD::MEMBARRIER_MCR", SDT_ARMMEMBARRIER, + [SDNPHasChain, SDNPSideEffect]>; +def ARMPreload : SDNode<"ARMISD::PRELOAD", SDT_ARMPREFETCH, + [SDNPHasChain, SDNPMayLoad, SDNPMayStore]>; + +def ARMtcret : SDNode<"ARMISD::TC_RETURN", SDT_ARMTCRET, + [SDNPHasChain, SDNPOptInGlue, SDNPVariadic]>; + +def ARMbfi : SDNode<"ARMISD::BFI", SDT_ARMBFI>; + +def ARMmemcopy : SDNode<"ARMISD::MEMCPY", SDT_ARMMEMCPY, + [SDNPHasChain, SDNPInGlue, SDNPOutGlue, + SDNPMayStore, SDNPMayLoad]>; + +def ARMsmulwb : SDNode<"ARMISD::SMULWB", SDTIntBinOp, []>; +def ARMsmulwt : SDNode<"ARMISD::SMULWT", SDTIntBinOp, []>; +def ARMsmlalbb : SDNode<"ARMISD::SMLALBB", SDT_LongMac, []>; +def ARMsmlalbt : SDNode<"ARMISD::SMLALBT", SDT_LongMac, []>; +def ARMsmlaltb : SDNode<"ARMISD::SMLALTB", SDT_LongMac, []>; +def ARMsmlaltt : SDNode<"ARMISD::SMLALTT", SDT_LongMac, []>; + +//===----------------------------------------------------------------------===// +// ARM Instruction Predicate Definitions. +// +def HasV4T : Predicate<"Subtarget->hasV4TOps()">, + AssemblerPredicate<"HasV4TOps", "armv4t">; +def NoV4T : Predicate<"!Subtarget->hasV4TOps()">; +def HasV5T : Predicate<"Subtarget->hasV5TOps()">, + AssemblerPredicate<"HasV5TOps", "armv5t">; +def HasV5TE : Predicate<"Subtarget->hasV5TEOps()">, + AssemblerPredicate<"HasV5TEOps", "armv5te">; +def HasV6 : Predicate<"Subtarget->hasV6Ops()">, + AssemblerPredicate<"HasV6Ops", "armv6">; +def NoV6 : Predicate<"!Subtarget->hasV6Ops()">; +def HasV6M : Predicate<"Subtarget->hasV6MOps()">, + AssemblerPredicate<"HasV6MOps", + "armv6m or armv6t2">; +def HasV8MBaseline : Predicate<"Subtarget->hasV8MBaselineOps()">, + AssemblerPredicate<"HasV8MBaselineOps", + "armv8m.base">; +def HasV8MMainline : Predicate<"Subtarget->hasV8MMainlineOps()">, + AssemblerPredicate<"HasV8MMainlineOps", + "armv8m.main">; +def HasV6T2 : Predicate<"Subtarget->hasV6T2Ops()">, + AssemblerPredicate<"HasV6T2Ops", "armv6t2">; +def NoV6T2 : Predicate<"!Subtarget->hasV6T2Ops()">; +def HasV6K : Predicate<"Subtarget->hasV6KOps()">, + AssemblerPredicate<"HasV6KOps", "armv6k">; +def NoV6K : Predicate<"!Subtarget->hasV6KOps()">; +def HasV7 : Predicate<"Subtarget->hasV7Ops()">, + AssemblerPredicate<"HasV7Ops", "armv7">; +def HasV8 : Predicate<"Subtarget->hasV8Ops()">, + AssemblerPredicate<"HasV8Ops", "armv8">; +def PreV8 : Predicate<"!Subtarget->hasV8Ops()">, + AssemblerPredicate<"!HasV8Ops", "armv7 or earlier">; +def HasV8_1a : Predicate<"Subtarget->hasV8_1aOps()">, + AssemblerPredicate<"HasV8_1aOps", "armv8.1a">; +def HasV8_2a : Predicate<"Subtarget->hasV8_2aOps()">, + AssemblerPredicate<"HasV8_2aOps", "armv8.2a">; +def HasV8_3a : Predicate<"Subtarget->hasV8_3aOps()">, + AssemblerPredicate<"HasV8_3aOps", "armv8.3a">; +def HasV8_4a : Predicate<"Subtarget->hasV8_4aOps()">, + AssemblerPredicate<"HasV8_4aOps", "armv8.4a">; +def NoVFP : Predicate<"!Subtarget->hasVFP2()">; +def HasVFP2 : Predicate<"Subtarget->hasVFP2()">, + AssemblerPredicate<"FeatureVFP2", "VFP2">; +def HasVFP3 : Predicate<"Subtarget->hasVFP3()">, + AssemblerPredicate<"FeatureVFP3", "VFP3">; +def HasVFP4 : Predicate<"Subtarget->hasVFP4()">, + AssemblerPredicate<"FeatureVFP4", "VFP4">; +def HasDPVFP : Predicate<"!Subtarget->isFPOnlySP()">, + AssemblerPredicate<"!FeatureVFPOnlySP", + "double precision VFP">; +def HasFPARMv8 : Predicate<"Subtarget->hasFPARMv8()">, + AssemblerPredicate<"FeatureFPARMv8", "FPARMv8">; +def HasNEON : Predicate<"Subtarget->hasNEON()">, + AssemblerPredicate<"FeatureNEON", "NEON">; +def HasSHA2 : Predicate<"Subtarget->hasSHA2()">, + AssemblerPredicate<"FeatureSHA2", "sha2">; +def HasAES : Predicate<"Subtarget->hasAES()">, + AssemblerPredicate<"FeatureAES", "aes">; +def HasCrypto : Predicate<"Subtarget->hasCrypto()">, + AssemblerPredicate<"FeatureCrypto", "crypto">; +def HasDotProd : Predicate<"Subtarget->hasDotProd()">, + AssemblerPredicate<"FeatureDotProd", "dotprod">; +def HasCRC : Predicate<"Subtarget->hasCRC()">, + AssemblerPredicate<"FeatureCRC", "crc">; +def HasRAS : Predicate<"Subtarget->hasRAS()">, + AssemblerPredicate<"FeatureRAS", "ras">; +def HasFP16 : Predicate<"Subtarget->hasFP16()">, + AssemblerPredicate<"FeatureFP16","half-float conversions">; +def HasFullFP16 : Predicate<"Subtarget->hasFullFP16()">, + AssemblerPredicate<"FeatureFullFP16","full half-float">; +def HasDivideInThumb : Predicate<"Subtarget->hasDivideInThumbMode()">, + AssemblerPredicate<"FeatureHWDivThumb", "divide in THUMB">; +def HasDivideInARM : Predicate<"Subtarget->hasDivideInARMMode()">, + AssemblerPredicate<"FeatureHWDivARM", "divide in ARM">; +def HasDSP : Predicate<"Subtarget->hasDSP()">, + AssemblerPredicate<"FeatureDSP", "dsp">; +def HasDB : Predicate<"Subtarget->hasDataBarrier()">, + AssemblerPredicate<"FeatureDB", + "data-barriers">; +def HasDFB : Predicate<"Subtarget->hasFullDataBarrier()">, + AssemblerPredicate<"FeatureDFB", + "full-data-barrier">; +def HasV7Clrex : Predicate<"Subtarget->hasV7Clrex()">, + AssemblerPredicate<"FeatureV7Clrex", + "v7 clrex">; +def HasAcquireRelease : Predicate<"Subtarget->hasAcquireRelease()">, + AssemblerPredicate<"FeatureAcquireRelease", + "acquire/release">; +def HasMP : Predicate<"Subtarget->hasMPExtension()">, + AssemblerPredicate<"FeatureMP", + "mp-extensions">; +def HasVirtualization: Predicate<"false">, + AssemblerPredicate<"FeatureVirtualization", + "virtualization-extensions">; +def HasTrustZone : Predicate<"Subtarget->hasTrustZone()">, + AssemblerPredicate<"FeatureTrustZone", + "TrustZone">; +def Has8MSecExt : Predicate<"Subtarget->has8MSecExt()">, + AssemblerPredicate<"Feature8MSecExt", + "ARMv8-M Security Extensions">; +def HasZCZ : Predicate<"Subtarget->hasZeroCycleZeroing()">; +def UseNEONForFP : Predicate<"Subtarget->useNEONForSinglePrecisionFP()">; +def DontUseNEONForFP : Predicate<"!Subtarget->useNEONForSinglePrecisionFP()">; +def IsThumb : Predicate<"Subtarget->isThumb()">, + AssemblerPredicate<"ModeThumb", "thumb">; +def IsThumb1Only : Predicate<"Subtarget->isThumb1Only()">; +def IsThumb2 : Predicate<"Subtarget->isThumb2()">, + AssemblerPredicate<"ModeThumb,FeatureThumb2", + "thumb2">; +def IsMClass : Predicate<"Subtarget->isMClass()">, + AssemblerPredicate<"FeatureMClass", "armv*m">; +def IsNotMClass : Predicate<"!Subtarget->isMClass()">, + AssemblerPredicate<"!FeatureMClass", + "!armv*m">; +def IsARM : Predicate<"!Subtarget->isThumb()">, + AssemblerPredicate<"!ModeThumb", "arm-mode">; +def IsMachO : Predicate<"Subtarget->isTargetMachO()">; +def IsNotMachO : Predicate<"!Subtarget->isTargetMachO()">; +def IsNaCl : Predicate<"Subtarget->isTargetNaCl()">; +def IsWindows : Predicate<"Subtarget->isTargetWindows()">; +def IsNotWindows : Predicate<"!Subtarget->isTargetWindows()">; +def IsReadTPHard : Predicate<"Subtarget->isReadTPHard()">; +def IsReadTPSoft : Predicate<"!Subtarget->isReadTPHard()">; +def UseNaClTrap : Predicate<"Subtarget->useNaClTrap()">, + AssemblerPredicate<"FeatureNaClTrap", "NaCl">; +def DontUseNaClTrap : Predicate<"!Subtarget->useNaClTrap()">; + +def UseNegativeImmediates : + Predicate<"false">, + AssemblerPredicate<"!FeatureNoNegativeImmediates", + "NegativeImmediates">; + +// FIXME: Eventually this will be just "hasV6T2Ops". +let RecomputePerFunction = 1 in { + def UseMovt : Predicate<"Subtarget->useMovt(*MF)">; + def DontUseMovt : Predicate<"!Subtarget->useMovt(*MF)">; + def UseMovtInPic : Predicate<"Subtarget->useMovt(*MF) && Subtarget->allowPositionIndependentMovt()">; + def DontUseMovtInPic : Predicate<"!Subtarget->useMovt(*MF) || !Subtarget->allowPositionIndependentMovt()">; +} +def UseFPVMLx : Predicate<"Subtarget->useFPVMLx()">; +def UseMulOps : Predicate<"Subtarget->useMulOps()">; + +// Prefer fused MAC for fp mul + add over fp VMLA / VMLS if they are available. +// But only select them if more precision in FP computation is allowed. +// Do not use them for Darwin platforms. +def UseFusedMAC : Predicate<"(TM.Options.AllowFPOpFusion ==" + " FPOpFusion::Fast && " + " Subtarget->hasVFP4()) && " + "!Subtarget->isTargetDarwin()">; +def DontUseFusedMAC : Predicate<"!(TM.Options.AllowFPOpFusion ==" + " FPOpFusion::Fast &&" + " Subtarget->hasVFP4()) || " + "Subtarget->isTargetDarwin()">; + +def HasFastVGETLNi32 : Predicate<"!Subtarget->hasSlowVGETLNi32()">; +def HasSlowVGETLNi32 : Predicate<"Subtarget->hasSlowVGETLNi32()">; + +def HasFastVDUP32 : Predicate<"!Subtarget->hasSlowVDUP32()">; +def HasSlowVDUP32 : Predicate<"Subtarget->hasSlowVDUP32()">; + +def UseVMOVSR : Predicate<"Subtarget->preferVMOVSR() ||" + "!Subtarget->useNEONForSinglePrecisionFP()">; +def DontUseVMOVSR : Predicate<"!Subtarget->preferVMOVSR() &&" + "Subtarget->useNEONForSinglePrecisionFP()">; + +let RecomputePerFunction = 1 in { + def IsLE : Predicate<"MF->getDataLayout().isLittleEndian()">; + def IsBE : Predicate<"MF->getDataLayout().isBigEndian()">; +} + +def GenExecuteOnly : Predicate<"Subtarget->genExecuteOnly()">; + +//===----------------------------------------------------------------------===// +// ARM Flag Definitions. + +class RegConstraint<string C> { + string Constraints = C; +} + +//===----------------------------------------------------------------------===// +// ARM specific transformation functions and pattern fragments. +// + +// imm_neg_XFORM - Return the negation of an i32 immediate value. +def imm_neg_XFORM : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant(-(int)N->getZExtValue(), SDLoc(N), MVT::i32); +}]>; + +// imm_not_XFORM - Return the complement of a i32 immediate value. +def imm_not_XFORM : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant(~(int)N->getZExtValue(), SDLoc(N), MVT::i32); +}]>; + +/// imm16_31 predicate - True if the 32-bit immediate is in the range [16,31]. +def imm16_31 : ImmLeaf<i32, [{ + return (int32_t)Imm >= 16 && (int32_t)Imm < 32; +}]>; + +// sext_16_node predicate - True if the SDNode is sign-extended 16 or more bits. +def sext_16_node : PatLeaf<(i32 GPR:$a), [{ + if (CurDAG->ComputeNumSignBits(SDValue(N,0)) >= 17) + return true; + + if (N->getOpcode() != ISD::SRA) + return false; + if (N->getOperand(0).getOpcode() != ISD::SHL) + return false; + + auto *ShiftVal = dyn_cast<ConstantSDNode>(N->getOperand(1)); + if (!ShiftVal || ShiftVal->getZExtValue() != 16) + return false; + + ShiftVal = dyn_cast<ConstantSDNode>(N->getOperand(0)->getOperand(1)); + if (!ShiftVal || ShiftVal->getZExtValue() != 16) + return false; + + return true; +}]>; + +/// Split a 32-bit immediate into two 16 bit parts. +def hi16 : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant((uint32_t)N->getZExtValue() >> 16, SDLoc(N), + MVT::i32); +}]>; + +def lo16AllZero : PatLeaf<(i32 imm), [{ + // Returns true if all low 16-bits are 0. + return (((uint32_t)N->getZExtValue()) & 0xFFFFUL) == 0; +}], hi16>; + +class BinOpFrag<dag res> : PatFrag<(ops node:$LHS, node:$RHS), res>; +class UnOpFrag <dag res> : PatFrag<(ops node:$Src), res>; + +// An 'and' node with a single use. +def and_su : PatFrag<(ops node:$lhs, node:$rhs), (and node:$lhs, node:$rhs), [{ + return N->hasOneUse(); +}]>; + +// An 'xor' node with a single use. +def xor_su : PatFrag<(ops node:$lhs, node:$rhs), (xor node:$lhs, node:$rhs), [{ + return N->hasOneUse(); +}]>; + +// An 'fmul' node with a single use. +def fmul_su : PatFrag<(ops node:$lhs, node:$rhs), (fmul node:$lhs, node:$rhs),[{ + return N->hasOneUse(); +}]>; + +// An 'fadd' node which checks for single non-hazardous use. +def fadd_mlx : PatFrag<(ops node:$lhs, node:$rhs),(fadd node:$lhs, node:$rhs),[{ + return hasNoVMLxHazardUse(N); +}]>; + +// An 'fsub' node which checks for single non-hazardous use. +def fsub_mlx : PatFrag<(ops node:$lhs, node:$rhs),(fsub node:$lhs, node:$rhs),[{ + return hasNoVMLxHazardUse(N); +}]>; + +//===----------------------------------------------------------------------===// +// Operand Definitions. +// + +// Immediate operands with a shared generic asm render method. +class ImmAsmOperand<int Low, int High> : AsmOperandClass { + let RenderMethod = "addImmOperands"; + let PredicateMethod = "isImmediate<" # Low # "," # High # ">"; + let DiagnosticString = "operand must be an immediate in the range [" # Low # "," # High # "]"; +} + +class ImmAsmOperandMinusOne<int Low, int High> : AsmOperandClass { + let PredicateMethod = "isImmediate<" # Low # "," # High # ">"; + let DiagnosticType = "ImmRange" # Low # "_" # High; + let DiagnosticString = "operand must be an immediate in the range [" # Low # "," # High # "]"; +} + +// Operands that are part of a memory addressing mode. +class MemOperand : Operand<i32> { let OperandType = "OPERAND_MEMORY"; } + +// Branch target. +// FIXME: rename brtarget to t2_brtarget +def brtarget : Operand<OtherVT> { + let EncoderMethod = "getBranchTargetOpValue"; + let OperandType = "OPERAND_PCREL"; + let DecoderMethod = "DecodeT2BROperand"; +} + +// Branches targeting ARM-mode must be divisible by 4 if they're a raw +// immediate. +def ARMBranchTarget : AsmOperandClass { + let Name = "ARMBranchTarget"; +} + +// Branches targeting Thumb-mode must be divisible by 2 if they're a raw +// immediate. +def ThumbBranchTarget : AsmOperandClass { + let Name = "ThumbBranchTarget"; +} + +def arm_br_target : Operand<OtherVT> { + let ParserMatchClass = ARMBranchTarget; + let EncoderMethod = "getARMBranchTargetOpValue"; + let OperandType = "OPERAND_PCREL"; +} + +// Call target for ARM. Handles conditional/unconditional +// FIXME: rename bl_target to t2_bltarget? +def arm_bl_target : Operand<i32> { + let ParserMatchClass = ARMBranchTarget; + let EncoderMethod = "getARMBLTargetOpValue"; + let OperandType = "OPERAND_PCREL"; +} + +// Target for BLX *from* ARM mode. +def arm_blx_target : Operand<i32> { + let ParserMatchClass = ThumbBranchTarget; + let EncoderMethod = "getARMBLXTargetOpValue"; + let OperandType = "OPERAND_PCREL"; +} + +// A list of registers separated by comma. Used by load/store multiple. +def RegListAsmOperand : AsmOperandClass { let Name = "RegList"; } +def reglist : Operand<i32> { + let EncoderMethod = "getRegisterListOpValue"; + let ParserMatchClass = RegListAsmOperand; + let PrintMethod = "printRegisterList"; + let DecoderMethod = "DecodeRegListOperand"; +} + +def GPRPairOp : RegisterOperand<GPRPair, "printGPRPairOperand">; + +def DPRRegListAsmOperand : AsmOperandClass { + let Name = "DPRRegList"; + let DiagnosticType = "DPR_RegList"; +} +def dpr_reglist : Operand<i32> { + let EncoderMethod = "getRegisterListOpValue"; + let ParserMatchClass = DPRRegListAsmOperand; + let PrintMethod = "printRegisterList"; + let DecoderMethod = "DecodeDPRRegListOperand"; +} + +def SPRRegListAsmOperand : AsmOperandClass { + let Name = "SPRRegList"; + let DiagnosticString = "operand must be a list of registers in range [s0, s31]"; +} +def spr_reglist : Operand<i32> { + let EncoderMethod = "getRegisterListOpValue"; + let ParserMatchClass = SPRRegListAsmOperand; + let PrintMethod = "printRegisterList"; + let DecoderMethod = "DecodeSPRRegListOperand"; +} + +// An operand for the CONSTPOOL_ENTRY pseudo-instruction. +def cpinst_operand : Operand<i32> { + let PrintMethod = "printCPInstOperand"; +} + +// Local PC labels. +def pclabel : Operand<i32> { + let PrintMethod = "printPCLabel"; +} + +// ADR instruction labels. +def AdrLabelAsmOperand : AsmOperandClass { let Name = "AdrLabel"; } +def adrlabel : Operand<i32> { + let EncoderMethod = "getAdrLabelOpValue"; + let ParserMatchClass = AdrLabelAsmOperand; + let PrintMethod = "printAdrLabelOperand<0>"; +} + +def neon_vcvt_imm32 : Operand<i32> { + let EncoderMethod = "getNEONVcvtImm32OpValue"; + let DecoderMethod = "DecodeVCVTImmOperand"; +} + +// rot_imm: An integer that encodes a rotate amount. Must be 8, 16, or 24. +def rot_imm_XFORM: SDNodeXForm<imm, [{ + switch (N->getZExtValue()){ + default: llvm_unreachable(nullptr); + case 0: return CurDAG->getTargetConstant(0, SDLoc(N), MVT::i32); + case 8: return CurDAG->getTargetConstant(1, SDLoc(N), MVT::i32); + case 16: return CurDAG->getTargetConstant(2, SDLoc(N), MVT::i32); + case 24: return CurDAG->getTargetConstant(3, SDLoc(N), MVT::i32); + } +}]>; +def RotImmAsmOperand : AsmOperandClass { + let Name = "RotImm"; + let ParserMethod = "parseRotImm"; +} +def rot_imm : Operand<i32>, PatLeaf<(i32 imm), [{ + int32_t v = N->getZExtValue(); + return v == 8 || v == 16 || v == 24; }], + rot_imm_XFORM> { + let PrintMethod = "printRotImmOperand"; + let ParserMatchClass = RotImmAsmOperand; +} + +// shift_imm: An integer that encodes a shift amount and the type of shift +// (asr or lsl). The 6-bit immediate encodes as: +// {5} 0 ==> lsl +// 1 asr +// {4-0} imm5 shift amount. +// asr #32 encoded as imm5 == 0. +def ShifterImmAsmOperand : AsmOperandClass { + let Name = "ShifterImm"; + let ParserMethod = "parseShifterImm"; +} +def shift_imm : Operand<i32> { + let PrintMethod = "printShiftImmOperand"; + let ParserMatchClass = ShifterImmAsmOperand; +} + +// shifter_operand operands: so_reg_reg, so_reg_imm, and mod_imm. +def ShiftedRegAsmOperand : AsmOperandClass { let Name = "RegShiftedReg"; } +def so_reg_reg : Operand<i32>, // reg reg imm + ComplexPattern<i32, 3, "SelectRegShifterOperand", + [shl, srl, sra, rotr]> { + let EncoderMethod = "getSORegRegOpValue"; + let PrintMethod = "printSORegRegOperand"; + let DecoderMethod = "DecodeSORegRegOperand"; + let ParserMatchClass = ShiftedRegAsmOperand; + let MIOperandInfo = (ops GPRnopc, GPRnopc, i32imm); +} + +def ShiftedImmAsmOperand : AsmOperandClass { let Name = "RegShiftedImm"; } +def so_reg_imm : Operand<i32>, // reg imm + ComplexPattern<i32, 2, "SelectImmShifterOperand", + [shl, srl, sra, rotr]> { + let EncoderMethod = "getSORegImmOpValue"; + let PrintMethod = "printSORegImmOperand"; + let DecoderMethod = "DecodeSORegImmOperand"; + let ParserMatchClass = ShiftedImmAsmOperand; + let MIOperandInfo = (ops GPR, i32imm); +} + +// FIXME: Does this need to be distinct from so_reg? +def shift_so_reg_reg : Operand<i32>, // reg reg imm + ComplexPattern<i32, 3, "SelectShiftRegShifterOperand", + [shl,srl,sra,rotr]> { + let EncoderMethod = "getSORegRegOpValue"; + let PrintMethod = "printSORegRegOperand"; + let DecoderMethod = "DecodeSORegRegOperand"; + let ParserMatchClass = ShiftedRegAsmOperand; + let MIOperandInfo = (ops GPR, GPR, i32imm); +} + +// FIXME: Does this need to be distinct from so_reg? +def shift_so_reg_imm : Operand<i32>, // reg reg imm + ComplexPattern<i32, 2, "SelectShiftImmShifterOperand", + [shl,srl,sra,rotr]> { + let EncoderMethod = "getSORegImmOpValue"; + let PrintMethod = "printSORegImmOperand"; + let DecoderMethod = "DecodeSORegImmOperand"; + let ParserMatchClass = ShiftedImmAsmOperand; + let MIOperandInfo = (ops GPR, i32imm); +} + +// mod_imm: match a 32-bit immediate operand, which can be encoded into +// a 12-bit immediate; an 8-bit integer and a 4-bit rotator (See ARMARM +// - "Modified Immediate Constants"). Within the MC layer we keep this +// immediate in its encoded form. +def ModImmAsmOperand: AsmOperandClass { + let Name = "ModImm"; + let ParserMethod = "parseModImm"; +} +def mod_imm : Operand<i32>, ImmLeaf<i32, [{ + return ARM_AM::getSOImmVal(Imm) != -1; + }]> { + let EncoderMethod = "getModImmOpValue"; + let PrintMethod = "printModImmOperand"; + let ParserMatchClass = ModImmAsmOperand; +} + +// Note: the patterns mod_imm_not and mod_imm_neg do not require an encoder +// method and such, as they are only used on aliases (Pat<> and InstAlias<>). +// The actual parsing, encoding, decoding are handled by the destination +// instructions, which use mod_imm. + +def ModImmNotAsmOperand : AsmOperandClass { let Name = "ModImmNot"; } +def mod_imm_not : Operand<i32>, PatLeaf<(imm), [{ + return ARM_AM::getSOImmVal(~(uint32_t)N->getZExtValue()) != -1; + }], imm_not_XFORM> { + let ParserMatchClass = ModImmNotAsmOperand; +} + +def ModImmNegAsmOperand : AsmOperandClass { let Name = "ModImmNeg"; } +def mod_imm_neg : Operand<i32>, PatLeaf<(imm), [{ + unsigned Value = -(unsigned)N->getZExtValue(); + return Value && ARM_AM::getSOImmVal(Value) != -1; + }], imm_neg_XFORM> { + let ParserMatchClass = ModImmNegAsmOperand; +} + +/// arm_i32imm - True for +V6T2, or when isSOImmTwoParVal() +def arm_i32imm : PatLeaf<(imm), [{ + if (Subtarget->useMovt(*MF)) + return true; + return ARM_AM::isSOImmTwoPartVal((unsigned)N->getZExtValue()); +}]>; + +/// imm0_1 predicate - Immediate in the range [0,1]. +def Imm0_1AsmOperand: ImmAsmOperand<0,1> { let Name = "Imm0_1"; } +def imm0_1 : Operand<i32> { let ParserMatchClass = Imm0_1AsmOperand; } + +/// imm0_3 predicate - Immediate in the range [0,3]. +def Imm0_3AsmOperand: ImmAsmOperand<0,3> { let Name = "Imm0_3"; } +def imm0_3 : Operand<i32> { let ParserMatchClass = Imm0_3AsmOperand; } + +/// imm0_7 predicate - Immediate in the range [0,7]. +def Imm0_7AsmOperand: ImmAsmOperand<0,7> { + let Name = "Imm0_7"; +} +def imm0_7 : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 0 && Imm < 8; +}]> { + let ParserMatchClass = Imm0_7AsmOperand; +} + +/// imm8_255 predicate - Immediate in the range [8,255]. +def Imm8_255AsmOperand: ImmAsmOperand<8,255> { let Name = "Imm8_255"; } +def imm8_255 : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 8 && Imm < 256; +}]> { + let ParserMatchClass = Imm8_255AsmOperand; +} + +/// imm8 predicate - Immediate is exactly 8. +def Imm8AsmOperand: ImmAsmOperand<8,8> { let Name = "Imm8"; } +def imm8 : Operand<i32>, ImmLeaf<i32, [{ return Imm == 8; }]> { + let ParserMatchClass = Imm8AsmOperand; +} + +/// imm16 predicate - Immediate is exactly 16. +def Imm16AsmOperand: ImmAsmOperand<16,16> { let Name = "Imm16"; } +def imm16 : Operand<i32>, ImmLeaf<i32, [{ return Imm == 16; }]> { + let ParserMatchClass = Imm16AsmOperand; +} + +/// imm32 predicate - Immediate is exactly 32. +def Imm32AsmOperand: ImmAsmOperand<32,32> { let Name = "Imm32"; } +def imm32 : Operand<i32>, ImmLeaf<i32, [{ return Imm == 32; }]> { + let ParserMatchClass = Imm32AsmOperand; +} + +def imm8_or_16 : ImmLeaf<i32, [{ return Imm == 8 || Imm == 16;}]>; + +/// imm1_7 predicate - Immediate in the range [1,7]. +def Imm1_7AsmOperand: ImmAsmOperand<1,7> { let Name = "Imm1_7"; } +def imm1_7 : Operand<i32>, ImmLeaf<i32, [{ return Imm > 0 && Imm < 8; }]> { + let ParserMatchClass = Imm1_7AsmOperand; +} + +/// imm1_15 predicate - Immediate in the range [1,15]. +def Imm1_15AsmOperand: ImmAsmOperand<1,15> { let Name = "Imm1_15"; } +def imm1_15 : Operand<i32>, ImmLeaf<i32, [{ return Imm > 0 && Imm < 16; }]> { + let ParserMatchClass = Imm1_15AsmOperand; +} + +/// imm1_31 predicate - Immediate in the range [1,31]. +def Imm1_31AsmOperand: ImmAsmOperand<1,31> { let Name = "Imm1_31"; } +def imm1_31 : Operand<i32>, ImmLeaf<i32, [{ return Imm > 0 && Imm < 32; }]> { + let ParserMatchClass = Imm1_31AsmOperand; +} + +/// imm0_15 predicate - Immediate in the range [0,15]. +def Imm0_15AsmOperand: ImmAsmOperand<0,15> { + let Name = "Imm0_15"; +} +def imm0_15 : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 0 && Imm < 16; +}]> { + let ParserMatchClass = Imm0_15AsmOperand; +} + +/// imm0_31 predicate - True if the 32-bit immediate is in the range [0,31]. +def Imm0_31AsmOperand: ImmAsmOperand<0,31> { let Name = "Imm0_31"; } +def imm0_31 : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 0 && Imm < 32; +}]> { + let ParserMatchClass = Imm0_31AsmOperand; +} + +/// imm0_32 predicate - True if the 32-bit immediate is in the range [0,32]. +def Imm0_32AsmOperand: ImmAsmOperand<0,32> { let Name = "Imm0_32"; } +def imm0_32 : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 0 && Imm < 33; +}]> { + let ParserMatchClass = Imm0_32AsmOperand; +} + +/// imm0_63 predicate - True if the 32-bit immediate is in the range [0,63]. +def Imm0_63AsmOperand: ImmAsmOperand<0,63> { let Name = "Imm0_63"; } +def imm0_63 : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 0 && Imm < 64; +}]> { + let ParserMatchClass = Imm0_63AsmOperand; +} + +/// imm0_239 predicate - Immediate in the range [0,239]. +def Imm0_239AsmOperand : ImmAsmOperand<0,239> { + let Name = "Imm0_239"; +} +def imm0_239 : Operand<i32>, ImmLeaf<i32, [{ return Imm >= 0 && Imm < 240; }]> { + let ParserMatchClass = Imm0_239AsmOperand; +} + +/// imm0_255 predicate - Immediate in the range [0,255]. +def Imm0_255AsmOperand : ImmAsmOperand<0,255> { let Name = "Imm0_255"; } +def imm0_255 : Operand<i32>, ImmLeaf<i32, [{ return Imm >= 0 && Imm < 256; }]> { + let ParserMatchClass = Imm0_255AsmOperand; +} + +/// imm0_65535 - An immediate is in the range [0,65535]. +def Imm0_65535AsmOperand: ImmAsmOperand<0,65535> { let Name = "Imm0_65535"; } +def imm0_65535 : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 0 && Imm < 65536; +}]> { + let ParserMatchClass = Imm0_65535AsmOperand; +} + +// imm0_65535_neg - An immediate whose negative value is in the range [0.65535]. +def imm0_65535_neg : Operand<i32>, ImmLeaf<i32, [{ + return -Imm >= 0 && -Imm < 65536; +}]>; + +// imm0_65535_expr - For movt/movw - 16-bit immediate that can also reference +// a relocatable expression. +// +// FIXME: This really needs a Thumb version separate from the ARM version. +// While the range is the same, and can thus use the same match class, +// the encoding is different so it should have a different encoder method. +def Imm0_65535ExprAsmOperand: AsmOperandClass { + let Name = "Imm0_65535Expr"; + let RenderMethod = "addImmOperands"; + let DiagnosticString = "operand must be an immediate in the range [0,0xffff] or a relocatable expression"; +} + +def imm0_65535_expr : Operand<i32> { + let EncoderMethod = "getHiLo16ImmOpValue"; + let ParserMatchClass = Imm0_65535ExprAsmOperand; +} + +def Imm256_65535ExprAsmOperand: ImmAsmOperand<256,65535> { let Name = "Imm256_65535Expr"; } +def imm256_65535_expr : Operand<i32> { + let ParserMatchClass = Imm256_65535ExprAsmOperand; +} + +/// imm24b - True if the 32-bit immediate is encodable in 24 bits. +def Imm24bitAsmOperand: ImmAsmOperand<0,0xffffff> { + let Name = "Imm24bit"; + let DiagnosticString = "operand must be an immediate in the range [0,0xffffff]"; +} +def imm24b : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 0 && Imm <= 0xffffff; +}]> { + let ParserMatchClass = Imm24bitAsmOperand; +} + + +/// bf_inv_mask_imm predicate - An AND mask to clear an arbitrary width bitfield +/// e.g., 0xf000ffff +def BitfieldAsmOperand : AsmOperandClass { + let Name = "Bitfield"; + let ParserMethod = "parseBitfield"; +} + +def bf_inv_mask_imm : Operand<i32>, + PatLeaf<(imm), [{ + return ARM::isBitFieldInvertedMask(N->getZExtValue()); +}] > { + let EncoderMethod = "getBitfieldInvertedMaskOpValue"; + let PrintMethod = "printBitfieldInvMaskImmOperand"; + let DecoderMethod = "DecodeBitfieldMaskOperand"; + let ParserMatchClass = BitfieldAsmOperand; + let GISelPredicateCode = [{ + // There's better methods of implementing this check. IntImmLeaf<> would be + // equivalent and have less boilerplate but we need a test for C++ + // predicates and this one causes new rules to be imported into GlobalISel + // without requiring additional features first. + const auto &MO = MI.getOperand(1); + if (!MO.isCImm()) + return false; + return ARM::isBitFieldInvertedMask(MO.getCImm()->getZExtValue()); + }]; +} + +def imm1_32_XFORM: SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant((int)N->getZExtValue() - 1, SDLoc(N), + MVT::i32); +}]>; +def Imm1_32AsmOperand: ImmAsmOperandMinusOne<1,32> { + let Name = "Imm1_32"; +} +def imm1_32 : Operand<i32>, PatLeaf<(imm), [{ + uint64_t Imm = N->getZExtValue(); + return Imm > 0 && Imm <= 32; + }], + imm1_32_XFORM> { + let PrintMethod = "printImmPlusOneOperand"; + let ParserMatchClass = Imm1_32AsmOperand; +} + +def imm1_16_XFORM: SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant((int)N->getZExtValue() - 1, SDLoc(N), + MVT::i32); +}]>; +def Imm1_16AsmOperand: ImmAsmOperandMinusOne<1,16> { let Name = "Imm1_16"; } +def imm1_16 : Operand<i32>, ImmLeaf<i32, [{ + return Imm > 0 && Imm <= 16; + }], + imm1_16_XFORM> { + let PrintMethod = "printImmPlusOneOperand"; + let ParserMatchClass = Imm1_16AsmOperand; +} + +// Define ARM specific addressing modes. +// addrmode_imm12 := reg +/- imm12 +// +def MemImm12OffsetAsmOperand : AsmOperandClass { let Name = "MemImm12Offset"; } +class AddrMode_Imm12 : MemOperand, + ComplexPattern<i32, 2, "SelectAddrModeImm12", []> { + // 12-bit immediate operand. Note that instructions using this encode + // #0 and #-0 differently. We flag #-0 as the magic value INT32_MIN. All other + // immediate values are as normal. + + let EncoderMethod = "getAddrModeImm12OpValue"; + let DecoderMethod = "DecodeAddrModeImm12Operand"; + let ParserMatchClass = MemImm12OffsetAsmOperand; + let MIOperandInfo = (ops GPR:$base, i32imm:$offsimm); +} + +def addrmode_imm12 : AddrMode_Imm12 { + let PrintMethod = "printAddrModeImm12Operand<false>"; +} + +def addrmode_imm12_pre : AddrMode_Imm12 { + let PrintMethod = "printAddrModeImm12Operand<true>"; +} + +// ldst_so_reg := reg +/- reg shop imm +// +def MemRegOffsetAsmOperand : AsmOperandClass { let Name = "MemRegOffset"; } +def ldst_so_reg : MemOperand, + ComplexPattern<i32, 3, "SelectLdStSOReg", []> { + let EncoderMethod = "getLdStSORegOpValue"; + // FIXME: Simplify the printer + let PrintMethod = "printAddrMode2Operand"; + let DecoderMethod = "DecodeSORegMemOperand"; + let ParserMatchClass = MemRegOffsetAsmOperand; + let MIOperandInfo = (ops GPR:$base, GPRnopc:$offsreg, i32imm:$shift); +} + +// postidx_imm8 := +/- [0,255] +// +// 9 bit value: +// {8} 1 is imm8 is non-negative. 0 otherwise. +// {7-0} [0,255] imm8 value. +def PostIdxImm8AsmOperand : AsmOperandClass { let Name = "PostIdxImm8"; } +def postidx_imm8 : MemOperand { + let PrintMethod = "printPostIdxImm8Operand"; + let ParserMatchClass = PostIdxImm8AsmOperand; + let MIOperandInfo = (ops i32imm); +} + +// postidx_imm8s4 := +/- [0,1020] +// +// 9 bit value: +// {8} 1 is imm8 is non-negative. 0 otherwise. +// {7-0} [0,255] imm8 value, scaled by 4. +def PostIdxImm8s4AsmOperand : AsmOperandClass { let Name = "PostIdxImm8s4"; } +def postidx_imm8s4 : MemOperand { + let PrintMethod = "printPostIdxImm8s4Operand"; + let ParserMatchClass = PostIdxImm8s4AsmOperand; + let MIOperandInfo = (ops i32imm); +} + + +// postidx_reg := +/- reg +// +def PostIdxRegAsmOperand : AsmOperandClass { + let Name = "PostIdxReg"; + let ParserMethod = "parsePostIdxReg"; +} +def postidx_reg : MemOperand { + let EncoderMethod = "getPostIdxRegOpValue"; + let DecoderMethod = "DecodePostIdxReg"; + let PrintMethod = "printPostIdxRegOperand"; + let ParserMatchClass = PostIdxRegAsmOperand; + let MIOperandInfo = (ops GPRnopc, i32imm); +} + +def PostIdxRegShiftedAsmOperand : AsmOperandClass { + let Name = "PostIdxRegShifted"; + let ParserMethod = "parsePostIdxReg"; +} +def am2offset_reg : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode2OffsetReg", + [], [SDNPWantRoot]> { + let EncoderMethod = "getAddrMode2OffsetOpValue"; + let PrintMethod = "printAddrMode2OffsetOperand"; + // When using this for assembly, it's always as a post-index offset. + let ParserMatchClass = PostIdxRegShiftedAsmOperand; + let MIOperandInfo = (ops GPRnopc, i32imm); +} + +// FIXME: am2offset_imm should only need the immediate, not the GPR. Having +// the GPR is purely vestigal at this point. +def AM2OffsetImmAsmOperand : AsmOperandClass { let Name = "AM2OffsetImm"; } +def am2offset_imm : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode2OffsetImm", + [], [SDNPWantRoot]> { + let EncoderMethod = "getAddrMode2OffsetOpValue"; + let PrintMethod = "printAddrMode2OffsetOperand"; + let ParserMatchClass = AM2OffsetImmAsmOperand; + let MIOperandInfo = (ops GPRnopc, i32imm); +} + + +// addrmode3 := reg +/- reg +// addrmode3 := reg +/- imm8 +// +// FIXME: split into imm vs. reg versions. +def AddrMode3AsmOperand : AsmOperandClass { let Name = "AddrMode3"; } +class AddrMode3 : MemOperand, + ComplexPattern<i32, 3, "SelectAddrMode3", []> { + let EncoderMethod = "getAddrMode3OpValue"; + let ParserMatchClass = AddrMode3AsmOperand; + let MIOperandInfo = (ops GPR:$base, GPR:$offsreg, i32imm:$offsimm); +} + +def addrmode3 : AddrMode3 +{ + let PrintMethod = "printAddrMode3Operand<false>"; +} + +def addrmode3_pre : AddrMode3 +{ + let PrintMethod = "printAddrMode3Operand<true>"; +} + +// FIXME: split into imm vs. reg versions. +// FIXME: parser method to handle +/- register. +def AM3OffsetAsmOperand : AsmOperandClass { + let Name = "AM3Offset"; + let ParserMethod = "parseAM3Offset"; +} +def am3offset : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode3Offset", + [], [SDNPWantRoot]> { + let EncoderMethod = "getAddrMode3OffsetOpValue"; + let PrintMethod = "printAddrMode3OffsetOperand"; + let ParserMatchClass = AM3OffsetAsmOperand; + let MIOperandInfo = (ops GPR, i32imm); +} + +// ldstm_mode := {ia, ib, da, db} +// +def ldstm_mode : OptionalDefOperand<OtherVT, (ops i32), (ops (i32 1))> { + let EncoderMethod = "getLdStmModeOpValue"; + let PrintMethod = "printLdStmModeOperand"; +} + +// addrmode5 := reg +/- imm8*4 +// +def AddrMode5AsmOperand : AsmOperandClass { let Name = "AddrMode5"; } +class AddrMode5 : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode5", []> { + let EncoderMethod = "getAddrMode5OpValue"; + let DecoderMethod = "DecodeAddrMode5Operand"; + let ParserMatchClass = AddrMode5AsmOperand; + let MIOperandInfo = (ops GPR:$base, i32imm); +} + +def addrmode5 : AddrMode5 { + let PrintMethod = "printAddrMode5Operand<false>"; +} + +def addrmode5_pre : AddrMode5 { + let PrintMethod = "printAddrMode5Operand<true>"; +} + +// addrmode5fp16 := reg +/- imm8*2 +// +def AddrMode5FP16AsmOperand : AsmOperandClass { let Name = "AddrMode5FP16"; } +class AddrMode5FP16 : Operand<i32>, + ComplexPattern<i32, 2, "SelectAddrMode5FP16", []> { + let EncoderMethod = "getAddrMode5FP16OpValue"; + let DecoderMethod = "DecodeAddrMode5FP16Operand"; + let ParserMatchClass = AddrMode5FP16AsmOperand; + let MIOperandInfo = (ops GPR:$base, i32imm); +} + +def addrmode5fp16 : AddrMode5FP16 { + let PrintMethod = "printAddrMode5FP16Operand<false>"; +} + +// addrmode6 := reg with optional alignment +// +def AddrMode6AsmOperand : AsmOperandClass { let Name = "AlignedMemory"; } +def addrmode6 : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode6", [], [SDNPWantParent]>{ + let PrintMethod = "printAddrMode6Operand"; + let MIOperandInfo = (ops GPR:$addr, i32imm:$align); + let EncoderMethod = "getAddrMode6AddressOpValue"; + let DecoderMethod = "DecodeAddrMode6Operand"; + let ParserMatchClass = AddrMode6AsmOperand; +} + +def am6offset : MemOperand, + ComplexPattern<i32, 1, "SelectAddrMode6Offset", + [], [SDNPWantRoot]> { + let PrintMethod = "printAddrMode6OffsetOperand"; + let MIOperandInfo = (ops GPR); + let EncoderMethod = "getAddrMode6OffsetOpValue"; + let DecoderMethod = "DecodeGPRRegisterClass"; +} + +// Special version of addrmode6 to handle alignment encoding for VST1/VLD1 +// (single element from one lane) for size 32. +def addrmode6oneL32 : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode6", [], [SDNPWantParent]>{ + let PrintMethod = "printAddrMode6Operand"; + let MIOperandInfo = (ops GPR:$addr, i32imm); + let EncoderMethod = "getAddrMode6OneLane32AddressOpValue"; +} + +// Base class for addrmode6 with specific alignment restrictions. +class AddrMode6Align : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode6", [], [SDNPWantParent]>{ + let PrintMethod = "printAddrMode6Operand"; + let MIOperandInfo = (ops GPR:$addr, i32imm:$align); + let EncoderMethod = "getAddrMode6AddressOpValue"; + let DecoderMethod = "DecodeAddrMode6Operand"; +} + +// Special version of addrmode6 to handle no allowed alignment encoding for +// VLD/VST instructions and checking the alignment is not specified. +def AddrMode6AlignNoneAsmOperand : AsmOperandClass { + let Name = "AlignedMemoryNone"; + let DiagnosticString = "alignment must be omitted"; +} +def addrmode6alignNone : AddrMode6Align { + // The alignment specifier can only be omitted. + let ParserMatchClass = AddrMode6AlignNoneAsmOperand; +} + +// Special version of addrmode6 to handle 16-bit alignment encoding for +// VLD/VST instructions and checking the alignment value. +def AddrMode6Align16AsmOperand : AsmOperandClass { + let Name = "AlignedMemory16"; + let DiagnosticString = "alignment must be 16 or omitted"; +} +def addrmode6align16 : AddrMode6Align { + // The alignment specifier can only be 16 or omitted. + let ParserMatchClass = AddrMode6Align16AsmOperand; +} + +// Special version of addrmode6 to handle 32-bit alignment encoding for +// VLD/VST instructions and checking the alignment value. +def AddrMode6Align32AsmOperand : AsmOperandClass { + let Name = "AlignedMemory32"; + let DiagnosticString = "alignment must be 32 or omitted"; +} +def addrmode6align32 : AddrMode6Align { + // The alignment specifier can only be 32 or omitted. + let ParserMatchClass = AddrMode6Align32AsmOperand; +} + +// Special version of addrmode6 to handle 64-bit alignment encoding for +// VLD/VST instructions and checking the alignment value. +def AddrMode6Align64AsmOperand : AsmOperandClass { + let Name = "AlignedMemory64"; + let DiagnosticString = "alignment must be 64 or omitted"; +} +def addrmode6align64 : AddrMode6Align { + // The alignment specifier can only be 64 or omitted. + let ParserMatchClass = AddrMode6Align64AsmOperand; +} + +// Special version of addrmode6 to handle 64-bit or 128-bit alignment encoding +// for VLD/VST instructions and checking the alignment value. +def AddrMode6Align64or128AsmOperand : AsmOperandClass { + let Name = "AlignedMemory64or128"; + let DiagnosticString = "alignment must be 64, 128 or omitted"; +} +def addrmode6align64or128 : AddrMode6Align { + // The alignment specifier can only be 64, 128 or omitted. + let ParserMatchClass = AddrMode6Align64or128AsmOperand; +} + +// Special version of addrmode6 to handle 64-bit, 128-bit or 256-bit alignment +// encoding for VLD/VST instructions and checking the alignment value. +def AddrMode6Align64or128or256AsmOperand : AsmOperandClass { + let Name = "AlignedMemory64or128or256"; + let DiagnosticString = "alignment must be 64, 128, 256 or omitted"; +} +def addrmode6align64or128or256 : AddrMode6Align { + // The alignment specifier can only be 64, 128, 256 or omitted. + let ParserMatchClass = AddrMode6Align64or128or256AsmOperand; +} + +// Special version of addrmode6 to handle alignment encoding for VLD-dup +// instructions, specifically VLD4-dup. +def addrmode6dup : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode6", [], [SDNPWantParent]>{ + let PrintMethod = "printAddrMode6Operand"; + let MIOperandInfo = (ops GPR:$addr, i32imm); + let EncoderMethod = "getAddrMode6DupAddressOpValue"; + // FIXME: This is close, but not quite right. The alignment specifier is + // different. + let ParserMatchClass = AddrMode6AsmOperand; +} + +// Base class for addrmode6dup with specific alignment restrictions. +class AddrMode6DupAlign : MemOperand, + ComplexPattern<i32, 2, "SelectAddrMode6", [], [SDNPWantParent]>{ + let PrintMethod = "printAddrMode6Operand"; + let MIOperandInfo = (ops GPR:$addr, i32imm); + let EncoderMethod = "getAddrMode6DupAddressOpValue"; +} + +// Special version of addrmode6 to handle no allowed alignment encoding for +// VLD-dup instruction and checking the alignment is not specified. +def AddrMode6dupAlignNoneAsmOperand : AsmOperandClass { + let Name = "DupAlignedMemoryNone"; + let DiagnosticString = "alignment must be omitted"; +} +def addrmode6dupalignNone : AddrMode6DupAlign { + // The alignment specifier can only be omitted. + let ParserMatchClass = AddrMode6dupAlignNoneAsmOperand; +} + +// Special version of addrmode6 to handle 16-bit alignment encoding for VLD-dup +// instruction and checking the alignment value. +def AddrMode6dupAlign16AsmOperand : AsmOperandClass { + let Name = "DupAlignedMemory16"; + let DiagnosticString = "alignment must be 16 or omitted"; +} +def addrmode6dupalign16 : AddrMode6DupAlign { + // The alignment specifier can only be 16 or omitted. + let ParserMatchClass = AddrMode6dupAlign16AsmOperand; +} + +// Special version of addrmode6 to handle 32-bit alignment encoding for VLD-dup +// instruction and checking the alignment value. +def AddrMode6dupAlign32AsmOperand : AsmOperandClass { + let Name = "DupAlignedMemory32"; + let DiagnosticString = "alignment must be 32 or omitted"; +} +def addrmode6dupalign32 : AddrMode6DupAlign { + // The alignment specifier can only be 32 or omitted. + let ParserMatchClass = AddrMode6dupAlign32AsmOperand; +} + +// Special version of addrmode6 to handle 64-bit alignment encoding for VLD +// instructions and checking the alignment value. +def AddrMode6dupAlign64AsmOperand : AsmOperandClass { + let Name = "DupAlignedMemory64"; + let DiagnosticString = "alignment must be 64 or omitted"; +} +def addrmode6dupalign64 : AddrMode6DupAlign { + // The alignment specifier can only be 64 or omitted. + let ParserMatchClass = AddrMode6dupAlign64AsmOperand; +} + +// Special version of addrmode6 to handle 64-bit or 128-bit alignment encoding +// for VLD instructions and checking the alignment value. +def AddrMode6dupAlign64or128AsmOperand : AsmOperandClass { + let Name = "DupAlignedMemory64or128"; + let DiagnosticString = "alignment must be 64, 128 or omitted"; +} +def addrmode6dupalign64or128 : AddrMode6DupAlign { + // The alignment specifier can only be 64, 128 or omitted. + let ParserMatchClass = AddrMode6dupAlign64or128AsmOperand; +} + +// addrmodepc := pc + reg +// +def addrmodepc : MemOperand, + ComplexPattern<i32, 2, "SelectAddrModePC", []> { + let PrintMethod = "printAddrModePCOperand"; + let MIOperandInfo = (ops GPR, i32imm); +} + +// addr_offset_none := reg +// +def MemNoOffsetAsmOperand : AsmOperandClass { let Name = "MemNoOffset"; } +def addr_offset_none : MemOperand, + ComplexPattern<i32, 1, "SelectAddrOffsetNone", []> { + let PrintMethod = "printAddrMode7Operand"; + let DecoderMethod = "DecodeAddrMode7Operand"; + let ParserMatchClass = MemNoOffsetAsmOperand; + let MIOperandInfo = (ops GPR:$base); +} + +def nohash_imm : Operand<i32> { + let PrintMethod = "printNoHashImmediate"; +} + +def CoprocNumAsmOperand : AsmOperandClass { + let Name = "CoprocNum"; + let ParserMethod = "parseCoprocNumOperand"; +} +def p_imm : Operand<i32> { + let PrintMethod = "printPImmediate"; + let ParserMatchClass = CoprocNumAsmOperand; + let DecoderMethod = "DecodeCoprocessor"; +} + +def CoprocRegAsmOperand : AsmOperandClass { + let Name = "CoprocReg"; + let ParserMethod = "parseCoprocRegOperand"; +} +def c_imm : Operand<i32> { + let PrintMethod = "printCImmediate"; + let ParserMatchClass = CoprocRegAsmOperand; +} +def CoprocOptionAsmOperand : AsmOperandClass { + let Name = "CoprocOption"; + let ParserMethod = "parseCoprocOptionOperand"; +} +def coproc_option_imm : Operand<i32> { + let PrintMethod = "printCoprocOptionImm"; + let ParserMatchClass = CoprocOptionAsmOperand; +} + +//===----------------------------------------------------------------------===// + +include "ARMInstrFormats.td" + +//===----------------------------------------------------------------------===// +// Multiclass helpers... +// + +/// AsI1_bin_irs - Defines a set of (op r, {mod_imm|r|so_reg}) patterns for a +/// binop that produces a value. +let TwoOperandAliasConstraint = "$Rn = $Rd" in +multiclass AsI1_bin_irs<bits<4> opcod, string opc, + InstrItinClass iii, InstrItinClass iir, InstrItinClass iis, + SDPatternOperator opnode, bit Commutable = 0> { + // The register-immediate version is re-materializable. This is useful + // in particular for taking the address of a local. + let isReMaterializable = 1 in { + def ri : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, mod_imm:$imm), DPFrm, + iii, opc, "\t$Rd, $Rn, $imm", + [(set GPR:$Rd, (opnode GPR:$Rn, mod_imm:$imm))]>, + Sched<[WriteALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> imm; + let Inst{25} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-0} = imm; + } + } + def rr : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), DPFrm, + iir, opc, "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (opnode GPR:$Rn, GPR:$Rm))]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + let Inst{25} = 0; + let isCommutable = Commutable; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-4} = 0b00000000; + let Inst{3-0} = Rm; + } + + def rsi : AsI1<opcod, (outs GPR:$Rd), + (ins GPR:$Rn, so_reg_imm:$shift), DPSoRegImmFrm, + iis, opc, "\t$Rd, $Rn, $shift", + [(set GPR:$Rd, (opnode GPR:$Rn, so_reg_imm:$shift))]>, + Sched<[WriteALUsi, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-5} = shift{11-5}; + let Inst{4} = 0; + let Inst{3-0} = shift{3-0}; + } + + def rsr : AsI1<opcod, (outs GPR:$Rd), + (ins GPR:$Rn, so_reg_reg:$shift), DPSoRegRegFrm, + iis, opc, "\t$Rd, $Rn, $shift", + [(set GPR:$Rd, (opnode GPR:$Rn, so_reg_reg:$shift))]>, + Sched<[WriteALUsr, ReadALUsr]> { + bits<4> Rd; + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-8} = shift{11-8}; + let Inst{7} = 0; + let Inst{6-5} = shift{6-5}; + let Inst{4} = 1; + let Inst{3-0} = shift{3-0}; + } +} + +/// AsI1_rbin_irs - Same as AsI1_bin_irs except the order of operands are +/// reversed. The 'rr' form is only defined for the disassembler; for codegen +/// it is equivalent to the AsI1_bin_irs counterpart. +let TwoOperandAliasConstraint = "$Rn = $Rd" in +multiclass AsI1_rbin_irs<bits<4> opcod, string opc, + InstrItinClass iii, InstrItinClass iir, InstrItinClass iis, + SDNode opnode, bit Commutable = 0> { + // The register-immediate version is re-materializable. This is useful + // in particular for taking the address of a local. + let isReMaterializable = 1 in { + def ri : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, mod_imm:$imm), DPFrm, + iii, opc, "\t$Rd, $Rn, $imm", + [(set GPR:$Rd, (opnode mod_imm:$imm, GPR:$Rn))]>, + Sched<[WriteALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> imm; + let Inst{25} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-0} = imm; + } + } + def rr : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), DPFrm, + iir, opc, "\t$Rd, $Rn, $Rm", + [/* pattern left blank */]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + let Inst{11-4} = 0b00000000; + let Inst{25} = 0; + let Inst{3-0} = Rm; + let Inst{15-12} = Rd; + let Inst{19-16} = Rn; + } + + def rsi : AsI1<opcod, (outs GPR:$Rd), + (ins GPR:$Rn, so_reg_imm:$shift), DPSoRegImmFrm, + iis, opc, "\t$Rd, $Rn, $shift", + [(set GPR:$Rd, (opnode so_reg_imm:$shift, GPR:$Rn))]>, + Sched<[WriteALUsi, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-5} = shift{11-5}; + let Inst{4} = 0; + let Inst{3-0} = shift{3-0}; + } + + def rsr : AsI1<opcod, (outs GPR:$Rd), + (ins GPR:$Rn, so_reg_reg:$shift), DPSoRegRegFrm, + iis, opc, "\t$Rd, $Rn, $shift", + [(set GPR:$Rd, (opnode so_reg_reg:$shift, GPR:$Rn))]>, + Sched<[WriteALUsr, ReadALUsr]> { + bits<4> Rd; + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-8} = shift{11-8}; + let Inst{7} = 0; + let Inst{6-5} = shift{6-5}; + let Inst{4} = 1; + let Inst{3-0} = shift{3-0}; + } +} + +/// AsI1_bin_s_irs - Same as AsI1_bin_irs except it sets the 's' bit by default. +/// +/// These opcodes will be converted to the real non-S opcodes by +/// AdjustInstrPostInstrSelection after giving them an optional CPSR operand. +let hasPostISelHook = 1, Defs = [CPSR] in { +multiclass AsI1_bin_s_irs<InstrItinClass iii, InstrItinClass iir, + InstrItinClass iis, SDNode opnode, + bit Commutable = 0> { + def ri : ARMPseudoInst<(outs GPR:$Rd), (ins GPR:$Rn, mod_imm:$imm, pred:$p), + 4, iii, + [(set GPR:$Rd, CPSR, (opnode GPR:$Rn, mod_imm:$imm))]>, + Sched<[WriteALU, ReadALU]>; + + def rr : ARMPseudoInst<(outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm, pred:$p), + 4, iir, + [(set GPR:$Rd, CPSR, (opnode GPR:$Rn, GPR:$Rm))]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + let isCommutable = Commutable; + } + def rsi : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$Rn, so_reg_imm:$shift, pred:$p), + 4, iis, + [(set GPR:$Rd, CPSR, (opnode GPR:$Rn, + so_reg_imm:$shift))]>, + Sched<[WriteALUsi, ReadALU]>; + + def rsr : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$Rn, so_reg_reg:$shift, pred:$p), + 4, iis, + [(set GPR:$Rd, CPSR, (opnode GPR:$Rn, + so_reg_reg:$shift))]>, + Sched<[WriteALUSsr, ReadALUsr]>; +} +} + +/// AsI1_rbin_s_is - Same as AsI1_bin_s_irs, except selection DAG +/// operands are reversed. +let hasPostISelHook = 1, Defs = [CPSR] in { +multiclass AsI1_rbin_s_is<InstrItinClass iii, InstrItinClass iir, + InstrItinClass iis, SDNode opnode, + bit Commutable = 0> { + def ri : ARMPseudoInst<(outs GPR:$Rd), (ins GPR:$Rn, mod_imm:$imm, pred:$p), + 4, iii, + [(set GPR:$Rd, CPSR, (opnode mod_imm:$imm, GPR:$Rn))]>, + Sched<[WriteALU, ReadALU]>; + + def rsi : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$Rn, so_reg_imm:$shift, pred:$p), + 4, iis, + [(set GPR:$Rd, CPSR, (opnode so_reg_imm:$shift, + GPR:$Rn))]>, + Sched<[WriteALUsi, ReadALU]>; + + def rsr : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$Rn, so_reg_reg:$shift, pred:$p), + 4, iis, + [(set GPR:$Rd, CPSR, (opnode so_reg_reg:$shift, + GPR:$Rn))]>, + Sched<[WriteALUSsr, ReadALUsr]>; +} +} + +/// AI1_cmp_irs - Defines a set of (op r, {mod_imm|r|so_reg}) cmp / test +/// patterns. Similar to AsI1_bin_irs except the instruction does not produce +/// a explicit result, only implicitly set CPSR. +let isCompare = 1, Defs = [CPSR] in { +multiclass AI1_cmp_irs<bits<4> opcod, string opc, + InstrItinClass iii, InstrItinClass iir, InstrItinClass iis, + SDPatternOperator opnode, bit Commutable = 0, + string rrDecoderMethod = ""> { + def ri : AI1<opcod, (outs), (ins GPR:$Rn, mod_imm:$imm), DPFrm, iii, + opc, "\t$Rn, $imm", + [(opnode GPR:$Rn, mod_imm:$imm)]>, + Sched<[WriteCMP, ReadALU]> { + bits<4> Rn; + bits<12> imm; + let Inst{25} = 1; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b0000; + let Inst{11-0} = imm; + + let Unpredictable{15-12} = 0b1111; + } + def rr : AI1<opcod, (outs), (ins GPR:$Rn, GPR:$Rm), DPFrm, iir, + opc, "\t$Rn, $Rm", + [(opnode GPR:$Rn, GPR:$Rm)]>, + Sched<[WriteCMP, ReadALU, ReadALU]> { + bits<4> Rn; + bits<4> Rm; + let isCommutable = Commutable; + let Inst{25} = 0; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b0000; + let Inst{11-4} = 0b00000000; + let Inst{3-0} = Rm; + let DecoderMethod = rrDecoderMethod; + + let Unpredictable{15-12} = 0b1111; + } + def rsi : AI1<opcod, (outs), + (ins GPR:$Rn, so_reg_imm:$shift), DPSoRegImmFrm, iis, + opc, "\t$Rn, $shift", + [(opnode GPR:$Rn, so_reg_imm:$shift)]>, + Sched<[WriteCMPsi, ReadALU]> { + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b0000; + let Inst{11-5} = shift{11-5}; + let Inst{4} = 0; + let Inst{3-0} = shift{3-0}; + + let Unpredictable{15-12} = 0b1111; + } + def rsr : AI1<opcod, (outs), + (ins GPRnopc:$Rn, so_reg_reg:$shift), DPSoRegRegFrm, iis, + opc, "\t$Rn, $shift", + [(opnode GPRnopc:$Rn, so_reg_reg:$shift)]>, + Sched<[WriteCMPsr, ReadALU]> { + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b0000; + let Inst{11-8} = shift{11-8}; + let Inst{7} = 0; + let Inst{6-5} = shift{6-5}; + let Inst{4} = 1; + let Inst{3-0} = shift{3-0}; + + let Unpredictable{15-12} = 0b1111; + } + +} +} + +/// AI_ext_rrot - A unary operation with two forms: one whose operand is a +/// register and one whose operand is a register rotated by 8/16/24. +/// FIXME: Remove the 'r' variant. Its rot_imm is zero. +class AI_ext_rrot<bits<8> opcod, string opc, PatFrag opnode> + : AExtI<opcod, (outs GPRnopc:$Rd), (ins GPRnopc:$Rm, rot_imm:$rot), + IIC_iEXTr, opc, "\t$Rd, $Rm$rot", + [(set GPRnopc:$Rd, (opnode (rotr GPRnopc:$Rm, rot_imm:$rot)))]>, + Requires<[IsARM, HasV6]>, Sched<[WriteALUsi]> { + bits<4> Rd; + bits<4> Rm; + bits<2> rot; + let Inst{19-16} = 0b1111; + let Inst{15-12} = Rd; + let Inst{11-10} = rot; + let Inst{3-0} = Rm; +} + +class AI_ext_rrot_np<bits<8> opcod, string opc> + : AExtI<opcod, (outs GPRnopc:$Rd), (ins GPRnopc:$Rm, rot_imm:$rot), + IIC_iEXTr, opc, "\t$Rd, $Rm$rot", []>, + Requires<[IsARM, HasV6]>, Sched<[WriteALUsi]> { + bits<2> rot; + let Inst{19-16} = 0b1111; + let Inst{11-10} = rot; + } + +/// AI_exta_rrot - A binary operation with two forms: one whose operand is a +/// register and one whose operand is a register rotated by 8/16/24. +class AI_exta_rrot<bits<8> opcod, string opc, PatFrag opnode> + : AExtI<opcod, (outs GPRnopc:$Rd), (ins GPR:$Rn, GPRnopc:$Rm, rot_imm:$rot), + IIC_iEXTAr, opc, "\t$Rd, $Rn, $Rm$rot", + [(set GPRnopc:$Rd, (opnode GPR:$Rn, + (rotr GPRnopc:$Rm, rot_imm:$rot)))]>, + Requires<[IsARM, HasV6]>, Sched<[WriteALUsr]> { + bits<4> Rd; + bits<4> Rm; + bits<4> Rn; + bits<2> rot; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-10} = rot; + let Inst{9-4} = 0b000111; + let Inst{3-0} = Rm; +} + +class AI_exta_rrot_np<bits<8> opcod, string opc> + : AExtI<opcod, (outs GPRnopc:$Rd), (ins GPR:$Rn, GPRnopc:$Rm, rot_imm:$rot), + IIC_iEXTAr, opc, "\t$Rd, $Rn, $Rm$rot", []>, + Requires<[IsARM, HasV6]>, Sched<[WriteALUsr]> { + bits<4> Rn; + bits<2> rot; + let Inst{19-16} = Rn; + let Inst{11-10} = rot; +} + +/// AI1_adde_sube_irs - Define instructions and patterns for adde and sube. +let TwoOperandAliasConstraint = "$Rn = $Rd" in +multiclass AI1_adde_sube_irs<bits<4> opcod, string opc, SDNode opnode, + bit Commutable = 0> { + let hasPostISelHook = 1, Defs = [CPSR], Uses = [CPSR] in { + def ri : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, mod_imm:$imm), + DPFrm, IIC_iALUi, opc, "\t$Rd, $Rn, $imm", + [(set GPR:$Rd, CPSR, (opnode GPR:$Rn, mod_imm:$imm, CPSR))]>, + Requires<[IsARM]>, + Sched<[WriteALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> imm; + let Inst{25} = 1; + let Inst{15-12} = Rd; + let Inst{19-16} = Rn; + let Inst{11-0} = imm; + } + def rr : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + DPFrm, IIC_iALUr, opc, "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, CPSR, (opnode GPR:$Rn, GPR:$Rm, CPSR))]>, + Requires<[IsARM]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + let Inst{11-4} = 0b00000000; + let Inst{25} = 0; + let isCommutable = Commutable; + let Inst{3-0} = Rm; + let Inst{15-12} = Rd; + let Inst{19-16} = Rn; + } + def rsi : AsI1<opcod, (outs GPR:$Rd), + (ins GPR:$Rn, so_reg_imm:$shift), + DPSoRegImmFrm, IIC_iALUsr, opc, "\t$Rd, $Rn, $shift", + [(set GPR:$Rd, CPSR, (opnode GPR:$Rn, so_reg_imm:$shift, CPSR))]>, + Requires<[IsARM]>, + Sched<[WriteALUsi, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-5} = shift{11-5}; + let Inst{4} = 0; + let Inst{3-0} = shift{3-0}; + } + def rsr : AsI1<opcod, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, so_reg_reg:$shift), + DPSoRegRegFrm, IIC_iALUsr, opc, "\t$Rd, $Rn, $shift", + [(set GPRnopc:$Rd, CPSR, + (opnode GPRnopc:$Rn, so_reg_reg:$shift, CPSR))]>, + Requires<[IsARM]>, + Sched<[WriteALUsr, ReadALUsr]> { + bits<4> Rd; + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-8} = shift{11-8}; + let Inst{7} = 0; + let Inst{6-5} = shift{6-5}; + let Inst{4} = 1; + let Inst{3-0} = shift{3-0}; + } + } +} + +/// AI1_rsc_irs - Define instructions and patterns for rsc +let TwoOperandAliasConstraint = "$Rn = $Rd" in +multiclass AI1_rsc_irs<bits<4> opcod, string opc, SDNode opnode> { + let hasPostISelHook = 1, Defs = [CPSR], Uses = [CPSR] in { + def ri : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, mod_imm:$imm), + DPFrm, IIC_iALUi, opc, "\t$Rd, $Rn, $imm", + [(set GPR:$Rd, CPSR, (opnode mod_imm:$imm, GPR:$Rn, CPSR))]>, + Requires<[IsARM]>, + Sched<[WriteALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> imm; + let Inst{25} = 1; + let Inst{15-12} = Rd; + let Inst{19-16} = Rn; + let Inst{11-0} = imm; + } + def rr : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + DPFrm, IIC_iALUr, opc, "\t$Rd, $Rn, $Rm", + [/* pattern left blank */]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + let Inst{11-4} = 0b00000000; + let Inst{25} = 0; + let Inst{3-0} = Rm; + let Inst{15-12} = Rd; + let Inst{19-16} = Rn; + } + def rsi : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, so_reg_imm:$shift), + DPSoRegImmFrm, IIC_iALUsr, opc, "\t$Rd, $Rn, $shift", + [(set GPR:$Rd, CPSR, (opnode so_reg_imm:$shift, GPR:$Rn, CPSR))]>, + Requires<[IsARM]>, + Sched<[WriteALUsi, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-5} = shift{11-5}; + let Inst{4} = 0; + let Inst{3-0} = shift{3-0}; + } + def rsr : AsI1<opcod, (outs GPR:$Rd), (ins GPR:$Rn, so_reg_reg:$shift), + DPSoRegRegFrm, IIC_iALUsr, opc, "\t$Rd, $Rn, $shift", + [(set GPR:$Rd, CPSR, (opnode so_reg_reg:$shift, GPR:$Rn, CPSR))]>, + Requires<[IsARM]>, + Sched<[WriteALUsr, ReadALUsr]> { + bits<4> Rd; + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-8} = shift{11-8}; + let Inst{7} = 0; + let Inst{6-5} = shift{6-5}; + let Inst{4} = 1; + let Inst{3-0} = shift{3-0}; + } + } +} + +let canFoldAsLoad = 1, isReMaterializable = 1 in { +multiclass AI_ldr1<bit isByte, string opc, InstrItinClass iii, + InstrItinClass iir, PatFrag opnode> { + // Note: We use the complex addrmode_imm12 rather than just an input + // GPR and a constrained immediate so that we can use this to match + // frame index references and avoid matching constant pool references. + def i12: AI2ldst<0b010, 1, isByte, (outs GPR:$Rt), (ins addrmode_imm12:$addr), + AddrMode_i12, LdFrm, iii, opc, "\t$Rt, $addr", + [(set GPR:$Rt, (opnode addrmode_imm12:$addr))]> { + bits<4> Rt; + bits<17> addr; + let Inst{23} = addr{12}; // U (add = ('U' == 1)) + let Inst{19-16} = addr{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = addr{11-0}; // imm12 + } + def rs : AI2ldst<0b011, 1, isByte, (outs GPR:$Rt), (ins ldst_so_reg:$shift), + AddrModeNone, LdFrm, iir, opc, "\t$Rt, $shift", + [(set GPR:$Rt, (opnode ldst_so_reg:$shift))]> { + bits<4> Rt; + bits<17> shift; + let shift{4} = 0; // Inst{4} = 0 + let Inst{23} = shift{12}; // U (add = ('U' == 1)) + let Inst{19-16} = shift{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = shift{11-0}; + } +} +} + +let canFoldAsLoad = 1, isReMaterializable = 1 in { +multiclass AI_ldr1nopc<bit isByte, string opc, InstrItinClass iii, + InstrItinClass iir, PatFrag opnode> { + // Note: We use the complex addrmode_imm12 rather than just an input + // GPR and a constrained immediate so that we can use this to match + // frame index references and avoid matching constant pool references. + def i12: AI2ldst<0b010, 1, isByte, (outs GPRnopc:$Rt), + (ins addrmode_imm12:$addr), + AddrMode_i12, LdFrm, iii, opc, "\t$Rt, $addr", + [(set GPRnopc:$Rt, (opnode addrmode_imm12:$addr))]> { + bits<4> Rt; + bits<17> addr; + let Inst{23} = addr{12}; // U (add = ('U' == 1)) + let Inst{19-16} = addr{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = addr{11-0}; // imm12 + } + def rs : AI2ldst<0b011, 1, isByte, (outs GPRnopc:$Rt), + (ins ldst_so_reg:$shift), + AddrModeNone, LdFrm, iir, opc, "\t$Rt, $shift", + [(set GPRnopc:$Rt, (opnode ldst_so_reg:$shift))]> { + bits<4> Rt; + bits<17> shift; + let shift{4} = 0; // Inst{4} = 0 + let Inst{23} = shift{12}; // U (add = ('U' == 1)) + let Inst{19-16} = shift{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = shift{11-0}; + } +} +} + + +multiclass AI_str1<bit isByte, string opc, InstrItinClass iii, + InstrItinClass iir, PatFrag opnode> { + // Note: We use the complex addrmode_imm12 rather than just an input + // GPR and a constrained immediate so that we can use this to match + // frame index references and avoid matching constant pool references. + def i12 : AI2ldst<0b010, 0, isByte, (outs), + (ins GPR:$Rt, addrmode_imm12:$addr), + AddrMode_i12, StFrm, iii, opc, "\t$Rt, $addr", + [(opnode GPR:$Rt, addrmode_imm12:$addr)]> { + bits<4> Rt; + bits<17> addr; + let Inst{23} = addr{12}; // U (add = ('U' == 1)) + let Inst{19-16} = addr{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = addr{11-0}; // imm12 + } + def rs : AI2ldst<0b011, 0, isByte, (outs), (ins GPR:$Rt, ldst_so_reg:$shift), + AddrModeNone, StFrm, iir, opc, "\t$Rt, $shift", + [(opnode GPR:$Rt, ldst_so_reg:$shift)]> { + bits<4> Rt; + bits<17> shift; + let shift{4} = 0; // Inst{4} = 0 + let Inst{23} = shift{12}; // U (add = ('U' == 1)) + let Inst{19-16} = shift{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = shift{11-0}; + } +} + +multiclass AI_str1nopc<bit isByte, string opc, InstrItinClass iii, + InstrItinClass iir, PatFrag opnode> { + // Note: We use the complex addrmode_imm12 rather than just an input + // GPR and a constrained immediate so that we can use this to match + // frame index references and avoid matching constant pool references. + def i12 : AI2ldst<0b010, 0, isByte, (outs), + (ins GPRnopc:$Rt, addrmode_imm12:$addr), + AddrMode_i12, StFrm, iii, opc, "\t$Rt, $addr", + [(opnode GPRnopc:$Rt, addrmode_imm12:$addr)]> { + bits<4> Rt; + bits<17> addr; + let Inst{23} = addr{12}; // U (add = ('U' == 1)) + let Inst{19-16} = addr{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = addr{11-0}; // imm12 + } + def rs : AI2ldst<0b011, 0, isByte, (outs), + (ins GPRnopc:$Rt, ldst_so_reg:$shift), + AddrModeNone, StFrm, iir, opc, "\t$Rt, $shift", + [(opnode GPRnopc:$Rt, ldst_so_reg:$shift)]> { + bits<4> Rt; + bits<17> shift; + let shift{4} = 0; // Inst{4} = 0 + let Inst{23} = shift{12}; // U (add = ('U' == 1)) + let Inst{19-16} = shift{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = shift{11-0}; + } +} + + +//===----------------------------------------------------------------------===// +// Instructions +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// Miscellaneous Instructions. +// + +/// CONSTPOOL_ENTRY - This instruction represents a floating constant pool in +/// the function. The first operand is the ID# for this instruction, the second +/// is the index into the MachineConstantPool that this is, the third is the +/// size in bytes of this constant pool entry. +let hasSideEffects = 0, isNotDuplicable = 1 in +def CONSTPOOL_ENTRY : +PseudoInst<(outs), (ins cpinst_operand:$instid, cpinst_operand:$cpidx, + i32imm:$size), NoItinerary, []>; + +/// A jumptable consisting of direct 32-bit addresses of the destination basic +/// blocks (either absolute, or relative to the start of the jump-table in PIC +/// mode). Used mostly in ARM and Thumb-1 modes. +def JUMPTABLE_ADDRS : +PseudoInst<(outs), (ins cpinst_operand:$instid, cpinst_operand:$cpidx, + i32imm:$size), NoItinerary, []>; + +/// A jumptable consisting of 32-bit jump instructions. Used for Thumb-2 tables +/// that cannot be optimised to use TBB or TBH. +def JUMPTABLE_INSTS : +PseudoInst<(outs), (ins cpinst_operand:$instid, cpinst_operand:$cpidx, + i32imm:$size), NoItinerary, []>; + +/// A jumptable consisting of 8-bit unsigned integers representing offsets from +/// a TBB instruction. +def JUMPTABLE_TBB : +PseudoInst<(outs), (ins cpinst_operand:$instid, cpinst_operand:$cpidx, + i32imm:$size), NoItinerary, []>; + +/// A jumptable consisting of 16-bit unsigned integers representing offsets from +/// a TBH instruction. +def JUMPTABLE_TBH : +PseudoInst<(outs), (ins cpinst_operand:$instid, cpinst_operand:$cpidx, + i32imm:$size), NoItinerary, []>; + + +// FIXME: Marking these as hasSideEffects is necessary to prevent machine DCE +// from removing one half of the matched pairs. That breaks PEI, which assumes +// these will always be in pairs, and asserts if it finds otherwise. Better way? +let Defs = [SP], Uses = [SP], hasSideEffects = 1 in { +def ADJCALLSTACKUP : +PseudoInst<(outs), (ins i32imm:$amt1, i32imm:$amt2, pred:$p), NoItinerary, + [(ARMcallseq_end timm:$amt1, timm:$amt2)]>; + +def ADJCALLSTACKDOWN : +PseudoInst<(outs), (ins i32imm:$amt, i32imm:$amt2, pred:$p), NoItinerary, + [(ARMcallseq_start timm:$amt, timm:$amt2)]>; +} + +def HINT : AI<(outs), (ins imm0_239:$imm), MiscFrm, NoItinerary, + "hint", "\t$imm", [(int_arm_hint imm0_239:$imm)]>, + Requires<[IsARM, HasV6]> { + bits<8> imm; + let Inst{27-8} = 0b00110010000011110000; + let Inst{7-0} = imm; + let DecoderMethod = "DecodeHINTInstruction"; +} + +def : InstAlias<"nop$p", (HINT 0, pred:$p)>, Requires<[IsARM, HasV6K]>; +def : InstAlias<"yield$p", (HINT 1, pred:$p)>, Requires<[IsARM, HasV6K]>; +def : InstAlias<"wfe$p", (HINT 2, pred:$p)>, Requires<[IsARM, HasV6K]>; +def : InstAlias<"wfi$p", (HINT 3, pred:$p)>, Requires<[IsARM, HasV6K]>; +def : InstAlias<"sev$p", (HINT 4, pred:$p)>, Requires<[IsARM, HasV6K]>; +def : InstAlias<"sevl$p", (HINT 5, pred:$p)>, Requires<[IsARM, HasV8]>; +def : InstAlias<"esb$p", (HINT 16, pred:$p)>, Requires<[IsARM, HasRAS]>; +def : InstAlias<"csdb$p", (HINT 20, pred:$p)>, Requires<[IsARM, HasV6K]>; + +def SEL : AI<(outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), DPFrm, NoItinerary, "sel", + "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (int_arm_sel GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasV6]> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + let Inst{3-0} = Rm; + let Inst{15-12} = Rd; + let Inst{19-16} = Rn; + let Inst{27-20} = 0b01101000; + let Inst{7-4} = 0b1011; + let Inst{11-8} = 0b1111; + let Unpredictable{11-8} = 0b1111; +} + +// The 16-bit operand $val can be used by a debugger to store more information +// about the breakpoint. +def BKPT : AInoP<(outs), (ins imm0_65535:$val), MiscFrm, NoItinerary, + "bkpt", "\t$val", []>, Requires<[IsARM]> { + bits<16> val; + let Inst{3-0} = val{3-0}; + let Inst{19-8} = val{15-4}; + let Inst{27-20} = 0b00010010; + let Inst{31-28} = 0xe; // AL + let Inst{7-4} = 0b0111; +} +// default immediate for breakpoint mnemonic +def : InstAlias<"bkpt", (BKPT 0), 0>, Requires<[IsARM]>; + +def HLT : AInoP<(outs), (ins imm0_65535:$val), MiscFrm, NoItinerary, + "hlt", "\t$val", []>, Requires<[IsARM, HasV8]> { + bits<16> val; + let Inst{3-0} = val{3-0}; + let Inst{19-8} = val{15-4}; + let Inst{27-20} = 0b00010000; + let Inst{31-28} = 0xe; // AL + let Inst{7-4} = 0b0111; +} + +// Change Processor State +// FIXME: We should use InstAlias to handle the optional operands. +class CPS<dag iops, string asm_ops> + : AXI<(outs), iops, MiscFrm, NoItinerary, !strconcat("cps", asm_ops), + []>, Requires<[IsARM]> { + bits<2> imod; + bits<3> iflags; + bits<5> mode; + bit M; + + let Inst{31-28} = 0b1111; + let Inst{27-20} = 0b00010000; + let Inst{19-18} = imod; + let Inst{17} = M; // Enabled if mode is set; + let Inst{16-9} = 0b00000000; + let Inst{8-6} = iflags; + let Inst{5} = 0; + let Inst{4-0} = mode; +} + +let DecoderMethod = "DecodeCPSInstruction" in { +let M = 1 in + def CPS3p : CPS<(ins imod_op:$imod, iflags_op:$iflags, imm0_31:$mode), + "$imod\t$iflags, $mode">; +let mode = 0, M = 0 in + def CPS2p : CPS<(ins imod_op:$imod, iflags_op:$iflags), "$imod\t$iflags">; + +let imod = 0, iflags = 0, M = 1 in + def CPS1p : CPS<(ins imm0_31:$mode), "\t$mode">; +} + +// Preload signals the memory system of possible future data/instruction access. +multiclass APreLoad<bits<1> read, bits<1> data, string opc> { + + def i12 : AXIM<(outs), (ins addrmode_imm12:$addr), AddrMode_i12, MiscFrm, + IIC_Preload, !strconcat(opc, "\t$addr"), + [(ARMPreload addrmode_imm12:$addr, (i32 read), (i32 data))]>, + Sched<[WritePreLd]> { + bits<4> Rt; + bits<17> addr; + let Inst{31-26} = 0b111101; + let Inst{25} = 0; // 0 for immediate form + let Inst{24} = data; + let Inst{23} = addr{12}; // U (add = ('U' == 1)) + let Inst{22} = read; + let Inst{21-20} = 0b01; + let Inst{19-16} = addr{16-13}; // Rn + let Inst{15-12} = 0b1111; + let Inst{11-0} = addr{11-0}; // imm12 + } + + def rs : AXI<(outs), (ins ldst_so_reg:$shift), MiscFrm, IIC_Preload, + !strconcat(opc, "\t$shift"), + [(ARMPreload ldst_so_reg:$shift, (i32 read), (i32 data))]>, + Sched<[WritePreLd]> { + bits<17> shift; + let Inst{31-26} = 0b111101; + let Inst{25} = 1; // 1 for register form + let Inst{24} = data; + let Inst{23} = shift{12}; // U (add = ('U' == 1)) + let Inst{22} = read; + let Inst{21-20} = 0b01; + let Inst{19-16} = shift{16-13}; // Rn + let Inst{15-12} = 0b1111; + let Inst{11-0} = shift{11-0}; + let Inst{4} = 0; + } +} + +defm PLD : APreLoad<1, 1, "pld">, Requires<[IsARM]>; +defm PLDW : APreLoad<0, 1, "pldw">, Requires<[IsARM,HasV7,HasMP]>; +defm PLI : APreLoad<1, 0, "pli">, Requires<[IsARM,HasV7]>; + +def SETEND : AXI<(outs), (ins setend_op:$end), MiscFrm, NoItinerary, + "setend\t$end", []>, Requires<[IsARM]>, Deprecated<HasV8Ops> { + bits<1> end; + let Inst{31-10} = 0b1111000100000001000000; + let Inst{9} = end; + let Inst{8-0} = 0; +} + +def DBG : AI<(outs), (ins imm0_15:$opt), MiscFrm, NoItinerary, "dbg", "\t$opt", + [(int_arm_dbg imm0_15:$opt)]>, Requires<[IsARM, HasV7]> { + bits<4> opt; + let Inst{27-4} = 0b001100100000111100001111; + let Inst{3-0} = opt; +} + +// A8.8.247 UDF - Undefined (Encoding A1) +def UDF : AInoP<(outs), (ins imm0_65535:$imm16), MiscFrm, NoItinerary, + "udf", "\t$imm16", [(int_arm_undefined imm0_65535:$imm16)]> { + bits<16> imm16; + let Inst{31-28} = 0b1110; // AL + let Inst{27-25} = 0b011; + let Inst{24-20} = 0b11111; + let Inst{19-8} = imm16{15-4}; + let Inst{7-4} = 0b1111; + let Inst{3-0} = imm16{3-0}; +} + +/* + * A5.4 Permanently UNDEFINED instructions. + * + * For most targets use UDF #65006, for which the OS will generate SIGTRAP. + * Other UDF encodings generate SIGILL. + * + * NaCl's OS instead chooses an ARM UDF encoding that's also a UDF in Thumb. + * Encoding A1: + * 1110 0111 1111 iiii iiii iiii 1111 iiii + * Encoding T1: + * 1101 1110 iiii iiii + * It uses the following encoding: + * 1110 0111 1111 1110 1101 1110 1111 0000 + * - In ARM: UDF #60896; + * - In Thumb: UDF #254 followed by a branch-to-self. + */ +let isBarrier = 1, isTerminator = 1 in +def TRAPNaCl : AXI<(outs), (ins), MiscFrm, NoItinerary, + "trap", [(trap)]>, + Requires<[IsARM,UseNaClTrap]> { + let Inst = 0xe7fedef0; +} +let isBarrier = 1, isTerminator = 1 in +def TRAP : AXI<(outs), (ins), MiscFrm, NoItinerary, + "trap", [(trap)]>, + Requires<[IsARM,DontUseNaClTrap]> { + let Inst = 0xe7ffdefe; +} + +// Address computation and loads and stores in PIC mode. +let isNotDuplicable = 1 in { +def PICADD : ARMPseudoInst<(outs GPR:$dst), (ins GPR:$a, pclabel:$cp, pred:$p), + 4, IIC_iALUr, + [(set GPR:$dst, (ARMpic_add GPR:$a, imm:$cp))]>, + Sched<[WriteALU, ReadALU]>; + +let AddedComplexity = 10 in { +def PICLDR : ARMPseudoInst<(outs GPR:$dst), (ins addrmodepc:$addr, pred:$p), + 4, IIC_iLoad_r, + [(set GPR:$dst, (load addrmodepc:$addr))]>; + +def PICLDRH : ARMPseudoInst<(outs GPR:$Rt), (ins addrmodepc:$addr, pred:$p), + 4, IIC_iLoad_bh_r, + [(set GPR:$Rt, (zextloadi16 addrmodepc:$addr))]>; + +def PICLDRB : ARMPseudoInst<(outs GPR:$Rt), (ins addrmodepc:$addr, pred:$p), + 4, IIC_iLoad_bh_r, + [(set GPR:$Rt, (zextloadi8 addrmodepc:$addr))]>; + +def PICLDRSH : ARMPseudoInst<(outs GPR:$Rt), (ins addrmodepc:$addr, pred:$p), + 4, IIC_iLoad_bh_r, + [(set GPR:$Rt, (sextloadi16 addrmodepc:$addr))]>; + +def PICLDRSB : ARMPseudoInst<(outs GPR:$Rt), (ins addrmodepc:$addr, pred:$p), + 4, IIC_iLoad_bh_r, + [(set GPR:$Rt, (sextloadi8 addrmodepc:$addr))]>; +} +let AddedComplexity = 10 in { +def PICSTR : ARMPseudoInst<(outs), (ins GPR:$src, addrmodepc:$addr, pred:$p), + 4, IIC_iStore_r, [(store GPR:$src, addrmodepc:$addr)]>; + +def PICSTRH : ARMPseudoInst<(outs), (ins GPR:$src, addrmodepc:$addr, pred:$p), + 4, IIC_iStore_bh_r, [(truncstorei16 GPR:$src, + addrmodepc:$addr)]>; + +def PICSTRB : ARMPseudoInst<(outs), (ins GPR:$src, addrmodepc:$addr, pred:$p), + 4, IIC_iStore_bh_r, [(truncstorei8 GPR:$src, addrmodepc:$addr)]>; +} +} // isNotDuplicable = 1 + + +// LEApcrel - Load a pc-relative address into a register without offending the +// assembler. +let hasSideEffects = 0, isReMaterializable = 1 in +// The 'adr' mnemonic encodes differently if the label is before or after +// the instruction. The {24-21} opcode bits are set by the fixup, as we don't +// know until then which form of the instruction will be used. +def ADR : AI1<{0,?,?,0}, (outs GPR:$Rd), (ins adrlabel:$label), + MiscFrm, IIC_iALUi, "adr", "\t$Rd, $label", []>, + Sched<[WriteALU, ReadALU]> { + bits<4> Rd; + bits<14> label; + let Inst{27-25} = 0b001; + let Inst{24} = 0; + let Inst{23-22} = label{13-12}; + let Inst{21} = 0; + let Inst{20} = 0; + let Inst{19-16} = 0b1111; + let Inst{15-12} = Rd; + let Inst{11-0} = label{11-0}; +} + +let hasSideEffects = 1 in { +def LEApcrel : ARMPseudoInst<(outs GPR:$Rd), (ins i32imm:$label, pred:$p), + 4, IIC_iALUi, []>, Sched<[WriteALU, ReadALU]>; + +def LEApcrelJT : ARMPseudoInst<(outs GPR:$Rd), + (ins i32imm:$label, pred:$p), + 4, IIC_iALUi, []>, Sched<[WriteALU, ReadALU]>; +} + +//===----------------------------------------------------------------------===// +// Control Flow Instructions. +// + +let isReturn = 1, isTerminator = 1, isBarrier = 1 in { + // ARMV4T and above + def BX_RET : AI<(outs), (ins), BrMiscFrm, IIC_Br, + "bx", "\tlr", [(ARMretflag)]>, + Requires<[IsARM, HasV4T]>, Sched<[WriteBr]> { + let Inst{27-0} = 0b0001001011111111111100011110; + } + + // ARMV4 only + def MOVPCLR : AI<(outs), (ins), BrMiscFrm, IIC_Br, + "mov", "\tpc, lr", [(ARMretflag)]>, + Requires<[IsARM, NoV4T]>, Sched<[WriteBr]> { + let Inst{27-0} = 0b0001101000001111000000001110; + } + + // Exception return: N.b. doesn't set CPSR as far as we're concerned (it sets + // the user-space one). + def SUBS_PC_LR : ARMPseudoInst<(outs), (ins i32imm:$offset, pred:$p), + 4, IIC_Br, + [(ARMintretflag imm:$offset)]>; +} + +// Indirect branches +let isBranch = 1, isTerminator = 1, isBarrier = 1, isIndirectBranch = 1 in { + // ARMV4T and above + def BX : AXI<(outs), (ins GPR:$dst), BrMiscFrm, IIC_Br, "bx\t$dst", + [(brind GPR:$dst)]>, + Requires<[IsARM, HasV4T]>, Sched<[WriteBr]> { + bits<4> dst; + let Inst{31-4} = 0b1110000100101111111111110001; + let Inst{3-0} = dst; + } + + def BX_pred : AI<(outs), (ins GPR:$dst), BrMiscFrm, IIC_Br, + "bx", "\t$dst", [/* pattern left blank */]>, + Requires<[IsARM, HasV4T]>, Sched<[WriteBr]> { + bits<4> dst; + let Inst{27-4} = 0b000100101111111111110001; + let Inst{3-0} = dst; + } +} + +// SP is marked as a use to prevent stack-pointer assignments that appear +// immediately before calls from potentially appearing dead. +let isCall = 1, + // FIXME: Do we really need a non-predicated version? If so, it should + // at least be a pseudo instruction expanding to the predicated version + // at MC lowering time. + Defs = [LR], Uses = [SP] in { + def BL : ABXI<0b1011, (outs), (ins arm_bl_target:$func), + IIC_Br, "bl\t$func", + [(ARMcall tglobaladdr:$func)]>, + Requires<[IsARM]>, Sched<[WriteBrL]> { + let Inst{31-28} = 0b1110; + bits<24> func; + let Inst{23-0} = func; + let DecoderMethod = "DecodeBranchImmInstruction"; + } + + def BL_pred : ABI<0b1011, (outs), (ins arm_bl_target:$func), + IIC_Br, "bl", "\t$func", + [(ARMcall_pred tglobaladdr:$func)]>, + Requires<[IsARM]>, Sched<[WriteBrL]> { + bits<24> func; + let Inst{23-0} = func; + let DecoderMethod = "DecodeBranchImmInstruction"; + } + + // ARMv5T and above + def BLX : AXI<(outs), (ins GPR:$func), BrMiscFrm, + IIC_Br, "blx\t$func", + [(ARMcall GPR:$func)]>, + Requires<[IsARM, HasV5T]>, Sched<[WriteBrL]> { + bits<4> func; + let Inst{31-4} = 0b1110000100101111111111110011; + let Inst{3-0} = func; + } + + def BLX_pred : AI<(outs), (ins GPR:$func), BrMiscFrm, + IIC_Br, "blx", "\t$func", + [(ARMcall_pred GPR:$func)]>, + Requires<[IsARM, HasV5T]>, Sched<[WriteBrL]> { + bits<4> func; + let Inst{27-4} = 0b000100101111111111110011; + let Inst{3-0} = func; + } + + // ARMv4T + // Note: Restrict $func to the tGPR regclass to prevent it being in LR. + def BX_CALL : ARMPseudoInst<(outs), (ins tGPR:$func), + 8, IIC_Br, [(ARMcall_nolink tGPR:$func)]>, + Requires<[IsARM, HasV4T]>, Sched<[WriteBr]>; + + // ARMv4 + def BMOVPCRX_CALL : ARMPseudoInst<(outs), (ins tGPR:$func), + 8, IIC_Br, [(ARMcall_nolink tGPR:$func)]>, + Requires<[IsARM, NoV4T]>, Sched<[WriteBr]>; + + // mov lr, pc; b if callee is marked noreturn to avoid confusing the + // return stack predictor. + def BMOVPCB_CALL : ARMPseudoInst<(outs), (ins arm_bl_target:$func), + 8, IIC_Br, [(ARMcall_nolink tglobaladdr:$func)]>, + Requires<[IsARM]>, Sched<[WriteBr]>; +} + +let isBranch = 1, isTerminator = 1 in { + // FIXME: should be able to write a pattern for ARMBrcond, but can't use + // a two-value operand where a dag node expects two operands. :( + def Bcc : ABI<0b1010, (outs), (ins arm_br_target:$target), + IIC_Br, "b", "\t$target", + [/*(ARMbrcond bb:$target, imm:$cc, CCR:$ccr)*/]>, + Sched<[WriteBr]> { + bits<24> target; + let Inst{23-0} = target; + let DecoderMethod = "DecodeBranchImmInstruction"; + } + + let isBarrier = 1 in { + // B is "predicable" since it's just a Bcc with an 'always' condition. + let isPredicable = 1 in + // FIXME: We shouldn't need this pseudo at all. Just using Bcc directly + // should be sufficient. + // FIXME: Is B really a Barrier? That doesn't seem right. + def B : ARMPseudoExpand<(outs), (ins arm_br_target:$target), 4, IIC_Br, + [(br bb:$target)], (Bcc arm_br_target:$target, + (ops 14, zero_reg))>, + Sched<[WriteBr]>; + + let Size = 4, isNotDuplicable = 1, isIndirectBranch = 1 in { + def BR_JTr : ARMPseudoInst<(outs), + (ins GPR:$target, i32imm:$jt), + 0, IIC_Br, + [(ARMbrjt GPR:$target, tjumptable:$jt)]>, + Sched<[WriteBr]>; + def BR_JTm_i12 : ARMPseudoInst<(outs), + (ins addrmode_imm12:$target, i32imm:$jt), + 0, IIC_Br, + [(ARMbrjt (i32 (load addrmode_imm12:$target)), + tjumptable:$jt)]>, Sched<[WriteBrTbl]>; + def BR_JTm_rs : ARMPseudoInst<(outs), + (ins ldst_so_reg:$target, i32imm:$jt), + 0, IIC_Br, + [(ARMbrjt (i32 (load ldst_so_reg:$target)), + tjumptable:$jt)]>, Sched<[WriteBrTbl]>; + def BR_JTadd : ARMPseudoInst<(outs), + (ins GPR:$target, GPR:$idx, i32imm:$jt), + 0, IIC_Br, + [(ARMbrjt (add GPR:$target, GPR:$idx), tjumptable:$jt)]>, + Sched<[WriteBrTbl]>; + } // isNotDuplicable = 1, isIndirectBranch = 1 + } // isBarrier = 1 + +} + +// BLX (immediate) +def BLXi : AXI<(outs), (ins arm_blx_target:$target), BrMiscFrm, NoItinerary, + "blx\t$target", []>, + Requires<[IsARM, HasV5T]>, Sched<[WriteBrL]> { + let Inst{31-25} = 0b1111101; + bits<25> target; + let Inst{23-0} = target{24-1}; + let Inst{24} = target{0}; + let isCall = 1; +} + +// Branch and Exchange Jazelle +def BXJ : ABI<0b0001, (outs), (ins GPR:$func), NoItinerary, "bxj", "\t$func", + [/* pattern left blank */]>, Sched<[WriteBr]> { + bits<4> func; + let Inst{23-20} = 0b0010; + let Inst{19-8} = 0xfff; + let Inst{7-4} = 0b0010; + let Inst{3-0} = func; + let isBranch = 1; +} + +// Tail calls. + +let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, Uses = [SP] in { + def TCRETURNdi : PseudoInst<(outs), (ins i32imm:$dst), IIC_Br, []>, + Sched<[WriteBr]>; + + def TCRETURNri : PseudoInst<(outs), (ins tcGPR:$dst), IIC_Br, []>, + Sched<[WriteBr]>; + + def TAILJMPd : ARMPseudoExpand<(outs), (ins arm_br_target:$dst), + 4, IIC_Br, [], + (Bcc arm_br_target:$dst, (ops 14, zero_reg))>, + Requires<[IsARM]>, Sched<[WriteBr]>; + + def TAILJMPr : ARMPseudoExpand<(outs), (ins tcGPR:$dst), + 4, IIC_Br, [], + (BX GPR:$dst)>, Sched<[WriteBr]>, + Requires<[IsARM, HasV4T]>; +} + +// Secure Monitor Call is a system instruction. +def SMC : ABI<0b0001, (outs), (ins imm0_15:$opt), NoItinerary, "smc", "\t$opt", + []>, Requires<[IsARM, HasTrustZone]> { + bits<4> opt; + let Inst{23-4} = 0b01100000000000000111; + let Inst{3-0} = opt; +} +def : MnemonicAlias<"smi", "smc">; + +// Supervisor Call (Software Interrupt) +let isCall = 1, Uses = [SP] in { +def SVC : ABI<0b1111, (outs), (ins imm24b:$svc), IIC_Br, "svc", "\t$svc", []>, + Sched<[WriteBr]> { + bits<24> svc; + let Inst{23-0} = svc; +} +} + +// Store Return State +class SRSI<bit wb, string asm> + : XI<(outs), (ins imm0_31:$mode), AddrModeNone, 4, IndexModeNone, BrFrm, + NoItinerary, asm, "", []> { + bits<5> mode; + let Inst{31-28} = 0b1111; + let Inst{27-25} = 0b100; + let Inst{22} = 1; + let Inst{21} = wb; + let Inst{20} = 0; + let Inst{19-16} = 0b1101; // SP + let Inst{15-5} = 0b00000101000; + let Inst{4-0} = mode; +} + +def SRSDA : SRSI<0, "srsda\tsp, $mode"> { + let Inst{24-23} = 0; +} +def SRSDA_UPD : SRSI<1, "srsda\tsp!, $mode"> { + let Inst{24-23} = 0; +} +def SRSDB : SRSI<0, "srsdb\tsp, $mode"> { + let Inst{24-23} = 0b10; +} +def SRSDB_UPD : SRSI<1, "srsdb\tsp!, $mode"> { + let Inst{24-23} = 0b10; +} +def SRSIA : SRSI<0, "srsia\tsp, $mode"> { + let Inst{24-23} = 0b01; +} +def SRSIA_UPD : SRSI<1, "srsia\tsp!, $mode"> { + let Inst{24-23} = 0b01; +} +def SRSIB : SRSI<0, "srsib\tsp, $mode"> { + let Inst{24-23} = 0b11; +} +def SRSIB_UPD : SRSI<1, "srsib\tsp!, $mode"> { + let Inst{24-23} = 0b11; +} + +def : ARMInstAlias<"srsda $mode", (SRSDA imm0_31:$mode)>; +def : ARMInstAlias<"srsda $mode!", (SRSDA_UPD imm0_31:$mode)>; + +def : ARMInstAlias<"srsdb $mode", (SRSDB imm0_31:$mode)>; +def : ARMInstAlias<"srsdb $mode!", (SRSDB_UPD imm0_31:$mode)>; + +def : ARMInstAlias<"srsia $mode", (SRSIA imm0_31:$mode)>; +def : ARMInstAlias<"srsia $mode!", (SRSIA_UPD imm0_31:$mode)>; + +def : ARMInstAlias<"srsib $mode", (SRSIB imm0_31:$mode)>; +def : ARMInstAlias<"srsib $mode!", (SRSIB_UPD imm0_31:$mode)>; + +// Return From Exception +class RFEI<bit wb, string asm> + : XI<(outs), (ins GPR:$Rn), AddrModeNone, 4, IndexModeNone, BrFrm, + NoItinerary, asm, "", []> { + bits<4> Rn; + let Inst{31-28} = 0b1111; + let Inst{27-25} = 0b100; + let Inst{22} = 0; + let Inst{21} = wb; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-0} = 0xa00; +} + +def RFEDA : RFEI<0, "rfeda\t$Rn"> { + let Inst{24-23} = 0; +} +def RFEDA_UPD : RFEI<1, "rfeda\t$Rn!"> { + let Inst{24-23} = 0; +} +def RFEDB : RFEI<0, "rfedb\t$Rn"> { + let Inst{24-23} = 0b10; +} +def RFEDB_UPD : RFEI<1, "rfedb\t$Rn!"> { + let Inst{24-23} = 0b10; +} +def RFEIA : RFEI<0, "rfeia\t$Rn"> { + let Inst{24-23} = 0b01; +} +def RFEIA_UPD : RFEI<1, "rfeia\t$Rn!"> { + let Inst{24-23} = 0b01; +} +def RFEIB : RFEI<0, "rfeib\t$Rn"> { + let Inst{24-23} = 0b11; +} +def RFEIB_UPD : RFEI<1, "rfeib\t$Rn!"> { + let Inst{24-23} = 0b11; +} + +// Hypervisor Call is a system instruction +let isCall = 1 in { +def HVC : AInoP< (outs), (ins imm0_65535:$imm), BrFrm, NoItinerary, + "hvc", "\t$imm", []>, + Requires<[IsARM, HasVirtualization]> { + bits<16> imm; + + // Even though HVC isn't predicable, it's encoding includes a condition field. + // The instruction is undefined if the condition field is 0xf otherwise it is + // unpredictable if it isn't condition AL (0xe). + let Inst{31-28} = 0b1110; + let Unpredictable{31-28} = 0b1111; + let Inst{27-24} = 0b0001; + let Inst{23-20} = 0b0100; + let Inst{19-8} = imm{15-4}; + let Inst{7-4} = 0b0111; + let Inst{3-0} = imm{3-0}; +} +} + +// Return from exception in Hypervisor mode. +let isReturn = 1, isBarrier = 1, isTerminator = 1, Defs = [PC] in +def ERET : ABI<0b0001, (outs), (ins), NoItinerary, "eret", "", []>, + Requires<[IsARM, HasVirtualization]> { + let Inst{23-0} = 0b011000000000000001101110; +} + +//===----------------------------------------------------------------------===// +// Load / Store Instructions. +// + +// Load + + +defm LDR : AI_ldr1<0, "ldr", IIC_iLoad_r, IIC_iLoad_si, load>; +defm LDRB : AI_ldr1nopc<1, "ldrb", IIC_iLoad_bh_r, IIC_iLoad_bh_si, + zextloadi8>; +defm STR : AI_str1<0, "str", IIC_iStore_r, IIC_iStore_si, store>; +defm STRB : AI_str1nopc<1, "strb", IIC_iStore_bh_r, IIC_iStore_bh_si, + truncstorei8>; + +// Special LDR for loads from non-pc-relative constpools. +let canFoldAsLoad = 1, mayLoad = 1, hasSideEffects = 0, + isReMaterializable = 1, isCodeGenOnly = 1 in +def LDRcp : AI2ldst<0b010, 1, 0, (outs GPR:$Rt), (ins addrmode_imm12:$addr), + AddrMode_i12, LdFrm, IIC_iLoad_r, "ldr", "\t$Rt, $addr", + []> { + bits<4> Rt; + bits<17> addr; + let Inst{23} = addr{12}; // U (add = ('U' == 1)) + let Inst{19-16} = 0b1111; + let Inst{15-12} = Rt; + let Inst{11-0} = addr{11-0}; // imm12 +} + +// Loads with zero extension +def LDRH : AI3ld<0b1011, 1, (outs GPR:$Rt), (ins addrmode3:$addr), LdMiscFrm, + IIC_iLoad_bh_r, "ldrh", "\t$Rt, $addr", + [(set GPR:$Rt, (zextloadi16 addrmode3:$addr))]>; + +// Loads with sign extension +def LDRSH : AI3ld<0b1111, 1, (outs GPR:$Rt), (ins addrmode3:$addr), LdMiscFrm, + IIC_iLoad_bh_r, "ldrsh", "\t$Rt, $addr", + [(set GPR:$Rt, (sextloadi16 addrmode3:$addr))]>; + +def LDRSB : AI3ld<0b1101, 1, (outs GPR:$Rt), (ins addrmode3:$addr), LdMiscFrm, + IIC_iLoad_bh_r, "ldrsb", "\t$Rt, $addr", + [(set GPR:$Rt, (sextloadi8 addrmode3:$addr))]>; + +let mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 in { + // Load doubleword + def LDRD : AI3ld<0b1101, 0, (outs GPR:$Rt, GPR:$Rt2), (ins addrmode3:$addr), + LdMiscFrm, IIC_iLoad_d_r, "ldrd", "\t$Rt, $Rt2, $addr", []>, + Requires<[IsARM, HasV5TE]>; +} + +def LDA : AIldracq<0b00, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "lda", "\t$Rt, $addr", []>; +def LDAB : AIldracq<0b10, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "ldab", "\t$Rt, $addr", []>; +def LDAH : AIldracq<0b11, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "ldah", "\t$Rt, $addr", []>; + +// Indexed loads +multiclass AI2_ldridx<bit isByte, string opc, + InstrItinClass iii, InstrItinClass iir> { + def _PRE_IMM : AI2ldstidx<1, isByte, 1, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addrmode_imm12_pre:$addr), IndexModePre, LdFrm, iii, + opc, "\t$Rt, $addr!", "$addr.base = $Rn_wb", []> { + bits<17> addr; + let Inst{25} = 0; + let Inst{23} = addr{12}; + let Inst{19-16} = addr{16-13}; + let Inst{11-0} = addr{11-0}; + let DecoderMethod = "DecodeLDRPreImm"; + } + + def _PRE_REG : AI2ldstidx<1, isByte, 1, (outs GPR:$Rt, GPR:$Rn_wb), + (ins ldst_so_reg:$addr), IndexModePre, LdFrm, iir, + opc, "\t$Rt, $addr!", "$addr.base = $Rn_wb", []> { + bits<17> addr; + let Inst{25} = 1; + let Inst{23} = addr{12}; + let Inst{19-16} = addr{16-13}; + let Inst{11-0} = addr{11-0}; + let Inst{4} = 0; + let DecoderMethod = "DecodeLDRPreReg"; + } + + def _POST_REG : AI2ldstidx<1, isByte, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$addr, am2offset_reg:$offset), + IndexModePost, LdFrm, iir, + opc, "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 1; + let Inst{23} = offset{12}; + let Inst{19-16} = addr; + let Inst{11-0} = offset{11-0}; + let Inst{4} = 0; + + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; + } + + def _POST_IMM : AI2ldstidx<1, isByte, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$addr, am2offset_imm:$offset), + IndexModePost, LdFrm, iii, + opc, "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 0; + let Inst{23} = offset{12}; + let Inst{19-16} = addr; + let Inst{11-0} = offset{11-0}; + + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; + } + +} + +let mayLoad = 1, hasSideEffects = 0 in { +// FIXME: for LDR_PRE_REG etc. the itineray should be either IIC_iLoad_ru or +// IIC_iLoad_siu depending on whether it the offset register is shifted. +defm LDR : AI2_ldridx<0, "ldr", IIC_iLoad_iu, IIC_iLoad_ru>; +defm LDRB : AI2_ldridx<1, "ldrb", IIC_iLoad_bh_iu, IIC_iLoad_bh_ru>; +} + +multiclass AI3_ldridx<bits<4> op, string opc, InstrItinClass itin> { + def _PRE : AI3ldstidx<op, 1, 1, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addrmode3_pre:$addr), IndexModePre, + LdMiscFrm, itin, + opc, "\t$Rt, $addr!", "$addr.base = $Rn_wb", []> { + bits<14> addr; + let Inst{23} = addr{8}; // U bit + let Inst{22} = addr{13}; // 1 == imm8, 0 == Rm + let Inst{19-16} = addr{12-9}; // Rn + let Inst{11-8} = addr{7-4}; // imm7_4/zero + let Inst{3-0} = addr{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; + } + def _POST : AI3ldstidx<op, 1, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$addr, am3offset:$offset), + IndexModePost, LdMiscFrm, itin, + opc, "\t$Rt, $addr, $offset", "$addr.base = $Rn_wb", + []> { + bits<10> offset; + bits<4> addr; + let Inst{23} = offset{8}; // U bit + let Inst{22} = offset{9}; // 1 == imm8, 0 == Rm + let Inst{19-16} = addr; + let Inst{11-8} = offset{7-4}; // imm7_4/zero + let Inst{3-0} = offset{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; + } +} + +let mayLoad = 1, hasSideEffects = 0 in { +defm LDRH : AI3_ldridx<0b1011, "ldrh", IIC_iLoad_bh_ru>; +defm LDRSH : AI3_ldridx<0b1111, "ldrsh", IIC_iLoad_bh_ru>; +defm LDRSB : AI3_ldridx<0b1101, "ldrsb", IIC_iLoad_bh_ru>; +let hasExtraDefRegAllocReq = 1 in { +def LDRD_PRE : AI3ldstidx<0b1101, 0, 1, (outs GPR:$Rt, GPR:$Rt2, GPR:$Rn_wb), + (ins addrmode3_pre:$addr), IndexModePre, + LdMiscFrm, IIC_iLoad_d_ru, + "ldrd", "\t$Rt, $Rt2, $addr!", + "$addr.base = $Rn_wb", []> { + bits<14> addr; + let Inst{23} = addr{8}; // U bit + let Inst{22} = addr{13}; // 1 == imm8, 0 == Rm + let Inst{19-16} = addr{12-9}; // Rn + let Inst{11-8} = addr{7-4}; // imm7_4/zero + let Inst{3-0} = addr{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; +} +def LDRD_POST: AI3ldstidx<0b1101, 0, 0, (outs GPR:$Rt, GPR:$Rt2, GPR:$Rn_wb), + (ins addr_offset_none:$addr, am3offset:$offset), + IndexModePost, LdMiscFrm, IIC_iLoad_d_ru, + "ldrd", "\t$Rt, $Rt2, $addr, $offset", + "$addr.base = $Rn_wb", []> { + bits<10> offset; + bits<4> addr; + let Inst{23} = offset{8}; // U bit + let Inst{22} = offset{9}; // 1 == imm8, 0 == Rm + let Inst{19-16} = addr; + let Inst{11-8} = offset{7-4}; // imm7_4/zero + let Inst{3-0} = offset{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; +} +} // hasExtraDefRegAllocReq = 1 +} // mayLoad = 1, hasSideEffects = 0 + +// LDRT, LDRBT, LDRSBT, LDRHT, LDRSHT. +let mayLoad = 1, hasSideEffects = 0 in { +def LDRT_POST_REG : AI2ldstidx<1, 0, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$addr, am2offset_reg:$offset), + IndexModePost, LdFrm, IIC_iLoad_ru, + "ldrt", "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 1; + let Inst{23} = offset{12}; + let Inst{21} = 1; // overwrite + let Inst{19-16} = addr; + let Inst{11-5} = offset{11-5}; + let Inst{4} = 0; + let Inst{3-0} = offset{3-0}; + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; +} + +def LDRT_POST_IMM + : AI2ldstidx<1, 0, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$addr, am2offset_imm:$offset), + IndexModePost, LdFrm, IIC_iLoad_ru, + "ldrt", "\t$Rt, $addr, $offset", "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 0; + let Inst{23} = offset{12}; + let Inst{21} = 1; // overwrite + let Inst{19-16} = addr; + let Inst{11-0} = offset{11-0}; + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; +} + +def LDRBT_POST_REG : AI2ldstidx<1, 1, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$addr, am2offset_reg:$offset), + IndexModePost, LdFrm, IIC_iLoad_bh_ru, + "ldrbt", "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 1; + let Inst{23} = offset{12}; + let Inst{21} = 1; // overwrite + let Inst{19-16} = addr; + let Inst{11-5} = offset{11-5}; + let Inst{4} = 0; + let Inst{3-0} = offset{3-0}; + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; +} + +def LDRBT_POST_IMM + : AI2ldstidx<1, 1, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$addr, am2offset_imm:$offset), + IndexModePost, LdFrm, IIC_iLoad_bh_ru, + "ldrbt", "\t$Rt, $addr, $offset", "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 0; + let Inst{23} = offset{12}; + let Inst{21} = 1; // overwrite + let Inst{19-16} = addr; + let Inst{11-0} = offset{11-0}; + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; +} + +multiclass AI3ldrT<bits<4> op, string opc> { + def i : AI3ldstidxT<op, 1, (outs GPR:$Rt, GPR:$base_wb), + (ins addr_offset_none:$addr, postidx_imm8:$offset), + IndexModePost, LdMiscFrm, IIC_iLoad_bh_ru, opc, + "\t$Rt, $addr, $offset", "$addr.base = $base_wb", []> { + bits<9> offset; + let Inst{23} = offset{8}; + let Inst{22} = 1; + let Inst{11-8} = offset{7-4}; + let Inst{3-0} = offset{3-0}; + } + def r : AI3ldstidxT<op, 1, (outs GPRnopc:$Rt, GPRnopc:$base_wb), + (ins addr_offset_none:$addr, postidx_reg:$Rm), + IndexModePost, LdMiscFrm, IIC_iLoad_bh_ru, opc, + "\t$Rt, $addr, $Rm", "$addr.base = $base_wb", []> { + bits<5> Rm; + let Inst{23} = Rm{4}; + let Inst{22} = 0; + let Inst{11-8} = 0; + let Unpredictable{11-8} = 0b1111; + let Inst{3-0} = Rm{3-0}; + let DecoderMethod = "DecodeLDR"; + } +} + +defm LDRSBT : AI3ldrT<0b1101, "ldrsbt">; +defm LDRHT : AI3ldrT<0b1011, "ldrht">; +defm LDRSHT : AI3ldrT<0b1111, "ldrsht">; +} + +def LDRT_POST + : ARMAsmPseudo<"ldrt${q} $Rt, $addr", (ins addr_offset_none:$addr, pred:$q), + (outs GPR:$Rt)>; + +def LDRBT_POST + : ARMAsmPseudo<"ldrbt${q} $Rt, $addr", (ins addr_offset_none:$addr, pred:$q), + (outs GPR:$Rt)>; + +// Pseudo instruction ldr Rt, =immediate +def LDRConstPool + : ARMAsmPseudo<"ldr${q} $Rt, $immediate", + (ins const_pool_asm_imm:$immediate, pred:$q), + (outs GPR:$Rt)>; + +// Store + +// Stores with truncate +def STRH : AI3str<0b1011, (outs), (ins GPR:$Rt, addrmode3:$addr), StMiscFrm, + IIC_iStore_bh_r, "strh", "\t$Rt, $addr", + [(truncstorei16 GPR:$Rt, addrmode3:$addr)]>; + +// Store doubleword +let mayStore = 1, hasSideEffects = 0, hasExtraSrcRegAllocReq = 1 in { + def STRD : AI3str<0b1111, (outs), (ins GPR:$Rt, GPR:$Rt2, addrmode3:$addr), + StMiscFrm, IIC_iStore_d_r, "strd", "\t$Rt, $Rt2, $addr", []>, + Requires<[IsARM, HasV5TE]> { + let Inst{21} = 0; + } +} + +// Indexed stores +multiclass AI2_stridx<bit isByte, string opc, + InstrItinClass iii, InstrItinClass iir> { + def _PRE_IMM : AI2ldstidx<0, isByte, 1, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addrmode_imm12_pre:$addr), IndexModePre, + StFrm, iii, + opc, "\t$Rt, $addr!", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", []> { + bits<17> addr; + let Inst{25} = 0; + let Inst{23} = addr{12}; // U (add = ('U' == 1)) + let Inst{19-16} = addr{16-13}; // Rn + let Inst{11-0} = addr{11-0}; // imm12 + let DecoderMethod = "DecodeSTRPreImm"; + } + + def _PRE_REG : AI2ldstidx<0, isByte, 1, (outs GPR:$Rn_wb), + (ins GPR:$Rt, ldst_so_reg:$addr), + IndexModePre, StFrm, iir, + opc, "\t$Rt, $addr!", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", []> { + bits<17> addr; + let Inst{25} = 1; + let Inst{23} = addr{12}; // U (add = ('U' == 1)) + let Inst{19-16} = addr{16-13}; // Rn + let Inst{11-0} = addr{11-0}; + let Inst{4} = 0; // Inst{4} = 0 + let DecoderMethod = "DecodeSTRPreReg"; + } + def _POST_REG : AI2ldstidx<0, isByte, 0, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addr_offset_none:$addr, am2offset_reg:$offset), + IndexModePost, StFrm, iir, + opc, "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 1; + let Inst{23} = offset{12}; + let Inst{19-16} = addr; + let Inst{11-0} = offset{11-0}; + let Inst{4} = 0; + + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; + } + + def _POST_IMM : AI2ldstidx<0, isByte, 0, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addr_offset_none:$addr, am2offset_imm:$offset), + IndexModePost, StFrm, iii, + opc, "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 0; + let Inst{23} = offset{12}; + let Inst{19-16} = addr; + let Inst{11-0} = offset{11-0}; + + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; + } +} + +let mayStore = 1, hasSideEffects = 0 in { +// FIXME: for STR_PRE_REG etc. the itineray should be either IIC_iStore_ru or +// IIC_iStore_siu depending on whether it the offset register is shifted. +defm STR : AI2_stridx<0, "str", IIC_iStore_iu, IIC_iStore_ru>; +defm STRB : AI2_stridx<1, "strb", IIC_iStore_bh_iu, IIC_iStore_bh_ru>; +} + +def : ARMPat<(post_store GPR:$Rt, addr_offset_none:$addr, + am2offset_reg:$offset), + (STR_POST_REG GPR:$Rt, addr_offset_none:$addr, + am2offset_reg:$offset)>; +def : ARMPat<(post_store GPR:$Rt, addr_offset_none:$addr, + am2offset_imm:$offset), + (STR_POST_IMM GPR:$Rt, addr_offset_none:$addr, + am2offset_imm:$offset)>; +def : ARMPat<(post_truncsti8 GPR:$Rt, addr_offset_none:$addr, + am2offset_reg:$offset), + (STRB_POST_REG GPR:$Rt, addr_offset_none:$addr, + am2offset_reg:$offset)>; +def : ARMPat<(post_truncsti8 GPR:$Rt, addr_offset_none:$addr, + am2offset_imm:$offset), + (STRB_POST_IMM GPR:$Rt, addr_offset_none:$addr, + am2offset_imm:$offset)>; + +// Pseudo-instructions for pattern matching the pre-indexed stores. We can't +// put the patterns on the instruction definitions directly as ISel wants +// the address base and offset to be separate operands, not a single +// complex operand like we represent the instructions themselves. The +// pseudos map between the two. +let usesCustomInserter = 1, + Constraints = "$Rn = $Rn_wb,@earlyclobber $Rn_wb" in { +def STRi_preidx: ARMPseudoInst<(outs GPR:$Rn_wb), + (ins GPR:$Rt, GPR:$Rn, am2offset_imm:$offset, pred:$p), + 4, IIC_iStore_ru, + [(set GPR:$Rn_wb, + (pre_store GPR:$Rt, GPR:$Rn, am2offset_imm:$offset))]>; +def STRr_preidx: ARMPseudoInst<(outs GPR:$Rn_wb), + (ins GPR:$Rt, GPR:$Rn, am2offset_reg:$offset, pred:$p), + 4, IIC_iStore_ru, + [(set GPR:$Rn_wb, + (pre_store GPR:$Rt, GPR:$Rn, am2offset_reg:$offset))]>; +def STRBi_preidx: ARMPseudoInst<(outs GPR:$Rn_wb), + (ins GPR:$Rt, GPR:$Rn, am2offset_imm:$offset, pred:$p), + 4, IIC_iStore_ru, + [(set GPR:$Rn_wb, + (pre_truncsti8 GPR:$Rt, GPR:$Rn, am2offset_imm:$offset))]>; +def STRBr_preidx: ARMPseudoInst<(outs GPR:$Rn_wb), + (ins GPR:$Rt, GPR:$Rn, am2offset_reg:$offset, pred:$p), + 4, IIC_iStore_ru, + [(set GPR:$Rn_wb, + (pre_truncsti8 GPR:$Rt, GPR:$Rn, am2offset_reg:$offset))]>; +def STRH_preidx: ARMPseudoInst<(outs GPR:$Rn_wb), + (ins GPR:$Rt, GPR:$Rn, am3offset:$offset, pred:$p), + 4, IIC_iStore_ru, + [(set GPR:$Rn_wb, + (pre_truncsti16 GPR:$Rt, GPR:$Rn, am3offset:$offset))]>; +} + + + +def STRH_PRE : AI3ldstidx<0b1011, 0, 1, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addrmode3_pre:$addr), IndexModePre, + StMiscFrm, IIC_iStore_bh_ru, + "strh", "\t$Rt, $addr!", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", []> { + bits<14> addr; + let Inst{23} = addr{8}; // U bit + let Inst{22} = addr{13}; // 1 == imm8, 0 == Rm + let Inst{19-16} = addr{12-9}; // Rn + let Inst{11-8} = addr{7-4}; // imm7_4/zero + let Inst{3-0} = addr{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; +} + +def STRH_POST : AI3ldstidx<0b1011, 0, 0, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addr_offset_none:$addr, am3offset:$offset), + IndexModePost, StMiscFrm, IIC_iStore_bh_ru, + "strh", "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", + [(set GPR:$Rn_wb, (post_truncsti16 GPR:$Rt, + addr_offset_none:$addr, + am3offset:$offset))]> { + bits<10> offset; + bits<4> addr; + let Inst{23} = offset{8}; // U bit + let Inst{22} = offset{9}; // 1 == imm8, 0 == Rm + let Inst{19-16} = addr; + let Inst{11-8} = offset{7-4}; // imm7_4/zero + let Inst{3-0} = offset{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; +} + +let mayStore = 1, hasSideEffects = 0, hasExtraSrcRegAllocReq = 1 in { +def STRD_PRE : AI3ldstidx<0b1111, 0, 1, (outs GPR:$Rn_wb), + (ins GPR:$Rt, GPR:$Rt2, addrmode3_pre:$addr), + IndexModePre, StMiscFrm, IIC_iStore_d_ru, + "strd", "\t$Rt, $Rt2, $addr!", + "$addr.base = $Rn_wb", []> { + bits<14> addr; + let Inst{23} = addr{8}; // U bit + let Inst{22} = addr{13}; // 1 == imm8, 0 == Rm + let Inst{19-16} = addr{12-9}; // Rn + let Inst{11-8} = addr{7-4}; // imm7_4/zero + let Inst{3-0} = addr{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; +} + +def STRD_POST: AI3ldstidx<0b1111, 0, 0, (outs GPR:$Rn_wb), + (ins GPR:$Rt, GPR:$Rt2, addr_offset_none:$addr, + am3offset:$offset), + IndexModePost, StMiscFrm, IIC_iStore_d_ru, + "strd", "\t$Rt, $Rt2, $addr, $offset", + "$addr.base = $Rn_wb", []> { + bits<10> offset; + bits<4> addr; + let Inst{23} = offset{8}; // U bit + let Inst{22} = offset{9}; // 1 == imm8, 0 == Rm + let Inst{19-16} = addr; + let Inst{11-8} = offset{7-4}; // imm7_4/zero + let Inst{3-0} = offset{3-0}; // imm3_0/Rm + let DecoderMethod = "DecodeAddrMode3Instruction"; +} +} // mayStore = 1, hasSideEffects = 0, hasExtraSrcRegAllocReq = 1 + +// STRT, STRBT, and STRHT + +def STRBT_POST_REG : AI2ldstidx<0, 1, 0, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addr_offset_none:$addr, am2offset_reg:$offset), + IndexModePost, StFrm, IIC_iStore_bh_ru, + "strbt", "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 1; + let Inst{23} = offset{12}; + let Inst{21} = 1; // overwrite + let Inst{19-16} = addr; + let Inst{11-5} = offset{11-5}; + let Inst{4} = 0; + let Inst{3-0} = offset{3-0}; + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; +} + +def STRBT_POST_IMM + : AI2ldstidx<0, 1, 0, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addr_offset_none:$addr, am2offset_imm:$offset), + IndexModePost, StFrm, IIC_iStore_bh_ru, + "strbt", "\t$Rt, $addr, $offset", "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 0; + let Inst{23} = offset{12}; + let Inst{21} = 1; // overwrite + let Inst{19-16} = addr; + let Inst{11-0} = offset{11-0}; + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; +} + +def STRBT_POST + : ARMAsmPseudo<"strbt${q} $Rt, $addr", + (ins GPR:$Rt, addr_offset_none:$addr, pred:$q)>; + +let mayStore = 1, hasSideEffects = 0 in { +def STRT_POST_REG : AI2ldstidx<0, 0, 0, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addr_offset_none:$addr, am2offset_reg:$offset), + IndexModePost, StFrm, IIC_iStore_ru, + "strt", "\t$Rt, $addr, $offset", + "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 1; + let Inst{23} = offset{12}; + let Inst{21} = 1; // overwrite + let Inst{19-16} = addr; + let Inst{11-5} = offset{11-5}; + let Inst{4} = 0; + let Inst{3-0} = offset{3-0}; + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; +} + +def STRT_POST_IMM + : AI2ldstidx<0, 0, 0, (outs GPR:$Rn_wb), + (ins GPR:$Rt, addr_offset_none:$addr, am2offset_imm:$offset), + IndexModePost, StFrm, IIC_iStore_ru, + "strt", "\t$Rt, $addr, $offset", "$addr.base = $Rn_wb", []> { + // {12} isAdd + // {11-0} imm12/Rm + bits<14> offset; + bits<4> addr; + let Inst{25} = 0; + let Inst{23} = offset{12}; + let Inst{21} = 1; // overwrite + let Inst{19-16} = addr; + let Inst{11-0} = offset{11-0}; + let DecoderMethod = "DecodeAddrMode2IdxInstruction"; +} +} + +def STRT_POST + : ARMAsmPseudo<"strt${q} $Rt, $addr", + (ins GPR:$Rt, addr_offset_none:$addr, pred:$q)>; + +multiclass AI3strT<bits<4> op, string opc> { + def i : AI3ldstidxT<op, 0, (outs GPR:$base_wb), + (ins GPR:$Rt, addr_offset_none:$addr, postidx_imm8:$offset), + IndexModePost, StMiscFrm, IIC_iStore_bh_ru, opc, + "\t$Rt, $addr, $offset", "$addr.base = $base_wb", []> { + bits<9> offset; + let Inst{23} = offset{8}; + let Inst{22} = 1; + let Inst{11-8} = offset{7-4}; + let Inst{3-0} = offset{3-0}; + } + def r : AI3ldstidxT<op, 0, (outs GPR:$base_wb), + (ins GPR:$Rt, addr_offset_none:$addr, postidx_reg:$Rm), + IndexModePost, StMiscFrm, IIC_iStore_bh_ru, opc, + "\t$Rt, $addr, $Rm", "$addr.base = $base_wb", []> { + bits<5> Rm; + let Inst{23} = Rm{4}; + let Inst{22} = 0; + let Inst{11-8} = 0; + let Inst{3-0} = Rm{3-0}; + } +} + + +defm STRHT : AI3strT<0b1011, "strht">; + +def STL : AIstrrel<0b00, (outs), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "stl", "\t$Rt, $addr", []>; +def STLB : AIstrrel<0b10, (outs), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "stlb", "\t$Rt, $addr", []>; +def STLH : AIstrrel<0b11, (outs), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "stlh", "\t$Rt, $addr", []>; + +//===----------------------------------------------------------------------===// +// Load / store multiple Instructions. +// + +multiclass arm_ldst_mult<string asm, string sfx, bit L_bit, bit P_bit, Format f, + InstrItinClass itin, InstrItinClass itin_upd> { + // IA is the default, so no need for an explicit suffix on the + // mnemonic here. Without it is the canonical spelling. + def IA : + AXI4<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IndexModeNone, f, itin, + !strconcat(asm, "${p}\t$Rn, $regs", sfx), "", []> { + let Inst{24-23} = 0b01; // Increment After + let Inst{22} = P_bit; + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + } + def IA_UPD : + AXI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IndexModeUpd, f, itin_upd, + !strconcat(asm, "${p}\t$Rn!, $regs", sfx), "$Rn = $wb", []> { + let Inst{24-23} = 0b01; // Increment After + let Inst{22} = P_bit; + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + + let DecoderMethod = "DecodeMemMultipleWritebackInstruction"; + } + def DA : + AXI4<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IndexModeNone, f, itin, + !strconcat(asm, "da${p}\t$Rn, $regs", sfx), "", []> { + let Inst{24-23} = 0b00; // Decrement After + let Inst{22} = P_bit; + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + } + def DA_UPD : + AXI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IndexModeUpd, f, itin_upd, + !strconcat(asm, "da${p}\t$Rn!, $regs", sfx), "$Rn = $wb", []> { + let Inst{24-23} = 0b00; // Decrement After + let Inst{22} = P_bit; + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + + let DecoderMethod = "DecodeMemMultipleWritebackInstruction"; + } + def DB : + AXI4<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IndexModeNone, f, itin, + !strconcat(asm, "db${p}\t$Rn, $regs", sfx), "", []> { + let Inst{24-23} = 0b10; // Decrement Before + let Inst{22} = P_bit; + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + } + def DB_UPD : + AXI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IndexModeUpd, f, itin_upd, + !strconcat(asm, "db${p}\t$Rn!, $regs", sfx), "$Rn = $wb", []> { + let Inst{24-23} = 0b10; // Decrement Before + let Inst{22} = P_bit; + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + + let DecoderMethod = "DecodeMemMultipleWritebackInstruction"; + } + def IB : + AXI4<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IndexModeNone, f, itin, + !strconcat(asm, "ib${p}\t$Rn, $regs", sfx), "", []> { + let Inst{24-23} = 0b11; // Increment Before + let Inst{22} = P_bit; + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + } + def IB_UPD : + AXI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IndexModeUpd, f, itin_upd, + !strconcat(asm, "ib${p}\t$Rn!, $regs", sfx), "$Rn = $wb", []> { + let Inst{24-23} = 0b11; // Increment Before + let Inst{22} = P_bit; + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + + let DecoderMethod = "DecodeMemMultipleWritebackInstruction"; + } +} + +let hasSideEffects = 0 in { + +let mayLoad = 1, hasExtraDefRegAllocReq = 1 in +defm LDM : arm_ldst_mult<"ldm", "", 1, 0, LdStMulFrm, IIC_iLoad_m, + IIC_iLoad_mu>, ComplexDeprecationPredicate<"ARMLoad">; + +let mayStore = 1, hasExtraSrcRegAllocReq = 1 in +defm STM : arm_ldst_mult<"stm", "", 0, 0, LdStMulFrm, IIC_iStore_m, + IIC_iStore_mu>, + ComplexDeprecationPredicate<"ARMStore">; + +} // hasSideEffects + +// FIXME: remove when we have a way to marking a MI with these properties. +// FIXME: Should pc be an implicit operand like PICADD, etc? +let isReturn = 1, isTerminator = 1, isBarrier = 1, mayLoad = 1, + hasExtraDefRegAllocReq = 1, isCodeGenOnly = 1 in +def LDMIA_RET : ARMPseudoExpand<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, + reglist:$regs, variable_ops), + 4, IIC_iLoad_mBr, [], + (LDMIA_UPD GPR:$wb, GPR:$Rn, pred:$p, reglist:$regs)>, + RegConstraint<"$Rn = $wb">; + +let mayLoad = 1, hasExtraDefRegAllocReq = 1 in +defm sysLDM : arm_ldst_mult<"ldm", " ^", 1, 1, LdStMulFrm, IIC_iLoad_m, + IIC_iLoad_mu>; + +let mayStore = 1, hasExtraSrcRegAllocReq = 1 in +defm sysSTM : arm_ldst_mult<"stm", " ^", 0, 1, LdStMulFrm, IIC_iStore_m, + IIC_iStore_mu>; + + + +//===----------------------------------------------------------------------===// +// Move Instructions. +// + +let hasSideEffects = 0, isMoveReg = 1 in +def MOVr : AsI1<0b1101, (outs GPR:$Rd), (ins GPR:$Rm), DPFrm, IIC_iMOVr, + "mov", "\t$Rd, $Rm", []>, UnaryDP, Sched<[WriteALU]> { + bits<4> Rd; + bits<4> Rm; + + let Inst{19-16} = 0b0000; + let Inst{11-4} = 0b00000000; + let Inst{25} = 0; + let Inst{3-0} = Rm; + let Inst{15-12} = Rd; +} + +// A version for the smaller set of tail call registers. +let hasSideEffects = 0 in +def MOVr_TC : AsI1<0b1101, (outs tcGPR:$Rd), (ins tcGPR:$Rm), DPFrm, + IIC_iMOVr, "mov", "\t$Rd, $Rm", []>, UnaryDP, Sched<[WriteALU]> { + bits<4> Rd; + bits<4> Rm; + + let Inst{11-4} = 0b00000000; + let Inst{25} = 0; + let Inst{3-0} = Rm; + let Inst{15-12} = Rd; +} + +def MOVsr : AsI1<0b1101, (outs GPRnopc:$Rd), (ins shift_so_reg_reg:$src), + DPSoRegRegFrm, IIC_iMOVsr, + "mov", "\t$Rd, $src", + [(set GPRnopc:$Rd, shift_so_reg_reg:$src)]>, UnaryDP, + Sched<[WriteALU]> { + bits<4> Rd; + bits<12> src; + let Inst{15-12} = Rd; + let Inst{19-16} = 0b0000; + let Inst{11-8} = src{11-8}; + let Inst{7} = 0; + let Inst{6-5} = src{6-5}; + let Inst{4} = 1; + let Inst{3-0} = src{3-0}; + let Inst{25} = 0; +} + +def MOVsi : AsI1<0b1101, (outs GPR:$Rd), (ins shift_so_reg_imm:$src), + DPSoRegImmFrm, IIC_iMOVsr, + "mov", "\t$Rd, $src", [(set GPR:$Rd, shift_so_reg_imm:$src)]>, + UnaryDP, Sched<[WriteALU]> { + bits<4> Rd; + bits<12> src; + let Inst{15-12} = Rd; + let Inst{19-16} = 0b0000; + let Inst{11-5} = src{11-5}; + let Inst{4} = 0; + let Inst{3-0} = src{3-0}; + let Inst{25} = 0; +} + +let isReMaterializable = 1, isAsCheapAsAMove = 1, isMoveImm = 1 in +def MOVi : AsI1<0b1101, (outs GPR:$Rd), (ins mod_imm:$imm), DPFrm, IIC_iMOVi, + "mov", "\t$Rd, $imm", [(set GPR:$Rd, mod_imm:$imm)]>, UnaryDP, + Sched<[WriteALU]> { + bits<4> Rd; + bits<12> imm; + let Inst{25} = 1; + let Inst{15-12} = Rd; + let Inst{19-16} = 0b0000; + let Inst{11-0} = imm; +} + +let isReMaterializable = 1, isAsCheapAsAMove = 1, isMoveImm = 1 in +def MOVi16 : AI1<0b1000, (outs GPR:$Rd), (ins imm0_65535_expr:$imm), + DPFrm, IIC_iMOVi, + "movw", "\t$Rd, $imm", + [(set GPR:$Rd, imm0_65535:$imm)]>, + Requires<[IsARM, HasV6T2]>, UnaryDP, Sched<[WriteALU]> { + bits<4> Rd; + bits<16> imm; + let Inst{15-12} = Rd; + let Inst{11-0} = imm{11-0}; + let Inst{19-16} = imm{15-12}; + let Inst{20} = 0; + let Inst{25} = 1; + let DecoderMethod = "DecodeArmMOVTWInstruction"; +} + +def : InstAlias<"mov${p} $Rd, $imm", + (MOVi16 GPR:$Rd, imm0_65535_expr:$imm, pred:$p), 0>, + Requires<[IsARM, HasV6T2]>; + +def MOVi16_ga_pcrel : PseudoInst<(outs GPR:$Rd), + (ins i32imm:$addr, pclabel:$id), IIC_iMOVi, []>, + Sched<[WriteALU]>; + +let Constraints = "$src = $Rd" in { +def MOVTi16 : AI1<0b1010, (outs GPRnopc:$Rd), + (ins GPR:$src, imm0_65535_expr:$imm), + DPFrm, IIC_iMOVi, + "movt", "\t$Rd, $imm", + [(set GPRnopc:$Rd, + (or (and GPR:$src, 0xffff), + lo16AllZero:$imm))]>, UnaryDP, + Requires<[IsARM, HasV6T2]>, Sched<[WriteALU]> { + bits<4> Rd; + bits<16> imm; + let Inst{15-12} = Rd; + let Inst{11-0} = imm{11-0}; + let Inst{19-16} = imm{15-12}; + let Inst{20} = 0; + let Inst{25} = 1; + let DecoderMethod = "DecodeArmMOVTWInstruction"; +} + +def MOVTi16_ga_pcrel : PseudoInst<(outs GPR:$Rd), + (ins GPR:$src, i32imm:$addr, pclabel:$id), IIC_iMOVi, []>, + Sched<[WriteALU]>; + +} // Constraints + +def : ARMPat<(or GPR:$src, 0xffff0000), (MOVTi16 GPR:$src, 0xffff)>, + Requires<[IsARM, HasV6T2]>; + +let Uses = [CPSR] in +def RRX: PseudoInst<(outs GPR:$Rd), (ins GPR:$Rm), IIC_iMOVsi, + [(set GPR:$Rd, (ARMrrx GPR:$Rm))]>, UnaryDP, + Requires<[IsARM]>, Sched<[WriteALU]>; + +// These aren't really mov instructions, but we have to define them this way +// due to flag operands. + +let Defs = [CPSR] in { +def MOVsrl_flag : PseudoInst<(outs GPR:$dst), (ins GPR:$src), IIC_iMOVsi, + [(set GPR:$dst, (ARMsrl_flag GPR:$src))]>, UnaryDP, + Sched<[WriteALU]>, Requires<[IsARM]>; +def MOVsra_flag : PseudoInst<(outs GPR:$dst), (ins GPR:$src), IIC_iMOVsi, + [(set GPR:$dst, (ARMsra_flag GPR:$src))]>, UnaryDP, + Sched<[WriteALU]>, Requires<[IsARM]>; +} + +//===----------------------------------------------------------------------===// +// Extend Instructions. +// + +// Sign extenders + +def SXTB : AI_ext_rrot<0b01101010, + "sxtb", UnOpFrag<(sext_inreg node:$Src, i8)>>; +def SXTH : AI_ext_rrot<0b01101011, + "sxth", UnOpFrag<(sext_inreg node:$Src, i16)>>; + +def SXTAB : AI_exta_rrot<0b01101010, + "sxtab", BinOpFrag<(add node:$LHS, (sext_inreg node:$RHS, i8))>>; +def SXTAH : AI_exta_rrot<0b01101011, + "sxtah", BinOpFrag<(add node:$LHS, (sext_inreg node:$RHS,i16))>>; + +def : ARMV6Pat<(add rGPR:$Rn, (sext_inreg (srl rGPR:$Rm, rot_imm:$rot), i8)), + (SXTAB rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : ARMV6Pat<(add rGPR:$Rn, (sext_inreg (srl rGPR:$Rm, imm8_or_16:$rot), + i16)), + (SXTAH rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; + +def SXTB16 : AI_ext_rrot_np<0b01101000, "sxtb16">; +def : ARMV6Pat<(int_arm_sxtb16 GPR:$Src), + (SXTB16 GPR:$Src, 0)>; + +def SXTAB16 : AI_exta_rrot_np<0b01101000, "sxtab16">; +def : ARMV6Pat<(int_arm_sxtab16 GPR:$LHS, GPR:$RHS), + (SXTAB16 GPR:$LHS, GPR:$RHS, 0)>; + +// Zero extenders + +let AddedComplexity = 16 in { +def UXTB : AI_ext_rrot<0b01101110, + "uxtb" , UnOpFrag<(and node:$Src, 0x000000FF)>>; +def UXTH : AI_ext_rrot<0b01101111, + "uxth" , UnOpFrag<(and node:$Src, 0x0000FFFF)>>; +def UXTB16 : AI_ext_rrot<0b01101100, + "uxtb16", UnOpFrag<(and node:$Src, 0x00FF00FF)>>; + +// FIXME: This pattern incorrectly assumes the shl operator is a rotate. +// The transformation should probably be done as a combiner action +// instead so we can include a check for masking back in the upper +// eight bits of the source into the lower eight bits of the result. +//def : ARMV6Pat<(and (shl GPR:$Src, (i32 8)), 0xFF00FF), +// (UXTB16r_rot GPR:$Src, 3)>; +def : ARMV6Pat<(and (srl GPR:$Src, (i32 8)), 0xFF00FF), + (UXTB16 GPR:$Src, 1)>; +def : ARMV6Pat<(int_arm_uxtb16 GPR:$Src), + (UXTB16 GPR:$Src, 0)>; + +def UXTAB : AI_exta_rrot<0b01101110, "uxtab", + BinOpFrag<(add node:$LHS, (and node:$RHS, 0x00FF))>>; +def UXTAH : AI_exta_rrot<0b01101111, "uxtah", + BinOpFrag<(add node:$LHS, (and node:$RHS, 0xFFFF))>>; + +def : ARMV6Pat<(add rGPR:$Rn, (and (srl rGPR:$Rm, rot_imm:$rot), 0xFF)), + (UXTAB rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : ARMV6Pat<(add rGPR:$Rn, (and (srl rGPR:$Rm, imm8_or_16:$rot), 0xFFFF)), + (UXTAH rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +} + +// This isn't safe in general, the add is two 16-bit units, not a 32-bit add. +def UXTAB16 : AI_exta_rrot_np<0b01101100, "uxtab16">; +def : ARMV6Pat<(int_arm_uxtab16 GPR:$LHS, GPR:$RHS), + (UXTAB16 GPR:$LHS, GPR:$RHS, 0)>; + + +def SBFX : I<(outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, imm0_31:$lsb, imm1_32:$width), + AddrMode1, 4, IndexModeNone, DPFrm, IIC_iUNAsi, + "sbfx", "\t$Rd, $Rn, $lsb, $width", "", []>, + Requires<[IsARM, HasV6T2]> { + bits<4> Rd; + bits<4> Rn; + bits<5> lsb; + bits<5> width; + let Inst{27-21} = 0b0111101; + let Inst{6-4} = 0b101; + let Inst{20-16} = width; + let Inst{15-12} = Rd; + let Inst{11-7} = lsb; + let Inst{3-0} = Rn; +} + +def UBFX : I<(outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, imm0_31:$lsb, imm1_32:$width), + AddrMode1, 4, IndexModeNone, DPFrm, IIC_iUNAsi, + "ubfx", "\t$Rd, $Rn, $lsb, $width", "", []>, + Requires<[IsARM, HasV6T2]> { + bits<4> Rd; + bits<4> Rn; + bits<5> lsb; + bits<5> width; + let Inst{27-21} = 0b0111111; + let Inst{6-4} = 0b101; + let Inst{20-16} = width; + let Inst{15-12} = Rd; + let Inst{11-7} = lsb; + let Inst{3-0} = Rn; +} + +//===----------------------------------------------------------------------===// +// Arithmetic Instructions. +// + +let isAdd = 1 in +defm ADD : AsI1_bin_irs<0b0100, "add", + IIC_iALUi, IIC_iALUr, IIC_iALUsr, add, 1>; +defm SUB : AsI1_bin_irs<0b0010, "sub", + IIC_iALUi, IIC_iALUr, IIC_iALUsr, sub>; + +// ADD and SUB with 's' bit set. +// +// Currently, ADDS/SUBS are pseudo opcodes that exist only in the +// selection DAG. They are "lowered" to real ADD/SUB opcodes by +// AdjustInstrPostInstrSelection where we determine whether or not to +// set the "s" bit based on CPSR liveness. +// +// FIXME: Eliminate ADDS/SUBS pseudo opcodes after adding tablegen +// support for an optional CPSR definition that corresponds to the DAG +// node's second value. We can then eliminate the implicit def of CPSR. +let isAdd = 1 in +defm ADDS : AsI1_bin_s_irs<IIC_iALUi, IIC_iALUr, IIC_iALUsr, ARMaddc, 1>; +defm SUBS : AsI1_bin_s_irs<IIC_iALUi, IIC_iALUr, IIC_iALUsr, ARMsubc>; + +let isAdd = 1 in +defm ADC : AI1_adde_sube_irs<0b0101, "adc", ARMadde, 1>; +defm SBC : AI1_adde_sube_irs<0b0110, "sbc", ARMsube>; + +defm RSB : AsI1_rbin_irs<0b0011, "rsb", + IIC_iALUi, IIC_iALUr, IIC_iALUsr, + sub>; + +// FIXME: Eliminate them if we can write def : Pat patterns which defines +// CPSR and the implicit def of CPSR is not needed. +defm RSBS : AsI1_rbin_s_is<IIC_iALUi, IIC_iALUr, IIC_iALUsr, ARMsubc>; + +defm RSC : AI1_rsc_irs<0b0111, "rsc", ARMsube>; + +// (sub X, imm) gets canonicalized to (add X, -imm). Match this form. +// The assume-no-carry-in form uses the negation of the input since add/sub +// assume opposite meanings of the carry flag (i.e., carry == !borrow). +// See the definition of AddWithCarry() in the ARM ARM A2.2.1 for the gory +// details. +def : ARMPat<(add GPR:$src, mod_imm_neg:$imm), + (SUBri GPR:$src, mod_imm_neg:$imm)>; +def : ARMPat<(ARMaddc GPR:$src, mod_imm_neg:$imm), + (SUBSri GPR:$src, mod_imm_neg:$imm)>; + +def : ARMPat<(add GPR:$src, imm0_65535_neg:$imm), + (SUBrr GPR:$src, (MOVi16 (imm_neg_XFORM imm:$imm)))>, + Requires<[IsARM, HasV6T2]>; +def : ARMPat<(ARMaddc GPR:$src, imm0_65535_neg:$imm), + (SUBSrr GPR:$src, (MOVi16 (imm_neg_XFORM imm:$imm)))>, + Requires<[IsARM, HasV6T2]>; + +// The with-carry-in form matches bitwise not instead of the negation. +// Effectively, the inverse interpretation of the carry flag already accounts +// for part of the negation. +def : ARMPat<(ARMadde GPR:$src, mod_imm_not:$imm, CPSR), + (SBCri GPR:$src, mod_imm_not:$imm)>; +def : ARMPat<(ARMadde GPR:$src, imm0_65535_neg:$imm, CPSR), + (SBCrr GPR:$src, (MOVi16 (imm_not_XFORM imm:$imm)))>, + Requires<[IsARM, HasV6T2]>; + +// Note: These are implemented in C++ code, because they have to generate +// ADD/SUBrs instructions, which use a complex pattern that a xform function +// cannot produce. +// (mul X, 2^n+1) -> (add (X << n), X) +// (mul X, 2^n-1) -> (rsb X, (X << n)) + +// ARM Arithmetic Instruction +// GPR:$dst = GPR:$a op GPR:$b +class AAI<bits<8> op27_20, bits<8> op11_4, string opc, + list<dag> pattern = [], + dag iops = (ins GPRnopc:$Rn, GPRnopc:$Rm), + string asm = "\t$Rd, $Rn, $Rm"> + : AI<(outs GPRnopc:$Rd), iops, DPFrm, IIC_iALUr, opc, asm, pattern>, + Sched<[WriteALU, ReadALU, ReadALU]> { + bits<4> Rn; + bits<4> Rd; + bits<4> Rm; + let Inst{27-20} = op27_20; + let Inst{11-4} = op11_4; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{3-0} = Rm; + + let Unpredictable{11-8} = 0b1111; +} + +// Wrappers around the AAI class +class AAIRevOpr<bits<8> op27_20, bits<8> op11_4, string opc, + list<dag> pattern = []> + : AAI<op27_20, op11_4, opc, + pattern, + (ins GPRnopc:$Rm, GPRnopc:$Rn), + "\t$Rd, $Rm, $Rn">; + +class AAIIntrinsic<bits<8> op27_20, bits<8> op11_4, string opc, + Intrinsic intrinsic> + : AAI<op27_20, op11_4, opc, + [(set GPRnopc:$Rd, (intrinsic GPRnopc:$Rn, GPRnopc:$Rm))]>; + +// Saturating add/subtract +let hasSideEffects = 1 in { +def QADD8 : AAIIntrinsic<0b01100010, 0b11111001, "qadd8", int_arm_qadd8>; +def QADD16 : AAIIntrinsic<0b01100010, 0b11110001, "qadd16", int_arm_qadd16>; +def QSUB16 : AAIIntrinsic<0b01100010, 0b11110111, "qsub16", int_arm_qsub16>; +def QSUB8 : AAIIntrinsic<0b01100010, 0b11111111, "qsub8", int_arm_qsub8>; + +def QDADD : AAIRevOpr<0b00010100, 0b00000101, "qdadd", + [(set GPRnopc:$Rd, (int_arm_qadd (int_arm_qadd GPRnopc:$Rm, + GPRnopc:$Rm), + GPRnopc:$Rn))]>; +def QDSUB : AAIRevOpr<0b00010110, 0b00000101, "qdsub", + [(set GPRnopc:$Rd, (int_arm_qsub GPRnopc:$Rm, + (int_arm_qadd GPRnopc:$Rn, GPRnopc:$Rn)))]>; +def QSUB : AAIRevOpr<0b00010010, 0b00000101, "qsub", + [(set GPRnopc:$Rd, (int_arm_qsub GPRnopc:$Rm, GPRnopc:$Rn))]>; +let DecoderMethod = "DecodeQADDInstruction" in + def QADD : AAIRevOpr<0b00010000, 0b00000101, "qadd", + [(set GPRnopc:$Rd, (int_arm_qadd GPRnopc:$Rm, GPRnopc:$Rn))]>; +} + +def UQADD16 : AAIIntrinsic<0b01100110, 0b11110001, "uqadd16", int_arm_uqadd16>; +def UQADD8 : AAIIntrinsic<0b01100110, 0b11111001, "uqadd8", int_arm_uqadd8>; +def UQSUB16 : AAIIntrinsic<0b01100110, 0b11110111, "uqsub16", int_arm_uqsub16>; +def UQSUB8 : AAIIntrinsic<0b01100110, 0b11111111, "uqsub8", int_arm_uqsub8>; +def QASX : AAIIntrinsic<0b01100010, 0b11110011, "qasx", int_arm_qasx>; +def QSAX : AAIIntrinsic<0b01100010, 0b11110101, "qsax", int_arm_qsax>; +def UQASX : AAIIntrinsic<0b01100110, 0b11110011, "uqasx", int_arm_uqasx>; +def UQSAX : AAIIntrinsic<0b01100110, 0b11110101, "uqsax", int_arm_uqsax>; + +// Signed/Unsigned add/subtract + +def SASX : AAIIntrinsic<0b01100001, 0b11110011, "sasx", int_arm_sasx>; +def SADD16 : AAIIntrinsic<0b01100001, 0b11110001, "sadd16", int_arm_sadd16>; +def SADD8 : AAIIntrinsic<0b01100001, 0b11111001, "sadd8", int_arm_sadd8>; +def SSAX : AAIIntrinsic<0b01100001, 0b11110101, "ssax", int_arm_ssax>; +def SSUB16 : AAIIntrinsic<0b01100001, 0b11110111, "ssub16", int_arm_ssub16>; +def SSUB8 : AAIIntrinsic<0b01100001, 0b11111111, "ssub8", int_arm_ssub8>; +def UASX : AAIIntrinsic<0b01100101, 0b11110011, "uasx", int_arm_uasx>; +def UADD16 : AAIIntrinsic<0b01100101, 0b11110001, "uadd16", int_arm_uadd16>; +def UADD8 : AAIIntrinsic<0b01100101, 0b11111001, "uadd8", int_arm_uadd8>; +def USAX : AAIIntrinsic<0b01100101, 0b11110101, "usax", int_arm_usax>; +def USUB16 : AAIIntrinsic<0b01100101, 0b11110111, "usub16", int_arm_usub16>; +def USUB8 : AAIIntrinsic<0b01100101, 0b11111111, "usub8", int_arm_usub8>; + +// Signed/Unsigned halving add/subtract + +def SHASX : AAIIntrinsic<0b01100011, 0b11110011, "shasx", int_arm_shasx>; +def SHADD16 : AAIIntrinsic<0b01100011, 0b11110001, "shadd16", int_arm_shadd16>; +def SHADD8 : AAIIntrinsic<0b01100011, 0b11111001, "shadd8", int_arm_shadd8>; +def SHSAX : AAIIntrinsic<0b01100011, 0b11110101, "shsax", int_arm_shsax>; +def SHSUB16 : AAIIntrinsic<0b01100011, 0b11110111, "shsub16", int_arm_shsub16>; +def SHSUB8 : AAIIntrinsic<0b01100011, 0b11111111, "shsub8", int_arm_shsub8>; +def UHASX : AAIIntrinsic<0b01100111, 0b11110011, "uhasx", int_arm_uhasx>; +def UHADD16 : AAIIntrinsic<0b01100111, 0b11110001, "uhadd16", int_arm_uhadd16>; +def UHADD8 : AAIIntrinsic<0b01100111, 0b11111001, "uhadd8", int_arm_uhadd8>; +def UHSAX : AAIIntrinsic<0b01100111, 0b11110101, "uhsax", int_arm_uhsax>; +def UHSUB16 : AAIIntrinsic<0b01100111, 0b11110111, "uhsub16", int_arm_uhsub16>; +def UHSUB8 : AAIIntrinsic<0b01100111, 0b11111111, "uhsub8", int_arm_uhsub8>; + +// Unsigned Sum of Absolute Differences [and Accumulate]. + +def USAD8 : AI<(outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + MulFrm /* for convenience */, NoItinerary, "usad8", + "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (int_arm_usad8 GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasV6]>, Sched<[WriteALU, ReadALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + let Inst{27-20} = 0b01111000; + let Inst{15-12} = 0b1111; + let Inst{7-4} = 0b0001; + let Inst{19-16} = Rd; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} +def USADA8 : AI<(outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm, GPR:$Ra), + MulFrm /* for convenience */, NoItinerary, "usada8", + "\t$Rd, $Rn, $Rm, $Ra", + [(set GPR:$Rd, (int_arm_usada8 GPR:$Rn, GPR:$Rm, GPR:$Ra))]>, + Requires<[IsARM, HasV6]>, Sched<[WriteALU, ReadALU, ReadALU]>{ + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + bits<4> Ra; + let Inst{27-20} = 0b01111000; + let Inst{7-4} = 0b0001; + let Inst{19-16} = Rd; + let Inst{15-12} = Ra; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} + +// Signed/Unsigned saturate +def SSAT : AI<(outs GPRnopc:$Rd), + (ins imm1_32:$sat_imm, GPRnopc:$Rn, shift_imm:$sh), + SatFrm, NoItinerary, "ssat", "\t$Rd, $sat_imm, $Rn$sh", []>, + Requires<[IsARM,HasV6]>{ + bits<4> Rd; + bits<5> sat_imm; + bits<4> Rn; + bits<8> sh; + let Inst{27-21} = 0b0110101; + let Inst{5-4} = 0b01; + let Inst{20-16} = sat_imm; + let Inst{15-12} = Rd; + let Inst{11-7} = sh{4-0}; + let Inst{6} = sh{5}; + let Inst{3-0} = Rn; +} + +def SSAT16 : AI<(outs GPRnopc:$Rd), + (ins imm1_16:$sat_imm, GPRnopc:$Rn), SatFrm, + NoItinerary, "ssat16", "\t$Rd, $sat_imm, $Rn", []>, + Requires<[IsARM,HasV6]>{ + bits<4> Rd; + bits<4> sat_imm; + bits<4> Rn; + let Inst{27-20} = 0b01101010; + let Inst{11-4} = 0b11110011; + let Inst{15-12} = Rd; + let Inst{19-16} = sat_imm; + let Inst{3-0} = Rn; +} + +def USAT : AI<(outs GPRnopc:$Rd), + (ins imm0_31:$sat_imm, GPRnopc:$Rn, shift_imm:$sh), + SatFrm, NoItinerary, "usat", "\t$Rd, $sat_imm, $Rn$sh", []>, + Requires<[IsARM,HasV6]> { + bits<4> Rd; + bits<5> sat_imm; + bits<4> Rn; + bits<8> sh; + let Inst{27-21} = 0b0110111; + let Inst{5-4} = 0b01; + let Inst{15-12} = Rd; + let Inst{11-7} = sh{4-0}; + let Inst{6} = sh{5}; + let Inst{20-16} = sat_imm; + let Inst{3-0} = Rn; +} + +def USAT16 : AI<(outs GPRnopc:$Rd), + (ins imm0_15:$sat_imm, GPRnopc:$Rn), SatFrm, + NoItinerary, "usat16", "\t$Rd, $sat_imm, $Rn", []>, + Requires<[IsARM,HasV6]>{ + bits<4> Rd; + bits<4> sat_imm; + bits<4> Rn; + let Inst{27-20} = 0b01101110; + let Inst{11-4} = 0b11110011; + let Inst{15-12} = Rd; + let Inst{19-16} = sat_imm; + let Inst{3-0} = Rn; +} + +def : ARMV6Pat<(int_arm_ssat GPRnopc:$a, imm1_32:$pos), + (SSAT imm1_32:$pos, GPRnopc:$a, 0)>; +def : ARMV6Pat<(int_arm_usat GPRnopc:$a, imm0_31:$pos), + (USAT imm0_31:$pos, GPRnopc:$a, 0)>; +def : ARMPat<(ARMssatnoshift GPRnopc:$Rn, imm0_31:$imm), + (SSAT imm0_31:$imm, GPRnopc:$Rn, 0)>; +def : ARMPat<(ARMusatnoshift GPRnopc:$Rn, imm0_31:$imm), + (USAT imm0_31:$imm, GPRnopc:$Rn, 0)>; +def : ARMV6Pat<(int_arm_ssat16 GPRnopc:$a, imm1_16:$pos), + (SSAT16 imm1_16:$pos, GPRnopc:$a)>; +def : ARMV6Pat<(int_arm_usat16 GPRnopc:$a, imm0_15:$pos), + (USAT16 imm0_15:$pos, GPRnopc:$a)>; + +//===----------------------------------------------------------------------===// +// Bitwise Instructions. +// + +defm AND : AsI1_bin_irs<0b0000, "and", + IIC_iBITi, IIC_iBITr, IIC_iBITsr, and, 1>; +defm ORR : AsI1_bin_irs<0b1100, "orr", + IIC_iBITi, IIC_iBITr, IIC_iBITsr, or, 1>; +defm EOR : AsI1_bin_irs<0b0001, "eor", + IIC_iBITi, IIC_iBITr, IIC_iBITsr, xor, 1>; +defm BIC : AsI1_bin_irs<0b1110, "bic", + IIC_iBITi, IIC_iBITr, IIC_iBITsr, + BinOpFrag<(and node:$LHS, (not node:$RHS))>>; + +// FIXME: bf_inv_mask_imm should be two operands, the lsb and the msb, just +// like in the actual instruction encoding. The complexity of mapping the mask +// to the lsb/msb pair should be handled by ISel, not encapsulated in the +// instruction description. +def BFC : I<(outs GPR:$Rd), (ins GPR:$src, bf_inv_mask_imm:$imm), + AddrMode1, 4, IndexModeNone, DPFrm, IIC_iUNAsi, + "bfc", "\t$Rd, $imm", "$src = $Rd", + [(set GPR:$Rd, (and GPR:$src, bf_inv_mask_imm:$imm))]>, + Requires<[IsARM, HasV6T2]> { + bits<4> Rd; + bits<10> imm; + let Inst{27-21} = 0b0111110; + let Inst{6-0} = 0b0011111; + let Inst{15-12} = Rd; + let Inst{11-7} = imm{4-0}; // lsb + let Inst{20-16} = imm{9-5}; // msb +} + +// A8.6.18 BFI - Bitfield insert (Encoding A1) +def BFI:I<(outs GPRnopc:$Rd), (ins GPRnopc:$src, GPR:$Rn, bf_inv_mask_imm:$imm), + AddrMode1, 4, IndexModeNone, DPFrm, IIC_iUNAsi, + "bfi", "\t$Rd, $Rn, $imm", "$src = $Rd", + [(set GPRnopc:$Rd, (ARMbfi GPRnopc:$src, GPR:$Rn, + bf_inv_mask_imm:$imm))]>, + Requires<[IsARM, HasV6T2]> { + bits<4> Rd; + bits<4> Rn; + bits<10> imm; + let Inst{27-21} = 0b0111110; + let Inst{6-4} = 0b001; // Rn: Inst{3-0} != 15 + let Inst{15-12} = Rd; + let Inst{11-7} = imm{4-0}; // lsb + let Inst{20-16} = imm{9-5}; // width + let Inst{3-0} = Rn; +} + +def MVNr : AsI1<0b1111, (outs GPR:$Rd), (ins GPR:$Rm), DPFrm, IIC_iMVNr, + "mvn", "\t$Rd, $Rm", + [(set GPR:$Rd, (not GPR:$Rm))]>, UnaryDP, Sched<[WriteALU]> { + bits<4> Rd; + bits<4> Rm; + let Inst{25} = 0; + let Inst{19-16} = 0b0000; + let Inst{11-4} = 0b00000000; + let Inst{15-12} = Rd; + let Inst{3-0} = Rm; + + let Unpredictable{19-16} = 0b1111; +} +def MVNsi : AsI1<0b1111, (outs GPR:$Rd), (ins so_reg_imm:$shift), + DPSoRegImmFrm, IIC_iMVNsr, "mvn", "\t$Rd, $shift", + [(set GPR:$Rd, (not so_reg_imm:$shift))]>, UnaryDP, + Sched<[WriteALU]> { + bits<4> Rd; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = 0b0000; + let Inst{15-12} = Rd; + let Inst{11-5} = shift{11-5}; + let Inst{4} = 0; + let Inst{3-0} = shift{3-0}; + + let Unpredictable{19-16} = 0b1111; +} +def MVNsr : AsI1<0b1111, (outs GPRnopc:$Rd), (ins so_reg_reg:$shift), + DPSoRegRegFrm, IIC_iMVNsr, "mvn", "\t$Rd, $shift", + [(set GPRnopc:$Rd, (not so_reg_reg:$shift))]>, UnaryDP, + Sched<[WriteALU]> { + bits<4> Rd; + bits<12> shift; + let Inst{25} = 0; + let Inst{19-16} = 0b0000; + let Inst{15-12} = Rd; + let Inst{11-8} = shift{11-8}; + let Inst{7} = 0; + let Inst{6-5} = shift{6-5}; + let Inst{4} = 1; + let Inst{3-0} = shift{3-0}; + + let Unpredictable{19-16} = 0b1111; +} +let isReMaterializable = 1, isAsCheapAsAMove = 1, isMoveImm = 1 in +def MVNi : AsI1<0b1111, (outs GPR:$Rd), (ins mod_imm:$imm), DPFrm, + IIC_iMVNi, "mvn", "\t$Rd, $imm", + [(set GPR:$Rd, mod_imm_not:$imm)]>,UnaryDP, Sched<[WriteALU]> { + bits<4> Rd; + bits<12> imm; + let Inst{25} = 1; + let Inst{19-16} = 0b0000; + let Inst{15-12} = Rd; + let Inst{11-0} = imm; +} + +let AddedComplexity = 1 in +def : ARMPat<(and GPR:$src, mod_imm_not:$imm), + (BICri GPR:$src, mod_imm_not:$imm)>; + +//===----------------------------------------------------------------------===// +// Multiply Instructions. +// +class AsMul1I32<bits<7> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AsMul1I<opcod, oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rm; + bits<4> Rn; + let Inst{19-16} = Rd; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} +class AsMul1I64<bits<7> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AsMul1I<opcod, oops, iops, itin, opc, asm, pattern> { + bits<4> RdLo; + bits<4> RdHi; + bits<4> Rm; + bits<4> Rn; + let Inst{19-16} = RdHi; + let Inst{15-12} = RdLo; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} +class AsMla1I64<bits<7> opcod, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AsMul1I<opcod, oops, iops, itin, opc, asm, pattern> { + bits<4> RdLo; + bits<4> RdHi; + bits<4> Rm; + bits<4> Rn; + let Inst{19-16} = RdHi; + let Inst{15-12} = RdLo; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} + +// FIXME: The v5 pseudos are only necessary for the additional Constraint +// property. Remove them when it's possible to add those properties +// on an individual MachineInstr, not just an instruction description. +let isCommutable = 1, TwoOperandAliasConstraint = "$Rn = $Rd" in { +def MUL : AsMul1I32<0b0000000, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm), + IIC_iMUL32, "mul", "\t$Rd, $Rn, $Rm", + [(set GPRnopc:$Rd, (mul GPRnopc:$Rn, GPRnopc:$Rm))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]> { + let Inst{15-12} = 0b0000; + let Unpredictable{15-12} = 0b1111; +} + +let Constraints = "@earlyclobber $Rd" in +def MULv5: ARMPseudoExpand<(outs GPRnopc:$Rd), (ins GPRnopc:$Rn, GPRnopc:$Rm, + pred:$p, cc_out:$s), + 4, IIC_iMUL32, + [(set GPRnopc:$Rd, (mul GPRnopc:$Rn, GPRnopc:$Rm))], + (MUL GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, pred:$p, cc_out:$s)>, + Requires<[IsARM, NoV6, UseMulOps]>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; +} + +def MLA : AsMul1I32<0b0000001, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$Ra), + IIC_iMAC32, "mla", "\t$Rd, $Rn, $Rm, $Ra", + [(set GPRnopc:$Rd, (add (mul GPRnopc:$Rn, GPRnopc:$Rm), GPRnopc:$Ra))]>, + Requires<[IsARM, HasV6, UseMulOps]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]> { + bits<4> Ra; + let Inst{15-12} = Ra; +} + +let Constraints = "@earlyclobber $Rd" in +def MLAv5: ARMPseudoExpand<(outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$Ra, + pred:$p, cc_out:$s), 4, IIC_iMAC32, + [(set GPRnopc:$Rd, (add (mul GPRnopc:$Rn, GPRnopc:$Rm), GPRnopc:$Ra))], + (MLA GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$Ra, pred:$p, cc_out:$s)>, + Requires<[IsARM, NoV6]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]>; + +def MLS : AMul1I<0b0000011, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm, GPR:$Ra), + IIC_iMAC32, "mls", "\t$Rd, $Rn, $Rm, $Ra", + [(set GPR:$Rd, (sub GPR:$Ra, (mul GPR:$Rn, GPR:$Rm)))]>, + Requires<[IsARM, HasV6T2, UseMulOps]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]> { + bits<4> Rd; + bits<4> Rm; + bits<4> Rn; + bits<4> Ra; + let Inst{19-16} = Rd; + let Inst{15-12} = Ra; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} + +// Extra precision multiplies with low / high results +let hasSideEffects = 0 in { +let isCommutable = 1 in { +def SMULL : AsMul1I64<0b0000110, (outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm), IIC_iMUL64, + "smull", "\t$RdLo, $RdHi, $Rn, $Rm", + [(set GPR:$RdLo, GPR:$RdHi, + (smullohi GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteMUL64Lo, WriteMUL64Hi, ReadMUL, ReadMUL]>; + +def UMULL : AsMul1I64<0b0000100, (outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm), IIC_iMUL64, + "umull", "\t$RdLo, $RdHi, $Rn, $Rm", + [(set GPR:$RdLo, GPR:$RdHi, + (umullohi GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL]>; + +let Constraints = "@earlyclobber $RdLo,@earlyclobber $RdHi" in { +def SMULLv5 : ARMPseudoExpand<(outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm, pred:$p, cc_out:$s), + 4, IIC_iMUL64, + [(set GPR:$RdLo, GPR:$RdHi, + (smullohi GPR:$Rn, GPR:$Rm))], + (SMULL GPR:$RdLo, GPR:$RdHi, GPR:$Rn, GPR:$Rm, pred:$p, cc_out:$s)>, + Requires<[IsARM, NoV6]>, + Sched<[WriteMUL64Lo, WriteMUL64Hi, ReadMUL, ReadMUL]>; + +def UMULLv5 : ARMPseudoExpand<(outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm, pred:$p, cc_out:$s), + 4, IIC_iMUL64, + [(set GPR:$RdLo, GPR:$RdHi, + (umullohi GPR:$Rn, GPR:$Rm))], + (UMULL GPR:$RdLo, GPR:$RdHi, GPR:$Rn, GPR:$Rm, pred:$p, cc_out:$s)>, + Requires<[IsARM, NoV6]>, + Sched<[WriteMUL64Lo, WriteMUL64Hi, ReadMUL, ReadMUL]>; +} +} + +// Multiply + accumulate +def SMLAL : AsMla1I64<0b0000111, (outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), IIC_iMAC64, + "smlal", "\t$RdLo, $RdHi, $Rn, $Rm", []>, + RegConstraint<"$RLo = $RdLo, $RHi = $RdHi">, Requires<[IsARM, HasV6]>, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]>; +def UMLAL : AsMla1I64<0b0000101, (outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), IIC_iMAC64, + "umlal", "\t$RdLo, $RdHi, $Rn, $Rm", []>, + RegConstraint<"$RLo = $RdLo, $RHi = $RdHi">, Requires<[IsARM, HasV6]>, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]>; + +def UMAAL : AMul1I <0b0000010, (outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + IIC_iMAC64, + "umaal", "\t$RdLo, $RdHi, $Rn, $Rm", []>, + RegConstraint<"$RLo = $RdLo, $RHi = $RdHi">, Requires<[IsARM, HasV6]>, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]> { + bits<4> RdLo; + bits<4> RdHi; + bits<4> Rm; + bits<4> Rn; + let Inst{19-16} = RdHi; + let Inst{15-12} = RdLo; + let Inst{11-8} = Rm; + let Inst{3-0} = Rn; +} + +let Constraints = + "@earlyclobber $RdLo,@earlyclobber $RdHi,$RLo = $RdLo,$RHi = $RdHi" in { +def SMLALv5 : ARMPseudoExpand<(outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi, pred:$p, cc_out:$s), + 4, IIC_iMAC64, [], + (SMLAL GPR:$RdLo, GPR:$RdHi, GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi, + pred:$p, cc_out:$s)>, + Requires<[IsARM, NoV6]>, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]>; +def UMLALv5 : ARMPseudoExpand<(outs GPR:$RdLo, GPR:$RdHi), + (ins GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi, pred:$p, cc_out:$s), + 4, IIC_iMAC64, [], + (UMLAL GPR:$RdLo, GPR:$RdHi, GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi, + pred:$p, cc_out:$s)>, + Requires<[IsARM, NoV6]>, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]>; +} + +} // hasSideEffects + +// Most significant word multiply +def SMMUL : AMul2I <0b0111010, 0b0001, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + IIC_iMUL32, "smmul", "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (mulhs GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]> { + let Inst{15-12} = 0b1111; +} + +def SMMULR : AMul2I <0b0111010, 0b0011, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + IIC_iMUL32, "smmulr", "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (ARMsmmlar GPR:$Rn, GPR:$Rm, (i32 0)))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]> { + let Inst{15-12} = 0b1111; +} + +def SMMLA : AMul2Ia <0b0111010, 0b0001, (outs GPR:$Rd), + (ins GPR:$Rn, GPR:$Rm, GPR:$Ra), + IIC_iMAC32, "smmla", "\t$Rd, $Rn, $Rm, $Ra", + [(set GPR:$Rd, (add (mulhs GPR:$Rn, GPR:$Rm), GPR:$Ra))]>, + Requires<[IsARM, HasV6, UseMulOps]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]>; + +def SMMLAR : AMul2Ia <0b0111010, 0b0011, (outs GPR:$Rd), + (ins GPR:$Rn, GPR:$Rm, GPR:$Ra), + IIC_iMAC32, "smmlar", "\t$Rd, $Rn, $Rm, $Ra", + [(set GPR:$Rd, (ARMsmmlar GPR:$Rn, GPR:$Rm, GPR:$Ra))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]>; + +def SMMLS : AMul2Ia <0b0111010, 0b1101, (outs GPR:$Rd), + (ins GPR:$Rn, GPR:$Rm, GPR:$Ra), + IIC_iMAC32, "smmls", "\t$Rd, $Rn, $Rm, $Ra", []>, + Requires<[IsARM, HasV6, UseMulOps]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]>; + +def SMMLSR : AMul2Ia <0b0111010, 0b1111, (outs GPR:$Rd), + (ins GPR:$Rn, GPR:$Rm, GPR:$Ra), + IIC_iMAC32, "smmlsr", "\t$Rd, $Rn, $Rm, $Ra", + [(set GPR:$Rd, (ARMsmmlsr GPR:$Rn, GPR:$Rm, GPR:$Ra))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]>; + +multiclass AI_smul<string opc> { + def BB : AMulxyI<0b0001011, 0b00, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + IIC_iMUL16, !strconcat(opc, "bb"), "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (mul (sext_inreg GPR:$Rn, i16), + (sext_inreg GPR:$Rm, i16)))]>, + Requires<[IsARM, HasV5TE]>, + Sched<[WriteMUL16, ReadMUL, ReadMUL]>; + + def BT : AMulxyI<0b0001011, 0b10, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + IIC_iMUL16, !strconcat(opc, "bt"), "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (mul (sext_inreg GPR:$Rn, i16), + (sra GPR:$Rm, (i32 16))))]>, + Requires<[IsARM, HasV5TE]>, + Sched<[WriteMUL16, ReadMUL, ReadMUL]>; + + def TB : AMulxyI<0b0001011, 0b01, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + IIC_iMUL16, !strconcat(opc, "tb"), "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (mul (sra GPR:$Rn, (i32 16)), + (sext_inreg GPR:$Rm, i16)))]>, + Requires<[IsARM, HasV5TE]>, + Sched<[WriteMUL16, ReadMUL, ReadMUL]>; + + def TT : AMulxyI<0b0001011, 0b11, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + IIC_iMUL16, !strconcat(opc, "tt"), "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (mul (sra GPR:$Rn, (i32 16)), + (sra GPR:$Rm, (i32 16))))]>, + Requires<[IsARM, HasV5TE]>, + Sched<[WriteMUL16, ReadMUL, ReadMUL]>; + + def WB : AMulxyI<0b0001001, 0b01, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + IIC_iMUL16, !strconcat(opc, "wb"), "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (ARMsmulwb GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasV5TE]>, + Sched<[WriteMUL16, ReadMUL, ReadMUL]>; + + def WT : AMulxyI<0b0001001, 0b11, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + IIC_iMUL16, !strconcat(opc, "wt"), "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (ARMsmulwt GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasV5TE]>, + Sched<[WriteMUL16, ReadMUL, ReadMUL]>; +} + + +multiclass AI_smla<string opc> { + let DecoderMethod = "DecodeSMLAInstruction" in { + def BB : AMulxyIa<0b0001000, 0b00, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + IIC_iMAC16, !strconcat(opc, "bb"), "\t$Rd, $Rn, $Rm, $Ra", + [(set GPRnopc:$Rd, (add GPR:$Ra, + (mul (sext_inreg GPRnopc:$Rn, i16), + (sext_inreg GPRnopc:$Rm, i16))))]>, + Requires<[IsARM, HasV5TE, UseMulOps]>, + Sched<[WriteMAC16, ReadMUL, ReadMUL, ReadMAC]>; + + def BT : AMulxyIa<0b0001000, 0b10, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + IIC_iMAC16, !strconcat(opc, "bt"), "\t$Rd, $Rn, $Rm, $Ra", + [(set GPRnopc:$Rd, + (add GPR:$Ra, (mul (sext_inreg GPRnopc:$Rn, i16), + (sra GPRnopc:$Rm, (i32 16)))))]>, + Requires<[IsARM, HasV5TE, UseMulOps]>, + Sched<[WriteMAC16, ReadMUL, ReadMUL, ReadMAC]>; + + def TB : AMulxyIa<0b0001000, 0b01, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + IIC_iMAC16, !strconcat(opc, "tb"), "\t$Rd, $Rn, $Rm, $Ra", + [(set GPRnopc:$Rd, + (add GPR:$Ra, (mul (sra GPRnopc:$Rn, (i32 16)), + (sext_inreg GPRnopc:$Rm, i16))))]>, + Requires<[IsARM, HasV5TE, UseMulOps]>, + Sched<[WriteMAC16, ReadMUL, ReadMUL, ReadMAC]>; + + def TT : AMulxyIa<0b0001000, 0b11, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + IIC_iMAC16, !strconcat(opc, "tt"), "\t$Rd, $Rn, $Rm, $Ra", + [(set GPRnopc:$Rd, + (add GPR:$Ra, (mul (sra GPRnopc:$Rn, (i32 16)), + (sra GPRnopc:$Rm, (i32 16)))))]>, + Requires<[IsARM, HasV5TE, UseMulOps]>, + Sched<[WriteMAC16, ReadMUL, ReadMUL, ReadMAC]>; + + def WB : AMulxyIa<0b0001001, 0b00, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + IIC_iMAC16, !strconcat(opc, "wb"), "\t$Rd, $Rn, $Rm, $Ra", + [(set GPRnopc:$Rd, + (add GPR:$Ra, (ARMsmulwb GPRnopc:$Rn, GPRnopc:$Rm)))]>, + Requires<[IsARM, HasV5TE, UseMulOps]>, + Sched<[WriteMAC16, ReadMUL, ReadMUL, ReadMAC]>; + + def WT : AMulxyIa<0b0001001, 0b10, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + IIC_iMAC16, !strconcat(opc, "wt"), "\t$Rd, $Rn, $Rm, $Ra", + [(set GPRnopc:$Rd, + (add GPR:$Ra, (ARMsmulwt GPRnopc:$Rn, GPRnopc:$Rm)))]>, + Requires<[IsARM, HasV5TE, UseMulOps]>, + Sched<[WriteMAC16, ReadMUL, ReadMUL, ReadMAC]>; + } +} + +defm SMUL : AI_smul<"smul">; +defm SMLA : AI_smla<"smla">; + +// Halfword multiply accumulate long: SMLAL<x><y>. +class SMLAL<bits<2> opc1, string asm> + : AMulxyI64<0b0001010, opc1, + (outs GPRnopc:$RdLo, GPRnopc:$RdHi), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi), + IIC_iMAC64, asm, "\t$RdLo, $RdHi, $Rn, $Rm", []>, + RegConstraint<"$RLo = $RdLo, $RHi = $RdHi">, + Requires<[IsARM, HasV5TE]>, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]>; + +def SMLALBB : SMLAL<0b00, "smlalbb">; +def SMLALBT : SMLAL<0b10, "smlalbt">; +def SMLALTB : SMLAL<0b01, "smlaltb">; +def SMLALTT : SMLAL<0b11, "smlaltt">; + +def : ARMV5TEPat<(ARMsmlalbb GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + (SMLALBB $Rn, $Rm, $RLo, $RHi)>; +def : ARMV5TEPat<(ARMsmlalbt GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + (SMLALBT $Rn, $Rm, $RLo, $RHi)>; +def : ARMV5TEPat<(ARMsmlaltb GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + (SMLALTB $Rn, $Rm, $RLo, $RHi)>; +def : ARMV5TEPat<(ARMsmlaltt GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + (SMLALTT $Rn, $Rm, $RLo, $RHi)>; + +// Helper class for AI_smld. +class AMulDualIbase<bit long, bit sub, bit swap, dag oops, dag iops, + InstrItinClass itin, string opc, string asm> + : AI<oops, iops, MulFrm, itin, opc, asm, []>, + Requires<[IsARM, HasV6]> { + bits<4> Rn; + bits<4> Rm; + let Inst{27-23} = 0b01110; + let Inst{22} = long; + let Inst{21-20} = 0b00; + let Inst{11-8} = Rm; + let Inst{7} = 0; + let Inst{6} = sub; + let Inst{5} = swap; + let Inst{4} = 1; + let Inst{3-0} = Rn; +} +class AMulDualI<bit long, bit sub, bit swap, dag oops, dag iops, + InstrItinClass itin, string opc, string asm> + : AMulDualIbase<long, sub, swap, oops, iops, itin, opc, asm> { + bits<4> Rd; + let Inst{15-12} = 0b1111; + let Inst{19-16} = Rd; +} +class AMulDualIa<bit long, bit sub, bit swap, dag oops, dag iops, + InstrItinClass itin, string opc, string asm> + : AMulDualIbase<long, sub, swap, oops, iops, itin, opc, asm> { + bits<4> Ra; + bits<4> Rd; + let Inst{19-16} = Rd; + let Inst{15-12} = Ra; +} +class AMulDualI64<bit long, bit sub, bit swap, dag oops, dag iops, + InstrItinClass itin, string opc, string asm> + : AMulDualIbase<long, sub, swap, oops, iops, itin, opc, asm> { + bits<4> RdLo; + bits<4> RdHi; + let Inst{19-16} = RdHi; + let Inst{15-12} = RdLo; +} + +multiclass AI_smld<bit sub, string opc> { + + def D : AMulDualIa<0, sub, 0, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + NoItinerary, !strconcat(opc, "d"), "\t$Rd, $Rn, $Rm, $Ra">, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]>; + + def DX: AMulDualIa<0, sub, 1, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + NoItinerary, !strconcat(opc, "dx"), "\t$Rd, $Rn, $Rm, $Ra">, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]>; + + def LD: AMulDualI64<1, sub, 0, (outs GPRnopc:$RdLo, GPRnopc:$RdHi), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi), + NoItinerary, + !strconcat(opc, "ld"), "\t$RdLo, $RdHi, $Rn, $Rm">, + RegConstraint<"$RLo = $RdLo, $RHi = $RdHi">, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]>; + + def LDX : AMulDualI64<1, sub, 1, (outs GPRnopc:$RdLo, GPRnopc:$RdHi), + (ins GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi), + NoItinerary, + !strconcat(opc, "ldx"),"\t$RdLo, $RdHi, $Rn, $Rm">, + RegConstraint<"$RLo = $RdLo, $RHi = $RdHi">, + Sched<[WriteMUL64Lo, WriteMUL64Hi, ReadMUL, ReadMUL]>; +} + +defm SMLA : AI_smld<0, "smla">; +defm SMLS : AI_smld<1, "smls">; + +def : ARMV6Pat<(int_arm_smlad GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + (SMLAD GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$Ra)>; +def : ARMV6Pat<(int_arm_smladx GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + (SMLADX GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$Ra)>; +def : ARMV6Pat<(int_arm_smlsd GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + (SMLSD GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$Ra)>; +def : ARMV6Pat<(int_arm_smlsdx GPRnopc:$Rn, GPRnopc:$Rm, GPR:$Ra), + (SMLSDX GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$Ra)>; +def : ARMV6Pat<(ARMSmlald GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi), + (SMLALD GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi)>; +def : ARMV6Pat<(ARMSmlaldx GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi), + (SMLALDX GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi)>; +def : ARMV6Pat<(ARMSmlsld GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi), + (SMLSLD GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi)>; +def : ARMV6Pat<(ARMSmlsldx GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi), + (SMLSLDX GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$RLo, GPRnopc:$RHi)>; + +multiclass AI_sdml<bit sub, string opc> { + + def D:AMulDualI<0, sub, 0, (outs GPRnopc:$Rd), (ins GPRnopc:$Rn, GPRnopc:$Rm), + NoItinerary, !strconcat(opc, "d"), "\t$Rd, $Rn, $Rm">, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; + def DX:AMulDualI<0, sub, 1, (outs GPRnopc:$Rd),(ins GPRnopc:$Rn, GPRnopc:$Rm), + NoItinerary, !strconcat(opc, "dx"), "\t$Rd, $Rn, $Rm">, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; +} + +defm SMUA : AI_sdml<0, "smua">; +defm SMUS : AI_sdml<1, "smus">; + +def : ARMV6Pat<(int_arm_smuad GPRnopc:$Rn, GPRnopc:$Rm), + (SMUAD GPRnopc:$Rn, GPRnopc:$Rm)>; +def : ARMV6Pat<(int_arm_smuadx GPRnopc:$Rn, GPRnopc:$Rm), + (SMUADX GPRnopc:$Rn, GPRnopc:$Rm)>; +def : ARMV6Pat<(int_arm_smusd GPRnopc:$Rn, GPRnopc:$Rm), + (SMUSD GPRnopc:$Rn, GPRnopc:$Rm)>; +def : ARMV6Pat<(int_arm_smusdx GPRnopc:$Rn, GPRnopc:$Rm), + (SMUSDX GPRnopc:$Rn, GPRnopc:$Rm)>; + +//===----------------------------------------------------------------------===// +// Division Instructions (ARMv7-A with virtualization extension) +// +def SDIV : ADivA1I<0b001, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), IIC_iDIV, + "sdiv", "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (sdiv GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasDivideInARM]>, + Sched<[WriteDIV]>; + +def UDIV : ADivA1I<0b011, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), IIC_iDIV, + "udiv", "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (udiv GPR:$Rn, GPR:$Rm))]>, + Requires<[IsARM, HasDivideInARM]>, + Sched<[WriteDIV]>; + +//===----------------------------------------------------------------------===// +// Misc. Arithmetic Instructions. +// + +def CLZ : AMiscA1I<0b00010110, 0b0001, (outs GPR:$Rd), (ins GPR:$Rm), + IIC_iUNAr, "clz", "\t$Rd, $Rm", + [(set GPR:$Rd, (ctlz GPR:$Rm))]>, Requires<[IsARM, HasV5T]>, + Sched<[WriteALU]>; + +def RBIT : AMiscA1I<0b01101111, 0b0011, (outs GPR:$Rd), (ins GPR:$Rm), + IIC_iUNAr, "rbit", "\t$Rd, $Rm", + [(set GPR:$Rd, (bitreverse GPR:$Rm))]>, + Requires<[IsARM, HasV6T2]>, + Sched<[WriteALU]>; + +def REV : AMiscA1I<0b01101011, 0b0011, (outs GPR:$Rd), (ins GPR:$Rm), + IIC_iUNAr, "rev", "\t$Rd, $Rm", + [(set GPR:$Rd, (bswap GPR:$Rm))]>, Requires<[IsARM, HasV6]>, + Sched<[WriteALU]>; + +let AddedComplexity = 5 in +def REV16 : AMiscA1I<0b01101011, 0b1011, (outs GPR:$Rd), (ins GPR:$Rm), + IIC_iUNAr, "rev16", "\t$Rd, $Rm", + [(set GPR:$Rd, (rotr (bswap GPR:$Rm), (i32 16)))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteALU]>; + +def : ARMV6Pat<(srl (bswap (extloadi16 addrmode3:$addr)), (i32 16)), + (REV16 (LDRH addrmode3:$addr))>; +def : ARMV6Pat<(truncstorei16 (srl (bswap GPR:$Rn), (i32 16)), addrmode3:$addr), + (STRH (REV16 GPR:$Rn), addrmode3:$addr)>; + +let AddedComplexity = 5 in +def REVSH : AMiscA1I<0b01101111, 0b1011, (outs GPR:$Rd), (ins GPR:$Rm), + IIC_iUNAr, "revsh", "\t$Rd, $Rm", + [(set GPR:$Rd, (sra (bswap GPR:$Rm), (i32 16)))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteALU]>; + +def : ARMV6Pat<(or (sra (shl GPR:$Rm, (i32 24)), (i32 16)), + (and (srl GPR:$Rm, (i32 8)), 0xFF)), + (REVSH GPR:$Rm)>; + +def PKHBT : APKHI<0b01101000, 0, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, pkh_lsl_amt:$sh), + IIC_iALUsi, "pkhbt", "\t$Rd, $Rn, $Rm$sh", + [(set GPRnopc:$Rd, (or (and GPRnopc:$Rn, 0xFFFF), + (and (shl GPRnopc:$Rm, pkh_lsl_amt:$sh), + 0xFFFF0000)))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteALUsi, ReadALU]>; + +// Alternate cases for PKHBT where identities eliminate some nodes. +def : ARMV6Pat<(or (and GPRnopc:$Rn, 0xFFFF), (and GPRnopc:$Rm, 0xFFFF0000)), + (PKHBT GPRnopc:$Rn, GPRnopc:$Rm, 0)>; +def : ARMV6Pat<(or (and GPRnopc:$Rn, 0xFFFF), (shl GPRnopc:$Rm, imm16_31:$sh)), + (PKHBT GPRnopc:$Rn, GPRnopc:$Rm, imm16_31:$sh)>; + +// Note: Shifts of 1-15 bits will be transformed to srl instead of sra and +// will match the pattern below. +def PKHTB : APKHI<0b01101000, 1, (outs GPRnopc:$Rd), + (ins GPRnopc:$Rn, GPRnopc:$Rm, pkh_asr_amt:$sh), + IIC_iBITsi, "pkhtb", "\t$Rd, $Rn, $Rm$sh", + [(set GPRnopc:$Rd, (or (and GPRnopc:$Rn, 0xFFFF0000), + (and (sra GPRnopc:$Rm, pkh_asr_amt:$sh), + 0xFFFF)))]>, + Requires<[IsARM, HasV6]>, + Sched<[WriteALUsi, ReadALU]>; + +// Alternate cases for PKHTB where identities eliminate some nodes. Note that +// a shift amount of 0 is *not legal* here, it is PKHBT instead. +// We also can not replace a srl (17..31) by an arithmetic shift we would use in +// pkhtb src1, src2, asr (17..31). +def : ARMV6Pat<(or (and GPRnopc:$src1, 0xFFFF0000), + (srl GPRnopc:$src2, imm16:$sh)), + (PKHTB GPRnopc:$src1, GPRnopc:$src2, imm16:$sh)>; +def : ARMV6Pat<(or (and GPRnopc:$src1, 0xFFFF0000), + (sra GPRnopc:$src2, imm16_31:$sh)), + (PKHTB GPRnopc:$src1, GPRnopc:$src2, imm16_31:$sh)>; +def : ARMV6Pat<(or (and GPRnopc:$src1, 0xFFFF0000), + (and (srl GPRnopc:$src2, imm1_15:$sh), 0xFFFF)), + (PKHTB GPRnopc:$src1, GPRnopc:$src2, imm1_15:$sh)>; + +//===----------------------------------------------------------------------===// +// CRC Instructions +// +// Polynomials: +// + CRC32{B,H,W} 0x04C11DB7 +// + CRC32C{B,H,W} 0x1EDC6F41 +// + +class AI_crc32<bit C, bits<2> sz, string suffix, SDPatternOperator builtin> + : AInoP<(outs GPRnopc:$Rd), (ins GPRnopc:$Rn, GPRnopc:$Rm), MiscFrm, NoItinerary, + !strconcat("crc32", suffix), "\t$Rd, $Rn, $Rm", + [(set GPRnopc:$Rd, (builtin GPRnopc:$Rn, GPRnopc:$Rm))]>, + Requires<[IsARM, HasV8, HasCRC]> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + + let Inst{31-28} = 0b1110; + let Inst{27-23} = 0b00010; + let Inst{22-21} = sz; + let Inst{20} = 0; + let Inst{19-16} = Rn; + let Inst{15-12} = Rd; + let Inst{11-10} = 0b00; + let Inst{9} = C; + let Inst{8} = 0; + let Inst{7-4} = 0b0100; + let Inst{3-0} = Rm; + + let Unpredictable{11-8} = 0b1101; +} + +def CRC32B : AI_crc32<0, 0b00, "b", int_arm_crc32b>; +def CRC32CB : AI_crc32<1, 0b00, "cb", int_arm_crc32cb>; +def CRC32H : AI_crc32<0, 0b01, "h", int_arm_crc32h>; +def CRC32CH : AI_crc32<1, 0b01, "ch", int_arm_crc32ch>; +def CRC32W : AI_crc32<0, 0b10, "w", int_arm_crc32w>; +def CRC32CW : AI_crc32<1, 0b10, "cw", int_arm_crc32cw>; + +//===----------------------------------------------------------------------===// +// ARMv8.1a Privilege Access Never extension +// +// SETPAN #imm1 + +def SETPAN : AInoP<(outs), (ins imm0_1:$imm), MiscFrm, NoItinerary, "setpan", + "\t$imm", []>, Requires<[IsARM, HasV8, HasV8_1a]> { + bits<1> imm; + + let Inst{31-28} = 0b1111; + let Inst{27-20} = 0b00010001; + let Inst{19-16} = 0b0000; + let Inst{15-10} = 0b000000; + let Inst{9} = imm; + let Inst{8} = 0b0; + let Inst{7-4} = 0b0000; + let Inst{3-0} = 0b0000; + + let Unpredictable{19-16} = 0b1111; + let Unpredictable{15-10} = 0b111111; + let Unpredictable{8} = 0b1; + let Unpredictable{3-0} = 0b1111; +} + +//===----------------------------------------------------------------------===// +// Comparison Instructions... +// + +defm CMP : AI1_cmp_irs<0b1010, "cmp", + IIC_iCMPi, IIC_iCMPr, IIC_iCMPsr, ARMcmp>; + +// ARMcmpZ can re-use the above instruction definitions. +def : ARMPat<(ARMcmpZ GPR:$src, mod_imm:$imm), + (CMPri GPR:$src, mod_imm:$imm)>; +def : ARMPat<(ARMcmpZ GPR:$src, GPR:$rhs), + (CMPrr GPR:$src, GPR:$rhs)>; +def : ARMPat<(ARMcmpZ GPR:$src, so_reg_imm:$rhs), + (CMPrsi GPR:$src, so_reg_imm:$rhs)>; +def : ARMPat<(ARMcmpZ GPR:$src, so_reg_reg:$rhs), + (CMPrsr GPR:$src, so_reg_reg:$rhs)>; + +// CMN register-integer +let isCompare = 1, Defs = [CPSR] in { +def CMNri : AI1<0b1011, (outs), (ins GPR:$Rn, mod_imm:$imm), DPFrm, IIC_iCMPi, + "cmn", "\t$Rn, $imm", + [(ARMcmn GPR:$Rn, mod_imm:$imm)]>, + Sched<[WriteCMP, ReadALU]> { + bits<4> Rn; + bits<12> imm; + let Inst{25} = 1; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b0000; + let Inst{11-0} = imm; + + let Unpredictable{15-12} = 0b1111; +} + +// CMN register-register/shift +def CMNzrr : AI1<0b1011, (outs), (ins GPR:$Rn, GPR:$Rm), DPFrm, IIC_iCMPr, + "cmn", "\t$Rn, $Rm", + [(BinOpFrag<(ARMcmpZ node:$LHS,(ineg node:$RHS))> + GPR:$Rn, GPR:$Rm)]>, Sched<[WriteCMP, ReadALU, ReadALU]> { + bits<4> Rn; + bits<4> Rm; + let isCommutable = 1; + let Inst{25} = 0; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b0000; + let Inst{11-4} = 0b00000000; + let Inst{3-0} = Rm; + + let Unpredictable{15-12} = 0b1111; +} + +def CMNzrsi : AI1<0b1011, (outs), + (ins GPR:$Rn, so_reg_imm:$shift), DPSoRegImmFrm, IIC_iCMPsr, + "cmn", "\t$Rn, $shift", + [(BinOpFrag<(ARMcmpZ node:$LHS,(ineg node:$RHS))> + GPR:$Rn, so_reg_imm:$shift)]>, + Sched<[WriteCMPsi, ReadALU]> { + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b0000; + let Inst{11-5} = shift{11-5}; + let Inst{4} = 0; + let Inst{3-0} = shift{3-0}; + + let Unpredictable{15-12} = 0b1111; +} + +def CMNzrsr : AI1<0b1011, (outs), + (ins GPRnopc:$Rn, so_reg_reg:$shift), DPSoRegRegFrm, IIC_iCMPsr, + "cmn", "\t$Rn, $shift", + [(BinOpFrag<(ARMcmpZ node:$LHS,(ineg node:$RHS))> + GPRnopc:$Rn, so_reg_reg:$shift)]>, + Sched<[WriteCMPsr, ReadALU]> { + bits<4> Rn; + bits<12> shift; + let Inst{25} = 0; + let Inst{20} = 1; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b0000; + let Inst{11-8} = shift{11-8}; + let Inst{7} = 0; + let Inst{6-5} = shift{6-5}; + let Inst{4} = 1; + let Inst{3-0} = shift{3-0}; + + let Unpredictable{15-12} = 0b1111; +} + +} + +def : ARMPat<(ARMcmp GPR:$src, mod_imm_neg:$imm), + (CMNri GPR:$src, mod_imm_neg:$imm)>; + +def : ARMPat<(ARMcmpZ GPR:$src, mod_imm_neg:$imm), + (CMNri GPR:$src, mod_imm_neg:$imm)>; + +// Note that TST/TEQ don't set all the same flags that CMP does! +defm TST : AI1_cmp_irs<0b1000, "tst", + IIC_iTSTi, IIC_iTSTr, IIC_iTSTsr, + BinOpFrag<(ARMcmpZ (and_su node:$LHS, node:$RHS), 0)>, 1, + "DecodeTSTInstruction">; +defm TEQ : AI1_cmp_irs<0b1001, "teq", + IIC_iTSTi, IIC_iTSTr, IIC_iTSTsr, + BinOpFrag<(ARMcmpZ (xor_su node:$LHS, node:$RHS), 0)>, 1>; + +// Pseudo i64 compares for some floating point compares. +let usesCustomInserter = 1, isBranch = 1, isTerminator = 1, + Defs = [CPSR] in { +def BCCi64 : PseudoInst<(outs), + (ins i32imm:$cc, GPR:$lhs1, GPR:$lhs2, GPR:$rhs1, GPR:$rhs2, brtarget:$dst), + IIC_Br, + [(ARMBcci64 imm:$cc, GPR:$lhs1, GPR:$lhs2, GPR:$rhs1, GPR:$rhs2, bb:$dst)]>, + Sched<[WriteBr]>; + +def BCCZi64 : PseudoInst<(outs), + (ins i32imm:$cc, GPR:$lhs1, GPR:$lhs2, brtarget:$dst), IIC_Br, + [(ARMBcci64 imm:$cc, GPR:$lhs1, GPR:$lhs2, 0, 0, bb:$dst)]>, + Sched<[WriteBr]>; +} // usesCustomInserter + + +// Conditional moves +let hasSideEffects = 0 in { + +let isCommutable = 1, isSelect = 1 in +def MOVCCr : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$false, GPR:$Rm, cmovpred:$p), + 4, IIC_iCMOVr, + [(set GPR:$Rd, (ARMcmov GPR:$false, GPR:$Rm, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + +def MOVCCsi : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$false, so_reg_imm:$shift, cmovpred:$p), + 4, IIC_iCMOVsr, + [(set GPR:$Rd, + (ARMcmov GPR:$false, so_reg_imm:$shift, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; +def MOVCCsr : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$false, so_reg_reg:$shift, cmovpred:$p), + 4, IIC_iCMOVsr, + [(set GPR:$Rd, (ARMcmov GPR:$false, so_reg_reg:$shift, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + + +let isMoveImm = 1 in +def MOVCCi16 + : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$false, imm0_65535_expr:$imm, cmovpred:$p), + 4, IIC_iMOVi, + [(set GPR:$Rd, (ARMcmov GPR:$false, imm0_65535:$imm, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Requires<[IsARM, HasV6T2]>, + Sched<[WriteALU]>; + +let isMoveImm = 1 in +def MOVCCi : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$false, mod_imm:$imm, cmovpred:$p), + 4, IIC_iCMOVi, + [(set GPR:$Rd, (ARMcmov GPR:$false, mod_imm:$imm, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + +// Two instruction predicate mov immediate. +let isMoveImm = 1 in +def MOVCCi32imm + : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$false, i32imm:$src, cmovpred:$p), + 8, IIC_iCMOVix2, + [(set GPR:$Rd, (ARMcmov GPR:$false, imm:$src, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Requires<[IsARM, HasV6T2]>; + +let isMoveImm = 1 in +def MVNCCi : ARMPseudoInst<(outs GPR:$Rd), + (ins GPR:$false, mod_imm:$imm, cmovpred:$p), + 4, IIC_iCMOVi, + [(set GPR:$Rd, (ARMcmov GPR:$false, mod_imm_not:$imm, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + +} // hasSideEffects + + +//===----------------------------------------------------------------------===// +// Atomic operations intrinsics +// + +def MemBarrierOptOperand : AsmOperandClass { + let Name = "MemBarrierOpt"; + let ParserMethod = "parseMemBarrierOptOperand"; +} +def memb_opt : Operand<i32> { + let PrintMethod = "printMemBOption"; + let ParserMatchClass = MemBarrierOptOperand; + let DecoderMethod = "DecodeMemBarrierOption"; +} + +def InstSyncBarrierOptOperand : AsmOperandClass { + let Name = "InstSyncBarrierOpt"; + let ParserMethod = "parseInstSyncBarrierOptOperand"; +} +def instsyncb_opt : Operand<i32> { + let PrintMethod = "printInstSyncBOption"; + let ParserMatchClass = InstSyncBarrierOptOperand; + let DecoderMethod = "DecodeInstSyncBarrierOption"; +} + +def TraceSyncBarrierOptOperand : AsmOperandClass { + let Name = "TraceSyncBarrierOpt"; + let ParserMethod = "parseTraceSyncBarrierOptOperand"; +} +def tsb_opt : Operand<i32> { + let PrintMethod = "printTraceSyncBOption"; + let ParserMatchClass = TraceSyncBarrierOptOperand; +} + +// Memory barriers protect the atomic sequences +let hasSideEffects = 1 in { +def DMB : AInoP<(outs), (ins memb_opt:$opt), MiscFrm, NoItinerary, + "dmb", "\t$opt", [(int_arm_dmb (i32 imm0_15:$opt))]>, + Requires<[IsARM, HasDB]> { + bits<4> opt; + let Inst{31-4} = 0xf57ff05; + let Inst{3-0} = opt; +} + +def DSB : AInoP<(outs), (ins memb_opt:$opt), MiscFrm, NoItinerary, + "dsb", "\t$opt", [(int_arm_dsb (i32 imm0_15:$opt))]>, + Requires<[IsARM, HasDB]> { + bits<4> opt; + let Inst{31-4} = 0xf57ff04; + let Inst{3-0} = opt; +} + +// ISB has only full system option +def ISB : AInoP<(outs), (ins instsyncb_opt:$opt), MiscFrm, NoItinerary, + "isb", "\t$opt", [(int_arm_isb (i32 imm0_15:$opt))]>, + Requires<[IsARM, HasDB]> { + bits<4> opt; + let Inst{31-4} = 0xf57ff06; + let Inst{3-0} = opt; +} + +let hasNoSchedulingInfo = 1 in +def TSB : AInoP<(outs), (ins tsb_opt:$opt), MiscFrm, NoItinerary, + "tsb", "\t$opt", []>, Requires<[IsARM, HasV8_4a]> { + let Inst{31-0} = 0xe320f012; +} + +} + +let usesCustomInserter = 1, Defs = [CPSR] in { + +// Pseudo instruction that combines movs + predicated rsbmi +// to implement integer ABS + def ABS : ARMPseudoInst<(outs GPR:$dst), (ins GPR:$src), 8, NoItinerary, []>; +} + +let usesCustomInserter = 1 in { + def COPY_STRUCT_BYVAL_I32 : PseudoInst< + (outs), (ins GPR:$dst, GPR:$src, i32imm:$size, i32imm:$alignment), + NoItinerary, + [(ARMcopystructbyval GPR:$dst, GPR:$src, imm:$size, imm:$alignment)]>; +} + +let hasPostISelHook = 1, Constraints = "$newdst = $dst, $newsrc = $src" in { + // %newsrc, %newdst = MEMCPY %dst, %src, N, ...N scratch regs... + // Copies N registers worth of memory from address %src to address %dst + // and returns the incremented addresses. N scratch register will + // be attached for the copy to use. + def MEMCPY : PseudoInst< + (outs GPR:$newdst, GPR:$newsrc), + (ins GPR:$dst, GPR:$src, i32imm:$nreg, variable_ops), + NoItinerary, + [(set GPR:$newdst, GPR:$newsrc, + (ARMmemcopy GPR:$dst, GPR:$src, imm:$nreg))]>; +} + +def ldrex_1 : PatFrag<(ops node:$ptr), (int_arm_ldrex node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i8; +}]>; + +def ldrex_2 : PatFrag<(ops node:$ptr), (int_arm_ldrex node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i16; +}]>; + +def ldrex_4 : PatFrag<(ops node:$ptr), (int_arm_ldrex node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32; +}]>; + +def strex_1 : PatFrag<(ops node:$val, node:$ptr), + (int_arm_strex node:$val, node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i8; +}]>; + +def strex_2 : PatFrag<(ops node:$val, node:$ptr), + (int_arm_strex node:$val, node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i16; +}]>; + +def strex_4 : PatFrag<(ops node:$val, node:$ptr), + (int_arm_strex node:$val, node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32; +}]>; + +def ldaex_1 : PatFrag<(ops node:$ptr), (int_arm_ldaex node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i8; +}]>; + +def ldaex_2 : PatFrag<(ops node:$ptr), (int_arm_ldaex node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i16; +}]>; + +def ldaex_4 : PatFrag<(ops node:$ptr), (int_arm_ldaex node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32; +}]>; + +def stlex_1 : PatFrag<(ops node:$val, node:$ptr), + (int_arm_stlex node:$val, node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i8; +}]>; + +def stlex_2 : PatFrag<(ops node:$val, node:$ptr), + (int_arm_stlex node:$val, node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i16; +}]>; + +def stlex_4 : PatFrag<(ops node:$val, node:$ptr), + (int_arm_stlex node:$val, node:$ptr), [{ + return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32; +}]>; + +let mayLoad = 1 in { +def LDREXB : AIldrex<0b10, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "ldrexb", "\t$Rt, $addr", + [(set GPR:$Rt, (ldrex_1 addr_offset_none:$addr))]>; +def LDREXH : AIldrex<0b11, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "ldrexh", "\t$Rt, $addr", + [(set GPR:$Rt, (ldrex_2 addr_offset_none:$addr))]>; +def LDREX : AIldrex<0b00, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "ldrex", "\t$Rt, $addr", + [(set GPR:$Rt, (ldrex_4 addr_offset_none:$addr))]>; +let hasExtraDefRegAllocReq = 1 in +def LDREXD : AIldrex<0b01, (outs GPRPairOp:$Rt),(ins addr_offset_none:$addr), + NoItinerary, "ldrexd", "\t$Rt, $addr", []> { + let DecoderMethod = "DecodeDoubleRegLoad"; +} + +def LDAEXB : AIldaex<0b10, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "ldaexb", "\t$Rt, $addr", + [(set GPR:$Rt, (ldaex_1 addr_offset_none:$addr))]>; +def LDAEXH : AIldaex<0b11, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "ldaexh", "\t$Rt, $addr", + [(set GPR:$Rt, (ldaex_2 addr_offset_none:$addr))]>; +def LDAEX : AIldaex<0b00, (outs GPR:$Rt), (ins addr_offset_none:$addr), + NoItinerary, "ldaex", "\t$Rt, $addr", + [(set GPR:$Rt, (ldaex_4 addr_offset_none:$addr))]>; +let hasExtraDefRegAllocReq = 1 in +def LDAEXD : AIldaex<0b01, (outs GPRPairOp:$Rt),(ins addr_offset_none:$addr), + NoItinerary, "ldaexd", "\t$Rt, $addr", []> { + let DecoderMethod = "DecodeDoubleRegLoad"; +} +} + +let mayStore = 1, Constraints = "@earlyclobber $Rd" in { +def STREXB: AIstrex<0b10, (outs GPR:$Rd), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "strexb", "\t$Rd, $Rt, $addr", + [(set GPR:$Rd, (strex_1 GPR:$Rt, + addr_offset_none:$addr))]>; +def STREXH: AIstrex<0b11, (outs GPR:$Rd), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "strexh", "\t$Rd, $Rt, $addr", + [(set GPR:$Rd, (strex_2 GPR:$Rt, + addr_offset_none:$addr))]>; +def STREX : AIstrex<0b00, (outs GPR:$Rd), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "strex", "\t$Rd, $Rt, $addr", + [(set GPR:$Rd, (strex_4 GPR:$Rt, + addr_offset_none:$addr))]>; +let hasExtraSrcRegAllocReq = 1 in +def STREXD : AIstrex<0b01, (outs GPR:$Rd), + (ins GPRPairOp:$Rt, addr_offset_none:$addr), + NoItinerary, "strexd", "\t$Rd, $Rt, $addr", []> { + let DecoderMethod = "DecodeDoubleRegStore"; +} +def STLEXB: AIstlex<0b10, (outs GPR:$Rd), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "stlexb", "\t$Rd, $Rt, $addr", + [(set GPR:$Rd, + (stlex_1 GPR:$Rt, addr_offset_none:$addr))]>; +def STLEXH: AIstlex<0b11, (outs GPR:$Rd), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "stlexh", "\t$Rd, $Rt, $addr", + [(set GPR:$Rd, + (stlex_2 GPR:$Rt, addr_offset_none:$addr))]>; +def STLEX : AIstlex<0b00, (outs GPR:$Rd), (ins GPR:$Rt, addr_offset_none:$addr), + NoItinerary, "stlex", "\t$Rd, $Rt, $addr", + [(set GPR:$Rd, + (stlex_4 GPR:$Rt, addr_offset_none:$addr))]>; +let hasExtraSrcRegAllocReq = 1 in +def STLEXD : AIstlex<0b01, (outs GPR:$Rd), + (ins GPRPairOp:$Rt, addr_offset_none:$addr), + NoItinerary, "stlexd", "\t$Rd, $Rt, $addr", []> { + let DecoderMethod = "DecodeDoubleRegStore"; +} +} + +def CLREX : AXI<(outs), (ins), MiscFrm, NoItinerary, "clrex", + [(int_arm_clrex)]>, + Requires<[IsARM, HasV6K]> { + let Inst{31-0} = 0b11110101011111111111000000011111; +} + +def : ARMPat<(strex_1 (and GPR:$Rt, 0xff), addr_offset_none:$addr), + (STREXB GPR:$Rt, addr_offset_none:$addr)>; +def : ARMPat<(strex_2 (and GPR:$Rt, 0xffff), addr_offset_none:$addr), + (STREXH GPR:$Rt, addr_offset_none:$addr)>; + +def : ARMPat<(stlex_1 (and GPR:$Rt, 0xff), addr_offset_none:$addr), + (STLEXB GPR:$Rt, addr_offset_none:$addr)>; +def : ARMPat<(stlex_2 (and GPR:$Rt, 0xffff), addr_offset_none:$addr), + (STLEXH GPR:$Rt, addr_offset_none:$addr)>; + +class acquiring_load<PatFrag base> + : PatFrag<(ops node:$ptr), (base node:$ptr), [{ + AtomicOrdering Ordering = cast<AtomicSDNode>(N)->getOrdering(); + return isAcquireOrStronger(Ordering); +}]>; + +def atomic_load_acquire_8 : acquiring_load<atomic_load_8>; +def atomic_load_acquire_16 : acquiring_load<atomic_load_16>; +def atomic_load_acquire_32 : acquiring_load<atomic_load_32>; + +class releasing_store<PatFrag base> + : PatFrag<(ops node:$ptr, node:$val), (base node:$ptr, node:$val), [{ + AtomicOrdering Ordering = cast<AtomicSDNode>(N)->getOrdering(); + return isReleaseOrStronger(Ordering); +}]>; + +def atomic_store_release_8 : releasing_store<atomic_store_8>; +def atomic_store_release_16 : releasing_store<atomic_store_16>; +def atomic_store_release_32 : releasing_store<atomic_store_32>; + +let AddedComplexity = 8 in { + def : ARMPat<(atomic_load_acquire_8 addr_offset_none:$addr), (LDAB addr_offset_none:$addr)>; + def : ARMPat<(atomic_load_acquire_16 addr_offset_none:$addr), (LDAH addr_offset_none:$addr)>; + def : ARMPat<(atomic_load_acquire_32 addr_offset_none:$addr), (LDA addr_offset_none:$addr)>; + def : ARMPat<(atomic_store_release_8 addr_offset_none:$addr, GPR:$val), (STLB GPR:$val, addr_offset_none:$addr)>; + def : ARMPat<(atomic_store_release_16 addr_offset_none:$addr, GPR:$val), (STLH GPR:$val, addr_offset_none:$addr)>; + def : ARMPat<(atomic_store_release_32 addr_offset_none:$addr, GPR:$val), (STL GPR:$val, addr_offset_none:$addr)>; +} + +// SWP/SWPB are deprecated in V6/V7 and optional in v7VE. +// FIXME Use InstAlias to generate LDREX/STREX pairs instead. +let mayLoad = 1, mayStore = 1 in { +def SWP : AIswp<0, (outs GPRnopc:$Rt), + (ins GPRnopc:$Rt2, addr_offset_none:$addr), "swp", []>, + Requires<[IsARM,PreV8]>; +def SWPB: AIswp<1, (outs GPRnopc:$Rt), + (ins GPRnopc:$Rt2, addr_offset_none:$addr), "swpb", []>, + Requires<[IsARM,PreV8]>; +} + +//===----------------------------------------------------------------------===// +// Coprocessor Instructions. +// + +def CDP : ABI<0b1110, (outs), (ins p_imm:$cop, imm0_15:$opc1, + c_imm:$CRd, c_imm:$CRn, c_imm:$CRm, imm0_7:$opc2), + NoItinerary, "cdp", "\t$cop, $opc1, $CRd, $CRn, $CRm, $opc2", + [(int_arm_cdp imm:$cop, imm:$opc1, imm:$CRd, imm:$CRn, + imm:$CRm, imm:$opc2)]>, + Requires<[IsARM,PreV8]> { + bits<4> opc1; + bits<4> CRn; + bits<4> CRd; + bits<4> cop; + bits<3> opc2; + bits<4> CRm; + + let Inst{3-0} = CRm; + let Inst{4} = 0; + let Inst{7-5} = opc2; + let Inst{11-8} = cop; + let Inst{15-12} = CRd; + let Inst{19-16} = CRn; + let Inst{23-20} = opc1; + + let DecoderNamespace = "CoProc"; +} + +def CDP2 : ABXI<0b1110, (outs), (ins p_imm:$cop, imm0_15:$opc1, + c_imm:$CRd, c_imm:$CRn, c_imm:$CRm, imm0_7:$opc2), + NoItinerary, "cdp2\t$cop, $opc1, $CRd, $CRn, $CRm, $opc2", + [(int_arm_cdp2 imm:$cop, imm:$opc1, imm:$CRd, imm:$CRn, + imm:$CRm, imm:$opc2)]>, + Requires<[IsARM,PreV8]> { + let Inst{31-28} = 0b1111; + bits<4> opc1; + bits<4> CRn; + bits<4> CRd; + bits<4> cop; + bits<3> opc2; + bits<4> CRm; + + let Inst{3-0} = CRm; + let Inst{4} = 0; + let Inst{7-5} = opc2; + let Inst{11-8} = cop; + let Inst{15-12} = CRd; + let Inst{19-16} = CRn; + let Inst{23-20} = opc1; + + let DecoderNamespace = "CoProc"; +} + +class ACI<dag oops, dag iops, string opc, string asm, + list<dag> pattern, IndexMode im = IndexModeNone> + : I<oops, iops, AddrModeNone, 4, im, BrFrm, NoItinerary, + opc, asm, "", pattern> { + let Inst{27-25} = 0b110; +} +class ACInoP<dag oops, dag iops, string opc, string asm, + list<dag> pattern, IndexMode im = IndexModeNone> + : InoP<oops, iops, AddrModeNone, 4, im, BrFrm, NoItinerary, + opc, asm, "", pattern> { + let Inst{31-28} = 0b1111; + let Inst{27-25} = 0b110; +} + +let DecoderNamespace = "CoProc" in { +multiclass LdStCop<bit load, bit Dbit, string asm, list<dag> pattern> { + def _OFFSET : ACI<(outs), (ins p_imm:$cop, c_imm:$CRd, addrmode5:$addr), + asm, "\t$cop, $CRd, $addr", pattern> { + bits<13> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 1; // P = 1 + let Inst{23} = addr{8}; + let Inst{22} = Dbit; + let Inst{21} = 0; // W = 0 + let Inst{20} = load; + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = addr{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _PRE : ACI<(outs), (ins p_imm:$cop, c_imm:$CRd, addrmode5_pre:$addr), + asm, "\t$cop, $CRd, $addr!", [], IndexModePre> { + bits<13> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 1; // P = 1 + let Inst{23} = addr{8}; + let Inst{22} = Dbit; + let Inst{21} = 1; // W = 1 + let Inst{20} = load; + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = addr{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _POST: ACI<(outs), (ins p_imm:$cop, c_imm:$CRd, addr_offset_none:$addr, + postidx_imm8s4:$offset), + asm, "\t$cop, $CRd, $addr, $offset", [], IndexModePost> { + bits<9> offset; + bits<4> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 0; // P = 0 + let Inst{23} = offset{8}; + let Inst{22} = Dbit; + let Inst{21} = 1; // W = 1 + let Inst{20} = load; + let Inst{19-16} = addr; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = offset{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _OPTION : ACI<(outs), + (ins p_imm:$cop, c_imm:$CRd, addr_offset_none:$addr, + coproc_option_imm:$option), + asm, "\t$cop, $CRd, $addr, $option", []> { + bits<8> option; + bits<4> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 0; // P = 0 + let Inst{23} = 1; // U = 1 + let Inst{22} = Dbit; + let Inst{21} = 0; // W = 0 + let Inst{20} = load; + let Inst{19-16} = addr; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = option; + let DecoderMethod = "DecodeCopMemInstruction"; + } +} +multiclass LdSt2Cop<bit load, bit Dbit, string asm, list<dag> pattern> { + def _OFFSET : ACInoP<(outs), (ins p_imm:$cop, c_imm:$CRd, addrmode5:$addr), + asm, "\t$cop, $CRd, $addr", pattern> { + bits<13> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 1; // P = 1 + let Inst{23} = addr{8}; + let Inst{22} = Dbit; + let Inst{21} = 0; // W = 0 + let Inst{20} = load; + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = addr{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _PRE : ACInoP<(outs), (ins p_imm:$cop, c_imm:$CRd, addrmode5_pre:$addr), + asm, "\t$cop, $CRd, $addr!", [], IndexModePre> { + bits<13> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 1; // P = 1 + let Inst{23} = addr{8}; + let Inst{22} = Dbit; + let Inst{21} = 1; // W = 1 + let Inst{20} = load; + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = addr{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _POST: ACInoP<(outs), (ins p_imm:$cop, c_imm:$CRd, addr_offset_none:$addr, + postidx_imm8s4:$offset), + asm, "\t$cop, $CRd, $addr, $offset", [], IndexModePost> { + bits<9> offset; + bits<4> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 0; // P = 0 + let Inst{23} = offset{8}; + let Inst{22} = Dbit; + let Inst{21} = 1; // W = 1 + let Inst{20} = load; + let Inst{19-16} = addr; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = offset{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _OPTION : ACInoP<(outs), + (ins p_imm:$cop, c_imm:$CRd, addr_offset_none:$addr, + coproc_option_imm:$option), + asm, "\t$cop, $CRd, $addr, $option", []> { + bits<8> option; + bits<4> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 0; // P = 0 + let Inst{23} = 1; // U = 1 + let Inst{22} = Dbit; + let Inst{21} = 0; // W = 0 + let Inst{20} = load; + let Inst{19-16} = addr; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = option; + let DecoderMethod = "DecodeCopMemInstruction"; + } +} + +defm LDC : LdStCop <1, 0, "ldc", [(int_arm_ldc imm:$cop, imm:$CRd, addrmode5:$addr)]>; +defm LDCL : LdStCop <1, 1, "ldcl", [(int_arm_ldcl imm:$cop, imm:$CRd, addrmode5:$addr)]>; +defm LDC2 : LdSt2Cop<1, 0, "ldc2", [(int_arm_ldc2 imm:$cop, imm:$CRd, addrmode5:$addr)]>, Requires<[IsARM,PreV8]>; +defm LDC2L : LdSt2Cop<1, 1, "ldc2l", [(int_arm_ldc2l imm:$cop, imm:$CRd, addrmode5:$addr)]>, Requires<[IsARM,PreV8]>; + +defm STC : LdStCop <0, 0, "stc", [(int_arm_stc imm:$cop, imm:$CRd, addrmode5:$addr)]>; +defm STCL : LdStCop <0, 1, "stcl", [(int_arm_stcl imm:$cop, imm:$CRd, addrmode5:$addr)]>; +defm STC2 : LdSt2Cop<0, 0, "stc2", [(int_arm_stc2 imm:$cop, imm:$CRd, addrmode5:$addr)]>, Requires<[IsARM,PreV8]>; +defm STC2L : LdSt2Cop<0, 1, "stc2l", [(int_arm_stc2l imm:$cop, imm:$CRd, addrmode5:$addr)]>, Requires<[IsARM,PreV8]>; + +} // DecoderNamespace = "CoProc" + +//===----------------------------------------------------------------------===// +// Move between coprocessor and ARM core register. +// + +class MovRCopro<string opc, bit direction, dag oops, dag iops, + list<dag> pattern> + : ABI<0b1110, oops, iops, NoItinerary, opc, + "\t$cop, $opc1, $Rt, $CRn, $CRm, $opc2", pattern> { + let Inst{20} = direction; + let Inst{4} = 1; + + bits<4> Rt; + bits<4> cop; + bits<3> opc1; + bits<3> opc2; + bits<4> CRm; + bits<4> CRn; + + let Inst{15-12} = Rt; + let Inst{11-8} = cop; + let Inst{23-21} = opc1; + let Inst{7-5} = opc2; + let Inst{3-0} = CRm; + let Inst{19-16} = CRn; + + let DecoderNamespace = "CoProc"; +} + +def MCR : MovRCopro<"mcr", 0 /* from ARM core register to coprocessor */, + (outs), + (ins p_imm:$cop, imm0_7:$opc1, GPR:$Rt, c_imm:$CRn, + c_imm:$CRm, imm0_7:$opc2), + [(int_arm_mcr imm:$cop, imm:$opc1, GPR:$Rt, imm:$CRn, + imm:$CRm, imm:$opc2)]>, + ComplexDeprecationPredicate<"MCR">; +def : ARMInstAlias<"mcr${p} $cop, $opc1, $Rt, $CRn, $CRm", + (MCR p_imm:$cop, imm0_7:$opc1, GPR:$Rt, c_imm:$CRn, + c_imm:$CRm, 0, pred:$p)>; +def MRC : MovRCopro<"mrc", 1 /* from coprocessor to ARM core register */, + (outs GPRwithAPSR:$Rt), + (ins p_imm:$cop, imm0_7:$opc1, c_imm:$CRn, c_imm:$CRm, + imm0_7:$opc2), []>; +def : ARMInstAlias<"mrc${p} $cop, $opc1, $Rt, $CRn, $CRm", + (MRC GPRwithAPSR:$Rt, p_imm:$cop, imm0_7:$opc1, c_imm:$CRn, + c_imm:$CRm, 0, pred:$p)>; + +def : ARMPat<(int_arm_mrc imm:$cop, imm:$opc1, imm:$CRn, imm:$CRm, imm:$opc2), + (MRC imm:$cop, imm:$opc1, imm:$CRn, imm:$CRm, imm:$opc2)>; + +class MovRCopro2<string opc, bit direction, dag oops, dag iops, + list<dag> pattern> + : ABXI<0b1110, oops, iops, NoItinerary, + !strconcat(opc, "\t$cop, $opc1, $Rt, $CRn, $CRm, $opc2"), pattern> { + let Inst{31-24} = 0b11111110; + let Inst{20} = direction; + let Inst{4} = 1; + + bits<4> Rt; + bits<4> cop; + bits<3> opc1; + bits<3> opc2; + bits<4> CRm; + bits<4> CRn; + + let Inst{15-12} = Rt; + let Inst{11-8} = cop; + let Inst{23-21} = opc1; + let Inst{7-5} = opc2; + let Inst{3-0} = CRm; + let Inst{19-16} = CRn; + + let DecoderNamespace = "CoProc"; +} + +def MCR2 : MovRCopro2<"mcr2", 0 /* from ARM core register to coprocessor */, + (outs), + (ins p_imm:$cop, imm0_7:$opc1, GPR:$Rt, c_imm:$CRn, + c_imm:$CRm, imm0_7:$opc2), + [(int_arm_mcr2 imm:$cop, imm:$opc1, GPR:$Rt, imm:$CRn, + imm:$CRm, imm:$opc2)]>, + Requires<[IsARM,PreV8]>; +def : ARMInstAlias<"mcr2 $cop, $opc1, $Rt, $CRn, $CRm", + (MCR2 p_imm:$cop, imm0_7:$opc1, GPR:$Rt, c_imm:$CRn, + c_imm:$CRm, 0)>; +def MRC2 : MovRCopro2<"mrc2", 1 /* from coprocessor to ARM core register */, + (outs GPRwithAPSR:$Rt), + (ins p_imm:$cop, imm0_7:$opc1, c_imm:$CRn, c_imm:$CRm, + imm0_7:$opc2), []>, + Requires<[IsARM,PreV8]>; +def : ARMInstAlias<"mrc2 $cop, $opc1, $Rt, $CRn, $CRm", + (MRC2 GPRwithAPSR:$Rt, p_imm:$cop, imm0_7:$opc1, c_imm:$CRn, + c_imm:$CRm, 0)>; + +def : ARMV5TPat<(int_arm_mrc2 imm:$cop, imm:$opc1, imm:$CRn, + imm:$CRm, imm:$opc2), + (MRC2 imm:$cop, imm:$opc1, imm:$CRn, imm:$CRm, imm:$opc2)>; + +class MovRRCopro<string opc, bit direction, dag oops, dag iops, list<dag> + pattern = []> + : ABI<0b1100, oops, iops, NoItinerary, opc, "\t$cop, $opc1, $Rt, $Rt2, $CRm", + pattern> { + + let Inst{23-21} = 0b010; + let Inst{20} = direction; + + bits<4> Rt; + bits<4> Rt2; + bits<4> cop; + bits<4> opc1; + bits<4> CRm; + + let Inst{15-12} = Rt; + let Inst{19-16} = Rt2; + let Inst{11-8} = cop; + let Inst{7-4} = opc1; + let Inst{3-0} = CRm; +} + +def MCRR : MovRRCopro<"mcrr", 0 /* from ARM core register to coprocessor */, + (outs), (ins p_imm:$cop, imm0_15:$opc1, GPRnopc:$Rt, + GPRnopc:$Rt2, c_imm:$CRm), + [(int_arm_mcrr imm:$cop, imm:$opc1, GPRnopc:$Rt, + GPRnopc:$Rt2, imm:$CRm)]>; +def MRRC : MovRRCopro<"mrrc", 1 /* from coprocessor to ARM core register */, + (outs GPRnopc:$Rt, GPRnopc:$Rt2), + (ins p_imm:$cop, imm0_15:$opc1, c_imm:$CRm), []>; + +class MovRRCopro2<string opc, bit direction, dag oops, dag iops, + list<dag> pattern = []> + : ABXI<0b1100, oops, iops, NoItinerary, + !strconcat(opc, "\t$cop, $opc1, $Rt, $Rt2, $CRm"), pattern>, + Requires<[IsARM,PreV8]> { + let Inst{31-28} = 0b1111; + let Inst{23-21} = 0b010; + let Inst{20} = direction; + + bits<4> Rt; + bits<4> Rt2; + bits<4> cop; + bits<4> opc1; + bits<4> CRm; + + let Inst{15-12} = Rt; + let Inst{19-16} = Rt2; + let Inst{11-8} = cop; + let Inst{7-4} = opc1; + let Inst{3-0} = CRm; + + let DecoderMethod = "DecoderForMRRC2AndMCRR2"; +} + +def MCRR2 : MovRRCopro2<"mcrr2", 0 /* from ARM core register to coprocessor */, + (outs), (ins p_imm:$cop, imm0_15:$opc1, GPRnopc:$Rt, + GPRnopc:$Rt2, c_imm:$CRm), + [(int_arm_mcrr2 imm:$cop, imm:$opc1, GPRnopc:$Rt, + GPRnopc:$Rt2, imm:$CRm)]>; + +def MRRC2 : MovRRCopro2<"mrrc2", 1 /* from coprocessor to ARM core register */, + (outs GPRnopc:$Rt, GPRnopc:$Rt2), + (ins p_imm:$cop, imm0_15:$opc1, c_imm:$CRm), []>; + +//===----------------------------------------------------------------------===// +// Move between special register and ARM core register +// + +// Move to ARM core register from Special Register +def MRS : ABI<0b0001, (outs GPRnopc:$Rd), (ins), NoItinerary, + "mrs", "\t$Rd, apsr", []> { + bits<4> Rd; + let Inst{23-16} = 0b00001111; + let Unpredictable{19-17} = 0b111; + + let Inst{15-12} = Rd; + + let Inst{11-0} = 0b000000000000; + let Unpredictable{11-0} = 0b110100001111; +} + +def : InstAlias<"mrs${p} $Rd, cpsr", (MRS GPRnopc:$Rd, pred:$p), 0>, + Requires<[IsARM]>; + +// The MRSsys instruction is the MRS instruction from the ARM ARM, +// section B9.3.9, with the R bit set to 1. +def MRSsys : ABI<0b0001, (outs GPRnopc:$Rd), (ins), NoItinerary, + "mrs", "\t$Rd, spsr", []> { + bits<4> Rd; + let Inst{23-16} = 0b01001111; + let Unpredictable{19-16} = 0b1111; + + let Inst{15-12} = Rd; + + let Inst{11-0} = 0b000000000000; + let Unpredictable{11-0} = 0b110100001111; +} + +// However, the MRS (banked register) system instruction (ARMv7VE) *does* have a +// separate encoding (distinguished by bit 5. +def MRSbanked : ABI<0b0001, (outs GPRnopc:$Rd), (ins banked_reg:$banked), + NoItinerary, "mrs", "\t$Rd, $banked", []>, + Requires<[IsARM, HasVirtualization]> { + bits<6> banked; + bits<4> Rd; + + let Inst{23} = 0; + let Inst{22} = banked{5}; // R bit + let Inst{21-20} = 0b00; + let Inst{19-16} = banked{3-0}; + let Inst{15-12} = Rd; + let Inst{11-9} = 0b001; + let Inst{8} = banked{4}; + let Inst{7-0} = 0b00000000; +} + +// Move from ARM core register to Special Register +// +// No need to have both system and application versions of MSR (immediate) or +// MSR (register), the encodings are the same and the assembly parser has no way +// to distinguish between them. The mask operand contains the special register +// (R Bit) in bit 4 and bits 3-0 contains the mask with the fields to be +// accessed in the special register. +let Defs = [CPSR] in +def MSR : ABI<0b0001, (outs), (ins msr_mask:$mask, GPR:$Rn), NoItinerary, + "msr", "\t$mask, $Rn", []> { + bits<5> mask; + bits<4> Rn; + + let Inst{23} = 0; + let Inst{22} = mask{4}; // R bit + let Inst{21-20} = 0b10; + let Inst{19-16} = mask{3-0}; + let Inst{15-12} = 0b1111; + let Inst{11-4} = 0b00000000; + let Inst{3-0} = Rn; +} + +let Defs = [CPSR] in +def MSRi : ABI<0b0011, (outs), (ins msr_mask:$mask, mod_imm:$imm), NoItinerary, + "msr", "\t$mask, $imm", []> { + bits<5> mask; + bits<12> imm; + + let Inst{23} = 0; + let Inst{22} = mask{4}; // R bit + let Inst{21-20} = 0b10; + let Inst{19-16} = mask{3-0}; + let Inst{15-12} = 0b1111; + let Inst{11-0} = imm; +} + +// However, the MSR (banked register) system instruction (ARMv7VE) *does* have a +// separate encoding (distinguished by bit 5. +def MSRbanked : ABI<0b0001, (outs), (ins banked_reg:$banked, GPRnopc:$Rn), + NoItinerary, "msr", "\t$banked, $Rn", []>, + Requires<[IsARM, HasVirtualization]> { + bits<6> banked; + bits<4> Rn; + + let Inst{23} = 0; + let Inst{22} = banked{5}; // R bit + let Inst{21-20} = 0b10; + let Inst{19-16} = banked{3-0}; + let Inst{15-12} = 0b1111; + let Inst{11-9} = 0b001; + let Inst{8} = banked{4}; + let Inst{7-4} = 0b0000; + let Inst{3-0} = Rn; +} + +// Dynamic stack allocation yields a _chkstk for Windows targets. These calls +// are needed to probe the stack when allocating more than +// 4k bytes in one go. Touching the stack at 4K increments is necessary to +// ensure that the guard pages used by the OS virtual memory manager are +// allocated in correct sequence. +// The main point of having separate instruction are extra unmodelled effects +// (compared to ordinary calls) like stack pointer change. + +def win__chkstk : SDNode<"ARMISD::WIN__CHKSTK", SDTNone, + [SDNPHasChain, SDNPSideEffect]>; +let usesCustomInserter = 1, Uses = [R4], Defs = [R4, SP] in + def WIN__CHKSTK : PseudoInst<(outs), (ins), NoItinerary, [(win__chkstk)]>; + +def win__dbzchk : SDNode<"ARMISD::WIN__DBZCHK", SDT_WIN__DBZCHK, + [SDNPHasChain, SDNPSideEffect, SDNPOutGlue]>; +let usesCustomInserter = 1, Defs = [CPSR] in + def WIN__DBZCHK : PseudoInst<(outs), (ins tGPR:$divisor), NoItinerary, + [(win__dbzchk tGPR:$divisor)]>; + +//===----------------------------------------------------------------------===// +// TLS Instructions +// + +// __aeabi_read_tp preserves the registers r1-r3. +// This is a pseudo inst so that we can get the encoding right, +// complete with fixup for the aeabi_read_tp function. +// TPsoft is valid for ARM mode only, in case of Thumb mode a tTPsoft pattern +// is defined in "ARMInstrThumb.td". +let isCall = 1, + Defs = [R0, R12, LR, CPSR], Uses = [SP] in { + def TPsoft : ARMPseudoInst<(outs), (ins), 4, IIC_Br, + [(set R0, ARMthread_pointer)]>, Sched<[WriteBr]>, + Requires<[IsARM, IsReadTPSoft]>; +} + +// Reading thread pointer from coprocessor register +def : ARMPat<(ARMthread_pointer), (MRC 15, 0, 13, 0, 3)>, + Requires<[IsARM, IsReadTPHard]>; + +//===----------------------------------------------------------------------===// +// SJLJ Exception handling intrinsics +// eh_sjlj_setjmp() is an instruction sequence to store the return +// address and save #0 in R0 for the non-longjmp case. +// Since by its nature we may be coming from some other function to get +// here, and we're using the stack frame for the containing function to +// save/restore registers, we can't keep anything live in regs across +// the eh_sjlj_setjmp(), else it will almost certainly have been tromped upon +// when we get here from a longjmp(). We force everything out of registers +// except for our own input by listing the relevant registers in Defs. By +// doing so, we also cause the prologue/epilogue code to actively preserve +// all of the callee-saved resgisters, which is exactly what we want. +// A constant value is passed in $val, and we use the location as a scratch. +// +// These are pseudo-instructions and are lowered to individual MC-insts, so +// no encoding information is necessary. +let Defs = + [ R0, R1, R2, R3, R4, R5, R6, R7, R8, R9, R10, R11, R12, LR, CPSR, + Q0, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, Q10, Q11, Q12, Q13, Q14, Q15 ], + hasSideEffects = 1, isBarrier = 1, usesCustomInserter = 1 in { + def Int_eh_sjlj_setjmp : PseudoInst<(outs), (ins GPR:$src, GPR:$val), + NoItinerary, + [(set R0, (ARMeh_sjlj_setjmp GPR:$src, GPR:$val))]>, + Requires<[IsARM, HasVFP2]>; +} + +let Defs = + [ R0, R1, R2, R3, R4, R5, R6, R7, R8, R9, R10, R11, R12, LR, CPSR ], + hasSideEffects = 1, isBarrier = 1, usesCustomInserter = 1 in { + def Int_eh_sjlj_setjmp_nofp : PseudoInst<(outs), (ins GPR:$src, GPR:$val), + NoItinerary, + [(set R0, (ARMeh_sjlj_setjmp GPR:$src, GPR:$val))]>, + Requires<[IsARM, NoVFP]>; +} + +// FIXME: Non-IOS version(s) +let isBarrier = 1, hasSideEffects = 1, isTerminator = 1, + Defs = [ R7, LR, SP ] in { +def Int_eh_sjlj_longjmp : PseudoInst<(outs), (ins GPR:$src, GPR:$scratch), + NoItinerary, + [(ARMeh_sjlj_longjmp GPR:$src, GPR:$scratch)]>, + Requires<[IsARM]>; +} + +let isBarrier = 1, hasSideEffects = 1, usesCustomInserter = 1 in +def Int_eh_sjlj_setup_dispatch : PseudoInst<(outs), (ins), NoItinerary, + [(ARMeh_sjlj_setup_dispatch)]>; + +// eh.sjlj.dispatchsetup pseudo-instruction. +// This pseudo is used for both ARM and Thumb. Any differences are handled when +// the pseudo is expanded (which happens before any passes that need the +// instruction size). +let isBarrier = 1 in +def Int_eh_sjlj_dispatchsetup : PseudoInst<(outs), (ins), NoItinerary, []>; + + +//===----------------------------------------------------------------------===// +// Non-Instruction Patterns +// + +// ARMv4 indirect branch using (MOVr PC, dst) +let isBranch = 1, isTerminator = 1, isBarrier = 1, isIndirectBranch = 1 in + def MOVPCRX : ARMPseudoExpand<(outs), (ins GPR:$dst), + 4, IIC_Br, [(brind GPR:$dst)], + (MOVr PC, GPR:$dst, (ops 14, zero_reg), zero_reg)>, + Requires<[IsARM, NoV4T]>, Sched<[WriteBr]>; + +let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, Uses = [SP] in + def TAILJMPr4 : ARMPseudoExpand<(outs), (ins GPR:$dst), + 4, IIC_Br, [], + (MOVr PC, GPR:$dst, (ops 14, zero_reg), zero_reg)>, + Requires<[IsARM, NoV4T]>, Sched<[WriteBr]>; + +// Large immediate handling. + +// 32-bit immediate using two piece mod_imms or movw + movt. +// This is a single pseudo instruction, the benefit is that it can be remat'd +// as a single unit instead of having to handle reg inputs. +// FIXME: Remove this when we can do generalized remat. +let isReMaterializable = 1, isMoveImm = 1 in +def MOVi32imm : PseudoInst<(outs GPR:$dst), (ins i32imm:$src), IIC_iMOVix2, + [(set GPR:$dst, (arm_i32imm:$src))]>, + Requires<[IsARM]>; + +def LDRLIT_ga_abs : PseudoInst<(outs GPR:$dst), (ins i32imm:$src), IIC_iLoad_i, + [(set GPR:$dst, (ARMWrapper tglobaladdr:$src))]>, + Requires<[IsARM, DontUseMovt]>; + +// Pseudo instruction that combines movw + movt + add pc (if PIC). +// It also makes it possible to rematerialize the instructions. +// FIXME: Remove this when we can do generalized remat and when machine licm +// can properly the instructions. +let isReMaterializable = 1 in { +def MOV_ga_pcrel : PseudoInst<(outs GPR:$dst), (ins i32imm:$addr), + IIC_iMOVix2addpc, + [(set GPR:$dst, (ARMWrapperPIC tglobaladdr:$addr))]>, + Requires<[IsARM, UseMovtInPic]>; + +def LDRLIT_ga_pcrel : PseudoInst<(outs GPR:$dst), (ins i32imm:$addr), + IIC_iLoadiALU, + [(set GPR:$dst, + (ARMWrapperPIC tglobaladdr:$addr))]>, + Requires<[IsARM, DontUseMovtInPic]>; + +let AddedComplexity = 10 in +def LDRLIT_ga_pcrel_ldr : PseudoInst<(outs GPR:$dst), (ins i32imm:$addr), + NoItinerary, + [(set GPR:$dst, + (load (ARMWrapperPIC tglobaladdr:$addr)))]>, + Requires<[IsARM, DontUseMovtInPic]>; + +let AddedComplexity = 10 in +def MOV_ga_pcrel_ldr : PseudoInst<(outs GPR:$dst), (ins i32imm:$addr), + IIC_iMOVix2ld, + [(set GPR:$dst, (load (ARMWrapperPIC tglobaladdr:$addr)))]>, + Requires<[IsARM, UseMovtInPic]>; +} // isReMaterializable + +// The many different faces of TLS access. +def : ARMPat<(ARMWrapper tglobaltlsaddr :$dst), + (MOVi32imm tglobaltlsaddr :$dst)>, + Requires<[IsARM, UseMovt]>; + +def : Pat<(ARMWrapper tglobaltlsaddr:$src), + (LDRLIT_ga_abs tglobaltlsaddr:$src)>, + Requires<[IsARM, DontUseMovt]>; + +def : Pat<(ARMWrapperPIC tglobaltlsaddr:$addr), + (MOV_ga_pcrel tglobaltlsaddr:$addr)>, Requires<[IsARM, UseMovtInPic]>; + +def : Pat<(ARMWrapperPIC tglobaltlsaddr:$addr), + (LDRLIT_ga_pcrel tglobaltlsaddr:$addr)>, + Requires<[IsARM, DontUseMovtInPic]>; +let AddedComplexity = 10 in +def : Pat<(load (ARMWrapperPIC tglobaltlsaddr:$addr)), + (MOV_ga_pcrel_ldr tglobaltlsaddr:$addr)>, + Requires<[IsARM, UseMovtInPic]>; + + +// ConstantPool, GlobalAddress, and JumpTable +def : ARMPat<(ARMWrapper tconstpool :$dst), (LEApcrel tconstpool :$dst)>; +def : ARMPat<(ARMWrapper tglobaladdr :$dst), (MOVi32imm tglobaladdr :$dst)>, + Requires<[IsARM, UseMovt]>; +def : ARMPat<(ARMWrapper texternalsym :$dst), (MOVi32imm texternalsym :$dst)>, + Requires<[IsARM, UseMovt]>; +def : ARMPat<(ARMWrapperJT tjumptable:$dst), + (LEApcrelJT tjumptable:$dst)>; + +// TODO: add,sub,and, 3-instr forms? + +// Tail calls. These patterns also apply to Thumb mode. +def : Pat<(ARMtcret tcGPR:$dst), (TCRETURNri tcGPR:$dst)>; +def : Pat<(ARMtcret (i32 tglobaladdr:$dst)), (TCRETURNdi texternalsym:$dst)>; +def : Pat<(ARMtcret (i32 texternalsym:$dst)), (TCRETURNdi texternalsym:$dst)>; + +// Direct calls +def : ARMPat<(ARMcall texternalsym:$func), (BL texternalsym:$func)>; +def : ARMPat<(ARMcall_nolink texternalsym:$func), + (BMOVPCB_CALL texternalsym:$func)>; + +// zextload i1 -> zextload i8 +def : ARMPat<(zextloadi1 addrmode_imm12:$addr), (LDRBi12 addrmode_imm12:$addr)>; +def : ARMPat<(zextloadi1 ldst_so_reg:$addr), (LDRBrs ldst_so_reg:$addr)>; + +// extload -> zextload +def : ARMPat<(extloadi1 addrmode_imm12:$addr), (LDRBi12 addrmode_imm12:$addr)>; +def : ARMPat<(extloadi1 ldst_so_reg:$addr), (LDRBrs ldst_so_reg:$addr)>; +def : ARMPat<(extloadi8 addrmode_imm12:$addr), (LDRBi12 addrmode_imm12:$addr)>; +def : ARMPat<(extloadi8 ldst_so_reg:$addr), (LDRBrs ldst_so_reg:$addr)>; + +def : ARMPat<(extloadi16 addrmode3:$addr), (LDRH addrmode3:$addr)>; + +def : ARMPat<(extloadi8 addrmodepc:$addr), (PICLDRB addrmodepc:$addr)>; +def : ARMPat<(extloadi16 addrmodepc:$addr), (PICLDRH addrmodepc:$addr)>; + +// smul* and smla* +def : ARMV5TEPat<(mul sext_16_node:$a, sext_16_node:$b), + (SMULBB GPR:$a, GPR:$b)>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; +def : ARMV5TEPat<(mul sext_16_node:$a, (sra GPR:$b, (i32 16))), + (SMULBT GPR:$a, GPR:$b)>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; +def : ARMV5TEPat<(mul (sra GPR:$a, (i32 16)), sext_16_node:$b), + (SMULTB GPR:$a, GPR:$b)>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; +def : ARMV5MOPat<(add GPR:$acc, + (mul sext_16_node:$a, sext_16_node:$b)), + (SMLABB GPR:$a, GPR:$b, GPR:$acc)>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; +def : ARMV5MOPat<(add GPR:$acc, + (mul sext_16_node:$a, (sra GPR:$b, (i32 16)))), + (SMLABT GPR:$a, GPR:$b, GPR:$acc)>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; +def : ARMV5MOPat<(add GPR:$acc, + (mul (sra GPR:$a, (i32 16)), sext_16_node:$b)), + (SMLATB GPR:$a, GPR:$b, GPR:$acc)>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]>; + +def : ARMV5TEPat<(int_arm_smulbb GPR:$a, GPR:$b), + (SMULBB GPR:$a, GPR:$b)>; +def : ARMV5TEPat<(int_arm_smulbt GPR:$a, GPR:$b), + (SMULBT GPR:$a, GPR:$b)>; +def : ARMV5TEPat<(int_arm_smultb GPR:$a, GPR:$b), + (SMULTB GPR:$a, GPR:$b)>; +def : ARMV5TEPat<(int_arm_smultt GPR:$a, GPR:$b), + (SMULTT GPR:$a, GPR:$b)>; +def : ARMV5TEPat<(int_arm_smulwb GPR:$a, GPR:$b), + (SMULWB GPR:$a, GPR:$b)>; +def : ARMV5TEPat<(int_arm_smulwt GPR:$a, GPR:$b), + (SMULWT GPR:$a, GPR:$b)>; + +def : ARMV5TEPat<(int_arm_smlabb GPR:$a, GPR:$b, GPR:$acc), + (SMLABB GPR:$a, GPR:$b, GPR:$acc)>; +def : ARMV5TEPat<(int_arm_smlabt GPR:$a, GPR:$b, GPR:$acc), + (SMLABT GPR:$a, GPR:$b, GPR:$acc)>; +def : ARMV5TEPat<(int_arm_smlatb GPR:$a, GPR:$b, GPR:$acc), + (SMLATB GPR:$a, GPR:$b, GPR:$acc)>; +def : ARMV5TEPat<(int_arm_smlatt GPR:$a, GPR:$b, GPR:$acc), + (SMLATT GPR:$a, GPR:$b, GPR:$acc)>; +def : ARMV5TEPat<(int_arm_smlawb GPR:$a, GPR:$b, GPR:$acc), + (SMLAWB GPR:$a, GPR:$b, GPR:$acc)>; +def : ARMV5TEPat<(int_arm_smlawt GPR:$a, GPR:$b, GPR:$acc), + (SMLAWT GPR:$a, GPR:$b, GPR:$acc)>; + +// Pre-v7 uses MCR for synchronization barriers. +def : ARMPat<(ARMMemBarrierMCR GPR:$zero), (MCR 15, 0, GPR:$zero, 7, 10, 5)>, + Requires<[IsARM, HasV6]>; + +// SXT/UXT with no rotate +let AddedComplexity = 16 in { +def : ARMV6Pat<(and GPR:$Src, 0x000000FF), (UXTB GPR:$Src, 0)>; +def : ARMV6Pat<(and GPR:$Src, 0x0000FFFF), (UXTH GPR:$Src, 0)>; +def : ARMV6Pat<(and GPR:$Src, 0x00FF00FF), (UXTB16 GPR:$Src, 0)>; +def : ARMV6Pat<(add GPR:$Rn, (and GPR:$Rm, 0x00FF)), + (UXTAB GPR:$Rn, GPR:$Rm, 0)>; +def : ARMV6Pat<(add GPR:$Rn, (and GPR:$Rm, 0xFFFF)), + (UXTAH GPR:$Rn, GPR:$Rm, 0)>; +} + +def : ARMV6Pat<(sext_inreg GPR:$Src, i8), (SXTB GPR:$Src, 0)>; +def : ARMV6Pat<(sext_inreg GPR:$Src, i16), (SXTH GPR:$Src, 0)>; + +def : ARMV6Pat<(add GPR:$Rn, (sext_inreg GPRnopc:$Rm, i8)), + (SXTAB GPR:$Rn, GPRnopc:$Rm, 0)>; +def : ARMV6Pat<(add GPR:$Rn, (sext_inreg GPRnopc:$Rm, i16)), + (SXTAH GPR:$Rn, GPRnopc:$Rm, 0)>; + +// Atomic load/store patterns +def : ARMPat<(atomic_load_8 ldst_so_reg:$src), + (LDRBrs ldst_so_reg:$src)>; +def : ARMPat<(atomic_load_8 addrmode_imm12:$src), + (LDRBi12 addrmode_imm12:$src)>; +def : ARMPat<(atomic_load_16 addrmode3:$src), + (LDRH addrmode3:$src)>; +def : ARMPat<(atomic_load_32 ldst_so_reg:$src), + (LDRrs ldst_so_reg:$src)>; +def : ARMPat<(atomic_load_32 addrmode_imm12:$src), + (LDRi12 addrmode_imm12:$src)>; +def : ARMPat<(atomic_store_8 ldst_so_reg:$ptr, GPR:$val), + (STRBrs GPR:$val, ldst_so_reg:$ptr)>; +def : ARMPat<(atomic_store_8 addrmode_imm12:$ptr, GPR:$val), + (STRBi12 GPR:$val, addrmode_imm12:$ptr)>; +def : ARMPat<(atomic_store_16 addrmode3:$ptr, GPR:$val), + (STRH GPR:$val, addrmode3:$ptr)>; +def : ARMPat<(atomic_store_32 ldst_so_reg:$ptr, GPR:$val), + (STRrs GPR:$val, ldst_so_reg:$ptr)>; +def : ARMPat<(atomic_store_32 addrmode_imm12:$ptr, GPR:$val), + (STRi12 GPR:$val, addrmode_imm12:$ptr)>; + + +//===----------------------------------------------------------------------===// +// Thumb Support +// + +include "ARMInstrThumb.td" + +//===----------------------------------------------------------------------===// +// Thumb2 Support +// + +include "ARMInstrThumb2.td" + +//===----------------------------------------------------------------------===// +// Floating Point Support +// + +include "ARMInstrVFP.td" + +//===----------------------------------------------------------------------===// +// Advanced SIMD (NEON) Support +// + +include "ARMInstrNEON.td" + +//===----------------------------------------------------------------------===// +// Assembler aliases +// + +// Memory barriers +def : InstAlias<"dmb", (DMB 0xf), 0>, Requires<[IsARM, HasDB]>; +def : InstAlias<"dsb", (DSB 0xf), 0>, Requires<[IsARM, HasDB]>; +def : InstAlias<"isb", (ISB 0xf), 0>, Requires<[IsARM, HasDB]>; +// Armv8-R 'Data Full Barrier' +def : InstAlias<"dfb", (DSB 0xc), 1>, Requires<[IsARM, HasDFB]>; + +// System instructions +def : MnemonicAlias<"swi", "svc">; + +// Load / Store Multiple +def : MnemonicAlias<"ldmfd", "ldm">; +def : MnemonicAlias<"ldmia", "ldm">; +def : MnemonicAlias<"ldmea", "ldmdb">; +def : MnemonicAlias<"stmfd", "stmdb">; +def : MnemonicAlias<"stmia", "stm">; +def : MnemonicAlias<"stmea", "stm">; + +// PKHBT/PKHTB with default shift amount. PKHTB is equivalent to PKHBT with the +// input operands swapped when the shift amount is zero (i.e., unspecified). +def : InstAlias<"pkhbt${p} $Rd, $Rn, $Rm", + (PKHBT GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, 0, pred:$p), 0>, + Requires<[IsARM, HasV6]>; +def : InstAlias<"pkhtb${p} $Rd, $Rn, $Rm", + (PKHBT GPRnopc:$Rd, GPRnopc:$Rm, GPRnopc:$Rn, 0, pred:$p), 0>, + Requires<[IsARM, HasV6]>; + +// PUSH/POP aliases for STM/LDM +def : ARMInstAlias<"push${p} $regs", (STMDB_UPD SP, pred:$p, reglist:$regs)>; +def : ARMInstAlias<"pop${p} $regs", (LDMIA_UPD SP, pred:$p, reglist:$regs)>; + +// SSAT/USAT optional shift operand. +def : ARMInstAlias<"ssat${p} $Rd, $sat_imm, $Rn", + (SSAT GPRnopc:$Rd, imm1_32:$sat_imm, GPRnopc:$Rn, 0, pred:$p)>; +def : ARMInstAlias<"usat${p} $Rd, $sat_imm, $Rn", + (USAT GPRnopc:$Rd, imm0_31:$sat_imm, GPRnopc:$Rn, 0, pred:$p)>; + + +// Extend instruction optional rotate operand. +def : ARMInstAlias<"sxtab${p} $Rd, $Rn, $Rm", + (SXTAB GPRnopc:$Rd, GPR:$Rn, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"sxtah${p} $Rd, $Rn, $Rm", + (SXTAH GPRnopc:$Rd, GPR:$Rn, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"sxtab16${p} $Rd, $Rn, $Rm", + (SXTAB16 GPRnopc:$Rd, GPR:$Rn, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"sxtb${p} $Rd, $Rm", + (SXTB GPRnopc:$Rd, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"sxtb16${p} $Rd, $Rm", + (SXTB16 GPRnopc:$Rd, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"sxth${p} $Rd, $Rm", + (SXTH GPRnopc:$Rd, GPRnopc:$Rm, 0, pred:$p)>; + +def : ARMInstAlias<"uxtab${p} $Rd, $Rn, $Rm", + (UXTAB GPRnopc:$Rd, GPR:$Rn, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"uxtah${p} $Rd, $Rn, $Rm", + (UXTAH GPRnopc:$Rd, GPR:$Rn, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"uxtab16${p} $Rd, $Rn, $Rm", + (UXTAB16 GPRnopc:$Rd, GPR:$Rn, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"uxtb${p} $Rd, $Rm", + (UXTB GPRnopc:$Rd, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"uxtb16${p} $Rd, $Rm", + (UXTB16 GPRnopc:$Rd, GPRnopc:$Rm, 0, pred:$p)>; +def : ARMInstAlias<"uxth${p} $Rd, $Rm", + (UXTH GPRnopc:$Rd, GPRnopc:$Rm, 0, pred:$p)>; + + +// RFE aliases +def : MnemonicAlias<"rfefa", "rfeda">; +def : MnemonicAlias<"rfeea", "rfedb">; +def : MnemonicAlias<"rfefd", "rfeia">; +def : MnemonicAlias<"rfeed", "rfeib">; +def : MnemonicAlias<"rfe", "rfeia">; + +// SRS aliases +def : MnemonicAlias<"srsfa", "srsib">; +def : MnemonicAlias<"srsea", "srsia">; +def : MnemonicAlias<"srsfd", "srsdb">; +def : MnemonicAlias<"srsed", "srsda">; +def : MnemonicAlias<"srs", "srsia">; + +// QSAX == QSUBADDX +def : MnemonicAlias<"qsubaddx", "qsax">; +// SASX == SADDSUBX +def : MnemonicAlias<"saddsubx", "sasx">; +// SHASX == SHADDSUBX +def : MnemonicAlias<"shaddsubx", "shasx">; +// SHSAX == SHSUBADDX +def : MnemonicAlias<"shsubaddx", "shsax">; +// SSAX == SSUBADDX +def : MnemonicAlias<"ssubaddx", "ssax">; +// UASX == UADDSUBX +def : MnemonicAlias<"uaddsubx", "uasx">; +// UHASX == UHADDSUBX +def : MnemonicAlias<"uhaddsubx", "uhasx">; +// UHSAX == UHSUBADDX +def : MnemonicAlias<"uhsubaddx", "uhsax">; +// UQASX == UQADDSUBX +def : MnemonicAlias<"uqaddsubx", "uqasx">; +// UQSAX == UQSUBADDX +def : MnemonicAlias<"uqsubaddx", "uqsax">; +// USAX == USUBADDX +def : MnemonicAlias<"usubaddx", "usax">; + +// "mov Rd, mod_imm_not" can be handled via "mvn" in assembly, just like +// for isel. +def : ARMInstSubst<"mov${s}${p} $Rd, $imm", + (MVNi rGPR:$Rd, mod_imm_not:$imm, pred:$p, cc_out:$s)>; +def : ARMInstSubst<"mvn${s}${p} $Rd, $imm", + (MOVi rGPR:$Rd, mod_imm_not:$imm, pred:$p, cc_out:$s)>; +// Same for AND <--> BIC +def : ARMInstSubst<"bic${s}${p} $Rd, $Rn, $imm", + (ANDri GPR:$Rd, GPR:$Rn, mod_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : ARMInstSubst<"bic${s}${p} $Rdn, $imm", + (ANDri GPR:$Rdn, GPR:$Rdn, mod_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : ARMInstSubst<"and${s}${p} $Rd, $Rn, $imm", + (BICri GPR:$Rd, GPR:$Rn, mod_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : ARMInstSubst<"and${s}${p} $Rdn, $imm", + (BICri GPR:$Rdn, GPR:$Rdn, mod_imm_not:$imm, + pred:$p, cc_out:$s)>; + +// Likewise, "add Rd, mod_imm_neg" -> sub +def : ARMInstSubst<"add${s}${p} $Rd, $Rn, $imm", + (SUBri GPR:$Rd, GPR:$Rn, mod_imm_neg:$imm, pred:$p, cc_out:$s)>; +def : ARMInstSubst<"add${s}${p} $Rd, $imm", + (SUBri GPR:$Rd, GPR:$Rd, mod_imm_neg:$imm, pred:$p, cc_out:$s)>; +// Likewise, "sub Rd, mod_imm_neg" -> add +def : ARMInstSubst<"sub${s}${p} $Rd, $Rn, $imm", + (ADDri GPR:$Rd, GPR:$Rn, mod_imm_neg:$imm, pred:$p, cc_out:$s)>; +def : ARMInstSubst<"sub${s}${p} $Rd, $imm", + (ADDri GPR:$Rd, GPR:$Rd, mod_imm_neg:$imm, pred:$p, cc_out:$s)>; + + +def : ARMInstSubst<"adc${s}${p} $Rd, $Rn, $imm", + (SBCri GPR:$Rd, GPR:$Rn, mod_imm_not:$imm, pred:$p, cc_out:$s)>; +def : ARMInstSubst<"adc${s}${p} $Rdn, $imm", + (SBCri GPR:$Rdn, GPR:$Rdn, mod_imm_not:$imm, pred:$p, cc_out:$s)>; +def : ARMInstSubst<"sbc${s}${p} $Rd, $Rn, $imm", + (ADCri GPR:$Rd, GPR:$Rn, mod_imm_not:$imm, pred:$p, cc_out:$s)>; +def : ARMInstSubst<"sbc${s}${p} $Rdn, $imm", + (ADCri GPR:$Rdn, GPR:$Rdn, mod_imm_not:$imm, pred:$p, cc_out:$s)>; + +// Same for CMP <--> CMN via mod_imm_neg +def : ARMInstSubst<"cmp${p} $Rd, $imm", + (CMNri rGPR:$Rd, mod_imm_neg:$imm, pred:$p)>; +def : ARMInstSubst<"cmn${p} $Rd, $imm", + (CMPri rGPR:$Rd, mod_imm_neg:$imm, pred:$p)>; + +// The shifter forms of the MOV instruction are aliased to the ASR, LSL, +// LSR, ROR, and RRX instructions. +// FIXME: We need C++ parser hooks to map the alias to the MOV +// encoding. It seems we should be able to do that sort of thing +// in tblgen, but it could get ugly. +let TwoOperandAliasConstraint = "$Rm = $Rd" in { +def ASRi : ARMAsmPseudo<"asr${s}${p} $Rd, $Rm, $imm", + (ins GPR:$Rd, GPR:$Rm, imm0_32:$imm, pred:$p, + cc_out:$s)>; +def LSRi : ARMAsmPseudo<"lsr${s}${p} $Rd, $Rm, $imm", + (ins GPR:$Rd, GPR:$Rm, imm0_32:$imm, pred:$p, + cc_out:$s)>; +def LSLi : ARMAsmPseudo<"lsl${s}${p} $Rd, $Rm, $imm", + (ins GPR:$Rd, GPR:$Rm, imm0_31:$imm, pred:$p, + cc_out:$s)>; +def RORi : ARMAsmPseudo<"ror${s}${p} $Rd, $Rm, $imm", + (ins GPR:$Rd, GPR:$Rm, imm0_31:$imm, pred:$p, + cc_out:$s)>; +} +def RRXi : ARMAsmPseudo<"rrx${s}${p} $Rd, $Rm", + (ins GPR:$Rd, GPR:$Rm, pred:$p, cc_out:$s)>; +let TwoOperandAliasConstraint = "$Rn = $Rd" in { +def ASRr : ARMAsmPseudo<"asr${s}${p} $Rd, $Rn, $Rm", + (ins GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, pred:$p, + cc_out:$s)>; +def LSRr : ARMAsmPseudo<"lsr${s}${p} $Rd, $Rn, $Rm", + (ins GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, pred:$p, + cc_out:$s)>; +def LSLr : ARMAsmPseudo<"lsl${s}${p} $Rd, $Rn, $Rm", + (ins GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, pred:$p, + cc_out:$s)>; +def RORr : ARMAsmPseudo<"ror${s}${p} $Rd, $Rn, $Rm", + (ins GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, pred:$p, + cc_out:$s)>; +} + +// "neg" is and alias for "rsb rd, rn, #0" +def : ARMInstAlias<"neg${s}${p} $Rd, $Rm", + (RSBri GPR:$Rd, GPR:$Rm, 0, pred:$p, cc_out:$s)>; + +// Pre-v6, 'mov r0, r0' was used as a NOP encoding. +def : InstAlias<"nop${p}", (MOVr R0, R0, pred:$p, zero_reg)>, + Requires<[IsARM, NoV6]>; + +// MUL/UMLAL/SMLAL/UMULL/SMULL are available on all arches, but +// the instruction definitions need difference constraints pre-v6. +// Use these aliases for the assembly parsing on pre-v6. +def : InstAlias<"mul${s}${p} $Rd, $Rn, $Rm", + (MUL GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, pred:$p, cc_out:$s), 0>, + Requires<[IsARM, NoV6]>; +def : InstAlias<"mla${s}${p} $Rd, $Rn, $Rm, $Ra", + (MLA GPRnopc:$Rd, GPRnopc:$Rn, GPRnopc:$Rm, GPRnopc:$Ra, + pred:$p, cc_out:$s), 0>, + Requires<[IsARM, NoV6]>; +def : InstAlias<"smlal${s}${p} $RdLo, $RdHi, $Rn, $Rm", + (SMLAL GPR:$RdLo, GPR:$RdHi, GPR:$Rn, GPR:$Rm, pred:$p, cc_out:$s), 0>, + Requires<[IsARM, NoV6]>; +def : InstAlias<"umlal${s}${p} $RdLo, $RdHi, $Rn, $Rm", + (UMLAL GPR:$RdLo, GPR:$RdHi, GPR:$Rn, GPR:$Rm, pred:$p, cc_out:$s), 0>, + Requires<[IsARM, NoV6]>; +def : InstAlias<"smull${s}${p} $RdLo, $RdHi, $Rn, $Rm", + (SMULL GPR:$RdLo, GPR:$RdHi, GPR:$Rn, GPR:$Rm, pred:$p, cc_out:$s), 0>, + Requires<[IsARM, NoV6]>; +def : InstAlias<"umull${s}${p} $RdLo, $RdHi, $Rn, $Rm", + (UMULL GPR:$RdLo, GPR:$RdHi, GPR:$Rn, GPR:$Rm, pred:$p, cc_out:$s), 0>, + Requires<[IsARM, NoV6]>; + +// 'it' blocks in ARM mode just validate the predicates. The IT itself +// is discarded. +def ITasm : ARMAsmPseudo<"it$mask $cc", (ins it_pred:$cc, it_mask:$mask)>, + ComplexDeprecationPredicate<"IT">; + +let mayLoad = 1, mayStore =1, hasSideEffects = 1 in +def SPACE : PseudoInst<(outs GPR:$Rd), (ins i32imm:$size, GPR:$Rn), + NoItinerary, + [(set GPR:$Rd, (int_arm_space imm:$size, GPR:$Rn))]>; + +//===---------------------------------- +// Atomic cmpxchg for -O0 +//===---------------------------------- + +// The fast register allocator used during -O0 inserts spills to cover any VRegs +// live across basic block boundaries. When this happens between an LDXR and an +// STXR it can clear the exclusive monitor, causing all cmpxchg attempts to +// fail. + +// Unfortunately, this means we have to have an alternative (expanded +// post-regalloc) path for -O0 compilations. Fortunately this path can be +// significantly more naive than the standard expansion: we conservatively +// assume seq_cst, strong cmpxchg and omit clrex on failure. + +let Constraints = "@earlyclobber $Rd,@earlyclobber $temp", + mayLoad = 1, mayStore = 1 in { +def CMP_SWAP_8 : PseudoInst<(outs GPR:$Rd, GPR:$temp), + (ins GPR:$addr, GPR:$desired, GPR:$new), + NoItinerary, []>, Sched<[]>; + +def CMP_SWAP_16 : PseudoInst<(outs GPR:$Rd, GPR:$temp), + (ins GPR:$addr, GPR:$desired, GPR:$new), + NoItinerary, []>, Sched<[]>; + +def CMP_SWAP_32 : PseudoInst<(outs GPR:$Rd, GPR:$temp), + (ins GPR:$addr, GPR:$desired, GPR:$new), + NoItinerary, []>, Sched<[]>; + +def CMP_SWAP_64 : PseudoInst<(outs GPRPair:$Rd, GPR:$temp), + (ins GPR:$addr, GPRPair:$desired, GPRPair:$new), + NoItinerary, []>, Sched<[]>; +} + +def CompilerBarrier : PseudoInst<(outs), (ins i32imm:$ordering), NoItinerary, + [(atomic_fence imm:$ordering, 0)]> { + let hasSideEffects = 1; + let Size = 0; + let AsmString = "@ COMPILER BARRIER"; +} diff --git a/capstone/suite/synctools/tablegen/ARM/ARMInstrNEON.td b/capstone/suite/synctools/tablegen/ARM/ARMInstrNEON.td new file mode 100644 index 000000000..4525eec8d --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMInstrNEON.td @@ -0,0 +1,8545 @@ +//===-- ARMInstrNEON.td - NEON support for ARM -------------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file describes the ARM NEON instruction set. +// +//===----------------------------------------------------------------------===// + + +//===----------------------------------------------------------------------===// +// NEON-specific Operands. +//===----------------------------------------------------------------------===// +def nModImm : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; +} + +def nImmSplatI8AsmOperand : AsmOperandClass { let Name = "NEONi8splat"; } +def nImmSplatI8 : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; + let ParserMatchClass = nImmSplatI8AsmOperand; +} +def nImmSplatI16AsmOperand : AsmOperandClass { let Name = "NEONi16splat"; } +def nImmSplatI16 : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; + let ParserMatchClass = nImmSplatI16AsmOperand; +} +def nImmSplatI32AsmOperand : AsmOperandClass { let Name = "NEONi32splat"; } +def nImmSplatI32 : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; + let ParserMatchClass = nImmSplatI32AsmOperand; +} +def nImmSplatNotI16AsmOperand : AsmOperandClass { let Name = "NEONi16splatNot"; } +def nImmSplatNotI16 : Operand<i32> { + let ParserMatchClass = nImmSplatNotI16AsmOperand; +} +def nImmSplatNotI32AsmOperand : AsmOperandClass { let Name = "NEONi32splatNot"; } +def nImmSplatNotI32 : Operand<i32> { + let ParserMatchClass = nImmSplatNotI32AsmOperand; +} +def nImmVMOVI32AsmOperand : AsmOperandClass { let Name = "NEONi32vmov"; } +def nImmVMOVI32 : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; + let ParserMatchClass = nImmVMOVI32AsmOperand; +} + +class nImmVMOVIAsmOperandReplicate<ValueType From, ValueType To> + : AsmOperandClass { + let Name = "NEONi" # To.Size # "vmovi" # From.Size # "Replicate"; + let PredicateMethod = "isNEONmovReplicate<" # From.Size # ", " # To.Size # ">"; + let RenderMethod = "addNEONvmovi" # From.Size # "ReplicateOperands"; +} + +class nImmVINVIAsmOperandReplicate<ValueType From, ValueType To> + : AsmOperandClass { + let Name = "NEONi" # To.Size # "invi" # From.Size # "Replicate"; + let PredicateMethod = "isNEONinvReplicate<" # From.Size # ", " # To.Size # ">"; + let RenderMethod = "addNEONinvi" # From.Size # "ReplicateOperands"; +} + +class nImmVMOVIReplicate<ValueType From, ValueType To> : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; + let ParserMatchClass = nImmVMOVIAsmOperandReplicate<From, To>; +} + +class nImmVINVIReplicate<ValueType From, ValueType To> : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; + let ParserMatchClass = nImmVINVIAsmOperandReplicate<From, To>; +} + +def nImmVMOVI32NegAsmOperand : AsmOperandClass { let Name = "NEONi32vmovNeg"; } +def nImmVMOVI32Neg : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; + let ParserMatchClass = nImmVMOVI32NegAsmOperand; +} +def nImmVMOVF32 : Operand<i32> { + let PrintMethod = "printFPImmOperand"; + let ParserMatchClass = FPImmOperand; +} +def nImmSplatI64AsmOperand : AsmOperandClass { let Name = "NEONi64splat"; } +def nImmSplatI64 : Operand<i32> { + let PrintMethod = "printNEONModImmOperand"; + let ParserMatchClass = nImmSplatI64AsmOperand; +} + +def VectorIndex8Operand : AsmOperandClass { let Name = "VectorIndex8"; } +def VectorIndex16Operand : AsmOperandClass { let Name = "VectorIndex16"; } +def VectorIndex32Operand : AsmOperandClass { let Name = "VectorIndex32"; } +def VectorIndex64Operand : AsmOperandClass { let Name = "VectorIndex64"; } +def VectorIndex8 : Operand<i32>, ImmLeaf<i32, [{ + return ((uint64_t)Imm) < 8; +}]> { + let ParserMatchClass = VectorIndex8Operand; + let PrintMethod = "printVectorIndex"; + let MIOperandInfo = (ops i32imm); +} +def VectorIndex16 : Operand<i32>, ImmLeaf<i32, [{ + return ((uint64_t)Imm) < 4; +}]> { + let ParserMatchClass = VectorIndex16Operand; + let PrintMethod = "printVectorIndex"; + let MIOperandInfo = (ops i32imm); +} +def VectorIndex32 : Operand<i32>, ImmLeaf<i32, [{ + return ((uint64_t)Imm) < 2; +}]> { + let ParserMatchClass = VectorIndex32Operand; + let PrintMethod = "printVectorIndex"; + let MIOperandInfo = (ops i32imm); +} +def VectorIndex64 : Operand<i32>, ImmLeaf<i32, [{ + return ((uint64_t)Imm) < 1; +}]> { + let ParserMatchClass = VectorIndex64Operand; + let PrintMethod = "printVectorIndex"; + let MIOperandInfo = (ops i32imm); +} + +// Register list of one D register. +def VecListOneDAsmOperand : AsmOperandClass { + let Name = "VecListOneD"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListOneD : RegisterOperand<DPR, "printVectorListOne"> { + let ParserMatchClass = VecListOneDAsmOperand; +} +// Register list of two sequential D registers. +def VecListDPairAsmOperand : AsmOperandClass { + let Name = "VecListDPair"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListDPair : RegisterOperand<DPair, "printVectorListTwo"> { + let ParserMatchClass = VecListDPairAsmOperand; +} +// Register list of three sequential D registers. +def VecListThreeDAsmOperand : AsmOperandClass { + let Name = "VecListThreeD"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListThreeD : RegisterOperand<DPR, "printVectorListThree"> { + let ParserMatchClass = VecListThreeDAsmOperand; +} +// Register list of four sequential D registers. +def VecListFourDAsmOperand : AsmOperandClass { + let Name = "VecListFourD"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListFourD : RegisterOperand<DPR, "printVectorListFour"> { + let ParserMatchClass = VecListFourDAsmOperand; +} +// Register list of two D registers spaced by 2 (two sequential Q registers). +def VecListDPairSpacedAsmOperand : AsmOperandClass { + let Name = "VecListDPairSpaced"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListDPairSpaced : RegisterOperand<DPair, "printVectorListTwoSpaced"> { + let ParserMatchClass = VecListDPairSpacedAsmOperand; +} +// Register list of three D registers spaced by 2 (three Q registers). +def VecListThreeQAsmOperand : AsmOperandClass { + let Name = "VecListThreeQ"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListThreeQ : RegisterOperand<DPR, "printVectorListThreeSpaced"> { + let ParserMatchClass = VecListThreeQAsmOperand; +} +// Register list of three D registers spaced by 2 (three Q registers). +def VecListFourQAsmOperand : AsmOperandClass { + let Name = "VecListFourQ"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListFourQ : RegisterOperand<DPR, "printVectorListFourSpaced"> { + let ParserMatchClass = VecListFourQAsmOperand; +} + +// Register list of one D register, with "all lanes" subscripting. +def VecListOneDAllLanesAsmOperand : AsmOperandClass { + let Name = "VecListOneDAllLanes"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListOneDAllLanes : RegisterOperand<DPR, "printVectorListOneAllLanes"> { + let ParserMatchClass = VecListOneDAllLanesAsmOperand; +} +// Register list of two D registers, with "all lanes" subscripting. +def VecListDPairAllLanesAsmOperand : AsmOperandClass { + let Name = "VecListDPairAllLanes"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListDPairAllLanes : RegisterOperand<DPair, + "printVectorListTwoAllLanes"> { + let ParserMatchClass = VecListDPairAllLanesAsmOperand; +} +// Register list of two D registers spaced by 2 (two sequential Q registers). +def VecListDPairSpacedAllLanesAsmOperand : AsmOperandClass { + let Name = "VecListDPairSpacedAllLanes"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListDPairSpacedAllLanes : RegisterOperand<DPairSpc, + "printVectorListTwoSpacedAllLanes"> { + let ParserMatchClass = VecListDPairSpacedAllLanesAsmOperand; +} +// Register list of three D registers, with "all lanes" subscripting. +def VecListThreeDAllLanesAsmOperand : AsmOperandClass { + let Name = "VecListThreeDAllLanes"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListThreeDAllLanes : RegisterOperand<DPR, + "printVectorListThreeAllLanes"> { + let ParserMatchClass = VecListThreeDAllLanesAsmOperand; +} +// Register list of three D registers spaced by 2 (three sequential Q regs). +def VecListThreeQAllLanesAsmOperand : AsmOperandClass { + let Name = "VecListThreeQAllLanes"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListThreeQAllLanes : RegisterOperand<DPR, + "printVectorListThreeSpacedAllLanes"> { + let ParserMatchClass = VecListThreeQAllLanesAsmOperand; +} +// Register list of four D registers, with "all lanes" subscripting. +def VecListFourDAllLanesAsmOperand : AsmOperandClass { + let Name = "VecListFourDAllLanes"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListFourDAllLanes : RegisterOperand<DPR, "printVectorListFourAllLanes"> { + let ParserMatchClass = VecListFourDAllLanesAsmOperand; +} +// Register list of four D registers spaced by 2 (four sequential Q regs). +def VecListFourQAllLanesAsmOperand : AsmOperandClass { + let Name = "VecListFourQAllLanes"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListOperands"; +} +def VecListFourQAllLanes : RegisterOperand<DPR, + "printVectorListFourSpacedAllLanes"> { + let ParserMatchClass = VecListFourQAllLanesAsmOperand; +} + + +// Register list of one D register, with byte lane subscripting. +def VecListOneDByteIndexAsmOperand : AsmOperandClass { + let Name = "VecListOneDByteIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListOneDByteIndexed : Operand<i32> { + let ParserMatchClass = VecListOneDByteIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with half-word lane subscripting. +def VecListOneDHWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListOneDHWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListOneDHWordIndexed : Operand<i32> { + let ParserMatchClass = VecListOneDHWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with word lane subscripting. +def VecListOneDWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListOneDWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListOneDWordIndexed : Operand<i32> { + let ParserMatchClass = VecListOneDWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} + +// Register list of two D registers with byte lane subscripting. +def VecListTwoDByteIndexAsmOperand : AsmOperandClass { + let Name = "VecListTwoDByteIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListTwoDByteIndexed : Operand<i32> { + let ParserMatchClass = VecListTwoDByteIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with half-word lane subscripting. +def VecListTwoDHWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListTwoDHWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListTwoDHWordIndexed : Operand<i32> { + let ParserMatchClass = VecListTwoDHWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with word lane subscripting. +def VecListTwoDWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListTwoDWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListTwoDWordIndexed : Operand<i32> { + let ParserMatchClass = VecListTwoDWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// Register list of two Q registers with half-word lane subscripting. +def VecListTwoQHWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListTwoQHWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListTwoQHWordIndexed : Operand<i32> { + let ParserMatchClass = VecListTwoQHWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with word lane subscripting. +def VecListTwoQWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListTwoQWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListTwoQWordIndexed : Operand<i32> { + let ParserMatchClass = VecListTwoQWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} + + +// Register list of three D registers with byte lane subscripting. +def VecListThreeDByteIndexAsmOperand : AsmOperandClass { + let Name = "VecListThreeDByteIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListThreeDByteIndexed : Operand<i32> { + let ParserMatchClass = VecListThreeDByteIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with half-word lane subscripting. +def VecListThreeDHWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListThreeDHWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListThreeDHWordIndexed : Operand<i32> { + let ParserMatchClass = VecListThreeDHWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with word lane subscripting. +def VecListThreeDWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListThreeDWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListThreeDWordIndexed : Operand<i32> { + let ParserMatchClass = VecListThreeDWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// Register list of three Q registers with half-word lane subscripting. +def VecListThreeQHWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListThreeQHWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListThreeQHWordIndexed : Operand<i32> { + let ParserMatchClass = VecListThreeQHWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with word lane subscripting. +def VecListThreeQWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListThreeQWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListThreeQWordIndexed : Operand<i32> { + let ParserMatchClass = VecListThreeQWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} + +// Register list of four D registers with byte lane subscripting. +def VecListFourDByteIndexAsmOperand : AsmOperandClass { + let Name = "VecListFourDByteIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListFourDByteIndexed : Operand<i32> { + let ParserMatchClass = VecListFourDByteIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with half-word lane subscripting. +def VecListFourDHWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListFourDHWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListFourDHWordIndexed : Operand<i32> { + let ParserMatchClass = VecListFourDHWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with word lane subscripting. +def VecListFourDWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListFourDWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListFourDWordIndexed : Operand<i32> { + let ParserMatchClass = VecListFourDWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// Register list of four Q registers with half-word lane subscripting. +def VecListFourQHWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListFourQHWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListFourQHWordIndexed : Operand<i32> { + let ParserMatchClass = VecListFourQHWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} +// ...with word lane subscripting. +def VecListFourQWordIndexAsmOperand : AsmOperandClass { + let Name = "VecListFourQWordIndexed"; + let ParserMethod = "parseVectorList"; + let RenderMethod = "addVecListIndexedOperands"; +} +def VecListFourQWordIndexed : Operand<i32> { + let ParserMatchClass = VecListFourQWordIndexAsmOperand; + let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx); +} + +def dword_alignedload : PatFrag<(ops node:$ptr), (load node:$ptr), [{ + return cast<LoadSDNode>(N)->getAlignment() >= 8; +}]>; +def dword_alignedstore : PatFrag<(ops node:$val, node:$ptr), + (store node:$val, node:$ptr), [{ + return cast<StoreSDNode>(N)->getAlignment() >= 8; +}]>; +def word_alignedload : PatFrag<(ops node:$ptr), (load node:$ptr), [{ + return cast<LoadSDNode>(N)->getAlignment() == 4; +}]>; +def word_alignedstore : PatFrag<(ops node:$val, node:$ptr), + (store node:$val, node:$ptr), [{ + return cast<StoreSDNode>(N)->getAlignment() == 4; +}]>; +def hword_alignedload : PatFrag<(ops node:$ptr), (load node:$ptr), [{ + return cast<LoadSDNode>(N)->getAlignment() == 2; +}]>; +def hword_alignedstore : PatFrag<(ops node:$val, node:$ptr), + (store node:$val, node:$ptr), [{ + return cast<StoreSDNode>(N)->getAlignment() == 2; +}]>; +def byte_alignedload : PatFrag<(ops node:$ptr), (load node:$ptr), [{ + return cast<LoadSDNode>(N)->getAlignment() == 1; +}]>; +def byte_alignedstore : PatFrag<(ops node:$val, node:$ptr), + (store node:$val, node:$ptr), [{ + return cast<StoreSDNode>(N)->getAlignment() == 1; +}]>; +def non_word_alignedload : PatFrag<(ops node:$ptr), (load node:$ptr), [{ + return cast<LoadSDNode>(N)->getAlignment() < 4; +}]>; +def non_word_alignedstore : PatFrag<(ops node:$val, node:$ptr), + (store node:$val, node:$ptr), [{ + return cast<StoreSDNode>(N)->getAlignment() < 4; +}]>; + +//===----------------------------------------------------------------------===// +// NEON-specific DAG Nodes. +//===----------------------------------------------------------------------===// + +def SDTARMVCMP : SDTypeProfile<1, 2, [SDTCisInt<0>, SDTCisSameAs<1, 2>]>; +def SDTARMVCMPZ : SDTypeProfile<1, 1, []>; + +def NEONvceq : SDNode<"ARMISD::VCEQ", SDTARMVCMP>; +def NEONvceqz : SDNode<"ARMISD::VCEQZ", SDTARMVCMPZ>; +def NEONvcge : SDNode<"ARMISD::VCGE", SDTARMVCMP>; +def NEONvcgez : SDNode<"ARMISD::VCGEZ", SDTARMVCMPZ>; +def NEONvclez : SDNode<"ARMISD::VCLEZ", SDTARMVCMPZ>; +def NEONvcgeu : SDNode<"ARMISD::VCGEU", SDTARMVCMP>; +def NEONvcgt : SDNode<"ARMISD::VCGT", SDTARMVCMP>; +def NEONvcgtz : SDNode<"ARMISD::VCGTZ", SDTARMVCMPZ>; +def NEONvcltz : SDNode<"ARMISD::VCLTZ", SDTARMVCMPZ>; +def NEONvcgtu : SDNode<"ARMISD::VCGTU", SDTARMVCMP>; +def NEONvtst : SDNode<"ARMISD::VTST", SDTARMVCMP>; + +// Types for vector shift by immediates. The "SHX" version is for long and +// narrow operations where the source and destination vectors have different +// types. The "SHINS" version is for shift and insert operations. +def SDTARMVSH : SDTypeProfile<1, 2, [SDTCisInt<0>, SDTCisSameAs<0, 1>, + SDTCisVT<2, i32>]>; +def SDTARMVSHX : SDTypeProfile<1, 2, [SDTCisInt<0>, SDTCisInt<1>, + SDTCisVT<2, i32>]>; +def SDTARMVSHINS : SDTypeProfile<1, 3, [SDTCisInt<0>, SDTCisSameAs<0, 1>, + SDTCisSameAs<0, 2>, SDTCisVT<3, i32>]>; + +def NEONvshl : SDNode<"ARMISD::VSHL", SDTARMVSH>; +def NEONvshrs : SDNode<"ARMISD::VSHRs", SDTARMVSH>; +def NEONvshru : SDNode<"ARMISD::VSHRu", SDTARMVSH>; +def NEONvshrn : SDNode<"ARMISD::VSHRN", SDTARMVSHX>; + +def NEONvrshrs : SDNode<"ARMISD::VRSHRs", SDTARMVSH>; +def NEONvrshru : SDNode<"ARMISD::VRSHRu", SDTARMVSH>; +def NEONvrshrn : SDNode<"ARMISD::VRSHRN", SDTARMVSHX>; + +def NEONvqshls : SDNode<"ARMISD::VQSHLs", SDTARMVSH>; +def NEONvqshlu : SDNode<"ARMISD::VQSHLu", SDTARMVSH>; +def NEONvqshlsu : SDNode<"ARMISD::VQSHLsu", SDTARMVSH>; +def NEONvqshrns : SDNode<"ARMISD::VQSHRNs", SDTARMVSHX>; +def NEONvqshrnu : SDNode<"ARMISD::VQSHRNu", SDTARMVSHX>; +def NEONvqshrnsu : SDNode<"ARMISD::VQSHRNsu", SDTARMVSHX>; + +def NEONvqrshrns : SDNode<"ARMISD::VQRSHRNs", SDTARMVSHX>; +def NEONvqrshrnu : SDNode<"ARMISD::VQRSHRNu", SDTARMVSHX>; +def NEONvqrshrnsu : SDNode<"ARMISD::VQRSHRNsu", SDTARMVSHX>; + +def NEONvsli : SDNode<"ARMISD::VSLI", SDTARMVSHINS>; +def NEONvsri : SDNode<"ARMISD::VSRI", SDTARMVSHINS>; + +def SDTARMVGETLN : SDTypeProfile<1, 2, [SDTCisVT<0, i32>, SDTCisInt<1>, + SDTCisVT<2, i32>]>; +def NEONvgetlaneu : SDNode<"ARMISD::VGETLANEu", SDTARMVGETLN>; +def NEONvgetlanes : SDNode<"ARMISD::VGETLANEs", SDTARMVGETLN>; + +def SDTARMVMOVIMM : SDTypeProfile<1, 1, [SDTCisVec<0>, SDTCisVT<1, i32>]>; +def NEONvmovImm : SDNode<"ARMISD::VMOVIMM", SDTARMVMOVIMM>; +def NEONvmvnImm : SDNode<"ARMISD::VMVNIMM", SDTARMVMOVIMM>; +def NEONvmovFPImm : SDNode<"ARMISD::VMOVFPIMM", SDTARMVMOVIMM>; + +def SDTARMVORRIMM : SDTypeProfile<1, 2, [SDTCisVec<0>, SDTCisSameAs<0, 1>, + SDTCisVT<2, i32>]>; +def NEONvorrImm : SDNode<"ARMISD::VORRIMM", SDTARMVORRIMM>; +def NEONvbicImm : SDNode<"ARMISD::VBICIMM", SDTARMVORRIMM>; + +def NEONvbsl : SDNode<"ARMISD::VBSL", + SDTypeProfile<1, 3, [SDTCisVec<0>, + SDTCisSameAs<0, 1>, + SDTCisSameAs<0, 2>, + SDTCisSameAs<0, 3>]>>; + +def NEONvdup : SDNode<"ARMISD::VDUP", SDTypeProfile<1, 1, [SDTCisVec<0>]>>; + +// VDUPLANE can produce a quad-register result from a double-register source, +// so the result is not constrained to match the source. +def NEONvduplane : SDNode<"ARMISD::VDUPLANE", + SDTypeProfile<1, 2, [SDTCisVec<0>, SDTCisVec<1>, + SDTCisVT<2, i32>]>>; + +def SDTARMVEXT : SDTypeProfile<1, 3, [SDTCisVec<0>, SDTCisSameAs<0, 1>, + SDTCisSameAs<0, 2>, SDTCisVT<3, i32>]>; +def NEONvext : SDNode<"ARMISD::VEXT", SDTARMVEXT>; + +def SDTARMVSHUF : SDTypeProfile<1, 1, [SDTCisVec<0>, SDTCisSameAs<0, 1>]>; +def NEONvrev64 : SDNode<"ARMISD::VREV64", SDTARMVSHUF>; +def NEONvrev32 : SDNode<"ARMISD::VREV32", SDTARMVSHUF>; +def NEONvrev16 : SDNode<"ARMISD::VREV16", SDTARMVSHUF>; + +def SDTARMVSHUF2 : SDTypeProfile<2, 2, [SDTCisVec<0>, SDTCisSameAs<0, 1>, + SDTCisSameAs<0, 2>, + SDTCisSameAs<0, 3>]>; +def NEONzip : SDNode<"ARMISD::VZIP", SDTARMVSHUF2>; +def NEONuzp : SDNode<"ARMISD::VUZP", SDTARMVSHUF2>; +def NEONtrn : SDNode<"ARMISD::VTRN", SDTARMVSHUF2>; + +def SDTARMVMULL : SDTypeProfile<1, 2, [SDTCisInt<0>, SDTCisInt<1>, + SDTCisSameAs<1, 2>]>; +def NEONvmulls : SDNode<"ARMISD::VMULLs", SDTARMVMULL>; +def NEONvmullu : SDNode<"ARMISD::VMULLu", SDTARMVMULL>; + +def SDTARMVTBL1 : SDTypeProfile<1, 2, [SDTCisVT<0, v8i8>, SDTCisVT<1, v8i8>, + SDTCisVT<2, v8i8>]>; +def SDTARMVTBL2 : SDTypeProfile<1, 3, [SDTCisVT<0, v8i8>, SDTCisVT<1, v8i8>, + SDTCisVT<2, v8i8>, SDTCisVT<3, v8i8>]>; +def NEONvtbl1 : SDNode<"ARMISD::VTBL1", SDTARMVTBL1>; +def NEONvtbl2 : SDNode<"ARMISD::VTBL2", SDTARMVTBL2>; + + +def NEONimmAllZerosV: PatLeaf<(NEONvmovImm (i32 timm)), [{ + ConstantSDNode *ConstVal = cast<ConstantSDNode>(N->getOperand(0)); + unsigned EltBits = 0; + uint64_t EltVal = ARM_AM::decodeNEONModImm(ConstVal->getZExtValue(), EltBits); + return (EltBits == 32 && EltVal == 0); +}]>; + +def NEONimmAllOnesV: PatLeaf<(NEONvmovImm (i32 timm)), [{ + ConstantSDNode *ConstVal = cast<ConstantSDNode>(N->getOperand(0)); + unsigned EltBits = 0; + uint64_t EltVal = ARM_AM::decodeNEONModImm(ConstVal->getZExtValue(), EltBits); + return (EltBits == 8 && EltVal == 0xff); +}]>; + +//===----------------------------------------------------------------------===// +// NEON load / store instructions +//===----------------------------------------------------------------------===// + +// Use VLDM to load a Q register as a D register pair. +// This is a pseudo instruction that is expanded to VLDMD after reg alloc. +def VLDMQIA + : PseudoVFPLdStM<(outs DPair:$dst), (ins GPR:$Rn), + IIC_fpLoad_m, "", + [(set DPair:$dst, (v2f64 (word_alignedload GPR:$Rn)))]>; + +// Use VSTM to store a Q register as a D register pair. +// This is a pseudo instruction that is expanded to VSTMD after reg alloc. +def VSTMQIA + : PseudoVFPLdStM<(outs), (ins DPair:$src, GPR:$Rn), + IIC_fpStore_m, "", + [(word_alignedstore (v2f64 DPair:$src), GPR:$Rn)]>; + +// Classes for VLD* pseudo-instructions with multi-register operands. +// These are expanded to real instructions after register allocation. +class VLDQPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QPR:$dst), (ins addrmode6:$addr), itin, "">; +class VLDQWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QPR:$dst, GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset), itin, + "$addr.addr = $wb">; +class VLDQWBfixedPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QPR:$dst, GPR:$wb), + (ins addrmode6:$addr), itin, + "$addr.addr = $wb">; +class VLDQWBregisterPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QPR:$dst, GPR:$wb), + (ins addrmode6:$addr, rGPR:$offset), itin, + "$addr.addr = $wb">; + +class VLDQQPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQPR:$dst), (ins addrmode6:$addr), itin, "">; +class VLDQQWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQPR:$dst, GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset), itin, + "$addr.addr = $wb">; +class VLDQQWBfixedPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQPR:$dst, GPR:$wb), + (ins addrmode6:$addr), itin, + "$addr.addr = $wb">; +class VLDQQWBregisterPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQPR:$dst, GPR:$wb), + (ins addrmode6:$addr, rGPR:$offset), itin, + "$addr.addr = $wb">; + + +class VLDQQQQPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQQQPR:$dst), (ins addrmode6:$addr, QQQQPR:$src),itin, + "$src = $dst">; +class VLDQQQQWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQQQPR:$dst, GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QQQQPR:$src), itin, + "$addr.addr = $wb, $src = $dst">; + +let mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 in { + +// VLD1 : Vector Load (multiple single elements) +class VLD1D<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<0,0b10,0b0111,op7_4, (outs VecListOneD:$Vd), + (ins AddrMode:$Rn), IIC_VLD1, + "vld1", Dt, "$Vd, $Rn", "", []>, Sched<[WriteVLD1]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; +} +class VLD1Q<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<0,0b10,0b1010,op7_4, (outs VecListDPair:$Vd), + (ins AddrMode:$Rn), IIC_VLD1x2, + "vld1", Dt, "$Vd, $Rn", "", []>, Sched<[WriteVLD2]> { + let Rm = 0b1111; + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; +} + +def VLD1d8 : VLD1D<{0,0,0,?}, "8", addrmode6align64>; +def VLD1d16 : VLD1D<{0,1,0,?}, "16", addrmode6align64>; +def VLD1d32 : VLD1D<{1,0,0,?}, "32", addrmode6align64>; +def VLD1d64 : VLD1D<{1,1,0,?}, "64", addrmode6align64>; + +def VLD1q8 : VLD1Q<{0,0,?,?}, "8", addrmode6align64or128>; +def VLD1q16 : VLD1Q<{0,1,?,?}, "16", addrmode6align64or128>; +def VLD1q32 : VLD1Q<{1,0,?,?}, "32", addrmode6align64or128>; +def VLD1q64 : VLD1Q<{1,1,?,?}, "64", addrmode6align64or128>; + +// ...with address register writeback: +multiclass VLD1DWB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0,0b10, 0b0111,op7_4, (outs VecListOneD:$Vd, GPR:$wb), + (ins AddrMode:$Rn), IIC_VLD1u, + "vld1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVLD1]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } + def _register : NLdSt<0,0b10,0b0111,op7_4, (outs VecListOneD:$Vd, GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm), IIC_VLD1u, + "vld1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD1]> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } +} +multiclass VLD1QWB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0,0b10,0b1010,op7_4, (outs VecListDPair:$Vd, GPR:$wb), + (ins AddrMode:$Rn), IIC_VLD1x2u, + "vld1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVLD2]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } + def _register : NLdSt<0,0b10,0b1010,op7_4, (outs VecListDPair:$Vd, GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm), IIC_VLD1x2u, + "vld1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD2]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } +} + +defm VLD1d8wb : VLD1DWB<{0,0,0,?}, "8", addrmode6align64>; +defm VLD1d16wb : VLD1DWB<{0,1,0,?}, "16", addrmode6align64>; +defm VLD1d32wb : VLD1DWB<{1,0,0,?}, "32", addrmode6align64>; +defm VLD1d64wb : VLD1DWB<{1,1,0,?}, "64", addrmode6align64>; +defm VLD1q8wb : VLD1QWB<{0,0,?,?}, "8", addrmode6align64or128>; +defm VLD1q16wb : VLD1QWB<{0,1,?,?}, "16", addrmode6align64or128>; +defm VLD1q32wb : VLD1QWB<{1,0,?,?}, "32", addrmode6align64or128>; +defm VLD1q64wb : VLD1QWB<{1,1,?,?}, "64", addrmode6align64or128>; + +// ...with 3 registers +class VLD1D3<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<0,0b10,0b0110,op7_4, (outs VecListThreeD:$Vd), + (ins AddrMode:$Rn), IIC_VLD1x3, "vld1", Dt, + "$Vd, $Rn", "", []>, Sched<[WriteVLD3]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; +} +multiclass VLD1D3WB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0,0b10,0b0110, op7_4, (outs VecListThreeD:$Vd, GPR:$wb), + (ins AddrMode:$Rn), IIC_VLD1x2u, + "vld1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVLD3]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } + def _register : NLdSt<0,0b10,0b0110,op7_4, (outs VecListThreeD:$Vd, GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm), IIC_VLD1x2u, + "vld1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD3]> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } +} + +def VLD1d8T : VLD1D3<{0,0,0,?}, "8", addrmode6align64>; +def VLD1d16T : VLD1D3<{0,1,0,?}, "16", addrmode6align64>; +def VLD1d32T : VLD1D3<{1,0,0,?}, "32", addrmode6align64>; +def VLD1d64T : VLD1D3<{1,1,0,?}, "64", addrmode6align64>; + +defm VLD1d8Twb : VLD1D3WB<{0,0,0,?}, "8", addrmode6align64>; +defm VLD1d16Twb : VLD1D3WB<{0,1,0,?}, "16", addrmode6align64>; +defm VLD1d32Twb : VLD1D3WB<{1,0,0,?}, "32", addrmode6align64>; +defm VLD1d64Twb : VLD1D3WB<{1,1,0,?}, "64", addrmode6align64>; + +def VLD1d8TPseudo : VLDQQPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1d16TPseudo : VLDQQPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1d32TPseudo : VLDQQPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1d64TPseudo : VLDQQPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1d64TPseudoWB_fixed : VLDQQWBfixedPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1d64TPseudoWB_register : VLDQQWBregisterPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; + +def VLD1q8HighTPseudo : VLDQQQQPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1q8LowTPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1q16HighTPseudo : VLDQQQQPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1q16LowTPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1q32HighTPseudo : VLDQQQQPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1q32LowTPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1q64HighTPseudo : VLDQQQQPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; +def VLD1q64LowTPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD1x3>, Sched<[WriteVLD3]>; + +// ...with 4 registers +class VLD1D4<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<0, 0b10, 0b0010, op7_4, (outs VecListFourD:$Vd), + (ins AddrMode:$Rn), IIC_VLD1x4, "vld1", Dt, + "$Vd, $Rn", "", []>, Sched<[WriteVLD4]> { + let Rm = 0b1111; + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; +} +multiclass VLD1D4WB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0,0b10,0b0010, op7_4, (outs VecListFourD:$Vd, GPR:$wb), + (ins AddrMode:$Rn), IIC_VLD1x2u, + "vld1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVLD4]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } + def _register : NLdSt<0,0b10,0b0010,op7_4, (outs VecListFourD:$Vd, GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm), IIC_VLD1x2u, + "vld1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD4]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } +} + +def VLD1d8Q : VLD1D4<{0,0,?,?}, "8", addrmode6align64or128or256>; +def VLD1d16Q : VLD1D4<{0,1,?,?}, "16", addrmode6align64or128or256>; +def VLD1d32Q : VLD1D4<{1,0,?,?}, "32", addrmode6align64or128or256>; +def VLD1d64Q : VLD1D4<{1,1,?,?}, "64", addrmode6align64or128or256>; + +defm VLD1d8Qwb : VLD1D4WB<{0,0,?,?}, "8", addrmode6align64or128or256>; +defm VLD1d16Qwb : VLD1D4WB<{0,1,?,?}, "16", addrmode6align64or128or256>; +defm VLD1d32Qwb : VLD1D4WB<{1,0,?,?}, "32", addrmode6align64or128or256>; +defm VLD1d64Qwb : VLD1D4WB<{1,1,?,?}, "64", addrmode6align64or128or256>; + +def VLD1d8QPseudo : VLDQQPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1d16QPseudo : VLDQQPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1d32QPseudo : VLDQQPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1d64QPseudo : VLDQQPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1d64QPseudoWB_fixed : VLDQQWBfixedPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1d64QPseudoWB_register : VLDQQWBregisterPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; + +def VLD1q8LowQPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1q8HighQPseudo : VLDQQQQPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1q16LowQPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1q16HighQPseudo : VLDQQQQPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1q32LowQPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1q32HighQPseudo : VLDQQQQPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1q64LowQPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; +def VLD1q64HighQPseudo : VLDQQQQPseudo<IIC_VLD1x4>, Sched<[WriteVLD4]>; + +// VLD2 : Vector Load (multiple 2-element structures) +class VLD2<bits<4> op11_8, bits<4> op7_4, string Dt, RegisterOperand VdTy, + InstrItinClass itin, Operand AddrMode> + : NLdSt<0, 0b10, op11_8, op7_4, (outs VdTy:$Vd), + (ins AddrMode:$Rn), itin, + "vld2", Dt, "$Vd, $Rn", "", []> { + let Rm = 0b1111; + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST2Instruction"; +} + +def VLD2d8 : VLD2<0b1000, {0,0,?,?}, "8", VecListDPair, IIC_VLD2, + addrmode6align64or128>, Sched<[WriteVLD2]>; +def VLD2d16 : VLD2<0b1000, {0,1,?,?}, "16", VecListDPair, IIC_VLD2, + addrmode6align64or128>, Sched<[WriteVLD2]>; +def VLD2d32 : VLD2<0b1000, {1,0,?,?}, "32", VecListDPair, IIC_VLD2, + addrmode6align64or128>, Sched<[WriteVLD2]>; + +def VLD2q8 : VLD2<0b0011, {0,0,?,?}, "8", VecListFourD, IIC_VLD2x2, + addrmode6align64or128or256>, Sched<[WriteVLD4]>; +def VLD2q16 : VLD2<0b0011, {0,1,?,?}, "16", VecListFourD, IIC_VLD2x2, + addrmode6align64or128or256>, Sched<[WriteVLD4]>; +def VLD2q32 : VLD2<0b0011, {1,0,?,?}, "32", VecListFourD, IIC_VLD2x2, + addrmode6align64or128or256>, Sched<[WriteVLD4]>; + +def VLD2q8Pseudo : VLDQQPseudo<IIC_VLD2x2>, Sched<[WriteVLD4]>; +def VLD2q16Pseudo : VLDQQPseudo<IIC_VLD2x2>, Sched<[WriteVLD4]>; +def VLD2q32Pseudo : VLDQQPseudo<IIC_VLD2x2>, Sched<[WriteVLD4]>; + +// ...with address register writeback: +multiclass VLD2WB<bits<4> op11_8, bits<4> op7_4, string Dt, + RegisterOperand VdTy, InstrItinClass itin, Operand AddrMode> { + def _fixed : NLdSt<0, 0b10, op11_8, op7_4, (outs VdTy:$Vd, GPR:$wb), + (ins AddrMode:$Rn), itin, + "vld2", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST2Instruction"; + } + def _register : NLdSt<0, 0b10, op11_8, op7_4, (outs VdTy:$Vd, GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm), itin, + "vld2", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST2Instruction"; + } +} + +defm VLD2d8wb : VLD2WB<0b1000, {0,0,?,?}, "8", VecListDPair, IIC_VLD2u, + addrmode6align64or128>, Sched<[WriteVLD2]>; +defm VLD2d16wb : VLD2WB<0b1000, {0,1,?,?}, "16", VecListDPair, IIC_VLD2u, + addrmode6align64or128>, Sched<[WriteVLD2]>; +defm VLD2d32wb : VLD2WB<0b1000, {1,0,?,?}, "32", VecListDPair, IIC_VLD2u, + addrmode6align64or128>, Sched<[WriteVLD2]>; + +defm VLD2q8wb : VLD2WB<0b0011, {0,0,?,?}, "8", VecListFourD, IIC_VLD2x2u, + addrmode6align64or128or256>, Sched<[WriteVLD4]>; +defm VLD2q16wb : VLD2WB<0b0011, {0,1,?,?}, "16", VecListFourD, IIC_VLD2x2u, + addrmode6align64or128or256>, Sched<[WriteVLD4]>; +defm VLD2q32wb : VLD2WB<0b0011, {1,0,?,?}, "32", VecListFourD, IIC_VLD2x2u, + addrmode6align64or128or256>, Sched<[WriteVLD4]>; + +def VLD2q8PseudoWB_fixed : VLDQQWBfixedPseudo<IIC_VLD2x2u>, Sched<[WriteVLD4]>; +def VLD2q16PseudoWB_fixed : VLDQQWBfixedPseudo<IIC_VLD2x2u>, Sched<[WriteVLD4]>; +def VLD2q32PseudoWB_fixed : VLDQQWBfixedPseudo<IIC_VLD2x2u>, Sched<[WriteVLD4]>; +def VLD2q8PseudoWB_register : VLDQQWBregisterPseudo<IIC_VLD2x2u>, Sched<[WriteVLD4]>; +def VLD2q16PseudoWB_register : VLDQQWBregisterPseudo<IIC_VLD2x2u>, Sched<[WriteVLD4]>; +def VLD2q32PseudoWB_register : VLDQQWBregisterPseudo<IIC_VLD2x2u>, Sched<[WriteVLD4]>; + +// ...with double-spaced registers +def VLD2b8 : VLD2<0b1001, {0,0,?,?}, "8", VecListDPairSpaced, IIC_VLD2, + addrmode6align64or128>, Sched<[WriteVLD2]>; +def VLD2b16 : VLD2<0b1001, {0,1,?,?}, "16", VecListDPairSpaced, IIC_VLD2, + addrmode6align64or128>, Sched<[WriteVLD2]>; +def VLD2b32 : VLD2<0b1001, {1,0,?,?}, "32", VecListDPairSpaced, IIC_VLD2, + addrmode6align64or128>, Sched<[WriteVLD2]>; +defm VLD2b8wb : VLD2WB<0b1001, {0,0,?,?}, "8", VecListDPairSpaced, IIC_VLD2u, + addrmode6align64or128>, Sched<[WriteVLD2]>; +defm VLD2b16wb : VLD2WB<0b1001, {0,1,?,?}, "16", VecListDPairSpaced, IIC_VLD2u, + addrmode6align64or128>, Sched<[WriteVLD2]>; +defm VLD2b32wb : VLD2WB<0b1001, {1,0,?,?}, "32", VecListDPairSpaced, IIC_VLD2u, + addrmode6align64or128>, Sched<[WriteVLD2]>; + +// VLD3 : Vector Load (multiple 3-element structures) +class VLD3D<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdSt<0, 0b10, op11_8, op7_4, (outs DPR:$Vd, DPR:$dst2, DPR:$dst3), + (ins addrmode6:$Rn), IIC_VLD3, + "vld3", Dt, "\\{$Vd, $dst2, $dst3\\}, $Rn", "", []>, Sched<[WriteVLD3]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST3Instruction"; +} + +def VLD3d8 : VLD3D<0b0100, {0,0,0,?}, "8">; +def VLD3d16 : VLD3D<0b0100, {0,1,0,?}, "16">; +def VLD3d32 : VLD3D<0b0100, {1,0,0,?}, "32">; + +def VLD3d8Pseudo : VLDQQPseudo<IIC_VLD3>, Sched<[WriteVLD3]>; +def VLD3d16Pseudo : VLDQQPseudo<IIC_VLD3>, Sched<[WriteVLD3]>; +def VLD3d32Pseudo : VLDQQPseudo<IIC_VLD3>, Sched<[WriteVLD3]>; + +// ...with address register writeback: +class VLD3DWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdSt<0, 0b10, op11_8, op7_4, + (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm), IIC_VLD3u, + "vld3", Dt, "\\{$Vd, $dst2, $dst3\\}, $Rn$Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD3]> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST3Instruction"; +} + +def VLD3d8_UPD : VLD3DWB<0b0100, {0,0,0,?}, "8">; +def VLD3d16_UPD : VLD3DWB<0b0100, {0,1,0,?}, "16">; +def VLD3d32_UPD : VLD3DWB<0b0100, {1,0,0,?}, "32">; + +def VLD3d8Pseudo_UPD : VLDQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; +def VLD3d16Pseudo_UPD : VLDQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; +def VLD3d32Pseudo_UPD : VLDQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; + +// ...with double-spaced registers: +def VLD3q8 : VLD3D<0b0101, {0,0,0,?}, "8">; +def VLD3q16 : VLD3D<0b0101, {0,1,0,?}, "16">; +def VLD3q32 : VLD3D<0b0101, {1,0,0,?}, "32">; +def VLD3q8_UPD : VLD3DWB<0b0101, {0,0,0,?}, "8">; +def VLD3q16_UPD : VLD3DWB<0b0101, {0,1,0,?}, "16">; +def VLD3q32_UPD : VLD3DWB<0b0101, {1,0,0,?}, "32">; + +def VLD3q8Pseudo_UPD : VLDQQQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; +def VLD3q16Pseudo_UPD : VLDQQQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; +def VLD3q32Pseudo_UPD : VLDQQQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; + +// ...alternate versions to be allocated odd register numbers: +def VLD3q8oddPseudo : VLDQQQQPseudo<IIC_VLD3>, Sched<[WriteVLD3]>; +def VLD3q16oddPseudo : VLDQQQQPseudo<IIC_VLD3>, Sched<[WriteVLD3]>; +def VLD3q32oddPseudo : VLDQQQQPseudo<IIC_VLD3>, Sched<[WriteVLD3]>; + +def VLD3q8oddPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; +def VLD3q16oddPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; +def VLD3q32oddPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD3u>, Sched<[WriteVLD3]>; + +// VLD4 : Vector Load (multiple 4-element structures) +class VLD4D<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdSt<0, 0b10, op11_8, op7_4, + (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4), + (ins addrmode6:$Rn), IIC_VLD4, + "vld4", Dt, "\\{$Vd, $dst2, $dst3, $dst4\\}, $Rn", "", []>, + Sched<[WriteVLD4]> { + let Rm = 0b1111; + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST4Instruction"; +} + +def VLD4d8 : VLD4D<0b0000, {0,0,?,?}, "8">; +def VLD4d16 : VLD4D<0b0000, {0,1,?,?}, "16">; +def VLD4d32 : VLD4D<0b0000, {1,0,?,?}, "32">; + +def VLD4d8Pseudo : VLDQQPseudo<IIC_VLD4>, Sched<[WriteVLD4]>; +def VLD4d16Pseudo : VLDQQPseudo<IIC_VLD4>, Sched<[WriteVLD4]>; +def VLD4d32Pseudo : VLDQQPseudo<IIC_VLD4>, Sched<[WriteVLD4]>; + +// ...with address register writeback: +class VLD4DWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdSt<0, 0b10, op11_8, op7_4, + (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4, GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm), IIC_VLD4u, + "vld4", Dt, "\\{$Vd, $dst2, $dst3, $dst4\\}, $Rn$Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD4]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST4Instruction"; +} + +def VLD4d8_UPD : VLD4DWB<0b0000, {0,0,?,?}, "8">; +def VLD4d16_UPD : VLD4DWB<0b0000, {0,1,?,?}, "16">; +def VLD4d32_UPD : VLD4DWB<0b0000, {1,0,?,?}, "32">; + +def VLD4d8Pseudo_UPD : VLDQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; +def VLD4d16Pseudo_UPD : VLDQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; +def VLD4d32Pseudo_UPD : VLDQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; + +// ...with double-spaced registers: +def VLD4q8 : VLD4D<0b0001, {0,0,?,?}, "8">; +def VLD4q16 : VLD4D<0b0001, {0,1,?,?}, "16">; +def VLD4q32 : VLD4D<0b0001, {1,0,?,?}, "32">; +def VLD4q8_UPD : VLD4DWB<0b0001, {0,0,?,?}, "8">; +def VLD4q16_UPD : VLD4DWB<0b0001, {0,1,?,?}, "16">; +def VLD4q32_UPD : VLD4DWB<0b0001, {1,0,?,?}, "32">; + +def VLD4q8Pseudo_UPD : VLDQQQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; +def VLD4q16Pseudo_UPD : VLDQQQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; +def VLD4q32Pseudo_UPD : VLDQQQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; + +// ...alternate versions to be allocated odd register numbers: +def VLD4q8oddPseudo : VLDQQQQPseudo<IIC_VLD4>, Sched<[WriteVLD4]>; +def VLD4q16oddPseudo : VLDQQQQPseudo<IIC_VLD4>, Sched<[WriteVLD4]>; +def VLD4q32oddPseudo : VLDQQQQPseudo<IIC_VLD4>, Sched<[WriteVLD4]>; + +def VLD4q8oddPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; +def VLD4q16oddPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; +def VLD4q32oddPseudo_UPD : VLDQQQQWBPseudo<IIC_VLD4u>, Sched<[WriteVLD4]>; + +} // mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 + +// Classes for VLD*LN pseudo-instructions with multi-register operands. +// These are expanded to real instructions after register allocation. +class VLDQLNPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QPR:$dst), + (ins addrmode6:$addr, QPR:$src, nohash_imm:$lane), + itin, "$src = $dst">; +class VLDQLNWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QPR:$dst, GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QPR:$src, + nohash_imm:$lane), itin, "$addr.addr = $wb, $src = $dst">; +class VLDQQLNPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQPR:$dst), + (ins addrmode6:$addr, QQPR:$src, nohash_imm:$lane), + itin, "$src = $dst">; +class VLDQQLNWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQPR:$dst, GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QQPR:$src, + nohash_imm:$lane), itin, "$addr.addr = $wb, $src = $dst">; +class VLDQQQQLNPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQQQPR:$dst), + (ins addrmode6:$addr, QQQQPR:$src, nohash_imm:$lane), + itin, "$src = $dst">; +class VLDQQQQLNWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs QQQQPR:$dst, GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QQQQPR:$src, + nohash_imm:$lane), itin, "$addr.addr = $wb, $src = $dst">; + +// VLD1LN : Vector Load (single element to one lane) +class VLD1LN<bits<4> op11_8, bits<4> op7_4, string Dt, ValueType Ty, + PatFrag LoadOp> + : NLdStLn<1, 0b10, op11_8, op7_4, (outs DPR:$Vd), + (ins addrmode6:$Rn, DPR:$src, nohash_imm:$lane), + IIC_VLD1ln, "vld1", Dt, "\\{$Vd[$lane]\\}, $Rn", + "$src = $Vd", + [(set DPR:$Vd, (vector_insert (Ty DPR:$src), + (i32 (LoadOp addrmode6:$Rn)), + imm:$lane))]> { + let Rm = 0b1111; + let DecoderMethod = "DecodeVLD1LN"; +} +class VLD1LN32<bits<4> op11_8, bits<4> op7_4, string Dt, ValueType Ty, + PatFrag LoadOp> + : NLdStLn<1, 0b10, op11_8, op7_4, (outs DPR:$Vd), + (ins addrmode6oneL32:$Rn, DPR:$src, nohash_imm:$lane), + IIC_VLD1ln, "vld1", Dt, "\\{$Vd[$lane]\\}, $Rn", + "$src = $Vd", + [(set DPR:$Vd, (vector_insert (Ty DPR:$src), + (i32 (LoadOp addrmode6oneL32:$Rn)), + imm:$lane))]>, Sched<[WriteVLD1]> { + let Rm = 0b1111; + let DecoderMethod = "DecodeVLD1LN"; +} +class VLD1QLNPseudo<ValueType Ty, PatFrag LoadOp> : VLDQLNPseudo<IIC_VLD1ln>, + Sched<[WriteVLD1]> { + let Pattern = [(set QPR:$dst, (vector_insert (Ty QPR:$src), + (i32 (LoadOp addrmode6:$addr)), + imm:$lane))]; +} + +def VLD1LNd8 : VLD1LN<0b0000, {?,?,?,0}, "8", v8i8, extloadi8> { + let Inst{7-5} = lane{2-0}; +} +def VLD1LNd16 : VLD1LN<0b0100, {?,?,0,?}, "16", v4i16, extloadi16> { + let Inst{7-6} = lane{1-0}; + let Inst{5-4} = Rn{5-4}; +} +def VLD1LNd32 : VLD1LN32<0b1000, {?,0,?,?}, "32", v2i32, load> { + let Inst{7} = lane{0}; + let Inst{5-4} = Rn{5-4}; +} + +def VLD1LNq8Pseudo : VLD1QLNPseudo<v16i8, extloadi8>; +def VLD1LNq16Pseudo : VLD1QLNPseudo<v8i16, extloadi16>; +def VLD1LNq32Pseudo : VLD1QLNPseudo<v4i32, load>; + +def : Pat<(vector_insert (v2f32 DPR:$src), + (f32 (load addrmode6:$addr)), imm:$lane), + (VLD1LNd32 addrmode6:$addr, DPR:$src, imm:$lane)>; +def : Pat<(vector_insert (v4f32 QPR:$src), + (f32 (load addrmode6:$addr)), imm:$lane), + (VLD1LNq32Pseudo addrmode6:$addr, QPR:$src, imm:$lane)>; + +// A 64-bit subvector insert to the first 128-bit vector position +// is a subregister copy that needs no instruction. +def : Pat<(insert_subvector undef, (v1i64 DPR:$src), (i32 0)), + (INSERT_SUBREG (v2i64 (IMPLICIT_DEF)), DPR:$src, dsub_0)>; +def : Pat<(insert_subvector undef, (v2i32 DPR:$src), (i32 0)), + (INSERT_SUBREG (v4i32 (IMPLICIT_DEF)), DPR:$src, dsub_0)>; +def : Pat<(insert_subvector undef, (v2f32 DPR:$src), (i32 0)), + (INSERT_SUBREG (v4f32 (IMPLICIT_DEF)), DPR:$src, dsub_0)>; +def : Pat<(insert_subvector undef, (v4i16 DPR:$src), (i32 0)), + (INSERT_SUBREG (v8i16 (IMPLICIT_DEF)), DPR:$src, dsub_0)>; +def : Pat<(insert_subvector undef, (v4f16 DPR:$src), (i32 0)), + (INSERT_SUBREG (v8f16 (IMPLICIT_DEF)), DPR:$src, dsub_0)>; +def : Pat<(insert_subvector (v16i8 undef), (v8i8 DPR:$src), (i32 0)), + (INSERT_SUBREG (v16i8 (IMPLICIT_DEF)), DPR:$src, dsub_0)>; + + +let mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 in { + +// ...with address register writeback: +class VLD1LNWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b10, op11_8, op7_4, (outs DPR:$Vd, GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$src, nohash_imm:$lane), IIC_VLD1lnu, "vld1", Dt, + "\\{$Vd[$lane]\\}, $Rn$Rm", + "$src = $Vd, $Rn.addr = $wb", []>, Sched<[WriteVLD1]> { + let DecoderMethod = "DecodeVLD1LN"; +} + +def VLD1LNd8_UPD : VLD1LNWB<0b0000, {?,?,?,0}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VLD1LNd16_UPD : VLD1LNWB<0b0100, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; + let Inst{4} = Rn{4}; +} +def VLD1LNd32_UPD : VLD1LNWB<0b1000, {?,0,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{4}; + let Inst{4} = Rn{4}; +} + +def VLD1LNq8Pseudo_UPD : VLDQLNWBPseudo<IIC_VLD1lnu>, Sched<[WriteVLD1]>; +def VLD1LNq16Pseudo_UPD : VLDQLNWBPseudo<IIC_VLD1lnu>, Sched<[WriteVLD1]>; +def VLD1LNq32Pseudo_UPD : VLDQLNWBPseudo<IIC_VLD1lnu>, Sched<[WriteVLD1]>; + +// VLD2LN : Vector Load (single 2-element structure to one lane) +class VLD2LN<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b10, op11_8, op7_4, (outs DPR:$Vd, DPR:$dst2), + (ins addrmode6:$Rn, DPR:$src1, DPR:$src2, nohash_imm:$lane), + IIC_VLD2ln, "vld2", Dt, "\\{$Vd[$lane], $dst2[$lane]\\}, $Rn", + "$src1 = $Vd, $src2 = $dst2", []>, Sched<[WriteVLD1]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD2LN"; +} + +def VLD2LNd8 : VLD2LN<0b0001, {?,?,?,?}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VLD2LNd16 : VLD2LN<0b0101, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD2LNd32 : VLD2LN<0b1001, {?,0,0,?}, "32"> { + let Inst{7} = lane{0}; +} + +def VLD2LNd8Pseudo : VLDQLNPseudo<IIC_VLD2ln>, Sched<[WriteVLD1]>; +def VLD2LNd16Pseudo : VLDQLNPseudo<IIC_VLD2ln>, Sched<[WriteVLD1]>; +def VLD2LNd32Pseudo : VLDQLNPseudo<IIC_VLD2ln>, Sched<[WriteVLD1]>; + +// ...with double-spaced registers: +def VLD2LNq16 : VLD2LN<0b0101, {?,?,1,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD2LNq32 : VLD2LN<0b1001, {?,1,0,?}, "32"> { + let Inst{7} = lane{0}; +} + +def VLD2LNq16Pseudo : VLDQQLNPseudo<IIC_VLD2ln>, Sched<[WriteVLD1]>; +def VLD2LNq32Pseudo : VLDQQLNPseudo<IIC_VLD2ln>, Sched<[WriteVLD1]>; + +// ...with address register writeback: +class VLD2LNWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b10, op11_8, op7_4, (outs DPR:$Vd, DPR:$dst2, GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$src1, DPR:$src2, nohash_imm:$lane), IIC_VLD2lnu, "vld2", Dt, + "\\{$Vd[$lane], $dst2[$lane]\\}, $Rn$Rm", + "$src1 = $Vd, $src2 = $dst2, $Rn.addr = $wb", []> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD2LN"; +} + +def VLD2LNd8_UPD : VLD2LNWB<0b0001, {?,?,?,?}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VLD2LNd16_UPD : VLD2LNWB<0b0101, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD2LNd32_UPD : VLD2LNWB<0b1001, {?,0,0,?}, "32"> { + let Inst{7} = lane{0}; +} + +def VLD2LNd8Pseudo_UPD : VLDQLNWBPseudo<IIC_VLD2lnu>, Sched<[WriteVLD1]>; +def VLD2LNd16Pseudo_UPD : VLDQLNWBPseudo<IIC_VLD2lnu>, Sched<[WriteVLD1]>; +def VLD2LNd32Pseudo_UPD : VLDQLNWBPseudo<IIC_VLD2lnu>, Sched<[WriteVLD1]>; + +def VLD2LNq16_UPD : VLD2LNWB<0b0101, {?,?,1,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD2LNq32_UPD : VLD2LNWB<0b1001, {?,1,0,?}, "32"> { + let Inst{7} = lane{0}; +} + +def VLD2LNq16Pseudo_UPD : VLDQQLNWBPseudo<IIC_VLD2lnu>, Sched<[WriteVLD1]>; +def VLD2LNq32Pseudo_UPD : VLDQQLNWBPseudo<IIC_VLD2lnu>, Sched<[WriteVLD1]>; + +// VLD3LN : Vector Load (single 3-element structure to one lane) +class VLD3LN<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b10, op11_8, op7_4, (outs DPR:$Vd, DPR:$dst2, DPR:$dst3), + (ins addrmode6:$Rn, DPR:$src1, DPR:$src2, DPR:$src3, + nohash_imm:$lane), IIC_VLD3ln, "vld3", Dt, + "\\{$Vd[$lane], $dst2[$lane], $dst3[$lane]\\}, $Rn", + "$src1 = $Vd, $src2 = $dst2, $src3 = $dst3", []>, Sched<[WriteVLD2]> { + let Rm = 0b1111; + let DecoderMethod = "DecodeVLD3LN"; +} + +def VLD3LNd8 : VLD3LN<0b0010, {?,?,?,0}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VLD3LNd16 : VLD3LN<0b0110, {?,?,0,0}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD3LNd32 : VLD3LN<0b1010, {?,0,0,0}, "32"> { + let Inst{7} = lane{0}; +} + +def VLD3LNd8Pseudo : VLDQQLNPseudo<IIC_VLD3ln>, Sched<[WriteVLD2]>; +def VLD3LNd16Pseudo : VLDQQLNPseudo<IIC_VLD3ln>, Sched<[WriteVLD2]>; +def VLD3LNd32Pseudo : VLDQQLNPseudo<IIC_VLD3ln>, Sched<[WriteVLD2]>; + +// ...with double-spaced registers: +def VLD3LNq16 : VLD3LN<0b0110, {?,?,1,0}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD3LNq32 : VLD3LN<0b1010, {?,1,0,0}, "32"> { + let Inst{7} = lane{0}; +} + +def VLD3LNq16Pseudo : VLDQQQQLNPseudo<IIC_VLD3ln>, Sched<[WriteVLD2]>; +def VLD3LNq32Pseudo : VLDQQQQLNPseudo<IIC_VLD3ln>, Sched<[WriteVLD2]>; + +// ...with address register writeback: +class VLD3LNWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b10, op11_8, op7_4, + (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$src1, DPR:$src2, DPR:$src3, nohash_imm:$lane), + IIC_VLD3lnu, "vld3", Dt, + "\\{$Vd[$lane], $dst2[$lane], $dst3[$lane]\\}, $Rn$Rm", + "$src1 = $Vd, $src2 = $dst2, $src3 = $dst3, $Rn.addr = $wb", + []>, Sched<[WriteVLD2]> { + let DecoderMethod = "DecodeVLD3LN"; +} + +def VLD3LNd8_UPD : VLD3LNWB<0b0010, {?,?,?,0}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VLD3LNd16_UPD : VLD3LNWB<0b0110, {?,?,0,0}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD3LNd32_UPD : VLD3LNWB<0b1010, {?,0,0,0}, "32"> { + let Inst{7} = lane{0}; +} + +def VLD3LNd8Pseudo_UPD : VLDQQLNWBPseudo<IIC_VLD3lnu>, Sched<[WriteVLD2]>; +def VLD3LNd16Pseudo_UPD : VLDQQLNWBPseudo<IIC_VLD3lnu>, Sched<[WriteVLD2]>; +def VLD3LNd32Pseudo_UPD : VLDQQLNWBPseudo<IIC_VLD3lnu>, Sched<[WriteVLD2]>; + +def VLD3LNq16_UPD : VLD3LNWB<0b0110, {?,?,1,0}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD3LNq32_UPD : VLD3LNWB<0b1010, {?,1,0,0}, "32"> { + let Inst{7} = lane{0}; +} + +def VLD3LNq16Pseudo_UPD : VLDQQQQLNWBPseudo<IIC_VLD3lnu>, Sched<[WriteVLD2]>; +def VLD3LNq32Pseudo_UPD : VLDQQQQLNWBPseudo<IIC_VLD3lnu>, Sched<[WriteVLD2]>; + +// VLD4LN : Vector Load (single 4-element structure to one lane) +class VLD4LN<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b10, op11_8, op7_4, + (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4), + (ins addrmode6:$Rn, DPR:$src1, DPR:$src2, DPR:$src3, DPR:$src4, + nohash_imm:$lane), IIC_VLD4ln, "vld4", Dt, + "\\{$Vd[$lane], $dst2[$lane], $dst3[$lane], $dst4[$lane]\\}, $Rn", + "$src1 = $Vd, $src2 = $dst2, $src3 = $dst3, $src4 = $dst4", []>, + Sched<[WriteVLD2]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD4LN"; +} + +def VLD4LNd8 : VLD4LN<0b0011, {?,?,?,?}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VLD4LNd16 : VLD4LN<0b0111, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD4LNd32 : VLD4LN<0b1011, {?,0,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{5}; +} + +def VLD4LNd8Pseudo : VLDQQLNPseudo<IIC_VLD4ln>, Sched<[WriteVLD2]>; +def VLD4LNd16Pseudo : VLDQQLNPseudo<IIC_VLD4ln>, Sched<[WriteVLD2]>; +def VLD4LNd32Pseudo : VLDQQLNPseudo<IIC_VLD4ln>, Sched<[WriteVLD2]>; + +// ...with double-spaced registers: +def VLD4LNq16 : VLD4LN<0b0111, {?,?,1,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD4LNq32 : VLD4LN<0b1011, {?,1,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{5}; +} + +def VLD4LNq16Pseudo : VLDQQQQLNPseudo<IIC_VLD4ln>, Sched<[WriteVLD2]>; +def VLD4LNq32Pseudo : VLDQQQQLNPseudo<IIC_VLD4ln>, Sched<[WriteVLD2]>; + +// ...with address register writeback: +class VLD4LNWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b10, op11_8, op7_4, + (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4, GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$src1, DPR:$src2, DPR:$src3, DPR:$src4, nohash_imm:$lane), + IIC_VLD4lnu, "vld4", Dt, +"\\{$Vd[$lane], $dst2[$lane], $dst3[$lane], $dst4[$lane]\\}, $Rn$Rm", +"$src1 = $Vd, $src2 = $dst2, $src3 = $dst3, $src4 = $dst4, $Rn.addr = $wb", + []> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD4LN" ; +} + +def VLD4LNd8_UPD : VLD4LNWB<0b0011, {?,?,?,?}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VLD4LNd16_UPD : VLD4LNWB<0b0111, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD4LNd32_UPD : VLD4LNWB<0b1011, {?,0,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{5}; +} + +def VLD4LNd8Pseudo_UPD : VLDQQLNWBPseudo<IIC_VLD4lnu>, Sched<[WriteVLD2]>; +def VLD4LNd16Pseudo_UPD : VLDQQLNWBPseudo<IIC_VLD4lnu>, Sched<[WriteVLD2]>; +def VLD4LNd32Pseudo_UPD : VLDQQLNWBPseudo<IIC_VLD4lnu>, Sched<[WriteVLD2]>; + +def VLD4LNq16_UPD : VLD4LNWB<0b0111, {?,?,1,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VLD4LNq32_UPD : VLD4LNWB<0b1011, {?,1,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{5}; +} + +def VLD4LNq16Pseudo_UPD : VLDQQQQLNWBPseudo<IIC_VLD4lnu>, Sched<[WriteVLD2]>; +def VLD4LNq32Pseudo_UPD : VLDQQQQLNWBPseudo<IIC_VLD4lnu>, Sched<[WriteVLD2]>; + +} // mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 + +// VLD1DUP : Vector Load (single element to all lanes) +class VLD1DUP<bits<4> op7_4, string Dt, ValueType Ty, PatFrag LoadOp, + Operand AddrMode> + : NLdSt<1, 0b10, 0b1100, op7_4, (outs VecListOneDAllLanes:$Vd), + (ins AddrMode:$Rn), + IIC_VLD1dup, "vld1", Dt, "$Vd, $Rn", "", + [(set VecListOneDAllLanes:$Vd, + (Ty (NEONvdup (i32 (LoadOp AddrMode:$Rn)))))]>, + Sched<[WriteVLD2]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD1DupInstruction"; +} +def VLD1DUPd8 : VLD1DUP<{0,0,0,?}, "8", v8i8, extloadi8, + addrmode6dupalignNone>; +def VLD1DUPd16 : VLD1DUP<{0,1,0,?}, "16", v4i16, extloadi16, + addrmode6dupalign16>; +def VLD1DUPd32 : VLD1DUP<{1,0,0,?}, "32", v2i32, load, + addrmode6dupalign32>; + +def : Pat<(v2f32 (NEONvdup (f32 (load addrmode6dup:$addr)))), + (VLD1DUPd32 addrmode6:$addr)>; + +class VLD1QDUP<bits<4> op7_4, string Dt, ValueType Ty, PatFrag LoadOp, + Operand AddrMode> + : NLdSt<1, 0b10, 0b1100, op7_4, (outs VecListDPairAllLanes:$Vd), + (ins AddrMode:$Rn), IIC_VLD1dup, + "vld1", Dt, "$Vd, $Rn", "", + [(set VecListDPairAllLanes:$Vd, + (Ty (NEONvdup (i32 (LoadOp AddrMode:$Rn)))))]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD1DupInstruction"; +} + +def VLD1DUPq8 : VLD1QDUP<{0,0,1,0}, "8", v16i8, extloadi8, + addrmode6dupalignNone>; +def VLD1DUPq16 : VLD1QDUP<{0,1,1,?}, "16", v8i16, extloadi16, + addrmode6dupalign16>; +def VLD1DUPq32 : VLD1QDUP<{1,0,1,?}, "32", v4i32, load, + addrmode6dupalign32>; + +def : Pat<(v4f32 (NEONvdup (f32 (load addrmode6dup:$addr)))), + (VLD1DUPq32 addrmode6:$addr)>; + +let mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 in { +// ...with address register writeback: +multiclass VLD1DUPWB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<1, 0b10, 0b1100, op7_4, + (outs VecListOneDAllLanes:$Vd, GPR:$wb), + (ins AddrMode:$Rn), IIC_VLD1dupu, + "vld1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD1DupInstruction"; + } + def _register : NLdSt<1, 0b10, 0b1100, op7_4, + (outs VecListOneDAllLanes:$Vd, GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm), IIC_VLD1dupu, + "vld1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD1DupInstruction"; + } +} +multiclass VLD1QDUPWB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<1, 0b10, 0b1100, op7_4, + (outs VecListDPairAllLanes:$Vd, GPR:$wb), + (ins AddrMode:$Rn), IIC_VLD1dupu, + "vld1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVLD1]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD1DupInstruction"; + } + def _register : NLdSt<1, 0b10, 0b1100, op7_4, + (outs VecListDPairAllLanes:$Vd, GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm), IIC_VLD1dupu, + "vld1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD1DupInstruction"; + } +} + +defm VLD1DUPd8wb : VLD1DUPWB<{0,0,0,0}, "8", addrmode6dupalignNone>; +defm VLD1DUPd16wb : VLD1DUPWB<{0,1,0,?}, "16", addrmode6dupalign16>; +defm VLD1DUPd32wb : VLD1DUPWB<{1,0,0,?}, "32", addrmode6dupalign32>; + +defm VLD1DUPq8wb : VLD1QDUPWB<{0,0,1,0}, "8", addrmode6dupalignNone>; +defm VLD1DUPq16wb : VLD1QDUPWB<{0,1,1,?}, "16", addrmode6dupalign16>; +defm VLD1DUPq32wb : VLD1QDUPWB<{1,0,1,?}, "32", addrmode6dupalign32>; + +// VLD2DUP : Vector Load (single 2-element structure to all lanes) +class VLD2DUP<bits<4> op7_4, string Dt, RegisterOperand VdTy, Operand AddrMode> + : NLdSt<1, 0b10, 0b1101, op7_4, (outs VdTy:$Vd), + (ins AddrMode:$Rn), IIC_VLD2dup, + "vld2", Dt, "$Vd, $Rn", "", []> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD2DupInstruction"; +} + +def VLD2DUPd8 : VLD2DUP<{0,0,0,?}, "8", VecListDPairAllLanes, + addrmode6dupalign16>; +def VLD2DUPd16 : VLD2DUP<{0,1,0,?}, "16", VecListDPairAllLanes, + addrmode6dupalign32>; +def VLD2DUPd32 : VLD2DUP<{1,0,0,?}, "32", VecListDPairAllLanes, + addrmode6dupalign64>; + +// HACK this one, VLD2DUPd8x2 must be changed at the same time with VLD2b8 or +// "vld2.8 {d0[], d2[]}, [r4:32]" will become "vld2.8 {d0, d2}, [r4:32]". +// ...with double-spaced registers +def VLD2DUPd8x2 : VLD2DUP<{0,0,1,?}, "8", VecListDPairSpacedAllLanes, + addrmode6dupalign16>; +def VLD2DUPd16x2 : VLD2DUP<{0,1,1,?}, "16", VecListDPairSpacedAllLanes, + addrmode6dupalign32>; +def VLD2DUPd32x2 : VLD2DUP<{1,0,1,?}, "32", VecListDPairSpacedAllLanes, + addrmode6dupalign64>; + +def VLD2DUPq8EvenPseudo : VLDQQPseudo<IIC_VLD2dup>, Sched<[WriteVLD2]>; +def VLD2DUPq8OddPseudo : VLDQQPseudo<IIC_VLD2dup>, Sched<[WriteVLD2]>; +def VLD2DUPq16EvenPseudo : VLDQQPseudo<IIC_VLD2dup>, Sched<[WriteVLD2]>; +def VLD2DUPq16OddPseudo : VLDQQPseudo<IIC_VLD2dup>, Sched<[WriteVLD2]>; +def VLD2DUPq32EvenPseudo : VLDQQPseudo<IIC_VLD2dup>, Sched<[WriteVLD2]>; +def VLD2DUPq32OddPseudo : VLDQQPseudo<IIC_VLD2dup>, Sched<[WriteVLD2]>; + +// ...with address register writeback: +multiclass VLD2DUPWB<bits<4> op7_4, string Dt, RegisterOperand VdTy, + Operand AddrMode> { + def _fixed : NLdSt<1, 0b10, 0b1101, op7_4, + (outs VdTy:$Vd, GPR:$wb), + (ins AddrMode:$Rn), IIC_VLD2dupu, + "vld2", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVLD1]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD2DupInstruction"; + } + def _register : NLdSt<1, 0b10, 0b1101, op7_4, + (outs VdTy:$Vd, GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm), IIC_VLD2dupu, + "vld2", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD1]> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD2DupInstruction"; + } +} + +defm VLD2DUPd8wb : VLD2DUPWB<{0,0,0,0}, "8", VecListDPairAllLanes, + addrmode6dupalign16>; +defm VLD2DUPd16wb : VLD2DUPWB<{0,1,0,?}, "16", VecListDPairAllLanes, + addrmode6dupalign32>; +defm VLD2DUPd32wb : VLD2DUPWB<{1,0,0,?}, "32", VecListDPairAllLanes, + addrmode6dupalign64>; + +defm VLD2DUPd8x2wb : VLD2DUPWB<{0,0,1,0}, "8", VecListDPairSpacedAllLanes, + addrmode6dupalign16>; +defm VLD2DUPd16x2wb : VLD2DUPWB<{0,1,1,?}, "16", VecListDPairSpacedAllLanes, + addrmode6dupalign32>; +defm VLD2DUPd32x2wb : VLD2DUPWB<{1,0,1,?}, "32", VecListDPairSpacedAllLanes, + addrmode6dupalign64>; + +// VLD3DUP : Vector Load (single 3-element structure to all lanes) +class VLD3DUP<bits<4> op7_4, string Dt> + : NLdSt<1, 0b10, 0b1110, op7_4, (outs DPR:$Vd, DPR:$dst2, DPR:$dst3), + (ins addrmode6dup:$Rn), IIC_VLD3dup, + "vld3", Dt, "\\{$Vd[], $dst2[], $dst3[]\\}, $Rn", "", []>, + Sched<[WriteVLD2]> { + let Rm = 0b1111; + let Inst{4} = 0; + let DecoderMethod = "DecodeVLD3DupInstruction"; +} + +def VLD3DUPd8 : VLD3DUP<{0,0,0,?}, "8">; +def VLD3DUPd16 : VLD3DUP<{0,1,0,?}, "16">; +def VLD3DUPd32 : VLD3DUP<{1,0,0,?}, "32">; + +def VLD3DUPd8Pseudo : VLDQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; +def VLD3DUPd16Pseudo : VLDQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; +def VLD3DUPd32Pseudo : VLDQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; + +// ...with double-spaced registers (not used for codegen): +def VLD3DUPq8 : VLD3DUP<{0,0,1,?}, "8">; +def VLD3DUPq16 : VLD3DUP<{0,1,1,?}, "16">; +def VLD3DUPq32 : VLD3DUP<{1,0,1,?}, "32">; + +def VLD3DUPq8EvenPseudo : VLDQQQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; +def VLD3DUPq8OddPseudo : VLDQQQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; +def VLD3DUPq16EvenPseudo : VLDQQQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; +def VLD3DUPq16OddPseudo : VLDQQQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; +def VLD3DUPq32EvenPseudo : VLDQQQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; +def VLD3DUPq32OddPseudo : VLDQQQQPseudo<IIC_VLD3dup>, Sched<[WriteVLD2]>; + +// ...with address register writeback: +class VLD3DUPWB<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<1, 0b10, 0b1110, op7_4, (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, GPR:$wb), + (ins AddrMode:$Rn, am6offset:$Rm), IIC_VLD3dupu, + "vld3", Dt, "\\{$Vd[], $dst2[], $dst3[]\\}, $Rn$Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD2]> { + let Inst{4} = 0; + let DecoderMethod = "DecodeVLD3DupInstruction"; +} + +def VLD3DUPd8_UPD : VLD3DUPWB<{0,0,0,0}, "8", addrmode6dupalign64>; +def VLD3DUPd16_UPD : VLD3DUPWB<{0,1,0,?}, "16", addrmode6dupalign64>; +def VLD3DUPd32_UPD : VLD3DUPWB<{1,0,0,?}, "32", addrmode6dupalign64>; + +def VLD3DUPq8_UPD : VLD3DUPWB<{0,0,1,0}, "8", addrmode6dupalign64>; +def VLD3DUPq16_UPD : VLD3DUPWB<{0,1,1,?}, "16", addrmode6dupalign64>; +def VLD3DUPq32_UPD : VLD3DUPWB<{1,0,1,?}, "32", addrmode6dupalign64>; + +def VLD3DUPd8Pseudo_UPD : VLDQQWBPseudo<IIC_VLD3dupu>, Sched<[WriteVLD2]>; +def VLD3DUPd16Pseudo_UPD : VLDQQWBPseudo<IIC_VLD3dupu>, Sched<[WriteVLD2]>; +def VLD3DUPd32Pseudo_UPD : VLDQQWBPseudo<IIC_VLD3dupu>, Sched<[WriteVLD2]>; + +// VLD4DUP : Vector Load (single 4-element structure to all lanes) +class VLD4DUP<bits<4> op7_4, string Dt> + : NLdSt<1, 0b10, 0b1111, op7_4, + (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4), + (ins addrmode6dup:$Rn), IIC_VLD4dup, + "vld4", Dt, "\\{$Vd[], $dst2[], $dst3[], $dst4[]\\}, $Rn", "", []> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD4DupInstruction"; +} + +def VLD4DUPd8 : VLD4DUP<{0,0,0,?}, "8">; +def VLD4DUPd16 : VLD4DUP<{0,1,0,?}, "16">; +def VLD4DUPd32 : VLD4DUP<{1,?,0,?}, "32"> { let Inst{6} = Rn{5}; } + +def VLD4DUPd8Pseudo : VLDQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; +def VLD4DUPd16Pseudo : VLDQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; +def VLD4DUPd32Pseudo : VLDQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; + +// ...with double-spaced registers (not used for codegen): +def VLD4DUPq8 : VLD4DUP<{0,0,1,?}, "8">; +def VLD4DUPq16 : VLD4DUP<{0,1,1,?}, "16">; +def VLD4DUPq32 : VLD4DUP<{1,?,1,?}, "32"> { let Inst{6} = Rn{5}; } + +def VLD4DUPq8EvenPseudo : VLDQQQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; +def VLD4DUPq8OddPseudo : VLDQQQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; +def VLD4DUPq16EvenPseudo : VLDQQQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; +def VLD4DUPq16OddPseudo : VLDQQQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; +def VLD4DUPq32EvenPseudo : VLDQQQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; +def VLD4DUPq32OddPseudo : VLDQQQQPseudo<IIC_VLD4dup>, Sched<[WriteVLD2]>; + +// ...with address register writeback: +class VLD4DUPWB<bits<4> op7_4, string Dt> + : NLdSt<1, 0b10, 0b1111, op7_4, + (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4, GPR:$wb), + (ins addrmode6dup:$Rn, am6offset:$Rm), IIC_VLD4dupu, + "vld4", Dt, "\\{$Vd[], $dst2[], $dst3[], $dst4[]\\}, $Rn$Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVLD2]> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLD4DupInstruction"; +} + +def VLD4DUPd8_UPD : VLD4DUPWB<{0,0,0,0}, "8">; +def VLD4DUPd16_UPD : VLD4DUPWB<{0,1,0,?}, "16">; +def VLD4DUPd32_UPD : VLD4DUPWB<{1,?,0,?}, "32"> { let Inst{6} = Rn{5}; } + +def VLD4DUPq8_UPD : VLD4DUPWB<{0,0,1,0}, "8">; +def VLD4DUPq16_UPD : VLD4DUPWB<{0,1,1,?}, "16">; +def VLD4DUPq32_UPD : VLD4DUPWB<{1,?,1,?}, "32"> { let Inst{6} = Rn{5}; } + +def VLD4DUPd8Pseudo_UPD : VLDQQWBPseudo<IIC_VLD4dupu>, Sched<[WriteVLD2]>; +def VLD4DUPd16Pseudo_UPD : VLDQQWBPseudo<IIC_VLD4dupu>, Sched<[WriteVLD2]>; +def VLD4DUPd32Pseudo_UPD : VLDQQWBPseudo<IIC_VLD4dupu>, Sched<[WriteVLD2]>; + +} // mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 + +let mayStore = 1, hasSideEffects = 0, hasExtraSrcRegAllocReq = 1 in { + +// Classes for VST* pseudo-instructions with multi-register operands. +// These are expanded to real instructions after register allocation. +class VSTQPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs), (ins addrmode6:$addr, QPR:$src), itin, "">; +class VSTQWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QPR:$src), itin, + "$addr.addr = $wb">; +class VSTQWBfixedPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, QPR:$src), itin, + "$addr.addr = $wb">; +class VSTQWBregisterPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, rGPR:$offset, QPR:$src), itin, + "$addr.addr = $wb">; +class VSTQQPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs), (ins addrmode6:$addr, QQPR:$src), itin, "">; +class VSTQQWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QQPR:$src), itin, + "$addr.addr = $wb">; +class VSTQQWBfixedPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, QQPR:$src), itin, + "$addr.addr = $wb">; +class VSTQQWBregisterPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, rGPR:$offset, QQPR:$src), itin, + "$addr.addr = $wb">; + +class VSTQQQQPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs), (ins addrmode6:$addr, QQQQPR:$src), itin, "">; +class VSTQQQQWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QQQQPR:$src), itin, + "$addr.addr = $wb">; + +// VST1 : Vector Store (multiple single elements) +class VST1D<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<0,0b00,0b0111,op7_4, (outs), (ins AddrMode:$Rn, VecListOneD:$Vd), + IIC_VST1, "vst1", Dt, "$Vd, $Rn", "", []>, Sched<[WriteVST1]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; +} +class VST1Q<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<0,0b00,0b1010,op7_4, (outs), (ins AddrMode:$Rn, VecListDPair:$Vd), + IIC_VST1x2, "vst1", Dt, "$Vd, $Rn", "", []>, Sched<[WriteVST2]> { + let Rm = 0b1111; + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; +} + +def VST1d8 : VST1D<{0,0,0,?}, "8", addrmode6align64>; +def VST1d16 : VST1D<{0,1,0,?}, "16", addrmode6align64>; +def VST1d32 : VST1D<{1,0,0,?}, "32", addrmode6align64>; +def VST1d64 : VST1D<{1,1,0,?}, "64", addrmode6align64>; + +def VST1q8 : VST1Q<{0,0,?,?}, "8", addrmode6align64or128>; +def VST1q16 : VST1Q<{0,1,?,?}, "16", addrmode6align64or128>; +def VST1q32 : VST1Q<{1,0,?,?}, "32", addrmode6align64or128>; +def VST1q64 : VST1Q<{1,1,?,?}, "64", addrmode6align64or128>; + +// ...with address register writeback: +multiclass VST1DWB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0,0b00, 0b0111,op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, VecListOneD:$Vd), IIC_VLD1u, + "vst1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVST1]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } + def _register : NLdSt<0,0b00,0b0111,op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm, VecListOneD:$Vd), + IIC_VLD1u, + "vst1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVST1]> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } +} +multiclass VST1QWB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0,0b00,0b1010,op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, VecListDPair:$Vd), IIC_VLD1x2u, + "vst1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVST2]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } + def _register : NLdSt<0,0b00,0b1010,op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm, VecListDPair:$Vd), + IIC_VLD1x2u, + "vst1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVST2]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } +} + +defm VST1d8wb : VST1DWB<{0,0,0,?}, "8", addrmode6align64>; +defm VST1d16wb : VST1DWB<{0,1,0,?}, "16", addrmode6align64>; +defm VST1d32wb : VST1DWB<{1,0,0,?}, "32", addrmode6align64>; +defm VST1d64wb : VST1DWB<{1,1,0,?}, "64", addrmode6align64>; + +defm VST1q8wb : VST1QWB<{0,0,?,?}, "8", addrmode6align64or128>; +defm VST1q16wb : VST1QWB<{0,1,?,?}, "16", addrmode6align64or128>; +defm VST1q32wb : VST1QWB<{1,0,?,?}, "32", addrmode6align64or128>; +defm VST1q64wb : VST1QWB<{1,1,?,?}, "64", addrmode6align64or128>; + +// ...with 3 registers +class VST1D3<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<0, 0b00, 0b0110, op7_4, (outs), + (ins AddrMode:$Rn, VecListThreeD:$Vd), + IIC_VST1x3, "vst1", Dt, "$Vd, $Rn", "", []>, Sched<[WriteVST3]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST1Instruction"; +} +multiclass VST1D3WB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0,0b00,0b0110,op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, VecListThreeD:$Vd), IIC_VLD1x3u, + "vst1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVST3]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } + def _register : NLdSt<0,0b00,0b0110,op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm, VecListThreeD:$Vd), + IIC_VLD1x3u, + "vst1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVST3]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } +} + +def VST1d8T : VST1D3<{0,0,0,?}, "8", addrmode6align64>; +def VST1d16T : VST1D3<{0,1,0,?}, "16", addrmode6align64>; +def VST1d32T : VST1D3<{1,0,0,?}, "32", addrmode6align64>; +def VST1d64T : VST1D3<{1,1,0,?}, "64", addrmode6align64>; + +defm VST1d8Twb : VST1D3WB<{0,0,0,?}, "8", addrmode6align64>; +defm VST1d16Twb : VST1D3WB<{0,1,0,?}, "16", addrmode6align64>; +defm VST1d32Twb : VST1D3WB<{1,0,0,?}, "32", addrmode6align64>; +defm VST1d64Twb : VST1D3WB<{1,1,0,?}, "64", addrmode6align64>; + +def VST1d8TPseudo : VSTQQPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1d16TPseudo : VSTQQPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1d32TPseudo : VSTQQPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1d64TPseudo : VSTQQPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1d64TPseudoWB_fixed : VSTQQWBfixedPseudo<IIC_VST1x3u>, Sched<[WriteVST3]>; +def VST1d64TPseudoWB_register : VSTQQWBPseudo<IIC_VST1x3u>, Sched<[WriteVST3]>; + +def VST1q8HighTPseudo : VSTQQQQPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1q8LowTPseudo_UPD : VSTQQQQWBPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1q16HighTPseudo : VSTQQQQPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1q16LowTPseudo_UPD : VSTQQQQWBPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1q32HighTPseudo : VSTQQQQPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1q32LowTPseudo_UPD : VSTQQQQWBPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1q64HighTPseudo : VSTQQQQPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; +def VST1q64LowTPseudo_UPD : VSTQQQQWBPseudo<IIC_VST1x3>, Sched<[WriteVST3]>; + +// ...with 4 registers +class VST1D4<bits<4> op7_4, string Dt, Operand AddrMode> + : NLdSt<0, 0b00, 0b0010, op7_4, (outs), + (ins AddrMode:$Rn, VecListFourD:$Vd), + IIC_VST1x4, "vst1", Dt, "$Vd, $Rn", "", + []>, Sched<[WriteVST4]> { + let Rm = 0b1111; + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; +} +multiclass VST1D4WB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0,0b00,0b0010,op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, VecListFourD:$Vd), IIC_VLD1x4u, + "vst1", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVST4]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } + def _register : NLdSt<0,0b00,0b0010,op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm, VecListFourD:$Vd), + IIC_VLD1x4u, + "vst1", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVST4]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST1Instruction"; + } +} + +def VST1d8Q : VST1D4<{0,0,?,?}, "8", addrmode6align64or128or256>; +def VST1d16Q : VST1D4<{0,1,?,?}, "16", addrmode6align64or128or256>; +def VST1d32Q : VST1D4<{1,0,?,?}, "32", addrmode6align64or128or256>; +def VST1d64Q : VST1D4<{1,1,?,?}, "64", addrmode6align64or128or256>; + +defm VST1d8Qwb : VST1D4WB<{0,0,?,?}, "8", addrmode6align64or128or256>; +defm VST1d16Qwb : VST1D4WB<{0,1,?,?}, "16", addrmode6align64or128or256>; +defm VST1d32Qwb : VST1D4WB<{1,0,?,?}, "32", addrmode6align64or128or256>; +defm VST1d64Qwb : VST1D4WB<{1,1,?,?}, "64", addrmode6align64or128or256>; + +def VST1d8QPseudo : VSTQQPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1d16QPseudo : VSTQQPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1d32QPseudo : VSTQQPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1d64QPseudo : VSTQQPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1d64QPseudoWB_fixed : VSTQQWBfixedPseudo<IIC_VST1x4u>, Sched<[WriteVST4]>; +def VST1d64QPseudoWB_register : VSTQQWBPseudo<IIC_VST1x4u>, Sched<[WriteVST4]>; + +def VST1q8HighQPseudo : VSTQQQQPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1q8LowQPseudo_UPD : VSTQQQQWBPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1q16HighQPseudo : VSTQQQQPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1q16LowQPseudo_UPD : VSTQQQQWBPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1q32HighQPseudo : VSTQQQQPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1q32LowQPseudo_UPD : VSTQQQQWBPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1q64HighQPseudo : VSTQQQQPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; +def VST1q64LowQPseudo_UPD : VSTQQQQWBPseudo<IIC_VST1x4>, Sched<[WriteVST4]>; + +// VST2 : Vector Store (multiple 2-element structures) +class VST2<bits<4> op11_8, bits<4> op7_4, string Dt, RegisterOperand VdTy, + InstrItinClass itin, Operand AddrMode> + : NLdSt<0, 0b00, op11_8, op7_4, (outs), (ins AddrMode:$Rn, VdTy:$Vd), + itin, "vst2", Dt, "$Vd, $Rn", "", []> { + let Rm = 0b1111; + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST2Instruction"; +} + +def VST2d8 : VST2<0b1000, {0,0,?,?}, "8", VecListDPair, IIC_VST2, + addrmode6align64or128>, Sched<[WriteVST2]>; +def VST2d16 : VST2<0b1000, {0,1,?,?}, "16", VecListDPair, IIC_VST2, + addrmode6align64or128>, Sched<[WriteVST2]>; +def VST2d32 : VST2<0b1000, {1,0,?,?}, "32", VecListDPair, IIC_VST2, + addrmode6align64or128>, Sched<[WriteVST2]>; + +def VST2q8 : VST2<0b0011, {0,0,?,?}, "8", VecListFourD, IIC_VST2x2, + addrmode6align64or128or256>, Sched<[WriteVST4]>; +def VST2q16 : VST2<0b0011, {0,1,?,?}, "16", VecListFourD, IIC_VST2x2, + addrmode6align64or128or256>, Sched<[WriteVST4]>; +def VST2q32 : VST2<0b0011, {1,0,?,?}, "32", VecListFourD, IIC_VST2x2, + addrmode6align64or128or256>, Sched<[WriteVST4]>; + +def VST2q8Pseudo : VSTQQPseudo<IIC_VST2x2>, Sched<[WriteVST4]>; +def VST2q16Pseudo : VSTQQPseudo<IIC_VST2x2>, Sched<[WriteVST4]>; +def VST2q32Pseudo : VSTQQPseudo<IIC_VST2x2>, Sched<[WriteVST4]>; + +// ...with address register writeback: +multiclass VST2DWB<bits<4> op11_8, bits<4> op7_4, string Dt, + RegisterOperand VdTy, Operand AddrMode> { + def _fixed : NLdSt<0, 0b00, op11_8, op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, VdTy:$Vd), IIC_VLD1u, + "vst2", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVST2]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST2Instruction"; + } + def _register : NLdSt<0, 0b00, op11_8, op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm, VdTy:$Vd), IIC_VLD1u, + "vst2", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVST2]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST2Instruction"; + } +} +multiclass VST2QWB<bits<4> op7_4, string Dt, Operand AddrMode> { + def _fixed : NLdSt<0, 0b00, 0b0011, op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, VecListFourD:$Vd), IIC_VLD1u, + "vst2", Dt, "$Vd, $Rn!", + "$Rn.addr = $wb", []>, Sched<[WriteVST4]> { + let Rm = 0b1101; // NLdSt will assign to the right encoding bits. + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST2Instruction"; + } + def _register : NLdSt<0, 0b00, 0b0011, op7_4, (outs GPR:$wb), + (ins AddrMode:$Rn, rGPR:$Rm, VecListFourD:$Vd), + IIC_VLD1u, + "vst2", Dt, "$Vd, $Rn, $Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVST4]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST2Instruction"; + } +} + +defm VST2d8wb : VST2DWB<0b1000, {0,0,?,?}, "8", VecListDPair, + addrmode6align64or128>; +defm VST2d16wb : VST2DWB<0b1000, {0,1,?,?}, "16", VecListDPair, + addrmode6align64or128>; +defm VST2d32wb : VST2DWB<0b1000, {1,0,?,?}, "32", VecListDPair, + addrmode6align64or128>; + +defm VST2q8wb : VST2QWB<{0,0,?,?}, "8", addrmode6align64or128or256>; +defm VST2q16wb : VST2QWB<{0,1,?,?}, "16", addrmode6align64or128or256>; +defm VST2q32wb : VST2QWB<{1,0,?,?}, "32", addrmode6align64or128or256>; + +def VST2q8PseudoWB_fixed : VSTQQWBfixedPseudo<IIC_VST2x2u>, Sched<[WriteVST4]>; +def VST2q16PseudoWB_fixed : VSTQQWBfixedPseudo<IIC_VST2x2u>, Sched<[WriteVST4]>; +def VST2q32PseudoWB_fixed : VSTQQWBfixedPseudo<IIC_VST2x2u>, Sched<[WriteVST4]>; +def VST2q8PseudoWB_register : VSTQQWBregisterPseudo<IIC_VST2x2u>, Sched<[WriteVST4]>; +def VST2q16PseudoWB_register : VSTQQWBregisterPseudo<IIC_VST2x2u>, Sched<[WriteVST4]>; +def VST2q32PseudoWB_register : VSTQQWBregisterPseudo<IIC_VST2x2u>, Sched<[WriteVST4]>; + +// ...with double-spaced registers +def VST2b8 : VST2<0b1001, {0,0,?,?}, "8", VecListDPairSpaced, IIC_VST2, + addrmode6align64or128>; +def VST2b16 : VST2<0b1001, {0,1,?,?}, "16", VecListDPairSpaced, IIC_VST2, + addrmode6align64or128>; +def VST2b32 : VST2<0b1001, {1,0,?,?}, "32", VecListDPairSpaced, IIC_VST2, + addrmode6align64or128>; +defm VST2b8wb : VST2DWB<0b1001, {0,0,?,?}, "8", VecListDPairSpaced, + addrmode6align64or128>; +defm VST2b16wb : VST2DWB<0b1001, {0,1,?,?}, "16", VecListDPairSpaced, + addrmode6align64or128>; +defm VST2b32wb : VST2DWB<0b1001, {1,0,?,?}, "32", VecListDPairSpaced, + addrmode6align64or128>; + +// VST3 : Vector Store (multiple 3-element structures) +class VST3D<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdSt<0, 0b00, op11_8, op7_4, (outs), + (ins addrmode6:$Rn, DPR:$Vd, DPR:$src2, DPR:$src3), IIC_VST3, + "vst3", Dt, "\\{$Vd, $src2, $src3\\}, $Rn", "", []>, Sched<[WriteVST3]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST3Instruction"; +} + +def VST3d8 : VST3D<0b0100, {0,0,0,?}, "8">; +def VST3d16 : VST3D<0b0100, {0,1,0,?}, "16">; +def VST3d32 : VST3D<0b0100, {1,0,0,?}, "32">; + +def VST3d8Pseudo : VSTQQPseudo<IIC_VST3>, Sched<[WriteVST3]>; +def VST3d16Pseudo : VSTQQPseudo<IIC_VST3>, Sched<[WriteVST3]>; +def VST3d32Pseudo : VSTQQPseudo<IIC_VST3>, Sched<[WriteVST3]>; + +// ...with address register writeback: +class VST3DWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdSt<0, 0b00, op11_8, op7_4, (outs GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$Vd, DPR:$src2, DPR:$src3), IIC_VST3u, + "vst3", Dt, "\\{$Vd, $src2, $src3\\}, $Rn$Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVST3]> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVLDST3Instruction"; +} + +def VST3d8_UPD : VST3DWB<0b0100, {0,0,0,?}, "8">; +def VST3d16_UPD : VST3DWB<0b0100, {0,1,0,?}, "16">; +def VST3d32_UPD : VST3DWB<0b0100, {1,0,0,?}, "32">; + +def VST3d8Pseudo_UPD : VSTQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; +def VST3d16Pseudo_UPD : VSTQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; +def VST3d32Pseudo_UPD : VSTQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; + +// ...with double-spaced registers: +def VST3q8 : VST3D<0b0101, {0,0,0,?}, "8">; +def VST3q16 : VST3D<0b0101, {0,1,0,?}, "16">; +def VST3q32 : VST3D<0b0101, {1,0,0,?}, "32">; +def VST3q8_UPD : VST3DWB<0b0101, {0,0,0,?}, "8">; +def VST3q16_UPD : VST3DWB<0b0101, {0,1,0,?}, "16">; +def VST3q32_UPD : VST3DWB<0b0101, {1,0,0,?}, "32">; + +def VST3q8Pseudo_UPD : VSTQQQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; +def VST3q16Pseudo_UPD : VSTQQQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; +def VST3q32Pseudo_UPD : VSTQQQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; + +// ...alternate versions to be allocated odd register numbers: +def VST3q8oddPseudo : VSTQQQQPseudo<IIC_VST3>, Sched<[WriteVST3]>; +def VST3q16oddPseudo : VSTQQQQPseudo<IIC_VST3>, Sched<[WriteVST3]>; +def VST3q32oddPseudo : VSTQQQQPseudo<IIC_VST3>, Sched<[WriteVST3]>; + +def VST3q8oddPseudo_UPD : VSTQQQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; +def VST3q16oddPseudo_UPD : VSTQQQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; +def VST3q32oddPseudo_UPD : VSTQQQQWBPseudo<IIC_VST3u>, Sched<[WriteVST3]>; + +// VST4 : Vector Store (multiple 4-element structures) +class VST4D<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdSt<0, 0b00, op11_8, op7_4, (outs), + (ins addrmode6:$Rn, DPR:$Vd, DPR:$src2, DPR:$src3, DPR:$src4), + IIC_VST4, "vst4", Dt, "\\{$Vd, $src2, $src3, $src4\\}, $Rn", + "", []>, Sched<[WriteVST4]> { + let Rm = 0b1111; + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST4Instruction"; +} + +def VST4d8 : VST4D<0b0000, {0,0,?,?}, "8">; +def VST4d16 : VST4D<0b0000, {0,1,?,?}, "16">; +def VST4d32 : VST4D<0b0000, {1,0,?,?}, "32">; + +def VST4d8Pseudo : VSTQQPseudo<IIC_VST4>, Sched<[WriteVST4]>; +def VST4d16Pseudo : VSTQQPseudo<IIC_VST4>, Sched<[WriteVST4]>; +def VST4d32Pseudo : VSTQQPseudo<IIC_VST4>, Sched<[WriteVST4]>; + +// ...with address register writeback: +class VST4DWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdSt<0, 0b00, op11_8, op7_4, (outs GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$Vd, DPR:$src2, DPR:$src3, DPR:$src4), IIC_VST4u, + "vst4", Dt, "\\{$Vd, $src2, $src3, $src4\\}, $Rn$Rm", + "$Rn.addr = $wb", []>, Sched<[WriteVST4]> { + let Inst{5-4} = Rn{5-4}; + let DecoderMethod = "DecodeVLDST4Instruction"; +} + +def VST4d8_UPD : VST4DWB<0b0000, {0,0,?,?}, "8">; +def VST4d16_UPD : VST4DWB<0b0000, {0,1,?,?}, "16">; +def VST4d32_UPD : VST4DWB<0b0000, {1,0,?,?}, "32">; + +def VST4d8Pseudo_UPD : VSTQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; +def VST4d16Pseudo_UPD : VSTQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; +def VST4d32Pseudo_UPD : VSTQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; + +// ...with double-spaced registers: +def VST4q8 : VST4D<0b0001, {0,0,?,?}, "8">; +def VST4q16 : VST4D<0b0001, {0,1,?,?}, "16">; +def VST4q32 : VST4D<0b0001, {1,0,?,?}, "32">; +def VST4q8_UPD : VST4DWB<0b0001, {0,0,?,?}, "8">; +def VST4q16_UPD : VST4DWB<0b0001, {0,1,?,?}, "16">; +def VST4q32_UPD : VST4DWB<0b0001, {1,0,?,?}, "32">; + +def VST4q8Pseudo_UPD : VSTQQQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; +def VST4q16Pseudo_UPD : VSTQQQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; +def VST4q32Pseudo_UPD : VSTQQQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; + +// ...alternate versions to be allocated odd register numbers: +def VST4q8oddPseudo : VSTQQQQPseudo<IIC_VST4>, Sched<[WriteVST4]>; +def VST4q16oddPseudo : VSTQQQQPseudo<IIC_VST4>, Sched<[WriteVST4]>; +def VST4q32oddPseudo : VSTQQQQPseudo<IIC_VST4>, Sched<[WriteVST4]>; + +def VST4q8oddPseudo_UPD : VSTQQQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; +def VST4q16oddPseudo_UPD : VSTQQQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; +def VST4q32oddPseudo_UPD : VSTQQQQWBPseudo<IIC_VST4u>, Sched<[WriteVST4]>; + +} // mayStore = 1, hasSideEffects = 0, hasExtraSrcRegAllocReq = 1 + +// Classes for VST*LN pseudo-instructions with multi-register operands. +// These are expanded to real instructions after register allocation. +class VSTQLNPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs), (ins addrmode6:$addr, QPR:$src, nohash_imm:$lane), + itin, "">; +class VSTQLNWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QPR:$src, + nohash_imm:$lane), itin, "$addr.addr = $wb">; +class VSTQQLNPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs), (ins addrmode6:$addr, QQPR:$src, nohash_imm:$lane), + itin, "">; +class VSTQQLNWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QQPR:$src, + nohash_imm:$lane), itin, "$addr.addr = $wb">; +class VSTQQQQLNPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs), (ins addrmode6:$addr, QQQQPR:$src, nohash_imm:$lane), + itin, "">; +class VSTQQQQLNWBPseudo<InstrItinClass itin> + : PseudoNLdSt<(outs GPR:$wb), + (ins addrmode6:$addr, am6offset:$offset, QQQQPR:$src, + nohash_imm:$lane), itin, "$addr.addr = $wb">; + +// VST1LN : Vector Store (single element from one lane) +class VST1LN<bits<4> op11_8, bits<4> op7_4, string Dt, ValueType Ty, + PatFrag StoreOp, SDNode ExtractOp, Operand AddrMode> + : NLdStLn<1, 0b00, op11_8, op7_4, (outs), + (ins AddrMode:$Rn, DPR:$Vd, nohash_imm:$lane), + IIC_VST1ln, "vst1", Dt, "\\{$Vd[$lane]\\}, $Rn", "", + [(StoreOp (ExtractOp (Ty DPR:$Vd), imm:$lane), AddrMode:$Rn)]>, + Sched<[WriteVST1]> { + let Rm = 0b1111; + let DecoderMethod = "DecodeVST1LN"; +} +class VST1QLNPseudo<ValueType Ty, PatFrag StoreOp, SDNode ExtractOp> + : VSTQLNPseudo<IIC_VST1ln>, Sched<[WriteVST1]> { + let Pattern = [(StoreOp (ExtractOp (Ty QPR:$src), imm:$lane), + addrmode6:$addr)]; +} + +def VST1LNd8 : VST1LN<0b0000, {?,?,?,0}, "8", v8i8, truncstorei8, + NEONvgetlaneu, addrmode6> { + let Inst{7-5} = lane{2-0}; +} +def VST1LNd16 : VST1LN<0b0100, {?,?,0,?}, "16", v4i16, truncstorei16, + NEONvgetlaneu, addrmode6> { + let Inst{7-6} = lane{1-0}; + let Inst{4} = Rn{4}; +} + +def VST1LNd32 : VST1LN<0b1000, {?,0,?,?}, "32", v2i32, store, extractelt, + addrmode6oneL32> { + let Inst{7} = lane{0}; + let Inst{5-4} = Rn{5-4}; +} + +def VST1LNq8Pseudo : VST1QLNPseudo<v16i8, truncstorei8, NEONvgetlaneu>; +def VST1LNq16Pseudo : VST1QLNPseudo<v8i16, truncstorei16, NEONvgetlaneu>; +def VST1LNq32Pseudo : VST1QLNPseudo<v4i32, store, extractelt>; + +def : Pat<(store (extractelt (v2f32 DPR:$src), imm:$lane), addrmode6:$addr), + (VST1LNd32 addrmode6:$addr, DPR:$src, imm:$lane)>; +def : Pat<(store (extractelt (v4f32 QPR:$src), imm:$lane), addrmode6:$addr), + (VST1LNq32Pseudo addrmode6:$addr, QPR:$src, imm:$lane)>; + +// ...with address register writeback: +class VST1LNWB<bits<4> op11_8, bits<4> op7_4, string Dt, ValueType Ty, + PatFrag StoreOp, SDNode ExtractOp, Operand AdrMode> + : NLdStLn<1, 0b00, op11_8, op7_4, (outs GPR:$wb), + (ins AdrMode:$Rn, am6offset:$Rm, + DPR:$Vd, nohash_imm:$lane), IIC_VST1lnu, "vst1", Dt, + "\\{$Vd[$lane]\\}, $Rn$Rm", + "$Rn.addr = $wb", + [(set GPR:$wb, (StoreOp (ExtractOp (Ty DPR:$Vd), imm:$lane), + AdrMode:$Rn, am6offset:$Rm))]>, + Sched<[WriteVST1]> { + let DecoderMethod = "DecodeVST1LN"; +} +class VST1QLNWBPseudo<ValueType Ty, PatFrag StoreOp, SDNode ExtractOp> + : VSTQLNWBPseudo<IIC_VST1lnu>, Sched<[WriteVST1]> { + let Pattern = [(set GPR:$wb, (StoreOp (ExtractOp (Ty QPR:$src), imm:$lane), + addrmode6:$addr, am6offset:$offset))]; +} + +def VST1LNd8_UPD : VST1LNWB<0b0000, {?,?,?,0}, "8", v8i8, post_truncsti8, + NEONvgetlaneu, addrmode6> { + let Inst{7-5} = lane{2-0}; +} +def VST1LNd16_UPD : VST1LNWB<0b0100, {?,?,0,?}, "16", v4i16, post_truncsti16, + NEONvgetlaneu, addrmode6> { + let Inst{7-6} = lane{1-0}; + let Inst{4} = Rn{4}; +} +def VST1LNd32_UPD : VST1LNWB<0b1000, {?,0,?,?}, "32", v2i32, post_store, + extractelt, addrmode6oneL32> { + let Inst{7} = lane{0}; + let Inst{5-4} = Rn{5-4}; +} + +def VST1LNq8Pseudo_UPD : VST1QLNWBPseudo<v16i8, post_truncsti8, NEONvgetlaneu>; +def VST1LNq16Pseudo_UPD : VST1QLNWBPseudo<v8i16, post_truncsti16,NEONvgetlaneu>; +def VST1LNq32Pseudo_UPD : VST1QLNWBPseudo<v4i32, post_store, extractelt>; + +let mayStore = 1, hasSideEffects = 0, hasExtraSrcRegAllocReq = 1 in { + +// VST2LN : Vector Store (single 2-element structure from one lane) +class VST2LN<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b00, op11_8, op7_4, (outs), + (ins addrmode6:$Rn, DPR:$Vd, DPR:$src2, nohash_imm:$lane), + IIC_VST2ln, "vst2", Dt, "\\{$Vd[$lane], $src2[$lane]\\}, $Rn", + "", []>, Sched<[WriteVST1]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVST2LN"; +} + +def VST2LNd8 : VST2LN<0b0001, {?,?,?,?}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VST2LNd16 : VST2LN<0b0101, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST2LNd32 : VST2LN<0b1001, {?,0,0,?}, "32"> { + let Inst{7} = lane{0}; +} + +def VST2LNd8Pseudo : VSTQLNPseudo<IIC_VST2ln>, Sched<[WriteVST1]>; +def VST2LNd16Pseudo : VSTQLNPseudo<IIC_VST2ln>, Sched<[WriteVST1]>; +def VST2LNd32Pseudo : VSTQLNPseudo<IIC_VST2ln>, Sched<[WriteVST1]>; + +// ...with double-spaced registers: +def VST2LNq16 : VST2LN<0b0101, {?,?,1,?}, "16"> { + let Inst{7-6} = lane{1-0}; + let Inst{4} = Rn{4}; +} +def VST2LNq32 : VST2LN<0b1001, {?,1,0,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{4} = Rn{4}; +} + +def VST2LNq16Pseudo : VSTQQLNPseudo<IIC_VST2ln>, Sched<[WriteVST1]>; +def VST2LNq32Pseudo : VSTQQLNPseudo<IIC_VST2ln>, Sched<[WriteVST1]>; + +// ...with address register writeback: +class VST2LNWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b00, op11_8, op7_4, (outs GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$Vd, DPR:$src2, nohash_imm:$lane), IIC_VST2lnu, "vst2", Dt, + "\\{$Vd[$lane], $src2[$lane]\\}, $Rn$Rm", + "$Rn.addr = $wb", []> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVST2LN"; +} + +def VST2LNd8_UPD : VST2LNWB<0b0001, {?,?,?,?}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VST2LNd16_UPD : VST2LNWB<0b0101, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST2LNd32_UPD : VST2LNWB<0b1001, {?,0,0,?}, "32"> { + let Inst{7} = lane{0}; +} + +def VST2LNd8Pseudo_UPD : VSTQLNWBPseudo<IIC_VST2lnu>, Sched<[WriteVST1]>; +def VST2LNd16Pseudo_UPD : VSTQLNWBPseudo<IIC_VST2lnu>, Sched<[WriteVST1]>; +def VST2LNd32Pseudo_UPD : VSTQLNWBPseudo<IIC_VST2lnu>, Sched<[WriteVST1]>; + +def VST2LNq16_UPD : VST2LNWB<0b0101, {?,?,1,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST2LNq32_UPD : VST2LNWB<0b1001, {?,1,0,?}, "32"> { + let Inst{7} = lane{0}; +} + +def VST2LNq16Pseudo_UPD : VSTQQLNWBPseudo<IIC_VST2lnu>, Sched<[WriteVST1]>; +def VST2LNq32Pseudo_UPD : VSTQQLNWBPseudo<IIC_VST2lnu>, Sched<[WriteVST1]>; + +// VST3LN : Vector Store (single 3-element structure from one lane) +class VST3LN<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b00, op11_8, op7_4, (outs), + (ins addrmode6:$Rn, DPR:$Vd, DPR:$src2, DPR:$src3, + nohash_imm:$lane), IIC_VST3ln, "vst3", Dt, + "\\{$Vd[$lane], $src2[$lane], $src3[$lane]\\}, $Rn", "", []>, + Sched<[WriteVST2]> { + let Rm = 0b1111; + let DecoderMethod = "DecodeVST3LN"; +} + +def VST3LNd8 : VST3LN<0b0010, {?,?,?,0}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VST3LNd16 : VST3LN<0b0110, {?,?,0,0}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST3LNd32 : VST3LN<0b1010, {?,0,0,0}, "32"> { + let Inst{7} = lane{0}; +} + +def VST3LNd8Pseudo : VSTQQLNPseudo<IIC_VST3ln>, Sched<[WriteVST2]>; +def VST3LNd16Pseudo : VSTQQLNPseudo<IIC_VST3ln>, Sched<[WriteVST2]>; +def VST3LNd32Pseudo : VSTQQLNPseudo<IIC_VST3ln>, Sched<[WriteVST2]>; + +// ...with double-spaced registers: +def VST3LNq16 : VST3LN<0b0110, {?,?,1,0}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST3LNq32 : VST3LN<0b1010, {?,1,0,0}, "32"> { + let Inst{7} = lane{0}; +} + +def VST3LNq16Pseudo : VSTQQQQLNPseudo<IIC_VST3ln>; +def VST3LNq32Pseudo : VSTQQQQLNPseudo<IIC_VST3ln>; + +// ...with address register writeback: +class VST3LNWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b00, op11_8, op7_4, (outs GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$Vd, DPR:$src2, DPR:$src3, nohash_imm:$lane), + IIC_VST3lnu, "vst3", Dt, + "\\{$Vd[$lane], $src2[$lane], $src3[$lane]\\}, $Rn$Rm", + "$Rn.addr = $wb", []> { + let DecoderMethod = "DecodeVST3LN"; +} + +def VST3LNd8_UPD : VST3LNWB<0b0010, {?,?,?,0}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VST3LNd16_UPD : VST3LNWB<0b0110, {?,?,0,0}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST3LNd32_UPD : VST3LNWB<0b1010, {?,0,0,0}, "32"> { + let Inst{7} = lane{0}; +} + +def VST3LNd8Pseudo_UPD : VSTQQLNWBPseudo<IIC_VST3lnu>, Sched<[WriteVST2]>; +def VST3LNd16Pseudo_UPD : VSTQQLNWBPseudo<IIC_VST3lnu>, Sched<[WriteVST2]>; +def VST3LNd32Pseudo_UPD : VSTQQLNWBPseudo<IIC_VST3lnu>, Sched<[WriteVST2]>; + +def VST3LNq16_UPD : VST3LNWB<0b0110, {?,?,1,0}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST3LNq32_UPD : VST3LNWB<0b1010, {?,1,0,0}, "32"> { + let Inst{7} = lane{0}; +} + +def VST3LNq16Pseudo_UPD : VSTQQQQLNWBPseudo<IIC_VST3lnu>, Sched<[WriteVST2]>; +def VST3LNq32Pseudo_UPD : VSTQQQQLNWBPseudo<IIC_VST3lnu>, Sched<[WriteVST2]>; + +// VST4LN : Vector Store (single 4-element structure from one lane) +class VST4LN<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b00, op11_8, op7_4, (outs), + (ins addrmode6:$Rn, DPR:$Vd, DPR:$src2, DPR:$src3, DPR:$src4, + nohash_imm:$lane), IIC_VST4ln, "vst4", Dt, + "\\{$Vd[$lane], $src2[$lane], $src3[$lane], $src4[$lane]\\}, $Rn", + "", []>, Sched<[WriteVST2]> { + let Rm = 0b1111; + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVST4LN"; +} + +def VST4LNd8 : VST4LN<0b0011, {?,?,?,?}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VST4LNd16 : VST4LN<0b0111, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST4LNd32 : VST4LN<0b1011, {?,0,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{5}; +} + +def VST4LNd8Pseudo : VSTQQLNPseudo<IIC_VST4ln>, Sched<[WriteVST2]>; +def VST4LNd16Pseudo : VSTQQLNPseudo<IIC_VST4ln>, Sched<[WriteVST2]>; +def VST4LNd32Pseudo : VSTQQLNPseudo<IIC_VST4ln>, Sched<[WriteVST2]>; + +// ...with double-spaced registers: +def VST4LNq16 : VST4LN<0b0111, {?,?,1,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST4LNq32 : VST4LN<0b1011, {?,1,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{5}; +} + +def VST4LNq16Pseudo : VSTQQQQLNPseudo<IIC_VST4ln>, Sched<[WriteVST2]>; +def VST4LNq32Pseudo : VSTQQQQLNPseudo<IIC_VST4ln>, Sched<[WriteVST2]>; + +// ...with address register writeback: +class VST4LNWB<bits<4> op11_8, bits<4> op7_4, string Dt> + : NLdStLn<1, 0b00, op11_8, op7_4, (outs GPR:$wb), + (ins addrmode6:$Rn, am6offset:$Rm, + DPR:$Vd, DPR:$src2, DPR:$src3, DPR:$src4, nohash_imm:$lane), + IIC_VST4lnu, "vst4", Dt, + "\\{$Vd[$lane], $src2[$lane], $src3[$lane], $src4[$lane]\\}, $Rn$Rm", + "$Rn.addr = $wb", []> { + let Inst{4} = Rn{4}; + let DecoderMethod = "DecodeVST4LN"; +} + +def VST4LNd8_UPD : VST4LNWB<0b0011, {?,?,?,?}, "8"> { + let Inst{7-5} = lane{2-0}; +} +def VST4LNd16_UPD : VST4LNWB<0b0111, {?,?,0,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST4LNd32_UPD : VST4LNWB<0b1011, {?,0,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{5}; +} + +def VST4LNd8Pseudo_UPD : VSTQQLNWBPseudo<IIC_VST4lnu>, Sched<[WriteVST2]>; +def VST4LNd16Pseudo_UPD : VSTQQLNWBPseudo<IIC_VST4lnu>, Sched<[WriteVST2]>; +def VST4LNd32Pseudo_UPD : VSTQQLNWBPseudo<IIC_VST4lnu>, Sched<[WriteVST2]>; + +def VST4LNq16_UPD : VST4LNWB<0b0111, {?,?,1,?}, "16"> { + let Inst{7-6} = lane{1-0}; +} +def VST4LNq32_UPD : VST4LNWB<0b1011, {?,1,?,?}, "32"> { + let Inst{7} = lane{0}; + let Inst{5} = Rn{5}; +} + +def VST4LNq16Pseudo_UPD : VSTQQQQLNWBPseudo<IIC_VST4lnu>, Sched<[WriteVST2]>; +def VST4LNq32Pseudo_UPD : VSTQQQQLNWBPseudo<IIC_VST4lnu>, Sched<[WriteVST2]>; + +} // mayStore = 1, hasSideEffects = 0, hasExtraSrcRegAllocReq = 1 + +// Use vld1/vst1 for unaligned f64 load / store +def : Pat<(f64 (hword_alignedload addrmode6:$addr)), + (VLD1d16 addrmode6:$addr)>, Requires<[IsLE]>; +def : Pat<(hword_alignedstore (f64 DPR:$value), addrmode6:$addr), + (VST1d16 addrmode6:$addr, DPR:$value)>, Requires<[IsLE]>; +def : Pat<(f64 (byte_alignedload addrmode6:$addr)), + (VLD1d8 addrmode6:$addr)>, Requires<[IsLE]>; +def : Pat<(byte_alignedstore (f64 DPR:$value), addrmode6:$addr), + (VST1d8 addrmode6:$addr, DPR:$value)>, Requires<[IsLE]>; +def : Pat<(f64 (non_word_alignedload addrmode6:$addr)), + (VLD1d64 addrmode6:$addr)>, Requires<[IsBE]>; +def : Pat<(non_word_alignedstore (f64 DPR:$value), addrmode6:$addr), + (VST1d64 addrmode6:$addr, DPR:$value)>, Requires<[IsBE]>; + +// Use vld1/vst1 for Q and QQ. Also use them for unaligned v2f64 +// load / store if it's legal. +def : Pat<(v2f64 (dword_alignedload addrmode6:$addr)), + (VLD1q64 addrmode6:$addr)>; +def : Pat<(dword_alignedstore (v2f64 QPR:$value), addrmode6:$addr), + (VST1q64 addrmode6:$addr, QPR:$value)>; +def : Pat<(v2f64 (word_alignedload addrmode6:$addr)), + (VLD1q32 addrmode6:$addr)>, Requires<[IsLE]>; +def : Pat<(word_alignedstore (v2f64 QPR:$value), addrmode6:$addr), + (VST1q32 addrmode6:$addr, QPR:$value)>, Requires<[IsLE]>; +def : Pat<(v2f64 (hword_alignedload addrmode6:$addr)), + (VLD1q16 addrmode6:$addr)>, Requires<[IsLE]>; +def : Pat<(hword_alignedstore (v2f64 QPR:$value), addrmode6:$addr), + (VST1q16 addrmode6:$addr, QPR:$value)>, Requires<[IsLE]>; +def : Pat<(v2f64 (byte_alignedload addrmode6:$addr)), + (VLD1q8 addrmode6:$addr)>, Requires<[IsLE]>; +def : Pat<(byte_alignedstore (v2f64 QPR:$value), addrmode6:$addr), + (VST1q8 addrmode6:$addr, QPR:$value)>, Requires<[IsLE]>; + +//===----------------------------------------------------------------------===// +// NEON pattern fragments +//===----------------------------------------------------------------------===// + +// Extract D sub-registers of Q registers. +def DSubReg_i8_reg : SDNodeXForm<imm, [{ + assert(ARM::dsub_7 == ARM::dsub_0+7 && "Unexpected subreg numbering"); + return CurDAG->getTargetConstant(ARM::dsub_0 + N->getZExtValue()/8, SDLoc(N), + MVT::i32); +}]>; +def DSubReg_i16_reg : SDNodeXForm<imm, [{ + assert(ARM::dsub_7 == ARM::dsub_0+7 && "Unexpected subreg numbering"); + return CurDAG->getTargetConstant(ARM::dsub_0 + N->getZExtValue()/4, SDLoc(N), + MVT::i32); +}]>; +def DSubReg_i32_reg : SDNodeXForm<imm, [{ + assert(ARM::dsub_7 == ARM::dsub_0+7 && "Unexpected subreg numbering"); + return CurDAG->getTargetConstant(ARM::dsub_0 + N->getZExtValue()/2, SDLoc(N), + MVT::i32); +}]>; +def DSubReg_f64_reg : SDNodeXForm<imm, [{ + assert(ARM::dsub_7 == ARM::dsub_0+7 && "Unexpected subreg numbering"); + return CurDAG->getTargetConstant(ARM::dsub_0 + N->getZExtValue(), SDLoc(N), + MVT::i32); +}]>; + +// Extract S sub-registers of Q/D registers. +def SSubReg_f32_reg : SDNodeXForm<imm, [{ + assert(ARM::ssub_3 == ARM::ssub_0+3 && "Unexpected subreg numbering"); + return CurDAG->getTargetConstant(ARM::ssub_0 + N->getZExtValue(), SDLoc(N), + MVT::i32); +}]>; + +// Translate lane numbers from Q registers to D subregs. +def SubReg_i8_lane : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant(N->getZExtValue() & 7, SDLoc(N), MVT::i32); +}]>; +def SubReg_i16_lane : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant(N->getZExtValue() & 3, SDLoc(N), MVT::i32); +}]>; +def SubReg_i32_lane : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant(N->getZExtValue() & 1, SDLoc(N), MVT::i32); +}]>; + +//===----------------------------------------------------------------------===// +// Instruction Classes +//===----------------------------------------------------------------------===// + +// Basic 2-register operations: double- and quad-register. +class N2VD<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op4, string OpcodeStr, + string Dt, ValueType ResTy, ValueType OpTy, SDNode OpNode> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, 0, op4, (outs DPR:$Vd), + (ins DPR:$Vm), IIC_VUNAD, OpcodeStr, Dt,"$Vd, $Vm", "", + [(set DPR:$Vd, (ResTy (OpNode (OpTy DPR:$Vm))))]>; +class N2VQ<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op4, string OpcodeStr, + string Dt, ValueType ResTy, ValueType OpTy, SDNode OpNode> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, 1, op4, (outs QPR:$Vd), + (ins QPR:$Vm), IIC_VUNAQ, OpcodeStr, Dt,"$Vd, $Vm", "", + [(set QPR:$Vd, (ResTy (OpNode (OpTy QPR:$Vm))))]>; + +// Basic 2-register intrinsics, both double- and quad-register. +class N2VDInt<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, 0, op4, (outs DPR:$Vd), + (ins DPR:$Vm), itin, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vm))))]>; +class N2VQInt<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, 1, op4, (outs QPR:$Vd), + (ins QPR:$Vm), itin, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vm))))]>; + +// Same as above, but not predicated. +class N2VDIntnp<bits<2> op19_18, bits<2> op17_16, bits<3> op10_8, bit op7, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2Vnp<op19_18, op17_16, op10_8, op7, 0, (outs DPR:$Vd), (ins DPR:$Vm), + itin, OpcodeStr, Dt, + [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vm))))]>; + +class N2VQIntnp<bits<2> op19_18, bits<2> op17_16, bits<3> op10_8, bit op7, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2Vnp<op19_18, op17_16, op10_8, op7, 1, (outs QPR:$Vd), (ins QPR:$Vm), + itin, OpcodeStr, Dt, + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vm))))]>; + +// Similar to NV2VQIntnp with some more encoding bits exposed (crypto). +class N2VQIntXnp<bits<2> op19_18, bits<2> op17_16, bits<3> op10_8, bit op6, + bit op7, InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2Vnp<op19_18, op17_16, op10_8, op7, op6, (outs QPR:$Vd), (ins QPR:$Vm), + itin, OpcodeStr, Dt, + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vm))))]>; + +// Same as N2VQIntXnp but with Vd as a src register. +class N2VQIntX2np<bits<2> op19_18, bits<2> op17_16, bits<3> op10_8, bit op6, + bit op7, InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2Vnp<op19_18, op17_16, op10_8, op7, op6, + (outs QPR:$Vd), (ins QPR:$src, QPR:$Vm), + itin, OpcodeStr, Dt, + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$src), (OpTy QPR:$Vm))))]> { + let Constraints = "$src = $Vd"; +} + +// Narrow 2-register operations. +class N2VN<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op6, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyD, ValueType TyQ, SDNode OpNode> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, op6, op4, (outs DPR:$Vd), + (ins QPR:$Vm), itin, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set DPR:$Vd, (TyD (OpNode (TyQ QPR:$Vm))))]>; + +// Narrow 2-register intrinsics. +class N2VNInt<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op6, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyD, ValueType TyQ, SDPatternOperator IntOp> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, op6, op4, (outs DPR:$Vd), + (ins QPR:$Vm), itin, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set DPR:$Vd, (TyD (IntOp (TyQ QPR:$Vm))))]>; + +// Long 2-register operations (currently only used for VMOVL). +class N2VL<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op6, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDNode OpNode> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, op6, op4, (outs QPR:$Vd), + (ins DPR:$Vm), itin, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set QPR:$Vd, (TyQ (OpNode (TyD DPR:$Vm))))]>; + +// Long 2-register intrinsics. +class N2VLInt<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op6, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDPatternOperator IntOp> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, op6, op4, (outs QPR:$Vd), + (ins DPR:$Vm), itin, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set QPR:$Vd, (TyQ (IntOp (TyD DPR:$Vm))))]>; + +// 2-register shuffles (VTRN/VZIP/VUZP), both double- and quad-register. +class N2VDShuffle<bits<2> op19_18, bits<5> op11_7, string OpcodeStr, string Dt> + : N2V<0b11, 0b11, op19_18, 0b10, op11_7, 0, 0, (outs DPR:$Vd, DPR:$Vm), + (ins DPR:$src1, DPR:$src2), IIC_VPERMD, + OpcodeStr, Dt, "$Vd, $Vm", + "$src1 = $Vd, $src2 = $Vm", []>; +class N2VQShuffle<bits<2> op19_18, bits<5> op11_7, + InstrItinClass itin, string OpcodeStr, string Dt> + : N2V<0b11, 0b11, op19_18, 0b10, op11_7, 1, 0, (outs QPR:$Vd, QPR:$Vm), + (ins QPR:$src1, QPR:$src2), itin, OpcodeStr, Dt, "$Vd, $Vm", + "$src1 = $Vd, $src2 = $Vm", []>; + +// Basic 3-register operations: double- and quad-register. +class N3VD<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDNode OpNode, bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set DPR:$Vd, (ResTy (OpNode (OpTy DPR:$Vn), (OpTy DPR:$Vm))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = Commutable; +} +// Same as N3VD but no data type. +class N3VDX<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, + ValueType ResTy, ValueType OpTy, + SDNode OpNode, bit Commutable> + : N3VX<op24, op23, op21_20, op11_8, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, "$Vd, $Vn, $Vm", "", + [(set DPR:$Vd, (ResTy (OpNode (OpTy DPR:$Vn), (OpTy DPR:$Vm))))]>{ + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = Commutable; +} + +class N3VDSL<bits<2> op21_20, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType Ty, SDNode ShOp> + : N3VLane32<0, 1, op21_20, op11_8, 1, 0, + (outs DPR:$Vd), (ins DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set (Ty DPR:$Vd), + (Ty (ShOp (Ty DPR:$Vn), + (Ty (NEONvduplane (Ty DPR_VFP2:$Vm),imm:$lane)))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = 0; +} +class N3VDSL16<bits<2> op21_20, bits<4> op11_8, + string OpcodeStr, string Dt, ValueType Ty, SDNode ShOp> + : N3VLane16<0, 1, op21_20, op11_8, 1, 0, + (outs DPR:$Vd), (ins DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, IIC_VMULi16D, OpcodeStr, Dt,"$Vd, $Vn, $Vm$lane","", + [(set (Ty DPR:$Vd), + (Ty (ShOp (Ty DPR:$Vn), + (Ty (NEONvduplane (Ty DPR_8:$Vm), imm:$lane)))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = 0; +} + +class N3VQ<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDNode OpNode, bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vn, QPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (ResTy (OpNode (OpTy QPR:$Vn), (OpTy QPR:$Vm))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = Commutable; +} +class N3VQX<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, + ValueType ResTy, ValueType OpTy, SDNode OpNode, bit Commutable> + : N3VX<op24, op23, op21_20, op11_8, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vn, QPR:$Vm), N3RegFrm, itin, + OpcodeStr, "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (ResTy (OpNode (OpTy QPR:$Vn), (OpTy QPR:$Vm))))]>{ + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = Commutable; +} +class N3VQSL<bits<2> op21_20, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDNode ShOp> + : N3VLane32<1, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), (ins QPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set (ResTy QPR:$Vd), + (ResTy (ShOp (ResTy QPR:$Vn), + (ResTy (NEONvduplane (OpTy DPR_VFP2:$Vm), + imm:$lane)))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = 0; +} +class N3VQSL16<bits<2> op21_20, bits<4> op11_8, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDNode ShOp> + : N3VLane16<1, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), (ins QPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, IIC_VMULi16Q, OpcodeStr, Dt,"$Vd, $Vn, $Vm$lane", "", + [(set (ResTy QPR:$Vd), + (ResTy (ShOp (ResTy QPR:$Vn), + (ResTy (NEONvduplane (OpTy DPR_8:$Vm), + imm:$lane)))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = 0; +} + +// Basic 3-register intrinsics, both double- and quad-register. +class N3VDInt<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + Format f, InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp, bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vn, DPR:$Vm), f, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vn), (OpTy DPR:$Vm))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = Commutable; +} + +class N3VDIntnp<bits<5> op27_23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, Format f, InstrItinClass itin, string OpcodeStr, + string Dt, ValueType ResTy, ValueType OpTy, + SDPatternOperator IntOp, bit Commutable> + : N3Vnp<op27_23, op21_20, op11_8, op6, op4, + (outs DPR:$Vd), (ins DPR:$Vn, DPR:$Vm), N3RegFrm, itin, OpcodeStr, Dt, + [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vn), (OpTy DPR:$Vm))))]>; + +class N3VDIntSL<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, ValueType Ty, SDPatternOperator IntOp> + : N3VLane32<0, 1, op21_20, op11_8, 1, 0, + (outs DPR:$Vd), (ins DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set (Ty DPR:$Vd), + (Ty (IntOp (Ty DPR:$Vn), + (Ty (NEONvduplane (Ty DPR_VFP2:$Vm), + imm:$lane)))))]> { + let isCommutable = 0; +} + +class N3VDIntSL16<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, ValueType Ty, SDPatternOperator IntOp> + : N3VLane16<0, 1, op21_20, op11_8, 1, 0, + (outs DPR:$Vd), (ins DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set (Ty DPR:$Vd), + (Ty (IntOp (Ty DPR:$Vn), + (Ty (NEONvduplane (Ty DPR_8:$Vm), imm:$lane)))))]> { + let isCommutable = 0; +} +class N3VDIntSh<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + Format f, InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vm, DPR:$Vn), f, itin, + OpcodeStr, Dt, "$Vd, $Vm, $Vn", "", + [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vm), (OpTy DPR:$Vn))))]> { + let TwoOperandAliasConstraint = "$Vm = $Vd"; + let isCommutable = 0; +} + +class N3VQInt<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + Format f, InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp, bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vn, QPR:$Vm), f, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vn), (OpTy QPR:$Vm))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = Commutable; +} + +class N3VQIntnp<bits<5> op27_23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, Format f, InstrItinClass itin, string OpcodeStr, + string Dt, ValueType ResTy, ValueType OpTy, + SDPatternOperator IntOp, bit Commutable> + : N3Vnp<op27_23, op21_20, op11_8, op6, op4, + (outs QPR:$Vd), (ins QPR:$Vn, QPR:$Vm), f, itin, OpcodeStr, Dt, + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vn), (OpTy QPR:$Vm))))]>; + +// Same as N3VQIntnp but with Vd as a src register. +class N3VQInt3np<bits<5> op27_23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, Format f, InstrItinClass itin, string OpcodeStr, + string Dt, ValueType ResTy, ValueType OpTy, + SDPatternOperator IntOp, bit Commutable> + : N3Vnp<op27_23, op21_20, op11_8, op6, op4, + (outs QPR:$Vd), (ins QPR:$src, QPR:$Vn, QPR:$Vm), + f, itin, OpcodeStr, Dt, + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$src), (OpTy QPR:$Vn), + (OpTy QPR:$Vm))))]> { + let Constraints = "$src = $Vd"; +} + +class N3VQIntSL<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3VLane32<1, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), (ins QPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set (ResTy QPR:$Vd), + (ResTy (IntOp (ResTy QPR:$Vn), + (ResTy (NEONvduplane (OpTy DPR_VFP2:$Vm), + imm:$lane)))))]> { + let isCommutable = 0; +} +class N3VQIntSL16<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3VLane16<1, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), (ins QPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set (ResTy QPR:$Vd), + (ResTy (IntOp (ResTy QPR:$Vn), + (ResTy (NEONvduplane (OpTy DPR_8:$Vm), + imm:$lane)))))]> { + let isCommutable = 0; +} +class N3VQIntSh<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + Format f, InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3V<op24, op23, op21_20, op11_8, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vm, QPR:$Vn), f, itin, + OpcodeStr, Dt, "$Vd, $Vm, $Vn", "", + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vm), (OpTy QPR:$Vn))))]> { + let TwoOperandAliasConstraint = "$Vm = $Vd"; + let isCommutable = 0; +} + +// Multiply-Add/Sub operations: double- and quad-register. +class N3VDMulOp<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType Ty, SDPatternOperator MulOp, SDPatternOperator OpNode> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs DPR:$Vd), (ins DPR:$src1, DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set DPR:$Vd, (Ty (OpNode DPR:$src1, + (Ty (MulOp DPR:$Vn, DPR:$Vm)))))]>; + +class N3VDMulOpSL<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, + ValueType Ty, SDPatternOperator MulOp, SDPatternOperator ShOp> + : N3VLane32<0, 1, op21_20, op11_8, 1, 0, + (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd", + [(set (Ty DPR:$Vd), + (Ty (ShOp (Ty DPR:$src1), + (Ty (MulOp DPR:$Vn, + (Ty (NEONvduplane (Ty DPR_VFP2:$Vm), + imm:$lane)))))))]>; +class N3VDMulOpSL16<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, + ValueType Ty, SDPatternOperator MulOp, SDPatternOperator ShOp> + : N3VLane16<0, 1, op21_20, op11_8, 1, 0, + (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd", + [(set (Ty DPR:$Vd), + (Ty (ShOp (Ty DPR:$src1), + (Ty (MulOp DPR:$Vn, + (Ty (NEONvduplane (Ty DPR_8:$Vm), + imm:$lane)))))))]>; + +class N3VQMulOp<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, ValueType Ty, + SDPatternOperator MulOp, SDPatternOperator OpNode> + : N3V<op24, op23, op21_20, op11_8, 1, op4, + (outs QPR:$Vd), (ins QPR:$src1, QPR:$Vn, QPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set QPR:$Vd, (Ty (OpNode QPR:$src1, + (Ty (MulOp QPR:$Vn, QPR:$Vm)))))]>; +class N3VQMulOpSL<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, ValueType ResTy, ValueType OpTy, + SDPatternOperator MulOp, SDPatternOperator ShOp> + : N3VLane32<1, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd", + [(set (ResTy QPR:$Vd), + (ResTy (ShOp (ResTy QPR:$src1), + (ResTy (MulOp QPR:$Vn, + (ResTy (NEONvduplane (OpTy DPR_VFP2:$Vm), + imm:$lane)))))))]>; +class N3VQMulOpSL16<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, + SDPatternOperator MulOp, SDPatternOperator ShOp> + : N3VLane16<1, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd", + [(set (ResTy QPR:$Vd), + (ResTy (ShOp (ResTy QPR:$src1), + (ResTy (MulOp QPR:$Vn, + (ResTy (NEONvduplane (OpTy DPR_8:$Vm), + imm:$lane)))))))]>; + +// Neon Intrinsic-Op instructions (VABA): double- and quad-register. +class N3VDIntOp<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType Ty, SDPatternOperator IntOp, SDNode OpNode> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs DPR:$Vd), (ins DPR:$src1, DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set DPR:$Vd, (Ty (OpNode DPR:$src1, + (Ty (IntOp (Ty DPR:$Vn), (Ty DPR:$Vm))))))]>; +class N3VQIntOp<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType Ty, SDPatternOperator IntOp, SDNode OpNode> + : N3V<op24, op23, op21_20, op11_8, 1, op4, + (outs QPR:$Vd), (ins QPR:$src1, QPR:$Vn, QPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set QPR:$Vd, (Ty (OpNode QPR:$src1, + (Ty (IntOp (Ty QPR:$Vn), (Ty QPR:$Vm))))))]>; + +// Neon 3-argument intrinsics, both double- and quad-register. +// The destination register is also used as the first source operand register. +class N3VDInt3<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs DPR:$Vd), (ins DPR:$src1, DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$src1), + (OpTy DPR:$Vn), (OpTy DPR:$Vm))))]>; +class N3VQInt3<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3V<op24, op23, op21_20, op11_8, 1, op4, + (outs QPR:$Vd), (ins QPR:$src1, QPR:$Vn, QPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$src1), + (OpTy QPR:$Vn), (OpTy QPR:$Vm))))]>; + +// Long Multiply-Add/Sub operations. +class N3VLMulOp<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDNode MulOp, SDNode OpNode> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs QPR:$Vd), (ins QPR:$src1, DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set QPR:$Vd, (OpNode (TyQ QPR:$src1), + (TyQ (MulOp (TyD DPR:$Vn), + (TyD DPR:$Vm)))))]>; +class N3VLMulOpSL<bit op24, bits<2> op21_20, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDNode MulOp, SDNode OpNode> + : N3VLane32<op24, 1, op21_20, op11_8, 1, 0, (outs QPR:$Vd), + (ins QPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd", + [(set QPR:$Vd, + (OpNode (TyQ QPR:$src1), + (TyQ (MulOp (TyD DPR:$Vn), + (TyD (NEONvduplane (TyD DPR_VFP2:$Vm), + imm:$lane))))))]>; +class N3VLMulOpSL16<bit op24, bits<2> op21_20, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDNode MulOp, SDNode OpNode> + : N3VLane16<op24, 1, op21_20, op11_8, 1, 0, (outs QPR:$Vd), + (ins QPR:$src1, DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd", + [(set QPR:$Vd, + (OpNode (TyQ QPR:$src1), + (TyQ (MulOp (TyD DPR:$Vn), + (TyD (NEONvduplane (TyD DPR_8:$Vm), + imm:$lane))))))]>; + +// Long Intrinsic-Op vector operations with explicit extend (VABAL). +class N3VLIntExtOp<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDPatternOperator IntOp, SDNode ExtOp, + SDNode OpNode> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs QPR:$Vd), (ins QPR:$src1, DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set QPR:$Vd, (OpNode (TyQ QPR:$src1), + (TyQ (ExtOp (TyD (IntOp (TyD DPR:$Vn), + (TyD DPR:$Vm)))))))]>; + +// Neon Long 3-argument intrinsic. The destination register is +// a quad-register and is also used as the first source operand register. +class N3VLInt3<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDPatternOperator IntOp> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs QPR:$Vd), (ins QPR:$src1, DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set QPR:$Vd, + (TyQ (IntOp (TyQ QPR:$src1), (TyD DPR:$Vn), (TyD DPR:$Vm))))]>; +class N3VLInt3SL<bit op24, bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3VLane32<op24, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), + (ins QPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd", + [(set (ResTy QPR:$Vd), + (ResTy (IntOp (ResTy QPR:$src1), + (OpTy DPR:$Vn), + (OpTy (NEONvduplane (OpTy DPR_VFP2:$Vm), + imm:$lane)))))]>; +class N3VLInt3SL16<bit op24, bits<2> op21_20, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3VLane16<op24, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), + (ins QPR:$src1, DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd", + [(set (ResTy QPR:$Vd), + (ResTy (IntOp (ResTy QPR:$src1), + (OpTy DPR:$Vn), + (OpTy (NEONvduplane (OpTy DPR_8:$Vm), + imm:$lane)))))]>; + +// Narrowing 3-register intrinsics. +class N3VNInt<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + string OpcodeStr, string Dt, ValueType TyD, ValueType TyQ, + SDPatternOperator IntOp, bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs DPR:$Vd), (ins QPR:$Vn, QPR:$Vm), N3RegFrm, IIC_VBINi4D, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set DPR:$Vd, (TyD (IntOp (TyQ QPR:$Vn), (TyQ QPR:$Vm))))]> { + let isCommutable = Commutable; +} + +// Long 3-register operations. +class N3VL<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDNode OpNode, bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs QPR:$Vd), (ins DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (TyQ (OpNode (TyD DPR:$Vn), (TyD DPR:$Vm))))]> { + let isCommutable = Commutable; +} + +class N3VLSL<bit op24, bits<2> op21_20, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDNode OpNode> + : N3VLane32<op24, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), (ins DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set QPR:$Vd, + (TyQ (OpNode (TyD DPR:$Vn), + (TyD (NEONvduplane (TyD DPR_VFP2:$Vm),imm:$lane)))))]>; +class N3VLSL16<bit op24, bits<2> op21_20, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDNode OpNode> + : N3VLane16<op24, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), (ins DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set QPR:$Vd, + (TyQ (OpNode (TyD DPR:$Vn), + (TyD (NEONvduplane (TyD DPR_8:$Vm), imm:$lane)))))]>; + +// Long 3-register operations with explicitly extended operands. +class N3VLExt<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDNode OpNode, SDNode ExtOp, + bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs QPR:$Vd), (ins DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (OpNode (TyQ (ExtOp (TyD DPR:$Vn))), + (TyQ (ExtOp (TyD DPR:$Vm)))))]> { + let isCommutable = Commutable; +} + +// Long 3-register intrinsics with explicit extend (VABDL). +class N3VLIntExt<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDPatternOperator IntOp, SDNode ExtOp, + bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs QPR:$Vd), (ins DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (TyQ (ExtOp (TyD (IntOp (TyD DPR:$Vn), + (TyD DPR:$Vm))))))]> { + let isCommutable = Commutable; +} + +// Long 3-register intrinsics. +class N3VLInt<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType TyQ, ValueType TyD, SDPatternOperator IntOp, bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs QPR:$Vd), (ins DPR:$Vn, DPR:$Vm), N3RegFrm, itin, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (TyQ (IntOp (TyD DPR:$Vn), (TyD DPR:$Vm))))]> { + let isCommutable = Commutable; +} + +// Same as above, but not predicated. +class N3VLIntnp<bits<5> op27_23, bits<2> op21_20, bits<4> op11_8, bit op6, + bit op4, InstrItinClass itin, string OpcodeStr, + string Dt, ValueType ResTy, ValueType OpTy, + SDPatternOperator IntOp, bit Commutable> + : N3Vnp<op27_23, op21_20, op11_8, op6, op4, + (outs QPR:$Vd), (ins DPR:$Vn, DPR:$Vm), N3RegFrm, itin, OpcodeStr, Dt, + [(set QPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vn), (OpTy DPR:$Vm))))]>; + +class N3VLIntSL<bit op24, bits<2> op21_20, bits<4> op11_8, InstrItinClass itin, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3VLane32<op24, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), (ins DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set (ResTy QPR:$Vd), + (ResTy (IntOp (OpTy DPR:$Vn), + (OpTy (NEONvduplane (OpTy DPR_VFP2:$Vm), + imm:$lane)))))]>; +class N3VLIntSL16<bit op24, bits<2> op21_20, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N3VLane16<op24, 1, op21_20, op11_8, 1, 0, + (outs QPR:$Vd), (ins DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane), + NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "", + [(set (ResTy QPR:$Vd), + (ResTy (IntOp (OpTy DPR:$Vn), + (OpTy (NEONvduplane (OpTy DPR_8:$Vm), + imm:$lane)))))]>; + +// Wide 3-register operations. +class N3VW<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4, + string OpcodeStr, string Dt, ValueType TyQ, ValueType TyD, + SDNode OpNode, SDNode ExtOp, bit Commutable> + : N3V<op24, op23, op21_20, op11_8, 0, op4, + (outs QPR:$Vd), (ins QPR:$Vn, DPR:$Vm), N3RegFrm, IIC_VSUBiD, + OpcodeStr, Dt, "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (OpNode (TyQ QPR:$Vn), + (TyQ (ExtOp (TyD DPR:$Vm)))))]> { + // All of these have a two-operand InstAlias. + let TwoOperandAliasConstraint = "$Vn = $Vd"; + let isCommutable = Commutable; +} + +// Pairwise long 2-register intrinsics, both double- and quad-register. +class N2VDPLInt<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op4, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, 0, op4, (outs DPR:$Vd), + (ins DPR:$Vm), IIC_VSHLiD, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vm))))]>; +class N2VQPLInt<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op4, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, 1, op4, (outs QPR:$Vd), + (ins QPR:$Vm), IIC_VSHLiD, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vm))))]>; + +// Pairwise long 2-register accumulate intrinsics, +// both double- and quad-register. +// The destination register is also used as the first source operand register. +class N2VDPLInt2<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op4, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, 0, op4, + (outs DPR:$Vd), (ins DPR:$src1, DPR:$Vm), IIC_VPALiD, + OpcodeStr, Dt, "$Vd, $Vm", "$src1 = $Vd", + [(set DPR:$Vd, (ResTy (IntOp (ResTy DPR:$src1), (OpTy DPR:$Vm))))]>; +class N2VQPLInt2<bits<2> op24_23, bits<2> op21_20, bits<2> op19_18, + bits<2> op17_16, bits<5> op11_7, bit op4, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, SDPatternOperator IntOp> + : N2V<op24_23, op21_20, op19_18, op17_16, op11_7, 1, op4, + (outs QPR:$Vd), (ins QPR:$src1, QPR:$Vm), IIC_VPALiQ, + OpcodeStr, Dt, "$Vd, $Vm", "$src1 = $Vd", + [(set QPR:$Vd, (ResTy (IntOp (ResTy QPR:$src1), (OpTy QPR:$Vm))))]>; + +// Shift by immediate, +// both double- and quad-register. +let TwoOperandAliasConstraint = "$Vm = $Vd" in { +class N2VDSh<bit op24, bit op23, bits<4> op11_8, bit op7, bit op4, + Format f, InstrItinClass itin, Operand ImmTy, + string OpcodeStr, string Dt, ValueType Ty, SDNode OpNode> + : N2VImm<op24, op23, op11_8, op7, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vm, ImmTy:$SIMM), f, itin, + OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "", + [(set DPR:$Vd, (Ty (OpNode (Ty DPR:$Vm), (i32 imm:$SIMM))))]>; +class N2VQSh<bit op24, bit op23, bits<4> op11_8, bit op7, bit op4, + Format f, InstrItinClass itin, Operand ImmTy, + string OpcodeStr, string Dt, ValueType Ty, SDNode OpNode> + : N2VImm<op24, op23, op11_8, op7, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vm, ImmTy:$SIMM), f, itin, + OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "", + [(set QPR:$Vd, (Ty (OpNode (Ty QPR:$Vm), (i32 imm:$SIMM))))]>; +} + +// Long shift by immediate. +class N2VLSh<bit op24, bit op23, bits<4> op11_8, bit op7, bit op6, bit op4, + string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, Operand ImmTy, + SDPatternOperator OpNode> + : N2VImm<op24, op23, op11_8, op7, op6, op4, + (outs QPR:$Vd), (ins DPR:$Vm, ImmTy:$SIMM), N2RegVShLFrm, + IIC_VSHLiD, OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "", + [(set QPR:$Vd, (ResTy (OpNode (OpTy DPR:$Vm), ImmTy:$SIMM)))]>; + +// Narrow shift by immediate. +class N2VNSh<bit op24, bit op23, bits<4> op11_8, bit op7, bit op6, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, Operand ImmTy, + SDPatternOperator OpNode> + : N2VImm<op24, op23, op11_8, op7, op6, op4, + (outs DPR:$Vd), (ins QPR:$Vm, ImmTy:$SIMM), N2RegVShRFrm, itin, + OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "", + [(set DPR:$Vd, (ResTy (OpNode (OpTy QPR:$Vm), + (i32 ImmTy:$SIMM))))]>; + +// Shift right by immediate and accumulate, +// both double- and quad-register. +let TwoOperandAliasConstraint = "$Vm = $Vd" in { +class N2VDShAdd<bit op24, bit op23, bits<4> op11_8, bit op7, bit op4, + Operand ImmTy, string OpcodeStr, string Dt, + ValueType Ty, SDNode ShOp> + : N2VImm<op24, op23, op11_8, op7, 0, op4, (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vm, ImmTy:$SIMM), N2RegVShRFrm, IIC_VPALiD, + OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "$src1 = $Vd", + [(set DPR:$Vd, (Ty (add DPR:$src1, + (Ty (ShOp DPR:$Vm, (i32 imm:$SIMM))))))]>; +class N2VQShAdd<bit op24, bit op23, bits<4> op11_8, bit op7, bit op4, + Operand ImmTy, string OpcodeStr, string Dt, + ValueType Ty, SDNode ShOp> + : N2VImm<op24, op23, op11_8, op7, 1, op4, (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vm, ImmTy:$SIMM), N2RegVShRFrm, IIC_VPALiD, + OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "$src1 = $Vd", + [(set QPR:$Vd, (Ty (add QPR:$src1, + (Ty (ShOp QPR:$Vm, (i32 imm:$SIMM))))))]>; +} + +// Shift by immediate and insert, +// both double- and quad-register. +let TwoOperandAliasConstraint = "$Vm = $Vd" in { +class N2VDShIns<bit op24, bit op23, bits<4> op11_8, bit op7, bit op4, + Operand ImmTy, Format f, string OpcodeStr, string Dt, + ValueType Ty,SDNode ShOp> + : N2VImm<op24, op23, op11_8, op7, 0, op4, (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vm, ImmTy:$SIMM), f, IIC_VSHLiD, + OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "$src1 = $Vd", + [(set DPR:$Vd, (Ty (ShOp DPR:$src1, DPR:$Vm, (i32 imm:$SIMM))))]>; +class N2VQShIns<bit op24, bit op23, bits<4> op11_8, bit op7, bit op4, + Operand ImmTy, Format f, string OpcodeStr, string Dt, + ValueType Ty,SDNode ShOp> + : N2VImm<op24, op23, op11_8, op7, 1, op4, (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vm, ImmTy:$SIMM), f, IIC_VSHLiQ, + OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "$src1 = $Vd", + [(set QPR:$Vd, (Ty (ShOp QPR:$src1, QPR:$Vm, (i32 imm:$SIMM))))]>; +} + +// Convert, with fractional bits immediate, +// both double- and quad-register. +class N2VCvtD<bit op24, bit op23, bits<4> op11_8, bit op7, bit op4, + string OpcodeStr, string Dt, ValueType ResTy, ValueType OpTy, + SDPatternOperator IntOp> + : N2VImm<op24, op23, op11_8, op7, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vm, neon_vcvt_imm32:$SIMM), NVCVTFrm, + IIC_VUNAD, OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "", + [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vm), (i32 imm:$SIMM))))]>; +class N2VCvtQ<bit op24, bit op23, bits<4> op11_8, bit op7, bit op4, + string OpcodeStr, string Dt, ValueType ResTy, ValueType OpTy, + SDPatternOperator IntOp> + : N2VImm<op24, op23, op11_8, op7, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vm, neon_vcvt_imm32:$SIMM), NVCVTFrm, + IIC_VUNAQ, OpcodeStr, Dt, "$Vd, $Vm, $SIMM", "", + [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vm), (i32 imm:$SIMM))))]>; + +//===----------------------------------------------------------------------===// +// Multiclasses +//===----------------------------------------------------------------------===// + +// Abbreviations used in multiclass suffixes: +// Q = quarter int (8 bit) elements +// H = half int (16 bit) elements +// S = single int (32 bit) elements +// D = double int (64 bit) elements + +// Neon 2-register vector operations and intrinsics. + +// Neon 2-register comparisons. +// source operand element sizes of 8, 16 and 32 bits: +multiclass N2V_QHS_cmp<bits<2> op24_23, bits<2> op21_20, bits<2> op17_16, + bits<5> op11_7, bit op4, string opc, string Dt, + string asm, SDNode OpNode> { + // 64-bit vector types. + def v8i8 : N2V<op24_23, op21_20, 0b00, op17_16, op11_7, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vm), NoItinerary, + opc, !strconcat(Dt, "8"), asm, "", + [(set DPR:$Vd, (v8i8 (OpNode (v8i8 DPR:$Vm))))]>; + def v4i16 : N2V<op24_23, op21_20, 0b01, op17_16, op11_7, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vm), NoItinerary, + opc, !strconcat(Dt, "16"), asm, "", + [(set DPR:$Vd, (v4i16 (OpNode (v4i16 DPR:$Vm))))]>; + def v2i32 : N2V<op24_23, op21_20, 0b10, op17_16, op11_7, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vm), NoItinerary, + opc, !strconcat(Dt, "32"), asm, "", + [(set DPR:$Vd, (v2i32 (OpNode (v2i32 DPR:$Vm))))]>; + def v2f32 : N2V<op24_23, op21_20, 0b10, op17_16, op11_7, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vm), NoItinerary, + opc, "f32", asm, "", + [(set DPR:$Vd, (v2i32 (OpNode (v2f32 DPR:$Vm))))]> { + let Inst{10} = 1; // overwrite F = 1 + } + def v4f16 : N2V<op24_23, op21_20, 0b01, op17_16, op11_7, 0, op4, + (outs DPR:$Vd), (ins DPR:$Vm), NoItinerary, + opc, "f16", asm, "", + [(set DPR:$Vd, (v4i16 (OpNode (v4f16 DPR:$Vm))))]>, + Requires<[HasNEON,HasFullFP16]> { + let Inst{10} = 1; // overwrite F = 1 + } + + // 128-bit vector types. + def v16i8 : N2V<op24_23, op21_20, 0b00, op17_16, op11_7, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vm), NoItinerary, + opc, !strconcat(Dt, "8"), asm, "", + [(set QPR:$Vd, (v16i8 (OpNode (v16i8 QPR:$Vm))))]>; + def v8i16 : N2V<op24_23, op21_20, 0b01, op17_16, op11_7, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vm), NoItinerary, + opc, !strconcat(Dt, "16"), asm, "", + [(set QPR:$Vd, (v8i16 (OpNode (v8i16 QPR:$Vm))))]>; + def v4i32 : N2V<op24_23, op21_20, 0b10, op17_16, op11_7, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vm), NoItinerary, + opc, !strconcat(Dt, "32"), asm, "", + [(set QPR:$Vd, (v4i32 (OpNode (v4i32 QPR:$Vm))))]>; + def v4f32 : N2V<op24_23, op21_20, 0b10, op17_16, op11_7, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vm), NoItinerary, + opc, "f32", asm, "", + [(set QPR:$Vd, (v4i32 (OpNode (v4f32 QPR:$Vm))))]> { + let Inst{10} = 1; // overwrite F = 1 + } + def v8f16 : N2V<op24_23, op21_20, 0b01, op17_16, op11_7, 1, op4, + (outs QPR:$Vd), (ins QPR:$Vm), NoItinerary, + opc, "f16", asm, "", + [(set QPR:$Vd, (v8i16 (OpNode (v8f16 QPR:$Vm))))]>, + Requires<[HasNEON,HasFullFP16]> { + let Inst{10} = 1; // overwrite F = 1 + } +} + + +// Neon 2-register vector intrinsics, +// element sizes of 8, 16 and 32 bits: +multiclass N2VInt_QHS<bits<2> op24_23, bits<2> op21_20, bits<2> op17_16, + bits<5> op11_7, bit op4, + InstrItinClass itinD, InstrItinClass itinQ, + string OpcodeStr, string Dt, SDPatternOperator IntOp> { + // 64-bit vector types. + def v8i8 : N2VDInt<op24_23, op21_20, 0b00, op17_16, op11_7, op4, + itinD, OpcodeStr, !strconcat(Dt, "8"), v8i8, v8i8, IntOp>; + def v4i16 : N2VDInt<op24_23, op21_20, 0b01, op17_16, op11_7, op4, + itinD, OpcodeStr, !strconcat(Dt, "16"),v4i16,v4i16,IntOp>; + def v2i32 : N2VDInt<op24_23, op21_20, 0b10, op17_16, op11_7, op4, + itinD, OpcodeStr, !strconcat(Dt, "32"),v2i32,v2i32,IntOp>; + + // 128-bit vector types. + def v16i8 : N2VQInt<op24_23, op21_20, 0b00, op17_16, op11_7, op4, + itinQ, OpcodeStr, !strconcat(Dt, "8"), v16i8,v16i8,IntOp>; + def v8i16 : N2VQInt<op24_23, op21_20, 0b01, op17_16, op11_7, op4, + itinQ, OpcodeStr, !strconcat(Dt, "16"),v8i16,v8i16,IntOp>; + def v4i32 : N2VQInt<op24_23, op21_20, 0b10, op17_16, op11_7, op4, + itinQ, OpcodeStr, !strconcat(Dt, "32"),v4i32,v4i32,IntOp>; +} + + +// Neon Narrowing 2-register vector operations, +// source operand element sizes of 16, 32 and 64 bits: +multiclass N2VN_HSD<bits<2> op24_23, bits<2> op21_20, bits<2> op17_16, + bits<5> op11_7, bit op6, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + SDNode OpNode> { + def v8i8 : N2VN<op24_23, op21_20, 0b00, op17_16, op11_7, op6, op4, + itin, OpcodeStr, !strconcat(Dt, "16"), + v8i8, v8i16, OpNode>; + def v4i16 : N2VN<op24_23, op21_20, 0b01, op17_16, op11_7, op6, op4, + itin, OpcodeStr, !strconcat(Dt, "32"), + v4i16, v4i32, OpNode>; + def v2i32 : N2VN<op24_23, op21_20, 0b10, op17_16, op11_7, op6, op4, + itin, OpcodeStr, !strconcat(Dt, "64"), + v2i32, v2i64, OpNode>; +} + +// Neon Narrowing 2-register vector intrinsics, +// source operand element sizes of 16, 32 and 64 bits: +multiclass N2VNInt_HSD<bits<2> op24_23, bits<2> op21_20, bits<2> op17_16, + bits<5> op11_7, bit op6, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + SDPatternOperator IntOp> { + def v8i8 : N2VNInt<op24_23, op21_20, 0b00, op17_16, op11_7, op6, op4, + itin, OpcodeStr, !strconcat(Dt, "16"), + v8i8, v8i16, IntOp>; + def v4i16 : N2VNInt<op24_23, op21_20, 0b01, op17_16, op11_7, op6, op4, + itin, OpcodeStr, !strconcat(Dt, "32"), + v4i16, v4i32, IntOp>; + def v2i32 : N2VNInt<op24_23, op21_20, 0b10, op17_16, op11_7, op6, op4, + itin, OpcodeStr, !strconcat(Dt, "64"), + v2i32, v2i64, IntOp>; +} + + +// Neon Lengthening 2-register vector intrinsic (currently specific to VMOVL). +// source operand element sizes of 16, 32 and 64 bits: +multiclass N2VL_QHS<bits<2> op24_23, bits<5> op11_7, bit op6, bit op4, + string OpcodeStr, string Dt, SDNode OpNode> { + def v8i16 : N2VL<op24_23, 0b00, 0b10, 0b00, op11_7, op6, op4, IIC_VQUNAiD, + OpcodeStr, !strconcat(Dt, "8"), v8i16, v8i8, OpNode>; + def v4i32 : N2VL<op24_23, 0b01, 0b00, 0b00, op11_7, op6, op4, IIC_VQUNAiD, + OpcodeStr, !strconcat(Dt, "16"), v4i32, v4i16, OpNode>; + def v2i64 : N2VL<op24_23, 0b10, 0b00, 0b00, op11_7, op6, op4, IIC_VQUNAiD, + OpcodeStr, !strconcat(Dt, "32"), v2i64, v2i32, OpNode>; +} + + +// Neon 3-register vector operations. + +// First with only element sizes of 8, 16 and 32 bits: +multiclass N3V_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, + SDNode OpNode, bit Commutable = 0> { + // 64-bit vector types. + def v8i8 : N3VD<op24, op23, 0b00, op11_8, op4, itinD16, + OpcodeStr, !strconcat(Dt, "8"), + v8i8, v8i8, OpNode, Commutable>; + def v4i16 : N3VD<op24, op23, 0b01, op11_8, op4, itinD16, + OpcodeStr, !strconcat(Dt, "16"), + v4i16, v4i16, OpNode, Commutable>; + def v2i32 : N3VD<op24, op23, 0b10, op11_8, op4, itinD32, + OpcodeStr, !strconcat(Dt, "32"), + v2i32, v2i32, OpNode, Commutable>; + + // 128-bit vector types. + def v16i8 : N3VQ<op24, op23, 0b00, op11_8, op4, itinQ16, + OpcodeStr, !strconcat(Dt, "8"), + v16i8, v16i8, OpNode, Commutable>; + def v8i16 : N3VQ<op24, op23, 0b01, op11_8, op4, itinQ16, + OpcodeStr, !strconcat(Dt, "16"), + v8i16, v8i16, OpNode, Commutable>; + def v4i32 : N3VQ<op24, op23, 0b10, op11_8, op4, itinQ32, + OpcodeStr, !strconcat(Dt, "32"), + v4i32, v4i32, OpNode, Commutable>; +} + +multiclass N3VSL_HS<bits<4> op11_8, string OpcodeStr, SDNode ShOp> { + def v4i16 : N3VDSL16<0b01, op11_8, OpcodeStr, "i16", v4i16, ShOp>; + def v2i32 : N3VDSL<0b10, op11_8, IIC_VMULi32D, OpcodeStr, "i32", v2i32, ShOp>; + def v8i16 : N3VQSL16<0b01, op11_8, OpcodeStr, "i16", v8i16, v4i16, ShOp>; + def v4i32 : N3VQSL<0b10, op11_8, IIC_VMULi32Q, OpcodeStr, "i32", + v4i32, v2i32, ShOp>; +} + +// ....then also with element size 64 bits: +multiclass N3V_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itinD, InstrItinClass itinQ, + string OpcodeStr, string Dt, + SDNode OpNode, bit Commutable = 0> + : N3V_QHS<op24, op23, op11_8, op4, itinD, itinD, itinQ, itinQ, + OpcodeStr, Dt, OpNode, Commutable> { + def v1i64 : N3VD<op24, op23, 0b11, op11_8, op4, itinD, + OpcodeStr, !strconcat(Dt, "64"), + v1i64, v1i64, OpNode, Commutable>; + def v2i64 : N3VQ<op24, op23, 0b11, op11_8, op4, itinQ, + OpcodeStr, !strconcat(Dt, "64"), + v2i64, v2i64, OpNode, Commutable>; +} + + +// Neon 3-register vector intrinsics. + +// First with only element sizes of 16 and 32 bits: +multiclass N3VInt_HS<bit op24, bit op23, bits<4> op11_8, bit op4, Format f, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, + SDPatternOperator IntOp, bit Commutable = 0> { + // 64-bit vector types. + def v4i16 : N3VDInt<op24, op23, 0b01, op11_8, op4, f, itinD16, + OpcodeStr, !strconcat(Dt, "16"), + v4i16, v4i16, IntOp, Commutable>; + def v2i32 : N3VDInt<op24, op23, 0b10, op11_8, op4, f, itinD32, + OpcodeStr, !strconcat(Dt, "32"), + v2i32, v2i32, IntOp, Commutable>; + + // 128-bit vector types. + def v8i16 : N3VQInt<op24, op23, 0b01, op11_8, op4, f, itinQ16, + OpcodeStr, !strconcat(Dt, "16"), + v8i16, v8i16, IntOp, Commutable>; + def v4i32 : N3VQInt<op24, op23, 0b10, op11_8, op4, f, itinQ32, + OpcodeStr, !strconcat(Dt, "32"), + v4i32, v4i32, IntOp, Commutable>; +} +multiclass N3VInt_HSSh<bit op24, bit op23, bits<4> op11_8, bit op4, Format f, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, + SDPatternOperator IntOp> { + // 64-bit vector types. + def v4i16 : N3VDIntSh<op24, op23, 0b01, op11_8, op4, f, itinD16, + OpcodeStr, !strconcat(Dt, "16"), + v4i16, v4i16, IntOp>; + def v2i32 : N3VDIntSh<op24, op23, 0b10, op11_8, op4, f, itinD32, + OpcodeStr, !strconcat(Dt, "32"), + v2i32, v2i32, IntOp>; + + // 128-bit vector types. + def v8i16 : N3VQIntSh<op24, op23, 0b01, op11_8, op4, f, itinQ16, + OpcodeStr, !strconcat(Dt, "16"), + v8i16, v8i16, IntOp>; + def v4i32 : N3VQIntSh<op24, op23, 0b10, op11_8, op4, f, itinQ32, + OpcodeStr, !strconcat(Dt, "32"), + v4i32, v4i32, IntOp>; +} + +multiclass N3VIntSL_HS<bits<4> op11_8, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, SDPatternOperator IntOp> { + def v4i16 : N3VDIntSL16<0b01, op11_8, itinD16, + OpcodeStr, !strconcat(Dt, "16"), v4i16, IntOp>; + def v2i32 : N3VDIntSL<0b10, op11_8, itinD32, + OpcodeStr, !strconcat(Dt, "32"), v2i32, IntOp>; + def v8i16 : N3VQIntSL16<0b01, op11_8, itinQ16, + OpcodeStr, !strconcat(Dt, "16"), v8i16, v4i16, IntOp>; + def v4i32 : N3VQIntSL<0b10, op11_8, itinQ32, + OpcodeStr, !strconcat(Dt, "32"), v4i32, v2i32, IntOp>; +} + +// ....then also with element size of 8 bits: +multiclass N3VInt_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, Format f, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, + SDPatternOperator IntOp, bit Commutable = 0> + : N3VInt_HS<op24, op23, op11_8, op4, f, itinD16, itinD32, itinQ16, itinQ32, + OpcodeStr, Dt, IntOp, Commutable> { + def v8i8 : N3VDInt<op24, op23, 0b00, op11_8, op4, f, itinD16, + OpcodeStr, !strconcat(Dt, "8"), + v8i8, v8i8, IntOp, Commutable>; + def v16i8 : N3VQInt<op24, op23, 0b00, op11_8, op4, f, itinQ16, + OpcodeStr, !strconcat(Dt, "8"), + v16i8, v16i8, IntOp, Commutable>; +} +multiclass N3VInt_QHSSh<bit op24, bit op23, bits<4> op11_8, bit op4, Format f, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, + SDPatternOperator IntOp> + : N3VInt_HSSh<op24, op23, op11_8, op4, f, itinD16, itinD32, itinQ16, itinQ32, + OpcodeStr, Dt, IntOp> { + def v8i8 : N3VDIntSh<op24, op23, 0b00, op11_8, op4, f, itinD16, + OpcodeStr, !strconcat(Dt, "8"), + v8i8, v8i8, IntOp>; + def v16i8 : N3VQIntSh<op24, op23, 0b00, op11_8, op4, f, itinQ16, + OpcodeStr, !strconcat(Dt, "8"), + v16i8, v16i8, IntOp>; +} + + +// ....then also with element size of 64 bits: +multiclass N3VInt_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, Format f, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, + SDPatternOperator IntOp, bit Commutable = 0> + : N3VInt_QHS<op24, op23, op11_8, op4, f, itinD16, itinD32, itinQ16, itinQ32, + OpcodeStr, Dt, IntOp, Commutable> { + def v1i64 : N3VDInt<op24, op23, 0b11, op11_8, op4, f, itinD32, + OpcodeStr, !strconcat(Dt, "64"), + v1i64, v1i64, IntOp, Commutable>; + def v2i64 : N3VQInt<op24, op23, 0b11, op11_8, op4, f, itinQ32, + OpcodeStr, !strconcat(Dt, "64"), + v2i64, v2i64, IntOp, Commutable>; +} +multiclass N3VInt_QHSDSh<bit op24, bit op23, bits<4> op11_8, bit op4, Format f, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, + SDPatternOperator IntOp> + : N3VInt_QHSSh<op24, op23, op11_8, op4, f, itinD16, itinD32, itinQ16, itinQ32, + OpcodeStr, Dt, IntOp> { + def v1i64 : N3VDIntSh<op24, op23, 0b11, op11_8, op4, f, itinD32, + OpcodeStr, !strconcat(Dt, "64"), + v1i64, v1i64, IntOp>; + def v2i64 : N3VQIntSh<op24, op23, 0b11, op11_8, op4, f, itinQ32, + OpcodeStr, !strconcat(Dt, "64"), + v2i64, v2i64, IntOp>; +} + +// Neon Narrowing 3-register vector intrinsics, +// source operand element sizes of 16, 32 and 64 bits: +multiclass N3VNInt_HSD<bit op24, bit op23, bits<4> op11_8, bit op4, + string OpcodeStr, string Dt, + SDPatternOperator IntOp, bit Commutable = 0> { + def v8i8 : N3VNInt<op24, op23, 0b00, op11_8, op4, + OpcodeStr, !strconcat(Dt, "16"), + v8i8, v8i16, IntOp, Commutable>; + def v4i16 : N3VNInt<op24, op23, 0b01, op11_8, op4, + OpcodeStr, !strconcat(Dt, "32"), + v4i16, v4i32, IntOp, Commutable>; + def v2i32 : N3VNInt<op24, op23, 0b10, op11_8, op4, + OpcodeStr, !strconcat(Dt, "64"), + v2i32, v2i64, IntOp, Commutable>; +} + + +// Neon Long 3-register vector operations. + +multiclass N3VL_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin16, InstrItinClass itin32, + string OpcodeStr, string Dt, + SDNode OpNode, bit Commutable = 0> { + def v8i16 : N3VL<op24, op23, 0b00, op11_8, op4, itin16, + OpcodeStr, !strconcat(Dt, "8"), + v8i16, v8i8, OpNode, Commutable>; + def v4i32 : N3VL<op24, op23, 0b01, op11_8, op4, itin16, + OpcodeStr, !strconcat(Dt, "16"), + v4i32, v4i16, OpNode, Commutable>; + def v2i64 : N3VL<op24, op23, 0b10, op11_8, op4, itin32, + OpcodeStr, !strconcat(Dt, "32"), + v2i64, v2i32, OpNode, Commutable>; +} + +multiclass N3VLSL_HS<bit op24, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + SDNode OpNode> { + def v4i16 : N3VLSL16<op24, 0b01, op11_8, itin, OpcodeStr, + !strconcat(Dt, "16"), v4i32, v4i16, OpNode>; + def v2i32 : N3VLSL<op24, 0b10, op11_8, itin, OpcodeStr, + !strconcat(Dt, "32"), v2i64, v2i32, OpNode>; +} + +multiclass N3VLExt_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin16, InstrItinClass itin32, + string OpcodeStr, string Dt, + SDNode OpNode, SDNode ExtOp, bit Commutable = 0> { + def v8i16 : N3VLExt<op24, op23, 0b00, op11_8, op4, itin16, + OpcodeStr, !strconcat(Dt, "8"), + v8i16, v8i8, OpNode, ExtOp, Commutable>; + def v4i32 : N3VLExt<op24, op23, 0b01, op11_8, op4, itin16, + OpcodeStr, !strconcat(Dt, "16"), + v4i32, v4i16, OpNode, ExtOp, Commutable>; + def v2i64 : N3VLExt<op24, op23, 0b10, op11_8, op4, itin32, + OpcodeStr, !strconcat(Dt, "32"), + v2i64, v2i32, OpNode, ExtOp, Commutable>; +} + +// Neon Long 3-register vector intrinsics. + +// First with only element sizes of 16 and 32 bits: +multiclass N3VLInt_HS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin16, InstrItinClass itin32, + string OpcodeStr, string Dt, + SDPatternOperator IntOp, bit Commutable = 0> { + def v4i32 : N3VLInt<op24, op23, 0b01, op11_8, op4, itin16, + OpcodeStr, !strconcat(Dt, "16"), + v4i32, v4i16, IntOp, Commutable>; + def v2i64 : N3VLInt<op24, op23, 0b10, op11_8, op4, itin32, + OpcodeStr, !strconcat(Dt, "32"), + v2i64, v2i32, IntOp, Commutable>; +} + +multiclass N3VLIntSL_HS<bit op24, bits<4> op11_8, + InstrItinClass itin, string OpcodeStr, string Dt, + SDPatternOperator IntOp> { + def v4i16 : N3VLIntSL16<op24, 0b01, op11_8, itin, + OpcodeStr, !strconcat(Dt, "16"), v4i32, v4i16, IntOp>; + def v2i32 : N3VLIntSL<op24, 0b10, op11_8, itin, + OpcodeStr, !strconcat(Dt, "32"), v2i64, v2i32, IntOp>; +} + +// ....then also with element size of 8 bits: +multiclass N3VLInt_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin16, InstrItinClass itin32, + string OpcodeStr, string Dt, + SDPatternOperator IntOp, bit Commutable = 0> + : N3VLInt_HS<op24, op23, op11_8, op4, itin16, itin32, OpcodeStr, Dt, + IntOp, Commutable> { + def v8i16 : N3VLInt<op24, op23, 0b00, op11_8, op4, itin16, + OpcodeStr, !strconcat(Dt, "8"), + v8i16, v8i8, IntOp, Commutable>; +} + +// ....with explicit extend (VABDL). +multiclass N3VLIntExt_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + SDPatternOperator IntOp, SDNode ExtOp, bit Commutable = 0> { + def v8i16 : N3VLIntExt<op24, op23, 0b00, op11_8, op4, itin, + OpcodeStr, !strconcat(Dt, "8"), + v8i16, v8i8, IntOp, ExtOp, Commutable>; + def v4i32 : N3VLIntExt<op24, op23, 0b01, op11_8, op4, itin, + OpcodeStr, !strconcat(Dt, "16"), + v4i32, v4i16, IntOp, ExtOp, Commutable>; + def v2i64 : N3VLIntExt<op24, op23, 0b10, op11_8, op4, itin, + OpcodeStr, !strconcat(Dt, "32"), + v2i64, v2i32, IntOp, ExtOp, Commutable>; +} + + +// Neon Wide 3-register vector intrinsics, +// source operand element sizes of 8, 16 and 32 bits: +multiclass N3VW_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + string OpcodeStr, string Dt, + SDNode OpNode, SDNode ExtOp, bit Commutable = 0> { + def v8i16 : N3VW<op24, op23, 0b00, op11_8, op4, + OpcodeStr, !strconcat(Dt, "8"), + v8i16, v8i8, OpNode, ExtOp, Commutable>; + def v4i32 : N3VW<op24, op23, 0b01, op11_8, op4, + OpcodeStr, !strconcat(Dt, "16"), + v4i32, v4i16, OpNode, ExtOp, Commutable>; + def v2i64 : N3VW<op24, op23, 0b10, op11_8, op4, + OpcodeStr, !strconcat(Dt, "32"), + v2i64, v2i32, OpNode, ExtOp, Commutable>; +} + + +// Neon Multiply-Op vector operations, +// element sizes of 8, 16 and 32 bits: +multiclass N3VMulOp_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, SDNode OpNode> { + // 64-bit vector types. + def v8i8 : N3VDMulOp<op24, op23, 0b00, op11_8, op4, itinD16, + OpcodeStr, !strconcat(Dt, "8"), v8i8, mul, OpNode>; + def v4i16 : N3VDMulOp<op24, op23, 0b01, op11_8, op4, itinD16, + OpcodeStr, !strconcat(Dt, "16"), v4i16, mul, OpNode>; + def v2i32 : N3VDMulOp<op24, op23, 0b10, op11_8, op4, itinD32, + OpcodeStr, !strconcat(Dt, "32"), v2i32, mul, OpNode>; + + // 128-bit vector types. + def v16i8 : N3VQMulOp<op24, op23, 0b00, op11_8, op4, itinQ16, + OpcodeStr, !strconcat(Dt, "8"), v16i8, mul, OpNode>; + def v8i16 : N3VQMulOp<op24, op23, 0b01, op11_8, op4, itinQ16, + OpcodeStr, !strconcat(Dt, "16"), v8i16, mul, OpNode>; + def v4i32 : N3VQMulOp<op24, op23, 0b10, op11_8, op4, itinQ32, + OpcodeStr, !strconcat(Dt, "32"), v4i32, mul, OpNode>; +} + +multiclass N3VMulOpSL_HS<bits<4> op11_8, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, SDPatternOperator ShOp> { + def v4i16 : N3VDMulOpSL16<0b01, op11_8, itinD16, + OpcodeStr, !strconcat(Dt, "16"), v4i16, mul, ShOp>; + def v2i32 : N3VDMulOpSL<0b10, op11_8, itinD32, + OpcodeStr, !strconcat(Dt, "32"), v2i32, mul, ShOp>; + def v8i16 : N3VQMulOpSL16<0b01, op11_8, itinQ16, + OpcodeStr, !strconcat(Dt, "16"), v8i16, v4i16, + mul, ShOp>; + def v4i32 : N3VQMulOpSL<0b10, op11_8, itinQ32, + OpcodeStr, !strconcat(Dt, "32"), v4i32, v2i32, + mul, ShOp>; +} + +// Neon Intrinsic-Op vector operations, +// element sizes of 8, 16 and 32 bits: +multiclass N3VIntOp_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itinD, InstrItinClass itinQ, + string OpcodeStr, string Dt, SDPatternOperator IntOp, + SDNode OpNode> { + // 64-bit vector types. + def v8i8 : N3VDIntOp<op24, op23, 0b00, op11_8, op4, itinD, + OpcodeStr, !strconcat(Dt, "8"), v8i8, IntOp, OpNode>; + def v4i16 : N3VDIntOp<op24, op23, 0b01, op11_8, op4, itinD, + OpcodeStr, !strconcat(Dt, "16"), v4i16, IntOp, OpNode>; + def v2i32 : N3VDIntOp<op24, op23, 0b10, op11_8, op4, itinD, + OpcodeStr, !strconcat(Dt, "32"), v2i32, IntOp, OpNode>; + + // 128-bit vector types. + def v16i8 : N3VQIntOp<op24, op23, 0b00, op11_8, op4, itinQ, + OpcodeStr, !strconcat(Dt, "8"), v16i8, IntOp, OpNode>; + def v8i16 : N3VQIntOp<op24, op23, 0b01, op11_8, op4, itinQ, + OpcodeStr, !strconcat(Dt, "16"), v8i16, IntOp, OpNode>; + def v4i32 : N3VQIntOp<op24, op23, 0b10, op11_8, op4, itinQ, + OpcodeStr, !strconcat(Dt, "32"), v4i32, IntOp, OpNode>; +} + +// Neon 3-argument intrinsics, +// element sizes of 16 and 32 bits: +multiclass N3VInt3_HS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, SDPatternOperator IntOp> { + // 64-bit vector types. + def v4i16 : N3VDInt3<op24, op23, 0b01, op11_8, op4, itinD16, + OpcodeStr, !strconcat(Dt, "16"), v4i16, v4i16, IntOp>; + def v2i32 : N3VDInt3<op24, op23, 0b10, op11_8, op4, itinD32, + OpcodeStr, !strconcat(Dt, "32"), v2i32, v2i32, IntOp>; + + // 128-bit vector types. + def v8i16 : N3VQInt3<op24, op23, 0b01, op11_8, op4, itinQ16, + OpcodeStr, !strconcat(Dt, "16"), v8i16, v8i16, IntOp>; + def v4i32 : N3VQInt3<op24, op23, 0b10, op11_8, op4, itinQ32, + OpcodeStr, !strconcat(Dt, "32"), v4i32, v4i32, IntOp>; +} + +// element sizes of 8, 16 and 32 bits: +multiclass N3VInt3_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itinD16, InstrItinClass itinD32, + InstrItinClass itinQ16, InstrItinClass itinQ32, + string OpcodeStr, string Dt, SDPatternOperator IntOp> + :N3VInt3_HS <op24, op23, op11_8, op4, itinD16, itinD32, + itinQ16, itinQ32, OpcodeStr, Dt, IntOp>{ + // 64-bit vector types. + def v8i8 : N3VDInt3<op24, op23, 0b00, op11_8, op4, itinD16, + OpcodeStr, !strconcat(Dt, "8"), v8i8, v8i8, IntOp>; + // 128-bit vector types. + def v16i8 : N3VQInt3<op24, op23, 0b00, op11_8, op4, itinQ16, + OpcodeStr, !strconcat(Dt, "8"), v16i8, v16i8, IntOp>; +} + +// Neon Long Multiply-Op vector operations, +// element sizes of 8, 16 and 32 bits: +multiclass N3VLMulOp_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin16, InstrItinClass itin32, + string OpcodeStr, string Dt, SDNode MulOp, + SDNode OpNode> { + def v8i16 : N3VLMulOp<op24, op23, 0b00, op11_8, op4, itin16, OpcodeStr, + !strconcat(Dt, "8"), v8i16, v8i8, MulOp, OpNode>; + def v4i32 : N3VLMulOp<op24, op23, 0b01, op11_8, op4, itin16, OpcodeStr, + !strconcat(Dt, "16"), v4i32, v4i16, MulOp, OpNode>; + def v2i64 : N3VLMulOp<op24, op23, 0b10, op11_8, op4, itin32, OpcodeStr, + !strconcat(Dt, "32"), v2i64, v2i32, MulOp, OpNode>; +} + +multiclass N3VLMulOpSL_HS<bit op24, bits<4> op11_8, string OpcodeStr, + string Dt, SDNode MulOp, SDNode OpNode> { + def v4i16 : N3VLMulOpSL16<op24, 0b01, op11_8, IIC_VMACi16D, OpcodeStr, + !strconcat(Dt,"16"), v4i32, v4i16, MulOp, OpNode>; + def v2i32 : N3VLMulOpSL<op24, 0b10, op11_8, IIC_VMACi32D, OpcodeStr, + !strconcat(Dt, "32"), v2i64, v2i32, MulOp, OpNode>; +} + + +// Neon Long 3-argument intrinsics. + +// First with only element sizes of 16 and 32 bits: +multiclass N3VLInt3_HS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin16, InstrItinClass itin32, + string OpcodeStr, string Dt, SDPatternOperator IntOp> { + def v4i32 : N3VLInt3<op24, op23, 0b01, op11_8, op4, itin16, + OpcodeStr, !strconcat(Dt, "16"), v4i32, v4i16, IntOp>; + def v2i64 : N3VLInt3<op24, op23, 0b10, op11_8, op4, itin32, + OpcodeStr, !strconcat(Dt, "32"), v2i64, v2i32, IntOp>; +} + +multiclass N3VLInt3SL_HS<bit op24, bits<4> op11_8, + string OpcodeStr, string Dt, SDPatternOperator IntOp> { + def v4i16 : N3VLInt3SL16<op24, 0b01, op11_8, IIC_VMACi16D, + OpcodeStr, !strconcat(Dt,"16"), v4i32, v4i16, IntOp>; + def v2i32 : N3VLInt3SL<op24, 0b10, op11_8, IIC_VMACi32D, + OpcodeStr, !strconcat(Dt, "32"), v2i64, v2i32, IntOp>; +} + +// ....then also with element size of 8 bits: +multiclass N3VLInt3_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin16, InstrItinClass itin32, + string OpcodeStr, string Dt, SDPatternOperator IntOp> + : N3VLInt3_HS<op24, op23, op11_8, op4, itin16, itin32, OpcodeStr, Dt, IntOp> { + def v8i16 : N3VLInt3<op24, op23, 0b00, op11_8, op4, itin16, + OpcodeStr, !strconcat(Dt, "8"), v8i16, v8i8, IntOp>; +} + +// ....with explicit extend (VABAL). +multiclass N3VLIntExtOp_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + SDPatternOperator IntOp, SDNode ExtOp, SDNode OpNode> { + def v8i16 : N3VLIntExtOp<op24, op23, 0b00, op11_8, op4, itin, + OpcodeStr, !strconcat(Dt, "8"), v8i16, v8i8, + IntOp, ExtOp, OpNode>; + def v4i32 : N3VLIntExtOp<op24, op23, 0b01, op11_8, op4, itin, + OpcodeStr, !strconcat(Dt, "16"), v4i32, v4i16, + IntOp, ExtOp, OpNode>; + def v2i64 : N3VLIntExtOp<op24, op23, 0b10, op11_8, op4, itin, + OpcodeStr, !strconcat(Dt, "32"), v2i64, v2i32, + IntOp, ExtOp, OpNode>; +} + + +// Neon Pairwise long 2-register intrinsics, +// element sizes of 8, 16 and 32 bits: +multiclass N2VPLInt_QHS<bits<2> op24_23, bits<2> op21_20, bits<2> op17_16, + bits<5> op11_7, bit op4, + string OpcodeStr, string Dt, SDPatternOperator IntOp> { + // 64-bit vector types. + def v8i8 : N2VDPLInt<op24_23, op21_20, 0b00, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "8"), v4i16, v8i8, IntOp>; + def v4i16 : N2VDPLInt<op24_23, op21_20, 0b01, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "16"), v2i32, v4i16, IntOp>; + def v2i32 : N2VDPLInt<op24_23, op21_20, 0b10, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "32"), v1i64, v2i32, IntOp>; + + // 128-bit vector types. + def v16i8 : N2VQPLInt<op24_23, op21_20, 0b00, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "8"), v8i16, v16i8, IntOp>; + def v8i16 : N2VQPLInt<op24_23, op21_20, 0b01, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "16"), v4i32, v8i16, IntOp>; + def v4i32 : N2VQPLInt<op24_23, op21_20, 0b10, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "32"), v2i64, v4i32, IntOp>; +} + + +// Neon Pairwise long 2-register accumulate intrinsics, +// element sizes of 8, 16 and 32 bits: +multiclass N2VPLInt2_QHS<bits<2> op24_23, bits<2> op21_20, bits<2> op17_16, + bits<5> op11_7, bit op4, + string OpcodeStr, string Dt, SDPatternOperator IntOp> { + // 64-bit vector types. + def v8i8 : N2VDPLInt2<op24_23, op21_20, 0b00, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "8"), v4i16, v8i8, IntOp>; + def v4i16 : N2VDPLInt2<op24_23, op21_20, 0b01, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "16"), v2i32, v4i16, IntOp>; + def v2i32 : N2VDPLInt2<op24_23, op21_20, 0b10, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "32"), v1i64, v2i32, IntOp>; + + // 128-bit vector types. + def v16i8 : N2VQPLInt2<op24_23, op21_20, 0b00, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "8"), v8i16, v16i8, IntOp>; + def v8i16 : N2VQPLInt2<op24_23, op21_20, 0b01, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "16"), v4i32, v8i16, IntOp>; + def v4i32 : N2VQPLInt2<op24_23, op21_20, 0b10, op17_16, op11_7, op4, + OpcodeStr, !strconcat(Dt, "32"), v2i64, v4i32, IntOp>; +} + + +// Neon 2-register vector shift by immediate, +// with f of either N2RegVShLFrm or N2RegVShRFrm +// element sizes of 8, 16, 32 and 64 bits: +multiclass N2VShL_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + SDNode OpNode> { + // 64-bit vector types. + def v8i8 : N2VDSh<op24, op23, op11_8, 0, op4, N2RegVShLFrm, itin, i32imm, + OpcodeStr, !strconcat(Dt, "8"), v8i8, OpNode> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v4i16 : N2VDSh<op24, op23, op11_8, 0, op4, N2RegVShLFrm, itin, i32imm, + OpcodeStr, !strconcat(Dt, "16"), v4i16, OpNode> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v2i32 : N2VDSh<op24, op23, op11_8, 0, op4, N2RegVShLFrm, itin, i32imm, + OpcodeStr, !strconcat(Dt, "32"), v2i32, OpNode> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v1i64 : N2VDSh<op24, op23, op11_8, 1, op4, N2RegVShLFrm, itin, i32imm, + OpcodeStr, !strconcat(Dt, "64"), v1i64, OpNode>; + // imm6 = xxxxxx + + // 128-bit vector types. + def v16i8 : N2VQSh<op24, op23, op11_8, 0, op4, N2RegVShLFrm, itin, i32imm, + OpcodeStr, !strconcat(Dt, "8"), v16i8, OpNode> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v8i16 : N2VQSh<op24, op23, op11_8, 0, op4, N2RegVShLFrm, itin, i32imm, + OpcodeStr, !strconcat(Dt, "16"), v8i16, OpNode> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v4i32 : N2VQSh<op24, op23, op11_8, 0, op4, N2RegVShLFrm, itin, i32imm, + OpcodeStr, !strconcat(Dt, "32"), v4i32, OpNode> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v2i64 : N2VQSh<op24, op23, op11_8, 1, op4, N2RegVShLFrm, itin, i32imm, + OpcodeStr, !strconcat(Dt, "64"), v2i64, OpNode>; + // imm6 = xxxxxx +} +multiclass N2VShR_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, + InstrItinClass itin, string OpcodeStr, string Dt, + string baseOpc, SDNode OpNode> { + // 64-bit vector types. + def v8i8 : N2VDSh<op24, op23, op11_8, 0, op4, N2RegVShRFrm, itin, shr_imm8, + OpcodeStr, !strconcat(Dt, "8"), v8i8, OpNode> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v4i16 : N2VDSh<op24, op23, op11_8, 0, op4, N2RegVShRFrm, itin, shr_imm16, + OpcodeStr, !strconcat(Dt, "16"), v4i16, OpNode> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v2i32 : N2VDSh<op24, op23, op11_8, 0, op4, N2RegVShRFrm, itin, shr_imm32, + OpcodeStr, !strconcat(Dt, "32"), v2i32, OpNode> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v1i64 : N2VDSh<op24, op23, op11_8, 1, op4, N2RegVShRFrm, itin, shr_imm64, + OpcodeStr, !strconcat(Dt, "64"), v1i64, OpNode>; + // imm6 = xxxxxx + + // 128-bit vector types. + def v16i8 : N2VQSh<op24, op23, op11_8, 0, op4, N2RegVShRFrm, itin, shr_imm8, + OpcodeStr, !strconcat(Dt, "8"), v16i8, OpNode> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v8i16 : N2VQSh<op24, op23, op11_8, 0, op4, N2RegVShRFrm, itin, shr_imm16, + OpcodeStr, !strconcat(Dt, "16"), v8i16, OpNode> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v4i32 : N2VQSh<op24, op23, op11_8, 0, op4, N2RegVShRFrm, itin, shr_imm32, + OpcodeStr, !strconcat(Dt, "32"), v4i32, OpNode> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v2i64 : N2VQSh<op24, op23, op11_8, 1, op4, N2RegVShRFrm, itin, shr_imm64, + OpcodeStr, !strconcat(Dt, "64"), v2i64, OpNode>; + // imm6 = xxxxxx +} + +// Neon Shift-Accumulate vector operations, +// element sizes of 8, 16, 32 and 64 bits: +multiclass N2VShAdd_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, + string OpcodeStr, string Dt, SDNode ShOp> { + // 64-bit vector types. + def v8i8 : N2VDShAdd<op24, op23, op11_8, 0, op4, shr_imm8, + OpcodeStr, !strconcat(Dt, "8"), v8i8, ShOp> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v4i16 : N2VDShAdd<op24, op23, op11_8, 0, op4, shr_imm16, + OpcodeStr, !strconcat(Dt, "16"), v4i16, ShOp> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v2i32 : N2VDShAdd<op24, op23, op11_8, 0, op4, shr_imm32, + OpcodeStr, !strconcat(Dt, "32"), v2i32, ShOp> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v1i64 : N2VDShAdd<op24, op23, op11_8, 1, op4, shr_imm64, + OpcodeStr, !strconcat(Dt, "64"), v1i64, ShOp>; + // imm6 = xxxxxx + + // 128-bit vector types. + def v16i8 : N2VQShAdd<op24, op23, op11_8, 0, op4, shr_imm8, + OpcodeStr, !strconcat(Dt, "8"), v16i8, ShOp> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v8i16 : N2VQShAdd<op24, op23, op11_8, 0, op4, shr_imm16, + OpcodeStr, !strconcat(Dt, "16"), v8i16, ShOp> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v4i32 : N2VQShAdd<op24, op23, op11_8, 0, op4, shr_imm32, + OpcodeStr, !strconcat(Dt, "32"), v4i32, ShOp> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v2i64 : N2VQShAdd<op24, op23, op11_8, 1, op4, shr_imm64, + OpcodeStr, !strconcat(Dt, "64"), v2i64, ShOp>; + // imm6 = xxxxxx +} + +// Neon Shift-Insert vector operations, +// with f of either N2RegVShLFrm or N2RegVShRFrm +// element sizes of 8, 16, 32 and 64 bits: +multiclass N2VShInsL_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, + string OpcodeStr> { + // 64-bit vector types. + def v8i8 : N2VDShIns<op24, op23, op11_8, 0, op4, i32imm, + N2RegVShLFrm, OpcodeStr, "8", v8i8, NEONvsli> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v4i16 : N2VDShIns<op24, op23, op11_8, 0, op4, i32imm, + N2RegVShLFrm, OpcodeStr, "16", v4i16, NEONvsli> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v2i32 : N2VDShIns<op24, op23, op11_8, 0, op4, i32imm, + N2RegVShLFrm, OpcodeStr, "32", v2i32, NEONvsli> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v1i64 : N2VDShIns<op24, op23, op11_8, 1, op4, i32imm, + N2RegVShLFrm, OpcodeStr, "64", v1i64, NEONvsli>; + // imm6 = xxxxxx + + // 128-bit vector types. + def v16i8 : N2VQShIns<op24, op23, op11_8, 0, op4, i32imm, + N2RegVShLFrm, OpcodeStr, "8", v16i8, NEONvsli> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v8i16 : N2VQShIns<op24, op23, op11_8, 0, op4, i32imm, + N2RegVShLFrm, OpcodeStr, "16", v8i16, NEONvsli> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v4i32 : N2VQShIns<op24, op23, op11_8, 0, op4, i32imm, + N2RegVShLFrm, OpcodeStr, "32", v4i32, NEONvsli> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v2i64 : N2VQShIns<op24, op23, op11_8, 1, op4, i32imm, + N2RegVShLFrm, OpcodeStr, "64", v2i64, NEONvsli>; + // imm6 = xxxxxx +} +multiclass N2VShInsR_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, + string OpcodeStr> { + // 64-bit vector types. + def v8i8 : N2VDShIns<op24, op23, op11_8, 0, op4, shr_imm8, + N2RegVShRFrm, OpcodeStr, "8", v8i8, NEONvsri> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v4i16 : N2VDShIns<op24, op23, op11_8, 0, op4, shr_imm16, + N2RegVShRFrm, OpcodeStr, "16", v4i16, NEONvsri> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v2i32 : N2VDShIns<op24, op23, op11_8, 0, op4, shr_imm32, + N2RegVShRFrm, OpcodeStr, "32", v2i32, NEONvsri> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v1i64 : N2VDShIns<op24, op23, op11_8, 1, op4, shr_imm64, + N2RegVShRFrm, OpcodeStr, "64", v1i64, NEONvsri>; + // imm6 = xxxxxx + + // 128-bit vector types. + def v16i8 : N2VQShIns<op24, op23, op11_8, 0, op4, shr_imm8, + N2RegVShRFrm, OpcodeStr, "8", v16i8, NEONvsri> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v8i16 : N2VQShIns<op24, op23, op11_8, 0, op4, shr_imm16, + N2RegVShRFrm, OpcodeStr, "16", v8i16, NEONvsri> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v4i32 : N2VQShIns<op24, op23, op11_8, 0, op4, shr_imm32, + N2RegVShRFrm, OpcodeStr, "32", v4i32, NEONvsri> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } + def v2i64 : N2VQShIns<op24, op23, op11_8, 1, op4, shr_imm64, + N2RegVShRFrm, OpcodeStr, "64", v2i64, NEONvsri>; + // imm6 = xxxxxx +} + +// Neon Shift Long operations, +// element sizes of 8, 16, 32 bits: +multiclass N2VLSh_QHS<bit op24, bit op23, bits<4> op11_8, bit op7, bit op6, + bit op4, string OpcodeStr, string Dt, + SDPatternOperator OpNode> { + def v8i16 : N2VLSh<op24, op23, op11_8, op7, op6, op4, + OpcodeStr, !strconcat(Dt, "8"), v8i16, v8i8, imm1_7, OpNode> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v4i32 : N2VLSh<op24, op23, op11_8, op7, op6, op4, + OpcodeStr, !strconcat(Dt, "16"), v4i32, v4i16, imm1_15, OpNode> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v2i64 : N2VLSh<op24, op23, op11_8, op7, op6, op4, + OpcodeStr, !strconcat(Dt, "32"), v2i64, v2i32, imm1_31, OpNode> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } +} + +// Neon Shift Narrow operations, +// element sizes of 16, 32, 64 bits: +multiclass N2VNSh_HSD<bit op24, bit op23, bits<4> op11_8, bit op7, bit op6, + bit op4, InstrItinClass itin, string OpcodeStr, string Dt, + SDPatternOperator OpNode> { + def v8i8 : N2VNSh<op24, op23, op11_8, op7, op6, op4, itin, + OpcodeStr, !strconcat(Dt, "16"), + v8i8, v8i16, shr_imm8, OpNode> { + let Inst{21-19} = 0b001; // imm6 = 001xxx + } + def v4i16 : N2VNSh<op24, op23, op11_8, op7, op6, op4, itin, + OpcodeStr, !strconcat(Dt, "32"), + v4i16, v4i32, shr_imm16, OpNode> { + let Inst{21-20} = 0b01; // imm6 = 01xxxx + } + def v2i32 : N2VNSh<op24, op23, op11_8, op7, op6, op4, itin, + OpcodeStr, !strconcat(Dt, "64"), + v2i32, v2i64, shr_imm32, OpNode> { + let Inst{21} = 0b1; // imm6 = 1xxxxx + } +} + +//===----------------------------------------------------------------------===// +// Instruction Definitions. +//===----------------------------------------------------------------------===// + +// Vector Add Operations. + +// VADD : Vector Add (integer and floating-point) +defm VADD : N3V_QHSD<0, 0, 0b1000, 0, IIC_VBINiD, IIC_VBINiQ, "vadd", "i", + add, 1>; +def VADDfd : N3VD<0, 0, 0b00, 0b1101, 0, IIC_VBIND, "vadd", "f32", + v2f32, v2f32, fadd, 1>; +def VADDfq : N3VQ<0, 0, 0b00, 0b1101, 0, IIC_VBINQ, "vadd", "f32", + v4f32, v4f32, fadd, 1>; +def VADDhd : N3VD<0, 0, 0b01, 0b1101, 0, IIC_VBIND, "vadd", "f16", + v4f16, v4f16, fadd, 1>, + Requires<[HasNEON,HasFullFP16]>; +def VADDhq : N3VQ<0, 0, 0b01, 0b1101, 0, IIC_VBINQ, "vadd", "f16", + v8f16, v8f16, fadd, 1>, + Requires<[HasNEON,HasFullFP16]>; +// VADDL : Vector Add Long (Q = D + D) +defm VADDLs : N3VLExt_QHS<0,1,0b0000,0, IIC_VSHLiD, IIC_VSHLiD, + "vaddl", "s", add, sext, 1>; +defm VADDLu : N3VLExt_QHS<1,1,0b0000,0, IIC_VSHLiD, IIC_VSHLiD, + "vaddl", "u", add, zext, 1>; +// VADDW : Vector Add Wide (Q = Q + D) +defm VADDWs : N3VW_QHS<0,1,0b0001,0, "vaddw", "s", add, sext, 0>; +defm VADDWu : N3VW_QHS<1,1,0b0001,0, "vaddw", "u", add, zext, 0>; +// VHADD : Vector Halving Add +defm VHADDs : N3VInt_QHS<0, 0, 0b0000, 0, N3RegFrm, + IIC_VBINi4D, IIC_VBINi4D, IIC_VBINi4Q, IIC_VBINi4Q, + "vhadd", "s", int_arm_neon_vhadds, 1>; +defm VHADDu : N3VInt_QHS<1, 0, 0b0000, 0, N3RegFrm, + IIC_VBINi4D, IIC_VBINi4D, IIC_VBINi4Q, IIC_VBINi4Q, + "vhadd", "u", int_arm_neon_vhaddu, 1>; +// VRHADD : Vector Rounding Halving Add +defm VRHADDs : N3VInt_QHS<0, 0, 0b0001, 0, N3RegFrm, + IIC_VBINi4D, IIC_VBINi4D, IIC_VBINi4Q, IIC_VBINi4Q, + "vrhadd", "s", int_arm_neon_vrhadds, 1>; +defm VRHADDu : N3VInt_QHS<1, 0, 0b0001, 0, N3RegFrm, + IIC_VBINi4D, IIC_VBINi4D, IIC_VBINi4Q, IIC_VBINi4Q, + "vrhadd", "u", int_arm_neon_vrhaddu, 1>; +// VQADD : Vector Saturating Add +defm VQADDs : N3VInt_QHSD<0, 0, 0b0000, 1, N3RegFrm, + IIC_VBINi4D, IIC_VBINi4D, IIC_VBINi4Q, IIC_VBINi4Q, + "vqadd", "s", int_arm_neon_vqadds, 1>; +defm VQADDu : N3VInt_QHSD<1, 0, 0b0000, 1, N3RegFrm, + IIC_VBINi4D, IIC_VBINi4D, IIC_VBINi4Q, IIC_VBINi4Q, + "vqadd", "u", int_arm_neon_vqaddu, 1>; +// VADDHN : Vector Add and Narrow Returning High Half (D = Q + Q) +defm VADDHN : N3VNInt_HSD<0,1,0b0100,0, "vaddhn", "i", null_frag, 1>; +// VRADDHN : Vector Rounding Add and Narrow Returning High Half (D = Q + Q) +defm VRADDHN : N3VNInt_HSD<1,1,0b0100,0, "vraddhn", "i", + int_arm_neon_vraddhn, 1>; + +def : Pat<(v8i8 (trunc (NEONvshru (add (v8i16 QPR:$Vn), QPR:$Vm), 8))), + (VADDHNv8i8 QPR:$Vn, QPR:$Vm)>; +def : Pat<(v4i16 (trunc (NEONvshru (add (v4i32 QPR:$Vn), QPR:$Vm), 16))), + (VADDHNv4i16 QPR:$Vn, QPR:$Vm)>; +def : Pat<(v2i32 (trunc (NEONvshru (add (v2i64 QPR:$Vn), QPR:$Vm), 32))), + (VADDHNv2i32 QPR:$Vn, QPR:$Vm)>; + +// Vector Multiply Operations. + +// VMUL : Vector Multiply (integer, polynomial and floating-point) +defm VMUL : N3V_QHS<0, 0, 0b1001, 1, IIC_VMULi16D, IIC_VMULi32D, + IIC_VMULi16Q, IIC_VMULi32Q, "vmul", "i", mul, 1>; +def VMULpd : N3VDInt<1, 0, 0b00, 0b1001, 1, N3RegFrm, IIC_VMULi16D, "vmul", + "p8", v8i8, v8i8, int_arm_neon_vmulp, 1>; +def VMULpq : N3VQInt<1, 0, 0b00, 0b1001, 1, N3RegFrm, IIC_VMULi16Q, "vmul", + "p8", v16i8, v16i8, int_arm_neon_vmulp, 1>; +def VMULfd : N3VD<1, 0, 0b00, 0b1101, 1, IIC_VFMULD, "vmul", "f32", + v2f32, v2f32, fmul, 1>; +def VMULfq : N3VQ<1, 0, 0b00, 0b1101, 1, IIC_VFMULQ, "vmul", "f32", + v4f32, v4f32, fmul, 1>; +def VMULhd : N3VD<1, 0, 0b01, 0b1101, 1, IIC_VFMULD, "vmul", "f16", + v4f16, v4f16, fmul, 1>, + Requires<[HasNEON,HasFullFP16]>; +def VMULhq : N3VQ<1, 0, 0b01, 0b1101, 1, IIC_VFMULQ, "vmul", "f16", + v8f16, v8f16, fmul, 1>, + Requires<[HasNEON,HasFullFP16]>; +defm VMULsl : N3VSL_HS<0b1000, "vmul", mul>; +def VMULslfd : N3VDSL<0b10, 0b1001, IIC_VBIND, "vmul", "f32", v2f32, fmul>; +def VMULslfq : N3VQSL<0b10, 0b1001, IIC_VBINQ, "vmul", "f32", v4f32, + v2f32, fmul>; +def VMULslhd : N3VDSL16<0b01, 0b1001, "vmul", "f16", v4f16, fmul>, + Requires<[HasNEON,HasFullFP16]>; +def VMULslhq : N3VQSL16<0b01, 0b1001, "vmul", "f16", v8f16, + v4f16, fmul>, + Requires<[HasNEON,HasFullFP16]>; + +def : Pat<(v8i16 (mul (v8i16 QPR:$src1), + (v8i16 (NEONvduplane (v8i16 QPR:$src2), imm:$lane)))), + (v8i16 (VMULslv8i16 (v8i16 QPR:$src1), + (v4i16 (EXTRACT_SUBREG QPR:$src2, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane)))>; +def : Pat<(v4i32 (mul (v4i32 QPR:$src1), + (v4i32 (NEONvduplane (v4i32 QPR:$src2), imm:$lane)))), + (v4i32 (VMULslv4i32 (v4i32 QPR:$src1), + (v2i32 (EXTRACT_SUBREG QPR:$src2, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; +def : Pat<(v4f32 (fmul (v4f32 QPR:$src1), + (v4f32 (NEONvduplane (v4f32 QPR:$src2), imm:$lane)))), + (v4f32 (VMULslfq (v4f32 QPR:$src1), + (v2f32 (EXTRACT_SUBREG QPR:$src2, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; + + +def : Pat<(v2f32 (fmul DPR:$Rn, (NEONvdup (f32 SPR:$Rm)))), + (VMULslfd DPR:$Rn, + (INSERT_SUBREG (v2f32 (IMPLICIT_DEF)), SPR:$Rm, ssub_0), + (i32 0))>; +def : Pat<(v4f32 (fmul QPR:$Rn, (NEONvdup (f32 SPR:$Rm)))), + (VMULslfq QPR:$Rn, + (INSERT_SUBREG (v2f32 (IMPLICIT_DEF)), SPR:$Rm, ssub_0), + (i32 0))>; + + +// VQDMULH : Vector Saturating Doubling Multiply Returning High Half +defm VQDMULH : N3VInt_HS<0, 0, 0b1011, 0, N3RegFrm, IIC_VMULi16D, IIC_VMULi32D, + IIC_VMULi16Q, IIC_VMULi32Q, + "vqdmulh", "s", int_arm_neon_vqdmulh, 1>; +defm VQDMULHsl: N3VIntSL_HS<0b1100, IIC_VMULi16D, IIC_VMULi32D, + IIC_VMULi16Q, IIC_VMULi32Q, + "vqdmulh", "s", int_arm_neon_vqdmulh>; +def : Pat<(v8i16 (int_arm_neon_vqdmulh (v8i16 QPR:$src1), + (v8i16 (NEONvduplane (v8i16 QPR:$src2), + imm:$lane)))), + (v8i16 (VQDMULHslv8i16 (v8i16 QPR:$src1), + (v4i16 (EXTRACT_SUBREG QPR:$src2, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane)))>; +def : Pat<(v4i32 (int_arm_neon_vqdmulh (v4i32 QPR:$src1), + (v4i32 (NEONvduplane (v4i32 QPR:$src2), + imm:$lane)))), + (v4i32 (VQDMULHslv4i32 (v4i32 QPR:$src1), + (v2i32 (EXTRACT_SUBREG QPR:$src2, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; + +// VQRDMULH : Vector Rounding Saturating Doubling Multiply Returning High Half +defm VQRDMULH : N3VInt_HS<1, 0, 0b1011, 0, N3RegFrm, + IIC_VMULi16D,IIC_VMULi32D,IIC_VMULi16Q,IIC_VMULi32Q, + "vqrdmulh", "s", int_arm_neon_vqrdmulh, 1>; +defm VQRDMULHsl : N3VIntSL_HS<0b1101, IIC_VMULi16D, IIC_VMULi32D, + IIC_VMULi16Q, IIC_VMULi32Q, + "vqrdmulh", "s", int_arm_neon_vqrdmulh>; +def : Pat<(v8i16 (int_arm_neon_vqrdmulh (v8i16 QPR:$src1), + (v8i16 (NEONvduplane (v8i16 QPR:$src2), + imm:$lane)))), + (v8i16 (VQRDMULHslv8i16 (v8i16 QPR:$src1), + (v4i16 (EXTRACT_SUBREG QPR:$src2, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane)))>; +def : Pat<(v4i32 (int_arm_neon_vqrdmulh (v4i32 QPR:$src1), + (v4i32 (NEONvduplane (v4i32 QPR:$src2), + imm:$lane)))), + (v4i32 (VQRDMULHslv4i32 (v4i32 QPR:$src1), + (v2i32 (EXTRACT_SUBREG QPR:$src2, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; + +// VMULL : Vector Multiply Long (integer and polynomial) (Q = D * D) +let PostEncoderMethod = "NEONThumb2DataIPostEncoder", + DecoderNamespace = "NEONData" in { + defm VMULLs : N3VL_QHS<0,1,0b1100,0, IIC_VMULi16D, IIC_VMULi32D, + "vmull", "s", NEONvmulls, 1>; + defm VMULLu : N3VL_QHS<1,1,0b1100,0, IIC_VMULi16D, IIC_VMULi32D, + "vmull", "u", NEONvmullu, 1>; + def VMULLp8 : N3VLInt<0, 1, 0b00, 0b1110, 0, IIC_VMULi16D, "vmull", "p8", + v8i16, v8i8, int_arm_neon_vmullp, 1>; + def VMULLp64 : N3VLIntnp<0b00101, 0b10, 0b1110, 0, 0, NoItinerary, + "vmull", "p64", v2i64, v1i64, int_arm_neon_vmullp, 1>, + Requires<[HasV8, HasCrypto]>; +} +defm VMULLsls : N3VLSL_HS<0, 0b1010, IIC_VMULi16D, "vmull", "s", NEONvmulls>; +defm VMULLslu : N3VLSL_HS<1, 0b1010, IIC_VMULi16D, "vmull", "u", NEONvmullu>; + +// VQDMULL : Vector Saturating Doubling Multiply Long (Q = D * D) +defm VQDMULL : N3VLInt_HS<0,1,0b1101,0, IIC_VMULi16D, IIC_VMULi32D, + "vqdmull", "s", int_arm_neon_vqdmull, 1>; +defm VQDMULLsl: N3VLIntSL_HS<0, 0b1011, IIC_VMULi16D, + "vqdmull", "s", int_arm_neon_vqdmull>; + +// Vector Multiply-Accumulate and Multiply-Subtract Operations. + +// VMLA : Vector Multiply Accumulate (integer and floating-point) +defm VMLA : N3VMulOp_QHS<0, 0, 0b1001, 0, IIC_VMACi16D, IIC_VMACi32D, + IIC_VMACi16Q, IIC_VMACi32Q, "vmla", "i", add>; +def VMLAfd : N3VDMulOp<0, 0, 0b00, 0b1101, 1, IIC_VMACD, "vmla", "f32", + v2f32, fmul_su, fadd_mlx>, + Requires<[HasNEON, UseFPVMLx, DontUseFusedMAC]>; +def VMLAfq : N3VQMulOp<0, 0, 0b00, 0b1101, 1, IIC_VMACQ, "vmla", "f32", + v4f32, fmul_su, fadd_mlx>, + Requires<[HasNEON, UseFPVMLx, DontUseFusedMAC]>; +def VMLAhd : N3VDMulOp<0, 0, 0b01, 0b1101, 1, IIC_VMACD, "vmla", "f16", + v4f16, fmul_su, fadd_mlx>, + Requires<[HasNEON, HasFullFP16, UseFPVMLx, DontUseFusedMAC]>; +def VMLAhq : N3VQMulOp<0, 0, 0b01, 0b1101, 1, IIC_VMACQ, "vmla", "f16", + v8f16, fmul_su, fadd_mlx>, + Requires<[HasNEON, HasFullFP16, UseFPVMLx, DontUseFusedMAC]>; +defm VMLAsl : N3VMulOpSL_HS<0b0000, IIC_VMACi16D, IIC_VMACi32D, + IIC_VMACi16Q, IIC_VMACi32Q, "vmla", "i", add>; +def VMLAslfd : N3VDMulOpSL<0b10, 0b0001, IIC_VMACD, "vmla", "f32", + v2f32, fmul_su, fadd_mlx>, + Requires<[HasNEON, UseFPVMLx]>; +def VMLAslfq : N3VQMulOpSL<0b10, 0b0001, IIC_VMACQ, "vmla", "f32", + v4f32, v2f32, fmul_su, fadd_mlx>, + Requires<[HasNEON, UseFPVMLx]>; +def VMLAslhd : N3VDMulOpSL16<0b01, 0b0001, IIC_VMACD, "vmla", "f16", + v4f16, fmul, fadd>, + Requires<[HasNEON, HasFullFP16, UseFPVMLx]>; +def VMLAslhq : N3VQMulOpSL16<0b01, 0b0001, IIC_VMACQ, "vmla", "f16", + v8f16, v4f16, fmul, fadd>, + Requires<[HasNEON, HasFullFP16, UseFPVMLx]>; + +def : Pat<(v8i16 (add (v8i16 QPR:$src1), + (mul (v8i16 QPR:$src2), + (v8i16 (NEONvduplane (v8i16 QPR:$src3), imm:$lane))))), + (v8i16 (VMLAslv8i16 (v8i16 QPR:$src1), (v8i16 QPR:$src2), + (v4i16 (EXTRACT_SUBREG QPR:$src3, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane)))>; + +def : Pat<(v4i32 (add (v4i32 QPR:$src1), + (mul (v4i32 QPR:$src2), + (v4i32 (NEONvduplane (v4i32 QPR:$src3), imm:$lane))))), + (v4i32 (VMLAslv4i32 (v4i32 QPR:$src1), (v4i32 QPR:$src2), + (v2i32 (EXTRACT_SUBREG QPR:$src3, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; + +def : Pat<(v4f32 (fadd_mlx (v4f32 QPR:$src1), + (fmul_su (v4f32 QPR:$src2), + (v4f32 (NEONvduplane (v4f32 QPR:$src3), imm:$lane))))), + (v4f32 (VMLAslfq (v4f32 QPR:$src1), + (v4f32 QPR:$src2), + (v2f32 (EXTRACT_SUBREG QPR:$src3, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>, + Requires<[HasNEON, UseFPVMLx]>; + +// VMLAL : Vector Multiply Accumulate Long (Q += D * D) +defm VMLALs : N3VLMulOp_QHS<0,1,0b1000,0, IIC_VMACi16D, IIC_VMACi32D, + "vmlal", "s", NEONvmulls, add>; +defm VMLALu : N3VLMulOp_QHS<1,1,0b1000,0, IIC_VMACi16D, IIC_VMACi32D, + "vmlal", "u", NEONvmullu, add>; + +defm VMLALsls : N3VLMulOpSL_HS<0, 0b0010, "vmlal", "s", NEONvmulls, add>; +defm VMLALslu : N3VLMulOpSL_HS<1, 0b0010, "vmlal", "u", NEONvmullu, add>; + +let Predicates = [HasNEON, HasV8_1a] in { + // v8.1a Neon Rounding Double Multiply-Op vector operations, + // VQRDMLAH : Vector Saturating Rounding Doubling Multiply Accumulate Long + // (Q += D * D) + defm VQRDMLAH : N3VInt3_HS<1, 0, 0b1011, 1, IIC_VMACi16D, IIC_VMACi32D, + IIC_VMACi16Q, IIC_VMACi32Q, "vqrdmlah", "s", + null_frag>; + def : Pat<(v4i16 (int_arm_neon_vqadds + (v4i16 DPR:$src1), + (v4i16 (int_arm_neon_vqrdmulh (v4i16 DPR:$Vn), + (v4i16 DPR:$Vm))))), + (v4i16 (VQRDMLAHv4i16 DPR:$src1, DPR:$Vn, DPR:$Vm))>; + def : Pat<(v2i32 (int_arm_neon_vqadds + (v2i32 DPR:$src1), + (v2i32 (int_arm_neon_vqrdmulh (v2i32 DPR:$Vn), + (v2i32 DPR:$Vm))))), + (v2i32 (VQRDMLAHv2i32 DPR:$src1, DPR:$Vn, DPR:$Vm))>; + def : Pat<(v8i16 (int_arm_neon_vqadds + (v8i16 QPR:$src1), + (v8i16 (int_arm_neon_vqrdmulh (v8i16 QPR:$Vn), + (v8i16 QPR:$Vm))))), + (v8i16 (VQRDMLAHv8i16 QPR:$src1, QPR:$Vn, QPR:$Vm))>; + def : Pat<(v4i32 (int_arm_neon_vqadds + (v4i32 QPR:$src1), + (v4i32 (int_arm_neon_vqrdmulh (v4i32 QPR:$Vn), + (v4i32 QPR:$Vm))))), + (v4i32 (VQRDMLAHv4i32 QPR:$src1, QPR:$Vn, QPR:$Vm))>; + + defm VQRDMLAHsl : N3VMulOpSL_HS<0b1110, IIC_VMACi16D, IIC_VMACi32D, + IIC_VMACi16Q, IIC_VMACi32Q, "vqrdmlah", "s", + null_frag>; + def : Pat<(v4i16 (int_arm_neon_vqadds + (v4i16 DPR:$src1), + (v4i16 (int_arm_neon_vqrdmulh + (v4i16 DPR:$Vn), + (v4i16 (NEONvduplane (v4i16 DPR_8:$Vm), + imm:$lane)))))), + (v4i16 (VQRDMLAHslv4i16 DPR:$src1, DPR:$Vn, DPR_8:$Vm, + imm:$lane))>; + def : Pat<(v2i32 (int_arm_neon_vqadds + (v2i32 DPR:$src1), + (v2i32 (int_arm_neon_vqrdmulh + (v2i32 DPR:$Vn), + (v2i32 (NEONvduplane (v2i32 DPR_VFP2:$Vm), + imm:$lane)))))), + (v2i32 (VQRDMLAHslv2i32 DPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, + imm:$lane))>; + def : Pat<(v8i16 (int_arm_neon_vqadds + (v8i16 QPR:$src1), + (v8i16 (int_arm_neon_vqrdmulh + (v8i16 QPR:$src2), + (v8i16 (NEONvduplane (v8i16 QPR:$src3), + imm:$lane)))))), + (v8i16 (VQRDMLAHslv8i16 (v8i16 QPR:$src1), + (v8i16 QPR:$src2), + (v4i16 (EXTRACT_SUBREG + QPR:$src3, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane)))>; + def : Pat<(v4i32 (int_arm_neon_vqadds + (v4i32 QPR:$src1), + (v4i32 (int_arm_neon_vqrdmulh + (v4i32 QPR:$src2), + (v4i32 (NEONvduplane (v4i32 QPR:$src3), + imm:$lane)))))), + (v4i32 (VQRDMLAHslv4i32 (v4i32 QPR:$src1), + (v4i32 QPR:$src2), + (v2i32 (EXTRACT_SUBREG + QPR:$src3, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; + + // VQRDMLSH : Vector Saturating Rounding Doubling Multiply Subtract Long + // (Q -= D * D) + defm VQRDMLSH : N3VInt3_HS<1, 0, 0b1100, 1, IIC_VMACi16D, IIC_VMACi32D, + IIC_VMACi16Q, IIC_VMACi32Q, "vqrdmlsh", "s", + null_frag>; + def : Pat<(v4i16 (int_arm_neon_vqsubs + (v4i16 DPR:$src1), + (v4i16 (int_arm_neon_vqrdmulh (v4i16 DPR:$Vn), + (v4i16 DPR:$Vm))))), + (v4i16 (VQRDMLSHv4i16 DPR:$src1, DPR:$Vn, DPR:$Vm))>; + def : Pat<(v2i32 (int_arm_neon_vqsubs + (v2i32 DPR:$src1), + (v2i32 (int_arm_neon_vqrdmulh (v2i32 DPR:$Vn), + (v2i32 DPR:$Vm))))), + (v2i32 (VQRDMLSHv2i32 DPR:$src1, DPR:$Vn, DPR:$Vm))>; + def : Pat<(v8i16 (int_arm_neon_vqsubs + (v8i16 QPR:$src1), + (v8i16 (int_arm_neon_vqrdmulh (v8i16 QPR:$Vn), + (v8i16 QPR:$Vm))))), + (v8i16 (VQRDMLSHv8i16 QPR:$src1, QPR:$Vn, QPR:$Vm))>; + def : Pat<(v4i32 (int_arm_neon_vqsubs + (v4i32 QPR:$src1), + (v4i32 (int_arm_neon_vqrdmulh (v4i32 QPR:$Vn), + (v4i32 QPR:$Vm))))), + (v4i32 (VQRDMLSHv4i32 QPR:$src1, QPR:$Vn, QPR:$Vm))>; + + defm VQRDMLSHsl : N3VMulOpSL_HS<0b1111, IIC_VMACi16D, IIC_VMACi32D, + IIC_VMACi16Q, IIC_VMACi32Q, "vqrdmlsh", "s", + null_frag>; + def : Pat<(v4i16 (int_arm_neon_vqsubs + (v4i16 DPR:$src1), + (v4i16 (int_arm_neon_vqrdmulh + (v4i16 DPR:$Vn), + (v4i16 (NEONvduplane (v4i16 DPR_8:$Vm), + imm:$lane)))))), + (v4i16 (VQRDMLSHslv4i16 DPR:$src1, DPR:$Vn, DPR_8:$Vm, imm:$lane))>; + def : Pat<(v2i32 (int_arm_neon_vqsubs + (v2i32 DPR:$src1), + (v2i32 (int_arm_neon_vqrdmulh + (v2i32 DPR:$Vn), + (v2i32 (NEONvduplane (v2i32 DPR_VFP2:$Vm), + imm:$lane)))))), + (v2i32 (VQRDMLSHslv2i32 DPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, + imm:$lane))>; + def : Pat<(v8i16 (int_arm_neon_vqsubs + (v8i16 QPR:$src1), + (v8i16 (int_arm_neon_vqrdmulh + (v8i16 QPR:$src2), + (v8i16 (NEONvduplane (v8i16 QPR:$src3), + imm:$lane)))))), + (v8i16 (VQRDMLSHslv8i16 (v8i16 QPR:$src1), + (v8i16 QPR:$src2), + (v4i16 (EXTRACT_SUBREG + QPR:$src3, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane)))>; + def : Pat<(v4i32 (int_arm_neon_vqsubs + (v4i32 QPR:$src1), + (v4i32 (int_arm_neon_vqrdmulh + (v4i32 QPR:$src2), + (v4i32 (NEONvduplane (v4i32 QPR:$src3), + imm:$lane)))))), + (v4i32 (VQRDMLSHslv4i32 (v4i32 QPR:$src1), + (v4i32 QPR:$src2), + (v2i32 (EXTRACT_SUBREG + QPR:$src3, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; +} +// VQDMLAL : Vector Saturating Doubling Multiply Accumulate Long (Q += D * D) +defm VQDMLAL : N3VLInt3_HS<0, 1, 0b1001, 0, IIC_VMACi16D, IIC_VMACi32D, + "vqdmlal", "s", null_frag>; +defm VQDMLALsl: N3VLInt3SL_HS<0, 0b0011, "vqdmlal", "s", null_frag>; + +def : Pat<(v4i32 (int_arm_neon_vqadds (v4i32 QPR:$src1), + (v4i32 (int_arm_neon_vqdmull (v4i16 DPR:$Vn), + (v4i16 DPR:$Vm))))), + (VQDMLALv4i32 QPR:$src1, DPR:$Vn, DPR:$Vm)>; +def : Pat<(v2i64 (int_arm_neon_vqadds (v2i64 QPR:$src1), + (v2i64 (int_arm_neon_vqdmull (v2i32 DPR:$Vn), + (v2i32 DPR:$Vm))))), + (VQDMLALv2i64 QPR:$src1, DPR:$Vn, DPR:$Vm)>; +def : Pat<(v4i32 (int_arm_neon_vqadds (v4i32 QPR:$src1), + (v4i32 (int_arm_neon_vqdmull (v4i16 DPR:$Vn), + (v4i16 (NEONvduplane (v4i16 DPR_8:$Vm), + imm:$lane)))))), + (VQDMLALslv4i16 QPR:$src1, DPR:$Vn, DPR_8:$Vm, imm:$lane)>; +def : Pat<(v2i64 (int_arm_neon_vqadds (v2i64 QPR:$src1), + (v2i64 (int_arm_neon_vqdmull (v2i32 DPR:$Vn), + (v2i32 (NEONvduplane (v2i32 DPR_VFP2:$Vm), + imm:$lane)))))), + (VQDMLALslv2i32 QPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, imm:$lane)>; + +// VMLS : Vector Multiply Subtract (integer and floating-point) +defm VMLS : N3VMulOp_QHS<1, 0, 0b1001, 0, IIC_VMACi16D, IIC_VMACi32D, + IIC_VMACi16Q, IIC_VMACi32Q, "vmls", "i", sub>; +def VMLSfd : N3VDMulOp<0, 0, 0b10, 0b1101, 1, IIC_VMACD, "vmls", "f32", + v2f32, fmul_su, fsub_mlx>, + Requires<[HasNEON, UseFPVMLx, DontUseFusedMAC]>; +def VMLSfq : N3VQMulOp<0, 0, 0b10, 0b1101, 1, IIC_VMACQ, "vmls", "f32", + v4f32, fmul_su, fsub_mlx>, + Requires<[HasNEON, UseFPVMLx, DontUseFusedMAC]>; +def VMLShd : N3VDMulOp<0, 0, 0b11, 0b1101, 1, IIC_VMACD, "vmls", "f16", + v4f16, fmul, fsub>, + Requires<[HasNEON, HasFullFP16, UseFPVMLx, DontUseFusedMAC]>; +def VMLShq : N3VQMulOp<0, 0, 0b11, 0b1101, 1, IIC_VMACQ, "vmls", "f16", + v8f16, fmul, fsub>, + Requires<[HasNEON, HasFullFP16, UseFPVMLx, DontUseFusedMAC]>; +defm VMLSsl : N3VMulOpSL_HS<0b0100, IIC_VMACi16D, IIC_VMACi32D, + IIC_VMACi16Q, IIC_VMACi32Q, "vmls", "i", sub>; +def VMLSslfd : N3VDMulOpSL<0b10, 0b0101, IIC_VMACD, "vmls", "f32", + v2f32, fmul_su, fsub_mlx>, + Requires<[HasNEON, UseFPVMLx]>; +def VMLSslfq : N3VQMulOpSL<0b10, 0b0101, IIC_VMACQ, "vmls", "f32", + v4f32, v2f32, fmul_su, fsub_mlx>, + Requires<[HasNEON, UseFPVMLx]>; +def VMLSslhd : N3VDMulOpSL16<0b01, 0b0101, IIC_VMACD, "vmls", "f16", + v4f16, fmul, fsub>, + Requires<[HasNEON, HasFullFP16, UseFPVMLx]>; +def VMLSslhq : N3VQMulOpSL16<0b01, 0b0101, IIC_VMACQ, "vmls", "f16", + v8f16, v4f16, fmul, fsub>, + Requires<[HasNEON, HasFullFP16, UseFPVMLx]>; + +def : Pat<(v8i16 (sub (v8i16 QPR:$src1), + (mul (v8i16 QPR:$src2), + (v8i16 (NEONvduplane (v8i16 QPR:$src3), imm:$lane))))), + (v8i16 (VMLSslv8i16 (v8i16 QPR:$src1), (v8i16 QPR:$src2), + (v4i16 (EXTRACT_SUBREG QPR:$src3, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane)))>; + +def : Pat<(v4i32 (sub (v4i32 QPR:$src1), + (mul (v4i32 QPR:$src2), + (v4i32 (NEONvduplane (v4i32 QPR:$src3), imm:$lane))))), + (v4i32 (VMLSslv4i32 (v4i32 QPR:$src1), (v4i32 QPR:$src2), + (v2i32 (EXTRACT_SUBREG QPR:$src3, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; + +def : Pat<(v4f32 (fsub_mlx (v4f32 QPR:$src1), + (fmul_su (v4f32 QPR:$src2), + (v4f32 (NEONvduplane (v4f32 QPR:$src3), imm:$lane))))), + (v4f32 (VMLSslfq (v4f32 QPR:$src1), (v4f32 QPR:$src2), + (v2f32 (EXTRACT_SUBREG QPR:$src3, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>, + Requires<[HasNEON, UseFPVMLx]>; + +// VMLSL : Vector Multiply Subtract Long (Q -= D * D) +defm VMLSLs : N3VLMulOp_QHS<0,1,0b1010,0, IIC_VMACi16D, IIC_VMACi32D, + "vmlsl", "s", NEONvmulls, sub>; +defm VMLSLu : N3VLMulOp_QHS<1,1,0b1010,0, IIC_VMACi16D, IIC_VMACi32D, + "vmlsl", "u", NEONvmullu, sub>; + +defm VMLSLsls : N3VLMulOpSL_HS<0, 0b0110, "vmlsl", "s", NEONvmulls, sub>; +defm VMLSLslu : N3VLMulOpSL_HS<1, 0b0110, "vmlsl", "u", NEONvmullu, sub>; + +// VQDMLSL : Vector Saturating Doubling Multiply Subtract Long (Q -= D * D) +defm VQDMLSL : N3VLInt3_HS<0, 1, 0b1011, 0, IIC_VMACi16D, IIC_VMACi32D, + "vqdmlsl", "s", null_frag>; +defm VQDMLSLsl: N3VLInt3SL_HS<0, 0b0111, "vqdmlsl", "s", null_frag>; + +def : Pat<(v4i32 (int_arm_neon_vqsubs (v4i32 QPR:$src1), + (v4i32 (int_arm_neon_vqdmull (v4i16 DPR:$Vn), + (v4i16 DPR:$Vm))))), + (VQDMLSLv4i32 QPR:$src1, DPR:$Vn, DPR:$Vm)>; +def : Pat<(v2i64 (int_arm_neon_vqsubs (v2i64 QPR:$src1), + (v2i64 (int_arm_neon_vqdmull (v2i32 DPR:$Vn), + (v2i32 DPR:$Vm))))), + (VQDMLSLv2i64 QPR:$src1, DPR:$Vn, DPR:$Vm)>; +def : Pat<(v4i32 (int_arm_neon_vqsubs (v4i32 QPR:$src1), + (v4i32 (int_arm_neon_vqdmull (v4i16 DPR:$Vn), + (v4i16 (NEONvduplane (v4i16 DPR_8:$Vm), + imm:$lane)))))), + (VQDMLSLslv4i16 QPR:$src1, DPR:$Vn, DPR_8:$Vm, imm:$lane)>; +def : Pat<(v2i64 (int_arm_neon_vqsubs (v2i64 QPR:$src1), + (v2i64 (int_arm_neon_vqdmull (v2i32 DPR:$Vn), + (v2i32 (NEONvduplane (v2i32 DPR_VFP2:$Vm), + imm:$lane)))))), + (VQDMLSLslv2i32 QPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, imm:$lane)>; + +// Fused Vector Multiply-Accumulate and Fused Multiply-Subtract Operations. +def VFMAfd : N3VDMulOp<0, 0, 0b00, 0b1100, 1, IIC_VFMACD, "vfma", "f32", + v2f32, fmul_su, fadd_mlx>, + Requires<[HasNEON,HasVFP4,UseFusedMAC]>; + +def VFMAfq : N3VQMulOp<0, 0, 0b00, 0b1100, 1, IIC_VFMACQ, "vfma", "f32", + v4f32, fmul_su, fadd_mlx>, + Requires<[HasNEON,HasVFP4,UseFusedMAC]>; +def VFMAhd : N3VDMulOp<0, 0, 0b01, 0b1100, 1, IIC_VFMACD, "vfma", "f16", + v4f16, fmul, fadd>, + Requires<[HasNEON,HasFullFP16,UseFusedMAC]>; + +def VFMAhq : N3VQMulOp<0, 0, 0b01, 0b1100, 1, IIC_VFMACQ, "vfma", "f16", + v8f16, fmul, fadd>, + Requires<[HasNEON,HasFullFP16,UseFusedMAC]>; + +// Fused Vector Multiply Subtract (floating-point) +def VFMSfd : N3VDMulOp<0, 0, 0b10, 0b1100, 1, IIC_VFMACD, "vfms", "f32", + v2f32, fmul_su, fsub_mlx>, + Requires<[HasNEON,HasVFP4,UseFusedMAC]>; +def VFMSfq : N3VQMulOp<0, 0, 0b10, 0b1100, 1, IIC_VFMACQ, "vfms", "f32", + v4f32, fmul_su, fsub_mlx>, + Requires<[HasNEON,HasVFP4,UseFusedMAC]>; +def VFMShd : N3VDMulOp<0, 0, 0b11, 0b1100, 1, IIC_VFMACD, "vfms", "f16", + v4f16, fmul, fsub>, + Requires<[HasNEON,HasFullFP16,UseFusedMAC]>; +def VFMShq : N3VQMulOp<0, 0, 0b11, 0b1100, 1, IIC_VFMACQ, "vfms", "f16", + v8f16, fmul, fsub>, + Requires<[HasNEON,HasFullFP16,UseFusedMAC]>; + +// Match @llvm.fma.* intrinsics +def : Pat<(v2f32 (fma DPR:$Vn, DPR:$Vm, DPR:$src1)), + (VFMAfd DPR:$src1, DPR:$Vn, DPR:$Vm)>, + Requires<[HasVFP4]>; +def : Pat<(v4f32 (fma QPR:$Vn, QPR:$Vm, QPR:$src1)), + (VFMAfq QPR:$src1, QPR:$Vn, QPR:$Vm)>, + Requires<[HasVFP4]>; +def : Pat<(v2f32 (fma (fneg DPR:$Vn), DPR:$Vm, DPR:$src1)), + (VFMSfd DPR:$src1, DPR:$Vn, DPR:$Vm)>, + Requires<[HasVFP4]>; +def : Pat<(v4f32 (fma (fneg QPR:$Vn), QPR:$Vm, QPR:$src1)), + (VFMSfq QPR:$src1, QPR:$Vn, QPR:$Vm)>, + Requires<[HasVFP4]>; + +// ARMv8.2a dot product instructions. +// We put them in the VFPV8 decoder namespace because the ARM and Thumb +// encodings are the same and thus no further bit twiddling is necessary +// in the disassembler. +class VDOT<bit op6, bit op4, RegisterClass RegTy, string Asm, string AsmTy, + ValueType AccumTy, ValueType InputTy, + SDPatternOperator OpNode> : + N3Vnp<0b11000, 0b10, 0b1101, op6, op4, (outs RegTy:$dst), + (ins RegTy:$Vd, RegTy:$Vn, RegTy:$Vm), N3RegFrm, IIC_VDOTPROD, + Asm, AsmTy, + [(set (AccumTy RegTy:$dst), + (OpNode (AccumTy RegTy:$Vd), + (InputTy RegTy:$Vn), + (InputTy RegTy:$Vm)))]> { + let Predicates = [HasDotProd]; + let DecoderNamespace = "VFPV8"; + let Constraints = "$dst = $Vd"; +} + +def VUDOTD : VDOT<0, 1, DPR, "vudot", "u8", v2i32, v8i8, int_arm_neon_udot>; +def VSDOTD : VDOT<0, 0, DPR, "vsdot", "s8", v2i32, v8i8, int_arm_neon_sdot>; +def VUDOTQ : VDOT<1, 1, QPR, "vudot", "u8", v4i32, v16i8, int_arm_neon_udot>; +def VSDOTQ : VDOT<1, 0, QPR, "vsdot", "s8", v4i32, v16i8, int_arm_neon_sdot>; + +// Indexed dot product instructions: +multiclass DOTI<string opc, string dt, bit Q, bit U, RegisterClass Ty, + ValueType AccumType, ValueType InputType, SDPatternOperator OpNode, + dag RHS> { + def "" : N3Vnp<0b11100, 0b10, 0b1101, Q, U, (outs Ty:$dst), + (ins Ty:$Vd, Ty:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane), + N3RegFrm, IIC_VDOTPROD, opc, dt, []> { + bit lane; + let Inst{5} = lane; + let AsmString = !strconcat(opc, ".", dt, "\t$Vd, $Vn, $Vm$lane"); + let Constraints = "$dst = $Vd"; + let Predicates = [HasDotProd]; + let DecoderNamespace = "VFPV8"; + } + + def : Pat< + (AccumType (OpNode (AccumType Ty:$Vd), + (InputType Ty:$Vn), + (InputType (bitconvert (AccumType + (NEONvduplane (AccumType Ty:$Vm), + VectorIndex32:$lane)))))), + (!cast<Instruction>(NAME) Ty:$Vd, Ty:$Vn, RHS, VectorIndex32:$lane)>; +} + +defm VUDOTDI : DOTI<"vudot", "u8", 0b0, 0b1, DPR, v2i32, v8i8, + int_arm_neon_udot, (v2i32 DPR_VFP2:$Vm)>; +defm VSDOTDI : DOTI<"vsdot", "s8", 0b0, 0b0, DPR, v2i32, v8i8, + int_arm_neon_sdot, (v2i32 DPR_VFP2:$Vm)>; +defm VUDOTQI : DOTI<"vudot", "u8", 0b1, 0b1, QPR, v4i32, v16i8, + int_arm_neon_udot, (EXTRACT_SUBREG QPR:$Vm, dsub_0)>; +defm VSDOTQI : DOTI<"vsdot", "s8", 0b1, 0b0, QPR, v4i32, v16i8, + int_arm_neon_sdot, (EXTRACT_SUBREG QPR:$Vm, dsub_0)>; + + +// ARMv8.3 complex operations +class BaseN3VCP8ComplexTied<bit op21, bit op4, bit s, bit q, + InstrItinClass itin, dag oops, dag iops, + string opc, string dt, list<dag> pattern> + : N3VCP8<{?,?}, {op21,s}, q, op4, oops, + iops, itin, opc, dt, "$Vd, $Vn, $Vm, $rot", "$src1 = $Vd", pattern>{ + bits<2> rot; + let Inst{24-23} = rot; +} + +class BaseN3VCP8ComplexOdd<bit op23, bit op21, bit op4, bit s, bit q, + InstrItinClass itin, dag oops, dag iops, string opc, + string dt, list<dag> pattern> + : N3VCP8<{?,op23}, {op21,s}, q, op4, oops, + iops, itin, opc, dt, "$Vd, $Vn, $Vm, $rot", "", pattern> { + bits<1> rot; + let Inst{24} = rot; +} + +class BaseN3VCP8ComplexTiedLane32<bit op4, bit s, bit q, InstrItinClass itin, + dag oops, dag iops, string opc, string dt, + list<dag> pattern> + : N3VLaneCP8<s, {?,?}, q, op4, oops, iops, itin, opc, dt, + "$Vd, $Vn, $Vm$lane, $rot", "$src1 = $Vd", pattern> { + bits<2> rot; + bit lane; + + let Inst{21-20} = rot; + let Inst{5} = lane; +} + +class BaseN3VCP8ComplexTiedLane64<bit op4, bit s, bit q, InstrItinClass itin, + dag oops, dag iops, string opc, string dt, + list<dag> pattern> + : N3VLaneCP8<s, {?,?}, q, op4, oops, iops, itin, opc, dt, + "$Vd, $Vn, $Vm$lane, $rot", "$src1 = $Vd", pattern> { + bits<2> rot; + bit lane; + + let Inst{21-20} = rot; + let Inst{5} = Vm{4}; + // This is needed because the lane operand does not have any bits in the + // encoding (it only has one possible value), so we need to manually set it + // to it's default value. + let DecoderMethod = "DecodeNEONComplexLane64Instruction"; +} + +multiclass N3VCP8ComplexTied<bit op21, bit op4, + string OpcodeStr, SDPatternOperator Op> { + let Predicates = [HasNEON,HasV8_3a,HasFullFP16] in { + def v4f16 : BaseN3VCP8ComplexTied<op21, op4, 0, 0, IIC_VMACD, (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vn, DPR:$Vm, complexrotateop:$rot), + OpcodeStr, "f16", []>; + def v8f16 : BaseN3VCP8ComplexTied<op21, op4, 0, 1, IIC_VMACQ, (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vn, QPR:$Vm, complexrotateop:$rot), + OpcodeStr, "f16", []>; + } + let Predicates = [HasNEON,HasV8_3a] in { + def v2f32 : BaseN3VCP8ComplexTied<op21, op4, 1, 0, IIC_VMACD, (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vn, DPR:$Vm, complexrotateop:$rot), + OpcodeStr, "f32", []>; + def v4f32 : BaseN3VCP8ComplexTied<op21, op4, 1, 1, IIC_VMACQ, (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vn, QPR:$Vm, complexrotateop:$rot), + OpcodeStr, "f32", []>; + } +} + +multiclass N3VCP8ComplexOdd<bit op23, bit op21, bit op4, + string OpcodeStr, SDPatternOperator Op> { + let Predicates = [HasNEON,HasV8_3a,HasFullFP16] in { + def v4f16 : BaseN3VCP8ComplexOdd<op23, op21, op4, 0, 0, IIC_VMACD, + (outs DPR:$Vd), + (ins DPR:$Vn, DPR:$Vm, complexrotateopodd:$rot), + OpcodeStr, "f16", []>; + def v8f16 : BaseN3VCP8ComplexOdd<op23, op21, op4, 0, 1, IIC_VMACQ, + (outs QPR:$Vd), + (ins QPR:$Vn, QPR:$Vm, complexrotateopodd:$rot), + OpcodeStr, "f16", []>; + } + let Predicates = [HasNEON,HasV8_3a] in { + def v2f32 : BaseN3VCP8ComplexOdd<op23, op21, op4, 1, 0, IIC_VMACD, + (outs DPR:$Vd), + (ins DPR:$Vn, DPR:$Vm, complexrotateopodd:$rot), + OpcodeStr, "f32", []>; + def v4f32 : BaseN3VCP8ComplexOdd<op23, op21, op4, 1, 1, IIC_VMACQ, + (outs QPR:$Vd), + (ins QPR:$Vn, QPR:$Vm, complexrotateopodd:$rot), + OpcodeStr, "f32", []>; + } +} + +// These instructions index by pairs of lanes, so the VectorIndexes are twice +// as wide as the data types. +multiclass N3VCP8ComplexTiedLane<bit op4, string OpcodeStr, + SDPatternOperator Op> { + let Predicates = [HasNEON,HasV8_3a,HasFullFP16] in { + def v4f16_indexed : BaseN3VCP8ComplexTiedLane32<op4, 0, 0, IIC_VMACD, + (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, + VectorIndex32:$lane, complexrotateop:$rot), + OpcodeStr, "f16", []>; + def v8f16_indexed : BaseN3VCP8ComplexTiedLane32<op4, 0, 1, IIC_VMACQ, + (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vn, DPR_VFP2:$Vm, + VectorIndex32:$lane, complexrotateop:$rot), + OpcodeStr, "f16", []>; + } + let Predicates = [HasNEON,HasV8_3a] in { + def v2f32_indexed : BaseN3VCP8ComplexTiedLane64<op4, 1, 0, IIC_VMACD, + (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vn, DPR:$Vm, VectorIndex64:$lane, + complexrotateop:$rot), + OpcodeStr, "f32", []>; + def v4f32_indexed : BaseN3VCP8ComplexTiedLane64<op4, 1, 1, IIC_VMACQ, + (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vn, DPR:$Vm, VectorIndex64:$lane, + complexrotateop:$rot), + OpcodeStr, "f32", []>; + } +} + +defm VCMLA : N3VCP8ComplexTied<1, 0, "vcmla", null_frag>; +defm VCADD : N3VCP8ComplexOdd<1, 0, 0, "vcadd", null_frag>; +defm VCMLA : N3VCP8ComplexTiedLane<0, "vcmla", null_frag>; + +// Vector Subtract Operations. + +// VSUB : Vector Subtract (integer and floating-point) +defm VSUB : N3V_QHSD<1, 0, 0b1000, 0, IIC_VSUBiD, IIC_VSUBiQ, + "vsub", "i", sub, 0>; +def VSUBfd : N3VD<0, 0, 0b10, 0b1101, 0, IIC_VBIND, "vsub", "f32", + v2f32, v2f32, fsub, 0>; +def VSUBfq : N3VQ<0, 0, 0b10, 0b1101, 0, IIC_VBINQ, "vsub", "f32", + v4f32, v4f32, fsub, 0>; +def VSUBhd : N3VD<0, 0, 0b11, 0b1101, 0, IIC_VBIND, "vsub", "f16", + v4f16, v4f16, fsub, 0>, + Requires<[HasNEON,HasFullFP16]>; +def VSUBhq : N3VQ<0, 0, 0b11, 0b1101, 0, IIC_VBINQ, "vsub", "f16", + v8f16, v8f16, fsub, 0>, + Requires<[HasNEON,HasFullFP16]>; +// VSUBL : Vector Subtract Long (Q = D - D) +defm VSUBLs : N3VLExt_QHS<0,1,0b0010,0, IIC_VSHLiD, IIC_VSHLiD, + "vsubl", "s", sub, sext, 0>; +defm VSUBLu : N3VLExt_QHS<1,1,0b0010,0, IIC_VSHLiD, IIC_VSHLiD, + "vsubl", "u", sub, zext, 0>; +// VSUBW : Vector Subtract Wide (Q = Q - D) +defm VSUBWs : N3VW_QHS<0,1,0b0011,0, "vsubw", "s", sub, sext, 0>; +defm VSUBWu : N3VW_QHS<1,1,0b0011,0, "vsubw", "u", sub, zext, 0>; +// VHSUB : Vector Halving Subtract +defm VHSUBs : N3VInt_QHS<0, 0, 0b0010, 0, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vhsub", "s", int_arm_neon_vhsubs, 0>; +defm VHSUBu : N3VInt_QHS<1, 0, 0b0010, 0, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vhsub", "u", int_arm_neon_vhsubu, 0>; +// VQSUB : Vector Saturing Subtract +defm VQSUBs : N3VInt_QHSD<0, 0, 0b0010, 1, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vqsub", "s", int_arm_neon_vqsubs, 0>; +defm VQSUBu : N3VInt_QHSD<1, 0, 0b0010, 1, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vqsub", "u", int_arm_neon_vqsubu, 0>; +// VSUBHN : Vector Subtract and Narrow Returning High Half (D = Q - Q) +defm VSUBHN : N3VNInt_HSD<0,1,0b0110,0, "vsubhn", "i", null_frag, 0>; +// VRSUBHN : Vector Rounding Subtract and Narrow Returning High Half (D=Q-Q) +defm VRSUBHN : N3VNInt_HSD<1,1,0b0110,0, "vrsubhn", "i", + int_arm_neon_vrsubhn, 0>; + +def : Pat<(v8i8 (trunc (NEONvshru (sub (v8i16 QPR:$Vn), QPR:$Vm), 8))), + (VSUBHNv8i8 QPR:$Vn, QPR:$Vm)>; +def : Pat<(v4i16 (trunc (NEONvshru (sub (v4i32 QPR:$Vn), QPR:$Vm), 16))), + (VSUBHNv4i16 QPR:$Vn, QPR:$Vm)>; +def : Pat<(v2i32 (trunc (NEONvshru (sub (v2i64 QPR:$Vn), QPR:$Vm), 32))), + (VSUBHNv2i32 QPR:$Vn, QPR:$Vm)>; + +// Vector Comparisons. + +// VCEQ : Vector Compare Equal +defm VCEQ : N3V_QHS<1, 0, 0b1000, 1, IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, + IIC_VSUBi4Q, "vceq", "i", NEONvceq, 1>; +def VCEQfd : N3VD<0,0,0b00,0b1110,0, IIC_VBIND, "vceq", "f32", v2i32, v2f32, + NEONvceq, 1>; +def VCEQfq : N3VQ<0,0,0b00,0b1110,0, IIC_VBINQ, "vceq", "f32", v4i32, v4f32, + NEONvceq, 1>; +def VCEQhd : N3VD<0,0,0b01,0b1110,0, IIC_VBIND, "vceq", "f16", v4i16, v4f16, + NEONvceq, 1>, + Requires<[HasNEON, HasFullFP16]>; +def VCEQhq : N3VQ<0,0,0b01,0b1110,0, IIC_VBINQ, "vceq", "f16", v8i16, v8f16, + NEONvceq, 1>, + Requires<[HasNEON, HasFullFP16]>; + +let TwoOperandAliasConstraint = "$Vm = $Vd" in +defm VCEQz : N2V_QHS_cmp<0b11, 0b11, 0b01, 0b00010, 0, "vceq", "i", + "$Vd, $Vm, #0", NEONvceqz>; + +// VCGE : Vector Compare Greater Than or Equal +defm VCGEs : N3V_QHS<0, 0, 0b0011, 1, IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, + IIC_VSUBi4Q, "vcge", "s", NEONvcge, 0>; +defm VCGEu : N3V_QHS<1, 0, 0b0011, 1, IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, + IIC_VSUBi4Q, "vcge", "u", NEONvcgeu, 0>; +def VCGEfd : N3VD<1,0,0b00,0b1110,0, IIC_VBIND, "vcge", "f32", v2i32, v2f32, + NEONvcge, 0>; +def VCGEfq : N3VQ<1,0,0b00,0b1110,0, IIC_VBINQ, "vcge", "f32", v4i32, v4f32, + NEONvcge, 0>; +def VCGEhd : N3VD<1,0,0b01,0b1110,0, IIC_VBIND, "vcge", "f16", v4i16, v4f16, + NEONvcge, 0>, + Requires<[HasNEON, HasFullFP16]>; +def VCGEhq : N3VQ<1,0,0b01,0b1110,0, IIC_VBINQ, "vcge", "f16", v8i16, v8f16, + NEONvcge, 0>, + Requires<[HasNEON, HasFullFP16]>; + +let TwoOperandAliasConstraint = "$Vm = $Vd" in { +defm VCGEz : N2V_QHS_cmp<0b11, 0b11, 0b01, 0b00001, 0, "vcge", "s", + "$Vd, $Vm, #0", NEONvcgez>; +defm VCLEz : N2V_QHS_cmp<0b11, 0b11, 0b01, 0b00011, 0, "vcle", "s", + "$Vd, $Vm, #0", NEONvclez>; +} + +// VCGT : Vector Compare Greater Than +defm VCGTs : N3V_QHS<0, 0, 0b0011, 0, IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, + IIC_VSUBi4Q, "vcgt", "s", NEONvcgt, 0>; +defm VCGTu : N3V_QHS<1, 0, 0b0011, 0, IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, + IIC_VSUBi4Q, "vcgt", "u", NEONvcgtu, 0>; +def VCGTfd : N3VD<1,0,0b10,0b1110,0, IIC_VBIND, "vcgt", "f32", v2i32, v2f32, + NEONvcgt, 0>; +def VCGTfq : N3VQ<1,0,0b10,0b1110,0, IIC_VBINQ, "vcgt", "f32", v4i32, v4f32, + NEONvcgt, 0>; +def VCGThd : N3VD<1,0,0b11,0b1110,0, IIC_VBIND, "vcgt", "f16", v4i16, v4f16, + NEONvcgt, 0>, + Requires<[HasNEON, HasFullFP16]>; +def VCGThq : N3VQ<1,0,0b11,0b1110,0, IIC_VBINQ, "vcgt", "f16", v8i16, v8f16, + NEONvcgt, 0>, + Requires<[HasNEON, HasFullFP16]>; + +let TwoOperandAliasConstraint = "$Vm = $Vd" in { +defm VCGTz : N2V_QHS_cmp<0b11, 0b11, 0b01, 0b00000, 0, "vcgt", "s", + "$Vd, $Vm, #0", NEONvcgtz>; +defm VCLTz : N2V_QHS_cmp<0b11, 0b11, 0b01, 0b00100, 0, "vclt", "s", + "$Vd, $Vm, #0", NEONvcltz>; +} + +// VACGE : Vector Absolute Compare Greater Than or Equal (aka VCAGE) +def VACGEfd : N3VDInt<1, 0, 0b00, 0b1110, 1, N3RegFrm, IIC_VBIND, "vacge", + "f32", v2i32, v2f32, int_arm_neon_vacge, 0>; +def VACGEfq : N3VQInt<1, 0, 0b00, 0b1110, 1, N3RegFrm, IIC_VBINQ, "vacge", + "f32", v4i32, v4f32, int_arm_neon_vacge, 0>; +def VACGEhd : N3VDInt<1, 0, 0b01, 0b1110, 1, N3RegFrm, IIC_VBIND, "vacge", + "f16", v4i16, v4f16, int_arm_neon_vacge, 0>, + Requires<[HasNEON, HasFullFP16]>; +def VACGEhq : N3VQInt<1, 0, 0b01, 0b1110, 1, N3RegFrm, IIC_VBINQ, "vacge", + "f16", v8i16, v8f16, int_arm_neon_vacge, 0>, + Requires<[HasNEON, HasFullFP16]>; +// VACGT : Vector Absolute Compare Greater Than (aka VCAGT) +def VACGTfd : N3VDInt<1, 0, 0b10, 0b1110, 1, N3RegFrm, IIC_VBIND, "vacgt", + "f32", v2i32, v2f32, int_arm_neon_vacgt, 0>; +def VACGTfq : N3VQInt<1, 0, 0b10, 0b1110, 1, N3RegFrm, IIC_VBINQ, "vacgt", + "f32", v4i32, v4f32, int_arm_neon_vacgt, 0>; +def VACGThd : N3VDInt<1, 0, 0b11, 0b1110, 1, N3RegFrm, IIC_VBIND, "vacgt", + "f16", v4i16, v4f16, int_arm_neon_vacgt, 0>, + Requires<[HasNEON, HasFullFP16]>; +def VACGThq : N3VQInt<1, 0, 0b11, 0b1110, 1, N3RegFrm, IIC_VBINQ, "vacgt", + "f16", v8f16, v8f16, int_arm_neon_vacgt, 0>, + Requires<[HasNEON, HasFullFP16]>; +// VTST : Vector Test Bits +defm VTST : N3V_QHS<0, 0, 0b1000, 1, IIC_VBINi4D, IIC_VBINi4D, IIC_VBINi4Q, + IIC_VBINi4Q, "vtst", "", NEONvtst, 1>; + +def: NEONInstAlias<"vaclt${p}.f32 $Vd, $Vn, $Vm", + (VACGTfd DPR:$Vd, DPR:$Vm, DPR:$Vn, pred:$p)>; +def: NEONInstAlias<"vaclt${p}.f32 $Vd, $Vn, $Vm", + (VACGTfq QPR:$Vd, QPR:$Vm, QPR:$Vn, pred:$p)>; +def: NEONInstAlias<"vacle${p}.f32 $Vd, $Vn, $Vm", + (VACGEfd DPR:$Vd, DPR:$Vm, DPR:$Vn, pred:$p)>; +def: NEONInstAlias<"vacle${p}.f32 $Vd, $Vn, $Vm", + (VACGEfq QPR:$Vd, QPR:$Vm, QPR:$Vn, pred:$p)>; +let Predicates = [HasNEON, HasFullFP16] in { +def: NEONInstAlias<"vaclt${p}.f16 $Vd, $Vn, $Vm", + (VACGThd DPR:$Vd, DPR:$Vm, DPR:$Vn, pred:$p)>; +def: NEONInstAlias<"vaclt${p}.f16 $Vd, $Vn, $Vm", + (VACGThq QPR:$Vd, QPR:$Vm, QPR:$Vn, pred:$p)>; +def: NEONInstAlias<"vacle${p}.f16 $Vd, $Vn, $Vm", + (VACGEhd DPR:$Vd, DPR:$Vm, DPR:$Vn, pred:$p)>; +def: NEONInstAlias<"vacle${p}.f16 $Vd, $Vn, $Vm", + (VACGEhq QPR:$Vd, QPR:$Vm, QPR:$Vn, pred:$p)>; +} + +def: NEONInstAlias<"vaclt${p}.f32 $Vd, $Vm", + (VACGTfd DPR:$Vd, DPR:$Vm, DPR:$Vd, pred:$p)>; +def: NEONInstAlias<"vaclt${p}.f32 $Vd, $Vm", + (VACGTfq QPR:$Vd, QPR:$Vm, QPR:$Vd, pred:$p)>; +def: NEONInstAlias<"vacle${p}.f32 $Vd, $Vm", + (VACGEfd DPR:$Vd, DPR:$Vm, DPR:$Vd, pred:$p)>; +def: NEONInstAlias<"vacle${p}.f32 $Vd, $Vm", + (VACGEfq QPR:$Vd, QPR:$Vm, QPR:$Vd, pred:$p)>; +let Predicates = [HasNEON, HasFullFP16] in { +def: NEONInstAlias<"vaclt${p}.f16 $Vd, $Vm", + (VACGThd DPR:$Vd, DPR:$Vm, DPR:$Vd, pred:$p)>; +def: NEONInstAlias<"vaclt${p}.f16 $Vd, $Vm", + (VACGThq QPR:$Vd, QPR:$Vm, QPR:$Vd, pred:$p)>; +def: NEONInstAlias<"vacle${p}.f16 $Vd, $Vm", + (VACGEhd DPR:$Vd, DPR:$Vm, DPR:$Vd, pred:$p)>; +def: NEONInstAlias<"vacle${p}.f16 $Vd, $Vm", + (VACGEhq QPR:$Vd, QPR:$Vm, QPR:$Vd, pred:$p)>; +} + +// Vector Bitwise Operations. + +def vnotd : PatFrag<(ops node:$in), + (xor node:$in, (bitconvert (v8i8 NEONimmAllOnesV)))>; +def vnotq : PatFrag<(ops node:$in), + (xor node:$in, (bitconvert (v16i8 NEONimmAllOnesV)))>; + + +// VAND : Vector Bitwise AND +def VANDd : N3VDX<0, 0, 0b00, 0b0001, 1, IIC_VBINiD, "vand", + v2i32, v2i32, and, 1>; +def VANDq : N3VQX<0, 0, 0b00, 0b0001, 1, IIC_VBINiQ, "vand", + v4i32, v4i32, and, 1>; + +// VEOR : Vector Bitwise Exclusive OR +def VEORd : N3VDX<1, 0, 0b00, 0b0001, 1, IIC_VBINiD, "veor", + v2i32, v2i32, xor, 1>; +def VEORq : N3VQX<1, 0, 0b00, 0b0001, 1, IIC_VBINiQ, "veor", + v4i32, v4i32, xor, 1>; + +// VORR : Vector Bitwise OR +def VORRd : N3VDX<0, 0, 0b10, 0b0001, 1, IIC_VBINiD, "vorr", + v2i32, v2i32, or, 1>; +def VORRq : N3VQX<0, 0, 0b10, 0b0001, 1, IIC_VBINiQ, "vorr", + v4i32, v4i32, or, 1>; + +def VORRiv4i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 0, 0, 1, + (outs DPR:$Vd), (ins nImmSplatI16:$SIMM, DPR:$src), + IIC_VMOVImm, + "vorr", "i16", "$Vd, $SIMM", "$src = $Vd", + [(set DPR:$Vd, + (v4i16 (NEONvorrImm DPR:$src, timm:$SIMM)))]> { + let Inst{9} = SIMM{9}; +} + +def VORRiv2i32 : N1ModImm<1, 0b000, {0,?,?,1}, 0, 0, 0, 1, + (outs DPR:$Vd), (ins nImmSplatI32:$SIMM, DPR:$src), + IIC_VMOVImm, + "vorr", "i32", "$Vd, $SIMM", "$src = $Vd", + [(set DPR:$Vd, + (v2i32 (NEONvorrImm DPR:$src, timm:$SIMM)))]> { + let Inst{10-9} = SIMM{10-9}; +} + +def VORRiv8i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 1, 0, 1, + (outs QPR:$Vd), (ins nImmSplatI16:$SIMM, QPR:$src), + IIC_VMOVImm, + "vorr", "i16", "$Vd, $SIMM", "$src = $Vd", + [(set QPR:$Vd, + (v8i16 (NEONvorrImm QPR:$src, timm:$SIMM)))]> { + let Inst{9} = SIMM{9}; +} + +def VORRiv4i32 : N1ModImm<1, 0b000, {0,?,?,1}, 0, 1, 0, 1, + (outs QPR:$Vd), (ins nImmSplatI32:$SIMM, QPR:$src), + IIC_VMOVImm, + "vorr", "i32", "$Vd, $SIMM", "$src = $Vd", + [(set QPR:$Vd, + (v4i32 (NEONvorrImm QPR:$src, timm:$SIMM)))]> { + let Inst{10-9} = SIMM{10-9}; +} + + +// VBIC : Vector Bitwise Bit Clear (AND NOT) +let TwoOperandAliasConstraint = "$Vn = $Vd" in { +def VBICd : N3VX<0, 0, 0b01, 0b0001, 0, 1, (outs DPR:$Vd), + (ins DPR:$Vn, DPR:$Vm), N3RegFrm, IIC_VBINiD, + "vbic", "$Vd, $Vn, $Vm", "", + [(set DPR:$Vd, (v2i32 (and DPR:$Vn, + (vnotd DPR:$Vm))))]>; +def VBICq : N3VX<0, 0, 0b01, 0b0001, 1, 1, (outs QPR:$Vd), + (ins QPR:$Vn, QPR:$Vm), N3RegFrm, IIC_VBINiQ, + "vbic", "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (v4i32 (and QPR:$Vn, + (vnotq QPR:$Vm))))]>; +} + +def VBICiv4i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 0, 1, 1, + (outs DPR:$Vd), (ins nImmSplatI16:$SIMM, DPR:$src), + IIC_VMOVImm, + "vbic", "i16", "$Vd, $SIMM", "$src = $Vd", + [(set DPR:$Vd, + (v4i16 (NEONvbicImm DPR:$src, timm:$SIMM)))]> { + let Inst{9} = SIMM{9}; +} + +def VBICiv2i32 : N1ModImm<1, 0b000, {0,?,?,1}, 0, 0, 1, 1, + (outs DPR:$Vd), (ins nImmSplatI32:$SIMM, DPR:$src), + IIC_VMOVImm, + "vbic", "i32", "$Vd, $SIMM", "$src = $Vd", + [(set DPR:$Vd, + (v2i32 (NEONvbicImm DPR:$src, timm:$SIMM)))]> { + let Inst{10-9} = SIMM{10-9}; +} + +def VBICiv8i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 1, 1, 1, + (outs QPR:$Vd), (ins nImmSplatI16:$SIMM, QPR:$src), + IIC_VMOVImm, + "vbic", "i16", "$Vd, $SIMM", "$src = $Vd", + [(set QPR:$Vd, + (v8i16 (NEONvbicImm QPR:$src, timm:$SIMM)))]> { + let Inst{9} = SIMM{9}; +} + +def VBICiv4i32 : N1ModImm<1, 0b000, {0,?,?,1}, 0, 1, 1, 1, + (outs QPR:$Vd), (ins nImmSplatI32:$SIMM, QPR:$src), + IIC_VMOVImm, + "vbic", "i32", "$Vd, $SIMM", "$src = $Vd", + [(set QPR:$Vd, + (v4i32 (NEONvbicImm QPR:$src, timm:$SIMM)))]> { + let Inst{10-9} = SIMM{10-9}; +} + +// VORN : Vector Bitwise OR NOT +def VORNd : N3VX<0, 0, 0b11, 0b0001, 0, 1, (outs DPR:$Vd), + (ins DPR:$Vn, DPR:$Vm), N3RegFrm, IIC_VBINiD, + "vorn", "$Vd, $Vn, $Vm", "", + [(set DPR:$Vd, (v2i32 (or DPR:$Vn, + (vnotd DPR:$Vm))))]>; +def VORNq : N3VX<0, 0, 0b11, 0b0001, 1, 1, (outs QPR:$Vd), + (ins QPR:$Vn, QPR:$Vm), N3RegFrm, IIC_VBINiQ, + "vorn", "$Vd, $Vn, $Vm", "", + [(set QPR:$Vd, (v4i32 (or QPR:$Vn, + (vnotq QPR:$Vm))))]>; + +// VMVN : Vector Bitwise NOT (Immediate) + +let isReMaterializable = 1 in { + +def VMVNv4i16 : N1ModImm<1, 0b000, {1,0,?,0}, 0, 0, 1, 1, (outs DPR:$Vd), + (ins nImmSplatI16:$SIMM), IIC_VMOVImm, + "vmvn", "i16", "$Vd, $SIMM", "", + [(set DPR:$Vd, (v4i16 (NEONvmvnImm timm:$SIMM)))]> { + let Inst{9} = SIMM{9}; +} + +def VMVNv8i16 : N1ModImm<1, 0b000, {1,0,?,0}, 0, 1, 1, 1, (outs QPR:$Vd), + (ins nImmSplatI16:$SIMM), IIC_VMOVImm, + "vmvn", "i16", "$Vd, $SIMM", "", + [(set QPR:$Vd, (v8i16 (NEONvmvnImm timm:$SIMM)))]> { + let Inst{9} = SIMM{9}; +} + +def VMVNv2i32 : N1ModImm<1, 0b000, {?,?,?,?}, 0, 0, 1, 1, (outs DPR:$Vd), + (ins nImmVMOVI32:$SIMM), IIC_VMOVImm, + "vmvn", "i32", "$Vd, $SIMM", "", + [(set DPR:$Vd, (v2i32 (NEONvmvnImm timm:$SIMM)))]> { + let Inst{11-8} = SIMM{11-8}; +} + +def VMVNv4i32 : N1ModImm<1, 0b000, {?,?,?,?}, 0, 1, 1, 1, (outs QPR:$Vd), + (ins nImmVMOVI32:$SIMM), IIC_VMOVImm, + "vmvn", "i32", "$Vd, $SIMM", "", + [(set QPR:$Vd, (v4i32 (NEONvmvnImm timm:$SIMM)))]> { + let Inst{11-8} = SIMM{11-8}; +} +} + +// VMVN : Vector Bitwise NOT +def VMVNd : N2VX<0b11, 0b11, 0b00, 0b00, 0b01011, 0, 0, + (outs DPR:$Vd), (ins DPR:$Vm), IIC_VSUBiD, + "vmvn", "$Vd, $Vm", "", + [(set DPR:$Vd, (v2i32 (vnotd DPR:$Vm)))]>; +def VMVNq : N2VX<0b11, 0b11, 0b00, 0b00, 0b01011, 1, 0, + (outs QPR:$Vd), (ins QPR:$Vm), IIC_VSUBiD, + "vmvn", "$Vd, $Vm", "", + [(set QPR:$Vd, (v4i32 (vnotq QPR:$Vm)))]>; +def : Pat<(v2i32 (vnotd DPR:$src)), (VMVNd DPR:$src)>; +def : Pat<(v4i32 (vnotq QPR:$src)), (VMVNq QPR:$src)>; + +// VBSL : Vector Bitwise Select +def VBSLd : N3VX<1, 0, 0b01, 0b0001, 0, 1, (outs DPR:$Vd), + (ins DPR:$src1, DPR:$Vn, DPR:$Vm), + N3RegFrm, IIC_VCNTiD, + "vbsl", "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set DPR:$Vd, + (v2i32 (NEONvbsl DPR:$src1, DPR:$Vn, DPR:$Vm)))]>; +def : Pat<(v8i8 (int_arm_neon_vbsl (v8i8 DPR:$src1), + (v8i8 DPR:$Vn), (v8i8 DPR:$Vm))), + (VBSLd DPR:$src1, DPR:$Vn, DPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v4i16 (int_arm_neon_vbsl (v4i16 DPR:$src1), + (v4i16 DPR:$Vn), (v4i16 DPR:$Vm))), + (VBSLd DPR:$src1, DPR:$Vn, DPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v2i32 (int_arm_neon_vbsl (v2i32 DPR:$src1), + (v2i32 DPR:$Vn), (v2i32 DPR:$Vm))), + (VBSLd DPR:$src1, DPR:$Vn, DPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v2f32 (int_arm_neon_vbsl (v2f32 DPR:$src1), + (v2f32 DPR:$Vn), (v2f32 DPR:$Vm))), + (VBSLd DPR:$src1, DPR:$Vn, DPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v1i64 (int_arm_neon_vbsl (v1i64 DPR:$src1), + (v1i64 DPR:$Vn), (v1i64 DPR:$Vm))), + (VBSLd DPR:$src1, DPR:$Vn, DPR:$Vm)>, + Requires<[HasNEON]>; + +def : Pat<(v2i32 (or (and DPR:$Vn, DPR:$Vd), + (and DPR:$Vm, (vnotd DPR:$Vd)))), + (VBSLd DPR:$Vd, DPR:$Vn, DPR:$Vm)>, + Requires<[HasNEON]>; + +def : Pat<(v1i64 (or (and DPR:$Vn, DPR:$Vd), + (and DPR:$Vm, (vnotd DPR:$Vd)))), + (VBSLd DPR:$Vd, DPR:$Vn, DPR:$Vm)>, + Requires<[HasNEON]>; + +def VBSLq : N3VX<1, 0, 0b01, 0b0001, 1, 1, (outs QPR:$Vd), + (ins QPR:$src1, QPR:$Vn, QPR:$Vm), + N3RegFrm, IIC_VCNTiQ, + "vbsl", "$Vd, $Vn, $Vm", "$src1 = $Vd", + [(set QPR:$Vd, + (v4i32 (NEONvbsl QPR:$src1, QPR:$Vn, QPR:$Vm)))]>; + +def : Pat<(v16i8 (int_arm_neon_vbsl (v16i8 QPR:$src1), + (v16i8 QPR:$Vn), (v16i8 QPR:$Vm))), + (VBSLq QPR:$src1, QPR:$Vn, QPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v8i16 (int_arm_neon_vbsl (v8i16 QPR:$src1), + (v8i16 QPR:$Vn), (v8i16 QPR:$Vm))), + (VBSLq QPR:$src1, QPR:$Vn, QPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v4i32 (int_arm_neon_vbsl (v4i32 QPR:$src1), + (v4i32 QPR:$Vn), (v4i32 QPR:$Vm))), + (VBSLq QPR:$src1, QPR:$Vn, QPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v4f32 (int_arm_neon_vbsl (v4f32 QPR:$src1), + (v4f32 QPR:$Vn), (v4f32 QPR:$Vm))), + (VBSLq QPR:$src1, QPR:$Vn, QPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v2i64 (int_arm_neon_vbsl (v2i64 QPR:$src1), + (v2i64 QPR:$Vn), (v2i64 QPR:$Vm))), + (VBSLq QPR:$src1, QPR:$Vn, QPR:$Vm)>, + Requires<[HasNEON]>; + +def : Pat<(v4i32 (or (and QPR:$Vn, QPR:$Vd), + (and QPR:$Vm, (vnotq QPR:$Vd)))), + (VBSLq QPR:$Vd, QPR:$Vn, QPR:$Vm)>, + Requires<[HasNEON]>; +def : Pat<(v2i64 (or (and QPR:$Vn, QPR:$Vd), + (and QPR:$Vm, (vnotq QPR:$Vd)))), + (VBSLq QPR:$Vd, QPR:$Vn, QPR:$Vm)>, + Requires<[HasNEON]>; + +// VBIF : Vector Bitwise Insert if False +// like VBSL but with: "vbif $dst, $src3, $src1", "$src2 = $dst", +// FIXME: This instruction's encoding MAY NOT BE correct. +def VBIFd : N3VX<1, 0, 0b11, 0b0001, 0, 1, + (outs DPR:$Vd), (ins DPR:$src1, DPR:$Vn, DPR:$Vm), + N3RegFrm, IIC_VBINiD, + "vbif", "$Vd, $Vn, $Vm", "$src1 = $Vd", + []>; +def VBIFq : N3VX<1, 0, 0b11, 0b0001, 1, 1, + (outs QPR:$Vd), (ins QPR:$src1, QPR:$Vn, QPR:$Vm), + N3RegFrm, IIC_VBINiQ, + "vbif", "$Vd, $Vn, $Vm", "$src1 = $Vd", + []>; + +// VBIT : Vector Bitwise Insert if True +// like VBSL but with: "vbit $dst, $src2, $src1", "$src3 = $dst", +// FIXME: This instruction's encoding MAY NOT BE correct. +def VBITd : N3VX<1, 0, 0b10, 0b0001, 0, 1, + (outs DPR:$Vd), (ins DPR:$src1, DPR:$Vn, DPR:$Vm), + N3RegFrm, IIC_VBINiD, + "vbit", "$Vd, $Vn, $Vm", "$src1 = $Vd", + []>; +def VBITq : N3VX<1, 0, 0b10, 0b0001, 1, 1, + (outs QPR:$Vd), (ins QPR:$src1, QPR:$Vn, QPR:$Vm), + N3RegFrm, IIC_VBINiQ, + "vbit", "$Vd, $Vn, $Vm", "$src1 = $Vd", + []>; + +// VBIT/VBIF are not yet implemented. The TwoAddress pass will not go looking +// for equivalent operations with different register constraints; it just +// inserts copies. + +// Vector Absolute Differences. + +// VABD : Vector Absolute Difference +defm VABDs : N3VInt_QHS<0, 0, 0b0111, 0, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vabd", "s", int_arm_neon_vabds, 1>; +defm VABDu : N3VInt_QHS<1, 0, 0b0111, 0, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vabd", "u", int_arm_neon_vabdu, 1>; +def VABDfd : N3VDInt<1, 0, 0b10, 0b1101, 0, N3RegFrm, IIC_VBIND, + "vabd", "f32", v2f32, v2f32, int_arm_neon_vabds, 1>; +def VABDfq : N3VQInt<1, 0, 0b10, 0b1101, 0, N3RegFrm, IIC_VBINQ, + "vabd", "f32", v4f32, v4f32, int_arm_neon_vabds, 1>; +def VABDhd : N3VDInt<1, 0, 0b11, 0b1101, 0, N3RegFrm, IIC_VBIND, + "vabd", "f16", v4f16, v4f16, int_arm_neon_vabds, 1>, + Requires<[HasNEON, HasFullFP16]>; +def VABDhq : N3VQInt<1, 0, 0b11, 0b1101, 0, N3RegFrm, IIC_VBINQ, + "vabd", "f16", v8f16, v8f16, int_arm_neon_vabds, 1>, + Requires<[HasNEON, HasFullFP16]>; + +// VABDL : Vector Absolute Difference Long (Q = | D - D |) +defm VABDLs : N3VLIntExt_QHS<0,1,0b0111,0, IIC_VSUBi4Q, + "vabdl", "s", int_arm_neon_vabds, zext, 1>; +defm VABDLu : N3VLIntExt_QHS<1,1,0b0111,0, IIC_VSUBi4Q, + "vabdl", "u", int_arm_neon_vabdu, zext, 1>; + +def : Pat<(v8i16 (abs (sub (zext (v8i8 DPR:$opA)), (zext (v8i8 DPR:$opB))))), + (VABDLuv8i16 DPR:$opA, DPR:$opB)>; +def : Pat<(v4i32 (abs (sub (zext (v4i16 DPR:$opA)), (zext (v4i16 DPR:$opB))))), + (VABDLuv4i32 DPR:$opA, DPR:$opB)>; + +// ISD::ABS is not legal for v2i64, so VABDL needs to be matched from the +// shift/xor pattern for ABS. + +def abd_shr : + PatFrag<(ops node:$in1, node:$in2, node:$shift), + (NEONvshrs (sub (zext node:$in1), + (zext node:$in2)), (i32 $shift))>; + +def : Pat<(xor (v4i32 (bitconvert (v2i64 (abd_shr (v2i32 DPR:$opA), (v2i32 DPR:$opB), 63)))), + (v4i32 (bitconvert (v2i64 (add (sub (zext (v2i32 DPR:$opA)), + (zext (v2i32 DPR:$opB))), + (abd_shr (v2i32 DPR:$opA), (v2i32 DPR:$opB), 63)))))), + (VABDLuv2i64 DPR:$opA, DPR:$opB)>; + +// VABA : Vector Absolute Difference and Accumulate +defm VABAs : N3VIntOp_QHS<0,0,0b0111,1, IIC_VABAD, IIC_VABAQ, + "vaba", "s", int_arm_neon_vabds, add>; +defm VABAu : N3VIntOp_QHS<1,0,0b0111,1, IIC_VABAD, IIC_VABAQ, + "vaba", "u", int_arm_neon_vabdu, add>; + +// VABAL : Vector Absolute Difference and Accumulate Long (Q += | D - D |) +defm VABALs : N3VLIntExtOp_QHS<0,1,0b0101,0, IIC_VABAD, + "vabal", "s", int_arm_neon_vabds, zext, add>; +defm VABALu : N3VLIntExtOp_QHS<1,1,0b0101,0, IIC_VABAD, + "vabal", "u", int_arm_neon_vabdu, zext, add>; + +// Vector Maximum and Minimum. + +// VMAX : Vector Maximum +defm VMAXs : N3VInt_QHS<0, 0, 0b0110, 0, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vmax", "s", smax, 1>; +defm VMAXu : N3VInt_QHS<1, 0, 0b0110, 0, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vmax", "u", umax, 1>; +def VMAXfd : N3VDInt<0, 0, 0b00, 0b1111, 0, N3RegFrm, IIC_VBIND, + "vmax", "f32", + v2f32, v2f32, fmaxnan, 1>; +def VMAXfq : N3VQInt<0, 0, 0b00, 0b1111, 0, N3RegFrm, IIC_VBINQ, + "vmax", "f32", + v4f32, v4f32, fmaxnan, 1>; +def VMAXhd : N3VDInt<0, 0, 0b01, 0b1111, 0, N3RegFrm, IIC_VBIND, + "vmax", "f16", + v4f16, v4f16, fmaxnan, 1>, + Requires<[HasNEON, HasFullFP16]>; +def VMAXhq : N3VQInt<0, 0, 0b01, 0b1111, 0, N3RegFrm, IIC_VBINQ, + "vmax", "f16", + v8f16, v8f16, fmaxnan, 1>, + Requires<[HasNEON, HasFullFP16]>; + +// VMAXNM +let PostEncoderMethod = "NEONThumb2V8PostEncoder", DecoderNamespace = "v8NEON" in { + def VMAXNMNDf : N3VDIntnp<0b00110, 0b00, 0b1111, 0, 1, + N3RegFrm, NoItinerary, "vmaxnm", "f32", + v2f32, v2f32, fmaxnum, 1>, + Requires<[HasV8, HasNEON]>; + def VMAXNMNQf : N3VQIntnp<0b00110, 0b00, 0b1111, 1, 1, + N3RegFrm, NoItinerary, "vmaxnm", "f32", + v4f32, v4f32, fmaxnum, 1>, + Requires<[HasV8, HasNEON]>; + def VMAXNMNDh : N3VDIntnp<0b00110, 0b01, 0b1111, 0, 1, + N3RegFrm, NoItinerary, "vmaxnm", "f16", + v4f16, v4f16, fmaxnum, 1>, + Requires<[HasV8, HasNEON, HasFullFP16]>; + def VMAXNMNQh : N3VQIntnp<0b00110, 0b01, 0b1111, 1, 1, + N3RegFrm, NoItinerary, "vmaxnm", "f16", + v8f16, v8f16, fmaxnum, 1>, + Requires<[HasV8, HasNEON, HasFullFP16]>; +} + +// VMIN : Vector Minimum +defm VMINs : N3VInt_QHS<0, 0, 0b0110, 1, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vmin", "s", smin, 1>; +defm VMINu : N3VInt_QHS<1, 0, 0b0110, 1, N3RegFrm, + IIC_VSUBi4D, IIC_VSUBi4D, IIC_VSUBi4Q, IIC_VSUBi4Q, + "vmin", "u", umin, 1>; +def VMINfd : N3VDInt<0, 0, 0b10, 0b1111, 0, N3RegFrm, IIC_VBIND, + "vmin", "f32", + v2f32, v2f32, fminnan, 1>; +def VMINfq : N3VQInt<0, 0, 0b10, 0b1111, 0, N3RegFrm, IIC_VBINQ, + "vmin", "f32", + v4f32, v4f32, fminnan, 1>; +def VMINhd : N3VDInt<0, 0, 0b11, 0b1111, 0, N3RegFrm, IIC_VBIND, + "vmin", "f16", + v4f16, v4f16, fminnan, 1>, + Requires<[HasNEON, HasFullFP16]>; +def VMINhq : N3VQInt<0, 0, 0b11, 0b1111, 0, N3RegFrm, IIC_VBINQ, + "vmin", "f16", + v8f16, v8f16, fminnan, 1>, + Requires<[HasNEON, HasFullFP16]>; + +// VMINNM +let PostEncoderMethod = "NEONThumb2V8PostEncoder", DecoderNamespace = "v8NEON" in { + def VMINNMNDf : N3VDIntnp<0b00110, 0b10, 0b1111, 0, 1, + N3RegFrm, NoItinerary, "vminnm", "f32", + v2f32, v2f32, fminnum, 1>, + Requires<[HasV8, HasNEON]>; + def VMINNMNQf : N3VQIntnp<0b00110, 0b10, 0b1111, 1, 1, + N3RegFrm, NoItinerary, "vminnm", "f32", + v4f32, v4f32, fminnum, 1>, + Requires<[HasV8, HasNEON]>; + def VMINNMNDh : N3VDIntnp<0b00110, 0b11, 0b1111, 0, 1, + N3RegFrm, NoItinerary, "vminnm", "f16", + v4f16, v4f16, fminnum, 1>, + Requires<[HasV8, HasNEON, HasFullFP16]>; + def VMINNMNQh : N3VQIntnp<0b00110, 0b11, 0b1111, 1, 1, + N3RegFrm, NoItinerary, "vminnm", "f16", + v8f16, v8f16, fminnum, 1>, + Requires<[HasV8, HasNEON, HasFullFP16]>; +} + +// Vector Pairwise Operations. + +// VPADD : Vector Pairwise Add +def VPADDi8 : N3VDInt<0, 0, 0b00, 0b1011, 1, N3RegFrm, IIC_VSHLiD, + "vpadd", "i8", + v8i8, v8i8, int_arm_neon_vpadd, 0>; +def VPADDi16 : N3VDInt<0, 0, 0b01, 0b1011, 1, N3RegFrm, IIC_VSHLiD, + "vpadd", "i16", + v4i16, v4i16, int_arm_neon_vpadd, 0>; +def VPADDi32 : N3VDInt<0, 0, 0b10, 0b1011, 1, N3RegFrm, IIC_VSHLiD, + "vpadd", "i32", + v2i32, v2i32, int_arm_neon_vpadd, 0>; +def VPADDf : N3VDInt<1, 0, 0b00, 0b1101, 0, N3RegFrm, + IIC_VPBIND, "vpadd", "f32", + v2f32, v2f32, int_arm_neon_vpadd, 0>; +def VPADDh : N3VDInt<1, 0, 0b01, 0b1101, 0, N3RegFrm, + IIC_VPBIND, "vpadd", "f16", + v4f16, v4f16, int_arm_neon_vpadd, 0>, + Requires<[HasNEON, HasFullFP16]>; + +// VPADDL : Vector Pairwise Add Long +defm VPADDLs : N2VPLInt_QHS<0b11, 0b11, 0b00, 0b00100, 0, "vpaddl", "s", + int_arm_neon_vpaddls>; +defm VPADDLu : N2VPLInt_QHS<0b11, 0b11, 0b00, 0b00101, 0, "vpaddl", "u", + int_arm_neon_vpaddlu>; + +// VPADAL : Vector Pairwise Add and Accumulate Long +defm VPADALs : N2VPLInt2_QHS<0b11, 0b11, 0b00, 0b01100, 0, "vpadal", "s", + int_arm_neon_vpadals>; +defm VPADALu : N2VPLInt2_QHS<0b11, 0b11, 0b00, 0b01101, 0, "vpadal", "u", + int_arm_neon_vpadalu>; + +// VPMAX : Vector Pairwise Maximum +def VPMAXs8 : N3VDInt<0, 0, 0b00, 0b1010, 0, N3RegFrm, IIC_VSUBi4D, "vpmax", + "s8", v8i8, v8i8, int_arm_neon_vpmaxs, 0>; +def VPMAXs16 : N3VDInt<0, 0, 0b01, 0b1010, 0, N3RegFrm, IIC_VSUBi4D, "vpmax", + "s16", v4i16, v4i16, int_arm_neon_vpmaxs, 0>; +def VPMAXs32 : N3VDInt<0, 0, 0b10, 0b1010, 0, N3RegFrm, IIC_VSUBi4D, "vpmax", + "s32", v2i32, v2i32, int_arm_neon_vpmaxs, 0>; +def VPMAXu8 : N3VDInt<1, 0, 0b00, 0b1010, 0, N3RegFrm, IIC_VSUBi4D, "vpmax", + "u8", v8i8, v8i8, int_arm_neon_vpmaxu, 0>; +def VPMAXu16 : N3VDInt<1, 0, 0b01, 0b1010, 0, N3RegFrm, IIC_VSUBi4D, "vpmax", + "u16", v4i16, v4i16, int_arm_neon_vpmaxu, 0>; +def VPMAXu32 : N3VDInt<1, 0, 0b10, 0b1010, 0, N3RegFrm, IIC_VSUBi4D, "vpmax", + "u32", v2i32, v2i32, int_arm_neon_vpmaxu, 0>; +def VPMAXf : N3VDInt<1, 0, 0b00, 0b1111, 0, N3RegFrm, IIC_VPBIND, "vpmax", + "f32", v2f32, v2f32, int_arm_neon_vpmaxs, 0>; +def VPMAXh : N3VDInt<1, 0, 0b01, 0b1111, 0, N3RegFrm, IIC_VPBIND, "vpmax", + "f16", v4f16, v4f16, int_arm_neon_vpmaxs, 0>, + Requires<[HasNEON, HasFullFP16]>; + +// VPMIN : Vector Pairwise Minimum +def VPMINs8 : N3VDInt<0, 0, 0b00, 0b1010, 1, N3RegFrm, IIC_VSUBi4D, "vpmin", + "s8", v8i8, v8i8, int_arm_neon_vpmins, 0>; +def VPMINs16 : N3VDInt<0, 0, 0b01, 0b1010, 1, N3RegFrm, IIC_VSUBi4D, "vpmin", + "s16", v4i16, v4i16, int_arm_neon_vpmins, 0>; +def VPMINs32 : N3VDInt<0, 0, 0b10, 0b1010, 1, N3RegFrm, IIC_VSUBi4D, "vpmin", + "s32", v2i32, v2i32, int_arm_neon_vpmins, 0>; +def VPMINu8 : N3VDInt<1, 0, 0b00, 0b1010, 1, N3RegFrm, IIC_VSUBi4D, "vpmin", + "u8", v8i8, v8i8, int_arm_neon_vpminu, 0>; +def VPMINu16 : N3VDInt<1, 0, 0b01, 0b1010, 1, N3RegFrm, IIC_VSUBi4D, "vpmin", + "u16", v4i16, v4i16, int_arm_neon_vpminu, 0>; +def VPMINu32 : N3VDInt<1, 0, 0b10, 0b1010, 1, N3RegFrm, IIC_VSUBi4D, "vpmin", + "u32", v2i32, v2i32, int_arm_neon_vpminu, 0>; +def VPMINf : N3VDInt<1, 0, 0b10, 0b1111, 0, N3RegFrm, IIC_VPBIND, "vpmin", + "f32", v2f32, v2f32, int_arm_neon_vpmins, 0>; +def VPMINh : N3VDInt<1, 0, 0b11, 0b1111, 0, N3RegFrm, IIC_VPBIND, "vpmin", + "f16", v4f16, v4f16, int_arm_neon_vpmins, 0>, + Requires<[HasNEON, HasFullFP16]>; + +// Vector Reciprocal and Reciprocal Square Root Estimate and Step. + +// VRECPE : Vector Reciprocal Estimate +def VRECPEd : N2VDInt<0b11, 0b11, 0b10, 0b11, 0b01000, 0, + IIC_VUNAD, "vrecpe", "u32", + v2i32, v2i32, int_arm_neon_vrecpe>; +def VRECPEq : N2VQInt<0b11, 0b11, 0b10, 0b11, 0b01000, 0, + IIC_VUNAQ, "vrecpe", "u32", + v4i32, v4i32, int_arm_neon_vrecpe>; +def VRECPEfd : N2VDInt<0b11, 0b11, 0b10, 0b11, 0b01010, 0, + IIC_VUNAD, "vrecpe", "f32", + v2f32, v2f32, int_arm_neon_vrecpe>; +def VRECPEfq : N2VQInt<0b11, 0b11, 0b10, 0b11, 0b01010, 0, + IIC_VUNAQ, "vrecpe", "f32", + v4f32, v4f32, int_arm_neon_vrecpe>; +def VRECPEhd : N2VDInt<0b11, 0b11, 0b01, 0b11, 0b01010, 0, + IIC_VUNAD, "vrecpe", "f16", + v4f16, v4f16, int_arm_neon_vrecpe>, + Requires<[HasNEON, HasFullFP16]>; +def VRECPEhq : N2VQInt<0b11, 0b11, 0b01, 0b11, 0b01010, 0, + IIC_VUNAQ, "vrecpe", "f16", + v8f16, v8f16, int_arm_neon_vrecpe>, + Requires<[HasNEON, HasFullFP16]>; + +// VRECPS : Vector Reciprocal Step +def VRECPSfd : N3VDInt<0, 0, 0b00, 0b1111, 1, N3RegFrm, + IIC_VRECSD, "vrecps", "f32", + v2f32, v2f32, int_arm_neon_vrecps, 1>; +def VRECPSfq : N3VQInt<0, 0, 0b00, 0b1111, 1, N3RegFrm, + IIC_VRECSQ, "vrecps", "f32", + v4f32, v4f32, int_arm_neon_vrecps, 1>; +def VRECPShd : N3VDInt<0, 0, 0b01, 0b1111, 1, N3RegFrm, + IIC_VRECSD, "vrecps", "f16", + v4f16, v4f16, int_arm_neon_vrecps, 1>, + Requires<[HasNEON, HasFullFP16]>; +def VRECPShq : N3VQInt<0, 0, 0b01, 0b1111, 1, N3RegFrm, + IIC_VRECSQ, "vrecps", "f16", + v8f16, v8f16, int_arm_neon_vrecps, 1>, + Requires<[HasNEON, HasFullFP16]>; + +// VRSQRTE : Vector Reciprocal Square Root Estimate +def VRSQRTEd : N2VDInt<0b11, 0b11, 0b10, 0b11, 0b01001, 0, + IIC_VUNAD, "vrsqrte", "u32", + v2i32, v2i32, int_arm_neon_vrsqrte>; +def VRSQRTEq : N2VQInt<0b11, 0b11, 0b10, 0b11, 0b01001, 0, + IIC_VUNAQ, "vrsqrte", "u32", + v4i32, v4i32, int_arm_neon_vrsqrte>; +def VRSQRTEfd : N2VDInt<0b11, 0b11, 0b10, 0b11, 0b01011, 0, + IIC_VUNAD, "vrsqrte", "f32", + v2f32, v2f32, int_arm_neon_vrsqrte>; +def VRSQRTEfq : N2VQInt<0b11, 0b11, 0b10, 0b11, 0b01011, 0, + IIC_VUNAQ, "vrsqrte", "f32", + v4f32, v4f32, int_arm_neon_vrsqrte>; +def VRSQRTEhd : N2VDInt<0b11, 0b11, 0b01, 0b11, 0b01011, 0, + IIC_VUNAD, "vrsqrte", "f16", + v4f16, v4f16, int_arm_neon_vrsqrte>, + Requires<[HasNEON, HasFullFP16]>; +def VRSQRTEhq : N2VQInt<0b11, 0b11, 0b01, 0b11, 0b01011, 0, + IIC_VUNAQ, "vrsqrte", "f16", + v8f16, v8f16, int_arm_neon_vrsqrte>, + Requires<[HasNEON, HasFullFP16]>; + +// VRSQRTS : Vector Reciprocal Square Root Step +def VRSQRTSfd : N3VDInt<0, 0, 0b10, 0b1111, 1, N3RegFrm, + IIC_VRECSD, "vrsqrts", "f32", + v2f32, v2f32, int_arm_neon_vrsqrts, 1>; +def VRSQRTSfq : N3VQInt<0, 0, 0b10, 0b1111, 1, N3RegFrm, + IIC_VRECSQ, "vrsqrts", "f32", + v4f32, v4f32, int_arm_neon_vrsqrts, 1>; +def VRSQRTShd : N3VDInt<0, 0, 0b11, 0b1111, 1, N3RegFrm, + IIC_VRECSD, "vrsqrts", "f16", + v4f16, v4f16, int_arm_neon_vrsqrts, 1>, + Requires<[HasNEON, HasFullFP16]>; +def VRSQRTShq : N3VQInt<0, 0, 0b11, 0b1111, 1, N3RegFrm, + IIC_VRECSQ, "vrsqrts", "f16", + v8f16, v8f16, int_arm_neon_vrsqrts, 1>, + Requires<[HasNEON, HasFullFP16]>; + +// Vector Shifts. + +// VSHL : Vector Shift +defm VSHLs : N3VInt_QHSDSh<0, 0, 0b0100, 0, N3RegVShFrm, + IIC_VSHLiD, IIC_VSHLiD, IIC_VSHLiQ, IIC_VSHLiQ, + "vshl", "s", int_arm_neon_vshifts>; +defm VSHLu : N3VInt_QHSDSh<1, 0, 0b0100, 0, N3RegVShFrm, + IIC_VSHLiD, IIC_VSHLiD, IIC_VSHLiQ, IIC_VSHLiQ, + "vshl", "u", int_arm_neon_vshiftu>; + +// VSHL : Vector Shift Left (Immediate) +defm VSHLi : N2VShL_QHSD<0, 1, 0b0101, 1, IIC_VSHLiD, "vshl", "i", NEONvshl>; + +// VSHR : Vector Shift Right (Immediate) +defm VSHRs : N2VShR_QHSD<0, 1, 0b0000, 1, IIC_VSHLiD, "vshr", "s", "VSHRs", + NEONvshrs>; +defm VSHRu : N2VShR_QHSD<1, 1, 0b0000, 1, IIC_VSHLiD, "vshr", "u", "VSHRu", + NEONvshru>; + +// VSHLL : Vector Shift Left Long +defm VSHLLs : N2VLSh_QHS<0, 1, 0b1010, 0, 0, 1, "vshll", "s", + PatFrag<(ops node:$LHS, node:$RHS), (NEONvshl (sext node:$LHS), node:$RHS)>>; +defm VSHLLu : N2VLSh_QHS<1, 1, 0b1010, 0, 0, 1, "vshll", "u", + PatFrag<(ops node:$LHS, node:$RHS), (NEONvshl (zext node:$LHS), node:$RHS)>>; + +// VSHLL : Vector Shift Left Long (with maximum shift count) +class N2VLShMax<bit op24, bit op23, bits<6> op21_16, bits<4> op11_8, bit op7, + bit op6, bit op4, string OpcodeStr, string Dt, ValueType ResTy, + ValueType OpTy, Operand ImmTy> + : N2VLSh<op24, op23, op11_8, op7, op6, op4, OpcodeStr, Dt, + ResTy, OpTy, ImmTy, null_frag> { + let Inst{21-16} = op21_16; + let DecoderMethod = "DecodeVSHLMaxInstruction"; +} +def VSHLLi8 : N2VLShMax<1, 1, 0b110010, 0b0011, 0, 0, 0, "vshll", "i8", + v8i16, v8i8, imm8>; +def VSHLLi16 : N2VLShMax<1, 1, 0b110110, 0b0011, 0, 0, 0, "vshll", "i16", + v4i32, v4i16, imm16>; +def VSHLLi32 : N2VLShMax<1, 1, 0b111010, 0b0011, 0, 0, 0, "vshll", "i32", + v2i64, v2i32, imm32>; + +def : Pat<(v8i16 (NEONvshl (zext (v8i8 DPR:$Rn)), (i32 8))), + (VSHLLi8 DPR:$Rn, 8)>; +def : Pat<(v4i32 (NEONvshl (zext (v4i16 DPR:$Rn)), (i32 16))), + (VSHLLi16 DPR:$Rn, 16)>; +def : Pat<(v2i64 (NEONvshl (zext (v2i32 DPR:$Rn)), (i32 32))), + (VSHLLi32 DPR:$Rn, 32)>; +def : Pat<(v8i16 (NEONvshl (sext (v8i8 DPR:$Rn)), (i32 8))), + (VSHLLi8 DPR:$Rn, 8)>; +def : Pat<(v4i32 (NEONvshl (sext (v4i16 DPR:$Rn)), (i32 16))), + (VSHLLi16 DPR:$Rn, 16)>; +def : Pat<(v2i64 (NEONvshl (sext (v2i32 DPR:$Rn)), (i32 32))), + (VSHLLi32 DPR:$Rn, 32)>; +def : Pat<(v8i16 (NEONvshl (anyext (v8i8 DPR:$Rn)), (i32 8))), + (VSHLLi8 DPR:$Rn, 8)>; +def : Pat<(v4i32 (NEONvshl (anyext (v4i16 DPR:$Rn)), (i32 16))), + (VSHLLi16 DPR:$Rn, 16)>; +def : Pat<(v2i64 (NEONvshl (anyext (v2i32 DPR:$Rn)), (i32 32))), + (VSHLLi32 DPR:$Rn, 32)>; + +// VSHRN : Vector Shift Right and Narrow +defm VSHRN : N2VNSh_HSD<0,1,0b1000,0,0,1, IIC_VSHLiD, "vshrn", "i", + PatFrag<(ops node:$Rn, node:$amt), + (trunc (NEONvshrs node:$Rn, node:$amt))>>; + +def : Pat<(v8i8 (trunc (NEONvshru (v8i16 QPR:$Vn), shr_imm8:$amt))), + (VSHRNv8i8 QPR:$Vn, shr_imm8:$amt)>; +def : Pat<(v4i16 (trunc (NEONvshru (v4i32 QPR:$Vn), shr_imm16:$amt))), + (VSHRNv4i16 QPR:$Vn, shr_imm16:$amt)>; +def : Pat<(v2i32 (trunc (NEONvshru (v2i64 QPR:$Vn), shr_imm32:$amt))), + (VSHRNv2i32 QPR:$Vn, shr_imm32:$amt)>; + +// VRSHL : Vector Rounding Shift +defm VRSHLs : N3VInt_QHSDSh<0, 0, 0b0101, 0, N3RegVShFrm, + IIC_VSHLi4D, IIC_VSHLi4D, IIC_VSHLi4Q, IIC_VSHLi4Q, + "vrshl", "s", int_arm_neon_vrshifts>; +defm VRSHLu : N3VInt_QHSDSh<1, 0, 0b0101, 0, N3RegVShFrm, + IIC_VSHLi4D, IIC_VSHLi4D, IIC_VSHLi4Q, IIC_VSHLi4Q, + "vrshl", "u", int_arm_neon_vrshiftu>; +// VRSHR : Vector Rounding Shift Right +defm VRSHRs : N2VShR_QHSD<0,1,0b0010,1, IIC_VSHLi4D, "vrshr", "s", "VRSHRs", + NEONvrshrs>; +defm VRSHRu : N2VShR_QHSD<1,1,0b0010,1, IIC_VSHLi4D, "vrshr", "u", "VRSHRu", + NEONvrshru>; + +// VRSHRN : Vector Rounding Shift Right and Narrow +defm VRSHRN : N2VNSh_HSD<0, 1, 0b1000, 0, 1, 1, IIC_VSHLi4D, "vrshrn", "i", + NEONvrshrn>; + +// VQSHL : Vector Saturating Shift +defm VQSHLs : N3VInt_QHSDSh<0, 0, 0b0100, 1, N3RegVShFrm, + IIC_VSHLi4D, IIC_VSHLi4D, IIC_VSHLi4Q, IIC_VSHLi4Q, + "vqshl", "s", int_arm_neon_vqshifts>; +defm VQSHLu : N3VInt_QHSDSh<1, 0, 0b0100, 1, N3RegVShFrm, + IIC_VSHLi4D, IIC_VSHLi4D, IIC_VSHLi4Q, IIC_VSHLi4Q, + "vqshl", "u", int_arm_neon_vqshiftu>; +// VQSHL : Vector Saturating Shift Left (Immediate) +defm VQSHLsi : N2VShL_QHSD<0,1,0b0111,1, IIC_VSHLi4D, "vqshl", "s",NEONvqshls>; +defm VQSHLui : N2VShL_QHSD<1,1,0b0111,1, IIC_VSHLi4D, "vqshl", "u",NEONvqshlu>; + +// VQSHLU : Vector Saturating Shift Left (Immediate, Unsigned) +defm VQSHLsu : N2VShL_QHSD<1,1,0b0110,1, IIC_VSHLi4D,"vqshlu","s",NEONvqshlsu>; + +// VQSHRN : Vector Saturating Shift Right and Narrow +defm VQSHRNs : N2VNSh_HSD<0, 1, 0b1001, 0, 0, 1, IIC_VSHLi4D, "vqshrn", "s", + NEONvqshrns>; +defm VQSHRNu : N2VNSh_HSD<1, 1, 0b1001, 0, 0, 1, IIC_VSHLi4D, "vqshrn", "u", + NEONvqshrnu>; + +// VQSHRUN : Vector Saturating Shift Right and Narrow (Unsigned) +defm VQSHRUN : N2VNSh_HSD<1, 1, 0b1000, 0, 0, 1, IIC_VSHLi4D, "vqshrun", "s", + NEONvqshrnsu>; + +// VQRSHL : Vector Saturating Rounding Shift +defm VQRSHLs : N3VInt_QHSDSh<0, 0, 0b0101, 1, N3RegVShFrm, + IIC_VSHLi4D, IIC_VSHLi4D, IIC_VSHLi4Q, IIC_VSHLi4Q, + "vqrshl", "s", int_arm_neon_vqrshifts>; +defm VQRSHLu : N3VInt_QHSDSh<1, 0, 0b0101, 1, N3RegVShFrm, + IIC_VSHLi4D, IIC_VSHLi4D, IIC_VSHLi4Q, IIC_VSHLi4Q, + "vqrshl", "u", int_arm_neon_vqrshiftu>; + +// VQRSHRN : Vector Saturating Rounding Shift Right and Narrow +defm VQRSHRNs : N2VNSh_HSD<0, 1, 0b1001, 0, 1, 1, IIC_VSHLi4D, "vqrshrn", "s", + NEONvqrshrns>; +defm VQRSHRNu : N2VNSh_HSD<1, 1, 0b1001, 0, 1, 1, IIC_VSHLi4D, "vqrshrn", "u", + NEONvqrshrnu>; + +// VQRSHRUN : Vector Saturating Rounding Shift Right and Narrow (Unsigned) +defm VQRSHRUN : N2VNSh_HSD<1, 1, 0b1000, 0, 1, 1, IIC_VSHLi4D, "vqrshrun", "s", + NEONvqrshrnsu>; + +// VSRA : Vector Shift Right and Accumulate +defm VSRAs : N2VShAdd_QHSD<0, 1, 0b0001, 1, "vsra", "s", NEONvshrs>; +defm VSRAu : N2VShAdd_QHSD<1, 1, 0b0001, 1, "vsra", "u", NEONvshru>; +// VRSRA : Vector Rounding Shift Right and Accumulate +defm VRSRAs : N2VShAdd_QHSD<0, 1, 0b0011, 1, "vrsra", "s", NEONvrshrs>; +defm VRSRAu : N2VShAdd_QHSD<1, 1, 0b0011, 1, "vrsra", "u", NEONvrshru>; + +// VSLI : Vector Shift Left and Insert +defm VSLI : N2VShInsL_QHSD<1, 1, 0b0101, 1, "vsli">; + +// VSRI : Vector Shift Right and Insert +defm VSRI : N2VShInsR_QHSD<1, 1, 0b0100, 1, "vsri">; + +// Vector Absolute and Saturating Absolute. + +// VABS : Vector Absolute Value +defm VABS : N2VInt_QHS<0b11, 0b11, 0b01, 0b00110, 0, + IIC_VUNAiD, IIC_VUNAiQ, "vabs", "s", abs>; +def VABSfd : N2VD<0b11, 0b11, 0b10, 0b01, 0b01110, 0, + "vabs", "f32", + v2f32, v2f32, fabs>; +def VABSfq : N2VQ<0b11, 0b11, 0b10, 0b01, 0b01110, 0, + "vabs", "f32", + v4f32, v4f32, fabs>; +def VABShd : N2VD<0b11, 0b11, 0b01, 0b01, 0b01110, 0, + "vabs", "f16", + v4f16, v4f16, fabs>, + Requires<[HasNEON, HasFullFP16]>; +def VABShq : N2VQ<0b11, 0b11, 0b01, 0b01, 0b01110, 0, + "vabs", "f16", + v8f16, v8f16, fabs>, + Requires<[HasNEON, HasFullFP16]>; + +// VQABS : Vector Saturating Absolute Value +defm VQABS : N2VInt_QHS<0b11, 0b11, 0b00, 0b01110, 0, + IIC_VQUNAiD, IIC_VQUNAiQ, "vqabs", "s", + int_arm_neon_vqabs>; + +// Vector Negate. + +def vnegd : PatFrag<(ops node:$in), + (sub (bitconvert (v2i32 NEONimmAllZerosV)), node:$in)>; +def vnegq : PatFrag<(ops node:$in), + (sub (bitconvert (v4i32 NEONimmAllZerosV)), node:$in)>; + +class VNEGD<bits<2> size, string OpcodeStr, string Dt, ValueType Ty> + : N2V<0b11, 0b11, size, 0b01, 0b00111, 0, 0, (outs DPR:$Vd), (ins DPR:$Vm), + IIC_VSHLiD, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set DPR:$Vd, (Ty (vnegd DPR:$Vm)))]>; +class VNEGQ<bits<2> size, string OpcodeStr, string Dt, ValueType Ty> + : N2V<0b11, 0b11, size, 0b01, 0b00111, 1, 0, (outs QPR:$Vd), (ins QPR:$Vm), + IIC_VSHLiQ, OpcodeStr, Dt, "$Vd, $Vm", "", + [(set QPR:$Vd, (Ty (vnegq QPR:$Vm)))]>; + +// VNEG : Vector Negate (integer) +def VNEGs8d : VNEGD<0b00, "vneg", "s8", v8i8>; +def VNEGs16d : VNEGD<0b01, "vneg", "s16", v4i16>; +def VNEGs32d : VNEGD<0b10, "vneg", "s32", v2i32>; +def VNEGs8q : VNEGQ<0b00, "vneg", "s8", v16i8>; +def VNEGs16q : VNEGQ<0b01, "vneg", "s16", v8i16>; +def VNEGs32q : VNEGQ<0b10, "vneg", "s32", v4i32>; + +// VNEG : Vector Negate (floating-point) +def VNEGfd : N2V<0b11, 0b11, 0b10, 0b01, 0b01111, 0, 0, + (outs DPR:$Vd), (ins DPR:$Vm), IIC_VUNAD, + "vneg", "f32", "$Vd, $Vm", "", + [(set DPR:$Vd, (v2f32 (fneg DPR:$Vm)))]>; +def VNEGf32q : N2V<0b11, 0b11, 0b10, 0b01, 0b01111, 1, 0, + (outs QPR:$Vd), (ins QPR:$Vm), IIC_VUNAQ, + "vneg", "f32", "$Vd, $Vm", "", + [(set QPR:$Vd, (v4f32 (fneg QPR:$Vm)))]>; +def VNEGhd : N2V<0b11, 0b11, 0b01, 0b01, 0b01111, 0, 0, + (outs DPR:$Vd), (ins DPR:$Vm), IIC_VUNAD, + "vneg", "f16", "$Vd, $Vm", "", + [(set DPR:$Vd, (v4f16 (fneg DPR:$Vm)))]>, + Requires<[HasNEON, HasFullFP16]>; +def VNEGhq : N2V<0b11, 0b11, 0b01, 0b01, 0b01111, 1, 0, + (outs QPR:$Vd), (ins QPR:$Vm), IIC_VUNAQ, + "vneg", "f16", "$Vd, $Vm", "", + [(set QPR:$Vd, (v8f16 (fneg QPR:$Vm)))]>, + Requires<[HasNEON, HasFullFP16]>; + +def : Pat<(v8i8 (vnegd DPR:$src)), (VNEGs8d DPR:$src)>; +def : Pat<(v4i16 (vnegd DPR:$src)), (VNEGs16d DPR:$src)>; +def : Pat<(v2i32 (vnegd DPR:$src)), (VNEGs32d DPR:$src)>; +def : Pat<(v16i8 (vnegq QPR:$src)), (VNEGs8q QPR:$src)>; +def : Pat<(v8i16 (vnegq QPR:$src)), (VNEGs16q QPR:$src)>; +def : Pat<(v4i32 (vnegq QPR:$src)), (VNEGs32q QPR:$src)>; + +// VQNEG : Vector Saturating Negate +defm VQNEG : N2VInt_QHS<0b11, 0b11, 0b00, 0b01111, 0, + IIC_VQUNAiD, IIC_VQUNAiQ, "vqneg", "s", + int_arm_neon_vqneg>; + +// Vector Bit Counting Operations. + +// VCLS : Vector Count Leading Sign Bits +defm VCLS : N2VInt_QHS<0b11, 0b11, 0b00, 0b01000, 0, + IIC_VCNTiD, IIC_VCNTiQ, "vcls", "s", + int_arm_neon_vcls>; +// VCLZ : Vector Count Leading Zeros +defm VCLZ : N2VInt_QHS<0b11, 0b11, 0b00, 0b01001, 0, + IIC_VCNTiD, IIC_VCNTiQ, "vclz", "i", + ctlz>; +// VCNT : Vector Count One Bits +def VCNTd : N2VDInt<0b11, 0b11, 0b00, 0b00, 0b01010, 0, + IIC_VCNTiD, "vcnt", "8", + v8i8, v8i8, ctpop>; +def VCNTq : N2VQInt<0b11, 0b11, 0b00, 0b00, 0b01010, 0, + IIC_VCNTiQ, "vcnt", "8", + v16i8, v16i8, ctpop>; + +// Vector Swap +def VSWPd : N2VX<0b11, 0b11, 0b00, 0b10, 0b00000, 0, 0, + (outs DPR:$Vd, DPR:$Vm), (ins DPR:$in1, DPR:$in2), + NoItinerary, "vswp", "$Vd, $Vm", "$in1 = $Vd, $in2 = $Vm", + []>; +def VSWPq : N2VX<0b11, 0b11, 0b00, 0b10, 0b00000, 1, 0, + (outs QPR:$Vd, QPR:$Vm), (ins QPR:$in1, QPR:$in2), + NoItinerary, "vswp", "$Vd, $Vm", "$in1 = $Vd, $in2 = $Vm", + []>; + +// Vector Move Operations. + +// VMOV : Vector Move (Register) +def : NEONInstAlias<"vmov${p} $Vd, $Vm", + (VORRd DPR:$Vd, DPR:$Vm, DPR:$Vm, pred:$p)>; +def : NEONInstAlias<"vmov${p} $Vd, $Vm", + (VORRq QPR:$Vd, QPR:$Vm, QPR:$Vm, pred:$p)>; + +// VMOV : Vector Move (Immediate) + +// Although VMOVs are not strictly speaking cheap, they are as expensive +// as their copies counterpart (VORR), so we should prefer rematerialization +// over splitting when it applies. +let isReMaterializable = 1, isAsCheapAsAMove=1 in { +def VMOVv8i8 : N1ModImm<1, 0b000, 0b1110, 0, 0, 0, 1, (outs DPR:$Vd), + (ins nImmSplatI8:$SIMM), IIC_VMOVImm, + "vmov", "i8", "$Vd, $SIMM", "", + [(set DPR:$Vd, (v8i8 (NEONvmovImm timm:$SIMM)))]>; +def VMOVv16i8 : N1ModImm<1, 0b000, 0b1110, 0, 1, 0, 1, (outs QPR:$Vd), + (ins nImmSplatI8:$SIMM), IIC_VMOVImm, + "vmov", "i8", "$Vd, $SIMM", "", + [(set QPR:$Vd, (v16i8 (NEONvmovImm timm:$SIMM)))]>; + +def VMOVv4i16 : N1ModImm<1, 0b000, {1,0,?,0}, 0, 0, 0, 1, (outs DPR:$Vd), + (ins nImmSplatI16:$SIMM), IIC_VMOVImm, + "vmov", "i16", "$Vd, $SIMM", "", + [(set DPR:$Vd, (v4i16 (NEONvmovImm timm:$SIMM)))]> { + let Inst{9} = SIMM{9}; +} + +def VMOVv8i16 : N1ModImm<1, 0b000, {1,0,?,0}, 0, 1, 0, 1, (outs QPR:$Vd), + (ins nImmSplatI16:$SIMM), IIC_VMOVImm, + "vmov", "i16", "$Vd, $SIMM", "", + [(set QPR:$Vd, (v8i16 (NEONvmovImm timm:$SIMM)))]> { + let Inst{9} = SIMM{9}; +} + +def VMOVv2i32 : N1ModImm<1, 0b000, {?,?,?,?}, 0, 0, 0, 1, (outs DPR:$Vd), + (ins nImmVMOVI32:$SIMM), IIC_VMOVImm, + "vmov", "i32", "$Vd, $SIMM", "", + [(set DPR:$Vd, (v2i32 (NEONvmovImm timm:$SIMM)))]> { + let Inst{11-8} = SIMM{11-8}; +} + +def VMOVv4i32 : N1ModImm<1, 0b000, {?,?,?,?}, 0, 1, 0, 1, (outs QPR:$Vd), + (ins nImmVMOVI32:$SIMM), IIC_VMOVImm, + "vmov", "i32", "$Vd, $SIMM", "", + [(set QPR:$Vd, (v4i32 (NEONvmovImm timm:$SIMM)))]> { + let Inst{11-8} = SIMM{11-8}; +} + +def VMOVv1i64 : N1ModImm<1, 0b000, 0b1110, 0, 0, 1, 1, (outs DPR:$Vd), + (ins nImmSplatI64:$SIMM), IIC_VMOVImm, + "vmov", "i64", "$Vd, $SIMM", "", + [(set DPR:$Vd, (v1i64 (NEONvmovImm timm:$SIMM)))]>; +def VMOVv2i64 : N1ModImm<1, 0b000, 0b1110, 0, 1, 1, 1, (outs QPR:$Vd), + (ins nImmSplatI64:$SIMM), IIC_VMOVImm, + "vmov", "i64", "$Vd, $SIMM", "", + [(set QPR:$Vd, (v2i64 (NEONvmovImm timm:$SIMM)))]>; + +def VMOVv2f32 : N1ModImm<1, 0b000, 0b1111, 0, 0, 0, 1, (outs DPR:$Vd), + (ins nImmVMOVF32:$SIMM), IIC_VMOVImm, + "vmov", "f32", "$Vd, $SIMM", "", + [(set DPR:$Vd, (v2f32 (NEONvmovFPImm timm:$SIMM)))]>; +def VMOVv4f32 : N1ModImm<1, 0b000, 0b1111, 0, 1, 0, 1, (outs QPR:$Vd), + (ins nImmVMOVF32:$SIMM), IIC_VMOVImm, + "vmov", "f32", "$Vd, $SIMM", "", + [(set QPR:$Vd, (v4f32 (NEONvmovFPImm timm:$SIMM)))]>; +} // isReMaterializable, isAsCheapAsAMove + +// Add support for bytes replication feature, so it could be GAS compatible. +multiclass NEONImmReplicateI8InstAlias<ValueType To> { + // E.g. instructions below: + // "vmov.i32 d0, #0xffffffff" + // "vmov.i32 d0, #0xabababab" + // "vmov.i16 d0, #0xabab" + // are incorrect, but we could deal with such cases. + // For last two instructions, for example, it should emit: + // "vmov.i8 d0, #0xab" + def : NEONInstAlias<"vmov${p}.i" # To.Size # " $Vd, $Vm", + (VMOVv8i8 DPR:$Vd, nImmVMOVIReplicate<i8, To>:$Vm, pred:$p)>; + def : NEONInstAlias<"vmov${p}.i" # To.Size # " $Vd, $Vm", + (VMOVv16i8 QPR:$Vd, nImmVMOVIReplicate<i8, To>:$Vm, pred:$p)>; + // Also add same support for VMVN instructions. So instruction: + // "vmvn.i32 d0, #0xabababab" + // actually means: + // "vmov.i8 d0, #0x54" + def : NEONInstAlias<"vmvn${p}.i" # To.Size # " $Vd, $Vm", + (VMOVv8i8 DPR:$Vd, nImmVINVIReplicate<i8, To>:$Vm, pred:$p)>; + def : NEONInstAlias<"vmvn${p}.i" # To.Size # " $Vd, $Vm", + (VMOVv16i8 QPR:$Vd, nImmVINVIReplicate<i8, To>:$Vm, pred:$p)>; +} + +defm : NEONImmReplicateI8InstAlias<i16>; +defm : NEONImmReplicateI8InstAlias<i32>; +defm : NEONImmReplicateI8InstAlias<i64>; + +// Similar to above for types other than i8, e.g.: +// "vmov.i32 d0, #0xab00ab00" -> "vmov.i16 d0, #0xab00" +// "vmvn.i64 q0, #0xab000000ab000000" -> "vmvn.i32 q0, #0xab000000" +// In this case we do not canonicalize VMVN to VMOV +multiclass NEONImmReplicateInstAlias<ValueType From, NeonI V8, NeonI V16, + NeonI NV8, NeonI NV16, ValueType To> { + def : NEONInstAlias<"vmov${p}.i" # To.Size # " $Vd, $Vm", + (V8 DPR:$Vd, nImmVMOVIReplicate<From, To>:$Vm, pred:$p)>; + def : NEONInstAlias<"vmov${p}.i" # To.Size # " $Vd, $Vm", + (V16 QPR:$Vd, nImmVMOVIReplicate<From, To>:$Vm, pred:$p)>; + def : NEONInstAlias<"vmvn${p}.i" # To.Size # " $Vd, $Vm", + (NV8 DPR:$Vd, nImmVMOVIReplicate<From, To>:$Vm, pred:$p)>; + def : NEONInstAlias<"vmvn${p}.i" # To.Size # " $Vd, $Vm", + (NV16 QPR:$Vd, nImmVMOVIReplicate<From, To>:$Vm, pred:$p)>; +} + +defm : NEONImmReplicateInstAlias<i16, VMOVv4i16, VMOVv8i16, + VMVNv4i16, VMVNv8i16, i32>; +defm : NEONImmReplicateInstAlias<i16, VMOVv4i16, VMOVv8i16, + VMVNv4i16, VMVNv8i16, i64>; +defm : NEONImmReplicateInstAlias<i32, VMOVv2i32, VMOVv4i32, + VMVNv2i32, VMVNv4i32, i64>; +// TODO: add "VMOV <-> VMVN" conversion for cases like +// "vmov.i32 d0, #0xffaaffaa" -> "vmvn.i16 d0, #0x55" +// "vmvn.i32 d0, #0xaaffaaff" -> "vmov.i16 d0, #0xff00" + +// On some CPUs the two instructions "vmov.i32 dD, #0" and "vmov.i32 qD, #0" +// require zero cycles to execute so they should be used wherever possible for +// setting a register to zero. + +// Even without these pseudo-insts we would probably end up with the correct +// instruction, but we could not mark the general ones with "isAsCheapAsAMove" +// since they are sometimes rather expensive (in general). + +let AddedComplexity = 50, isAsCheapAsAMove = 1, isReMaterializable = 1 in { + def VMOVD0 : ARMPseudoExpand<(outs DPR:$Vd), (ins), 4, IIC_VMOVImm, + [(set DPR:$Vd, (v2i32 NEONimmAllZerosV))], + (VMOVv2i32 DPR:$Vd, 0, (ops 14, zero_reg))>, + Requires<[HasZCZ]>; + def VMOVQ0 : ARMPseudoExpand<(outs QPR:$Vd), (ins), 4, IIC_VMOVImm, + [(set QPR:$Vd, (v4i32 NEONimmAllZerosV))], + (VMOVv4i32 QPR:$Vd, 0, (ops 14, zero_reg))>, + Requires<[HasZCZ]>; +} + +// VMOV : Vector Get Lane (move scalar to ARM core register) + +def VGETLNs8 : NVGetLane<{1,1,1,0,0,1,?,1}, 0b1011, {?,?}, + (outs GPR:$R), (ins DPR:$V, VectorIndex8:$lane), + IIC_VMOVSI, "vmov", "s8", "$R, $V$lane", + [(set GPR:$R, (NEONvgetlanes (v8i8 DPR:$V), + imm:$lane))]> { + let Inst{21} = lane{2}; + let Inst{6-5} = lane{1-0}; +} +def VGETLNs16 : NVGetLane<{1,1,1,0,0,0,?,1}, 0b1011, {?,1}, + (outs GPR:$R), (ins DPR:$V, VectorIndex16:$lane), + IIC_VMOVSI, "vmov", "s16", "$R, $V$lane", + [(set GPR:$R, (NEONvgetlanes (v4i16 DPR:$V), + imm:$lane))]> { + let Inst{21} = lane{1}; + let Inst{6} = lane{0}; +} +def VGETLNu8 : NVGetLane<{1,1,1,0,1,1,?,1}, 0b1011, {?,?}, + (outs GPR:$R), (ins DPR:$V, VectorIndex8:$lane), + IIC_VMOVSI, "vmov", "u8", "$R, $V$lane", + [(set GPR:$R, (NEONvgetlaneu (v8i8 DPR:$V), + imm:$lane))]> { + let Inst{21} = lane{2}; + let Inst{6-5} = lane{1-0}; +} +def VGETLNu16 : NVGetLane<{1,1,1,0,1,0,?,1}, 0b1011, {?,1}, + (outs GPR:$R), (ins DPR:$V, VectorIndex16:$lane), + IIC_VMOVSI, "vmov", "u16", "$R, $V$lane", + [(set GPR:$R, (NEONvgetlaneu (v4i16 DPR:$V), + imm:$lane))]> { + let Inst{21} = lane{1}; + let Inst{6} = lane{0}; +} +def VGETLNi32 : NVGetLane<{1,1,1,0,0,0,?,1}, 0b1011, 0b00, + (outs GPR:$R), (ins DPR:$V, VectorIndex32:$lane), + IIC_VMOVSI, "vmov", "32", "$R, $V$lane", + [(set GPR:$R, (extractelt (v2i32 DPR:$V), + imm:$lane))]>, + Requires<[HasVFP2, HasFastVGETLNi32]> { + let Inst{21} = lane{0}; +} +// def VGETLNf32: see FMRDH and FMRDL in ARMInstrVFP.td +def : Pat<(NEONvgetlanes (v16i8 QPR:$src), imm:$lane), + (VGETLNs8 (v8i8 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i8_reg imm:$lane))), + (SubReg_i8_lane imm:$lane))>; +def : Pat<(NEONvgetlanes (v8i16 QPR:$src), imm:$lane), + (VGETLNs16 (v4i16 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane))>; +def : Pat<(NEONvgetlaneu (v16i8 QPR:$src), imm:$lane), + (VGETLNu8 (v8i8 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i8_reg imm:$lane))), + (SubReg_i8_lane imm:$lane))>; +def : Pat<(NEONvgetlaneu (v8i16 QPR:$src), imm:$lane), + (VGETLNu16 (v4i16 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane))>; +def : Pat<(extractelt (v4i32 QPR:$src), imm:$lane), + (VGETLNi32 (v2i32 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane))>, + Requires<[HasNEON, HasFastVGETLNi32]>; +def : Pat<(extractelt (v2i32 DPR:$src), imm:$lane), + (COPY_TO_REGCLASS + (i32 (EXTRACT_SUBREG DPR:$src, (SSubReg_f32_reg imm:$lane))), GPR)>, + Requires<[HasNEON, HasSlowVGETLNi32]>; +def : Pat<(extractelt (v4i32 QPR:$src), imm:$lane), + (COPY_TO_REGCLASS + (i32 (EXTRACT_SUBREG QPR:$src, (SSubReg_f32_reg imm:$lane))), GPR)>, + Requires<[HasNEON, HasSlowVGETLNi32]>; +def : Pat<(extractelt (v2f32 DPR:$src1), imm:$src2), + (EXTRACT_SUBREG (v2f32 (COPY_TO_REGCLASS (v2f32 DPR:$src1),DPR_VFP2)), + (SSubReg_f32_reg imm:$src2))>; +def : Pat<(extractelt (v4f32 QPR:$src1), imm:$src2), + (EXTRACT_SUBREG (v4f32 (COPY_TO_REGCLASS (v4f32 QPR:$src1),QPR_VFP2)), + (SSubReg_f32_reg imm:$src2))>; +//def : Pat<(extractelt (v2i64 QPR:$src1), imm:$src2), +// (EXTRACT_SUBREG QPR:$src1, (DSubReg_f64_reg imm:$src2))>; +def : Pat<(extractelt (v2f64 QPR:$src1), imm:$src2), + (EXTRACT_SUBREG QPR:$src1, (DSubReg_f64_reg imm:$src2))>; + + +// VMOV : Vector Set Lane (move ARM core register to scalar) + +let Constraints = "$src1 = $V" in { +def VSETLNi8 : NVSetLane<{1,1,1,0,0,1,?,0}, 0b1011, {?,?}, (outs DPR:$V), + (ins DPR:$src1, GPR:$R, VectorIndex8:$lane), + IIC_VMOVISL, "vmov", "8", "$V$lane, $R", + [(set DPR:$V, (vector_insert (v8i8 DPR:$src1), + GPR:$R, imm:$lane))]> { + let Inst{21} = lane{2}; + let Inst{6-5} = lane{1-0}; +} +def VSETLNi16 : NVSetLane<{1,1,1,0,0,0,?,0}, 0b1011, {?,1}, (outs DPR:$V), + (ins DPR:$src1, GPR:$R, VectorIndex16:$lane), + IIC_VMOVISL, "vmov", "16", "$V$lane, $R", + [(set DPR:$V, (vector_insert (v4i16 DPR:$src1), + GPR:$R, imm:$lane))]> { + let Inst{21} = lane{1}; + let Inst{6} = lane{0}; +} +def VSETLNi32 : NVSetLane<{1,1,1,0,0,0,?,0}, 0b1011, 0b00, (outs DPR:$V), + (ins DPR:$src1, GPR:$R, VectorIndex32:$lane), + IIC_VMOVISL, "vmov", "32", "$V$lane, $R", + [(set DPR:$V, (insertelt (v2i32 DPR:$src1), + GPR:$R, imm:$lane))]>, + Requires<[HasVFP2]> { + let Inst{21} = lane{0}; + // This instruction is equivalent as + // $V = INSERT_SUBREG $src1, $R, translateImmToSubIdx($imm) + let isInsertSubreg = 1; +} +} +def : Pat<(vector_insert (v16i8 QPR:$src1), GPR:$src2, imm:$lane), + (v16i8 (INSERT_SUBREG QPR:$src1, + (v8i8 (VSETLNi8 (v8i8 (EXTRACT_SUBREG QPR:$src1, + (DSubReg_i8_reg imm:$lane))), + GPR:$src2, (SubReg_i8_lane imm:$lane))), + (DSubReg_i8_reg imm:$lane)))>; +def : Pat<(vector_insert (v8i16 QPR:$src1), GPR:$src2, imm:$lane), + (v8i16 (INSERT_SUBREG QPR:$src1, + (v4i16 (VSETLNi16 (v4i16 (EXTRACT_SUBREG QPR:$src1, + (DSubReg_i16_reg imm:$lane))), + GPR:$src2, (SubReg_i16_lane imm:$lane))), + (DSubReg_i16_reg imm:$lane)))>; +def : Pat<(insertelt (v4i32 QPR:$src1), GPR:$src2, imm:$lane), + (v4i32 (INSERT_SUBREG QPR:$src1, + (v2i32 (VSETLNi32 (v2i32 (EXTRACT_SUBREG QPR:$src1, + (DSubReg_i32_reg imm:$lane))), + GPR:$src2, (SubReg_i32_lane imm:$lane))), + (DSubReg_i32_reg imm:$lane)))>; + +def : Pat<(v2f32 (insertelt DPR:$src1, SPR:$src2, imm:$src3)), + (INSERT_SUBREG (v2f32 (COPY_TO_REGCLASS DPR:$src1, DPR_VFP2)), + SPR:$src2, (SSubReg_f32_reg imm:$src3))>; +def : Pat<(v4f32 (insertelt QPR:$src1, SPR:$src2, imm:$src3)), + (INSERT_SUBREG (v4f32 (COPY_TO_REGCLASS QPR:$src1, QPR_VFP2)), + SPR:$src2, (SSubReg_f32_reg imm:$src3))>; + +//def : Pat<(v2i64 (insertelt QPR:$src1, DPR:$src2, imm:$src3)), +// (INSERT_SUBREG QPR:$src1, DPR:$src2, (DSubReg_f64_reg imm:$src3))>; +def : Pat<(v2f64 (insertelt QPR:$src1, DPR:$src2, imm:$src3)), + (INSERT_SUBREG QPR:$src1, DPR:$src2, (DSubReg_f64_reg imm:$src3))>; + +def : Pat<(v2f32 (scalar_to_vector SPR:$src)), + (INSERT_SUBREG (v2f32 (IMPLICIT_DEF)), SPR:$src, ssub_0)>; +def : Pat<(v2f64 (scalar_to_vector (f64 DPR:$src))), + (INSERT_SUBREG (v2f64 (IMPLICIT_DEF)), DPR:$src, dsub_0)>; +def : Pat<(v4f32 (scalar_to_vector SPR:$src)), + (INSERT_SUBREG (v4f32 (IMPLICIT_DEF)), SPR:$src, ssub_0)>; + +def : Pat<(v8i8 (scalar_to_vector GPR:$src)), + (VSETLNi8 (v8i8 (IMPLICIT_DEF)), GPR:$src, (i32 0))>; +def : Pat<(v4i16 (scalar_to_vector GPR:$src)), + (VSETLNi16 (v4i16 (IMPLICIT_DEF)), GPR:$src, (i32 0))>; +def : Pat<(v2i32 (scalar_to_vector GPR:$src)), + (VSETLNi32 (v2i32 (IMPLICIT_DEF)), GPR:$src, (i32 0))>; + +def : Pat<(v16i8 (scalar_to_vector GPR:$src)), + (INSERT_SUBREG (v16i8 (IMPLICIT_DEF)), + (VSETLNi8 (v8i8 (IMPLICIT_DEF)), GPR:$src, (i32 0)), + dsub_0)>; +def : Pat<(v8i16 (scalar_to_vector GPR:$src)), + (INSERT_SUBREG (v8i16 (IMPLICIT_DEF)), + (VSETLNi16 (v4i16 (IMPLICIT_DEF)), GPR:$src, (i32 0)), + dsub_0)>; +def : Pat<(v4i32 (scalar_to_vector GPR:$src)), + (INSERT_SUBREG (v4i32 (IMPLICIT_DEF)), + (VSETLNi32 (v2i32 (IMPLICIT_DEF)), GPR:$src, (i32 0)), + dsub_0)>; + +// VDUP : Vector Duplicate (from ARM core register to all elements) + +class VDUPD<bits<8> opcod1, bits<2> opcod3, string Dt, ValueType Ty> + : NVDup<opcod1, 0b1011, opcod3, (outs DPR:$V), (ins GPR:$R), + IIC_VMOVIS, "vdup", Dt, "$V, $R", + [(set DPR:$V, (Ty (NEONvdup (i32 GPR:$R))))]>; +class VDUPQ<bits<8> opcod1, bits<2> opcod3, string Dt, ValueType Ty> + : NVDup<opcod1, 0b1011, opcod3, (outs QPR:$V), (ins GPR:$R), + IIC_VMOVIS, "vdup", Dt, "$V, $R", + [(set QPR:$V, (Ty (NEONvdup (i32 GPR:$R))))]>; + +def VDUP8d : VDUPD<0b11101100, 0b00, "8", v8i8>; +def VDUP16d : VDUPD<0b11101000, 0b01, "16", v4i16>; +def VDUP32d : VDUPD<0b11101000, 0b00, "32", v2i32>, + Requires<[HasNEON, HasFastVDUP32]>; +def VDUP8q : VDUPQ<0b11101110, 0b00, "8", v16i8>; +def VDUP16q : VDUPQ<0b11101010, 0b01, "16", v8i16>; +def VDUP32q : VDUPQ<0b11101010, 0b00, "32", v4i32>; + +// NEONvdup patterns for uarchs with fast VDUP.32. +def : Pat<(v2f32 (NEONvdup (f32 (bitconvert GPR:$R)))), (VDUP32d GPR:$R)>, + Requires<[HasNEON,HasFastVDUP32]>; +def : Pat<(v4f32 (NEONvdup (f32 (bitconvert GPR:$R)))), (VDUP32q GPR:$R)>; + +// NEONvdup patterns for uarchs with slow VDUP.32 - use VMOVDRR instead. +def : Pat<(v2i32 (NEONvdup (i32 GPR:$R))), (VMOVDRR GPR:$R, GPR:$R)>, + Requires<[HasNEON,HasSlowVDUP32]>; +def : Pat<(v2f32 (NEONvdup (f32 (bitconvert GPR:$R)))), (VMOVDRR GPR:$R, GPR:$R)>, + Requires<[HasNEON,HasSlowVDUP32]>; + +// VDUP : Vector Duplicate Lane (from scalar to all elements) + +class VDUPLND<bits<4> op19_16, string OpcodeStr, string Dt, + ValueType Ty, Operand IdxTy> + : NVDupLane<op19_16, 0, (outs DPR:$Vd), (ins DPR:$Vm, IdxTy:$lane), + IIC_VMOVD, OpcodeStr, Dt, "$Vd, $Vm$lane", + [(set DPR:$Vd, (Ty (NEONvduplane (Ty DPR:$Vm), imm:$lane)))]>; + +class VDUPLNQ<bits<4> op19_16, string OpcodeStr, string Dt, + ValueType ResTy, ValueType OpTy, Operand IdxTy> + : NVDupLane<op19_16, 1, (outs QPR:$Vd), (ins DPR:$Vm, IdxTy:$lane), + IIC_VMOVQ, OpcodeStr, Dt, "$Vd, $Vm$lane", + [(set QPR:$Vd, (ResTy (NEONvduplane (OpTy DPR:$Vm), + VectorIndex32:$lane)))]>; + +// Inst{19-16} is partially specified depending on the element size. + +def VDUPLN8d : VDUPLND<{?,?,?,1}, "vdup", "8", v8i8, VectorIndex8> { + bits<3> lane; + let Inst{19-17} = lane{2-0}; +} +def VDUPLN16d : VDUPLND<{?,?,1,0}, "vdup", "16", v4i16, VectorIndex16> { + bits<2> lane; + let Inst{19-18} = lane{1-0}; +} +def VDUPLN32d : VDUPLND<{?,1,0,0}, "vdup", "32", v2i32, VectorIndex32> { + bits<1> lane; + let Inst{19} = lane{0}; +} +def VDUPLN8q : VDUPLNQ<{?,?,?,1}, "vdup", "8", v16i8, v8i8, VectorIndex8> { + bits<3> lane; + let Inst{19-17} = lane{2-0}; +} +def VDUPLN16q : VDUPLNQ<{?,?,1,0}, "vdup", "16", v8i16, v4i16, VectorIndex16> { + bits<2> lane; + let Inst{19-18} = lane{1-0}; +} +def VDUPLN32q : VDUPLNQ<{?,1,0,0}, "vdup", "32", v4i32, v2i32, VectorIndex32> { + bits<1> lane; + let Inst{19} = lane{0}; +} + +def : Pat<(v2f32 (NEONvduplane (v2f32 DPR:$Vm), imm:$lane)), + (VDUPLN32d DPR:$Vm, imm:$lane)>; + +def : Pat<(v4f32 (NEONvduplane (v2f32 DPR:$Vm), imm:$lane)), + (VDUPLN32q DPR:$Vm, imm:$lane)>; + +def : Pat<(v16i8 (NEONvduplane (v16i8 QPR:$src), imm:$lane)), + (v16i8 (VDUPLN8q (v8i8 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i8_reg imm:$lane))), + (SubReg_i8_lane imm:$lane)))>; +def : Pat<(v8i16 (NEONvduplane (v8i16 QPR:$src), imm:$lane)), + (v8i16 (VDUPLN16q (v4i16 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i16_reg imm:$lane))), + (SubReg_i16_lane imm:$lane)))>; +def : Pat<(v4i32 (NEONvduplane (v4i32 QPR:$src), imm:$lane)), + (v4i32 (VDUPLN32q (v2i32 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; +def : Pat<(v4f32 (NEONvduplane (v4f32 QPR:$src), imm:$lane)), + (v4f32 (VDUPLN32q (v2f32 (EXTRACT_SUBREG QPR:$src, + (DSubReg_i32_reg imm:$lane))), + (SubReg_i32_lane imm:$lane)))>; + +def : Pat<(v2f32 (NEONvdup (f32 SPR:$src))), + (v2f32 (VDUPLN32d (INSERT_SUBREG (v2f32 (IMPLICIT_DEF)), + SPR:$src, ssub_0), (i32 0)))>; +def : Pat<(v4f32 (NEONvdup (f32 SPR:$src))), + (v4f32 (VDUPLN32q (INSERT_SUBREG (v2f32 (IMPLICIT_DEF)), + SPR:$src, ssub_0), (i32 0)))>; + +// VMOVN : Vector Narrowing Move +defm VMOVN : N2VN_HSD<0b11,0b11,0b10,0b00100,0,0, IIC_VMOVN, + "vmovn", "i", trunc>; +// VQMOVN : Vector Saturating Narrowing Move +defm VQMOVNs : N2VNInt_HSD<0b11,0b11,0b10,0b00101,0,0, IIC_VQUNAiD, + "vqmovn", "s", int_arm_neon_vqmovns>; +defm VQMOVNu : N2VNInt_HSD<0b11,0b11,0b10,0b00101,1,0, IIC_VQUNAiD, + "vqmovn", "u", int_arm_neon_vqmovnu>; +defm VQMOVNsu : N2VNInt_HSD<0b11,0b11,0b10,0b00100,1,0, IIC_VQUNAiD, + "vqmovun", "s", int_arm_neon_vqmovnsu>; +// VMOVL : Vector Lengthening Move +defm VMOVLs : N2VL_QHS<0b01,0b10100,0,1, "vmovl", "s", sext>; +defm VMOVLu : N2VL_QHS<0b11,0b10100,0,1, "vmovl", "u", zext>; +def : Pat<(v8i16 (anyext (v8i8 DPR:$Vm))), (VMOVLuv8i16 DPR:$Vm)>; +def : Pat<(v4i32 (anyext (v4i16 DPR:$Vm))), (VMOVLuv4i32 DPR:$Vm)>; +def : Pat<(v2i64 (anyext (v2i32 DPR:$Vm))), (VMOVLuv2i64 DPR:$Vm)>; + +// Vector Conversions. + +// VCVT : Vector Convert Between Floating-Point and Integers +def VCVTf2sd : N2VD<0b11, 0b11, 0b10, 0b11, 0b01110, 0, "vcvt", "s32.f32", + v2i32, v2f32, fp_to_sint>; +def VCVTf2ud : N2VD<0b11, 0b11, 0b10, 0b11, 0b01111, 0, "vcvt", "u32.f32", + v2i32, v2f32, fp_to_uint>; +def VCVTs2fd : N2VD<0b11, 0b11, 0b10, 0b11, 0b01100, 0, "vcvt", "f32.s32", + v2f32, v2i32, sint_to_fp>; +def VCVTu2fd : N2VD<0b11, 0b11, 0b10, 0b11, 0b01101, 0, "vcvt", "f32.u32", + v2f32, v2i32, uint_to_fp>; + +def VCVTf2sq : N2VQ<0b11, 0b11, 0b10, 0b11, 0b01110, 0, "vcvt", "s32.f32", + v4i32, v4f32, fp_to_sint>; +def VCVTf2uq : N2VQ<0b11, 0b11, 0b10, 0b11, 0b01111, 0, "vcvt", "u32.f32", + v4i32, v4f32, fp_to_uint>; +def VCVTs2fq : N2VQ<0b11, 0b11, 0b10, 0b11, 0b01100, 0, "vcvt", "f32.s32", + v4f32, v4i32, sint_to_fp>; +def VCVTu2fq : N2VQ<0b11, 0b11, 0b10, 0b11, 0b01101, 0, "vcvt", "f32.u32", + v4f32, v4i32, uint_to_fp>; + +def VCVTh2sd : N2VD<0b11, 0b11, 0b01, 0b11, 0b01110, 0, "vcvt", "s16.f16", + v4i16, v4f16, fp_to_sint>, + Requires<[HasNEON, HasFullFP16]>; +def VCVTh2ud : N2VD<0b11, 0b11, 0b01, 0b11, 0b01111, 0, "vcvt", "u16.f16", + v4i16, v4f16, fp_to_uint>, + Requires<[HasNEON, HasFullFP16]>; +def VCVTs2hd : N2VD<0b11, 0b11, 0b01, 0b11, 0b01100, 0, "vcvt", "f16.s16", + v4f16, v4i16, sint_to_fp>, + Requires<[HasNEON, HasFullFP16]>; +def VCVTu2hd : N2VD<0b11, 0b11, 0b01, 0b11, 0b01101, 0, "vcvt", "f16.u16", + v4f16, v4i16, uint_to_fp>, + Requires<[HasNEON, HasFullFP16]>; + +def VCVTh2sq : N2VQ<0b11, 0b11, 0b01, 0b11, 0b01110, 0, "vcvt", "s16.f16", + v8i16, v8f16, fp_to_sint>, + Requires<[HasNEON, HasFullFP16]>; +def VCVTh2uq : N2VQ<0b11, 0b11, 0b01, 0b11, 0b01111, 0, "vcvt", "u16.f16", + v8i16, v8f16, fp_to_uint>, + Requires<[HasNEON, HasFullFP16]>; +def VCVTs2hq : N2VQ<0b11, 0b11, 0b01, 0b11, 0b01100, 0, "vcvt", "f16.s16", + v8f16, v8i16, sint_to_fp>, + Requires<[HasNEON, HasFullFP16]>; +def VCVTu2hq : N2VQ<0b11, 0b11, 0b01, 0b11, 0b01101, 0, "vcvt", "f16.u16", + v8f16, v8i16, uint_to_fp>, + Requires<[HasNEON, HasFullFP16]>; + +// VCVT{A, N, P, M} +multiclass VCVT_FPI<string op, bits<3> op10_8, SDPatternOperator IntS, + SDPatternOperator IntU> { + let PostEncoderMethod = "NEONThumb2V8PostEncoder", DecoderNamespace = "v8NEON" in { + def SDf : N2VDIntnp<0b10, 0b11, op10_8, 0, NoItinerary, !strconcat("vcvt", op), + "s32.f32", v2i32, v2f32, IntS>, Requires<[HasV8, HasNEON]>; + def SQf : N2VQIntnp<0b10, 0b11, op10_8, 0, NoItinerary, !strconcat("vcvt", op), + "s32.f32", v4i32, v4f32, IntS>, Requires<[HasV8, HasNEON]>; + def UDf : N2VDIntnp<0b10, 0b11, op10_8, 1, NoItinerary, !strconcat("vcvt", op), + "u32.f32", v2i32, v2f32, IntU>, Requires<[HasV8, HasNEON]>; + def UQf : N2VQIntnp<0b10, 0b11, op10_8, 1, NoItinerary, !strconcat("vcvt", op), + "u32.f32", v4i32, v4f32, IntU>, Requires<[HasV8, HasNEON]>; + def SDh : N2VDIntnp<0b01, 0b11, op10_8, 0, NoItinerary, !strconcat("vcvt", op), + "s16.f16", v4i16, v4f16, IntS>, + Requires<[HasV8, HasNEON, HasFullFP16]>; + def SQh : N2VQIntnp<0b01, 0b11, op10_8, 0, NoItinerary, !strconcat("vcvt", op), + "s16.f16", v8i16, v8f16, IntS>, + Requires<[HasV8, HasNEON, HasFullFP16]>; + def UDh : N2VDIntnp<0b01, 0b11, op10_8, 1, NoItinerary, !strconcat("vcvt", op), + "u16.f16", v4i16, v4f16, IntU>, + Requires<[HasV8, HasNEON, HasFullFP16]>; + def UQh : N2VQIntnp<0b01, 0b11, op10_8, 1, NoItinerary, !strconcat("vcvt", op), + "u16.f16", v8i16, v8f16, IntU>, + Requires<[HasV8, HasNEON, HasFullFP16]>; + } +} + +defm VCVTAN : VCVT_FPI<"a", 0b000, int_arm_neon_vcvtas, int_arm_neon_vcvtau>; +defm VCVTNN : VCVT_FPI<"n", 0b001, int_arm_neon_vcvtns, int_arm_neon_vcvtnu>; +defm VCVTPN : VCVT_FPI<"p", 0b010, int_arm_neon_vcvtps, int_arm_neon_vcvtpu>; +defm VCVTMN : VCVT_FPI<"m", 0b011, int_arm_neon_vcvtms, int_arm_neon_vcvtmu>; + +// VCVT : Vector Convert Between Floating-Point and Fixed-Point. +let DecoderMethod = "DecodeVCVTD" in { +def VCVTf2xsd : N2VCvtD<0, 1, 0b1111, 0, 1, "vcvt", "s32.f32", + v2i32, v2f32, int_arm_neon_vcvtfp2fxs>; +def VCVTf2xud : N2VCvtD<1, 1, 0b1111, 0, 1, "vcvt", "u32.f32", + v2i32, v2f32, int_arm_neon_vcvtfp2fxu>; +def VCVTxs2fd : N2VCvtD<0, 1, 0b1110, 0, 1, "vcvt", "f32.s32", + v2f32, v2i32, int_arm_neon_vcvtfxs2fp>; +def VCVTxu2fd : N2VCvtD<1, 1, 0b1110, 0, 1, "vcvt", "f32.u32", + v2f32, v2i32, int_arm_neon_vcvtfxu2fp>; +let Predicates = [HasNEON, HasFullFP16] in { +def VCVTh2xsd : N2VCvtD<0, 1, 0b1101, 0, 1, "vcvt", "s16.f16", + v4i16, v4f16, int_arm_neon_vcvtfp2fxs>; +def VCVTh2xud : N2VCvtD<1, 1, 0b1101, 0, 1, "vcvt", "u16.f16", + v4i16, v4f16, int_arm_neon_vcvtfp2fxu>; +def VCVTxs2hd : N2VCvtD<0, 1, 0b1100, 0, 1, "vcvt", "f16.s16", + v4f16, v4i16, int_arm_neon_vcvtfxs2fp>; +def VCVTxu2hd : N2VCvtD<1, 1, 0b1100, 0, 1, "vcvt", "f16.u16", + v4f16, v4i16, int_arm_neon_vcvtfxu2fp>; +} // Predicates = [HasNEON, HasFullFP16] +} + +let DecoderMethod = "DecodeVCVTQ" in { +def VCVTf2xsq : N2VCvtQ<0, 1, 0b1111, 0, 1, "vcvt", "s32.f32", + v4i32, v4f32, int_arm_neon_vcvtfp2fxs>; +def VCVTf2xuq : N2VCvtQ<1, 1, 0b1111, 0, 1, "vcvt", "u32.f32", + v4i32, v4f32, int_arm_neon_vcvtfp2fxu>; +def VCVTxs2fq : N2VCvtQ<0, 1, 0b1110, 0, 1, "vcvt", "f32.s32", + v4f32, v4i32, int_arm_neon_vcvtfxs2fp>; +def VCVTxu2fq : N2VCvtQ<1, 1, 0b1110, 0, 1, "vcvt", "f32.u32", + v4f32, v4i32, int_arm_neon_vcvtfxu2fp>; +let Predicates = [HasNEON, HasFullFP16] in { +def VCVTh2xsq : N2VCvtQ<0, 1, 0b1101, 0, 1, "vcvt", "s16.f16", + v8i16, v8f16, int_arm_neon_vcvtfp2fxs>; +def VCVTh2xuq : N2VCvtQ<1, 1, 0b1101, 0, 1, "vcvt", "u16.f16", + v8i16, v8f16, int_arm_neon_vcvtfp2fxu>; +def VCVTxs2hq : N2VCvtQ<0, 1, 0b1100, 0, 1, "vcvt", "f16.s16", + v8f16, v8i16, int_arm_neon_vcvtfxs2fp>; +def VCVTxu2hq : N2VCvtQ<1, 1, 0b1100, 0, 1, "vcvt", "f16.u16", + v8f16, v8i16, int_arm_neon_vcvtfxu2fp>; +} // Predicates = [HasNEON, HasFullFP16] +} + +def : NEONInstAlias<"vcvt${p}.s32.f32 $Dd, $Dm, #0", + (VCVTf2sd DPR:$Dd, DPR:$Dm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.u32.f32 $Dd, $Dm, #0", + (VCVTf2ud DPR:$Dd, DPR:$Dm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.f32.s32 $Dd, $Dm, #0", + (VCVTs2fd DPR:$Dd, DPR:$Dm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.f32.u32 $Dd, $Dm, #0", + (VCVTu2fd DPR:$Dd, DPR:$Dm, pred:$p)>; + +def : NEONInstAlias<"vcvt${p}.s32.f32 $Qd, $Qm, #0", + (VCVTf2sq QPR:$Qd, QPR:$Qm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.u32.f32 $Qd, $Qm, #0", + (VCVTf2uq QPR:$Qd, QPR:$Qm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.f32.s32 $Qd, $Qm, #0", + (VCVTs2fq QPR:$Qd, QPR:$Qm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.f32.u32 $Qd, $Qm, #0", + (VCVTu2fq QPR:$Qd, QPR:$Qm, pred:$p)>; + +def : NEONInstAlias<"vcvt${p}.s16.f16 $Dd, $Dm, #0", + (VCVTh2sd DPR:$Dd, DPR:$Dm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.u16.f16 $Dd, $Dm, #0", + (VCVTh2ud DPR:$Dd, DPR:$Dm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.f16.s16 $Dd, $Dm, #0", + (VCVTs2hd DPR:$Dd, DPR:$Dm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.f16.u16 $Dd, $Dm, #0", + (VCVTu2hd DPR:$Dd, DPR:$Dm, pred:$p)>; + +def : NEONInstAlias<"vcvt${p}.s16.f16 $Qd, $Qm, #0", + (VCVTh2sq QPR:$Qd, QPR:$Qm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.u16.f16 $Qd, $Qm, #0", + (VCVTh2uq QPR:$Qd, QPR:$Qm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.f16.s16 $Qd, $Qm, #0", + (VCVTs2hq QPR:$Qd, QPR:$Qm, pred:$p)>; +def : NEONInstAlias<"vcvt${p}.f16.u16 $Qd, $Qm, #0", + (VCVTu2hq QPR:$Qd, QPR:$Qm, pred:$p)>; + + +// VCVT : Vector Convert Between Half-Precision and Single-Precision. +def VCVTf2h : N2VNInt<0b11, 0b11, 0b01, 0b10, 0b01100, 0, 0, + IIC_VUNAQ, "vcvt", "f16.f32", + v4i16, v4f32, int_arm_neon_vcvtfp2hf>, + Requires<[HasNEON, HasFP16]>; +def VCVTh2f : N2VLInt<0b11, 0b11, 0b01, 0b10, 0b01110, 0, 0, + IIC_VUNAQ, "vcvt", "f32.f16", + v4f32, v4i16, int_arm_neon_vcvthf2fp>, + Requires<[HasNEON, HasFP16]>; + +// Vector Reverse. + +// VREV64 : Vector Reverse elements within 64-bit doublewords + +class VREV64D<bits<2> op19_18, string OpcodeStr, string Dt, ValueType Ty> + : N2V<0b11, 0b11, op19_18, 0b00, 0b00000, 0, 0, (outs DPR:$Vd), + (ins DPR:$Vm), IIC_VMOVD, + OpcodeStr, Dt, "$Vd, $Vm", "", + [(set DPR:$Vd, (Ty (NEONvrev64 (Ty DPR:$Vm))))]>; +class VREV64Q<bits<2> op19_18, string OpcodeStr, string Dt, ValueType Ty> + : N2V<0b11, 0b11, op19_18, 0b00, 0b00000, 1, 0, (outs QPR:$Vd), + (ins QPR:$Vm), IIC_VMOVQ, + OpcodeStr, Dt, "$Vd, $Vm", "", + [(set QPR:$Vd, (Ty (NEONvrev64 (Ty QPR:$Vm))))]>; + +def VREV64d8 : VREV64D<0b00, "vrev64", "8", v8i8>; +def VREV64d16 : VREV64D<0b01, "vrev64", "16", v4i16>; +def VREV64d32 : VREV64D<0b10, "vrev64", "32", v2i32>; +def : Pat<(v2f32 (NEONvrev64 (v2f32 DPR:$Vm))), (VREV64d32 DPR:$Vm)>; + +def VREV64q8 : VREV64Q<0b00, "vrev64", "8", v16i8>; +def VREV64q16 : VREV64Q<0b01, "vrev64", "16", v8i16>; +def VREV64q32 : VREV64Q<0b10, "vrev64", "32", v4i32>; +def : Pat<(v4f32 (NEONvrev64 (v4f32 QPR:$Vm))), (VREV64q32 QPR:$Vm)>; + +// VREV32 : Vector Reverse elements within 32-bit words + +class VREV32D<bits<2> op19_18, string OpcodeStr, string Dt, ValueType Ty> + : N2V<0b11, 0b11, op19_18, 0b00, 0b00001, 0, 0, (outs DPR:$Vd), + (ins DPR:$Vm), IIC_VMOVD, + OpcodeStr, Dt, "$Vd, $Vm", "", + [(set DPR:$Vd, (Ty (NEONvrev32 (Ty DPR:$Vm))))]>; +class VREV32Q<bits<2> op19_18, string OpcodeStr, string Dt, ValueType Ty> + : N2V<0b11, 0b11, op19_18, 0b00, 0b00001, 1, 0, (outs QPR:$Vd), + (ins QPR:$Vm), IIC_VMOVQ, + OpcodeStr, Dt, "$Vd, $Vm", "", + [(set QPR:$Vd, (Ty (NEONvrev32 (Ty QPR:$Vm))))]>; + +def VREV32d8 : VREV32D<0b00, "vrev32", "8", v8i8>; +def VREV32d16 : VREV32D<0b01, "vrev32", "16", v4i16>; + +def VREV32q8 : VREV32Q<0b00, "vrev32", "8", v16i8>; +def VREV32q16 : VREV32Q<0b01, "vrev32", "16", v8i16>; + +// VREV16 : Vector Reverse elements within 16-bit halfwords + +class VREV16D<bits<2> op19_18, string OpcodeStr, string Dt, ValueType Ty> + : N2V<0b11, 0b11, op19_18, 0b00, 0b00010, 0, 0, (outs DPR:$Vd), + (ins DPR:$Vm), IIC_VMOVD, + OpcodeStr, Dt, "$Vd, $Vm", "", + [(set DPR:$Vd, (Ty (NEONvrev16 (Ty DPR:$Vm))))]>; +class VREV16Q<bits<2> op19_18, string OpcodeStr, string Dt, ValueType Ty> + : N2V<0b11, 0b11, op19_18, 0b00, 0b00010, 1, 0, (outs QPR:$Vd), + (ins QPR:$Vm), IIC_VMOVQ, + OpcodeStr, Dt, "$Vd, $Vm", "", + [(set QPR:$Vd, (Ty (NEONvrev16 (Ty QPR:$Vm))))]>; + +def VREV16d8 : VREV16D<0b00, "vrev16", "8", v8i8>; +def VREV16q8 : VREV16Q<0b00, "vrev16", "8", v16i8>; + +// Other Vector Shuffles. + +// Aligned extractions: really just dropping registers + +class AlignedVEXTq<ValueType DestTy, ValueType SrcTy, SDNodeXForm LaneCVT> + : Pat<(DestTy (vector_extract_subvec (SrcTy QPR:$src), (i32 imm:$start))), + (EXTRACT_SUBREG (SrcTy QPR:$src), (LaneCVT imm:$start))>; + +def : AlignedVEXTq<v8i8, v16i8, DSubReg_i8_reg>; + +def : AlignedVEXTq<v4i16, v8i16, DSubReg_i16_reg>; + +def : AlignedVEXTq<v2i32, v4i32, DSubReg_i32_reg>; + +def : AlignedVEXTq<v1i64, v2i64, DSubReg_f64_reg>; + +def : AlignedVEXTq<v2f32, v4f32, DSubReg_i32_reg>; + + +// VEXT : Vector Extract + + +// All of these have a two-operand InstAlias. +let TwoOperandAliasConstraint = "$Vn = $Vd" in { +class VEXTd<string OpcodeStr, string Dt, ValueType Ty, Operand immTy> + : N3V<0,1,0b11,{?,?,?,?},0,0, (outs DPR:$Vd), + (ins DPR:$Vn, DPR:$Vm, immTy:$index), NVExtFrm, + IIC_VEXTD, OpcodeStr, Dt, "$Vd, $Vn, $Vm, $index", "", + [(set DPR:$Vd, (Ty (NEONvext (Ty DPR:$Vn), + (Ty DPR:$Vm), imm:$index)))]> { + bits<3> index; + let Inst{11} = 0b0; + let Inst{10-8} = index{2-0}; +} + +class VEXTq<string OpcodeStr, string Dt, ValueType Ty, Operand immTy> + : N3V<0,1,0b11,{?,?,?,?},1,0, (outs QPR:$Vd), + (ins QPR:$Vn, QPR:$Vm, imm0_15:$index), NVExtFrm, + IIC_VEXTQ, OpcodeStr, Dt, "$Vd, $Vn, $Vm, $index", "", + [(set QPR:$Vd, (Ty (NEONvext (Ty QPR:$Vn), + (Ty QPR:$Vm), imm:$index)))]> { + bits<4> index; + let Inst{11-8} = index{3-0}; +} +} + +def VEXTd8 : VEXTd<"vext", "8", v8i8, imm0_7> { + let Inst{10-8} = index{2-0}; +} +def VEXTd16 : VEXTd<"vext", "16", v4i16, imm0_3> { + let Inst{10-9} = index{1-0}; + let Inst{8} = 0b0; +} +def VEXTd32 : VEXTd<"vext", "32", v2i32, imm0_1> { + let Inst{10} = index{0}; + let Inst{9-8} = 0b00; +} +def : Pat<(v2f32 (NEONvext (v2f32 DPR:$Vn), + (v2f32 DPR:$Vm), + (i32 imm:$index))), + (VEXTd32 DPR:$Vn, DPR:$Vm, imm:$index)>; + +def VEXTq8 : VEXTq<"vext", "8", v16i8, imm0_15> { + let Inst{11-8} = index{3-0}; +} +def VEXTq16 : VEXTq<"vext", "16", v8i16, imm0_7> { + let Inst{11-9} = index{2-0}; + let Inst{8} = 0b0; +} +def VEXTq32 : VEXTq<"vext", "32", v4i32, imm0_3> { + let Inst{11-10} = index{1-0}; + let Inst{9-8} = 0b00; +} +def VEXTq64 : VEXTq<"vext", "64", v2i64, imm0_1> { + let Inst{11} = index{0}; + let Inst{10-8} = 0b000; +} +def : Pat<(v4f32 (NEONvext (v4f32 QPR:$Vn), + (v4f32 QPR:$Vm), + (i32 imm:$index))), + (VEXTq32 QPR:$Vn, QPR:$Vm, imm:$index)>; + +// VTRN : Vector Transpose + +def VTRNd8 : N2VDShuffle<0b00, 0b00001, "vtrn", "8">; +def VTRNd16 : N2VDShuffle<0b01, 0b00001, "vtrn", "16">; +def VTRNd32 : N2VDShuffle<0b10, 0b00001, "vtrn", "32">; + +def VTRNq8 : N2VQShuffle<0b00, 0b00001, IIC_VPERMQ, "vtrn", "8">; +def VTRNq16 : N2VQShuffle<0b01, 0b00001, IIC_VPERMQ, "vtrn", "16">; +def VTRNq32 : N2VQShuffle<0b10, 0b00001, IIC_VPERMQ, "vtrn", "32">; + +// VUZP : Vector Unzip (Deinterleave) + +def VUZPd8 : N2VDShuffle<0b00, 0b00010, "vuzp", "8">; +def VUZPd16 : N2VDShuffle<0b01, 0b00010, "vuzp", "16">; +// vuzp.32 Dd, Dm is a pseudo-instruction expanded to vtrn.32 Dd, Dm. +def : NEONInstAlias<"vuzp${p}.32 $Dd, $Dm", + (VTRNd32 DPR:$Dd, DPR:$Dm, pred:$p)>; + +def VUZPq8 : N2VQShuffle<0b00, 0b00010, IIC_VPERMQ3, "vuzp", "8">; +def VUZPq16 : N2VQShuffle<0b01, 0b00010, IIC_VPERMQ3, "vuzp", "16">; +def VUZPq32 : N2VQShuffle<0b10, 0b00010, IIC_VPERMQ3, "vuzp", "32">; + +// VZIP : Vector Zip (Interleave) + +def VZIPd8 : N2VDShuffle<0b00, 0b00011, "vzip", "8">; +def VZIPd16 : N2VDShuffle<0b01, 0b00011, "vzip", "16">; +// vzip.32 Dd, Dm is a pseudo-instruction expanded to vtrn.32 Dd, Dm. +def : NEONInstAlias<"vzip${p}.32 $Dd, $Dm", + (VTRNd32 DPR:$Dd, DPR:$Dm, pred:$p)>; + +def VZIPq8 : N2VQShuffle<0b00, 0b00011, IIC_VPERMQ3, "vzip", "8">; +def VZIPq16 : N2VQShuffle<0b01, 0b00011, IIC_VPERMQ3, "vzip", "16">; +def VZIPq32 : N2VQShuffle<0b10, 0b00011, IIC_VPERMQ3, "vzip", "32">; + +// Vector Table Lookup and Table Extension. + +// VTBL : Vector Table Lookup +let DecoderMethod = "DecodeTBLInstruction" in { +def VTBL1 + : N3V<1,1,0b11,0b1000,0,0, (outs DPR:$Vd), + (ins VecListOneD:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTB1, + "vtbl", "8", "$Vd, $Vn, $Vm", "", + [(set DPR:$Vd, (v8i8 (NEONvtbl1 VecListOneD:$Vn, DPR:$Vm)))]>; + +let hasExtraSrcRegAllocReq = 1 in { +def VTBL2 + : N3V<1,1,0b11,0b1001,0,0, (outs DPR:$Vd), + (ins VecListDPair:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTB2, + "vtbl", "8", "$Vd, $Vn, $Vm", "", []>; +def VTBL3 + : N3V<1,1,0b11,0b1010,0,0, (outs DPR:$Vd), + (ins VecListThreeD:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTB3, + "vtbl", "8", "$Vd, $Vn, $Vm", "", []>; +def VTBL4 + : N3V<1,1,0b11,0b1011,0,0, (outs DPR:$Vd), + (ins VecListFourD:$Vn, DPR:$Vm), + NVTBLFrm, IIC_VTB4, + "vtbl", "8", "$Vd, $Vn, $Vm", "", []>; +} // hasExtraSrcRegAllocReq = 1 + +def VTBL3Pseudo + : PseudoNeonI<(outs DPR:$dst), (ins QQPR:$tbl, DPR:$src), IIC_VTB3, "", []>; +def VTBL4Pseudo + : PseudoNeonI<(outs DPR:$dst), (ins QQPR:$tbl, DPR:$src), IIC_VTB4, "", []>; + +// VTBX : Vector Table Extension +def VTBX1 + : N3V<1,1,0b11,0b1000,1,0, (outs DPR:$Vd), + (ins DPR:$orig, VecListOneD:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTBX1, + "vtbx", "8", "$Vd, $Vn, $Vm", "$orig = $Vd", + [(set DPR:$Vd, (v8i8 (int_arm_neon_vtbx1 + DPR:$orig, VecListOneD:$Vn, DPR:$Vm)))]>; +let hasExtraSrcRegAllocReq = 1 in { +def VTBX2 + : N3V<1,1,0b11,0b1001,1,0, (outs DPR:$Vd), + (ins DPR:$orig, VecListDPair:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTBX2, + "vtbx", "8", "$Vd, $Vn, $Vm", "$orig = $Vd", []>; +def VTBX3 + : N3V<1,1,0b11,0b1010,1,0, (outs DPR:$Vd), + (ins DPR:$orig, VecListThreeD:$Vn, DPR:$Vm), + NVTBLFrm, IIC_VTBX3, + "vtbx", "8", "$Vd, $Vn, $Vm", + "$orig = $Vd", []>; +def VTBX4 + : N3V<1,1,0b11,0b1011,1,0, (outs DPR:$Vd), + (ins DPR:$orig, VecListFourD:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTBX4, + "vtbx", "8", "$Vd, $Vn, $Vm", + "$orig = $Vd", []>; +} // hasExtraSrcRegAllocReq = 1 + +def VTBX3Pseudo + : PseudoNeonI<(outs DPR:$dst), (ins DPR:$orig, QQPR:$tbl, DPR:$src), + IIC_VTBX3, "$orig = $dst", []>; +def VTBX4Pseudo + : PseudoNeonI<(outs DPR:$dst), (ins DPR:$orig, QQPR:$tbl, DPR:$src), + IIC_VTBX4, "$orig = $dst", []>; +} // DecoderMethod = "DecodeTBLInstruction" + +def : Pat<(v8i8 (NEONvtbl2 v8i8:$Vn0, v8i8:$Vn1, v8i8:$Vm)), + (v8i8 (VTBL2 (REG_SEQUENCE DPair, v8i8:$Vn0, dsub_0, + v8i8:$Vn1, dsub_1), + v8i8:$Vm))>; +def : Pat<(v8i8 (int_arm_neon_vtbx2 v8i8:$orig, v8i8:$Vn0, v8i8:$Vn1, + v8i8:$Vm)), + (v8i8 (VTBX2 v8i8:$orig, + (REG_SEQUENCE DPair, v8i8:$Vn0, dsub_0, + v8i8:$Vn1, dsub_1), + v8i8:$Vm))>; + +def : Pat<(v8i8 (int_arm_neon_vtbl3 v8i8:$Vn0, v8i8:$Vn1, + v8i8:$Vn2, v8i8:$Vm)), + (v8i8 (VTBL3Pseudo (REG_SEQUENCE QQPR, v8i8:$Vn0, dsub_0, + v8i8:$Vn1, dsub_1, + v8i8:$Vn2, dsub_2, + (v8i8 (IMPLICIT_DEF)), dsub_3), + v8i8:$Vm))>; +def : Pat<(v8i8 (int_arm_neon_vtbx3 v8i8:$orig, v8i8:$Vn0, v8i8:$Vn1, + v8i8:$Vn2, v8i8:$Vm)), + (v8i8 (VTBX3Pseudo v8i8:$orig, + (REG_SEQUENCE QQPR, v8i8:$Vn0, dsub_0, + v8i8:$Vn1, dsub_1, + v8i8:$Vn2, dsub_2, + (v8i8 (IMPLICIT_DEF)), dsub_3), + v8i8:$Vm))>; + +def : Pat<(v8i8 (int_arm_neon_vtbl4 v8i8:$Vn0, v8i8:$Vn1, + v8i8:$Vn2, v8i8:$Vn3, v8i8:$Vm)), + (v8i8 (VTBL4Pseudo (REG_SEQUENCE QQPR, v8i8:$Vn0, dsub_0, + v8i8:$Vn1, dsub_1, + v8i8:$Vn2, dsub_2, + v8i8:$Vn3, dsub_3), + v8i8:$Vm))>; +def : Pat<(v8i8 (int_arm_neon_vtbx4 v8i8:$orig, v8i8:$Vn0, v8i8:$Vn1, + v8i8:$Vn2, v8i8:$Vn3, v8i8:$Vm)), + (v8i8 (VTBX4Pseudo v8i8:$orig, + (REG_SEQUENCE QQPR, v8i8:$Vn0, dsub_0, + v8i8:$Vn1, dsub_1, + v8i8:$Vn2, dsub_2, + v8i8:$Vn3, dsub_3), + v8i8:$Vm))>; + +// VRINT : Vector Rounding +multiclass VRINT_FPI<string op, bits<3> op9_7, SDPatternOperator Int> { + let PostEncoderMethod = "NEONThumb2V8PostEncoder", DecoderNamespace = "v8NEON" in { + def Df : N2VDIntnp<0b10, 0b10, 0b100, 0, NoItinerary, + !strconcat("vrint", op), "f32", + v2f32, v2f32, Int>, Requires<[HasV8, HasNEON]> { + let Inst{9-7} = op9_7; + } + def Qf : N2VQIntnp<0b10, 0b10, 0b100, 0, NoItinerary, + !strconcat("vrint", op), "f32", + v4f32, v4f32, Int>, Requires<[HasV8, HasNEON]> { + let Inst{9-7} = op9_7; + } + def Dh : N2VDIntnp<0b01, 0b10, 0b100, 0, NoItinerary, + !strconcat("vrint", op), "f16", + v4f16, v4f16, Int>, + Requires<[HasV8, HasNEON, HasFullFP16]> { + let Inst{9-7} = op9_7; + } + def Qh : N2VQIntnp<0b01, 0b10, 0b100, 0, NoItinerary, + !strconcat("vrint", op), "f16", + v8f16, v8f16, Int>, + Requires<[HasV8, HasNEON, HasFullFP16]> { + let Inst{9-7} = op9_7; + } + } + + def : NEONInstAlias<!strconcat("vrint", op, ".f32.f32\t$Dd, $Dm"), + (!cast<Instruction>(NAME#"Df") DPR:$Dd, DPR:$Dm)>; + def : NEONInstAlias<!strconcat("vrint", op, ".f32.f32\t$Qd, $Qm"), + (!cast<Instruction>(NAME#"Qf") QPR:$Qd, QPR:$Qm)>; + let Predicates = [HasNEON, HasFullFP16] in { + def : NEONInstAlias<!strconcat("vrint", op, ".f16.f16\t$Dd, $Dm"), + (!cast<Instruction>(NAME#"Dh") DPR:$Dd, DPR:$Dm)>; + def : NEONInstAlias<!strconcat("vrint", op, ".f16.f16\t$Qd, $Qm"), + (!cast<Instruction>(NAME#"Qh") QPR:$Qd, QPR:$Qm)>; + } +} + +defm VRINTNN : VRINT_FPI<"n", 0b000, int_arm_neon_vrintn>; +defm VRINTXN : VRINT_FPI<"x", 0b001, int_arm_neon_vrintx>; +defm VRINTAN : VRINT_FPI<"a", 0b010, int_arm_neon_vrinta>; +defm VRINTZN : VRINT_FPI<"z", 0b011, int_arm_neon_vrintz>; +defm VRINTMN : VRINT_FPI<"m", 0b101, int_arm_neon_vrintm>; +defm VRINTPN : VRINT_FPI<"p", 0b111, int_arm_neon_vrintp>; + +// Cryptography instructions +let PostEncoderMethod = "NEONThumb2DataIPostEncoder", + DecoderNamespace = "v8Crypto", hasSideEffects = 0 in { + class AES<string op, bit op7, bit op6, SDPatternOperator Int> + : N2VQIntXnp<0b00, 0b00, 0b011, op6, op7, NoItinerary, + !strconcat("aes", op), "8", v16i8, v16i8, Int>, + Requires<[HasV8, HasCrypto]>; + class AES2Op<string op, bit op7, bit op6, SDPatternOperator Int> + : N2VQIntX2np<0b00, 0b00, 0b011, op6, op7, NoItinerary, + !strconcat("aes", op), "8", v16i8, v16i8, Int>, + Requires<[HasV8, HasCrypto]>; + class N2SHA<string op, bits<2> op17_16, bits<3> op10_8, bit op7, bit op6, + SDPatternOperator Int> + : N2VQIntXnp<0b10, op17_16, op10_8, op6, op7, NoItinerary, + !strconcat("sha", op), "32", v4i32, v4i32, Int>, + Requires<[HasV8, HasCrypto]>; + class N2SHA2Op<string op, bits<2> op17_16, bits<3> op10_8, bit op7, bit op6, + SDPatternOperator Int> + : N2VQIntX2np<0b10, op17_16, op10_8, op6, op7, NoItinerary, + !strconcat("sha", op), "32", v4i32, v4i32, Int>, + Requires<[HasV8, HasCrypto]>; + class N3SHA3Op<string op, bits<5> op27_23, bits<2> op21_20, SDPatternOperator Int> + : N3VQInt3np<op27_23, op21_20, 0b1100, 1, 0, N3RegFrm, NoItinerary, + !strconcat("sha", op), "32", v4i32, v4i32, Int, 0>, + Requires<[HasV8, HasCrypto]>; +} + +def AESD : AES2Op<"d", 0, 1, int_arm_neon_aesd>; +def AESE : AES2Op<"e", 0, 0, int_arm_neon_aese>; +def AESIMC : AES<"imc", 1, 1, int_arm_neon_aesimc>; +def AESMC : AES<"mc", 1, 0, int_arm_neon_aesmc>; + +def SHA1H : N2SHA<"1h", 0b01, 0b010, 1, 1, null_frag>; +def SHA1SU1 : N2SHA2Op<"1su1", 0b10, 0b011, 1, 0, int_arm_neon_sha1su1>; +def SHA256SU0 : N2SHA2Op<"256su0", 0b10, 0b011, 1, 1, int_arm_neon_sha256su0>; +def SHA1C : N3SHA3Op<"1c", 0b00100, 0b00, null_frag>; +def SHA1M : N3SHA3Op<"1m", 0b00100, 0b10, null_frag>; +def SHA1P : N3SHA3Op<"1p", 0b00100, 0b01, null_frag>; +def SHA1SU0 : N3SHA3Op<"1su0", 0b00100, 0b11, int_arm_neon_sha1su0>; +def SHA256H : N3SHA3Op<"256h", 0b00110, 0b00, int_arm_neon_sha256h>; +def SHA256H2 : N3SHA3Op<"256h2", 0b00110, 0b01, int_arm_neon_sha256h2>; +def SHA256SU1 : N3SHA3Op<"256su1", 0b00110, 0b10, int_arm_neon_sha256su1>; + +def : Pat<(i32 (int_arm_neon_sha1h i32:$Rn)), + (COPY_TO_REGCLASS (f32 (EXTRACT_SUBREG + (SHA1H (SUBREG_TO_REG (i64 0), + (f32 (COPY_TO_REGCLASS i32:$Rn, SPR)), + ssub_0)), + ssub_0)), GPR)>; + +def : Pat<(v4i32 (int_arm_neon_sha1c v4i32:$hash_abcd, i32:$hash_e, v4i32:$wk)), + (SHA1C v4i32:$hash_abcd, + (SUBREG_TO_REG (i64 0), + (f32 (COPY_TO_REGCLASS i32:$hash_e, SPR)), + ssub_0), + v4i32:$wk)>; + +def : Pat<(v4i32 (int_arm_neon_sha1m v4i32:$hash_abcd, i32:$hash_e, v4i32:$wk)), + (SHA1M v4i32:$hash_abcd, + (SUBREG_TO_REG (i64 0), + (f32 (COPY_TO_REGCLASS i32:$hash_e, SPR)), + ssub_0), + v4i32:$wk)>; + +def : Pat<(v4i32 (int_arm_neon_sha1p v4i32:$hash_abcd, i32:$hash_e, v4i32:$wk)), + (SHA1P v4i32:$hash_abcd, + (SUBREG_TO_REG (i64 0), + (f32 (COPY_TO_REGCLASS i32:$hash_e, SPR)), + ssub_0), + v4i32:$wk)>; + +//===----------------------------------------------------------------------===// +// NEON instructions for single-precision FP math +//===----------------------------------------------------------------------===// + +class N2VSPat<SDNode OpNode, NeonI Inst> + : NEONFPPat<(f32 (OpNode SPR:$a)), + (EXTRACT_SUBREG + (v2f32 (COPY_TO_REGCLASS (Inst + (INSERT_SUBREG + (v2f32 (COPY_TO_REGCLASS (v2f32 (IMPLICIT_DEF)), DPR_VFP2)), + SPR:$a, ssub_0)), DPR_VFP2)), ssub_0)>; + +class N3VSPat<SDNode OpNode, NeonI Inst> + : NEONFPPat<(f32 (OpNode SPR:$a, SPR:$b)), + (EXTRACT_SUBREG + (v2f32 (COPY_TO_REGCLASS (Inst + (INSERT_SUBREG + (v2f32 (COPY_TO_REGCLASS (v2f32 (IMPLICIT_DEF)), DPR_VFP2)), + SPR:$a, ssub_0), + (INSERT_SUBREG + (v2f32 (COPY_TO_REGCLASS (v2f32 (IMPLICIT_DEF)), DPR_VFP2)), + SPR:$b, ssub_0)), DPR_VFP2)), ssub_0)>; + +class N3VSPatFP16<SDNode OpNode, NeonI Inst> + : NEONFPPat<(f16 (OpNode HPR:$a, HPR:$b)), + (EXTRACT_SUBREG + (v4f16 (COPY_TO_REGCLASS (Inst + (INSERT_SUBREG + (v4f16 (COPY_TO_REGCLASS (v4f16 (IMPLICIT_DEF)), DPR_VFP2)), + HPR:$a, ssub_0), + (INSERT_SUBREG + (v4f16 (COPY_TO_REGCLASS (v4f16 (IMPLICIT_DEF)), DPR_VFP2)), + HPR:$b, ssub_0)), DPR_VFP2)), ssub_0)>; + +class N3VSMulOpPat<SDNode MulNode, SDNode OpNode, NeonI Inst> + : NEONFPPat<(f32 (OpNode SPR:$acc, (f32 (MulNode SPR:$a, SPR:$b)))), + (EXTRACT_SUBREG + (v2f32 (COPY_TO_REGCLASS (Inst + (INSERT_SUBREG + (v2f32 (COPY_TO_REGCLASS (v2f32 (IMPLICIT_DEF)), DPR_VFP2)), + SPR:$acc, ssub_0), + (INSERT_SUBREG + (v2f32 (COPY_TO_REGCLASS (v2f32 (IMPLICIT_DEF)), DPR_VFP2)), + SPR:$a, ssub_0), + (INSERT_SUBREG + (v2f32 (COPY_TO_REGCLASS (v2f32 (IMPLICIT_DEF)), DPR_VFP2)), + SPR:$b, ssub_0)), DPR_VFP2)), ssub_0)>; + +class NVCVTIFPat<SDNode OpNode, NeonI Inst> + : NEONFPPat<(f32 (OpNode GPR:$a)), + (f32 (EXTRACT_SUBREG + (v2f32 (Inst + (INSERT_SUBREG + (v2f32 (IMPLICIT_DEF)), + (i32 (COPY_TO_REGCLASS GPR:$a, SPR)), ssub_0))), + ssub_0))>; +class NVCVTFIPat<SDNode OpNode, NeonI Inst> + : NEONFPPat<(i32 (OpNode SPR:$a)), + (i32 (EXTRACT_SUBREG + (v2f32 (Inst (INSERT_SUBREG (v2f32 (IMPLICIT_DEF)), + SPR:$a, ssub_0))), + ssub_0))>; + +def : N3VSPat<fadd, VADDfd>; +def : N3VSPat<fsub, VSUBfd>; +def : N3VSPat<fmul, VMULfd>; +def : N3VSMulOpPat<fmul, fadd, VMLAfd>, + Requires<[HasNEON, UseNEONForFP, UseFPVMLx, DontUseFusedMAC]>; +def : N3VSMulOpPat<fmul, fsub, VMLSfd>, + Requires<[HasNEON, UseNEONForFP, UseFPVMLx, DontUseFusedMAC]>; +def : N3VSMulOpPat<fmul, fadd, VFMAfd>, + Requires<[HasVFP4, UseNEONForFP, UseFusedMAC]>; +def : N3VSMulOpPat<fmul, fsub, VFMSfd>, + Requires<[HasVFP4, UseNEONForFP, UseFusedMAC]>; +def : N2VSPat<fabs, VABSfd>; +def : N2VSPat<fneg, VNEGfd>; +def : N3VSPatFP16<fmaxnan, VMAXhd>, Requires<[HasFullFP16]>; +def : N3VSPatFP16<fminnan, VMINhd>, Requires<[HasFullFP16]>; +def : N3VSPat<fmaxnan, VMAXfd>, Requires<[HasNEON]>; +def : N3VSPat<fminnan, VMINfd>, Requires<[HasNEON]>; +def : NVCVTFIPat<fp_to_sint, VCVTf2sd>; +def : NVCVTFIPat<fp_to_uint, VCVTf2ud>; +def : NVCVTIFPat<sint_to_fp, VCVTs2fd>; +def : NVCVTIFPat<uint_to_fp, VCVTu2fd>; + +// NEON doesn't have any f64 conversions, so provide patterns to make +// sure the VFP conversions match when extracting from a vector. +def : VFPPat<(f64 (sint_to_fp (extractelt (v2i32 DPR:$src), imm:$lane))), + (VSITOD (EXTRACT_SUBREG DPR:$src, (SSubReg_f32_reg imm:$lane)))>; +def : VFPPat<(f64 (sint_to_fp (extractelt (v4i32 QPR:$src), imm:$lane))), + (VSITOD (EXTRACT_SUBREG QPR:$src, (SSubReg_f32_reg imm:$lane)))>; +def : VFPPat<(f64 (uint_to_fp (extractelt (v2i32 DPR:$src), imm:$lane))), + (VUITOD (EXTRACT_SUBREG DPR:$src, (SSubReg_f32_reg imm:$lane)))>; +def : VFPPat<(f64 (uint_to_fp (extractelt (v4i32 QPR:$src), imm:$lane))), + (VUITOD (EXTRACT_SUBREG QPR:$src, (SSubReg_f32_reg imm:$lane)))>; + + +// Prefer VMOVDRR for i32 -> f32 bitcasts, it can write all DPR registers. +def : Pat<(f32 (bitconvert GPR:$a)), + (EXTRACT_SUBREG (VMOVDRR GPR:$a, GPR:$a), ssub_0)>, + Requires<[HasNEON, DontUseVMOVSR]>; +def : Pat<(arm_vmovsr GPR:$a), + (EXTRACT_SUBREG (VMOVDRR GPR:$a, GPR:$a), ssub_0)>, + Requires<[HasNEON, DontUseVMOVSR]>; + +//===----------------------------------------------------------------------===// +// Non-Instruction Patterns +//===----------------------------------------------------------------------===// + +// bit_convert +let Predicates = [IsLE] in { + def : Pat<(v1i64 (bitconvert (v2i32 DPR:$src))), (v1i64 DPR:$src)>; + def : Pat<(v1i64 (bitconvert (v4i16 DPR:$src))), (v1i64 DPR:$src)>; + def : Pat<(v1i64 (bitconvert (v8i8 DPR:$src))), (v1i64 DPR:$src)>; +} +def : Pat<(v1i64 (bitconvert (f64 DPR:$src))), (v1i64 DPR:$src)>; +let Predicates = [IsLE] in { + def : Pat<(v1i64 (bitconvert (v2f32 DPR:$src))), (v1i64 DPR:$src)>; + def : Pat<(v2i32 (bitconvert (v1i64 DPR:$src))), (v2i32 DPR:$src)>; + def : Pat<(v2i32 (bitconvert (v4i16 DPR:$src))), (v2i32 DPR:$src)>; + def : Pat<(v2i32 (bitconvert (v8i8 DPR:$src))), (v2i32 DPR:$src)>; + def : Pat<(v2i32 (bitconvert (f64 DPR:$src))), (v2i32 DPR:$src)>; +} +def : Pat<(v2i32 (bitconvert (v2f32 DPR:$src))), (v2i32 DPR:$src)>; +let Predicates = [IsLE] in { + def : Pat<(v4i16 (bitconvert (v1i64 DPR:$src))), (v4i16 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (v2i32 DPR:$src))), (v4i16 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (v8i8 DPR:$src))), (v4i16 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (f64 DPR:$src))), (v4i16 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (v2f32 DPR:$src))), (v4i16 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (v1i64 DPR:$src))), (v8i8 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (v2i32 DPR:$src))), (v8i8 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (v4i16 DPR:$src))), (v8i8 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (f64 DPR:$src))), (v8i8 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (v2f32 DPR:$src))), (v8i8 DPR:$src)>; +} +def : Pat<(f64 (bitconvert (v1i64 DPR:$src))), (f64 DPR:$src)>; +let Predicates = [IsLE] in { + def : Pat<(f64 (bitconvert (v2i32 DPR:$src))), (f64 DPR:$src)>; + def : Pat<(f64 (bitconvert (v4i16 DPR:$src))), (f64 DPR:$src)>; + def : Pat<(f64 (bitconvert (v4f16 DPR:$src))), (f64 DPR:$src)>; + def : Pat<(f64 (bitconvert (v8i8 DPR:$src))), (f64 DPR:$src)>; + def : Pat<(f64 (bitconvert (v2f32 DPR:$src))), (f64 DPR:$src)>; + def : Pat<(v2f32 (bitconvert (f64 DPR:$src))), (v2f32 DPR:$src)>; + def : Pat<(v4f16 (bitconvert (f64 DPR:$src))), (v4f16 DPR:$src)>; + def : Pat<(v2f32 (bitconvert (v1i64 DPR:$src))), (v2f32 DPR:$src)>; +} +def : Pat<(v2f32 (bitconvert (v2i32 DPR:$src))), (v2f32 DPR:$src)>; +let Predicates = [IsLE] in { + def : Pat<(v2f32 (bitconvert (v4i16 DPR:$src))), (v2f32 DPR:$src)>; + def : Pat<(v2f32 (bitconvert (v8i8 DPR:$src))), (v2f32 DPR:$src)>; +} + +let Predicates = [IsLE] in { + def : Pat<(v2i64 (bitconvert (v4i32 QPR:$src))), (v2i64 QPR:$src)>; + def : Pat<(v2i64 (bitconvert (v8i16 QPR:$src))), (v2i64 QPR:$src)>; + def : Pat<(v2i64 (bitconvert (v16i8 QPR:$src))), (v2i64 QPR:$src)>; +} +def : Pat<(v2i64 (bitconvert (v2f64 QPR:$src))), (v2i64 QPR:$src)>; +let Predicates = [IsLE] in { + def : Pat<(v2i64 (bitconvert (v4f32 QPR:$src))), (v2i64 QPR:$src)>; + def : Pat<(v4i32 (bitconvert (v2i64 QPR:$src))), (v4i32 QPR:$src)>; + def : Pat<(v4i32 (bitconvert (v8i16 QPR:$src))), (v4i32 QPR:$src)>; + def : Pat<(v4i32 (bitconvert (v16i8 QPR:$src))), (v4i32 QPR:$src)>; + def : Pat<(v4i32 (bitconvert (v2f64 QPR:$src))), (v4i32 QPR:$src)>; +} +def : Pat<(v4i32 (bitconvert (v4f32 QPR:$src))), (v4i32 QPR:$src)>; +let Predicates = [IsLE] in { + def : Pat<(v8i16 (bitconvert (v2i64 QPR:$src))), (v8i16 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v4i32 QPR:$src))), (v8i16 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v16i8 QPR:$src))), (v8i16 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v2f64 QPR:$src))), (v8i16 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v4f32 QPR:$src))), (v8i16 QPR:$src)>; + def : Pat<(v8f16 (bitconvert (v2f64 QPR:$src))), (v8f16 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v2i64 QPR:$src))), (v16i8 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v4i32 QPR:$src))), (v16i8 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v8i16 QPR:$src))), (v16i8 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v2f64 QPR:$src))), (v16i8 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v4f32 QPR:$src))), (v16i8 QPR:$src)>; + def : Pat<(v4f32 (bitconvert (v2i64 QPR:$src))), (v4f32 QPR:$src)>; +} +def : Pat<(v4f32 (bitconvert (v4i32 QPR:$src))), (v4f32 QPR:$src)>; +let Predicates = [IsLE] in { + def : Pat<(v4f32 (bitconvert (v8i16 QPR:$src))), (v4f32 QPR:$src)>; + def : Pat<(v4f32 (bitconvert (v16i8 QPR:$src))), (v4f32 QPR:$src)>; + def : Pat<(v4f32 (bitconvert (v2f64 QPR:$src))), (v4f32 QPR:$src)>; +} +def : Pat<(v2f64 (bitconvert (v2i64 QPR:$src))), (v2f64 QPR:$src)>; +let Predicates = [IsLE] in { + def : Pat<(v2f64 (bitconvert (v4i32 QPR:$src))), (v2f64 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v8i16 QPR:$src))), (v2f64 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v8f16 QPR:$src))), (v2f64 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v16i8 QPR:$src))), (v2f64 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v4f32 QPR:$src))), (v2f64 QPR:$src)>; +} + +let Predicates = [IsBE] in { + // 64 bit conversions + def : Pat<(v1i64 (bitconvert (v2i32 DPR:$src))), (VREV64d32 DPR:$src)>; + def : Pat<(v1i64 (bitconvert (v4i16 DPR:$src))), (VREV64d16 DPR:$src)>; + def : Pat<(v1i64 (bitconvert (v8i8 DPR:$src))), (VREV64d8 DPR:$src)>; + def : Pat<(v1i64 (bitconvert (v2f32 DPR:$src))), (VREV64d32 DPR:$src)>; + def : Pat<(v2i32 (bitconvert (v1i64 DPR:$src))), (VREV64d32 DPR:$src)>; + def : Pat<(v2i32 (bitconvert (v4i16 DPR:$src))), (VREV32d16 DPR:$src)>; + def : Pat<(v2i32 (bitconvert (v8i8 DPR:$src))), (VREV32d8 DPR:$src)>; + def : Pat<(v2i32 (bitconvert (f64 DPR:$src))), (VREV64d32 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (v1i64 DPR:$src))), (VREV64d16 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (v2i32 DPR:$src))), (VREV32d16 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (v8i8 DPR:$src))), (VREV16d8 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (f64 DPR:$src))), (VREV64d16 DPR:$src)>; + def : Pat<(v4i16 (bitconvert (v2f32 DPR:$src))), (VREV32d16 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (v1i64 DPR:$src))), (VREV64d8 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (v2i32 DPR:$src))), (VREV32d8 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (v4i16 DPR:$src))), (VREV16d8 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (f64 DPR:$src))), (VREV64d8 DPR:$src)>; + def : Pat<(v8i8 (bitconvert (v2f32 DPR:$src))), (VREV32d8 DPR:$src)>; + def : Pat<(f64 (bitconvert (v2i32 DPR:$src))), (VREV64d32 DPR:$src)>; + def : Pat<(f64 (bitconvert (v4f16 DPR:$src))), (VREV64d16 DPR:$src)>; + def : Pat<(f64 (bitconvert (v4i16 DPR:$src))), (VREV64d16 DPR:$src)>; + def : Pat<(f64 (bitconvert (v8i8 DPR:$src))), (VREV64d8 DPR:$src)>; + def : Pat<(f64 (bitconvert (v2f32 DPR:$src))), (VREV64d32 DPR:$src)>; + def : Pat<(v2f32 (bitconvert (f64 DPR:$src))), (VREV64d32 DPR:$src)>; + def : Pat<(v2f32 (bitconvert (v1i64 DPR:$src))), (VREV64d32 DPR:$src)>; + def : Pat<(v2f32 (bitconvert (v4i16 DPR:$src))), (VREV32d16 DPR:$src)>; + def : Pat<(v2f32 (bitconvert (v8i8 DPR:$src))), (VREV32d8 DPR:$src)>; + + // 128 bit conversions + def : Pat<(v2i64 (bitconvert (v4i32 QPR:$src))), (VREV64q32 QPR:$src)>; + def : Pat<(v2i64 (bitconvert (v8i16 QPR:$src))), (VREV64q16 QPR:$src)>; + def : Pat<(v2i64 (bitconvert (v16i8 QPR:$src))), (VREV64q8 QPR:$src)>; + def : Pat<(v2i64 (bitconvert (v4f32 QPR:$src))), (VREV64q32 QPR:$src)>; + def : Pat<(v4i32 (bitconvert (v2i64 QPR:$src))), (VREV64q32 QPR:$src)>; + def : Pat<(v4i32 (bitconvert (v8i16 QPR:$src))), (VREV32q16 QPR:$src)>; + def : Pat<(v4i32 (bitconvert (v16i8 QPR:$src))), (VREV32q8 QPR:$src)>; + def : Pat<(v4i32 (bitconvert (v2f64 QPR:$src))), (VREV64q32 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v2i64 QPR:$src))), (VREV64q16 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v4i32 QPR:$src))), (VREV32q16 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v16i8 QPR:$src))), (VREV16q8 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v2f64 QPR:$src))), (VREV64q16 QPR:$src)>; + def : Pat<(v8f16 (bitconvert (v2f64 QPR:$src))), (VREV64q16 QPR:$src)>; + def : Pat<(v8i16 (bitconvert (v4f32 QPR:$src))), (VREV32q16 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v2i64 QPR:$src))), (VREV64q8 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v4i32 QPR:$src))), (VREV32q8 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v8i16 QPR:$src))), (VREV16q8 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v2f64 QPR:$src))), (VREV64q8 QPR:$src)>; + def : Pat<(v16i8 (bitconvert (v4f32 QPR:$src))), (VREV32q8 QPR:$src)>; + def : Pat<(v4f32 (bitconvert (v2i64 QPR:$src))), (VREV64q32 QPR:$src)>; + def : Pat<(v4f32 (bitconvert (v8i16 QPR:$src))), (VREV32q16 QPR:$src)>; + def : Pat<(v4f32 (bitconvert (v8f16 QPR:$src))), (VREV32q16 QPR:$src)>; + def : Pat<(v4f32 (bitconvert (v16i8 QPR:$src))), (VREV32q8 QPR:$src)>; + def : Pat<(v4f32 (bitconvert (v2f64 QPR:$src))), (VREV64q32 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v4i32 QPR:$src))), (VREV64q32 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v8i16 QPR:$src))), (VREV64q16 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v8f16 QPR:$src))), (VREV64q16 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v16i8 QPR:$src))), (VREV64q8 QPR:$src)>; + def : Pat<(v2f64 (bitconvert (v4f32 QPR:$src))), (VREV64q32 QPR:$src)>; +} + +// Use VLD1/VST1 + VREV for non-word-aligned v2f64 load/store on Big Endian +def : Pat<(v2f64 (byte_alignedload addrmode6:$addr)), + (VREV64q8 (VLD1q8 addrmode6:$addr))>, Requires<[IsBE]>; +def : Pat<(byte_alignedstore (v2f64 QPR:$value), addrmode6:$addr), + (VST1q8 addrmode6:$addr, (VREV64q8 QPR:$value))>, Requires<[IsBE]>; +def : Pat<(v2f64 (hword_alignedload addrmode6:$addr)), + (VREV64q16 (VLD1q16 addrmode6:$addr))>, Requires<[IsBE]>; +def : Pat<(hword_alignedstore (v2f64 QPR:$value), addrmode6:$addr), + (VST1q16 addrmode6:$addr, (VREV64q16 QPR:$value))>, Requires<[IsBE]>; + +// Fold extracting an element out of a v2i32 into a vfp register. +def : Pat<(f32 (bitconvert (i32 (extractelt (v2i32 DPR:$src), imm:$lane)))), + (f32 (EXTRACT_SUBREG DPR:$src, (SSubReg_f32_reg imm:$lane)))>; + +// Vector lengthening move with load, matching extending loads. + +// extload, zextload and sextload for a standard lengthening load. Example: +// Lengthen_Single<"8", "i16", "8"> = +// Pat<(v8i16 (extloadvi8 addrmode6:$addr)) +// (VMOVLuv8i16 (VLD1d8 addrmode6:$addr, +// (f64 (IMPLICIT_DEF)), (i32 0)))>; +multiclass Lengthen_Single<string DestLanes, string DestTy, string SrcTy> { + let AddedComplexity = 10 in { + def _Any : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("extloadvi" # SrcTy) addrmode6:$addr)), + (!cast<Instruction>("VMOVLuv" # DestLanes # DestTy) + (!cast<Instruction>("VLD1d" # SrcTy) addrmode6:$addr))>; + + def _Z : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("zextloadvi" # SrcTy) addrmode6:$addr)), + (!cast<Instruction>("VMOVLuv" # DestLanes # DestTy) + (!cast<Instruction>("VLD1d" # SrcTy) addrmode6:$addr))>; + + def _S : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("sextloadvi" # SrcTy) addrmode6:$addr)), + (!cast<Instruction>("VMOVLsv" # DestLanes # DestTy) + (!cast<Instruction>("VLD1d" # SrcTy) addrmode6:$addr))>; + } +} + +// extload, zextload and sextload for a lengthening load which only uses +// half the lanes available. Example: +// Lengthen_HalfSingle<"4", "i16", "8", "i16", "i8"> = +// Pat<(v4i16 (extloadvi8 addrmode6oneL32:$addr)), +// (EXTRACT_SUBREG (VMOVLuv8i16 (VLD1LNd32 addrmode6oneL32:$addr, +// (f64 (IMPLICIT_DEF)), (i32 0))), +// dsub_0)>; +multiclass Lengthen_HalfSingle<string DestLanes, string DestTy, string SrcTy, + string InsnLanes, string InsnTy> { + def _Any : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("extloadv" # SrcTy) addrmode6oneL32:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # InsnLanes # InsnTy) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0)>; + def _Z : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("zextloadv" # SrcTy) addrmode6oneL32:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # InsnLanes # InsnTy) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0)>; + def _S : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("sextloadv" # SrcTy) addrmode6oneL32:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLsv" # InsnLanes # InsnTy) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0)>; +} + +// The following class definition is basically a copy of the +// Lengthen_HalfSingle definition above, however with an additional parameter +// "RevLanes" to select the correct VREV32dXX instruction. This is to convert +// data loaded by VLD1LN into proper vector format in big endian mode. +multiclass Lengthen_HalfSingle_Big_Endian<string DestLanes, string DestTy, string SrcTy, + string InsnLanes, string InsnTy, string RevLanes> { + def _Any : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("extloadv" # SrcTy) addrmode6oneL32:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # InsnLanes # InsnTy) + (!cast<Instruction>("VREV32d" # RevLanes) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0)>; + def _Z : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("zextloadv" # SrcTy) addrmode6oneL32:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # InsnLanes # InsnTy) + (!cast<Instruction>("VREV32d" # RevLanes) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0)>; + def _S : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("sextloadv" # SrcTy) addrmode6oneL32:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLsv" # InsnLanes # InsnTy) + (!cast<Instruction>("VREV32d" # RevLanes) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0)>; +} + +// extload, zextload and sextload for a lengthening load followed by another +// lengthening load, to quadruple the initial length. +// +// Lengthen_Double<"4", "i32", "i8", "8", "i16", "4", "i32"> = +// Pat<(v4i32 (extloadvi8 addrmode6oneL32:$addr)) +// (EXTRACT_SUBREG (VMOVLuv4i32 +// (EXTRACT_SUBREG (VMOVLuv8i16 (VLD1LNd32 addrmode6oneL32:$addr, +// (f64 (IMPLICIT_DEF)), +// (i32 0))), +// dsub_0)), +// dsub_0)>; +multiclass Lengthen_Double<string DestLanes, string DestTy, string SrcTy, + string Insn1Lanes, string Insn1Ty, string Insn2Lanes, + string Insn2Ty> { + def _Any : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("extloadv" # SrcTy) addrmode6oneL32:$addr)), + (!cast<Instruction>("VMOVLuv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn1Lanes # Insn1Ty) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0))>; + def _Z : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("zextloadv" # SrcTy) addrmode6oneL32:$addr)), + (!cast<Instruction>("VMOVLuv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn1Lanes # Insn1Ty) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0))>; + def _S : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("sextloadv" # SrcTy) addrmode6oneL32:$addr)), + (!cast<Instruction>("VMOVLsv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLsv" # Insn1Lanes # Insn1Ty) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0))>; +} + +// The following class definition is basically a copy of the +// Lengthen_Double definition above, however with an additional parameter +// "RevLanes" to select the correct VREV32dXX instruction. This is to convert +// data loaded by VLD1LN into proper vector format in big endian mode. +multiclass Lengthen_Double_Big_Endian<string DestLanes, string DestTy, string SrcTy, + string Insn1Lanes, string Insn1Ty, string Insn2Lanes, + string Insn2Ty, string RevLanes> { + def _Any : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("extloadv" # SrcTy) addrmode6oneL32:$addr)), + (!cast<Instruction>("VMOVLuv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn1Lanes # Insn1Ty) + (!cast<Instruction>("VREV32d" # RevLanes) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0))>; + def _Z : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("zextloadv" # SrcTy) addrmode6oneL32:$addr)), + (!cast<Instruction>("VMOVLuv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn1Lanes # Insn1Ty) + (!cast<Instruction>("VREV32d" # RevLanes) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0))>; + def _S : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("sextloadv" # SrcTy) addrmode6oneL32:$addr)), + (!cast<Instruction>("VMOVLsv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLsv" # Insn1Lanes # Insn1Ty) + (!cast<Instruction>("VREV32d" # RevLanes) + (VLD1LNd32 addrmode6oneL32:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0))>; +} + +// extload, zextload and sextload for a lengthening load followed by another +// lengthening load, to quadruple the initial length, but which ends up only +// requiring half the available lanes (a 64-bit outcome instead of a 128-bit). +// +// Lengthen_HalfDouble<"2", "i32", "i8", "8", "i16", "4", "i32"> = +// Pat<(v2i32 (extloadvi8 addrmode6:$addr)) +// (EXTRACT_SUBREG (VMOVLuv4i32 +// (EXTRACT_SUBREG (VMOVLuv8i16 (VLD1LNd16 addrmode6:$addr, +// (f64 (IMPLICIT_DEF)), (i32 0))), +// dsub_0)), +// dsub_0)>; +multiclass Lengthen_HalfDouble<string DestLanes, string DestTy, string SrcTy, + string Insn1Lanes, string Insn1Ty, string Insn2Lanes, + string Insn2Ty> { + def _Any : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("extloadv" # SrcTy) addrmode6:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn1Lanes # Insn1Ty) + (VLD1LNd16 addrmode6:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0)), + dsub_0)>; + def _Z : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("zextloadv" # SrcTy) addrmode6:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn1Lanes # Insn1Ty) + (VLD1LNd16 addrmode6:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0)), + dsub_0)>; + def _S : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("sextloadv" # SrcTy) addrmode6:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLsv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLsv" # Insn1Lanes # Insn1Ty) + (VLD1LNd16 addrmode6:$addr, (f64 (IMPLICIT_DEF)), (i32 0))), + dsub_0)), + dsub_0)>; +} + +// The following class definition is basically a copy of the +// Lengthen_HalfDouble definition above, however with an additional VREV16d8 +// instruction to convert data loaded by VLD1LN into proper vector format +// in big endian mode. +multiclass Lengthen_HalfDouble_Big_Endian<string DestLanes, string DestTy, string SrcTy, + string Insn1Lanes, string Insn1Ty, string Insn2Lanes, + string Insn2Ty> { + def _Any : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("extloadv" # SrcTy) addrmode6:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn1Lanes # Insn1Ty) + (!cast<Instruction>("VREV16d8") + (VLD1LNd16 addrmode6:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0)), + dsub_0)>; + def _Z : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("zextloadv" # SrcTy) addrmode6:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLuv" # Insn1Lanes # Insn1Ty) + (!cast<Instruction>("VREV16d8") + (VLD1LNd16 addrmode6:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0)), + dsub_0)>; + def _S : Pat<(!cast<ValueType>("v" # DestLanes # DestTy) + (!cast<PatFrag>("sextloadv" # SrcTy) addrmode6:$addr)), + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLsv" # Insn2Lanes # Insn2Ty) + (EXTRACT_SUBREG (!cast<Instruction>("VMOVLsv" # Insn1Lanes # Insn1Ty) + (!cast<Instruction>("VREV16d8") + (VLD1LNd16 addrmode6:$addr, (f64 (IMPLICIT_DEF)), (i32 0)))), + dsub_0)), + dsub_0)>; +} + +defm : Lengthen_Single<"8", "i16", "8">; // v8i8 -> v8i16 +defm : Lengthen_Single<"4", "i32", "16">; // v4i16 -> v4i32 +defm : Lengthen_Single<"2", "i64", "32">; // v2i32 -> v2i64 + +let Predicates = [IsLE] in { + defm : Lengthen_HalfSingle<"4", "i16", "i8", "8", "i16">; // v4i8 -> v4i16 + defm : Lengthen_HalfSingle<"2", "i32", "i16", "4", "i32">; // v2i16 -> v2i32 + + // Double lengthening - v4i8 -> v4i16 -> v4i32 + defm : Lengthen_Double<"4", "i32", "i8", "8", "i16", "4", "i32">; + // v2i8 -> v2i16 -> v2i32 + defm : Lengthen_HalfDouble<"2", "i32", "i8", "8", "i16", "4", "i32">; + // v2i16 -> v2i32 -> v2i64 + defm : Lengthen_Double<"2", "i64", "i16", "4", "i32", "2", "i64">; +} + +let Predicates = [IsBE] in { + defm : Lengthen_HalfSingle_Big_Endian<"4", "i16", "i8", "8", "i16", "8">; // v4i8 -> v4i16 + defm : Lengthen_HalfSingle_Big_Endian<"2", "i32", "i16", "4", "i32", "16">; // v2i16 -> v2i32 + + // Double lengthening - v4i8 -> v4i16 -> v4i32 + defm : Lengthen_Double_Big_Endian<"4", "i32", "i8", "8", "i16", "4", "i32", "8">; + // v2i8 -> v2i16 -> v2i32 + defm : Lengthen_HalfDouble_Big_Endian<"2", "i32", "i8", "8", "i16", "4", "i32">; + // v2i16 -> v2i32 -> v2i64 + defm : Lengthen_Double_Big_Endian<"2", "i64", "i16", "4", "i32", "2", "i64", "16">; +} + +// Triple lengthening - v2i8 -> v2i16 -> v2i32 -> v2i64 +let Predicates = [IsLE] in { + def : Pat<(v2i64 (extloadvi8 addrmode6:$addr)), + (VMOVLuv2i64 (EXTRACT_SUBREG (VMOVLuv4i32 (EXTRACT_SUBREG (VMOVLuv8i16 + (VLD1LNd16 addrmode6:$addr, + (f64 (IMPLICIT_DEF)), (i32 0))), dsub_0)), dsub_0))>; + def : Pat<(v2i64 (zextloadvi8 addrmode6:$addr)), + (VMOVLuv2i64 (EXTRACT_SUBREG (VMOVLuv4i32 (EXTRACT_SUBREG (VMOVLuv8i16 + (VLD1LNd16 addrmode6:$addr, + (f64 (IMPLICIT_DEF)), (i32 0))), dsub_0)), dsub_0))>; + def : Pat<(v2i64 (sextloadvi8 addrmode6:$addr)), + (VMOVLsv2i64 (EXTRACT_SUBREG (VMOVLsv4i32 (EXTRACT_SUBREG (VMOVLsv8i16 + (VLD1LNd16 addrmode6:$addr, + (f64 (IMPLICIT_DEF)), (i32 0))), dsub_0)), dsub_0))>; +} +// The following patterns are basically a copy of the patterns above, +// however with an additional VREV16d instruction to convert data +// loaded by VLD1LN into proper vector format in big endian mode. +let Predicates = [IsBE] in { + def : Pat<(v2i64 (extloadvi8 addrmode6:$addr)), + (VMOVLuv2i64 (EXTRACT_SUBREG (VMOVLuv4i32 (EXTRACT_SUBREG (VMOVLuv8i16 + (!cast<Instruction>("VREV16d8") + (VLD1LNd16 addrmode6:$addr, + (f64 (IMPLICIT_DEF)), (i32 0)))), dsub_0)), dsub_0))>; + def : Pat<(v2i64 (zextloadvi8 addrmode6:$addr)), + (VMOVLuv2i64 (EXTRACT_SUBREG (VMOVLuv4i32 (EXTRACT_SUBREG (VMOVLuv8i16 + (!cast<Instruction>("VREV16d8") + (VLD1LNd16 addrmode6:$addr, + (f64 (IMPLICIT_DEF)), (i32 0)))), dsub_0)), dsub_0))>; + def : Pat<(v2i64 (sextloadvi8 addrmode6:$addr)), + (VMOVLsv2i64 (EXTRACT_SUBREG (VMOVLsv4i32 (EXTRACT_SUBREG (VMOVLsv8i16 + (!cast<Instruction>("VREV16d8") + (VLD1LNd16 addrmode6:$addr, + (f64 (IMPLICIT_DEF)), (i32 0)))), dsub_0)), dsub_0))>; +} + +def : Pat<(v2i64 (concat_vectors DPR:$Dn, DPR:$Dm)), + (REG_SEQUENCE QPR, DPR:$Dn, dsub_0, DPR:$Dm, dsub_1)>; +def : Pat<(v4i32 (concat_vectors DPR:$Dn, DPR:$Dm)), + (REG_SEQUENCE QPR, DPR:$Dn, dsub_0, DPR:$Dm, dsub_1)>; +def : Pat<(v8i16 (concat_vectors DPR:$Dn, DPR:$Dm)), + (REG_SEQUENCE QPR, DPR:$Dn, dsub_0, DPR:$Dm, dsub_1)>; +def : Pat<(v16i8 (concat_vectors DPR:$Dn, DPR:$Dm)), + (REG_SEQUENCE QPR, DPR:$Dn, dsub_0, DPR:$Dm, dsub_1)>; +def : Pat<(v4f32 (concat_vectors DPR:$Dn, DPR:$Dm)), + (REG_SEQUENCE QPR, DPR:$Dn, dsub_0, DPR:$Dm, dsub_1)>; + +//===----------------------------------------------------------------------===// +// Assembler aliases +// + +def : VFP2InstAlias<"fmdhr${p} $Dd, $Rn", + (VSETLNi32 DPR:$Dd, GPR:$Rn, 1, pred:$p)>; +def : VFP2InstAlias<"fmdlr${p} $Dd, $Rn", + (VSETLNi32 DPR:$Dd, GPR:$Rn, 0, pred:$p)>; + +// VAND/VBIC/VEOR/VORR accept but do not require a type suffix. +defm : NEONDTAnyInstAlias<"vand${p}", "$Vd, $Vn, $Vm", + (VANDd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vand${p}", "$Vd, $Vn, $Vm", + (VANDq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vbic${p}", "$Vd, $Vn, $Vm", + (VBICd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vbic${p}", "$Vd, $Vn, $Vm", + (VBICq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"veor${p}", "$Vd, $Vn, $Vm", + (VEORd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"veor${p}", "$Vd, $Vn, $Vm", + (VEORq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vorr${p}", "$Vd, $Vn, $Vm", + (VORRd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vorr${p}", "$Vd, $Vn, $Vm", + (VORRq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>; +// ... two-operand aliases +defm : NEONDTAnyInstAlias<"vand${p}", "$Vdn, $Vm", + (VANDd DPR:$Vdn, DPR:$Vdn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vand${p}", "$Vdn, $Vm", + (VANDq QPR:$Vdn, QPR:$Vdn, QPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"veor${p}", "$Vdn, $Vm", + (VEORd DPR:$Vdn, DPR:$Vdn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"veor${p}", "$Vdn, $Vm", + (VEORq QPR:$Vdn, QPR:$Vdn, QPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vorr${p}", "$Vdn, $Vm", + (VORRd DPR:$Vdn, DPR:$Vdn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vorr${p}", "$Vdn, $Vm", + (VORRq QPR:$Vdn, QPR:$Vdn, QPR:$Vm, pred:$p)>; +// ... immediates +def : NEONInstAlias<"vand${p}.i16 $Vd, $imm", + (VBICiv4i16 DPR:$Vd, nImmSplatNotI16:$imm, pred:$p)>; +def : NEONInstAlias<"vand${p}.i32 $Vd, $imm", + (VBICiv2i32 DPR:$Vd, nImmSplatNotI32:$imm, pred:$p)>; +def : NEONInstAlias<"vand${p}.i16 $Vd, $imm", + (VBICiv8i16 QPR:$Vd, nImmSplatNotI16:$imm, pred:$p)>; +def : NEONInstAlias<"vand${p}.i32 $Vd, $imm", + (VBICiv4i32 QPR:$Vd, nImmSplatNotI32:$imm, pred:$p)>; + + +// VLD1 single-lane pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VLD1LNdAsm_8 : NEONDataTypeAsmPseudoInst<"vld1${p}", ".8", "$list, $addr", + (ins VecListOneDByteIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD1LNdAsm_16 : NEONDataTypeAsmPseudoInst<"vld1${p}", ".16", "$list, $addr", + (ins VecListOneDHWordIndexed:$list, addrmode6align16:$addr, + pred:$p)>; +def VLD1LNdAsm_32 : NEONDataTypeAsmPseudoInst<"vld1${p}", ".32", "$list, $addr", + (ins VecListOneDWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; + +def VLD1LNdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld1${p}", ".8", "$list, $addr!", + (ins VecListOneDByteIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD1LNdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld1${p}", ".16", "$list, $addr!", + (ins VecListOneDHWordIndexed:$list, addrmode6align16:$addr, + pred:$p)>; +def VLD1LNdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld1${p}", ".32", "$list, $addr!", + (ins VecListOneDWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VLD1LNdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld1${p}", ".8", "$list, $addr, $Rm", + (ins VecListOneDByteIndexed:$list, addrmode6alignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VLD1LNdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld1${p}", ".16", "$list, $addr, $Rm", + (ins VecListOneDHWordIndexed:$list, addrmode6align16:$addr, + rGPR:$Rm, pred:$p)>; +def VLD1LNdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld1${p}", ".32", "$list, $addr, $Rm", + (ins VecListOneDWordIndexed:$list, addrmode6align32:$addr, + rGPR:$Rm, pred:$p)>; + + +// VST1 single-lane pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VST1LNdAsm_8 : NEONDataTypeAsmPseudoInst<"vst1${p}", ".8", "$list, $addr", + (ins VecListOneDByteIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST1LNdAsm_16 : NEONDataTypeAsmPseudoInst<"vst1${p}", ".16", "$list, $addr", + (ins VecListOneDHWordIndexed:$list, addrmode6align16:$addr, + pred:$p)>; +def VST1LNdAsm_32 : NEONDataTypeAsmPseudoInst<"vst1${p}", ".32", "$list, $addr", + (ins VecListOneDWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; + +def VST1LNdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst1${p}", ".8", "$list, $addr!", + (ins VecListOneDByteIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST1LNdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst1${p}", ".16", "$list, $addr!", + (ins VecListOneDHWordIndexed:$list, addrmode6align16:$addr, + pred:$p)>; +def VST1LNdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst1${p}", ".32", "$list, $addr!", + (ins VecListOneDWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VST1LNdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst1${p}", ".8", "$list, $addr, $Rm", + (ins VecListOneDByteIndexed:$list, addrmode6alignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VST1LNdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst1${p}", ".16", "$list, $addr, $Rm", + (ins VecListOneDHWordIndexed:$list, addrmode6align16:$addr, + rGPR:$Rm, pred:$p)>; +def VST1LNdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst1${p}", ".32", "$list, $addr, $Rm", + (ins VecListOneDWordIndexed:$list, addrmode6align32:$addr, + rGPR:$Rm, pred:$p)>; + +// VLD2 single-lane pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VLD2LNdAsm_8 : NEONDataTypeAsmPseudoInst<"vld2${p}", ".8", "$list, $addr", + (ins VecListTwoDByteIndexed:$list, addrmode6align16:$addr, + pred:$p)>; +def VLD2LNdAsm_16 : NEONDataTypeAsmPseudoInst<"vld2${p}", ".16", "$list, $addr", + (ins VecListTwoDHWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VLD2LNdAsm_32 : NEONDataTypeAsmPseudoInst<"vld2${p}", ".32", "$list, $addr", + (ins VecListTwoDWordIndexed:$list, addrmode6align64:$addr, pred:$p)>; +def VLD2LNqAsm_16 : NEONDataTypeAsmPseudoInst<"vld2${p}", ".16", "$list, $addr", + (ins VecListTwoQHWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VLD2LNqAsm_32 : NEONDataTypeAsmPseudoInst<"vld2${p}", ".32", "$list, $addr", + (ins VecListTwoQWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; + +def VLD2LNdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".8", "$list, $addr!", + (ins VecListTwoDByteIndexed:$list, addrmode6align16:$addr, + pred:$p)>; +def VLD2LNdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".16", "$list, $addr!", + (ins VecListTwoDHWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VLD2LNdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".32", "$list, $addr!", + (ins VecListTwoDWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VLD2LNqWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".16", "$list, $addr!", + (ins VecListTwoQHWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VLD2LNqWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".32", "$list, $addr!", + (ins VecListTwoQWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VLD2LNdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".8", "$list, $addr, $Rm", + (ins VecListTwoDByteIndexed:$list, addrmode6align16:$addr, + rGPR:$Rm, pred:$p)>; +def VLD2LNdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".16", "$list, $addr, $Rm", + (ins VecListTwoDHWordIndexed:$list, addrmode6align32:$addr, + rGPR:$Rm, pred:$p)>; +def VLD2LNdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".32", "$list, $addr, $Rm", + (ins VecListTwoDWordIndexed:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD2LNqWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".16", "$list, $addr, $Rm", + (ins VecListTwoQHWordIndexed:$list, addrmode6align32:$addr, + rGPR:$Rm, pred:$p)>; +def VLD2LNqWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld2${p}", ".32", "$list, $addr, $Rm", + (ins VecListTwoQWordIndexed:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; + + +// VST2 single-lane pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VST2LNdAsm_8 : NEONDataTypeAsmPseudoInst<"vst2${p}", ".8", "$list, $addr", + (ins VecListTwoDByteIndexed:$list, addrmode6align16:$addr, + pred:$p)>; +def VST2LNdAsm_16 : NEONDataTypeAsmPseudoInst<"vst2${p}", ".16", "$list, $addr", + (ins VecListTwoDHWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VST2LNdAsm_32 : NEONDataTypeAsmPseudoInst<"vst2${p}", ".32", "$list, $addr", + (ins VecListTwoDWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VST2LNqAsm_16 : NEONDataTypeAsmPseudoInst<"vst2${p}", ".16", "$list, $addr", + (ins VecListTwoQHWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VST2LNqAsm_32 : NEONDataTypeAsmPseudoInst<"vst2${p}", ".32", "$list, $addr", + (ins VecListTwoQWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; + +def VST2LNdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".8", "$list, $addr!", + (ins VecListTwoDByteIndexed:$list, addrmode6align16:$addr, + pred:$p)>; +def VST2LNdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".16", "$list, $addr!", + (ins VecListTwoDHWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VST2LNdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".32", "$list, $addr!", + (ins VecListTwoDWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VST2LNqWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".16", "$list, $addr!", + (ins VecListTwoQHWordIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VST2LNqWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".32", "$list, $addr!", + (ins VecListTwoQWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VST2LNdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".8", "$list, $addr, $Rm", + (ins VecListTwoDByteIndexed:$list, addrmode6align16:$addr, + rGPR:$Rm, pred:$p)>; +def VST2LNdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".16","$list, $addr, $Rm", + (ins VecListTwoDHWordIndexed:$list, addrmode6align32:$addr, + rGPR:$Rm, pred:$p)>; +def VST2LNdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".32", "$list, $addr, $Rm", + (ins VecListTwoDWordIndexed:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VST2LNqWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".16","$list, $addr, $Rm", + (ins VecListTwoQHWordIndexed:$list, addrmode6align32:$addr, + rGPR:$Rm, pred:$p)>; +def VST2LNqWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst2${p}", ".32", "$list, $addr, $Rm", + (ins VecListTwoQWordIndexed:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; + +// VLD3 all-lanes pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VLD3DUPdAsm_8 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPdAsm_16: NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPdAsm_32: NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPqAsm_8 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPqAsm_16: NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPqAsm_32: NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; + +def VLD3DUPdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr!", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr!", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr!", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPqWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr!", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPqWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr!", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPqWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr!", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + pred:$p)>; +def VLD3DUPdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr, $Rm", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3DUPdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3DUPdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeDAllLanes:$list, addrmode6dupalignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3DUPqWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr, $Rm", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3DUPqWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3DUPqWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeQAllLanes:$list, addrmode6dupalignNone:$addr, + rGPR:$Rm, pred:$p)>; + + +// VLD3 single-lane pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VLD3LNdAsm_8 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr", + (ins VecListThreeDByteIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNdAsm_16 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr", + (ins VecListThreeDHWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNdAsm_32 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr", + (ins VecListThreeDWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNqAsm_16 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr", + (ins VecListThreeQHWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNqAsm_32 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr", + (ins VecListThreeQWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; + +def VLD3LNdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr!", + (ins VecListThreeDByteIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr!", + (ins VecListThreeDHWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr!", + (ins VecListThreeDWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNqWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr!", + (ins VecListThreeQHWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNqWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr!", + (ins VecListThreeQWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VLD3LNdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr, $Rm", + (ins VecListThreeDByteIndexed:$list, addrmode6alignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3LNdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeDHWordIndexed:$list, + addrmode6alignNone:$addr, rGPR:$Rm, pred:$p)>; +def VLD3LNdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeDWordIndexed:$list, addrmode6alignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3LNqWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeQHWordIndexed:$list, + addrmode6alignNone:$addr, rGPR:$Rm, pred:$p)>; +def VLD3LNqWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeQWordIndexed:$list, addrmode6alignNone:$addr, + rGPR:$Rm, pred:$p)>; + +// VLD3 multiple structure pseudo-instructions. These need special handling for +// the vector operands that the normal instructions don't yet model. +// FIXME: Remove these when the register classes and instructions are updated. +def VLD3dAsm_8 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3dAsm_16 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3dAsm_32 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3qAsm_8 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3qAsm_16 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3qAsm_32 : NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; + +def VLD3dWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr!", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3dWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr!", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3dWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr!", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3qWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr!", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3qWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr!", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3qWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr!", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VLD3dWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr, $Rm", + (ins VecListThreeD:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3dWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeD:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3dWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeD:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3qWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".8", "$list, $addr, $Rm", + (ins VecListThreeQ:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3qWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeQ:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD3qWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeQ:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; + +// VST3 single-lane pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VST3LNdAsm_8 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr", + (ins VecListThreeDByteIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNdAsm_16 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr", + (ins VecListThreeDHWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNdAsm_32 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr", + (ins VecListThreeDWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNqAsm_16 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr", + (ins VecListThreeQHWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNqAsm_32 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr", + (ins VecListThreeQWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; + +def VST3LNdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr!", + (ins VecListThreeDByteIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr!", + (ins VecListThreeDHWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr!", + (ins VecListThreeDWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNqWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr!", + (ins VecListThreeQHWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNqWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr!", + (ins VecListThreeQWordIndexed:$list, addrmode6alignNone:$addr, + pred:$p)>; +def VST3LNdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr, $Rm", + (ins VecListThreeDByteIndexed:$list, addrmode6alignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VST3LNdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeDHWordIndexed:$list, + addrmode6alignNone:$addr, rGPR:$Rm, pred:$p)>; +def VST3LNdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeDWordIndexed:$list, addrmode6alignNone:$addr, + rGPR:$Rm, pred:$p)>; +def VST3LNqWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeQHWordIndexed:$list, + addrmode6alignNone:$addr, rGPR:$Rm, pred:$p)>; +def VST3LNqWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeQWordIndexed:$list, addrmode6alignNone:$addr, + rGPR:$Rm, pred:$p)>; + + +// VST3 multiple structure pseudo-instructions. These need special handling for +// the vector operands that the normal instructions don't yet model. +// FIXME: Remove these when the register classes and instructions are updated. +def VST3dAsm_8 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VST3dAsm_16 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VST3dAsm_32 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VST3qAsm_8 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VST3qAsm_16 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VST3qAsm_32 : NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; + +def VST3dWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr!", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VST3dWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr!", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VST3dWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr!", + (ins VecListThreeD:$list, addrmode6align64:$addr, pred:$p)>; +def VST3qWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr!", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VST3qWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr!", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VST3qWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr!", + (ins VecListThreeQ:$list, addrmode6align64:$addr, pred:$p)>; +def VST3dWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr, $Rm", + (ins VecListThreeD:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VST3dWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeD:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VST3dWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeD:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VST3qWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".8", "$list, $addr, $Rm", + (ins VecListThreeQ:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VST3qWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".16", "$list, $addr, $Rm", + (ins VecListThreeQ:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VST3qWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst3${p}", ".32", "$list, $addr, $Rm", + (ins VecListThreeQ:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; + +// VLD4 all-lanes pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VLD4DUPdAsm_8 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr", + (ins VecListFourDAllLanes:$list, addrmode6dupalign32:$addr, + pred:$p)>; +def VLD4DUPdAsm_16: NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr", + (ins VecListFourDAllLanes:$list, addrmode6dupalign64:$addr, + pred:$p)>; +def VLD4DUPdAsm_32: NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr", + (ins VecListFourDAllLanes:$list, addrmode6dupalign64or128:$addr, + pred:$p)>; +def VLD4DUPqAsm_8 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr", + (ins VecListFourQAllLanes:$list, addrmode6dupalign32:$addr, + pred:$p)>; +def VLD4DUPqAsm_16: NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr", + (ins VecListFourQAllLanes:$list, addrmode6dupalign64:$addr, + pred:$p)>; +def VLD4DUPqAsm_32: NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr", + (ins VecListFourQAllLanes:$list, addrmode6dupalign64or128:$addr, + pred:$p)>; + +def VLD4DUPdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr!", + (ins VecListFourDAllLanes:$list, addrmode6dupalign32:$addr, + pred:$p)>; +def VLD4DUPdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr!", + (ins VecListFourDAllLanes:$list, addrmode6dupalign64:$addr, + pred:$p)>; +def VLD4DUPdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr!", + (ins VecListFourDAllLanes:$list, addrmode6dupalign64or128:$addr, + pred:$p)>; +def VLD4DUPqWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr!", + (ins VecListFourQAllLanes:$list, addrmode6dupalign32:$addr, + pred:$p)>; +def VLD4DUPqWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr!", + (ins VecListFourQAllLanes:$list, addrmode6dupalign64:$addr, + pred:$p)>; +def VLD4DUPqWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr!", + (ins VecListFourQAllLanes:$list, addrmode6dupalign64or128:$addr, + pred:$p)>; +def VLD4DUPdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr, $Rm", + (ins VecListFourDAllLanes:$list, addrmode6dupalign32:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4DUPdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourDAllLanes:$list, addrmode6dupalign64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4DUPdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourDAllLanes:$list, + addrmode6dupalign64or128:$addr, rGPR:$Rm, pred:$p)>; +def VLD4DUPqWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr, $Rm", + (ins VecListFourQAllLanes:$list, addrmode6dupalign32:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4DUPqWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourQAllLanes:$list, addrmode6dupalign64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4DUPqWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourQAllLanes:$list, + addrmode6dupalign64or128:$addr, rGPR:$Rm, pred:$p)>; + + +// VLD4 single-lane pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VLD4LNdAsm_8 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr", + (ins VecListFourDByteIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VLD4LNdAsm_16 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr", + (ins VecListFourDHWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VLD4LNdAsm_32 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr", + (ins VecListFourDWordIndexed:$list, addrmode6align64or128:$addr, + pred:$p)>; +def VLD4LNqAsm_16 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr", + (ins VecListFourQHWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VLD4LNqAsm_32 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr", + (ins VecListFourQWordIndexed:$list, addrmode6align64or128:$addr, + pred:$p)>; + +def VLD4LNdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr!", + (ins VecListFourDByteIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VLD4LNdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr!", + (ins VecListFourDHWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VLD4LNdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr!", + (ins VecListFourDWordIndexed:$list, addrmode6align64or128:$addr, + pred:$p)>; +def VLD4LNqWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr!", + (ins VecListFourQHWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VLD4LNqWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr!", + (ins VecListFourQWordIndexed:$list, addrmode6align64or128:$addr, + pred:$p)>; +def VLD4LNdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr, $Rm", + (ins VecListFourDByteIndexed:$list, addrmode6align32:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4LNdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourDHWordIndexed:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4LNdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourDWordIndexed:$list, + addrmode6align64or128:$addr, rGPR:$Rm, pred:$p)>; +def VLD4LNqWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourQHWordIndexed:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4LNqWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourQWordIndexed:$list, + addrmode6align64or128:$addr, rGPR:$Rm, pred:$p)>; + + + +// VLD4 multiple structure pseudo-instructions. These need special handling for +// the vector operands that the normal instructions don't yet model. +// FIXME: Remove these when the register classes and instructions are updated. +def VLD4dAsm_8 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4dAsm_16 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4dAsm_32 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4qAsm_8 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4qAsm_16 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4qAsm_32 : NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; + +def VLD4dWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr!", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4dWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr!", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4dWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr!", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4qWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr!", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4qWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr!", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4qWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr!", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VLD4dWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr, $Rm", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4dWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4dWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4qWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".8", "$list, $addr, $Rm", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4qWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VLD4qWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vld4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; + +// VST4 single-lane pseudo-instructions. These need special handling for +// the lane index that an InstAlias can't handle, so we use these instead. +def VST4LNdAsm_8 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr", + (ins VecListFourDByteIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VST4LNdAsm_16 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr", + (ins VecListFourDHWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VST4LNdAsm_32 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr", + (ins VecListFourDWordIndexed:$list, addrmode6align64or128:$addr, + pred:$p)>; +def VST4LNqAsm_16 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr", + (ins VecListFourQHWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VST4LNqAsm_32 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr", + (ins VecListFourQWordIndexed:$list, addrmode6align64or128:$addr, + pred:$p)>; + +def VST4LNdWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr!", + (ins VecListFourDByteIndexed:$list, addrmode6align32:$addr, + pred:$p)>; +def VST4LNdWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr!", + (ins VecListFourDHWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VST4LNdWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr!", + (ins VecListFourDWordIndexed:$list, addrmode6align64or128:$addr, + pred:$p)>; +def VST4LNqWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr!", + (ins VecListFourQHWordIndexed:$list, addrmode6align64:$addr, + pred:$p)>; +def VST4LNqWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr!", + (ins VecListFourQWordIndexed:$list, addrmode6align64or128:$addr, + pred:$p)>; +def VST4LNdWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr, $Rm", + (ins VecListFourDByteIndexed:$list, addrmode6align32:$addr, + rGPR:$Rm, pred:$p)>; +def VST4LNdWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourDHWordIndexed:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VST4LNdWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourDWordIndexed:$list, + addrmode6align64or128:$addr, rGPR:$Rm, pred:$p)>; +def VST4LNqWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourQHWordIndexed:$list, addrmode6align64:$addr, + rGPR:$Rm, pred:$p)>; +def VST4LNqWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourQWordIndexed:$list, + addrmode6align64or128:$addr, rGPR:$Rm, pred:$p)>; + + +// VST4 multiple structure pseudo-instructions. These need special handling for +// the vector operands that the normal instructions don't yet model. +// FIXME: Remove these when the register classes and instructions are updated. +def VST4dAsm_8 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4dAsm_16 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4dAsm_32 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4qAsm_8 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4qAsm_16 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4qAsm_32 : NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; + +def VST4dWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr!", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4dWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr!", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4dWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr!", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4qWB_fixed_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr!", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4qWB_fixed_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr!", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4qWB_fixed_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr!", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + pred:$p)>; +def VST4dWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr, $Rm", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VST4dWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VST4dWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourD:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VST4qWB_register_Asm_8 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".8", "$list, $addr, $Rm", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VST4qWB_register_Asm_16 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".16", "$list, $addr, $Rm", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; +def VST4qWB_register_Asm_32 : + NEONDataTypeAsmPseudoInst<"vst4${p}", ".32", "$list, $addr, $Rm", + (ins VecListFourQ:$list, addrmode6align64or128or256:$addr, + rGPR:$Rm, pred:$p)>; + +// VMOV/VMVN takes an optional datatype suffix +defm : NEONDTAnyInstAlias<"vmov${p}", "$Vd, $Vm", + (VORRd DPR:$Vd, DPR:$Vm, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vmov${p}", "$Vd, $Vm", + (VORRq QPR:$Vd, QPR:$Vm, QPR:$Vm, pred:$p)>; + +defm : NEONDTAnyInstAlias<"vmvn${p}", "$Vd, $Vm", + (VMVNd DPR:$Vd, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vmvn${p}", "$Vd, $Vm", + (VMVNq QPR:$Vd, QPR:$Vm, pred:$p)>; + +// VCLT (register) is an assembler alias for VCGT w/ the operands reversed. +// D-register versions. +def : NEONInstAlias<"vcle${p}.s8 $Dd, $Dn, $Dm", + (VCGEsv8i8 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.s16 $Dd, $Dn, $Dm", + (VCGEsv4i16 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.s32 $Dd, $Dn, $Dm", + (VCGEsv2i32 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.u8 $Dd, $Dn, $Dm", + (VCGEuv8i8 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.u16 $Dd, $Dn, $Dm", + (VCGEuv4i16 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.u32 $Dd, $Dn, $Dm", + (VCGEuv2i32 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.f32 $Dd, $Dn, $Dm", + (VCGEfd DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +let Predicates = [HasNEON, HasFullFP16] in +def : NEONInstAlias<"vcle${p}.f16 $Dd, $Dn, $Dm", + (VCGEhd DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +// Q-register versions. +def : NEONInstAlias<"vcle${p}.s8 $Qd, $Qn, $Qm", + (VCGEsv16i8 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.s16 $Qd, $Qn, $Qm", + (VCGEsv8i16 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.s32 $Qd, $Qn, $Qm", + (VCGEsv4i32 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.u8 $Qd, $Qn, $Qm", + (VCGEuv16i8 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.u16 $Qd, $Qn, $Qm", + (VCGEuv8i16 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.u32 $Qd, $Qn, $Qm", + (VCGEuv4i32 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vcle${p}.f32 $Qd, $Qn, $Qm", + (VCGEfq QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +let Predicates = [HasNEON, HasFullFP16] in +def : NEONInstAlias<"vcle${p}.f16 $Qd, $Qn, $Qm", + (VCGEhq QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; + +// VCLT (register) is an assembler alias for VCGT w/ the operands reversed. +// D-register versions. +def : NEONInstAlias<"vclt${p}.s8 $Dd, $Dn, $Dm", + (VCGTsv8i8 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.s16 $Dd, $Dn, $Dm", + (VCGTsv4i16 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.s32 $Dd, $Dn, $Dm", + (VCGTsv2i32 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.u8 $Dd, $Dn, $Dm", + (VCGTuv8i8 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.u16 $Dd, $Dn, $Dm", + (VCGTuv4i16 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.u32 $Dd, $Dn, $Dm", + (VCGTuv2i32 DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.f32 $Dd, $Dn, $Dm", + (VCGTfd DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +let Predicates = [HasNEON, HasFullFP16] in +def : NEONInstAlias<"vclt${p}.f16 $Dd, $Dn, $Dm", + (VCGThd DPR:$Dd, DPR:$Dm, DPR:$Dn, pred:$p)>; +// Q-register versions. +def : NEONInstAlias<"vclt${p}.s8 $Qd, $Qn, $Qm", + (VCGTsv16i8 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.s16 $Qd, $Qn, $Qm", + (VCGTsv8i16 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.s32 $Qd, $Qn, $Qm", + (VCGTsv4i32 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.u8 $Qd, $Qn, $Qm", + (VCGTuv16i8 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.u16 $Qd, $Qn, $Qm", + (VCGTuv8i16 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.u32 $Qd, $Qn, $Qm", + (VCGTuv4i32 QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +def : NEONInstAlias<"vclt${p}.f32 $Qd, $Qn, $Qm", + (VCGTfq QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; +let Predicates = [HasNEON, HasFullFP16] in +def : NEONInstAlias<"vclt${p}.f16 $Qd, $Qn, $Qm", + (VCGThq QPR:$Qd, QPR:$Qm, QPR:$Qn, pred:$p)>; + +// VSWP allows, but does not require, a type suffix. +defm : NEONDTAnyInstAlias<"vswp${p}", "$Vd, $Vm", + (VSWPd DPR:$Vd, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vswp${p}", "$Vd, $Vm", + (VSWPq QPR:$Vd, QPR:$Vm, pred:$p)>; + +// VBIF, VBIT, and VBSL allow, but do not require, a type suffix. +defm : NEONDTAnyInstAlias<"vbif${p}", "$Vd, $Vn, $Vm", + (VBIFd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vbit${p}", "$Vd, $Vn, $Vm", + (VBITd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vbsl${p}", "$Vd, $Vn, $Vm", + (VBSLd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vbif${p}", "$Vd, $Vn, $Vm", + (VBIFq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vbit${p}", "$Vd, $Vn, $Vm", + (VBITq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>; +defm : NEONDTAnyInstAlias<"vbsl${p}", "$Vd, $Vn, $Vm", + (VBSLq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>; + +// "vmov Rd, #-imm" can be handled via "vmvn". +def : NEONInstAlias<"vmov${p}.i32 $Vd, $imm", + (VMVNv2i32 DPR:$Vd, nImmVMOVI32Neg:$imm, pred:$p)>; +def : NEONInstAlias<"vmov${p}.i32 $Vd, $imm", + (VMVNv4i32 QPR:$Vd, nImmVMOVI32Neg:$imm, pred:$p)>; +def : NEONInstAlias<"vmvn${p}.i32 $Vd, $imm", + (VMOVv2i32 DPR:$Vd, nImmVMOVI32Neg:$imm, pred:$p)>; +def : NEONInstAlias<"vmvn${p}.i32 $Vd, $imm", + (VMOVv4i32 QPR:$Vd, nImmVMOVI32Neg:$imm, pred:$p)>; + +// 'gas' compatibility aliases for quad-word instructions. Strictly speaking, +// these should restrict to just the Q register variants, but the register +// classes are enough to match correctly regardless, so we keep it simple +// and just use MnemonicAlias. +def : NEONMnemonicAlias<"vbicq", "vbic">; +def : NEONMnemonicAlias<"vandq", "vand">; +def : NEONMnemonicAlias<"veorq", "veor">; +def : NEONMnemonicAlias<"vorrq", "vorr">; + +def : NEONMnemonicAlias<"vmovq", "vmov">; +def : NEONMnemonicAlias<"vmvnq", "vmvn">; +// Explicit versions for floating point so that the FPImm variants get +// handled early. The parser gets confused otherwise. +def : NEONMnemonicAlias<"vmovq.f32", "vmov.f32">; +def : NEONMnemonicAlias<"vmovq.f64", "vmov.f64">; + +def : NEONMnemonicAlias<"vaddq", "vadd">; +def : NEONMnemonicAlias<"vsubq", "vsub">; + +def : NEONMnemonicAlias<"vminq", "vmin">; +def : NEONMnemonicAlias<"vmaxq", "vmax">; + +def : NEONMnemonicAlias<"vmulq", "vmul">; + +def : NEONMnemonicAlias<"vabsq", "vabs">; + +def : NEONMnemonicAlias<"vshlq", "vshl">; +def : NEONMnemonicAlias<"vshrq", "vshr">; + +def : NEONMnemonicAlias<"vcvtq", "vcvt">; + +def : NEONMnemonicAlias<"vcleq", "vcle">; +def : NEONMnemonicAlias<"vceqq", "vceq">; + +def : NEONMnemonicAlias<"vzipq", "vzip">; +def : NEONMnemonicAlias<"vswpq", "vswp">; + +def : NEONMnemonicAlias<"vrecpeq.f32", "vrecpe.f32">; +def : NEONMnemonicAlias<"vrecpeq.u32", "vrecpe.u32">; + + +// Alias for loading floating point immediates that aren't representable +// using the vmov.f32 encoding but the bitpattern is representable using +// the .i32 encoding. +def : NEONInstAlias<"vmov${p}.f32 $Vd, $imm", + (VMOVv4i32 QPR:$Vd, nImmVMOVI32:$imm, pred:$p)>; +def : NEONInstAlias<"vmov${p}.f32 $Vd, $imm", + (VMOVv2i32 DPR:$Vd, nImmVMOVI32:$imm, pred:$p)>; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMInstrThumb.td b/capstone/suite/synctools/tablegen/ARM/ARMInstrThumb.td new file mode 100644 index 000000000..88aab47a7 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMInstrThumb.td @@ -0,0 +1,1707 @@ +//===-- ARMInstrThumb.td - Thumb support for ARM -----------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file describes the Thumb instruction set. +// +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// Thumb specific DAG Nodes. +// + +def imm_sr_XFORM: SDNodeXForm<imm, [{ + unsigned Imm = N->getZExtValue(); + return CurDAG->getTargetConstant((Imm == 32 ? 0 : Imm), SDLoc(N), MVT::i32); +}]>; +def ThumbSRImmAsmOperand: ImmAsmOperand<1,32> { let Name = "ImmThumbSR"; } +def imm_sr : Operand<i32>, PatLeaf<(imm), [{ + uint64_t Imm = N->getZExtValue(); + return Imm > 0 && Imm <= 32; +}], imm_sr_XFORM> { + let PrintMethod = "printThumbSRImm"; + let ParserMatchClass = ThumbSRImmAsmOperand; +} + +def imm0_7_neg : PatLeaf<(i32 imm), [{ + return (uint32_t)-N->getZExtValue() < 8; +}], imm_neg_XFORM>; + +def ThumbModImmNeg1_7AsmOperand : AsmOperandClass { let Name = "ThumbModImmNeg1_7"; } +def mod_imm1_7_neg : Operand<i32>, PatLeaf<(imm), [{ + unsigned Value = -(unsigned)N->getZExtValue(); + return 0 < Value && Value < 8; + }], imm_neg_XFORM> { + let ParserMatchClass = ThumbModImmNeg1_7AsmOperand; +} + +def ThumbModImmNeg8_255AsmOperand : AsmOperandClass { let Name = "ThumbModImmNeg8_255"; } +def mod_imm8_255_neg : Operand<i32>, PatLeaf<(imm), [{ + unsigned Value = -(unsigned)N->getZExtValue(); + return 7 < Value && Value < 256; + }], imm_neg_XFORM> { + let ParserMatchClass = ThumbModImmNeg8_255AsmOperand; +} + + +def imm0_255_comp : PatLeaf<(i32 imm), [{ + return ~((uint32_t)N->getZExtValue()) < 256; +}]>; + +def imm8_255_neg : PatLeaf<(i32 imm), [{ + unsigned Val = -N->getZExtValue(); + return Val >= 8 && Val < 256; +}], imm_neg_XFORM>; + +// Break imm's up into two pieces: an immediate + a left shift. This uses +// thumb_immshifted to match and thumb_immshifted_val and thumb_immshifted_shamt +// to get the val/shift pieces. +def thumb_immshifted : PatLeaf<(imm), [{ + return ARM_AM::isThumbImmShiftedVal((unsigned)N->getZExtValue()); +}]>; + +def thumb_immshifted_val : SDNodeXForm<imm, [{ + unsigned V = ARM_AM::getThumbImmNonShiftedVal((unsigned)N->getZExtValue()); + return CurDAG->getTargetConstant(V, SDLoc(N), MVT::i32); +}]>; + +def thumb_immshifted_shamt : SDNodeXForm<imm, [{ + unsigned V = ARM_AM::getThumbImmValShift((unsigned)N->getZExtValue()); + return CurDAG->getTargetConstant(V, SDLoc(N), MVT::i32); +}]>; + +def imm256_510 : ImmLeaf<i32, [{ + return Imm >= 256 && Imm < 511; +}]>; + +def thumb_imm256_510_addend : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant(N->getZExtValue() - 255, SDLoc(N), MVT::i32); +}]>; + +// Scaled 4 immediate. +def t_imm0_1020s4_asmoperand: AsmOperandClass { let Name = "Imm0_1020s4"; } +def t_imm0_1020s4 : Operand<i32> { + let PrintMethod = "printThumbS4ImmOperand"; + let ParserMatchClass = t_imm0_1020s4_asmoperand; + let OperandType = "OPERAND_IMMEDIATE"; +} + +def t_imm0_508s4_asmoperand: AsmOperandClass { let Name = "Imm0_508s4"; } +def t_imm0_508s4 : Operand<i32> { + let PrintMethod = "printThumbS4ImmOperand"; + let ParserMatchClass = t_imm0_508s4_asmoperand; + let OperandType = "OPERAND_IMMEDIATE"; +} +// Alias use only, so no printer is necessary. +def t_imm0_508s4_neg_asmoperand: AsmOperandClass { let Name = "Imm0_508s4Neg"; } +def t_imm0_508s4_neg : Operand<i32> { + let ParserMatchClass = t_imm0_508s4_neg_asmoperand; + let OperandType = "OPERAND_IMMEDIATE"; +} + +// Define Thumb specific addressing modes. + +// unsigned 8-bit, 2-scaled memory offset +class OperandUnsignedOffset_b8s2 : AsmOperandClass { + let Name = "UnsignedOffset_b8s2"; + let PredicateMethod = "isUnsignedOffset<8, 2>"; +} + +def UnsignedOffset_b8s2 : OperandUnsignedOffset_b8s2; + +// thumb style PC relative operand. signed, 8 bits magnitude, +// two bits shift. can be represented as either [pc, #imm], #imm, +// or relocatable expression... +def ThumbMemPC : AsmOperandClass { + let Name = "ThumbMemPC"; +} + +let OperandType = "OPERAND_PCREL" in { +def t_brtarget : Operand<OtherVT> { + let EncoderMethod = "getThumbBRTargetOpValue"; + let DecoderMethod = "DecodeThumbBROperand"; +} + +// ADR instruction labels. +def t_adrlabel : Operand<i32> { + let EncoderMethod = "getThumbAdrLabelOpValue"; + let PrintMethod = "printAdrLabelOperand<2>"; + let ParserMatchClass = UnsignedOffset_b8s2; +} + + +def thumb_br_target : Operand<OtherVT> { + let ParserMatchClass = ThumbBranchTarget; + let EncoderMethod = "getThumbBranchTargetOpValue"; + let OperandType = "OPERAND_PCREL"; +} + +def thumb_bl_target : Operand<i32> { + let ParserMatchClass = ThumbBranchTarget; + let EncoderMethod = "getThumbBLTargetOpValue"; + let DecoderMethod = "DecodeThumbBLTargetOperand"; +} + +// Target for BLX *from* thumb mode. +def thumb_blx_target : Operand<i32> { + let ParserMatchClass = ARMBranchTarget; + let EncoderMethod = "getThumbBLXTargetOpValue"; + let DecoderMethod = "DecodeThumbBLXOffset"; +} + +def thumb_bcc_target : Operand<OtherVT> { + let ParserMatchClass = ThumbBranchTarget; + let EncoderMethod = "getThumbBCCTargetOpValue"; + let DecoderMethod = "DecodeThumbBCCTargetOperand"; +} + +def thumb_cb_target : Operand<OtherVT> { + let ParserMatchClass = ThumbBranchTarget; + let EncoderMethod = "getThumbCBTargetOpValue"; + let DecoderMethod = "DecodeThumbCmpBROperand"; +} + +// t_addrmode_pc := <label> => pc + imm8 * 4 +// +def t_addrmode_pc : MemOperand { + let EncoderMethod = "getAddrModePCOpValue"; + let DecoderMethod = "DecodeThumbAddrModePC"; + let PrintMethod = "printThumbLdrLabelOperand"; + let ParserMatchClass = ThumbMemPC; +} +} + +// t_addrmode_rr := reg + reg +// +def t_addrmode_rr_asm_operand : AsmOperandClass { let Name = "MemThumbRR"; } +def t_addrmode_rr : MemOperand, + ComplexPattern<i32, 2, "SelectThumbAddrModeRR", []> { + let EncoderMethod = "getThumbAddrModeRegRegOpValue"; + let PrintMethod = "printThumbAddrModeRROperand"; + let DecoderMethod = "DecodeThumbAddrModeRR"; + let ParserMatchClass = t_addrmode_rr_asm_operand; + let MIOperandInfo = (ops tGPR:$base, tGPR:$offsreg); +} + +// t_addrmode_rrs := reg + reg +// +// We use separate scaled versions because the Select* functions need +// to explicitly check for a matching constant and return false here so that +// the reg+imm forms will match instead. This is a horrible way to do that, +// as it forces tight coupling between the methods, but it's how selectiondag +// currently works. +def t_addrmode_rrs1 : MemOperand, + ComplexPattern<i32, 2, "SelectThumbAddrModeRI5S1", []> { + let EncoderMethod = "getThumbAddrModeRegRegOpValue"; + let PrintMethod = "printThumbAddrModeRROperand"; + let DecoderMethod = "DecodeThumbAddrModeRR"; + let ParserMatchClass = t_addrmode_rr_asm_operand; + let MIOperandInfo = (ops tGPR:$base, tGPR:$offsreg); +} +def t_addrmode_rrs2 : MemOperand, + ComplexPattern<i32, 2, "SelectThumbAddrModeRI5S2", []> { + let EncoderMethod = "getThumbAddrModeRegRegOpValue"; + let DecoderMethod = "DecodeThumbAddrModeRR"; + let PrintMethod = "printThumbAddrModeRROperand"; + let ParserMatchClass = t_addrmode_rr_asm_operand; + let MIOperandInfo = (ops tGPR:$base, tGPR:$offsreg); +} +def t_addrmode_rrs4 : MemOperand, + ComplexPattern<i32, 2, "SelectThumbAddrModeRI5S4", []> { + let EncoderMethod = "getThumbAddrModeRegRegOpValue"; + let DecoderMethod = "DecodeThumbAddrModeRR"; + let PrintMethod = "printThumbAddrModeRROperand"; + let ParserMatchClass = t_addrmode_rr_asm_operand; + let MIOperandInfo = (ops tGPR:$base, tGPR:$offsreg); +} + +// t_addrmode_is4 := reg + imm5 * 4 +// +def t_addrmode_is4_asm_operand : AsmOperandClass { let Name = "MemThumbRIs4"; } +def t_addrmode_is4 : MemOperand, + ComplexPattern<i32, 2, "SelectThumbAddrModeImm5S4", []> { + let EncoderMethod = "getAddrModeISOpValue"; + let DecoderMethod = "DecodeThumbAddrModeIS"; + let PrintMethod = "printThumbAddrModeImm5S4Operand"; + let ParserMatchClass = t_addrmode_is4_asm_operand; + let MIOperandInfo = (ops tGPR:$base, i32imm:$offsimm); +} + +// t_addrmode_is2 := reg + imm5 * 2 +// +def t_addrmode_is2_asm_operand : AsmOperandClass { let Name = "MemThumbRIs2"; } +def t_addrmode_is2 : MemOperand, + ComplexPattern<i32, 2, "SelectThumbAddrModeImm5S2", []> { + let EncoderMethod = "getAddrModeISOpValue"; + let DecoderMethod = "DecodeThumbAddrModeIS"; + let PrintMethod = "printThumbAddrModeImm5S2Operand"; + let ParserMatchClass = t_addrmode_is2_asm_operand; + let MIOperandInfo = (ops tGPR:$base, i32imm:$offsimm); +} + +// t_addrmode_is1 := reg + imm5 +// +def t_addrmode_is1_asm_operand : AsmOperandClass { let Name = "MemThumbRIs1"; } +def t_addrmode_is1 : MemOperand, + ComplexPattern<i32, 2, "SelectThumbAddrModeImm5S1", []> { + let EncoderMethod = "getAddrModeISOpValue"; + let DecoderMethod = "DecodeThumbAddrModeIS"; + let PrintMethod = "printThumbAddrModeImm5S1Operand"; + let ParserMatchClass = t_addrmode_is1_asm_operand; + let MIOperandInfo = (ops tGPR:$base, i32imm:$offsimm); +} + +// t_addrmode_sp := sp + imm8 * 4 +// +// FIXME: This really shouldn't have an explicit SP operand at all. It should +// be implicit, just like in the instruction encoding itself. +def t_addrmode_sp_asm_operand : AsmOperandClass { let Name = "MemThumbSPI"; } +def t_addrmode_sp : MemOperand, + ComplexPattern<i32, 2, "SelectThumbAddrModeSP", []> { + let EncoderMethod = "getAddrModeThumbSPOpValue"; + let DecoderMethod = "DecodeThumbAddrModeSP"; + let PrintMethod = "printThumbAddrModeSPOperand"; + let ParserMatchClass = t_addrmode_sp_asm_operand; + let MIOperandInfo = (ops GPR:$base, i32imm:$offsimm); +} + +// Inspects parent to determine whether an or instruction can be implemented as +// an add (i.e. whether we know overflow won't occur in the add). +def AddLikeOrOp : ComplexPattern<i32, 1, "SelectAddLikeOr", [], + [SDNPWantParent]>; + +// Pattern to exclude immediates from matching +def non_imm32 : PatLeaf<(i32 GPR), [{ return !isa<ConstantSDNode>(N); }]>; + +//===----------------------------------------------------------------------===// +// Miscellaneous Instructions. +// + +// FIXME: Marking these as hasSideEffects is necessary to prevent machine DCE +// from removing one half of the matched pairs. That breaks PEI, which assumes +// these will always be in pairs, and asserts if it finds otherwise. Better way? +let Defs = [SP], Uses = [SP], hasSideEffects = 1 in { +def tADJCALLSTACKUP : + PseudoInst<(outs), (ins i32imm:$amt1, i32imm:$amt2), NoItinerary, + [(ARMcallseq_end imm:$amt1, imm:$amt2)]>, + Requires<[IsThumb, IsThumb1Only]>; + +def tADJCALLSTACKDOWN : + PseudoInst<(outs), (ins i32imm:$amt, i32imm:$amt2), NoItinerary, + [(ARMcallseq_start imm:$amt, imm:$amt2)]>, + Requires<[IsThumb, IsThumb1Only]>; +} + +class T1SystemEncoding<bits<8> opc> + : T1Encoding<0b101111> { + let Inst{9-8} = 0b11; + let Inst{7-0} = opc; +} + +def tHINT : T1pI<(outs), (ins imm0_15:$imm), NoItinerary, "hint", "\t$imm", + [(int_arm_hint imm0_15:$imm)]>, + T1SystemEncoding<0x00>, + Requires<[IsThumb, HasV6M]> { + bits<4> imm; + let Inst{7-4} = imm; +} + +// Note: When EmitPriority == 1, the alias will be used for printing +class tHintAlias<string Asm, dag Result, bit EmitPriority = 0> : tInstAlias<Asm, Result, EmitPriority> { + let Predicates = [IsThumb, HasV6M]; +} + +def : tHintAlias<"nop$p", (tHINT 0, pred:$p), 1>; // A8.6.110 +def : tHintAlias<"yield$p", (tHINT 1, pred:$p), 1>; // A8.6.410 +def : tHintAlias<"wfe$p", (tHINT 2, pred:$p), 1>; // A8.6.408 +def : tHintAlias<"wfi$p", (tHINT 3, pred:$p), 1>; // A8.6.409 +def : tHintAlias<"sev$p", (tHINT 4, pred:$p), 1>; // A8.6.157 +def : tInstAlias<"sevl$p", (tHINT 5, pred:$p), 1> { + let Predicates = [IsThumb2, HasV8]; +} + +// The imm operand $val can be used by a debugger to store more information +// about the breakpoint. +def tBKPT : T1I<(outs), (ins imm0_255:$val), NoItinerary, "bkpt\t$val", + []>, + T1Encoding<0b101111> { + let Inst{9-8} = 0b10; + // A8.6.22 + bits<8> val; + let Inst{7-0} = val; +} +// default immediate for breakpoint mnemonic +def : InstAlias<"bkpt", (tBKPT 0), 0>, Requires<[IsThumb]>; + +def tHLT : T1I<(outs), (ins imm0_63:$val), NoItinerary, "hlt\t$val", + []>, T1Encoding<0b101110>, Requires<[IsThumb, HasV8]> { + let Inst{9-6} = 0b1010; + bits<6> val; + let Inst{5-0} = val; +} + +def tSETEND : T1I<(outs), (ins setend_op:$end), NoItinerary, "setend\t$end", + []>, T1Encoding<0b101101>, Requires<[IsThumb, IsNotMClass]>, Deprecated<HasV8Ops> { + bits<1> end; + // A8.6.156 + let Inst{9-5} = 0b10010; + let Inst{4} = 1; + let Inst{3} = end; + let Inst{2-0} = 0b000; +} + +// Change Processor State is a system instruction -- for disassembly only. +def tCPS : T1I<(outs), (ins imod_op:$imod, iflags_op:$iflags), + NoItinerary, "cps$imod $iflags", []>, + T1Misc<0b0110011> { + // A8.6.38 & B6.1.1 + bit imod; + bits<3> iflags; + + let Inst{4} = imod; + let Inst{3} = 0; + let Inst{2-0} = iflags; + let DecoderMethod = "DecodeThumbCPS"; +} + +// For both thumb1 and thumb2. +let isNotDuplicable = 1, isCodeGenOnly = 1 in +def tPICADD : TIt<(outs GPR:$dst), (ins GPR:$lhs, pclabel:$cp), IIC_iALUr, "", + [(set GPR:$dst, (ARMpic_add GPR:$lhs, imm:$cp))]>, + T1Special<{0,0,?,?}>, Sched<[WriteALU]> { + // A8.6.6 + bits<3> dst; + let Inst{6-3} = 0b1111; // Rm = pc + let Inst{2-0} = dst; +} + +// ADD <Rd>, sp, #<imm8> +// FIXME: This should not be marked as having side effects, and it should be +// rematerializable. Clearing the side effect bit causes miscompilations, +// probably because the instruction can be moved around. +def tADDrSPi : T1pI<(outs tGPR:$dst), (ins GPRsp:$sp, t_imm0_1020s4:$imm), + IIC_iALUi, "add", "\t$dst, $sp, $imm", []>, + T1Encoding<{1,0,1,0,1,?}>, Sched<[WriteALU]> { + // A6.2 & A8.6.8 + bits<3> dst; + bits<8> imm; + let Inst{10-8} = dst; + let Inst{7-0} = imm; + let DecoderMethod = "DecodeThumbAddSpecialReg"; +} + +// Thumb1 frame lowering is rather fragile, we hope to be able to use +// tADDrSPi, but we may need to insert a sequence that clobbers CPSR. +def tADDframe : PseudoInst<(outs tGPR:$dst), (ins i32imm:$base, i32imm:$offset), + NoItinerary, []>, + Requires<[IsThumb, IsThumb1Only]> { + let Defs = [CPSR]; +} + +// ADD sp, sp, #<imm7> +def tADDspi : T1pIt<(outs GPRsp:$Rdn), (ins GPRsp:$Rn, t_imm0_508s4:$imm), + IIC_iALUi, "add", "\t$Rdn, $imm", []>, + T1Misc<{0,0,0,0,0,?,?}>, Sched<[WriteALU]> { + // A6.2.5 & A8.6.8 + bits<7> imm; + let Inst{6-0} = imm; + let DecoderMethod = "DecodeThumbAddSPImm"; +} + +// SUB sp, sp, #<imm7> +// FIXME: The encoding and the ASM string don't match up. +def tSUBspi : T1pIt<(outs GPRsp:$Rdn), (ins GPRsp:$Rn, t_imm0_508s4:$imm), + IIC_iALUi, "sub", "\t$Rdn, $imm", []>, + T1Misc<{0,0,0,0,1,?,?}>, Sched<[WriteALU]> { + // A6.2.5 & A8.6.214 + bits<7> imm; + let Inst{6-0} = imm; + let DecoderMethod = "DecodeThumbAddSPImm"; +} + +def : tInstSubst<"add${p} sp, $imm", + (tSUBspi SP, t_imm0_508s4_neg:$imm, pred:$p)>; +def : tInstSubst<"add${p} sp, sp, $imm", + (tSUBspi SP, t_imm0_508s4_neg:$imm, pred:$p)>; + +// Can optionally specify SP as a three operand instruction. +def : tInstAlias<"add${p} sp, sp, $imm", + (tADDspi SP, t_imm0_508s4:$imm, pred:$p)>; +def : tInstAlias<"sub${p} sp, sp, $imm", + (tSUBspi SP, t_imm0_508s4:$imm, pred:$p)>; + +// ADD <Rm>, sp +def tADDrSP : T1pI<(outs GPR:$Rdn), (ins GPRsp:$sp, GPR:$Rn), IIC_iALUr, + "add", "\t$Rdn, $sp, $Rn", []>, + T1Special<{0,0,?,?}>, Sched<[WriteALU]> { + // A8.6.9 Encoding T1 + bits<4> Rdn; + let Inst{7} = Rdn{3}; + let Inst{6-3} = 0b1101; + let Inst{2-0} = Rdn{2-0}; + let DecoderMethod = "DecodeThumbAddSPReg"; +} + +// ADD sp, <Rm> +def tADDspr : T1pIt<(outs GPRsp:$Rdn), (ins GPRsp:$Rn, GPR:$Rm), IIC_iALUr, + "add", "\t$Rdn, $Rm", []>, + T1Special<{0,0,?,?}>, Sched<[WriteALU]> { + // A8.6.9 Encoding T2 + bits<4> Rm; + let Inst{7} = 1; + let Inst{6-3} = Rm; + let Inst{2-0} = 0b101; + let DecoderMethod = "DecodeThumbAddSPReg"; +} + +//===----------------------------------------------------------------------===// +// Control Flow Instructions. +// + +// Indirect branches +let isBranch = 1, isTerminator = 1, isBarrier = 1, isIndirectBranch = 1 in { + def tBX : TI<(outs), (ins GPR:$Rm, pred:$p), IIC_Br, "bx${p}\t$Rm", []>, + T1Special<{1,1,0,?}>, Sched<[WriteBr]> { + // A6.2.3 & A8.6.25 + bits<4> Rm; + let Inst{6-3} = Rm; + let Inst{2-0} = 0b000; + let Unpredictable{2-0} = 0b111; + } + def tBXNS : TI<(outs), (ins GPR:$Rm, pred:$p), IIC_Br, "bxns${p}\t$Rm", []>, + Requires<[IsThumb, Has8MSecExt]>, + T1Special<{1,1,0,?}>, Sched<[WriteBr]> { + bits<4> Rm; + let Inst{6-3} = Rm; + let Inst{2-0} = 0b100; + let Unpredictable{1-0} = 0b11; + } +} + +let isReturn = 1, isTerminator = 1, isBarrier = 1 in { + def tBX_RET : tPseudoExpand<(outs), (ins pred:$p), 2, IIC_Br, + [(ARMretflag)], (tBX LR, pred:$p)>, Sched<[WriteBr]>; + + // Alternative return instruction used by vararg functions. + def tBX_RET_vararg : tPseudoExpand<(outs), (ins tGPR:$Rm, pred:$p), + 2, IIC_Br, [], + (tBX GPR:$Rm, pred:$p)>, Sched<[WriteBr]>; +} + +// All calls clobber the non-callee saved registers. SP is marked as a use to +// prevent stack-pointer assignments that appear immediately before calls from +// potentially appearing dead. +let isCall = 1, + Defs = [LR], Uses = [SP] in { + // Also used for Thumb2 + def tBL : TIx2<0b11110, 0b11, 1, + (outs), (ins pred:$p, thumb_bl_target:$func), IIC_Br, + "bl${p}\t$func", + [(ARMcall tglobaladdr:$func)]>, + Requires<[IsThumb]>, Sched<[WriteBrL]> { + bits<24> func; + let Inst{26} = func{23}; + let Inst{25-16} = func{20-11}; + let Inst{13} = func{22}; + let Inst{11} = func{21}; + let Inst{10-0} = func{10-0}; + } + + // ARMv5T and above, also used for Thumb2 + def tBLXi : TIx2<0b11110, 0b11, 0, + (outs), (ins pred:$p, thumb_blx_target:$func), IIC_Br, + "blx${p}\t$func", []>, + Requires<[IsThumb, HasV5T, IsNotMClass]>, Sched<[WriteBrL]> { + bits<24> func; + let Inst{26} = func{23}; + let Inst{25-16} = func{20-11}; + let Inst{13} = func{22}; + let Inst{11} = func{21}; + let Inst{10-1} = func{10-1}; + let Inst{0} = 0; // func{0} is assumed zero + } + + // Also used for Thumb2 + def tBLXr : TI<(outs), (ins pred:$p, GPR:$func), IIC_Br, + "blx${p}\t$func", + [(ARMcall GPR:$func)]>, + Requires<[IsThumb, HasV5T]>, + T1Special<{1,1,1,?}>, Sched<[WriteBrL]> { // A6.2.3 & A8.6.24; + bits<4> func; + let Inst{6-3} = func; + let Inst{2-0} = 0b000; + } + + // ARMv8-M Security Extensions + def tBLXNSr : TI<(outs), (ins pred:$p, GPRnopc:$func), IIC_Br, + "blxns${p}\t$func", []>, + Requires<[IsThumb, Has8MSecExt]>, + T1Special<{1,1,1,?}>, Sched<[WriteBrL]> { + bits<4> func; + let Inst{6-3} = func; + let Inst{2-0} = 0b100; + let Unpredictable{1-0} = 0b11; + } + + // ARMv4T + def tBX_CALL : tPseudoInst<(outs), (ins tGPR:$func), + 4, IIC_Br, + [(ARMcall_nolink tGPR:$func)]>, + Requires<[IsThumb, IsThumb1Only]>, Sched<[WriteBr]>; +} + +let isBranch = 1, isTerminator = 1, isBarrier = 1 in { + let isPredicable = 1 in + def tB : T1pI<(outs), (ins t_brtarget:$target), IIC_Br, + "b", "\t$target", [(br bb:$target)]>, + T1Encoding<{1,1,1,0,0,?}>, Sched<[WriteBr]> { + bits<11> target; + let Inst{10-0} = target; + let AsmMatchConverter = "cvtThumbBranches"; + } + + // Far jump + // Just a pseudo for a tBL instruction. Needed to let regalloc know about + // the clobber of LR. + let Defs = [LR] in + def tBfar : tPseudoExpand<(outs), (ins thumb_bl_target:$target, pred:$p), + 4, IIC_Br, [], + (tBL pred:$p, thumb_bl_target:$target)>, + Sched<[WriteBrTbl]>; + + def tBR_JTr : tPseudoInst<(outs), + (ins tGPR:$target, i32imm:$jt), + 0, IIC_Br, + [(ARMbrjt tGPR:$target, tjumptable:$jt)]>, + Sched<[WriteBrTbl]> { + let Size = 2; + list<Predicate> Predicates = [IsThumb, IsThumb1Only]; + } +} + +// FIXME: should be able to write a pattern for ARMBrcond, but can't use +// a two-value operand where a dag node expects two operands. :( +let isBranch = 1, isTerminator = 1 in + def tBcc : T1I<(outs), (ins thumb_bcc_target:$target, pred:$p), IIC_Br, + "b${p}\t$target", + [/*(ARMbrcond bb:$target, imm:$cc)*/]>, + T1BranchCond<{1,1,0,1}>, Sched<[WriteBr]> { + bits<4> p; + bits<8> target; + let Inst{11-8} = p; + let Inst{7-0} = target; + let AsmMatchConverter = "cvtThumbBranches"; +} + + +// Tail calls +let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1 in { + // IOS versions. + let Uses = [SP] in { + def tTAILJMPr : tPseudoExpand<(outs), (ins tcGPR:$dst), + 4, IIC_Br, [], + (tBX GPR:$dst, (ops 14, zero_reg))>, + Requires<[IsThumb]>, Sched<[WriteBr]>; + } + // tTAILJMPd: MachO version uses a Thumb2 branch (no Thumb1 tail calls + // on MachO), so it's in ARMInstrThumb2.td. + // Non-MachO version: + let Uses = [SP] in { + def tTAILJMPdND : tPseudoExpand<(outs), + (ins t_brtarget:$dst, pred:$p), + 4, IIC_Br, [], + (tB t_brtarget:$dst, pred:$p)>, + Requires<[IsThumb, IsNotMachO]>, Sched<[WriteBr]>; + } +} + + +// A8.6.218 Supervisor Call (Software Interrupt) +// A8.6.16 B: Encoding T1 +// If Inst{11-8} == 0b1111 then SEE SVC +let isCall = 1, Uses = [SP] in +def tSVC : T1pI<(outs), (ins imm0_255:$imm), IIC_Br, + "svc", "\t$imm", []>, Encoding16, Sched<[WriteBr]> { + bits<8> imm; + let Inst{15-12} = 0b1101; + let Inst{11-8} = 0b1111; + let Inst{7-0} = imm; +} + +// The assembler uses 0xDEFE for a trap instruction. +let isBarrier = 1, isTerminator = 1 in +def tTRAP : TI<(outs), (ins), IIC_Br, + "trap", [(trap)]>, Encoding16, Sched<[WriteBr]> { + let Inst = 0xdefe; +} + +//===----------------------------------------------------------------------===// +// Load Store Instructions. +// + +// PC-relative loads need to be matched first as constant pool accesses need to +// always be PC-relative. We do this using AddedComplexity, as the pattern is +// simpler than the patterns of the other load instructions. +let canFoldAsLoad = 1, isReMaterializable = 1, AddedComplexity = 10 in +def tLDRpci : T1pIs<(outs tGPR:$Rt), (ins t_addrmode_pc:$addr), IIC_iLoad_i, + "ldr", "\t$Rt, $addr", + [(set tGPR:$Rt, (load (ARMWrapper tconstpool:$addr)))]>, + T1Encoding<{0,1,0,0,1,?}> { + // A6.2 & A8.6.59 + bits<3> Rt; + bits<8> addr; + let Inst{10-8} = Rt; + let Inst{7-0} = addr; +} + +// SP-relative loads should be matched before standard immediate-offset loads as +// it means we avoid having to move SP to another register. +let canFoldAsLoad = 1 in +def tLDRspi : T1pIs<(outs tGPR:$Rt), (ins t_addrmode_sp:$addr), IIC_iLoad_i, + "ldr", "\t$Rt, $addr", + [(set tGPR:$Rt, (load t_addrmode_sp:$addr))]>, + T1LdStSP<{1,?,?}> { + bits<3> Rt; + bits<8> addr; + let Inst{10-8} = Rt; + let Inst{7-0} = addr; +} + +// Loads: reg/reg and reg/imm5 +let canFoldAsLoad = 1, isReMaterializable = 1 in +multiclass thumb_ld_rr_ri_enc<bits<3> reg_opc, bits<4> imm_opc, + Operand AddrMode_r, Operand AddrMode_i, + AddrMode am, InstrItinClass itin_r, + InstrItinClass itin_i, string asm, + PatFrag opnode> { + // Immediate-offset loads should be matched before register-offset loads as + // when the offset is a constant it's simpler to first check if it fits in the + // immediate offset field then fall back to register-offset if it doesn't. + def i : // reg/imm5 + T1pILdStEncodeImm<imm_opc, 1 /* Load */, + (outs tGPR:$Rt), (ins AddrMode_i:$addr), + am, itin_i, asm, "\t$Rt, $addr", + [(set tGPR:$Rt, (opnode AddrMode_i:$addr))]>; + // Register-offset loads are matched last. + def r : // reg/reg + T1pILdStEncode<reg_opc, + (outs tGPR:$Rt), (ins AddrMode_r:$addr), + am, itin_r, asm, "\t$Rt, $addr", + [(set tGPR:$Rt, (opnode AddrMode_r:$addr))]>; +} +// Stores: reg/reg and reg/imm5 +multiclass thumb_st_rr_ri_enc<bits<3> reg_opc, bits<4> imm_opc, + Operand AddrMode_r, Operand AddrMode_i, + AddrMode am, InstrItinClass itin_r, + InstrItinClass itin_i, string asm, + PatFrag opnode> { + def i : // reg/imm5 + T1pILdStEncodeImm<imm_opc, 0 /* Store */, + (outs), (ins tGPR:$Rt, AddrMode_i:$addr), + am, itin_i, asm, "\t$Rt, $addr", + [(opnode tGPR:$Rt, AddrMode_i:$addr)]>; + def r : // reg/reg + T1pILdStEncode<reg_opc, + (outs), (ins tGPR:$Rt, AddrMode_r:$addr), + am, itin_r, asm, "\t$Rt, $addr", + [(opnode tGPR:$Rt, AddrMode_r:$addr)]>; +} + +// A8.6.57 & A8.6.60 +defm tLDR : thumb_ld_rr_ri_enc<0b100, 0b0110, t_addrmode_rr, + t_addrmode_is4, AddrModeT1_4, + IIC_iLoad_r, IIC_iLoad_i, "ldr", + load>; + +// A8.6.64 & A8.6.61 +defm tLDRB : thumb_ld_rr_ri_enc<0b110, 0b0111, t_addrmode_rr, + t_addrmode_is1, AddrModeT1_1, + IIC_iLoad_bh_r, IIC_iLoad_bh_i, "ldrb", + zextloadi8>; + +// A8.6.76 & A8.6.73 +defm tLDRH : thumb_ld_rr_ri_enc<0b101, 0b1000, t_addrmode_rr, + t_addrmode_is2, AddrModeT1_2, + IIC_iLoad_bh_r, IIC_iLoad_bh_i, "ldrh", + zextloadi16>; + +let AddedComplexity = 10 in +def tLDRSB : // A8.6.80 + T1pILdStEncode<0b011, (outs tGPR:$Rt), (ins t_addrmode_rr:$addr), + AddrModeT1_1, IIC_iLoad_bh_r, + "ldrsb", "\t$Rt, $addr", + [(set tGPR:$Rt, (sextloadi8 t_addrmode_rr:$addr))]>; + +let AddedComplexity = 10 in +def tLDRSH : // A8.6.84 + T1pILdStEncode<0b111, (outs tGPR:$Rt), (ins t_addrmode_rr:$addr), + AddrModeT1_2, IIC_iLoad_bh_r, + "ldrsh", "\t$Rt, $addr", + [(set tGPR:$Rt, (sextloadi16 t_addrmode_rr:$addr))]>; + + +def tSTRspi : T1pIs<(outs), (ins tGPR:$Rt, t_addrmode_sp:$addr), IIC_iStore_i, + "str", "\t$Rt, $addr", + [(store tGPR:$Rt, t_addrmode_sp:$addr)]>, + T1LdStSP<{0,?,?}> { + bits<3> Rt; + bits<8> addr; + let Inst{10-8} = Rt; + let Inst{7-0} = addr; +} + +// A8.6.194 & A8.6.192 +defm tSTR : thumb_st_rr_ri_enc<0b000, 0b0110, t_addrmode_rr, + t_addrmode_is4, AddrModeT1_4, + IIC_iStore_r, IIC_iStore_i, "str", + store>; + +// A8.6.197 & A8.6.195 +defm tSTRB : thumb_st_rr_ri_enc<0b010, 0b0111, t_addrmode_rr, + t_addrmode_is1, AddrModeT1_1, + IIC_iStore_bh_r, IIC_iStore_bh_i, "strb", + truncstorei8>; + +// A8.6.207 & A8.6.205 +defm tSTRH : thumb_st_rr_ri_enc<0b001, 0b1000, t_addrmode_rr, + t_addrmode_is2, AddrModeT1_2, + IIC_iStore_bh_r, IIC_iStore_bh_i, "strh", + truncstorei16>; + + +//===----------------------------------------------------------------------===// +// Load / store multiple Instructions. +// + +// These require base address to be written back or one of the loaded regs. +let hasSideEffects = 0 in { + +let mayLoad = 1, hasExtraDefRegAllocReq = 1 in +def tLDMIA : T1I<(outs), (ins tGPR:$Rn, pred:$p, reglist:$regs, variable_ops), + IIC_iLoad_m, "ldm${p}\t$Rn, $regs", []>, T1Encoding<{1,1,0,0,1,?}> { + bits<3> Rn; + bits<8> regs; + let Inst{10-8} = Rn; + let Inst{7-0} = regs; +} + +// Writeback version is just a pseudo, as there's no encoding difference. +// Writeback happens iff the base register is not in the destination register +// list. +let mayLoad = 1, hasExtraDefRegAllocReq = 1 in +def tLDMIA_UPD : + InstTemplate<AddrModeNone, 0, IndexModeNone, Pseudo, GenericDomain, + "$Rn = $wb", IIC_iLoad_mu>, + PseudoInstExpansion<(tLDMIA tGPR:$Rn, pred:$p, reglist:$regs)> { + let Size = 2; + let OutOperandList = (outs GPR:$wb); + let InOperandList = (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops); + let Pattern = []; + let isCodeGenOnly = 1; + let isPseudo = 1; + list<Predicate> Predicates = [IsThumb]; +} + +// There is no non-writeback version of STM for Thumb. +let mayStore = 1, hasExtraSrcRegAllocReq = 1 in +def tSTMIA_UPD : Thumb1I<(outs GPR:$wb), + (ins tGPR:$Rn, pred:$p, reglist:$regs, variable_ops), + AddrModeNone, 2, IIC_iStore_mu, + "stm${p}\t$Rn!, $regs", "$Rn = $wb", []>, + T1Encoding<{1,1,0,0,0,?}> { + bits<3> Rn; + bits<8> regs; + let Inst{10-8} = Rn; + let Inst{7-0} = regs; +} + +} // hasSideEffects + +def : InstAlias<"ldm${p} $Rn!, $regs", + (tLDMIA tGPR:$Rn, pred:$p, reglist:$regs), 0>, + Requires<[IsThumb, IsThumb1Only]>; + +let mayLoad = 1, Uses = [SP], Defs = [SP], hasExtraDefRegAllocReq = 1 in +def tPOP : T1I<(outs), (ins pred:$p, reglist:$regs, variable_ops), + IIC_iPop, + "pop${p}\t$regs", []>, + T1Misc<{1,1,0,?,?,?,?}> { + bits<16> regs; + let Inst{8} = regs{15}; + let Inst{7-0} = regs{7-0}; +} + +let mayStore = 1, Uses = [SP], Defs = [SP], hasExtraSrcRegAllocReq = 1 in +def tPUSH : T1I<(outs), (ins pred:$p, reglist:$regs, variable_ops), + IIC_iStore_m, + "push${p}\t$regs", []>, + T1Misc<{0,1,0,?,?,?,?}> { + bits<16> regs; + let Inst{8} = regs{14}; + let Inst{7-0} = regs{7-0}; +} + +//===----------------------------------------------------------------------===// +// Arithmetic Instructions. +// + +// Helper classes for encoding T1pI patterns: +class T1pIDPEncode<bits<4> opA, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T1pI<oops, iops, itin, opc, asm, pattern>, + T1DataProcessing<opA> { + bits<3> Rm; + bits<3> Rn; + let Inst{5-3} = Rm; + let Inst{2-0} = Rn; +} +class T1pIMiscEncode<bits<7> opA, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T1pI<oops, iops, itin, opc, asm, pattern>, + T1Misc<opA> { + bits<3> Rm; + bits<3> Rd; + let Inst{5-3} = Rm; + let Inst{2-0} = Rd; +} + +// Helper classes for encoding T1sI patterns: +class T1sIDPEncode<bits<4> opA, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T1sI<oops, iops, itin, opc, asm, pattern>, + T1DataProcessing<opA> { + bits<3> Rd; + bits<3> Rn; + let Inst{5-3} = Rn; + let Inst{2-0} = Rd; +} +class T1sIGenEncode<bits<5> opA, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T1sI<oops, iops, itin, opc, asm, pattern>, + T1General<opA> { + bits<3> Rm; + bits<3> Rn; + bits<3> Rd; + let Inst{8-6} = Rm; + let Inst{5-3} = Rn; + let Inst{2-0} = Rd; +} +class T1sIGenEncodeImm<bits<5> opA, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T1sI<oops, iops, itin, opc, asm, pattern>, + T1General<opA> { + bits<3> Rd; + bits<3> Rm; + let Inst{5-3} = Rm; + let Inst{2-0} = Rd; +} + +// Helper classes for encoding T1sIt patterns: +class T1sItDPEncode<bits<4> opA, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T1sIt<oops, iops, itin, opc, asm, pattern>, + T1DataProcessing<opA> { + bits<3> Rdn; + bits<3> Rm; + let Inst{5-3} = Rm; + let Inst{2-0} = Rdn; +} +class T1sItGenEncodeImm<bits<5> opA, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T1sIt<oops, iops, itin, opc, asm, pattern>, + T1General<opA> { + bits<3> Rdn; + bits<8> imm8; + let Inst{10-8} = Rdn; + let Inst{7-0} = imm8; +} + +let isAdd = 1 in { + // Add with carry register + let isCommutable = 1, Uses = [CPSR] in + def tADC : // A8.6.2 + T1sItDPEncode<0b0101, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), IIC_iALUr, + "adc", "\t$Rdn, $Rm", + []>, Sched<[WriteALU]>; + + // Add immediate + def tADDi3 : // A8.6.4 T1 + T1sIGenEncodeImm<0b01110, (outs tGPR:$Rd), (ins tGPR:$Rm, imm0_7:$imm3), + IIC_iALUi, + "add", "\t$Rd, $Rm, $imm3", + [(set tGPR:$Rd, (add tGPR:$Rm, imm0_7:$imm3))]>, + Sched<[WriteALU]> { + bits<3> imm3; + let Inst{8-6} = imm3; + } + + def tADDi8 : // A8.6.4 T2 + T1sItGenEncodeImm<{1,1,0,?,?}, (outs tGPR:$Rdn), + (ins tGPR:$Rn, imm0_255:$imm8), IIC_iALUi, + "add", "\t$Rdn, $imm8", + [(set tGPR:$Rdn, (add tGPR:$Rn, imm8_255:$imm8))]>, + Sched<[WriteALU]>; + + // Add register + let isCommutable = 1 in + def tADDrr : // A8.6.6 T1 + T1sIGenEncode<0b01100, (outs tGPR:$Rd), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iALUr, + "add", "\t$Rd, $Rn, $Rm", + [(set tGPR:$Rd, (add tGPR:$Rn, tGPR:$Rm))]>, Sched<[WriteALU]>; + + /// Similar to the above except these set the 's' bit so the + /// instruction modifies the CPSR register. + /// + /// These opcodes will be converted to the real non-S opcodes by + /// AdjustInstrPostInstrSelection after giving then an optional CPSR operand. + let hasPostISelHook = 1, Defs = [CPSR] in { + let isCommutable = 1, Uses = [CPSR] in + def tADCS : tPseudoInst<(outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + 2, IIC_iALUr, + [(set tGPR:$Rdn, CPSR, (ARMadde tGPR:$Rn, tGPR:$Rm, + CPSR))]>, + Requires<[IsThumb1Only]>, + Sched<[WriteALU]>; + + def tADDSi3 : tPseudoInst<(outs tGPR:$Rd), (ins tGPR:$Rm, imm0_7:$imm3), + 2, IIC_iALUi, + [(set tGPR:$Rd, CPSR, (ARMaddc tGPR:$Rm, + imm0_7:$imm3))]>, + Requires<[IsThumb1Only]>, + Sched<[WriteALU]>; + + def tADDSi8 : tPseudoInst<(outs tGPR:$Rdn), (ins tGPR:$Rn, imm0_255:$imm8), + 2, IIC_iALUi, + [(set tGPR:$Rdn, CPSR, (ARMaddc tGPR:$Rn, + imm8_255:$imm8))]>, + Requires<[IsThumb1Only]>, + Sched<[WriteALU]>; + + let isCommutable = 1 in + def tADDSrr : tPseudoInst<(outs tGPR:$Rd), (ins tGPR:$Rn, tGPR:$Rm), + 2, IIC_iALUr, + [(set tGPR:$Rd, CPSR, (ARMaddc tGPR:$Rn, + tGPR:$Rm))]>, + Requires<[IsThumb1Only]>, + Sched<[WriteALU]>; + } + + let hasSideEffects = 0 in + def tADDhirr : T1pIt<(outs GPR:$Rdn), (ins GPR:$Rn, GPR:$Rm), IIC_iALUr, + "add", "\t$Rdn, $Rm", []>, + T1Special<{0,0,?,?}>, Sched<[WriteALU]> { + // A8.6.6 T2 + bits<4> Rdn; + bits<4> Rm; + let Inst{7} = Rdn{3}; + let Inst{6-3} = Rm; + let Inst{2-0} = Rdn{2-0}; + } +} + +// Thumb has more flexible short encodings for ADD than ORR, so use those where +// possible. +def : T1Pat<(or AddLikeOrOp:$Rn, imm0_7:$imm), (tADDi3 $Rn, imm0_7:$imm)>; + +def : T1Pat<(or AddLikeOrOp:$Rn, imm8_255:$imm), (tADDi8 $Rn, imm8_255:$imm)>; + +def : T1Pat<(or AddLikeOrOp:$Rn, tGPR:$Rm), (tADDrr $Rn, $Rm)>; + + +def : tInstAlias <"add${s}${p} $Rdn, $Rm", + (tADDrr tGPR:$Rdn,s_cc_out:$s, tGPR:$Rdn, tGPR:$Rm, pred:$p)>; + +def : tInstSubst<"sub${s}${p} $rd, $rn, $imm", + (tADDi3 tGPR:$rd, s_cc_out:$s, tGPR:$rn, mod_imm1_7_neg:$imm, pred:$p)>; +def : tInstSubst<"sub${s}${p} $rdn, $imm", + (tADDi8 tGPR:$rdn, s_cc_out:$s, mod_imm8_255_neg:$imm, pred:$p)>; + + +// AND register +let isCommutable = 1 in +def tAND : // A8.6.12 + T1sItDPEncode<0b0000, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iBITr, + "and", "\t$Rdn, $Rm", + [(set tGPR:$Rdn, (and tGPR:$Rn, tGPR:$Rm))]>, Sched<[WriteALU]>; + +// ASR immediate +def tASRri : // A8.6.14 + T1sIGenEncodeImm<{0,1,0,?,?}, (outs tGPR:$Rd), (ins tGPR:$Rm, imm_sr:$imm5), + IIC_iMOVsi, + "asr", "\t$Rd, $Rm, $imm5", + [(set tGPR:$Rd, (sra tGPR:$Rm, (i32 imm_sr:$imm5)))]>, + Sched<[WriteALU]> { + bits<5> imm5; + let Inst{10-6} = imm5; +} + +// ASR register +def tASRrr : // A8.6.15 + T1sItDPEncode<0b0100, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iMOVsr, + "asr", "\t$Rdn, $Rm", + [(set tGPR:$Rdn, (sra tGPR:$Rn, tGPR:$Rm))]>, Sched<[WriteALU]>; + +// BIC register +def tBIC : // A8.6.20 + T1sItDPEncode<0b1110, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iBITr, + "bic", "\t$Rdn, $Rm", + [(set tGPR:$Rdn, (and tGPR:$Rn, (not tGPR:$Rm)))]>, + Sched<[WriteALU]>; + +// CMN register +let isCompare = 1, Defs = [CPSR] in { +//FIXME: Disable CMN, as CCodes are backwards from compare expectations +// Compare-to-zero still works out, just not the relationals +//def tCMN : // A8.6.33 +// T1pIDPEncode<0b1011, (outs), (ins tGPR:$lhs, tGPR:$rhs), +// IIC_iCMPr, +// "cmn", "\t$lhs, $rhs", +// [(ARMcmp tGPR:$lhs, (ineg tGPR:$rhs))]>; + +def tCMNz : // A8.6.33 + T1pIDPEncode<0b1011, (outs), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iCMPr, + "cmn", "\t$Rn, $Rm", + [(ARMcmpZ tGPR:$Rn, (ineg tGPR:$Rm))]>, Sched<[WriteCMP]>; + +} // isCompare = 1, Defs = [CPSR] + +// CMP immediate +let isCompare = 1, Defs = [CPSR] in { +def tCMPi8 : T1pI<(outs), (ins tGPR:$Rn, imm0_255:$imm8), IIC_iCMPi, + "cmp", "\t$Rn, $imm8", + [(ARMcmp tGPR:$Rn, imm0_255:$imm8)]>, + T1General<{1,0,1,?,?}>, Sched<[WriteCMP]> { + // A8.6.35 + bits<3> Rn; + bits<8> imm8; + let Inst{10-8} = Rn; + let Inst{7-0} = imm8; +} + +// CMP register +def tCMPr : // A8.6.36 T1 + T1pIDPEncode<0b1010, (outs), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iCMPr, + "cmp", "\t$Rn, $Rm", + [(ARMcmp tGPR:$Rn, tGPR:$Rm)]>, Sched<[WriteCMP]>; + +def tCMPhir : T1pI<(outs), (ins GPR:$Rn, GPR:$Rm), IIC_iCMPr, + "cmp", "\t$Rn, $Rm", []>, + T1Special<{0,1,?,?}>, Sched<[WriteCMP]> { + // A8.6.36 T2 + bits<4> Rm; + bits<4> Rn; + let Inst{7} = Rn{3}; + let Inst{6-3} = Rm; + let Inst{2-0} = Rn{2-0}; +} +} // isCompare = 1, Defs = [CPSR] + + +// XOR register +let isCommutable = 1 in +def tEOR : // A8.6.45 + T1sItDPEncode<0b0001, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iBITr, + "eor", "\t$Rdn, $Rm", + [(set tGPR:$Rdn, (xor tGPR:$Rn, tGPR:$Rm))]>, Sched<[WriteALU]>; + +// LSL immediate +def tLSLri : // A8.6.88 + T1sIGenEncodeImm<{0,0,0,?,?}, (outs tGPR:$Rd), (ins tGPR:$Rm, imm0_31:$imm5), + IIC_iMOVsi, + "lsl", "\t$Rd, $Rm, $imm5", + [(set tGPR:$Rd, (shl tGPR:$Rm, (i32 imm:$imm5)))]>, + Sched<[WriteALU]> { + bits<5> imm5; + let Inst{10-6} = imm5; +} + +// LSL register +def tLSLrr : // A8.6.89 + T1sItDPEncode<0b0010, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iMOVsr, + "lsl", "\t$Rdn, $Rm", + [(set tGPR:$Rdn, (shl tGPR:$Rn, tGPR:$Rm))]>, Sched<[WriteALU]>; + +// LSR immediate +def tLSRri : // A8.6.90 + T1sIGenEncodeImm<{0,0,1,?,?}, (outs tGPR:$Rd), (ins tGPR:$Rm, imm_sr:$imm5), + IIC_iMOVsi, + "lsr", "\t$Rd, $Rm, $imm5", + [(set tGPR:$Rd, (srl tGPR:$Rm, (i32 imm_sr:$imm5)))]>, + Sched<[WriteALU]> { + bits<5> imm5; + let Inst{10-6} = imm5; +} + +// LSR register +def tLSRrr : // A8.6.91 + T1sItDPEncode<0b0011, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iMOVsr, + "lsr", "\t$Rdn, $Rm", + [(set tGPR:$Rdn, (srl tGPR:$Rn, tGPR:$Rm))]>, Sched<[WriteALU]>; + +// Move register +let isMoveImm = 1 in +def tMOVi8 : T1sI<(outs tGPR:$Rd), (ins imm0_255:$imm8), IIC_iMOVi, + "mov", "\t$Rd, $imm8", + [(set tGPR:$Rd, imm0_255:$imm8)]>, + T1General<{1,0,0,?,?}>, Sched<[WriteALU]> { + // A8.6.96 + bits<3> Rd; + bits<8> imm8; + let Inst{10-8} = Rd; + let Inst{7-0} = imm8; +} +// Because we have an explicit tMOVSr below, we need an alias to handle +// the immediate "movs" form here. Blech. +def : tInstAlias <"movs $Rdn, $imm", + (tMOVi8 tGPR:$Rdn, CPSR, imm0_255:$imm, 14, 0)>; + +// A7-73: MOV(2) - mov setting flag. + +let hasSideEffects = 0, isMoveReg = 1 in { +def tMOVr : Thumb1pI<(outs GPR:$Rd), (ins GPR:$Rm), AddrModeNone, + 2, IIC_iMOVr, + "mov", "\t$Rd, $Rm", "", []>, + T1Special<{1,0,?,?}>, Sched<[WriteALU]> { + // A8.6.97 + bits<4> Rd; + bits<4> Rm; + let Inst{7} = Rd{3}; + let Inst{6-3} = Rm; + let Inst{2-0} = Rd{2-0}; +} +let Defs = [CPSR] in +def tMOVSr : T1I<(outs tGPR:$Rd), (ins tGPR:$Rm), IIC_iMOVr, + "movs\t$Rd, $Rm", []>, Encoding16, Sched<[WriteALU]> { + // A8.6.97 + bits<3> Rd; + bits<3> Rm; + let Inst{15-6} = 0b0000000000; + let Inst{5-3} = Rm; + let Inst{2-0} = Rd; +} +} // hasSideEffects + +// Multiply register +let isCommutable = 1 in +def tMUL : // A8.6.105 T1 + Thumb1sI<(outs tGPR:$Rd), (ins tGPR:$Rn, tGPR:$Rm), AddrModeNone, 2, + IIC_iMUL32, "mul", "\t$Rd, $Rn, $Rm", "$Rm = $Rd", + [(set tGPR:$Rd, (mul tGPR:$Rn, tGPR:$Rm))]>, + T1DataProcessing<0b1101> { + bits<3> Rd; + bits<3> Rn; + let Inst{5-3} = Rn; + let Inst{2-0} = Rd; + let AsmMatchConverter = "cvtThumbMultiply"; +} + +def :tInstAlias<"mul${s}${p} $Rdm, $Rn", (tMUL tGPR:$Rdm, s_cc_out:$s, tGPR:$Rn, + pred:$p)>; + +// Move inverse register +def tMVN : // A8.6.107 + T1sIDPEncode<0b1111, (outs tGPR:$Rd), (ins tGPR:$Rn), IIC_iMVNr, + "mvn", "\t$Rd, $Rn", + [(set tGPR:$Rd, (not tGPR:$Rn))]>, Sched<[WriteALU]>; + +// Bitwise or register +let isCommutable = 1 in +def tORR : // A8.6.114 + T1sItDPEncode<0b1100, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iBITr, + "orr", "\t$Rdn, $Rm", + [(set tGPR:$Rdn, (or tGPR:$Rn, tGPR:$Rm))]>, Sched<[WriteALU]>; + +// Swaps +def tREV : // A8.6.134 + T1pIMiscEncode<{1,0,1,0,0,0,?}, (outs tGPR:$Rd), (ins tGPR:$Rm), + IIC_iUNAr, + "rev", "\t$Rd, $Rm", + [(set tGPR:$Rd, (bswap tGPR:$Rm))]>, + Requires<[IsThumb, IsThumb1Only, HasV6]>, Sched<[WriteALU]>; + +def tREV16 : // A8.6.135 + T1pIMiscEncode<{1,0,1,0,0,1,?}, (outs tGPR:$Rd), (ins tGPR:$Rm), + IIC_iUNAr, + "rev16", "\t$Rd, $Rm", + [(set tGPR:$Rd, (rotr (bswap tGPR:$Rm), (i32 16)))]>, + Requires<[IsThumb, IsThumb1Only, HasV6]>, Sched<[WriteALU]>; + +def tREVSH : // A8.6.136 + T1pIMiscEncode<{1,0,1,0,1,1,?}, (outs tGPR:$Rd), (ins tGPR:$Rm), + IIC_iUNAr, + "revsh", "\t$Rd, $Rm", + [(set tGPR:$Rd, (sra (bswap tGPR:$Rm), (i32 16)))]>, + Requires<[IsThumb, IsThumb1Only, HasV6]>, Sched<[WriteALU]>; + +// Rotate right register +def tROR : // A8.6.139 + T1sItDPEncode<0b0111, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iMOVsr, + "ror", "\t$Rdn, $Rm", + [(set tGPR:$Rdn, (rotr tGPR:$Rn, tGPR:$Rm))]>, + Sched<[WriteALU]>; + +// Negate register +def tRSB : // A8.6.141 + T1sIDPEncode<0b1001, (outs tGPR:$Rd), (ins tGPR:$Rn), + IIC_iALUi, + "rsb", "\t$Rd, $Rn, #0", + [(set tGPR:$Rd, (ineg tGPR:$Rn))]>, Sched<[WriteALU]>; + +// Subtract with carry register +let Uses = [CPSR] in +def tSBC : // A8.6.151 + T1sItDPEncode<0b0110, (outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iALUr, + "sbc", "\t$Rdn, $Rm", + []>, + Sched<[WriteALU]>; + +// Subtract immediate +def tSUBi3 : // A8.6.210 T1 + T1sIGenEncodeImm<0b01111, (outs tGPR:$Rd), (ins tGPR:$Rm, imm0_7:$imm3), + IIC_iALUi, + "sub", "\t$Rd, $Rm, $imm3", + [(set tGPR:$Rd, (add tGPR:$Rm, imm0_7_neg:$imm3))]>, + Sched<[WriteALU]> { + bits<3> imm3; + let Inst{8-6} = imm3; +} + +def tSUBi8 : // A8.6.210 T2 + T1sItGenEncodeImm<{1,1,1,?,?}, (outs tGPR:$Rdn), + (ins tGPR:$Rn, imm0_255:$imm8), IIC_iALUi, + "sub", "\t$Rdn, $imm8", + [(set tGPR:$Rdn, (add tGPR:$Rn, imm8_255_neg:$imm8))]>, + Sched<[WriteALU]>; + +def : tInstSubst<"add${s}${p} $rd, $rn, $imm", + (tSUBi3 tGPR:$rd, s_cc_out:$s, tGPR:$rn, mod_imm1_7_neg:$imm, pred:$p)>; + + +def : tInstSubst<"add${s}${p} $rdn, $imm", + (tSUBi8 tGPR:$rdn, s_cc_out:$s, mod_imm8_255_neg:$imm, pred:$p)>; + + +// Subtract register +def tSUBrr : // A8.6.212 + T1sIGenEncode<0b01101, (outs tGPR:$Rd), (ins tGPR:$Rn, tGPR:$Rm), + IIC_iALUr, + "sub", "\t$Rd, $Rn, $Rm", + [(set tGPR:$Rd, (sub tGPR:$Rn, tGPR:$Rm))]>, + Sched<[WriteALU]>; + +def : tInstAlias <"sub${s}${p} $Rdn, $Rm", + (tSUBrr tGPR:$Rdn,s_cc_out:$s, tGPR:$Rdn, tGPR:$Rm, pred:$p)>; + +/// Similar to the above except these set the 's' bit so the +/// instruction modifies the CPSR register. +/// +/// These opcodes will be converted to the real non-S opcodes by +/// AdjustInstrPostInstrSelection after giving then an optional CPSR operand. +let hasPostISelHook = 1, Defs = [CPSR] in { + let Uses = [CPSR] in + def tSBCS : tPseudoInst<(outs tGPR:$Rdn), (ins tGPR:$Rn, tGPR:$Rm), + 2, IIC_iALUr, + [(set tGPR:$Rdn, CPSR, (ARMsube tGPR:$Rn, tGPR:$Rm, + CPSR))]>, + Requires<[IsThumb1Only]>, + Sched<[WriteALU]>; + + def tSUBSi3 : tPseudoInst<(outs tGPR:$Rd), (ins tGPR:$Rm, imm0_7:$imm3), + 2, IIC_iALUi, + [(set tGPR:$Rd, CPSR, (ARMsubc tGPR:$Rm, + imm0_7:$imm3))]>, + Requires<[IsThumb1Only]>, + Sched<[WriteALU]>; + + def tSUBSi8 : tPseudoInst<(outs tGPR:$Rdn), (ins tGPR:$Rn, imm0_255:$imm8), + 2, IIC_iALUi, + [(set tGPR:$Rdn, CPSR, (ARMsubc tGPR:$Rn, + imm8_255:$imm8))]>, + Requires<[IsThumb1Only]>, + Sched<[WriteALU]>; + + def tSUBSrr : tPseudoInst<(outs tGPR:$Rd), (ins tGPR:$Rn, tGPR:$Rm), + 2, IIC_iALUr, + [(set tGPR:$Rd, CPSR, (ARMsubc tGPR:$Rn, + tGPR:$Rm))]>, + Requires<[IsThumb1Only]>, + Sched<[WriteALU]>; +} + +// Sign-extend byte +def tSXTB : // A8.6.222 + T1pIMiscEncode<{0,0,1,0,0,1,?}, (outs tGPR:$Rd), (ins tGPR:$Rm), + IIC_iUNAr, + "sxtb", "\t$Rd, $Rm", + [(set tGPR:$Rd, (sext_inreg tGPR:$Rm, i8))]>, + Requires<[IsThumb, IsThumb1Only, HasV6]>, + Sched<[WriteALU]>; + +// Sign-extend short +def tSXTH : // A8.6.224 + T1pIMiscEncode<{0,0,1,0,0,0,?}, (outs tGPR:$Rd), (ins tGPR:$Rm), + IIC_iUNAr, + "sxth", "\t$Rd, $Rm", + [(set tGPR:$Rd, (sext_inreg tGPR:$Rm, i16))]>, + Requires<[IsThumb, IsThumb1Only, HasV6]>, + Sched<[WriteALU]>; + +// Test +let isCompare = 1, isCommutable = 1, Defs = [CPSR] in +def tTST : // A8.6.230 + T1pIDPEncode<0b1000, (outs), (ins tGPR:$Rn, tGPR:$Rm), IIC_iTSTr, + "tst", "\t$Rn, $Rm", + [(ARMcmpZ (and_su tGPR:$Rn, tGPR:$Rm), 0)]>, + Sched<[WriteALU]>; + +// A8.8.247 UDF - Undefined (Encoding T1) +def tUDF : TI<(outs), (ins imm0_255:$imm8), IIC_Br, "udf\t$imm8", + [(int_arm_undefined imm0_255:$imm8)]>, Encoding16 { + bits<8> imm8; + let Inst{15-12} = 0b1101; + let Inst{11-8} = 0b1110; + let Inst{7-0} = imm8; +} + +def t__brkdiv0 : TI<(outs), (ins), IIC_Br, "__brkdiv0", + [(int_arm_undefined 249)]>, Encoding16, + Requires<[IsThumb, IsWindows]> { + let Inst = 0xdef9; + let isTerminator = 1; +} + +// Zero-extend byte +def tUXTB : // A8.6.262 + T1pIMiscEncode<{0,0,1,0,1,1,?}, (outs tGPR:$Rd), (ins tGPR:$Rm), + IIC_iUNAr, + "uxtb", "\t$Rd, $Rm", + [(set tGPR:$Rd, (and tGPR:$Rm, 0xFF))]>, + Requires<[IsThumb, IsThumb1Only, HasV6]>, + Sched<[WriteALU]>; + +// Zero-extend short +def tUXTH : // A8.6.264 + T1pIMiscEncode<{0,0,1,0,1,0,?}, (outs tGPR:$Rd), (ins tGPR:$Rm), + IIC_iUNAr, + "uxth", "\t$Rd, $Rm", + [(set tGPR:$Rd, (and tGPR:$Rm, 0xFFFF))]>, + Requires<[IsThumb, IsThumb1Only, HasV6]>, Sched<[WriteALU]>; + +// Conditional move tMOVCCr - Used to implement the Thumb SELECT_CC operation. +// Expanded after instruction selection into a branch sequence. +let usesCustomInserter = 1 in // Expanded after instruction selection. + def tMOVCCr_pseudo : + PseudoInst<(outs tGPR:$dst), (ins tGPR:$false, tGPR:$true, cmovpred:$p), + NoItinerary, + [(set tGPR:$dst, (ARMcmov tGPR:$false, tGPR:$true, cmovpred:$p))]>; + +// tLEApcrel - Load a pc-relative address into a register without offending the +// assembler. + +def tADR : T1I<(outs tGPR:$Rd), (ins t_adrlabel:$addr, pred:$p), + IIC_iALUi, "adr{$p}\t$Rd, $addr", []>, + T1Encoding<{1,0,1,0,0,?}>, Sched<[WriteALU]> { + bits<3> Rd; + bits<8> addr; + let Inst{10-8} = Rd; + let Inst{7-0} = addr; + let DecoderMethod = "DecodeThumbAddSpecialReg"; +} + +let hasSideEffects = 0, isReMaterializable = 1 in +def tLEApcrel : tPseudoInst<(outs tGPR:$Rd), (ins i32imm:$label, pred:$p), + 2, IIC_iALUi, []>, Sched<[WriteALU]>; + +let hasSideEffects = 1 in +def tLEApcrelJT : tPseudoInst<(outs tGPR:$Rd), + (ins i32imm:$label, pred:$p), + 2, IIC_iALUi, []>, Sched<[WriteALU]>; + +// Thumb-1 doesn't have the TBB or TBH instructions, but we can synthesize them +// and make use of the same compressed jump table format as Thumb-2. +let Size = 2, isBranch = 1, isTerminator = 1, isBarrier = 1, + isIndirectBranch = 1 in { +def tTBB_JT : tPseudoInst<(outs), + (ins tGPRwithpc:$base, tGPR:$index, i32imm:$jt, i32imm:$pclbl), 0, + IIC_Br, []>, Sched<[WriteBr]>; + +def tTBH_JT : tPseudoInst<(outs), + (ins tGPRwithpc:$base, tGPR:$index, i32imm:$jt, i32imm:$pclbl), 0, + IIC_Br, []>, Sched<[WriteBr]>; +} + +//===----------------------------------------------------------------------===// +// TLS Instructions +// + +// __aeabi_read_tp preserves the registers r1-r3. +// This is a pseudo inst so that we can get the encoding right, +// complete with fixup for the aeabi_read_tp function. +let isCall = 1, Defs = [R0, R12, LR, CPSR], Uses = [SP] in +def tTPsoft : tPseudoInst<(outs), (ins), 4, IIC_Br, + [(set R0, ARMthread_pointer)]>, + Sched<[WriteBr]>; + +//===----------------------------------------------------------------------===// +// SJLJ Exception handling intrinsics +// + +// eh_sjlj_setjmp() is an instruction sequence to store the return address and +// save #0 in R0 for the non-longjmp case. Since by its nature we may be coming +// from some other function to get here, and we're using the stack frame for the +// containing function to save/restore registers, we can't keep anything live in +// regs across the eh_sjlj_setjmp(), else it will almost certainly have been +// tromped upon when we get here from a longjmp(). We force everything out of +// registers except for our own input by listing the relevant registers in +// Defs. By doing so, we also cause the prologue/epilogue code to actively +// preserve all of the callee-saved resgisters, which is exactly what we want. +// $val is a scratch register for our use. +let Defs = [ R0, R1, R2, R3, R4, R5, R6, R7, R12, CPSR ], + hasSideEffects = 1, isBarrier = 1, isCodeGenOnly = 1, + usesCustomInserter = 1 in +def tInt_eh_sjlj_setjmp : ThumbXI<(outs),(ins tGPR:$src, tGPR:$val), + AddrModeNone, 0, NoItinerary, "","", + [(set R0, (ARMeh_sjlj_setjmp tGPR:$src, tGPR:$val))]>; + +// FIXME: Non-IOS version(s) +let isBarrier = 1, hasSideEffects = 1, isTerminator = 1, isCodeGenOnly = 1, + Defs = [ R7, LR, SP ] in +def tInt_eh_sjlj_longjmp : XI<(outs), (ins GPR:$src, GPR:$scratch), + AddrModeNone, 0, IndexModeNone, + Pseudo, NoItinerary, "", "", + [(ARMeh_sjlj_longjmp GPR:$src, GPR:$scratch)]>, + Requires<[IsThumb,IsNotWindows]>; + +let isBarrier = 1, hasSideEffects = 1, isTerminator = 1, isCodeGenOnly = 1, + Defs = [ R11, LR, SP ] in +def tInt_WIN_eh_sjlj_longjmp + : XI<(outs), (ins GPR:$src, GPR:$scratch), AddrModeNone, 0, IndexModeNone, + Pseudo, NoItinerary, "", "", [(ARMeh_sjlj_longjmp GPR:$src, GPR:$scratch)]>, + Requires<[IsThumb,IsWindows]>; + +//===----------------------------------------------------------------------===// +// Non-Instruction Patterns +// + +// Comparisons +def : T1Pat<(ARMcmpZ tGPR:$Rn, imm0_255:$imm8), + (tCMPi8 tGPR:$Rn, imm0_255:$imm8)>; +def : T1Pat<(ARMcmpZ tGPR:$Rn, tGPR:$Rm), + (tCMPr tGPR:$Rn, tGPR:$Rm)>; + +// Bswap 16 with load/store +def : T1Pat<(srl (bswap (extloadi16 t_addrmode_is2:$addr)), (i32 16)), + (tREV16 (tLDRHi t_addrmode_is2:$addr))>; +def : T1Pat<(srl (bswap (extloadi16 t_addrmode_rr:$addr)), (i32 16)), + (tREV16 (tLDRHr t_addrmode_rr:$addr))>; +def : T1Pat<(truncstorei16 (srl (bswap tGPR:$Rn), (i32 16)), + t_addrmode_is2:$addr), + (tSTRHi(tREV16 tGPR:$Rn), t_addrmode_is2:$addr)>; +def : T1Pat<(truncstorei16 (srl (bswap tGPR:$Rn), (i32 16)), + t_addrmode_rr:$addr), + (tSTRHr (tREV16 tGPR:$Rn), t_addrmode_rr:$addr)>; + +// ConstantPool +def : T1Pat<(ARMWrapper tconstpool :$dst), (tLEApcrel tconstpool :$dst)>; + +// GlobalAddress +def tLDRLIT_ga_pcrel : PseudoInst<(outs tGPR:$dst), (ins i32imm:$addr), + IIC_iLoadiALU, + [(set tGPR:$dst, + (ARMWrapperPIC tglobaladdr:$addr))]>, + Requires<[IsThumb, DontUseMovtInPic]>; + +def tLDRLIT_ga_abs : PseudoInst<(outs tGPR:$dst), (ins i32imm:$src), + IIC_iLoad_i, + [(set tGPR:$dst, + (ARMWrapper tglobaladdr:$src))]>, + Requires<[IsThumb, DontUseMovt]>; + +// TLS globals +def : Pat<(ARMWrapperPIC tglobaltlsaddr:$addr), + (tLDRLIT_ga_pcrel tglobaltlsaddr:$addr)>, + Requires<[IsThumb, DontUseMovtInPic]>; +def : Pat<(ARMWrapper tglobaltlsaddr:$addr), + (tLDRLIT_ga_abs tglobaltlsaddr:$addr)>, + Requires<[IsThumb, DontUseMovt]>; + + +// JumpTable +def : T1Pat<(ARMWrapperJT tjumptable:$dst), + (tLEApcrelJT tjumptable:$dst)>; + +// Direct calls +def : T1Pat<(ARMcall texternalsym:$func), (tBL texternalsym:$func)>, + Requires<[IsThumb]>; + +// zextload i1 -> zextload i8 +def : T1Pat<(zextloadi1 t_addrmode_is1:$addr), + (tLDRBi t_addrmode_is1:$addr)>; +def : T1Pat<(zextloadi1 t_addrmode_rr:$addr), + (tLDRBr t_addrmode_rr:$addr)>; + +// extload from the stack -> word load from the stack, as it avoids having to +// materialize the base in a separate register. This only works when a word +// load puts the byte/halfword value in the same place in the register that the +// byte/halfword load would, i.e. when little-endian. +def : T1Pat<(extloadi1 t_addrmode_sp:$addr), (tLDRspi t_addrmode_sp:$addr)>, + Requires<[IsThumb, IsThumb1Only, IsLE]>; +def : T1Pat<(extloadi8 t_addrmode_sp:$addr), (tLDRspi t_addrmode_sp:$addr)>, + Requires<[IsThumb, IsThumb1Only, IsLE]>; +def : T1Pat<(extloadi16 t_addrmode_sp:$addr), (tLDRspi t_addrmode_sp:$addr)>, + Requires<[IsThumb, IsThumb1Only, IsLE]>; + +// extload -> zextload +def : T1Pat<(extloadi1 t_addrmode_is1:$addr), (tLDRBi t_addrmode_is1:$addr)>; +def : T1Pat<(extloadi1 t_addrmode_rr:$addr), (tLDRBr t_addrmode_rr:$addr)>; +def : T1Pat<(extloadi8 t_addrmode_is1:$addr), (tLDRBi t_addrmode_is1:$addr)>; +def : T1Pat<(extloadi8 t_addrmode_rr:$addr), (tLDRBr t_addrmode_rr:$addr)>; +def : T1Pat<(extloadi16 t_addrmode_is2:$addr), (tLDRHi t_addrmode_is2:$addr)>; +def : T1Pat<(extloadi16 t_addrmode_rr:$addr), (tLDRHr t_addrmode_rr:$addr)>; + +// post-inc loads and stores + +// post-inc LDR -> LDM r0!, {r1}. The way operands are layed out in LDMs is +// different to how ISel expects them for a post-inc load, so use a pseudo +// and expand it just after ISel. +let usesCustomInserter = 1, mayLoad =1, + Constraints = "$Rn = $Rn_wb,@earlyclobber $Rn_wb" in + def tLDR_postidx: tPseudoInst<(outs rGPR:$Rt, rGPR:$Rn_wb), + (ins rGPR:$Rn, pred:$p), + 4, IIC_iStore_ru, + []>; + +// post-inc STR -> STM r0!, {r1}. The layout of this (because it doesn't def +// multiple registers) is the same in ISel as MachineInstr, so there's no need +// for a pseudo. +def : T1Pat<(post_store rGPR:$Rt, rGPR:$Rn, 4), + (tSTMIA_UPD rGPR:$Rn, rGPR:$Rt)>; + +// If it's impossible to use [r,r] address mode for sextload, select to +// ldr{b|h} + sxt{b|h} instead. +def : T1Pat<(sextloadi8 t_addrmode_is1:$addr), + (tSXTB (tLDRBi t_addrmode_is1:$addr))>, + Requires<[IsThumb, IsThumb1Only, HasV6]>; +def : T1Pat<(sextloadi8 t_addrmode_rr:$addr), + (tSXTB (tLDRBr t_addrmode_rr:$addr))>, + Requires<[IsThumb, IsThumb1Only, HasV6]>; +def : T1Pat<(sextloadi16 t_addrmode_is2:$addr), + (tSXTH (tLDRHi t_addrmode_is2:$addr))>, + Requires<[IsThumb, IsThumb1Only, HasV6]>; +def : T1Pat<(sextloadi16 t_addrmode_rr:$addr), + (tSXTH (tLDRHr t_addrmode_rr:$addr))>, + Requires<[IsThumb, IsThumb1Only, HasV6]>; + +def : T1Pat<(sextloadi8 t_addrmode_is1:$addr), + (tASRri (tLSLri (tLDRBi t_addrmode_is1:$addr), 24), 24)>; +def : T1Pat<(sextloadi8 t_addrmode_rr:$addr), + (tASRri (tLSLri (tLDRBr t_addrmode_rr:$addr), 24), 24)>; +def : T1Pat<(sextloadi16 t_addrmode_is2:$addr), + (tASRri (tLSLri (tLDRHi t_addrmode_is2:$addr), 16), 16)>; +def : T1Pat<(sextloadi16 t_addrmode_rr:$addr), + (tASRri (tLSLri (tLDRHr t_addrmode_rr:$addr), 16), 16)>; + +def : T1Pat<(atomic_load_8 t_addrmode_is1:$src), + (tLDRBi t_addrmode_is1:$src)>; +def : T1Pat<(atomic_load_8 t_addrmode_rr:$src), + (tLDRBr t_addrmode_rr:$src)>; +def : T1Pat<(atomic_load_16 t_addrmode_is2:$src), + (tLDRHi t_addrmode_is2:$src)>; +def : T1Pat<(atomic_load_16 t_addrmode_rr:$src), + (tLDRHr t_addrmode_rr:$src)>; +def : T1Pat<(atomic_load_32 t_addrmode_is4:$src), + (tLDRi t_addrmode_is4:$src)>; +def : T1Pat<(atomic_load_32 t_addrmode_rr:$src), + (tLDRr t_addrmode_rr:$src)>; +def : T1Pat<(atomic_store_8 t_addrmode_is1:$ptr, tGPR:$val), + (tSTRBi tGPR:$val, t_addrmode_is1:$ptr)>; +def : T1Pat<(atomic_store_8 t_addrmode_rr:$ptr, tGPR:$val), + (tSTRBr tGPR:$val, t_addrmode_rr:$ptr)>; +def : T1Pat<(atomic_store_16 t_addrmode_is2:$ptr, tGPR:$val), + (tSTRHi tGPR:$val, t_addrmode_is2:$ptr)>; +def : T1Pat<(atomic_store_16 t_addrmode_rr:$ptr, tGPR:$val), + (tSTRHr tGPR:$val, t_addrmode_rr:$ptr)>; +def : T1Pat<(atomic_store_32 t_addrmode_is4:$ptr, tGPR:$val), + (tSTRi tGPR:$val, t_addrmode_is4:$ptr)>; +def : T1Pat<(atomic_store_32 t_addrmode_rr:$ptr, tGPR:$val), + (tSTRr tGPR:$val, t_addrmode_rr:$ptr)>; + +// Large immediate handling. + +// Two piece imms. +def : T1Pat<(i32 thumb_immshifted:$src), + (tLSLri (tMOVi8 (thumb_immshifted_val imm:$src)), + (thumb_immshifted_shamt imm:$src))>; + +def : T1Pat<(i32 imm0_255_comp:$src), + (tMVN (tMOVi8 (imm_not_XFORM imm:$src)))>; + +def : T1Pat<(i32 imm256_510:$src), + (tADDi8 (tMOVi8 255), + (thumb_imm256_510_addend imm:$src))>; + +// Pseudo instruction that combines ldr from constpool and add pc. This should +// be expanded into two instructions late to allow if-conversion and +// scheduling. +let isReMaterializable = 1 in +def tLDRpci_pic : PseudoInst<(outs GPR:$dst), (ins i32imm:$addr, pclabel:$cp), + NoItinerary, + [(set GPR:$dst, (ARMpic_add (load (ARMWrapper tconstpool:$addr)), + imm:$cp))]>, + Requires<[IsThumb, IsThumb1Only]>; + +// Pseudo-instruction for merged POP and return. +// FIXME: remove when we have a way to marking a MI with these properties. +let isReturn = 1, isTerminator = 1, isBarrier = 1, mayLoad = 1, + hasExtraDefRegAllocReq = 1 in +def tPOP_RET : tPseudoExpand<(outs), (ins pred:$p, reglist:$regs, variable_ops), + 2, IIC_iPop_Br, [], + (tPOP pred:$p, reglist:$regs)>, Sched<[WriteBrL]>; + +// Indirect branch using "mov pc, $Rm" +let isBranch = 1, isTerminator = 1, isBarrier = 1, isIndirectBranch = 1 in { + def tBRIND : tPseudoExpand<(outs), (ins GPR:$Rm, pred:$p), + 2, IIC_Br, [(brind GPR:$Rm)], + (tMOVr PC, GPR:$Rm, pred:$p)>, Sched<[WriteBr]>; +} + + +// In Thumb1, "nop" is encoded as a "mov r8, r8". Technically, the bf00 +// encoding is available on ARMv6K, but we don't differentiate that finely. +def : InstAlias<"nop", (tMOVr R8, R8, 14, 0), 0>, Requires<[IsThumb, IsThumb1Only]>; + + +// "neg" is and alias for "rsb rd, rn, #0" +def : tInstAlias<"neg${s}${p} $Rd, $Rm", + (tRSB tGPR:$Rd, s_cc_out:$s, tGPR:$Rm, pred:$p)>; + + +// Implied destination operand forms for shifts. +def : tInstAlias<"lsl${s}${p} $Rdm, $imm", + (tLSLri tGPR:$Rdm, cc_out:$s, tGPR:$Rdm, imm0_31:$imm, pred:$p)>; +def : tInstAlias<"lsr${s}${p} $Rdm, $imm", + (tLSRri tGPR:$Rdm, cc_out:$s, tGPR:$Rdm, imm_sr:$imm, pred:$p)>; +def : tInstAlias<"asr${s}${p} $Rdm, $imm", + (tASRri tGPR:$Rdm, cc_out:$s, tGPR:$Rdm, imm_sr:$imm, pred:$p)>; + +// Pseudo instruction ldr Rt, =immediate +def tLDRConstPool + : tAsmPseudo<"ldr${p} $Rt, $immediate", + (ins tGPR:$Rt, const_pool_asm_imm:$immediate, pred:$p)>; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMInstrThumb2.td b/capstone/suite/synctools/tablegen/ARM/ARMInstrThumb2.td new file mode 100644 index 000000000..f67075fbf --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMInstrThumb2.td @@ -0,0 +1,4867 @@ +//===-- ARMInstrThumb2.td - Thumb2 support for ARM ---------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file describes the Thumb2 instruction set. +// +//===----------------------------------------------------------------------===// + +// IT block predicate field +def it_pred_asmoperand : AsmOperandClass { + let Name = "ITCondCode"; + let ParserMethod = "parseITCondCode"; +} +def it_pred : Operand<i32> { + let PrintMethod = "printMandatoryPredicateOperand"; + let ParserMatchClass = it_pred_asmoperand; +} + +// IT block condition mask +def it_mask_asmoperand : AsmOperandClass { let Name = "ITMask"; } +def it_mask : Operand<i32> { + let PrintMethod = "printThumbITMask"; + let ParserMatchClass = it_mask_asmoperand; +} + +// t2_shift_imm: An integer that encodes a shift amount and the type of shift +// (asr or lsl). The 6-bit immediate encodes as: +// {5} 0 ==> lsl +// 1 asr +// {4-0} imm5 shift amount. +// asr #32 not allowed +def t2_shift_imm : Operand<i32> { + let PrintMethod = "printShiftImmOperand"; + let ParserMatchClass = ShifterImmAsmOperand; + let DecoderMethod = "DecodeT2ShifterImmOperand"; +} + +// Shifted operands. No register controlled shifts for Thumb2. +// Note: We do not support rrx shifted operands yet. +def t2_so_reg : Operand<i32>, // reg imm + ComplexPattern<i32, 2, "SelectShiftImmShifterOperand", + [shl,srl,sra,rotr]> { + let EncoderMethod = "getT2SORegOpValue"; + let PrintMethod = "printT2SOOperand"; + let DecoderMethod = "DecodeSORegImmOperand"; + let ParserMatchClass = ShiftedImmAsmOperand; + let MIOperandInfo = (ops rGPR, i32imm); +} + +// t2_so_imm_not_XFORM - Return the complement of a t2_so_imm value +def t2_so_imm_not_XFORM : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant(~((uint32_t)N->getZExtValue()), SDLoc(N), + MVT::i32); +}]>; + +// t2_so_imm_neg_XFORM - Return the negation of a t2_so_imm value +def t2_so_imm_neg_XFORM : SDNodeXForm<imm, [{ + return CurDAG->getTargetConstant(-((int)N->getZExtValue()), SDLoc(N), + MVT::i32); +}]>; + +// so_imm_notSext_XFORM - Return a so_imm value packed into the format +// described for so_imm_notSext def below, with sign extension from 16 +// bits. +def t2_so_imm_notSext16_XFORM : SDNodeXForm<imm, [{ + APInt apIntN = N->getAPIntValue(); + unsigned N16bitSignExt = apIntN.trunc(16).sext(32).getZExtValue(); + return CurDAG->getTargetConstant(~N16bitSignExt, SDLoc(N), MVT::i32); +}]>; + +// t2_so_imm - Match a 32-bit immediate operand, which is an +// 8-bit immediate rotated by an arbitrary number of bits, or an 8-bit +// immediate splatted into multiple bytes of the word. +def t2_so_imm_asmoperand : AsmOperandClass { + let Name = "T2SOImm"; + let RenderMethod = "addImmOperands"; + +} +def t2_so_imm : Operand<i32>, ImmLeaf<i32, [{ + return ARM_AM::getT2SOImmVal(Imm) != -1; + }]> { + let ParserMatchClass = t2_so_imm_asmoperand; + let EncoderMethod = "getT2SOImmOpValue"; + let DecoderMethod = "DecodeT2SOImm"; +} + +// t2_so_imm_not - Match an immediate that is a complement +// of a t2_so_imm. +// Note: this pattern doesn't require an encoder method and such, as it's +// only used on aliases (Pat<> and InstAlias<>). The actual encoding +// is handled by the destination instructions, which use t2_so_imm. +def t2_so_imm_not_asmoperand : AsmOperandClass { let Name = "T2SOImmNot"; } +def t2_so_imm_not : Operand<i32>, PatLeaf<(imm), [{ + return ARM_AM::getT2SOImmVal(~((uint32_t)N->getZExtValue())) != -1; +}], t2_so_imm_not_XFORM> { + let ParserMatchClass = t2_so_imm_not_asmoperand; +} + +// t2_so_imm_notSext - match an immediate that is a complement of a t2_so_imm +// if the upper 16 bits are zero. +def t2_so_imm_notSext : Operand<i32>, PatLeaf<(imm), [{ + APInt apIntN = N->getAPIntValue(); + if (!apIntN.isIntN(16)) return false; + unsigned N16bitSignExt = apIntN.trunc(16).sext(32).getZExtValue(); + return ARM_AM::getT2SOImmVal(~N16bitSignExt) != -1; + }], t2_so_imm_notSext16_XFORM> { + let ParserMatchClass = t2_so_imm_not_asmoperand; +} + +// t2_so_imm_neg - Match an immediate that is a negation of a t2_so_imm. +def t2_so_imm_neg_asmoperand : AsmOperandClass { let Name = "T2SOImmNeg"; } +def t2_so_imm_neg : Operand<i32>, ImmLeaf<i32, [{ + return Imm && ARM_AM::getT2SOImmVal(-(uint32_t)Imm) != -1; +}], t2_so_imm_neg_XFORM> { + let ParserMatchClass = t2_so_imm_neg_asmoperand; +} + +/// imm0_4095 predicate - True if the 32-bit immediate is in the range [0,4095]. +def imm0_4095_asmoperand: ImmAsmOperand<0,4095> { let Name = "Imm0_4095"; } +def imm0_4095 : Operand<i32>, ImmLeaf<i32, [{ + return Imm >= 0 && Imm < 4096; +}]> { + let ParserMatchClass = imm0_4095_asmoperand; +} + +def imm0_4095_neg_asmoperand: AsmOperandClass { let Name = "Imm0_4095Neg"; } +def imm0_4095_neg : Operand<i32>, PatLeaf<(i32 imm), [{ + return (uint32_t)(-N->getZExtValue()) < 4096; +}], imm_neg_XFORM> { + let ParserMatchClass = imm0_4095_neg_asmoperand; +} + +def imm1_255_neg : PatLeaf<(i32 imm), [{ + uint32_t Val = -N->getZExtValue(); + return (Val > 0 && Val < 255); +}], imm_neg_XFORM>; + +def imm0_255_not : PatLeaf<(i32 imm), [{ + return (uint32_t)(~N->getZExtValue()) < 255; +}], imm_not_XFORM>; + +def lo5AllOne : PatLeaf<(i32 imm), [{ + // Returns true if all low 5-bits are 1. + return (((uint32_t)N->getZExtValue()) & 0x1FUL) == 0x1FUL; +}]>; + +// Define Thumb2 specific addressing modes. + +// t2addrmode_imm12 := reg + imm12 +def t2addrmode_imm12_asmoperand : AsmOperandClass {let Name="MemUImm12Offset";} +def t2addrmode_imm12 : MemOperand, + ComplexPattern<i32, 2, "SelectT2AddrModeImm12", []> { + let PrintMethod = "printAddrModeImm12Operand<false>"; + let EncoderMethod = "getAddrModeImm12OpValue"; + let DecoderMethod = "DecodeT2AddrModeImm12"; + let ParserMatchClass = t2addrmode_imm12_asmoperand; + let MIOperandInfo = (ops GPR:$base, i32imm:$offsimm); +} + +// t2ldrlabel := imm12 +def t2ldrlabel : Operand<i32> { + let EncoderMethod = "getAddrModeImm12OpValue"; + let PrintMethod = "printThumbLdrLabelOperand"; +} + +def t2ldr_pcrel_imm12_asmoperand : AsmOperandClass {let Name = "MemPCRelImm12";} +def t2ldr_pcrel_imm12 : Operand<i32> { + let ParserMatchClass = t2ldr_pcrel_imm12_asmoperand; + // used for assembler pseudo instruction and maps to t2ldrlabel, so + // doesn't need encoder or print methods of its own. +} + +// ADR instruction labels. +def t2adrlabel : Operand<i32> { + let EncoderMethod = "getT2AdrLabelOpValue"; + let PrintMethod = "printAdrLabelOperand<0>"; +} + +// t2addrmode_posimm8 := reg + imm8 +def MemPosImm8OffsetAsmOperand : AsmOperandClass {let Name="MemPosImm8Offset";} +def t2addrmode_posimm8 : MemOperand { + let PrintMethod = "printT2AddrModeImm8Operand<false>"; + let EncoderMethod = "getT2AddrModeImm8OpValue"; + let DecoderMethod = "DecodeT2AddrModeImm8"; + let ParserMatchClass = MemPosImm8OffsetAsmOperand; + let MIOperandInfo = (ops GPR:$base, i32imm:$offsimm); +} + +// t2addrmode_negimm8 := reg - imm8 +def MemNegImm8OffsetAsmOperand : AsmOperandClass {let Name="MemNegImm8Offset";} +def t2addrmode_negimm8 : MemOperand, + ComplexPattern<i32, 2, "SelectT2AddrModeImm8", []> { + let PrintMethod = "printT2AddrModeImm8Operand<false>"; + let EncoderMethod = "getT2AddrModeImm8OpValue"; + let DecoderMethod = "DecodeT2AddrModeImm8"; + let ParserMatchClass = MemNegImm8OffsetAsmOperand; + let MIOperandInfo = (ops GPR:$base, i32imm:$offsimm); +} + +// t2addrmode_imm8 := reg +/- imm8 +def MemImm8OffsetAsmOperand : AsmOperandClass { let Name = "MemImm8Offset"; } +class T2AddrMode_Imm8 : MemOperand, + ComplexPattern<i32, 2, "SelectT2AddrModeImm8", []> { + let EncoderMethod = "getT2AddrModeImm8OpValue"; + let DecoderMethod = "DecodeT2AddrModeImm8"; + let ParserMatchClass = MemImm8OffsetAsmOperand; + let MIOperandInfo = (ops GPR:$base, i32imm:$offsimm); +} + +def t2addrmode_imm8 : T2AddrMode_Imm8 { + let PrintMethod = "printT2AddrModeImm8Operand<false>"; +} + +def t2addrmode_imm8_pre : T2AddrMode_Imm8 { + let PrintMethod = "printT2AddrModeImm8Operand<true>"; +} + +def t2am_imm8_offset : MemOperand, + ComplexPattern<i32, 1, "SelectT2AddrModeImm8Offset", + [], [SDNPWantRoot]> { + let PrintMethod = "printT2AddrModeImm8OffsetOperand"; + let EncoderMethod = "getT2AddrModeImm8OffsetOpValue"; + let DecoderMethod = "DecodeT2Imm8"; +} + +// t2addrmode_imm8s4 := reg +/- (imm8 << 2) +def MemImm8s4OffsetAsmOperand : AsmOperandClass {let Name = "MemImm8s4Offset";} +class T2AddrMode_Imm8s4 : MemOperand { + let EncoderMethod = "getT2AddrModeImm8s4OpValue"; + let DecoderMethod = "DecodeT2AddrModeImm8s4"; + let ParserMatchClass = MemImm8s4OffsetAsmOperand; + let MIOperandInfo = (ops GPR:$base, i32imm:$offsimm); +} + +def t2addrmode_imm8s4 : T2AddrMode_Imm8s4 { + let PrintMethod = "printT2AddrModeImm8s4Operand<false>"; +} + +def t2addrmode_imm8s4_pre : T2AddrMode_Imm8s4 { + let PrintMethod = "printT2AddrModeImm8s4Operand<true>"; +} + +def t2am_imm8s4_offset_asmoperand : AsmOperandClass { let Name = "Imm8s4"; } +def t2am_imm8s4_offset : MemOperand { + let PrintMethod = "printT2AddrModeImm8s4OffsetOperand"; + let EncoderMethod = "getT2Imm8s4OpValue"; + let DecoderMethod = "DecodeT2Imm8S4"; +} + +// t2addrmode_imm0_1020s4 := reg + (imm8 << 2) +def MemImm0_1020s4OffsetAsmOperand : AsmOperandClass { + let Name = "MemImm0_1020s4Offset"; +} +def t2addrmode_imm0_1020s4 : MemOperand, + ComplexPattern<i32, 2, "SelectT2AddrModeExclusive"> { + let PrintMethod = "printT2AddrModeImm0_1020s4Operand"; + let EncoderMethod = "getT2AddrModeImm0_1020s4OpValue"; + let DecoderMethod = "DecodeT2AddrModeImm0_1020s4"; + let ParserMatchClass = MemImm0_1020s4OffsetAsmOperand; + let MIOperandInfo = (ops GPRnopc:$base, i32imm:$offsimm); +} + +// t2addrmode_so_reg := reg + (reg << imm2) +def t2addrmode_so_reg_asmoperand : AsmOperandClass {let Name="T2MemRegOffset";} +def t2addrmode_so_reg : MemOperand, + ComplexPattern<i32, 3, "SelectT2AddrModeSoReg", []> { + let PrintMethod = "printT2AddrModeSoRegOperand"; + let EncoderMethod = "getT2AddrModeSORegOpValue"; + let DecoderMethod = "DecodeT2AddrModeSOReg"; + let ParserMatchClass = t2addrmode_so_reg_asmoperand; + let MIOperandInfo = (ops GPRnopc:$base, rGPR:$offsreg, i32imm:$offsimm); +} + +// Addresses for the TBB/TBH instructions. +def addrmode_tbb_asmoperand : AsmOperandClass { let Name = "MemTBB"; } +def addrmode_tbb : MemOperand { + let PrintMethod = "printAddrModeTBB"; + let ParserMatchClass = addrmode_tbb_asmoperand; + let MIOperandInfo = (ops GPR:$Rn, rGPR:$Rm); +} +def addrmode_tbh_asmoperand : AsmOperandClass { let Name = "MemTBH"; } +def addrmode_tbh : MemOperand { + let PrintMethod = "printAddrModeTBH"; + let ParserMatchClass = addrmode_tbh_asmoperand; + let MIOperandInfo = (ops GPR:$Rn, rGPR:$Rm); +} + +//===----------------------------------------------------------------------===// +// Multiclass helpers... +// + + +class T2OneRegImm<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<12> imm; + + let Inst{11-8} = Rd; + let Inst{26} = imm{11}; + let Inst{14-12} = imm{10-8}; + let Inst{7-0} = imm{7-0}; +} + + +class T2sOneRegImm<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2sI<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<12> imm; + + let Inst{11-8} = Rd; + let Inst{26} = imm{11}; + let Inst{14-12} = imm{10-8}; + let Inst{7-0} = imm{7-0}; +} + +class T2OneRegCmpImm<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rn; + bits<12> imm; + + let Inst{19-16} = Rn; + let Inst{26} = imm{11}; + let Inst{14-12} = imm{10-8}; + let Inst{7-0} = imm{7-0}; +} + + +class T2OneRegShiftedReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<12> ShiftedRm; + + let Inst{11-8} = Rd; + let Inst{3-0} = ShiftedRm{3-0}; + let Inst{5-4} = ShiftedRm{6-5}; + let Inst{14-12} = ShiftedRm{11-9}; + let Inst{7-6} = ShiftedRm{8-7}; +} + +class T2sOneRegShiftedReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2sI<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<12> ShiftedRm; + + let Inst{11-8} = Rd; + let Inst{3-0} = ShiftedRm{3-0}; + let Inst{5-4} = ShiftedRm{6-5}; + let Inst{14-12} = ShiftedRm{11-9}; + let Inst{7-6} = ShiftedRm{8-7}; +} + +class T2OneRegCmpShiftedReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rn; + bits<12> ShiftedRm; + + let Inst{19-16} = Rn; + let Inst{3-0} = ShiftedRm{3-0}; + let Inst{5-4} = ShiftedRm{6-5}; + let Inst{14-12} = ShiftedRm{11-9}; + let Inst{7-6} = ShiftedRm{8-7}; +} + +class T2TwoReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rm; + + let Inst{11-8} = Rd; + let Inst{3-0} = Rm; +} + +class T2sTwoReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2sI<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rm; + + let Inst{11-8} = Rd; + let Inst{3-0} = Rm; +} + +class T2TwoRegCmp<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rn; + bits<4> Rm; + + let Inst{19-16} = Rn; + let Inst{3-0} = Rm; +} + + +class T2TwoRegImm<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<12> imm; + + let Inst{11-8} = Rd; + let Inst{19-16} = Rn; + let Inst{26} = imm{11}; + let Inst{14-12} = imm{10-8}; + let Inst{7-0} = imm{7-0}; +} + +class T2sTwoRegImm<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2sI<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<12> imm; + + let Inst{11-8} = Rd; + let Inst{19-16} = Rn; + let Inst{26} = imm{11}; + let Inst{14-12} = imm{10-8}; + let Inst{7-0} = imm{7-0}; +} + +class T2TwoRegShiftImm<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rm; + bits<5> imm; + + let Inst{11-8} = Rd; + let Inst{3-0} = Rm; + let Inst{14-12} = imm{4-2}; + let Inst{7-6} = imm{1-0}; +} + +class T2sTwoRegShiftImm<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2sI<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rm; + bits<5> imm; + + let Inst{11-8} = Rd; + let Inst{3-0} = Rm; + let Inst{14-12} = imm{4-2}; + let Inst{7-6} = imm{1-0}; +} + +class T2ThreeReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + + let Inst{11-8} = Rd; + let Inst{19-16} = Rn; + let Inst{3-0} = Rm; +} + +class T2ThreeRegNoP<dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : T2XI<oops, iops, itin, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + + let Inst{11-8} = Rd; + let Inst{19-16} = Rn; + let Inst{3-0} = Rm; +} + +class T2sThreeReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2sI<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + + let Inst{11-8} = Rd; + let Inst{19-16} = Rn; + let Inst{3-0} = Rm; +} + +class T2TwoRegShiftedReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<12> ShiftedRm; + + let Inst{11-8} = Rd; + let Inst{19-16} = Rn; + let Inst{3-0} = ShiftedRm{3-0}; + let Inst{5-4} = ShiftedRm{6-5}; + let Inst{14-12} = ShiftedRm{11-9}; + let Inst{7-6} = ShiftedRm{8-7}; +} + +class T2sTwoRegShiftedReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2sI<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<12> ShiftedRm; + + let Inst{11-8} = Rd; + let Inst{19-16} = Rn; + let Inst{3-0} = ShiftedRm{3-0}; + let Inst{5-4} = ShiftedRm{6-5}; + let Inst{14-12} = ShiftedRm{11-9}; + let Inst{7-6} = ShiftedRm{8-7}; +} + +class T2FourReg<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + bits<4> Ra; + + let Inst{19-16} = Rn; + let Inst{15-12} = Ra; + let Inst{11-8} = Rd; + let Inst{3-0} = Rm; +} + +class T2MulLong<bits<3> opc22_20, bits<4> opc7_4, + string opc, list<dag> pattern> + : T2I<(outs rGPR:$RdLo, rGPR:$RdHi), (ins rGPR:$Rn, rGPR:$Rm), IIC_iMUL64, + opc, "\t$RdLo, $RdHi, $Rn, $Rm", pattern>, + Sched<[WriteMUL64Lo, WriteMUL64Hi, ReadMUL, ReadMUL]> { + bits<4> RdLo; + bits<4> RdHi; + bits<4> Rn; + bits<4> Rm; + + let Inst{31-23} = 0b111110111; + let Inst{22-20} = opc22_20; + let Inst{19-16} = Rn; + let Inst{15-12} = RdLo; + let Inst{11-8} = RdHi; + let Inst{7-4} = opc7_4; + let Inst{3-0} = Rm; +} +class T2MlaLong<bits<3> opc22_20, bits<4> opc7_4, string opc> + : T2I<(outs rGPR:$RdLo, rGPR:$RdHi), + (ins rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi), IIC_iMAC64, + opc, "\t$RdLo, $RdHi, $Rn, $Rm", []>, + RegConstraint<"$RLo = $RdLo, $RHi = $RdHi">, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]> { + bits<4> RdLo; + bits<4> RdHi; + bits<4> Rn; + bits<4> Rm; + + let Inst{31-23} = 0b111110111; + let Inst{22-20} = opc22_20; + let Inst{19-16} = Rn; + let Inst{15-12} = RdLo; + let Inst{11-8} = RdHi; + let Inst{7-4} = opc7_4; + let Inst{3-0} = Rm; +} + + +/// T2I_bin_irs - Defines a set of (op reg, {so_imm|r|so_reg}) patterns for a +/// binary operation that produces a value. These are predicable and can be +/// changed to modify CPSR. +multiclass T2I_bin_irs<bits<4> opcod, string opc, + InstrItinClass iii, InstrItinClass iir, InstrItinClass iis, + SDPatternOperator opnode, bit Commutable = 0, + string wide = ""> { + // shifted imm + def ri : T2sTwoRegImm< + (outs rGPR:$Rd), (ins rGPR:$Rn, t2_so_imm:$imm), iii, + opc, "\t$Rd, $Rn, $imm", + [(set rGPR:$Rd, (opnode rGPR:$Rn, t2_so_imm:$imm))]>, + Sched<[WriteALU, ReadALU]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 0; + let Inst{24-21} = opcod; + let Inst{15} = 0; + } + // register + def rr : T2sThreeReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), iir, + opc, !strconcat(wide, "\t$Rd, $Rn, $Rm"), + [(set rGPR:$Rd, (opnode rGPR:$Rn, rGPR:$Rm))]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + let isCommutable = Commutable; + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + let Inst{14-12} = 0b000; // imm3 + let Inst{7-6} = 0b00; // imm2 + let Inst{5-4} = 0b00; // type + } + // shifted register + def rs : T2sTwoRegShiftedReg< + (outs rGPR:$Rd), (ins rGPR:$Rn, t2_so_reg:$ShiftedRm), iis, + opc, !strconcat(wide, "\t$Rd, $Rn, $ShiftedRm"), + [(set rGPR:$Rd, (opnode rGPR:$Rn, t2_so_reg:$ShiftedRm))]>, + Sched<[WriteALUsi, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + } + // Assembly aliases for optional destination operand when it's the same + // as the source operand. + def : t2InstAlias<!strconcat(opc, "${s}${p} $Rdn, $imm"), + (!cast<Instruction>(NAME#"ri") rGPR:$Rdn, rGPR:$Rdn, + t2_so_imm:$imm, pred:$p, + cc_out:$s)>; + def : t2InstAlias<!strconcat(opc, "${s}${p}", wide, " $Rdn, $Rm"), + (!cast<Instruction>(NAME#"rr") rGPR:$Rdn, rGPR:$Rdn, + rGPR:$Rm, pred:$p, + cc_out:$s)>; + def : t2InstAlias<!strconcat(opc, "${s}${p}", wide, " $Rdn, $shift"), + (!cast<Instruction>(NAME#"rs") rGPR:$Rdn, rGPR:$Rdn, + t2_so_reg:$shift, pred:$p, + cc_out:$s)>; +} + +/// T2I_bin_w_irs - Same as T2I_bin_irs except these operations need +// the ".w" suffix to indicate that they are wide. +multiclass T2I_bin_w_irs<bits<4> opcod, string opc, + InstrItinClass iii, InstrItinClass iir, InstrItinClass iis, + SDPatternOperator opnode, bit Commutable = 0> : + T2I_bin_irs<opcod, opc, iii, iir, iis, opnode, Commutable, ".w"> { + // Assembler aliases w/ the ".w" suffix. + def : t2InstAlias<!strconcat(opc, "${s}${p}.w", " $Rd, $Rn, $imm"), + (!cast<Instruction>(NAME#"ri") rGPR:$Rd, rGPR:$Rn, t2_so_imm:$imm, pred:$p, + cc_out:$s)>; + // Assembler aliases w/o the ".w" suffix. + def : t2InstAlias<!strconcat(opc, "${s}${p}", " $Rd, $Rn, $Rm"), + (!cast<Instruction>(NAME#"rr") rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, pred:$p, + cc_out:$s)>; + def : t2InstAlias<!strconcat(opc, "${s}${p}", " $Rd, $Rn, $shift"), + (!cast<Instruction>(NAME#"rs") rGPR:$Rd, rGPR:$Rn, t2_so_reg:$shift, + pred:$p, cc_out:$s)>; + + // and with the optional destination operand, too. + def : t2InstAlias<!strconcat(opc, "${s}${p}.w", " $Rdn, $imm"), + (!cast<Instruction>(NAME#"ri") rGPR:$Rdn, rGPR:$Rdn, t2_so_imm:$imm, + pred:$p, cc_out:$s)>; + def : t2InstAlias<!strconcat(opc, "${s}${p}", " $Rdn, $Rm"), + (!cast<Instruction>(NAME#"rr") rGPR:$Rdn, rGPR:$Rdn, rGPR:$Rm, pred:$p, + cc_out:$s)>; + def : t2InstAlias<!strconcat(opc, "${s}${p}", " $Rdn, $shift"), + (!cast<Instruction>(NAME#"rs") rGPR:$Rdn, rGPR:$Rdn, t2_so_reg:$shift, + pred:$p, cc_out:$s)>; +} + +/// T2I_rbin_is - Same as T2I_bin_irs except the order of operands are +/// reversed. The 'rr' form is only defined for the disassembler; for codegen +/// it is equivalent to the T2I_bin_irs counterpart. +multiclass T2I_rbin_irs<bits<4> opcod, string opc, SDNode opnode> { + // shifted imm + def ri : T2sTwoRegImm< + (outs rGPR:$Rd), (ins rGPR:$Rn, t2_so_imm:$imm), IIC_iALUi, + opc, ".w\t$Rd, $Rn, $imm", + [(set rGPR:$Rd, (opnode t2_so_imm:$imm, rGPR:$Rn))]>, + Sched<[WriteALU, ReadALU]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 0; + let Inst{24-21} = opcod; + let Inst{15} = 0; + } + // register + def rr : T2sThreeReg< + (outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), IIC_iALUr, + opc, "\t$Rd, $Rn, $Rm", + [/* For disassembly only; pattern left blank */]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + let Inst{14-12} = 0b000; // imm3 + let Inst{7-6} = 0b00; // imm2 + let Inst{5-4} = 0b00; // type + } + // shifted register + def rs : T2sTwoRegShiftedReg< + (outs rGPR:$Rd), (ins rGPR:$Rn, t2_so_reg:$ShiftedRm), + IIC_iALUsir, opc, "\t$Rd, $Rn, $ShiftedRm", + [(set rGPR:$Rd, (opnode t2_so_reg:$ShiftedRm, rGPR:$Rn))]>, + Sched<[WriteALUsi, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + } +} + +/// T2I_bin_s_irs - Similar to T2I_bin_irs except it sets the 's' bit so the +/// instruction modifies the CPSR register. +/// +/// These opcodes will be converted to the real non-S opcodes by +/// AdjustInstrPostInstrSelection after giving then an optional CPSR operand. +let hasPostISelHook = 1, Defs = [CPSR] in { +multiclass T2I_bin_s_irs<InstrItinClass iii, InstrItinClass iir, + InstrItinClass iis, SDNode opnode, + bit Commutable = 0> { + // shifted imm + def ri : t2PseudoInst<(outs rGPR:$Rd), + (ins GPRnopc:$Rn, t2_so_imm:$imm, pred:$p), + 4, iii, + [(set rGPR:$Rd, CPSR, (opnode GPRnopc:$Rn, + t2_so_imm:$imm))]>, + Sched<[WriteALU, ReadALU]>; + // register + def rr : t2PseudoInst<(outs rGPR:$Rd), (ins GPRnopc:$Rn, rGPR:$Rm, pred:$p), + 4, iir, + [(set rGPR:$Rd, CPSR, (opnode GPRnopc:$Rn, + rGPR:$Rm))]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + let isCommutable = Commutable; + } + // shifted register + def rs : t2PseudoInst<(outs rGPR:$Rd), + (ins GPRnopc:$Rn, t2_so_reg:$ShiftedRm, pred:$p), + 4, iis, + [(set rGPR:$Rd, CPSR, (opnode GPRnopc:$Rn, + t2_so_reg:$ShiftedRm))]>, + Sched<[WriteALUsi, ReadALUsr]>; +} +} + +/// T2I_rbin_s_is - Same as T2I_bin_s_irs, except selection DAG +/// operands are reversed. +let hasPostISelHook = 1, Defs = [CPSR] in { +multiclass T2I_rbin_s_is<SDNode opnode> { + // shifted imm + def ri : t2PseudoInst<(outs rGPR:$Rd), + (ins rGPR:$Rn, t2_so_imm:$imm, pred:$p), + 4, IIC_iALUi, + [(set rGPR:$Rd, CPSR, (opnode t2_so_imm:$imm, + rGPR:$Rn))]>, + Sched<[WriteALU, ReadALU]>; + // shifted register + def rs : t2PseudoInst<(outs rGPR:$Rd), + (ins rGPR:$Rn, t2_so_reg:$ShiftedRm, pred:$p), + 4, IIC_iALUsi, + [(set rGPR:$Rd, CPSR, (opnode t2_so_reg:$ShiftedRm, + rGPR:$Rn))]>, + Sched<[WriteALUsi, ReadALU]>; +} +} + +/// T2I_bin_ii12rs - Defines a set of (op reg, {so_imm|imm0_4095|r|so_reg}) +/// patterns for a binary operation that produces a value. +multiclass T2I_bin_ii12rs<bits<3> op23_21, string opc, SDNode opnode, + bit Commutable = 0> { + // shifted imm + // The register-immediate version is re-materializable. This is useful + // in particular for taking the address of a local. + let isReMaterializable = 1 in { + def ri : T2sTwoRegImm< + (outs GPRnopc:$Rd), (ins GPRnopc:$Rn, t2_so_imm:$imm), IIC_iALUi, + opc, ".w\t$Rd, $Rn, $imm", + [(set GPRnopc:$Rd, (opnode GPRnopc:$Rn, t2_so_imm:$imm))]>, + Sched<[WriteALU, ReadALU]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 0; + let Inst{24} = 1; + let Inst{23-21} = op23_21; + let Inst{15} = 0; + } + } + // 12-bit imm + def ri12 : T2I< + (outs GPRnopc:$Rd), (ins GPR:$Rn, imm0_4095:$imm), IIC_iALUi, + !strconcat(opc, "w"), "\t$Rd, $Rn, $imm", + [(set GPRnopc:$Rd, (opnode GPR:$Rn, imm0_4095:$imm))]>, + Sched<[WriteALU, ReadALU]> { + bits<4> Rd; + bits<4> Rn; + bits<12> imm; + let Inst{31-27} = 0b11110; + let Inst{26} = imm{11}; + let Inst{25-24} = 0b10; + let Inst{23-21} = op23_21; + let Inst{20} = 0; // The S bit. + let Inst{19-16} = Rn; + let Inst{15} = 0; + let Inst{14-12} = imm{10-8}; + let Inst{11-8} = Rd; + let Inst{7-0} = imm{7-0}; + } + // register + def rr : T2sThreeReg<(outs GPRnopc:$Rd), (ins GPRnopc:$Rn, rGPR:$Rm), + IIC_iALUr, opc, ".w\t$Rd, $Rn, $Rm", + [(set GPRnopc:$Rd, (opnode GPRnopc:$Rn, rGPR:$Rm))]>, + Sched<[WriteALU, ReadALU, ReadALU]> { + let isCommutable = Commutable; + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24} = 1; + let Inst{23-21} = op23_21; + let Inst{14-12} = 0b000; // imm3 + let Inst{7-6} = 0b00; // imm2 + let Inst{5-4} = 0b00; // type + } + // shifted register + def rs : T2sTwoRegShiftedReg< + (outs GPRnopc:$Rd), (ins GPRnopc:$Rn, t2_so_reg:$ShiftedRm), + IIC_iALUsi, opc, ".w\t$Rd, $Rn, $ShiftedRm", + [(set GPRnopc:$Rd, (opnode GPRnopc:$Rn, t2_so_reg:$ShiftedRm))]>, + Sched<[WriteALUsi, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24} = 1; + let Inst{23-21} = op23_21; + } +} + +/// T2I_adde_sube_irs - Defines a set of (op reg, {so_imm|r|so_reg}) patterns +/// for a binary operation that produces a value and use the carry +/// bit. It's not predicable. +let Defs = [CPSR], Uses = [CPSR] in { +multiclass T2I_adde_sube_irs<bits<4> opcod, string opc, SDNode opnode, + bit Commutable = 0> { + // shifted imm + def ri : T2sTwoRegImm<(outs rGPR:$Rd), (ins rGPR:$Rn, t2_so_imm:$imm), + IIC_iALUi, opc, "\t$Rd, $Rn, $imm", + [(set rGPR:$Rd, CPSR, (opnode rGPR:$Rn, t2_so_imm:$imm, CPSR))]>, + Requires<[IsThumb2]>, Sched<[WriteALU, ReadALU]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 0; + let Inst{24-21} = opcod; + let Inst{15} = 0; + } + // register + def rr : T2sThreeReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), IIC_iALUr, + opc, ".w\t$Rd, $Rn, $Rm", + [(set rGPR:$Rd, CPSR, (opnode rGPR:$Rn, rGPR:$Rm, CPSR))]>, + Requires<[IsThumb2]>, Sched<[WriteALU, ReadALU, ReadALU]> { + let isCommutable = Commutable; + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + let Inst{14-12} = 0b000; // imm3 + let Inst{7-6} = 0b00; // imm2 + let Inst{5-4} = 0b00; // type + } + // shifted register + def rs : T2sTwoRegShiftedReg< + (outs rGPR:$Rd), (ins rGPR:$Rn, t2_so_reg:$ShiftedRm), + IIC_iALUsi, opc, ".w\t$Rd, $Rn, $ShiftedRm", + [(set rGPR:$Rd, CPSR, (opnode rGPR:$Rn, t2_so_reg:$ShiftedRm, CPSR))]>, + Requires<[IsThumb2]>, Sched<[WriteALUsi, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + } +} +} + +/// T2I_sh_ir - Defines a set of (op reg, {so_imm|r}) patterns for a shift / +// rotate operation that produces a value. +multiclass T2I_sh_ir<bits<2> opcod, string opc, Operand ty, SDNode opnode> { + // 5-bit imm + def ri : T2sTwoRegShiftImm< + (outs rGPR:$Rd), (ins rGPR:$Rm, ty:$imm), IIC_iMOVsi, + opc, ".w\t$Rd, $Rm, $imm", + [(set rGPR:$Rd, (opnode rGPR:$Rm, (i32 ty:$imm)))]>, + Sched<[WriteALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-21} = 0b010010; + let Inst{19-16} = 0b1111; // Rn + let Inst{5-4} = opcod; + } + // register + def rr : T2sThreeReg< + (outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), IIC_iMOVsr, + opc, ".w\t$Rd, $Rn, $Rm", + [(set rGPR:$Rd, (opnode rGPR:$Rn, rGPR:$Rm))]>, + Sched<[WriteALU]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0100; + let Inst{22-21} = opcod; + let Inst{15-12} = 0b1111; + let Inst{7-4} = 0b0000; + } + + // Optional destination register + def : t2InstAlias<!strconcat(opc, "${s}${p}", ".w $Rdn, $imm"), + (!cast<Instruction>(NAME#"ri") rGPR:$Rdn, rGPR:$Rdn, ty:$imm, pred:$p, + cc_out:$s)>; + def : t2InstAlias<!strconcat(opc, "${s}${p}", ".w $Rdn, $Rm"), + (!cast<Instruction>(NAME#"rr") rGPR:$Rdn, rGPR:$Rdn, rGPR:$Rm, pred:$p, + cc_out:$s)>; + + // Assembler aliases w/o the ".w" suffix. + def : t2InstAlias<!strconcat(opc, "${s}${p}", " $Rd, $Rn, $imm"), + (!cast<Instruction>(NAME#"ri") rGPR:$Rd, rGPR:$Rn, ty:$imm, pred:$p, + cc_out:$s)>; + def : t2InstAlias<!strconcat(opc, "${s}${p}", " $Rd, $Rn, $Rm"), + (!cast<Instruction>(NAME#"rr") rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, pred:$p, + cc_out:$s)>; + + // and with the optional destination operand, too. + def : t2InstAlias<!strconcat(opc, "${s}${p}", " $Rdn, $imm"), + (!cast<Instruction>(NAME#"ri") rGPR:$Rdn, rGPR:$Rdn, ty:$imm, pred:$p, + cc_out:$s)>; + def : t2InstAlias<!strconcat(opc, "${s}${p}", " $Rdn, $Rm"), + (!cast<Instruction>(NAME#"rr") rGPR:$Rdn, rGPR:$Rdn, rGPR:$Rm, pred:$p, + cc_out:$s)>; +} + +/// T2I_cmp_irs - Defines a set of (op r, {so_imm|r|so_reg}) cmp / test +/// patterns. Similar to T2I_bin_irs except the instruction does not produce +/// a explicit result, only implicitly set CPSR. +multiclass T2I_cmp_irs<bits<4> opcod, string opc, + InstrItinClass iii, InstrItinClass iir, InstrItinClass iis, + SDPatternOperator opnode> { +let isCompare = 1, Defs = [CPSR] in { + // shifted imm + def ri : T2OneRegCmpImm< + (outs), (ins GPRnopc:$Rn, t2_so_imm:$imm), iii, + opc, ".w\t$Rn, $imm", + [(opnode GPRnopc:$Rn, t2_so_imm:$imm)]>, Sched<[WriteCMP]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 0; + let Inst{24-21} = opcod; + let Inst{20} = 1; // The S bit. + let Inst{15} = 0; + let Inst{11-8} = 0b1111; // Rd + } + // register + def rr : T2TwoRegCmp< + (outs), (ins GPRnopc:$Rn, rGPR:$Rm), iir, + opc, ".w\t$Rn, $Rm", + [(opnode GPRnopc:$Rn, rGPR:$Rm)]>, Sched<[WriteCMP]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + let Inst{20} = 1; // The S bit. + let Inst{14-12} = 0b000; // imm3 + let Inst{11-8} = 0b1111; // Rd + let Inst{7-6} = 0b00; // imm2 + let Inst{5-4} = 0b00; // type + } + // shifted register + def rs : T2OneRegCmpShiftedReg< + (outs), (ins GPRnopc:$Rn, t2_so_reg:$ShiftedRm), iis, + opc, ".w\t$Rn, $ShiftedRm", + [(opnode GPRnopc:$Rn, t2_so_reg:$ShiftedRm)]>, + Sched<[WriteCMPsi]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + let Inst{20} = 1; // The S bit. + let Inst{11-8} = 0b1111; // Rd + } +} + + // Assembler aliases w/o the ".w" suffix. + // No alias here for 'rr' version as not all instantiations of this + // multiclass want one (CMP in particular, does not). + def : t2InstAlias<!strconcat(opc, "${p}", " $Rn, $imm"), + (!cast<Instruction>(NAME#"ri") GPRnopc:$Rn, t2_so_imm:$imm, pred:$p)>; + def : t2InstAlias<!strconcat(opc, "${p}", " $Rn, $shift"), + (!cast<Instruction>(NAME#"rs") GPRnopc:$Rn, t2_so_reg:$shift, pred:$p)>; +} + +/// T2I_ld - Defines a set of (op r, {imm12|imm8|so_reg}) load patterns. +multiclass T2I_ld<bit signed, bits<2> opcod, string opc, + InstrItinClass iii, InstrItinClass iis, RegisterClass target, + PatFrag opnode> { + def i12 : T2Ii12<(outs target:$Rt), (ins t2addrmode_imm12:$addr), iii, + opc, ".w\t$Rt, $addr", + [(set target:$Rt, (opnode t2addrmode_imm12:$addr))]>, + Sched<[WriteLd]> { + bits<4> Rt; + bits<17> addr; + let Inst{31-25} = 0b1111100; + let Inst{24} = signed; + let Inst{23} = 1; + let Inst{22-21} = opcod; + let Inst{20} = 1; // load + let Inst{19-16} = addr{16-13}; // Rn + let Inst{15-12} = Rt; + let Inst{11-0} = addr{11-0}; // imm + + let DecoderMethod = "DecodeT2LoadImm12"; + } + def i8 : T2Ii8 <(outs target:$Rt), (ins t2addrmode_negimm8:$addr), iii, + opc, "\t$Rt, $addr", + [(set target:$Rt, (opnode t2addrmode_negimm8:$addr))]>, + Sched<[WriteLd]> { + bits<4> Rt; + bits<13> addr; + let Inst{31-27} = 0b11111; + let Inst{26-25} = 0b00; + let Inst{24} = signed; + let Inst{23} = 0; + let Inst{22-21} = opcod; + let Inst{20} = 1; // load + let Inst{19-16} = addr{12-9}; // Rn + let Inst{15-12} = Rt; + let Inst{11} = 1; + // Offset: index==TRUE, wback==FALSE + let Inst{10} = 1; // The P bit. + let Inst{9} = addr{8}; // U + let Inst{8} = 0; // The W bit. + let Inst{7-0} = addr{7-0}; // imm + + let DecoderMethod = "DecodeT2LoadImm8"; + } + def s : T2Iso <(outs target:$Rt), (ins t2addrmode_so_reg:$addr), iis, + opc, ".w\t$Rt, $addr", + [(set target:$Rt, (opnode t2addrmode_so_reg:$addr))]>, + Sched<[WriteLd]> { + let Inst{31-27} = 0b11111; + let Inst{26-25} = 0b00; + let Inst{24} = signed; + let Inst{23} = 0; + let Inst{22-21} = opcod; + let Inst{20} = 1; // load + let Inst{11-6} = 0b000000; + + bits<4> Rt; + let Inst{15-12} = Rt; + + bits<10> addr; + let Inst{19-16} = addr{9-6}; // Rn + let Inst{3-0} = addr{5-2}; // Rm + let Inst{5-4} = addr{1-0}; // imm + + let DecoderMethod = "DecodeT2LoadShift"; + } + + // pci variant is very similar to i12, but supports negative offsets + // from the PC. + def pci : T2Ipc <(outs target:$Rt), (ins t2ldrlabel:$addr), iii, + opc, ".w\t$Rt, $addr", + [(set target:$Rt, (opnode (ARMWrapper tconstpool:$addr)))]>, + Sched<[WriteLd]> { + let isReMaterializable = 1; + let Inst{31-27} = 0b11111; + let Inst{26-25} = 0b00; + let Inst{24} = signed; + let Inst{22-21} = opcod; + let Inst{20} = 1; // load + let Inst{19-16} = 0b1111; // Rn + + bits<4> Rt; + let Inst{15-12} = Rt{3-0}; + + bits<13> addr; + let Inst{23} = addr{12}; // add = (U == '1') + let Inst{11-0} = addr{11-0}; + + let DecoderMethod = "DecodeT2LoadLabel"; + } +} + +/// T2I_st - Defines a set of (op r, {imm12|imm8|so_reg}) store patterns. +multiclass T2I_st<bits<2> opcod, string opc, + InstrItinClass iii, InstrItinClass iis, RegisterClass target, + PatFrag opnode> { + def i12 : T2Ii12<(outs), (ins target:$Rt, t2addrmode_imm12:$addr), iii, + opc, ".w\t$Rt, $addr", + [(opnode target:$Rt, t2addrmode_imm12:$addr)]>, + Sched<[WriteST]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0001; + let Inst{22-21} = opcod; + let Inst{20} = 0; // !load + + bits<4> Rt; + let Inst{15-12} = Rt; + + bits<17> addr; + let addr{12} = 1; // add = TRUE + let Inst{19-16} = addr{16-13}; // Rn + let Inst{23} = addr{12}; // U + let Inst{11-0} = addr{11-0}; // imm + } + def i8 : T2Ii8 <(outs), (ins target:$Rt, t2addrmode_negimm8:$addr), iii, + opc, "\t$Rt, $addr", + [(opnode target:$Rt, t2addrmode_negimm8:$addr)]>, + Sched<[WriteST]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0000; + let Inst{22-21} = opcod; + let Inst{20} = 0; // !load + let Inst{11} = 1; + // Offset: index==TRUE, wback==FALSE + let Inst{10} = 1; // The P bit. + let Inst{8} = 0; // The W bit. + + bits<4> Rt; + let Inst{15-12} = Rt; + + bits<13> addr; + let Inst{19-16} = addr{12-9}; // Rn + let Inst{9} = addr{8}; // U + let Inst{7-0} = addr{7-0}; // imm + } + def s : T2Iso <(outs), (ins target:$Rt, t2addrmode_so_reg:$addr), iis, + opc, ".w\t$Rt, $addr", + [(opnode target:$Rt, t2addrmode_so_reg:$addr)]>, + Sched<[WriteST]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0000; + let Inst{22-21} = opcod; + let Inst{20} = 0; // !load + let Inst{11-6} = 0b000000; + + bits<4> Rt; + let Inst{15-12} = Rt; + + bits<10> addr; + let Inst{19-16} = addr{9-6}; // Rn + let Inst{3-0} = addr{5-2}; // Rm + let Inst{5-4} = addr{1-0}; // imm + } +} + +/// T2I_ext_rrot - A unary operation with two forms: one whose operand is a +/// register and one whose operand is a register rotated by 8/16/24. +class T2I_ext_rrot_base<bits<3> opcod, dag iops, dag oops, + string opc, string oprs, + list<dag> pattern> + : T2TwoReg<iops, oops, IIC_iEXTr, opc, oprs, pattern> { + bits<2> rot; + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0100; + let Inst{22-20} = opcod; + let Inst{19-16} = 0b1111; // Rn + let Inst{15-12} = 0b1111; + let Inst{7} = 1; + let Inst{5-4} = rot; // rotate +} + +class T2I_ext_rrot<bits<3> opcod, string opc> + : T2I_ext_rrot_base<opcod, + (outs rGPR:$Rd), + (ins rGPR:$Rm, rot_imm:$rot), + opc, ".w\t$Rd, $Rm$rot", []>, + Requires<[IsThumb2]>, + Sched<[WriteALU, ReadALU]>; + +// UXTB16, SXTB16 - Requires HasDSP, does not need the .w qualifier. +class T2I_ext_rrot_xtb16<bits<3> opcod, string opc> + : T2I_ext_rrot_base<opcod, + (outs rGPR:$Rd), + (ins rGPR:$Rm, rot_imm:$rot), + opc, "\t$Rd, $Rm$rot", []>, + Requires<[HasDSP, IsThumb2]>, + Sched<[WriteALU, ReadALU]>; + +/// T2I_exta_rrot - A binary operation with two forms: one whose operand is a +/// register and one whose operand is a register rotated by 8/16/24. +class T2I_exta_rrot<bits<3> opcod, string opc> + : T2ThreeReg<(outs rGPR:$Rd), + (ins rGPR:$Rn, rGPR:$Rm, rot_imm:$rot), + IIC_iEXTAsr, opc, "\t$Rd, $Rn, $Rm$rot", []>, + Requires<[HasDSP, IsThumb2]>, + Sched<[WriteALU, ReadALU]> { + bits<2> rot; + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0100; + let Inst{22-20} = opcod; + let Inst{15-12} = 0b1111; + let Inst{7} = 1; + let Inst{5-4} = rot; +} + +//===----------------------------------------------------------------------===// +// Instructions +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// Miscellaneous Instructions. +// + +class T2PCOneRegImm<dag oops, dag iops, InstrItinClass itin, + string asm, list<dag> pattern> + : T2XI<oops, iops, itin, asm, pattern> { + bits<4> Rd; + bits<12> label; + + let Inst{11-8} = Rd; + let Inst{26} = label{11}; + let Inst{14-12} = label{10-8}; + let Inst{7-0} = label{7-0}; +} + +// LEApcrel - Load a pc-relative address into a register without offending the +// assembler. +def t2ADR : T2PCOneRegImm<(outs rGPR:$Rd), + (ins t2adrlabel:$addr, pred:$p), + IIC_iALUi, "adr{$p}.w\t$Rd, $addr", []>, + Sched<[WriteALU, ReadALU]> { + let Inst{31-27} = 0b11110; + let Inst{25-24} = 0b10; + // Inst{23:21} = '11' (add = FALSE) or '00' (add = TRUE) + let Inst{22} = 0; + let Inst{20} = 0; + let Inst{19-16} = 0b1111; // Rn + let Inst{15} = 0; + + bits<4> Rd; + bits<13> addr; + let Inst{11-8} = Rd; + let Inst{23} = addr{12}; + let Inst{21} = addr{12}; + let Inst{26} = addr{11}; + let Inst{14-12} = addr{10-8}; + let Inst{7-0} = addr{7-0}; + + let DecoderMethod = "DecodeT2Adr"; +} + +let hasSideEffects = 0, isReMaterializable = 1 in +def t2LEApcrel : t2PseudoInst<(outs rGPR:$Rd), (ins i32imm:$label, pred:$p), + 4, IIC_iALUi, []>, Sched<[WriteALU, ReadALU]>; +let hasSideEffects = 1 in +def t2LEApcrelJT : t2PseudoInst<(outs rGPR:$Rd), + (ins i32imm:$label, pred:$p), + 4, IIC_iALUi, + []>, Sched<[WriteALU, ReadALU]>; + + +//===----------------------------------------------------------------------===// +// Load / store Instructions. +// + +// Load +let canFoldAsLoad = 1, isReMaterializable = 1 in +defm t2LDR : T2I_ld<0, 0b10, "ldr", IIC_iLoad_i, IIC_iLoad_si, GPR, load>; + +// Loads with zero extension +defm t2LDRH : T2I_ld<0, 0b01, "ldrh", IIC_iLoad_bh_i, IIC_iLoad_bh_si, + GPRnopc, zextloadi16>; +defm t2LDRB : T2I_ld<0, 0b00, "ldrb", IIC_iLoad_bh_i, IIC_iLoad_bh_si, + GPRnopc, zextloadi8>; + +// Loads with sign extension +defm t2LDRSH : T2I_ld<1, 0b01, "ldrsh", IIC_iLoad_bh_i, IIC_iLoad_bh_si, + GPRnopc, sextloadi16>; +defm t2LDRSB : T2I_ld<1, 0b00, "ldrsb", IIC_iLoad_bh_i, IIC_iLoad_bh_si, + GPRnopc, sextloadi8>; + +let mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 in { +// Load doubleword +def t2LDRDi8 : T2Ii8s4<1, 0, 1, (outs rGPR:$Rt, rGPR:$Rt2), + (ins t2addrmode_imm8s4:$addr), + IIC_iLoad_d_i, "ldrd", "\t$Rt, $Rt2, $addr", "", []>, + Sched<[WriteLd]>; +} // mayLoad = 1, hasSideEffects = 0, hasExtraDefRegAllocReq = 1 + +// zextload i1 -> zextload i8 +def : T2Pat<(zextloadi1 t2addrmode_imm12:$addr), + (t2LDRBi12 t2addrmode_imm12:$addr)>; +def : T2Pat<(zextloadi1 t2addrmode_negimm8:$addr), + (t2LDRBi8 t2addrmode_negimm8:$addr)>; +def : T2Pat<(zextloadi1 t2addrmode_so_reg:$addr), + (t2LDRBs t2addrmode_so_reg:$addr)>; +def : T2Pat<(zextloadi1 (ARMWrapper tconstpool:$addr)), + (t2LDRBpci tconstpool:$addr)>; + +// extload -> zextload +// FIXME: Reduce the number of patterns by legalizing extload to zextload +// earlier? +def : T2Pat<(extloadi1 t2addrmode_imm12:$addr), + (t2LDRBi12 t2addrmode_imm12:$addr)>; +def : T2Pat<(extloadi1 t2addrmode_negimm8:$addr), + (t2LDRBi8 t2addrmode_negimm8:$addr)>; +def : T2Pat<(extloadi1 t2addrmode_so_reg:$addr), + (t2LDRBs t2addrmode_so_reg:$addr)>; +def : T2Pat<(extloadi1 (ARMWrapper tconstpool:$addr)), + (t2LDRBpci tconstpool:$addr)>; + +def : T2Pat<(extloadi8 t2addrmode_imm12:$addr), + (t2LDRBi12 t2addrmode_imm12:$addr)>; +def : T2Pat<(extloadi8 t2addrmode_negimm8:$addr), + (t2LDRBi8 t2addrmode_negimm8:$addr)>; +def : T2Pat<(extloadi8 t2addrmode_so_reg:$addr), + (t2LDRBs t2addrmode_so_reg:$addr)>; +def : T2Pat<(extloadi8 (ARMWrapper tconstpool:$addr)), + (t2LDRBpci tconstpool:$addr)>; + +def : T2Pat<(extloadi16 t2addrmode_imm12:$addr), + (t2LDRHi12 t2addrmode_imm12:$addr)>; +def : T2Pat<(extloadi16 t2addrmode_negimm8:$addr), + (t2LDRHi8 t2addrmode_negimm8:$addr)>; +def : T2Pat<(extloadi16 t2addrmode_so_reg:$addr), + (t2LDRHs t2addrmode_so_reg:$addr)>; +def : T2Pat<(extloadi16 (ARMWrapper tconstpool:$addr)), + (t2LDRHpci tconstpool:$addr)>; + +// FIXME: The destination register of the loads and stores can't be PC, but +// can be SP. We need another regclass (similar to rGPR) to represent +// that. Not a pressing issue since these are selected manually, +// not via pattern. + +// Indexed loads + +let mayLoad = 1, hasSideEffects = 0 in { +def t2LDR_PRE : T2Ipreldst<0, 0b10, 1, 1, (outs GPR:$Rt, GPR:$Rn_wb), + (ins t2addrmode_imm8_pre:$addr), + AddrModeT2_i8, IndexModePre, IIC_iLoad_iu, + "ldr", "\t$Rt, $addr!", "$addr.base = $Rn_wb", []>, + Sched<[WriteLd]>; + +def t2LDR_POST : T2Ipostldst<0, 0b10, 1, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$Rn, t2am_imm8_offset:$offset), + AddrModeT2_i8, IndexModePost, IIC_iLoad_iu, + "ldr", "\t$Rt, $Rn$offset", "$Rn = $Rn_wb", []>, + Sched<[WriteLd]>; + +def t2LDRB_PRE : T2Ipreldst<0, 0b00, 1, 1, (outs GPR:$Rt, GPR:$Rn_wb), + (ins t2addrmode_imm8_pre:$addr), + AddrModeT2_i8, IndexModePre, IIC_iLoad_bh_iu, + "ldrb", "\t$Rt, $addr!", "$addr.base = $Rn_wb", []>, + Sched<[WriteLd]>; + +def t2LDRB_POST : T2Ipostldst<0, 0b00, 1, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$Rn, t2am_imm8_offset:$offset), + AddrModeT2_i8, IndexModePost, IIC_iLoad_bh_iu, + "ldrb", "\t$Rt, $Rn$offset", "$Rn = $Rn_wb", []>; + +def t2LDRH_PRE : T2Ipreldst<0, 0b01, 1, 1, (outs GPR:$Rt, GPR:$Rn_wb), + (ins t2addrmode_imm8_pre:$addr), + AddrModeT2_i8, IndexModePre, IIC_iLoad_bh_iu, + "ldrh", "\t$Rt, $addr!", "$addr.base = $Rn_wb", []>, + Sched<[WriteLd]>; + +def t2LDRH_POST : T2Ipostldst<0, 0b01, 1, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$Rn, t2am_imm8_offset:$offset), + AddrModeT2_i8, IndexModePost, IIC_iLoad_bh_iu, + "ldrh", "\t$Rt, $Rn$offset", "$Rn = $Rn_wb", []>, + Sched<[WriteLd]>; + +def t2LDRSB_PRE : T2Ipreldst<1, 0b00, 1, 1, (outs GPR:$Rt, GPR:$Rn_wb), + (ins t2addrmode_imm8_pre:$addr), + AddrModeT2_i8, IndexModePre, IIC_iLoad_bh_iu, + "ldrsb", "\t$Rt, $addr!", "$addr.base = $Rn_wb", + []>, Sched<[WriteLd]>; + +def t2LDRSB_POST : T2Ipostldst<1, 0b00, 1, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$Rn, t2am_imm8_offset:$offset), + AddrModeT2_i8, IndexModePost, IIC_iLoad_bh_iu, + "ldrsb", "\t$Rt, $Rn$offset", "$Rn = $Rn_wb", []>, + Sched<[WriteLd]>; + +def t2LDRSH_PRE : T2Ipreldst<1, 0b01, 1, 1, (outs GPR:$Rt, GPR:$Rn_wb), + (ins t2addrmode_imm8_pre:$addr), + AddrModeT2_i8, IndexModePre, IIC_iLoad_bh_iu, + "ldrsh", "\t$Rt, $addr!", "$addr.base = $Rn_wb", + []>, Sched<[WriteLd]>; + +def t2LDRSH_POST : T2Ipostldst<1, 0b01, 1, 0, (outs GPR:$Rt, GPR:$Rn_wb), + (ins addr_offset_none:$Rn, t2am_imm8_offset:$offset), + AddrModeT2_i8, IndexModePost, IIC_iLoad_bh_iu, + "ldrsh", "\t$Rt, $Rn$offset", "$Rn = $Rn_wb", []>, + Sched<[WriteLd]>; +} // mayLoad = 1, hasSideEffects = 0 + +// LDRT, LDRBT, LDRHT, LDRSBT, LDRSHT all have offset mode (PUW=0b110). +// Ref: A8.6.57 LDR (immediate, Thumb) Encoding T4 +class T2IldT<bit signed, bits<2> type, string opc, InstrItinClass ii> + : T2Ii8<(outs rGPR:$Rt), (ins t2addrmode_posimm8:$addr), ii, opc, + "\t$Rt, $addr", []>, Sched<[WriteLd]> { + bits<4> Rt; + bits<13> addr; + let Inst{31-27} = 0b11111; + let Inst{26-25} = 0b00; + let Inst{24} = signed; + let Inst{23} = 0; + let Inst{22-21} = type; + let Inst{20} = 1; // load + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = Rt; + let Inst{11} = 1; + let Inst{10-8} = 0b110; // PUW. + let Inst{7-0} = addr{7-0}; + + let DecoderMethod = "DecodeT2LoadT"; +} + +def t2LDRT : T2IldT<0, 0b10, "ldrt", IIC_iLoad_i>; +def t2LDRBT : T2IldT<0, 0b00, "ldrbt", IIC_iLoad_bh_i>; +def t2LDRHT : T2IldT<0, 0b01, "ldrht", IIC_iLoad_bh_i>; +def t2LDRSBT : T2IldT<1, 0b00, "ldrsbt", IIC_iLoad_bh_i>; +def t2LDRSHT : T2IldT<1, 0b01, "ldrsht", IIC_iLoad_bh_i>; + +class T2Ildacq<bits<4> bits23_20, bits<2> bit54, dag oops, dag iops, + string opc, string asm, list<dag> pattern> + : Thumb2I<oops, iops, AddrModeNone, 4, NoItinerary, + opc, asm, "", pattern>, Requires<[IsThumb, HasAcquireRelease]> { + bits<4> Rt; + bits<4> addr; + + let Inst{31-27} = 0b11101; + let Inst{26-24} = 0b000; + let Inst{23-20} = bits23_20; + let Inst{11-6} = 0b111110; + let Inst{5-4} = bit54; + let Inst{3-0} = 0b1111; + + // Encode instruction operands + let Inst{19-16} = addr; + let Inst{15-12} = Rt; +} + +def t2LDA : T2Ildacq<0b1101, 0b10, (outs rGPR:$Rt), + (ins addr_offset_none:$addr), "lda", "\t$Rt, $addr", []>, + Sched<[WriteLd]>; +def t2LDAB : T2Ildacq<0b1101, 0b00, (outs rGPR:$Rt), + (ins addr_offset_none:$addr), "ldab", "\t$Rt, $addr", []>, + Sched<[WriteLd]>; +def t2LDAH : T2Ildacq<0b1101, 0b01, (outs rGPR:$Rt), + (ins addr_offset_none:$addr), "ldah", "\t$Rt, $addr", []>, + Sched<[WriteLd]>; + +// Store +defm t2STR :T2I_st<0b10,"str", IIC_iStore_i, IIC_iStore_si, GPR, store>; +defm t2STRB:T2I_st<0b00,"strb", IIC_iStore_bh_i, IIC_iStore_bh_si, + rGPR, truncstorei8>; +defm t2STRH:T2I_st<0b01,"strh", IIC_iStore_bh_i, IIC_iStore_bh_si, + rGPR, truncstorei16>; + +// Store doubleword +let mayStore = 1, hasSideEffects = 0, hasExtraSrcRegAllocReq = 1 in +def t2STRDi8 : T2Ii8s4<1, 0, 0, (outs), + (ins rGPR:$Rt, rGPR:$Rt2, t2addrmode_imm8s4:$addr), + IIC_iStore_d_r, "strd", "\t$Rt, $Rt2, $addr", "", []>, + Sched<[WriteST]>; + +// Indexed stores + +let mayStore = 1, hasSideEffects = 0 in { +def t2STR_PRE : T2Ipreldst<0, 0b10, 0, 1, (outs GPRnopc:$Rn_wb), + (ins GPRnopc:$Rt, t2addrmode_imm8_pre:$addr), + AddrModeT2_i8, IndexModePre, IIC_iStore_iu, + "str", "\t$Rt, $addr!", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", []>, + Sched<[WriteST]>; + +def t2STRH_PRE : T2Ipreldst<0, 0b01, 0, 1, (outs GPRnopc:$Rn_wb), + (ins rGPR:$Rt, t2addrmode_imm8_pre:$addr), + AddrModeT2_i8, IndexModePre, IIC_iStore_iu, + "strh", "\t$Rt, $addr!", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", []>, + Sched<[WriteST]>; + +def t2STRB_PRE : T2Ipreldst<0, 0b00, 0, 1, (outs GPRnopc:$Rn_wb), + (ins rGPR:$Rt, t2addrmode_imm8_pre:$addr), + AddrModeT2_i8, IndexModePre, IIC_iStore_bh_iu, + "strb", "\t$Rt, $addr!", + "$addr.base = $Rn_wb,@earlyclobber $Rn_wb", []>, + Sched<[WriteST]>; +} // mayStore = 1, hasSideEffects = 0 + +def t2STR_POST : T2Ipostldst<0, 0b10, 0, 0, (outs GPRnopc:$Rn_wb), + (ins GPRnopc:$Rt, addr_offset_none:$Rn, + t2am_imm8_offset:$offset), + AddrModeT2_i8, IndexModePost, IIC_iStore_iu, + "str", "\t$Rt, $Rn$offset", + "$Rn = $Rn_wb,@earlyclobber $Rn_wb", + [(set GPRnopc:$Rn_wb, + (post_store GPRnopc:$Rt, addr_offset_none:$Rn, + t2am_imm8_offset:$offset))]>, + Sched<[WriteST]>; + +def t2STRH_POST : T2Ipostldst<0, 0b01, 0, 0, (outs GPRnopc:$Rn_wb), + (ins rGPR:$Rt, addr_offset_none:$Rn, + t2am_imm8_offset:$offset), + AddrModeT2_i8, IndexModePost, IIC_iStore_bh_iu, + "strh", "\t$Rt, $Rn$offset", + "$Rn = $Rn_wb,@earlyclobber $Rn_wb", + [(set GPRnopc:$Rn_wb, + (post_truncsti16 rGPR:$Rt, addr_offset_none:$Rn, + t2am_imm8_offset:$offset))]>, + Sched<[WriteST]>; + +def t2STRB_POST : T2Ipostldst<0, 0b00, 0, 0, (outs GPRnopc:$Rn_wb), + (ins rGPR:$Rt, addr_offset_none:$Rn, + t2am_imm8_offset:$offset), + AddrModeT2_i8, IndexModePost, IIC_iStore_bh_iu, + "strb", "\t$Rt, $Rn$offset", + "$Rn = $Rn_wb,@earlyclobber $Rn_wb", + [(set GPRnopc:$Rn_wb, + (post_truncsti8 rGPR:$Rt, addr_offset_none:$Rn, + t2am_imm8_offset:$offset))]>, + Sched<[WriteST]>; + +// Pseudo-instructions for pattern matching the pre-indexed stores. We can't +// put the patterns on the instruction definitions directly as ISel wants +// the address base and offset to be separate operands, not a single +// complex operand like we represent the instructions themselves. The +// pseudos map between the two. +let usesCustomInserter = 1, + Constraints = "$Rn = $Rn_wb,@earlyclobber $Rn_wb" in { +def t2STR_preidx: t2PseudoInst<(outs GPRnopc:$Rn_wb), + (ins rGPR:$Rt, GPRnopc:$Rn, t2am_imm8_offset:$offset, pred:$p), + 4, IIC_iStore_ru, + [(set GPRnopc:$Rn_wb, + (pre_store rGPR:$Rt, GPRnopc:$Rn, t2am_imm8_offset:$offset))]>, + Sched<[WriteST]>; +def t2STRB_preidx: t2PseudoInst<(outs GPRnopc:$Rn_wb), + (ins rGPR:$Rt, GPRnopc:$Rn, t2am_imm8_offset:$offset, pred:$p), + 4, IIC_iStore_ru, + [(set GPRnopc:$Rn_wb, + (pre_truncsti8 rGPR:$Rt, GPRnopc:$Rn, t2am_imm8_offset:$offset))]>, + Sched<[WriteST]>; +def t2STRH_preidx: t2PseudoInst<(outs GPRnopc:$Rn_wb), + (ins rGPR:$Rt, GPRnopc:$Rn, t2am_imm8_offset:$offset, pred:$p), + 4, IIC_iStore_ru, + [(set GPRnopc:$Rn_wb, + (pre_truncsti16 rGPR:$Rt, GPRnopc:$Rn, t2am_imm8_offset:$offset))]>, + Sched<[WriteST]>; +} + +// STRT, STRBT, STRHT all have offset mode (PUW=0b110) and are for disassembly +// only. +// Ref: A8.6.193 STR (immediate, Thumb) Encoding T4 +class T2IstT<bits<2> type, string opc, InstrItinClass ii> + : T2Ii8<(outs rGPR:$Rt), (ins t2addrmode_imm8:$addr), ii, opc, + "\t$Rt, $addr", []>, Sched<[WriteST]> { + let Inst{31-27} = 0b11111; + let Inst{26-25} = 0b00; + let Inst{24} = 0; // not signed + let Inst{23} = 0; + let Inst{22-21} = type; + let Inst{20} = 0; // store + let Inst{11} = 1; + let Inst{10-8} = 0b110; // PUW + + bits<4> Rt; + bits<13> addr; + let Inst{15-12} = Rt; + let Inst{19-16} = addr{12-9}; + let Inst{7-0} = addr{7-0}; +} + +def t2STRT : T2IstT<0b10, "strt", IIC_iStore_i>; +def t2STRBT : T2IstT<0b00, "strbt", IIC_iStore_bh_i>; +def t2STRHT : T2IstT<0b01, "strht", IIC_iStore_bh_i>; + +// ldrd / strd pre / post variants + +let mayLoad = 1 in +def t2LDRD_PRE : T2Ii8s4<1, 1, 1, (outs rGPR:$Rt, rGPR:$Rt2, GPR:$wb), + (ins t2addrmode_imm8s4_pre:$addr), IIC_iLoad_d_ru, + "ldrd", "\t$Rt, $Rt2, $addr!", "$addr.base = $wb", []>, + Sched<[WriteLd]> { + let DecoderMethod = "DecodeT2LDRDPreInstruction"; +} + +let mayLoad = 1 in +def t2LDRD_POST : T2Ii8s4post<0, 1, 1, (outs rGPR:$Rt, rGPR:$Rt2, GPR:$wb), + (ins addr_offset_none:$addr, t2am_imm8s4_offset:$imm), + IIC_iLoad_d_ru, "ldrd", "\t$Rt, $Rt2, $addr$imm", + "$addr.base = $wb", []>, Sched<[WriteLd]>; + +let mayStore = 1 in +def t2STRD_PRE : T2Ii8s4<1, 1, 0, (outs GPR:$wb), + (ins rGPR:$Rt, rGPR:$Rt2, t2addrmode_imm8s4_pre:$addr), + IIC_iStore_d_ru, "strd", "\t$Rt, $Rt2, $addr!", + "$addr.base = $wb", []>, Sched<[WriteST]> { + let DecoderMethod = "DecodeT2STRDPreInstruction"; +} + +let mayStore = 1 in +def t2STRD_POST : T2Ii8s4post<0, 1, 0, (outs GPR:$wb), + (ins rGPR:$Rt, rGPR:$Rt2, addr_offset_none:$addr, + t2am_imm8s4_offset:$imm), + IIC_iStore_d_ru, "strd", "\t$Rt, $Rt2, $addr$imm", + "$addr.base = $wb", []>, Sched<[WriteST]>; + +class T2Istrrel<bits<2> bit54, dag oops, dag iops, + string opc, string asm, list<dag> pattern> + : Thumb2I<oops, iops, AddrModeNone, 4, NoItinerary, opc, + asm, "", pattern>, Requires<[IsThumb, HasAcquireRelease]>, + Sched<[WriteST]> { + bits<4> Rt; + bits<4> addr; + + let Inst{31-27} = 0b11101; + let Inst{26-20} = 0b0001100; + let Inst{11-6} = 0b111110; + let Inst{5-4} = bit54; + let Inst{3-0} = 0b1111; + + // Encode instruction operands + let Inst{19-16} = addr; + let Inst{15-12} = Rt; +} + +def t2STL : T2Istrrel<0b10, (outs), (ins rGPR:$Rt, addr_offset_none:$addr), + "stl", "\t$Rt, $addr", []>; +def t2STLB : T2Istrrel<0b00, (outs), (ins rGPR:$Rt, addr_offset_none:$addr), + "stlb", "\t$Rt, $addr", []>; +def t2STLH : T2Istrrel<0b01, (outs), (ins rGPR:$Rt, addr_offset_none:$addr), + "stlh", "\t$Rt, $addr", []>; + +// T2Ipl (Preload Data/Instruction) signals the memory system of possible future +// data/instruction access. +// instr_write is inverted for Thumb mode: (prefetch 3) -> (preload 0), +// (prefetch 1) -> (preload 2), (prefetch 2) -> (preload 1). +multiclass T2Ipl<bits<1> write, bits<1> instr, string opc> { + + def i12 : T2Ii12<(outs), (ins t2addrmode_imm12:$addr), IIC_Preload, opc, + "\t$addr", + [(ARMPreload t2addrmode_imm12:$addr, (i32 write), (i32 instr))]>, + Sched<[WritePreLd]> { + let Inst{31-25} = 0b1111100; + let Inst{24} = instr; + let Inst{23} = 1; + let Inst{22} = 0; + let Inst{21} = write; + let Inst{20} = 1; + let Inst{15-12} = 0b1111; + + bits<17> addr; + let Inst{19-16} = addr{16-13}; // Rn + let Inst{11-0} = addr{11-0}; // imm12 + + let DecoderMethod = "DecodeT2LoadImm12"; + } + + def i8 : T2Ii8<(outs), (ins t2addrmode_negimm8:$addr), IIC_Preload, opc, + "\t$addr", + [(ARMPreload t2addrmode_negimm8:$addr, (i32 write), (i32 instr))]>, + Sched<[WritePreLd]> { + let Inst{31-25} = 0b1111100; + let Inst{24} = instr; + let Inst{23} = 0; // U = 0 + let Inst{22} = 0; + let Inst{21} = write; + let Inst{20} = 1; + let Inst{15-12} = 0b1111; + let Inst{11-8} = 0b1100; + + bits<13> addr; + let Inst{19-16} = addr{12-9}; // Rn + let Inst{7-0} = addr{7-0}; // imm8 + + let DecoderMethod = "DecodeT2LoadImm8"; + } + + def s : T2Iso<(outs), (ins t2addrmode_so_reg:$addr), IIC_Preload, opc, + "\t$addr", + [(ARMPreload t2addrmode_so_reg:$addr, (i32 write), (i32 instr))]>, + Sched<[WritePreLd]> { + let Inst{31-25} = 0b1111100; + let Inst{24} = instr; + let Inst{23} = 0; // add = TRUE for T1 + let Inst{22} = 0; + let Inst{21} = write; + let Inst{20} = 1; + let Inst{15-12} = 0b1111; + let Inst{11-6} = 0b000000; + + bits<10> addr; + let Inst{19-16} = addr{9-6}; // Rn + let Inst{3-0} = addr{5-2}; // Rm + let Inst{5-4} = addr{1-0}; // imm2 + + let DecoderMethod = "DecodeT2LoadShift"; + } +} + +defm t2PLD : T2Ipl<0, 0, "pld">, Requires<[IsThumb2]>; +defm t2PLDW : T2Ipl<1, 0, "pldw">, Requires<[IsThumb2,HasV7,HasMP]>; +defm t2PLI : T2Ipl<0, 1, "pli">, Requires<[IsThumb2,HasV7]>; + +// pci variant is very similar to i12, but supports negative offsets +// from the PC. Only PLD and PLI have pci variants (not PLDW) +class T2Iplpci<bits<1> inst, string opc> : T2Iso<(outs), (ins t2ldrlabel:$addr), + IIC_Preload, opc, "\t$addr", + [(ARMPreload (ARMWrapper tconstpool:$addr), + (i32 0), (i32 inst))]>, Sched<[WritePreLd]> { + let Inst{31-25} = 0b1111100; + let Inst{24} = inst; + let Inst{22-20} = 0b001; + let Inst{19-16} = 0b1111; + let Inst{15-12} = 0b1111; + + bits<13> addr; + let Inst{23} = addr{12}; // add = (U == '1') + let Inst{11-0} = addr{11-0}; // imm12 + + let DecoderMethod = "DecodeT2LoadLabel"; +} + +def t2PLDpci : T2Iplpci<0, "pld">, Requires<[IsThumb2]>; +def t2PLIpci : T2Iplpci<1, "pli">, Requires<[IsThumb2,HasV7]>; + +//===----------------------------------------------------------------------===// +// Load / store multiple Instructions. +// + +multiclass thumb2_ld_mult<string asm, InstrItinClass itin, + InstrItinClass itin_upd, bit L_bit> { + def IA : + T2XI<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + itin, !strconcat(asm, "${p}.w\t$Rn, $regs"), []> { + bits<4> Rn; + bits<16> regs; + + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b00; + let Inst{24-23} = 0b01; // Increment After + let Inst{22} = 0; + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + let Inst{19-16} = Rn; + let Inst{15-0} = regs; + } + def IA_UPD : + T2XIt<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + itin_upd, !strconcat(asm, "${p}.w\t$Rn!, $regs"), "$Rn = $wb", []> { + bits<4> Rn; + bits<16> regs; + + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b00; + let Inst{24-23} = 0b01; // Increment After + let Inst{22} = 0; + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + let Inst{19-16} = Rn; + let Inst{15-0} = regs; + } + def DB : + T2XI<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + itin, !strconcat(asm, "db${p}\t$Rn, $regs"), []> { + bits<4> Rn; + bits<16> regs; + + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b00; + let Inst{24-23} = 0b10; // Decrement Before + let Inst{22} = 0; + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + let Inst{19-16} = Rn; + let Inst{15-0} = regs; + } + def DB_UPD : + T2XIt<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + itin_upd, !strconcat(asm, "db${p}\t$Rn!, $regs"), "$Rn = $wb", []> { + bits<4> Rn; + bits<16> regs; + + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b00; + let Inst{24-23} = 0b10; // Decrement Before + let Inst{22} = 0; + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + let Inst{19-16} = Rn; + let Inst{15-0} = regs; + } +} + +let hasSideEffects = 0 in { + +let mayLoad = 1, hasExtraDefRegAllocReq = 1 in +defm t2LDM : thumb2_ld_mult<"ldm", IIC_iLoad_m, IIC_iLoad_mu, 1>; + +multiclass thumb2_st_mult<string asm, InstrItinClass itin, + InstrItinClass itin_upd, bit L_bit> { + def IA : + T2XI<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + itin, !strconcat(asm, "${p}.w\t$Rn, $regs"), []> { + bits<4> Rn; + bits<16> regs; + + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b00; + let Inst{24-23} = 0b01; // Increment After + let Inst{22} = 0; + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + let Inst{19-16} = Rn; + let Inst{15} = 0; + let Inst{14} = regs{14}; + let Inst{13} = 0; + let Inst{12-0} = regs{12-0}; + } + def IA_UPD : + T2XIt<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + itin_upd, !strconcat(asm, "${p}.w\t$Rn!, $regs"), "$Rn = $wb", []> { + bits<4> Rn; + bits<16> regs; + + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b00; + let Inst{24-23} = 0b01; // Increment After + let Inst{22} = 0; + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + let Inst{19-16} = Rn; + let Inst{15} = 0; + let Inst{14} = regs{14}; + let Inst{13} = 0; + let Inst{12-0} = regs{12-0}; + } + def DB : + T2XI<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + itin, !strconcat(asm, "db${p}\t$Rn, $regs"), []> { + bits<4> Rn; + bits<16> regs; + + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b00; + let Inst{24-23} = 0b10; // Decrement Before + let Inst{22} = 0; + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + let Inst{19-16} = Rn; + let Inst{15} = 0; + let Inst{14} = regs{14}; + let Inst{13} = 0; + let Inst{12-0} = regs{12-0}; + } + def DB_UPD : + T2XIt<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops), + itin_upd, !strconcat(asm, "db${p}\t$Rn!, $regs"), "$Rn = $wb", []> { + bits<4> Rn; + bits<16> regs; + + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b00; + let Inst{24-23} = 0b10; // Decrement Before + let Inst{22} = 0; + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + let Inst{19-16} = Rn; + let Inst{15} = 0; + let Inst{14} = regs{14}; + let Inst{13} = 0; + let Inst{12-0} = regs{12-0}; + } +} + + +let mayStore = 1, hasExtraSrcRegAllocReq = 1 in +defm t2STM : thumb2_st_mult<"stm", IIC_iStore_m, IIC_iStore_mu, 0>; + +} // hasSideEffects + + +//===----------------------------------------------------------------------===// +// Move Instructions. +// + +let hasSideEffects = 0 in +def t2MOVr : T2sTwoReg<(outs GPRnopc:$Rd), (ins GPRnopc:$Rm), IIC_iMOVr, + "mov", ".w\t$Rd, $Rm", []>, Sched<[WriteALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = 0b0010; + let Inst{19-16} = 0b1111; // Rn + let Inst{14-12} = 0b000; + let Inst{7-4} = 0b0000; +} +def : t2InstAlias<"mov${p}.w $Rd, $Rm", (t2MOVr GPRnopc:$Rd, GPRnopc:$Rm, + pred:$p, zero_reg)>; +def : t2InstAlias<"movs${p}.w $Rd, $Rm", (t2MOVr GPRnopc:$Rd, GPRnopc:$Rm, + pred:$p, CPSR)>; +def : t2InstAlias<"movs${p} $Rd, $Rm", (t2MOVr GPRnopc:$Rd, GPRnopc:$Rm, + pred:$p, CPSR)>; + +// AddedComplexity to ensure isel tries t2MOVi before t2MOVi16. +let isReMaterializable = 1, isAsCheapAsAMove = 1, isMoveImm = 1, + AddedComplexity = 1 in +def t2MOVi : T2sOneRegImm<(outs rGPR:$Rd), (ins t2_so_imm:$imm), IIC_iMOVi, + "mov", ".w\t$Rd, $imm", + [(set rGPR:$Rd, t2_so_imm:$imm)]>, Sched<[WriteALU]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 0; + let Inst{24-21} = 0b0010; + let Inst{19-16} = 0b1111; // Rn + let Inst{15} = 0; +} + +// cc_out is handled as part of the explicit mnemonic in the parser for 'mov'. +// Use aliases to get that to play nice here. +def : t2InstAlias<"movs${p}.w $Rd, $imm", (t2MOVi rGPR:$Rd, t2_so_imm:$imm, + pred:$p, CPSR)>; +def : t2InstAlias<"movs${p} $Rd, $imm", (t2MOVi rGPR:$Rd, t2_so_imm:$imm, + pred:$p, CPSR)>; + +def : t2InstAlias<"mov${p}.w $Rd, $imm", (t2MOVi rGPR:$Rd, t2_so_imm:$imm, + pred:$p, zero_reg)>; +def : t2InstAlias<"mov${p} $Rd, $imm", (t2MOVi rGPR:$Rd, t2_so_imm:$imm, + pred:$p, zero_reg)>; + +let isReMaterializable = 1, isAsCheapAsAMove = 1, isMoveImm = 1 in +def t2MOVi16 : T2I<(outs rGPR:$Rd), (ins imm0_65535_expr:$imm), IIC_iMOVi, + "movw", "\t$Rd, $imm", + [(set rGPR:$Rd, imm0_65535:$imm)]>, Sched<[WriteALU]>, + Requires<[IsThumb, HasV8MBaseline]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 1; + let Inst{24-21} = 0b0010; + let Inst{20} = 0; // The S bit. + let Inst{15} = 0; + + bits<4> Rd; + bits<16> imm; + + let Inst{11-8} = Rd; + let Inst{19-16} = imm{15-12}; + let Inst{26} = imm{11}; + let Inst{14-12} = imm{10-8}; + let Inst{7-0} = imm{7-0}; + let DecoderMethod = "DecodeT2MOVTWInstruction"; +} + +def : InstAlias<"mov${p} $Rd, $imm", + (t2MOVi16 rGPR:$Rd, imm256_65535_expr:$imm, pred:$p), 0>, + Requires<[IsThumb, HasV8MBaseline]>, Sched<[WriteALU]>; + +def t2MOVi16_ga_pcrel : PseudoInst<(outs rGPR:$Rd), + (ins i32imm:$addr, pclabel:$id), IIC_iMOVi, []>, + Sched<[WriteALU]>; + +let Constraints = "$src = $Rd" in { +def t2MOVTi16 : T2I<(outs rGPR:$Rd), + (ins rGPR:$src, imm0_65535_expr:$imm), IIC_iMOVi, + "movt", "\t$Rd, $imm", + [(set rGPR:$Rd, + (or (and rGPR:$src, 0xffff), lo16AllZero:$imm))]>, + Sched<[WriteALU]>, + Requires<[IsThumb, HasV8MBaseline]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 1; + let Inst{24-21} = 0b0110; + let Inst{20} = 0; // The S bit. + let Inst{15} = 0; + + bits<4> Rd; + bits<16> imm; + + let Inst{11-8} = Rd; + let Inst{19-16} = imm{15-12}; + let Inst{26} = imm{11}; + let Inst{14-12} = imm{10-8}; + let Inst{7-0} = imm{7-0}; + let DecoderMethod = "DecodeT2MOVTWInstruction"; +} + +def t2MOVTi16_ga_pcrel : PseudoInst<(outs rGPR:$Rd), + (ins rGPR:$src, i32imm:$addr, pclabel:$id), IIC_iMOVi, []>, + Sched<[WriteALU]>, Requires<[IsThumb, HasV8MBaseline]>; +} // Constraints + +def : T2Pat<(or rGPR:$src, 0xffff0000), (t2MOVTi16 rGPR:$src, 0xffff)>; + +//===----------------------------------------------------------------------===// +// Extend Instructions. +// + +// Sign extenders + +def t2SXTB : T2I_ext_rrot<0b100, "sxtb">; +def t2SXTH : T2I_ext_rrot<0b000, "sxth">; +def t2SXTB16 : T2I_ext_rrot_xtb16<0b010, "sxtb16">; + +def t2SXTAB : T2I_exta_rrot<0b100, "sxtab">; +def t2SXTAH : T2I_exta_rrot<0b000, "sxtah">; +def t2SXTAB16 : T2I_exta_rrot<0b010, "sxtab16">; + +def : T2Pat<(sext_inreg (rotr rGPR:$Rn, rot_imm:$rot), i8), + (t2SXTB rGPR:$Rn, rot_imm:$rot)>; +def : T2Pat<(sext_inreg (rotr rGPR:$Rn, rot_imm:$rot), i16), + (t2SXTH rGPR:$Rn, rot_imm:$rot)>; +def : Thumb2DSPPat<(add rGPR:$Rn, + (sext_inreg (rotr rGPR:$Rm, rot_imm:$rot), i8)), + (t2SXTAB rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(add rGPR:$Rn, + (sext_inreg (rotr rGPR:$Rm, rot_imm:$rot), i16)), + (t2SXTAH rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(int_arm_sxtb16 rGPR:$Rn), + (t2SXTB16 rGPR:$Rn, 0)>; +def : Thumb2DSPPat<(int_arm_sxtab16 rGPR:$Rn, rGPR:$Rm), + (t2SXTAB16 rGPR:$Rn, rGPR:$Rm, 0)>; + + +// A simple right-shift can also be used in most cases (the exception is the +// SXTH operations with a rotate of 24: there the non-contiguous bits are +// relevant). +def : Thumb2DSPPat<(add rGPR:$Rn, (sext_inreg + (srl rGPR:$Rm, rot_imm:$rot), i8)), + (t2SXTAB rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(add rGPR:$Rn, (sext_inreg + (srl rGPR:$Rm, imm8_or_16:$rot), i16)), + (t2SXTAH rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(add rGPR:$Rn, (sext_inreg + (rotr rGPR:$Rm, (i32 24)), i16)), + (t2SXTAH rGPR:$Rn, rGPR:$Rm, (i32 3))>; +def : Thumb2DSPPat<(add rGPR:$Rn, (sext_inreg + (or (srl rGPR:$Rm, (i32 24)), + (shl rGPR:$Rm, (i32 8))), i16)), + (t2SXTAH rGPR:$Rn, rGPR:$Rm, (i32 3))>; + +// Zero extenders + +let AddedComplexity = 16 in { +def t2UXTB : T2I_ext_rrot<0b101, "uxtb">; +def t2UXTH : T2I_ext_rrot<0b001, "uxth">; +def t2UXTB16 : T2I_ext_rrot_xtb16<0b011, "uxtb16">; + +def : Thumb2DSPPat<(and (rotr rGPR:$Rm, rot_imm:$rot), 0x000000FF), + (t2UXTB rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(and (rotr rGPR:$Rm, rot_imm:$rot), 0x0000FFFF), + (t2UXTH rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(and (rotr rGPR:$Rm, rot_imm:$rot), 0x00FF00FF), + (t2UXTB16 rGPR:$Rm, rot_imm:$rot)>; + +def : Thumb2DSPPat<(int_arm_uxtb16 rGPR:$Rm), + (t2UXTB16 rGPR:$Rm, 0)>; + +// FIXME: This pattern incorrectly assumes the shl operator is a rotate. +// The transformation should probably be done as a combiner action +// instead so we can include a check for masking back in the upper +// eight bits of the source into the lower eight bits of the result. +//def : T2Pat<(and (shl rGPR:$Src, (i32 8)), 0xFF00FF), +// (t2UXTB16 rGPR:$Src, 3)>, +// Requires<[HasDSP, IsThumb2]>; +def : T2Pat<(and (srl rGPR:$Src, (i32 8)), 0xFF00FF), + (t2UXTB16 rGPR:$Src, 1)>, + Requires<[HasDSP, IsThumb2]>; + +def t2UXTAB : T2I_exta_rrot<0b101, "uxtab">; +def t2UXTAH : T2I_exta_rrot<0b001, "uxtah">; +def t2UXTAB16 : T2I_exta_rrot<0b011, "uxtab16">; + +def : Thumb2DSPPat<(add rGPR:$Rn, (and (rotr rGPR:$Rm, rot_imm:$rot), + 0x00FF)), + (t2UXTAB rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(add rGPR:$Rn, (and (rotr rGPR:$Rm, rot_imm:$rot), + 0xFFFF)), + (t2UXTAH rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(add rGPR:$Rn, (and (srl rGPR:$Rm, rot_imm:$rot), + 0xFF)), + (t2UXTAB rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(add rGPR:$Rn, (and (srl rGPR:$Rm, imm8_or_16:$rot), + 0xFFFF)), + (t2UXTAH rGPR:$Rn, rGPR:$Rm, rot_imm:$rot)>; +def : Thumb2DSPPat<(int_arm_uxtab16 rGPR:$Rn, rGPR:$Rm), + (t2UXTAB16 rGPR:$Rn, rGPR:$Rm, 0)>; +} + + +//===----------------------------------------------------------------------===// +// Arithmetic Instructions. +// + +let isAdd = 1 in +defm t2ADD : T2I_bin_ii12rs<0b000, "add", add, 1>; +defm t2SUB : T2I_bin_ii12rs<0b101, "sub", sub>; + +// ADD and SUB with 's' bit set. No 12-bit immediate (T4) variants. +// +// Currently, t2ADDS/t2SUBS are pseudo opcodes that exist only in the +// selection DAG. They are "lowered" to real t2ADD/t2SUB opcodes by +// AdjustInstrPostInstrSelection where we determine whether or not to +// set the "s" bit based on CPSR liveness. +// +// FIXME: Eliminate t2ADDS/t2SUBS pseudo opcodes after adding tablegen +// support for an optional CPSR definition that corresponds to the DAG +// node's second value. We can then eliminate the implicit def of CPSR. +defm t2ADDS : T2I_bin_s_irs <IIC_iALUi, IIC_iALUr, IIC_iALUsi, ARMaddc, 1>; +defm t2SUBS : T2I_bin_s_irs <IIC_iALUi, IIC_iALUr, IIC_iALUsi, ARMsubc>; + +let hasPostISelHook = 1 in { +defm t2ADC : T2I_adde_sube_irs<0b1010, "adc", ARMadde, 1>; +defm t2SBC : T2I_adde_sube_irs<0b1011, "sbc", ARMsube>; +} + +def : t2InstSubst<"adc${s}${p} $rd, $rn, $imm", + (t2SBCri rGPR:$rd, rGPR:$rn, t2_so_imm_not:$imm, pred:$p, s_cc_out:$s)>; +def : t2InstSubst<"sbc${s}${p} $rd, $rn, $imm", + (t2ADCri rGPR:$rd, rGPR:$rn, t2_so_imm_not:$imm, pred:$p, s_cc_out:$s)>; + +def : t2InstSubst<"add${s}${p}.w $rd, $rn, $imm", + (t2SUBri GPRnopc:$rd, GPRnopc:$rn, t2_so_imm_neg:$imm, pred:$p, s_cc_out:$s)>; +def : t2InstSubst<"addw${p} $rd, $rn, $imm", + (t2SUBri12 GPRnopc:$rd, GPR:$rn, t2_so_imm_neg:$imm, pred:$p)>; +def : t2InstSubst<"sub${s}${p}.w $rd, $rn, $imm", + (t2ADDri GPRnopc:$rd, GPRnopc:$rn, t2_so_imm_neg:$imm, pred:$p, s_cc_out:$s)>; +def : t2InstSubst<"subw${p} $rd, $rn, $imm", + (t2ADDri12 GPRnopc:$rd, GPR:$rn, t2_so_imm_neg:$imm, pred:$p)>; +def : t2InstSubst<"subw${p} $Rd, $Rn, $imm", + (t2ADDri12 GPRnopc:$Rd, GPR:$Rn, imm0_4095_neg:$imm, pred:$p)>; +def : t2InstSubst<"sub${s}${p} $rd, $rn, $imm", + (t2ADDri GPRnopc:$rd, GPRnopc:$rn, t2_so_imm_neg:$imm, pred:$p, s_cc_out:$s)>; +def : t2InstSubst<"sub${p} $rd, $rn, $imm", + (t2ADDri12 GPRnopc:$rd, GPR:$rn, t2_so_imm_neg:$imm, pred:$p)>; +// RSB +defm t2RSB : T2I_rbin_irs <0b1110, "rsb", sub>; + +// FIXME: Eliminate them if we can write def : Pat patterns which defines +// CPSR and the implicit def of CPSR is not needed. +defm t2RSBS : T2I_rbin_s_is <ARMsubc>; + +// (sub X, imm) gets canonicalized to (add X, -imm). Match this form. +// The assume-no-carry-in form uses the negation of the input since add/sub +// assume opposite meanings of the carry flag (i.e., carry == !borrow). +// See the definition of AddWithCarry() in the ARM ARM A2.2.1 for the gory +// details. +// The AddedComplexity preferences the first variant over the others since +// it can be shrunk to a 16-bit wide encoding, while the others cannot. +let AddedComplexity = 1 in +def : T2Pat<(add GPR:$src, imm1_255_neg:$imm), + (t2SUBri GPR:$src, imm1_255_neg:$imm)>; +def : T2Pat<(add GPR:$src, t2_so_imm_neg:$imm), + (t2SUBri GPR:$src, t2_so_imm_neg:$imm)>; +def : T2Pat<(add GPR:$src, imm0_4095_neg:$imm), + (t2SUBri12 GPR:$src, imm0_4095_neg:$imm)>; +def : T2Pat<(add GPR:$src, imm0_65535_neg:$imm), + (t2SUBrr GPR:$src, (t2MOVi16 (imm_neg_XFORM imm:$imm)))>; + +let AddedComplexity = 1 in +def : T2Pat<(ARMaddc rGPR:$src, imm1_255_neg:$imm), + (t2SUBSri rGPR:$src, imm1_255_neg:$imm)>; +def : T2Pat<(ARMaddc rGPR:$src, t2_so_imm_neg:$imm), + (t2SUBSri rGPR:$src, t2_so_imm_neg:$imm)>; +def : T2Pat<(ARMaddc rGPR:$src, imm0_65535_neg:$imm), + (t2SUBSrr rGPR:$src, (t2MOVi16 (imm_neg_XFORM imm:$imm)))>; +// The with-carry-in form matches bitwise not instead of the negation. +// Effectively, the inverse interpretation of the carry flag already accounts +// for part of the negation. +let AddedComplexity = 1 in +def : T2Pat<(ARMadde rGPR:$src, imm0_255_not:$imm, CPSR), + (t2SBCri rGPR:$src, imm0_255_not:$imm)>; +def : T2Pat<(ARMadde rGPR:$src, t2_so_imm_not:$imm, CPSR), + (t2SBCri rGPR:$src, t2_so_imm_not:$imm)>; +def : T2Pat<(ARMadde rGPR:$src, imm0_65535_neg:$imm, CPSR), + (t2SBCrr rGPR:$src, (t2MOVi16 (imm_not_XFORM imm:$imm)))>; + +def t2SEL : T2ThreeReg<(outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm), + NoItinerary, "sel", "\t$Rd, $Rn, $Rm", + [(set GPR:$Rd, (int_arm_sel GPR:$Rn, GPR:$Rm))]>, + Requires<[IsThumb2, HasDSP]> { + let Inst{31-27} = 0b11111; + let Inst{26-24} = 0b010; + let Inst{23} = 0b1; + let Inst{22-20} = 0b010; + let Inst{15-12} = 0b1111; + let Inst{7} = 0b1; + let Inst{6-4} = 0b000; +} + +// A6.3.13, A6.3.14, A6.3.15 Parallel addition and subtraction (signed/unsigned) +// And Miscellaneous operations -- for disassembly only +class T2I_pam<bits<3> op22_20, bits<4> op7_4, string opc, + list<dag> pat, dag iops, string asm> + : T2I<(outs rGPR:$Rd), iops, NoItinerary, opc, asm, pat>, + Requires<[IsThumb2, HasDSP]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0101; + let Inst{22-20} = op22_20; + let Inst{15-12} = 0b1111; + let Inst{7-4} = op7_4; + + bits<4> Rd; + bits<4> Rn; + bits<4> Rm; + + let Inst{11-8} = Rd; + let Inst{19-16} = Rn; + let Inst{3-0} = Rm; +} + +class T2I_pam_intrinsics<bits<3> op22_20, bits<4> op7_4, string opc, + Intrinsic intrinsic> + : T2I_pam<op22_20, op7_4, opc, + [(set rGPR:$Rd, (intrinsic rGPR:$Rn, rGPR:$Rm))], + (ins rGPR:$Rn, rGPR:$Rm), "\t$Rd, $Rn, $Rm">; + +class T2I_pam_intrinsics_rev<bits<3> op22_20, bits<4> op7_4, string opc> + : T2I_pam<op22_20, op7_4, opc, [], + (ins rGPR:$Rm, rGPR:$Rn), "\t$Rd, $Rm, $Rn">; + +// Saturating add/subtract +def t2QADD16 : T2I_pam_intrinsics<0b001, 0b0001, "qadd16", int_arm_qadd16>; +def t2QADD8 : T2I_pam_intrinsics<0b000, 0b0001, "qadd8", int_arm_qadd8>; +def t2QASX : T2I_pam_intrinsics<0b010, 0b0001, "qasx", int_arm_qasx>; +def t2UQSUB8 : T2I_pam_intrinsics<0b100, 0b0101, "uqsub8", int_arm_uqsub8>; +def t2QSAX : T2I_pam_intrinsics<0b110, 0b0001, "qsax", int_arm_qsax>; +def t2QSUB16 : T2I_pam_intrinsics<0b101, 0b0001, "qsub16", int_arm_qsub16>; +def t2QSUB8 : T2I_pam_intrinsics<0b100, 0b0001, "qsub8", int_arm_qsub8>; +def t2UQADD16 : T2I_pam_intrinsics<0b001, 0b0101, "uqadd16", int_arm_uqadd16>; +def t2UQADD8 : T2I_pam_intrinsics<0b000, 0b0101, "uqadd8", int_arm_uqadd8>; +def t2UQASX : T2I_pam_intrinsics<0b010, 0b0101, "uqasx", int_arm_uqasx>; +def t2UQSAX : T2I_pam_intrinsics<0b110, 0b0101, "uqsax", int_arm_uqsax>; +def t2UQSUB16 : T2I_pam_intrinsics<0b101, 0b0101, "uqsub16", int_arm_uqsub16>; +def t2QADD : T2I_pam_intrinsics_rev<0b000, 0b1000, "qadd">; +def t2QSUB : T2I_pam_intrinsics_rev<0b000, 0b1010, "qsub">; +def t2QDADD : T2I_pam_intrinsics_rev<0b000, 0b1001, "qdadd">; +def t2QDSUB : T2I_pam_intrinsics_rev<0b000, 0b1011, "qdsub">; + +def : Thumb2DSPPat<(int_arm_qadd rGPR:$Rm, rGPR:$Rn), + (t2QADD rGPR:$Rm, rGPR:$Rn)>; +def : Thumb2DSPPat<(int_arm_qsub rGPR:$Rm, rGPR:$Rn), + (t2QSUB rGPR:$Rm, rGPR:$Rn)>; +def : Thumb2DSPPat<(int_arm_qadd(int_arm_qadd rGPR:$Rm, rGPR:$Rm), rGPR:$Rn), + (t2QDADD rGPR:$Rm, rGPR:$Rn)>; +def : Thumb2DSPPat<(int_arm_qsub rGPR:$Rm, (int_arm_qadd rGPR:$Rn, rGPR:$Rn)), + (t2QDSUB rGPR:$Rm, rGPR:$Rn)>; + +// Signed/Unsigned add/subtract + +def t2SASX : T2I_pam_intrinsics<0b010, 0b0000, "sasx", int_arm_sasx>; +def t2SADD16 : T2I_pam_intrinsics<0b001, 0b0000, "sadd16", int_arm_sadd16>; +def t2SADD8 : T2I_pam_intrinsics<0b000, 0b0000, "sadd8", int_arm_sadd8>; +def t2SSAX : T2I_pam_intrinsics<0b110, 0b0000, "ssax", int_arm_ssax>; +def t2SSUB16 : T2I_pam_intrinsics<0b101, 0b0000, "ssub16", int_arm_ssub16>; +def t2SSUB8 : T2I_pam_intrinsics<0b100, 0b0000, "ssub8", int_arm_ssub8>; +def t2UASX : T2I_pam_intrinsics<0b010, 0b0100, "uasx", int_arm_uasx>; +def t2UADD16 : T2I_pam_intrinsics<0b001, 0b0100, "uadd16", int_arm_uadd16>; +def t2UADD8 : T2I_pam_intrinsics<0b000, 0b0100, "uadd8", int_arm_uadd8>; +def t2USAX : T2I_pam_intrinsics<0b110, 0b0100, "usax", int_arm_usax>; +def t2USUB16 : T2I_pam_intrinsics<0b101, 0b0100, "usub16", int_arm_usub16>; +def t2USUB8 : T2I_pam_intrinsics<0b100, 0b0100, "usub8", int_arm_usub8>; + +// Signed/Unsigned halving add/subtract + +def t2SHASX : T2I_pam_intrinsics<0b010, 0b0010, "shasx", int_arm_shasx>; +def t2SHADD16 : T2I_pam_intrinsics<0b001, 0b0010, "shadd16", int_arm_shadd16>; +def t2SHADD8 : T2I_pam_intrinsics<0b000, 0b0010, "shadd8", int_arm_shadd8>; +def t2SHSAX : T2I_pam_intrinsics<0b110, 0b0010, "shsax", int_arm_shsax>; +def t2SHSUB16 : T2I_pam_intrinsics<0b101, 0b0010, "shsub16", int_arm_shsub16>; +def t2SHSUB8 : T2I_pam_intrinsics<0b100, 0b0010, "shsub8", int_arm_shsub8>; +def t2UHASX : T2I_pam_intrinsics<0b010, 0b0110, "uhasx", int_arm_uhasx>; +def t2UHADD16 : T2I_pam_intrinsics<0b001, 0b0110, "uhadd16", int_arm_uhadd16>; +def t2UHADD8 : T2I_pam_intrinsics<0b000, 0b0110, "uhadd8", int_arm_uhadd8>; +def t2UHSAX : T2I_pam_intrinsics<0b110, 0b0110, "uhsax", int_arm_uhsax>; +def t2UHSUB16 : T2I_pam_intrinsics<0b101, 0b0110, "uhsub16", int_arm_uhsub16>; +def t2UHSUB8 : T2I_pam_intrinsics<0b100, 0b0110, "uhsub8", int_arm_uhsub8>; + +// Helper class for disassembly only +// A6.3.16 & A6.3.17 +// T2Imac - Thumb2 multiply [accumulate, and absolute difference] instructions. +class T2ThreeReg_mac<bit long, bits<3> op22_20, bits<4> op7_4, dag oops, + dag iops, InstrItinClass itin, string opc, string asm, list<dag> pattern> + : T2ThreeReg<oops, iops, itin, opc, asm, pattern> { + let Inst{31-27} = 0b11111; + let Inst{26-24} = 0b011; + let Inst{23} = long; + let Inst{22-20} = op22_20; + let Inst{7-4} = op7_4; +} + +class T2FourReg_mac<bit long, bits<3> op22_20, bits<4> op7_4, dag oops, + dag iops, InstrItinClass itin, string opc, string asm, list<dag> pattern> + : T2FourReg<oops, iops, itin, opc, asm, pattern> { + let Inst{31-27} = 0b11111; + let Inst{26-24} = 0b011; + let Inst{23} = long; + let Inst{22-20} = op22_20; + let Inst{7-4} = op7_4; +} + +// Unsigned Sum of Absolute Differences [and Accumulate]. +def t2USAD8 : T2ThreeReg_mac<0, 0b111, 0b0000, (outs rGPR:$Rd), + (ins rGPR:$Rn, rGPR:$Rm), + NoItinerary, "usad8", "\t$Rd, $Rn, $Rm", + [(set rGPR:$Rd, (int_arm_usad8 rGPR:$Rn, rGPR:$Rm))]>, + Requires<[IsThumb2, HasDSP]> { + let Inst{15-12} = 0b1111; +} +def t2USADA8 : T2FourReg_mac<0, 0b111, 0b0000, (outs rGPR:$Rd), + (ins rGPR:$Rn, rGPR:$Rm, rGPR:$Ra), NoItinerary, + "usada8", "\t$Rd, $Rn, $Rm, $Ra", + [(set rGPR:$Rd, (int_arm_usada8 rGPR:$Rn, rGPR:$Rm, rGPR:$Ra))]>, + Requires<[IsThumb2, HasDSP]>; + +// Signed/Unsigned saturate. +let hasSideEffects = 1 in +class T2SatI<dag iops, string opc, string asm> + : T2I<(outs rGPR:$Rd), iops, NoItinerary, opc, asm, []> { + bits<4> Rd; + bits<4> Rn; + bits<5> sat_imm; + bits<6> sh; + + let Inst{31-24} = 0b11110011; + let Inst{21} = sh{5}; + let Inst{20} = 0; + let Inst{19-16} = Rn; + let Inst{15} = 0; + let Inst{14-12} = sh{4-2}; + let Inst{11-8} = Rd; + let Inst{7-6} = sh{1-0}; + let Inst{5} = 0; + let Inst{4-0} = sat_imm; +} + +def t2SSAT: T2SatI<(ins imm1_32:$sat_imm, rGPR:$Rn, t2_shift_imm:$sh), + "ssat", "\t$Rd, $sat_imm, $Rn$sh">, + Requires<[IsThumb2]> { + let Inst{23-22} = 0b00; + let Inst{5} = 0; +} + +def t2SSAT16: T2SatI<(ins imm1_16:$sat_imm, rGPR:$Rn), + "ssat16", "\t$Rd, $sat_imm, $Rn">, + Requires<[IsThumb2, HasDSP]> { + let Inst{23-22} = 0b00; + let sh = 0b100000; + let Inst{4} = 0; +} + +def t2USAT: T2SatI<(ins imm0_31:$sat_imm, rGPR:$Rn, t2_shift_imm:$sh), + "usat", "\t$Rd, $sat_imm, $Rn$sh">, + Requires<[IsThumb2]> { + let Inst{23-22} = 0b10; +} + +def t2USAT16: T2SatI<(ins imm0_15:$sat_imm, rGPR:$Rn), + "usat16", "\t$Rd, $sat_imm, $Rn">, + Requires<[IsThumb2, HasDSP]> { + let Inst{23-22} = 0b10; + let sh = 0b100000; + let Inst{4} = 0; +} + +def : T2Pat<(ARMssatnoshift GPRnopc:$Rn, imm0_31:$imm), + (t2SSAT imm0_31:$imm, GPRnopc:$Rn, 0)>; +def : T2Pat<(ARMusatnoshift GPRnopc:$Rn, imm0_31:$imm), + (t2USAT imm0_31:$imm, GPRnopc:$Rn, 0)>; +def : T2Pat<(int_arm_ssat GPR:$a, imm1_32:$pos), + (t2SSAT imm1_32:$pos, GPR:$a, 0)>; +def : T2Pat<(int_arm_usat GPR:$a, imm0_31:$pos), + (t2USAT imm0_31:$pos, GPR:$a, 0)>; +def : T2Pat<(int_arm_ssat16 GPR:$a, imm1_16:$pos), + (t2SSAT16 imm1_16:$pos, GPR:$a)>; +def : T2Pat<(int_arm_usat16 GPR:$a, imm0_15:$pos), + (t2USAT16 imm0_15:$pos, GPR:$a)>; + +//===----------------------------------------------------------------------===// +// Shift and rotate Instructions. +// + +defm t2LSL : T2I_sh_ir<0b00, "lsl", imm1_31, shl>; +defm t2LSR : T2I_sh_ir<0b01, "lsr", imm_sr, srl>; +defm t2ASR : T2I_sh_ir<0b10, "asr", imm_sr, sra>; +defm t2ROR : T2I_sh_ir<0b11, "ror", imm0_31, rotr>; + +// LSL #0 is actually MOV, and has slightly different permitted registers to +// LSL with non-zero shift +def : t2InstAlias<"lsl${s}${p} $Rd, $Rm, #0", + (t2MOVr GPRnopc:$Rd, GPRnopc:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"lsl${s}${p}.w $Rd, $Rm, #0", + (t2MOVr GPRnopc:$Rd, GPRnopc:$Rm, pred:$p, cc_out:$s)>; + +// (rotr x, (and y, 0x...1f)) ==> (ROR x, y) +def : T2Pat<(rotr rGPR:$lhs, (and rGPR:$rhs, lo5AllOne)), + (t2RORrr rGPR:$lhs, rGPR:$rhs)>; + +let Uses = [CPSR] in { +def t2RRX : T2sTwoReg<(outs rGPR:$Rd), (ins rGPR:$Rm), IIC_iMOVsi, + "rrx", "\t$Rd, $Rm", + [(set rGPR:$Rd, (ARMrrx rGPR:$Rm))]>, Sched<[WriteALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = 0b0010; + let Inst{19-16} = 0b1111; // Rn + let Inst{14-12} = 0b000; + let Inst{7-4} = 0b0011; +} +} + +let isCodeGenOnly = 1, Defs = [CPSR] in { +def t2MOVsrl_flag : T2TwoRegShiftImm< + (outs rGPR:$Rd), (ins rGPR:$Rm), IIC_iMOVsi, + "lsrs", ".w\t$Rd, $Rm, #1", + [(set rGPR:$Rd, (ARMsrl_flag rGPR:$Rm))]>, + Sched<[WriteALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = 0b0010; + let Inst{20} = 1; // The S bit. + let Inst{19-16} = 0b1111; // Rn + let Inst{5-4} = 0b01; // Shift type. + // Shift amount = Inst{14-12:7-6} = 1. + let Inst{14-12} = 0b000; + let Inst{7-6} = 0b01; +} +def t2MOVsra_flag : T2TwoRegShiftImm< + (outs rGPR:$Rd), (ins rGPR:$Rm), IIC_iMOVsi, + "asrs", ".w\t$Rd, $Rm, #1", + [(set rGPR:$Rd, (ARMsra_flag rGPR:$Rm))]>, + Sched<[WriteALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = 0b0010; + let Inst{20} = 1; // The S bit. + let Inst{19-16} = 0b1111; // Rn + let Inst{5-4} = 0b10; // Shift type. + // Shift amount = Inst{14-12:7-6} = 1. + let Inst{14-12} = 0b000; + let Inst{7-6} = 0b01; +} +} + +//===----------------------------------------------------------------------===// +// Bitwise Instructions. +// + +defm t2AND : T2I_bin_w_irs<0b0000, "and", + IIC_iBITi, IIC_iBITr, IIC_iBITsi, and, 1>; +defm t2ORR : T2I_bin_w_irs<0b0010, "orr", + IIC_iBITi, IIC_iBITr, IIC_iBITsi, or, 1>; +defm t2EOR : T2I_bin_w_irs<0b0100, "eor", + IIC_iBITi, IIC_iBITr, IIC_iBITsi, xor, 1>; + +defm t2BIC : T2I_bin_w_irs<0b0001, "bic", + IIC_iBITi, IIC_iBITr, IIC_iBITsi, + BinOpFrag<(and node:$LHS, (not node:$RHS))>>; + +class T2BitFI<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern> { + bits<4> Rd; + bits<5> msb; + bits<5> lsb; + + let Inst{11-8} = Rd; + let Inst{4-0} = msb{4-0}; + let Inst{14-12} = lsb{4-2}; + let Inst{7-6} = lsb{1-0}; +} + +class T2TwoRegBitFI<dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2BitFI<oops, iops, itin, opc, asm, pattern> { + bits<4> Rn; + + let Inst{19-16} = Rn; +} + +let Constraints = "$src = $Rd" in +def t2BFC : T2BitFI<(outs rGPR:$Rd), (ins rGPR:$src, bf_inv_mask_imm:$imm), + IIC_iUNAsi, "bfc", "\t$Rd, $imm", + [(set rGPR:$Rd, (and rGPR:$src, bf_inv_mask_imm:$imm))]> { + let Inst{31-27} = 0b11110; + let Inst{26} = 0; // should be 0. + let Inst{25} = 1; + let Inst{24-20} = 0b10110; + let Inst{19-16} = 0b1111; // Rn + let Inst{15} = 0; + let Inst{5} = 0; // should be 0. + + bits<10> imm; + let msb{4-0} = imm{9-5}; + let lsb{4-0} = imm{4-0}; +} + +def t2SBFX: T2TwoRegBitFI< + (outs rGPR:$Rd), (ins rGPR:$Rn, imm0_31:$lsb, imm1_32:$msb), + IIC_iUNAsi, "sbfx", "\t$Rd, $Rn, $lsb, $msb", []> { + let Inst{31-27} = 0b11110; + let Inst{25} = 1; + let Inst{24-20} = 0b10100; + let Inst{15} = 0; +} + +def t2UBFX: T2TwoRegBitFI< + (outs rGPR:$Rd), (ins rGPR:$Rn, imm0_31:$lsb, imm1_32:$msb), + IIC_iUNAsi, "ubfx", "\t$Rd, $Rn, $lsb, $msb", []> { + let Inst{31-27} = 0b11110; + let Inst{25} = 1; + let Inst{24-20} = 0b11100; + let Inst{15} = 0; +} + +// A8.8.247 UDF - Undefined (Encoding T2) +def t2UDF : T2XI<(outs), (ins imm0_65535:$imm16), IIC_Br, "udf.w\t$imm16", + [(int_arm_undefined imm0_65535:$imm16)]> { + bits<16> imm16; + let Inst{31-29} = 0b111; + let Inst{28-27} = 0b10; + let Inst{26-20} = 0b1111111; + let Inst{19-16} = imm16{15-12}; + let Inst{15} = 0b1; + let Inst{14-12} = 0b010; + let Inst{11-0} = imm16{11-0}; +} + +// A8.6.18 BFI - Bitfield insert (Encoding T1) +let Constraints = "$src = $Rd" in { + def t2BFI : T2TwoRegBitFI<(outs rGPR:$Rd), + (ins rGPR:$src, rGPR:$Rn, bf_inv_mask_imm:$imm), + IIC_iBITi, "bfi", "\t$Rd, $Rn, $imm", + [(set rGPR:$Rd, (ARMbfi rGPR:$src, rGPR:$Rn, + bf_inv_mask_imm:$imm))]> { + let Inst{31-27} = 0b11110; + let Inst{26} = 0; // should be 0. + let Inst{25} = 1; + let Inst{24-20} = 0b10110; + let Inst{15} = 0; + let Inst{5} = 0; // should be 0. + + bits<10> imm; + let msb{4-0} = imm{9-5}; + let lsb{4-0} = imm{4-0}; + } +} + +defm t2ORN : T2I_bin_irs<0b0011, "orn", + IIC_iBITi, IIC_iBITr, IIC_iBITsi, + BinOpFrag<(or node:$LHS, (not node:$RHS))>, 0, "">; + +/// T2I_un_irs - Defines a set of (op reg, {so_imm|r|so_reg}) patterns for a +/// unary operation that produces a value. These are predicable and can be +/// changed to modify CPSR. +multiclass T2I_un_irs<bits<4> opcod, string opc, + InstrItinClass iii, InstrItinClass iir, InstrItinClass iis, + PatFrag opnode, + bit Cheap = 0, bit ReMat = 0, bit MoveImm = 0> { + // shifted imm + def i : T2sOneRegImm<(outs rGPR:$Rd), (ins t2_so_imm:$imm), iii, + opc, "\t$Rd, $imm", + [(set rGPR:$Rd, (opnode t2_so_imm:$imm))]>, Sched<[WriteALU]> { + let isAsCheapAsAMove = Cheap; + let isReMaterializable = ReMat; + let isMoveImm = MoveImm; + let Inst{31-27} = 0b11110; + let Inst{25} = 0; + let Inst{24-21} = opcod; + let Inst{19-16} = 0b1111; // Rn + let Inst{15} = 0; + } + // register + def r : T2sTwoReg<(outs rGPR:$Rd), (ins rGPR:$Rm), iir, + opc, ".w\t$Rd, $Rm", + [(set rGPR:$Rd, (opnode rGPR:$Rm))]>, Sched<[WriteALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + let Inst{19-16} = 0b1111; // Rn + let Inst{14-12} = 0b000; // imm3 + let Inst{7-6} = 0b00; // imm2 + let Inst{5-4} = 0b00; // type + } + // shifted register + def s : T2sOneRegShiftedReg<(outs rGPR:$Rd), (ins t2_so_reg:$ShiftedRm), iis, + opc, ".w\t$Rd, $ShiftedRm", + [(set rGPR:$Rd, (opnode t2_so_reg:$ShiftedRm))]>, + Sched<[WriteALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = opcod; + let Inst{19-16} = 0b1111; // Rn + } +} + +// Prefer over of t2EORri ra, rb, -1 because mvn has 16-bit version +let AddedComplexity = 1 in +defm t2MVN : T2I_un_irs <0b0011, "mvn", + IIC_iMVNi, IIC_iMVNr, IIC_iMVNsi, + not, 1, 1, 1>; + +let AddedComplexity = 1 in +def : T2Pat<(and rGPR:$src, t2_so_imm_not:$imm), + (t2BICri rGPR:$src, t2_so_imm_not:$imm)>; + +// top16Zero - answer true if the upper 16 bits of $src are 0, false otherwise +def top16Zero: PatLeaf<(i32 rGPR:$src), [{ + return CurDAG->MaskedValueIsZero(SDValue(N,0), APInt::getHighBitsSet(32, 16)); + }]>; + +// so_imm_notSext is needed instead of so_imm_not, as the value of imm +// will match the extended, not the original bitWidth for $src. +def : T2Pat<(and top16Zero:$src, t2_so_imm_notSext:$imm), + (t2BICri rGPR:$src, t2_so_imm_notSext:$imm)>; + + +// FIXME: Disable this pattern on Darwin to workaround an assembler bug. +def : T2Pat<(or rGPR:$src, t2_so_imm_not:$imm), + (t2ORNri rGPR:$src, t2_so_imm_not:$imm)>, + Requires<[IsThumb2]>; + +def : T2Pat<(t2_so_imm_not:$src), + (t2MVNi t2_so_imm_not:$src)>; + +// There are shorter Thumb encodings for ADD than ORR, so to increase +// Thumb2SizeReduction's chances later on we select a t2ADD for an or where +// possible. +def : T2Pat<(or AddLikeOrOp:$Rn, t2_so_imm:$imm), + (t2ADDri $Rn, t2_so_imm:$imm)>; + +def : T2Pat<(or AddLikeOrOp:$Rn, imm0_4095:$Rm), + (t2ADDri12 $Rn, imm0_4095:$Rm)>; + +def : T2Pat<(or AddLikeOrOp:$Rn, non_imm32:$Rm), + (t2ADDrr $Rn, $Rm)>; + +//===----------------------------------------------------------------------===// +// Multiply Instructions. +// +let isCommutable = 1 in +def t2MUL: T2ThreeReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), IIC_iMUL32, + "mul", "\t$Rd, $Rn, $Rm", + [(set rGPR:$Rd, (mul rGPR:$Rn, rGPR:$Rm))]>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0110; + let Inst{22-20} = 0b000; + let Inst{15-12} = 0b1111; // Ra = 0b1111 (no accumulate) + let Inst{7-4} = 0b0000; // Multiply +} + +class T2FourRegMLA<bits<4> op7_4, string opc, list<dag> pattern> + : T2FourReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm, rGPR:$Ra), IIC_iMAC32, + opc, "\t$Rd, $Rn, $Rm, $Ra", pattern>, + Requires<[IsThumb2, UseMulOps]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0110; + let Inst{22-20} = 0b000; + let Inst{7-4} = op7_4; +} + +def t2MLA : T2FourRegMLA<0b0000, "mla", + [(set rGPR:$Rd, (add (mul rGPR:$Rn, rGPR:$Rm), + rGPR:$Ra))]>; +def t2MLS: T2FourRegMLA<0b0001, "mls", + [(set rGPR:$Rd, (sub rGPR:$Ra, (mul rGPR:$Rn, + rGPR:$Rm)))]>; + +// Extra precision multiplies with low / high results +let hasSideEffects = 0 in { +let isCommutable = 1 in { +def t2SMULL : T2MulLong<0b000, 0b0000, "smull", + [(set rGPR:$RdLo, rGPR:$RdHi, + (smullohi rGPR:$Rn, rGPR:$Rm))]>; +def t2UMULL : T2MulLong<0b010, 0b0000, "umull", + [(set rGPR:$RdLo, rGPR:$RdHi, + (umullohi rGPR:$Rn, rGPR:$Rm))]>; +} // isCommutable + +// Multiply + accumulate +def t2SMLAL : T2MlaLong<0b100, 0b0000, "smlal">; +def t2UMLAL : T2MlaLong<0b110, 0b0000, "umlal">; +def t2UMAAL : T2MlaLong<0b110, 0b0110, "umaal">, Requires<[IsThumb2, HasDSP]>; +} // hasSideEffects + +// Rounding variants of the below included for disassembly only + +// Most significant word multiply +class T2SMMUL<bits<4> op7_4, string opc, list<dag> pattern> + : T2ThreeReg<(outs rGPR:$Rd), + (ins rGPR:$Rn, rGPR:$Rm), IIC_iMUL32, + opc, "\t$Rd, $Rn, $Rm", pattern>, + Requires<[IsThumb2, HasDSP]>, + Sched<[WriteMUL32, ReadMUL, ReadMUL]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0110; + let Inst{22-20} = 0b101; + let Inst{15-12} = 0b1111; // Ra = 0b1111 (no accumulate) + let Inst{7-4} = op7_4; +} +def t2SMMUL : T2SMMUL<0b0000, "smmul", [(set rGPR:$Rd, (mulhs rGPR:$Rn, + rGPR:$Rm))]>; +def t2SMMULR : + T2SMMUL<0b0001, "smmulr", + [(set rGPR:$Rd, (ARMsmmlar rGPR:$Rn, rGPR:$Rm, (i32 0)))]>; + +class T2FourRegSMMLA<bits<3> op22_20, bits<4> op7_4, string opc, + list<dag> pattern> + : T2FourReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm, rGPR:$Ra), IIC_iMAC32, + opc, "\t$Rd, $Rn, $Rm, $Ra", pattern>, + Requires<[IsThumb2, HasDSP, UseMulOps]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0110; + let Inst{22-20} = op22_20; + let Inst{7-4} = op7_4; +} + +def t2SMMLA : T2FourRegSMMLA<0b101, 0b0000, "smmla", + [(set rGPR:$Rd, (add (mulhs rGPR:$Rm, rGPR:$Rn), rGPR:$Ra))]>; +def t2SMMLAR: T2FourRegSMMLA<0b101, 0b0001, "smmlar", + [(set rGPR:$Rd, (ARMsmmlar rGPR:$Rn, rGPR:$Rm, rGPR:$Ra))]>; +def t2SMMLS: T2FourRegSMMLA<0b110, 0b0000, "smmls", []>; +def t2SMMLSR: T2FourRegSMMLA<0b110, 0b0001, "smmlsr", + [(set rGPR:$Rd, (ARMsmmlsr rGPR:$Rn, rGPR:$Rm, rGPR:$Ra))]>; + +class T2ThreeRegSMUL<bits<3> op22_20, bits<2> op5_4, string opc, + list<dag> pattern> + : T2ThreeReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), IIC_iMUL16, opc, + "\t$Rd, $Rn, $Rm", pattern>, + Requires<[IsThumb2, HasDSP]>, + Sched<[WriteMUL16, ReadMUL, ReadMUL]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0110; + let Inst{22-20} = op22_20; + let Inst{15-12} = 0b1111; // Ra = 0b1111 (no accumulate) + let Inst{7-6} = 0b00; + let Inst{5-4} = op5_4; +} + +def t2SMULBB : T2ThreeRegSMUL<0b001, 0b00, "smulbb", + [(set rGPR:$Rd, (mul (sext_inreg rGPR:$Rn, i16), + (sext_inreg rGPR:$Rm, i16)))]>; +def t2SMULBT : T2ThreeRegSMUL<0b001, 0b01, "smulbt", + [(set rGPR:$Rd, (mul (sext_inreg rGPR:$Rn, i16), + (sra rGPR:$Rm, (i32 16))))]>; +def t2SMULTB : T2ThreeRegSMUL<0b001, 0b10, "smultb", + [(set rGPR:$Rd, (mul (sra rGPR:$Rn, (i32 16)), + (sext_inreg rGPR:$Rm, i16)))]>; +def t2SMULTT : T2ThreeRegSMUL<0b001, 0b11, "smultt", + [(set rGPR:$Rd, (mul (sra rGPR:$Rn, (i32 16)), + (sra rGPR:$Rm, (i32 16))))]>; +def t2SMULWB : T2ThreeRegSMUL<0b011, 0b00, "smulwb", + [(set rGPR:$Rd, (ARMsmulwb rGPR:$Rn, rGPR:$Rm))]>; +def t2SMULWT : T2ThreeRegSMUL<0b011, 0b01, "smulwt", + [(set rGPR:$Rd, (ARMsmulwt rGPR:$Rn, rGPR:$Rm))]>; + +def : Thumb2DSPPat<(mul sext_16_node:$Rm, sext_16_node:$Rn), + (t2SMULBB rGPR:$Rm, rGPR:$Rn)>; +def : Thumb2DSPPat<(mul sext_16_node:$Rn, (sra rGPR:$Rm, (i32 16))), + (t2SMULBT rGPR:$Rn, rGPR:$Rm)>; +def : Thumb2DSPPat<(mul (sra rGPR:$Rn, (i32 16)), sext_16_node:$Rm), + (t2SMULTB rGPR:$Rn, rGPR:$Rm)>; +def : Thumb2DSPPat<(int_arm_smulbb rGPR:$Rn, rGPR:$Rm), + (t2SMULBB rGPR:$Rn, rGPR:$Rm)>; +def : Thumb2DSPPat<(int_arm_smulbt rGPR:$Rn, rGPR:$Rm), + (t2SMULBT rGPR:$Rn, rGPR:$Rm)>; +def : Thumb2DSPPat<(int_arm_smultb rGPR:$Rn, rGPR:$Rm), + (t2SMULTB rGPR:$Rn, rGPR:$Rm)>; +def : Thumb2DSPPat<(int_arm_smultt rGPR:$Rn, rGPR:$Rm), + (t2SMULTT rGPR:$Rn, rGPR:$Rm)>; +def : Thumb2DSPPat<(int_arm_smulwb rGPR:$Rn, rGPR:$Rm), + (t2SMULWB rGPR:$Rn, rGPR:$Rm)>; +def : Thumb2DSPPat<(int_arm_smulwt rGPR:$Rn, rGPR:$Rm), + (t2SMULWT rGPR:$Rn, rGPR:$Rm)>; + +class T2FourRegSMLA<bits<3> op22_20, bits<2> op5_4, string opc, + list<dag> pattern> + : T2FourReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm, rGPR:$Ra), IIC_iMUL16, + opc, "\t$Rd, $Rn, $Rm, $Ra", pattern>, + Requires<[IsThumb2, HasDSP, UseMulOps]>, + Sched<[WriteMAC16, ReadMUL, ReadMUL, ReadMAC]> { + let Inst{31-27} = 0b11111; + let Inst{26-23} = 0b0110; + let Inst{22-20} = op22_20; + let Inst{7-6} = 0b00; + let Inst{5-4} = op5_4; +} + +def t2SMLABB : T2FourRegSMLA<0b001, 0b00, "smlabb", + [(set rGPR:$Rd, (add rGPR:$Ra, + (mul (sext_inreg rGPR:$Rn, i16), + (sext_inreg rGPR:$Rm, i16))))]>; +def t2SMLABT : T2FourRegSMLA<0b001, 0b01, "smlabt", + [(set rGPR:$Rd, (add rGPR:$Ra, (mul (sext_inreg rGPR:$Rn, i16), + (sra rGPR:$Rm, (i32 16)))))]>; +def t2SMLATB : T2FourRegSMLA<0b001, 0b10, "smlatb", + [(set rGPR:$Rd, (add rGPR:$Ra, (mul (sra rGPR:$Rn, (i32 16)), + (sext_inreg rGPR:$Rm, i16))))]>; +def t2SMLATT : T2FourRegSMLA<0b001, 0b11, "smlatt", + [(set rGPR:$Rd, (add rGPR:$Ra, (mul (sra rGPR:$Rn, (i32 16)), + (sra rGPR:$Rm, (i32 16)))))]>; +def t2SMLAWB : T2FourRegSMLA<0b011, 0b00, "smlawb", + [(set rGPR:$Rd, (add rGPR:$Ra, (ARMsmulwb rGPR:$Rn, rGPR:$Rm)))]>; +def t2SMLAWT : T2FourRegSMLA<0b011, 0b01, "smlawt", + [(set rGPR:$Rd, (add rGPR:$Ra, (ARMsmulwt rGPR:$Rn, rGPR:$Rm)))]>; + +def : Thumb2DSPMulPat<(add rGPR:$Ra, (mul sext_16_node:$Rn, sext_16_node:$Rm)), + (t2SMLABB rGPR:$Rn, rGPR:$Rm, rGPR:$Ra)>; +def : Thumb2DSPMulPat<(add rGPR:$Ra, + (mul sext_16_node:$Rn, (sra rGPR:$Rm, (i32 16)))), + (t2SMLABT rGPR:$Rn, rGPR:$Rm, rGPR:$Ra)>; +def : Thumb2DSPMulPat<(add rGPR:$Ra, + (mul (sra rGPR:$Rn, (i32 16)), sext_16_node:$Rm)), + (t2SMLATB rGPR:$Rn, rGPR:$Rm, rGPR:$Ra)>; + +def : Thumb2DSPPat<(int_arm_smlabb GPR:$a, GPR:$b, GPR:$acc), + (t2SMLABB GPR:$a, GPR:$b, GPR:$acc)>; +def : Thumb2DSPPat<(int_arm_smlabt GPR:$a, GPR:$b, GPR:$acc), + (t2SMLABT GPR:$a, GPR:$b, GPR:$acc)>; +def : Thumb2DSPPat<(int_arm_smlatb GPR:$a, GPR:$b, GPR:$acc), + (t2SMLATB GPR:$a, GPR:$b, GPR:$acc)>; +def : Thumb2DSPPat<(int_arm_smlatt GPR:$a, GPR:$b, GPR:$acc), + (t2SMLATT GPR:$a, GPR:$b, GPR:$acc)>; +def : Thumb2DSPPat<(int_arm_smlawb GPR:$a, GPR:$b, GPR:$acc), + (t2SMLAWB GPR:$a, GPR:$b, GPR:$acc)>; +def : Thumb2DSPPat<(int_arm_smlawt GPR:$a, GPR:$b, GPR:$acc), + (t2SMLAWT GPR:$a, GPR:$b, GPR:$acc)>; + +// Halfword multiple accumulate long: SMLAL<x><y> +def t2SMLALBB : T2MlaLong<0b100, 0b1000, "smlalbb">, + Requires<[IsThumb2, HasDSP]>; +def t2SMLALBT : T2MlaLong<0b100, 0b1001, "smlalbt">, + Requires<[IsThumb2, HasDSP]>; +def t2SMLALTB : T2MlaLong<0b100, 0b1010, "smlaltb">, + Requires<[IsThumb2, HasDSP]>; +def t2SMLALTT : T2MlaLong<0b100, 0b1011, "smlaltt">, + Requires<[IsThumb2, HasDSP]>; + +def : Thumb2DSPPat<(ARMsmlalbb GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + (t2SMLALBB $Rn, $Rm, $RLo, $RHi)>; +def : Thumb2DSPPat<(ARMsmlalbt GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + (t2SMLALBT $Rn, $Rm, $RLo, $RHi)>; +def : Thumb2DSPPat<(ARMsmlaltb GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + (t2SMLALTB $Rn, $Rm, $RLo, $RHi)>; +def : Thumb2DSPPat<(ARMsmlaltt GPR:$Rn, GPR:$Rm, GPR:$RLo, GPR:$RHi), + (t2SMLALTT $Rn, $Rm, $RLo, $RHi)>; + +class T2DualHalfMul<bits<3> op22_20, bits<4> op7_4, string opc, + Intrinsic intrinsic> + : T2ThreeReg_mac<0, op22_20, op7_4, + (outs rGPR:$Rd), + (ins rGPR:$Rn, rGPR:$Rm), + IIC_iMAC32, opc, "\t$Rd, $Rn, $Rm", + [(set rGPR:$Rd, (intrinsic rGPR:$Rn, rGPR:$Rm))]>, + Requires<[IsThumb2, HasDSP]>, + Sched<[WriteMAC32, ReadMUL, ReadMUL, ReadMAC]> { + let Inst{15-12} = 0b1111; +} + +// Dual halfword multiple: SMUAD, SMUSD, SMLAD, SMLSD, SMLALD, SMLSLD +def t2SMUAD: T2DualHalfMul<0b010, 0b0000, "smuad", int_arm_smuad>; +def t2SMUADX: T2DualHalfMul<0b010, 0b0001, "smuadx", int_arm_smuadx>; +def t2SMUSD: T2DualHalfMul<0b100, 0b0000, "smusd", int_arm_smusd>; +def t2SMUSDX: T2DualHalfMul<0b100, 0b0001, "smusdx", int_arm_smusdx>; + +class T2DualHalfMulAdd<bits<3> op22_20, bits<4> op7_4, string opc, + Intrinsic intrinsic> + : T2FourReg_mac<0, op22_20, op7_4, + (outs rGPR:$Rd), + (ins rGPR:$Rn, rGPR:$Rm, rGPR:$Ra), + IIC_iMAC32, opc, "\t$Rd, $Rn, $Rm, $Ra", + [(set rGPR:$Rd, (intrinsic rGPR:$Rn, rGPR:$Rm, rGPR:$Ra))]>, + Requires<[IsThumb2, HasDSP]>; + +def t2SMLAD : T2DualHalfMulAdd<0b010, 0b0000, "smlad", int_arm_smlad>; +def t2SMLADX : T2DualHalfMulAdd<0b010, 0b0001, "smladx", int_arm_smladx>; +def t2SMLSD : T2DualHalfMulAdd<0b100, 0b0000, "smlsd", int_arm_smlsd>; +def t2SMLSDX : T2DualHalfMulAdd<0b100, 0b0001, "smlsdx", int_arm_smlsdx>; + +class T2DualHalfMulAddLong<bits<3> op22_20, bits<4> op7_4, string opc> + : T2FourReg_mac<1, op22_20, op7_4, + (outs rGPR:$Ra, rGPR:$Rd), + (ins rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi), + IIC_iMAC64, opc, "\t$Ra, $Rd, $Rn, $Rm", []>, + RegConstraint<"$Ra = $RLo, $Rd = $RHi">, + Requires<[IsThumb2, HasDSP]>, + Sched<[WriteMAC64Lo, WriteMAC64Hi, ReadMUL, ReadMUL, ReadMAC, ReadMAC]>; + +def t2SMLALD : T2DualHalfMulAddLong<0b100, 0b1100, "smlald">; +def t2SMLALDX : T2DualHalfMulAddLong<0b100, 0b1101, "smlaldx">; +def t2SMLSLD : T2DualHalfMulAddLong<0b101, 0b1100, "smlsld">; +def t2SMLSLDX : T2DualHalfMulAddLong<0b101, 0b1101, "smlsldx">; + +def : Thumb2DSPPat<(ARMSmlald rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi), + (t2SMLALD rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi)>; +def : Thumb2DSPPat<(ARMSmlaldx rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi), + (t2SMLALDX rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi)>; +def : Thumb2DSPPat<(ARMSmlsld rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi), + (t2SMLSLD rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi)>; +def : Thumb2DSPPat<(ARMSmlsldx rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi), + (t2SMLSLDX rGPR:$Rn, rGPR:$Rm, rGPR:$RLo, rGPR:$RHi)>; + +//===----------------------------------------------------------------------===// +// Division Instructions. +// Signed and unsigned division on v7-M +// +def t2SDIV : T2ThreeReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), IIC_iDIV, + "sdiv", "\t$Rd, $Rn, $Rm", + [(set rGPR:$Rd, (sdiv rGPR:$Rn, rGPR:$Rm))]>, + Requires<[HasDivideInThumb, IsThumb, HasV8MBaseline]>, + Sched<[WriteDIV]> { + let Inst{31-27} = 0b11111; + let Inst{26-21} = 0b011100; + let Inst{20} = 0b1; + let Inst{15-12} = 0b1111; + let Inst{7-4} = 0b1111; +} + +def t2UDIV : T2ThreeReg<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), IIC_iDIV, + "udiv", "\t$Rd, $Rn, $Rm", + [(set rGPR:$Rd, (udiv rGPR:$Rn, rGPR:$Rm))]>, + Requires<[HasDivideInThumb, IsThumb, HasV8MBaseline]>, + Sched<[WriteDIV]> { + let Inst{31-27} = 0b11111; + let Inst{26-21} = 0b011101; + let Inst{20} = 0b1; + let Inst{15-12} = 0b1111; + let Inst{7-4} = 0b1111; +} + +//===----------------------------------------------------------------------===// +// Misc. Arithmetic Instructions. +// + +class T2I_misc<bits<2> op1, bits<2> op2, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, list<dag> pattern> + : T2ThreeReg<oops, iops, itin, opc, asm, pattern> { + let Inst{31-27} = 0b11111; + let Inst{26-22} = 0b01010; + let Inst{21-20} = op1; + let Inst{15-12} = 0b1111; + let Inst{7-6} = 0b10; + let Inst{5-4} = op2; + let Rn{3-0} = Rm; +} + +def t2CLZ : T2I_misc<0b11, 0b00, (outs rGPR:$Rd), (ins rGPR:$Rm), IIC_iUNAr, + "clz", "\t$Rd, $Rm", [(set rGPR:$Rd, (ctlz rGPR:$Rm))]>, + Sched<[WriteALU]>; + +def t2RBIT : T2I_misc<0b01, 0b10, (outs rGPR:$Rd), (ins rGPR:$Rm), IIC_iUNAr, + "rbit", "\t$Rd, $Rm", + [(set rGPR:$Rd, (bitreverse rGPR:$Rm))]>, + Sched<[WriteALU]>; + +def t2REV : T2I_misc<0b01, 0b00, (outs rGPR:$Rd), (ins rGPR:$Rm), IIC_iUNAr, + "rev", ".w\t$Rd, $Rm", [(set rGPR:$Rd, (bswap rGPR:$Rm))]>, + Sched<[WriteALU]>; + +def t2REV16 : T2I_misc<0b01, 0b01, (outs rGPR:$Rd), (ins rGPR:$Rm), IIC_iUNAr, + "rev16", ".w\t$Rd, $Rm", + [(set rGPR:$Rd, (rotr (bswap rGPR:$Rm), (i32 16)))]>, + Sched<[WriteALU]>; + +def t2REVSH : T2I_misc<0b01, 0b11, (outs rGPR:$Rd), (ins rGPR:$Rm), IIC_iUNAr, + "revsh", ".w\t$Rd, $Rm", + [(set rGPR:$Rd, (sra (bswap rGPR:$Rm), (i32 16)))]>, + Sched<[WriteALU]>; + +def : T2Pat<(or (sra (shl rGPR:$Rm, (i32 24)), (i32 16)), + (and (srl rGPR:$Rm, (i32 8)), 0xFF)), + (t2REVSH rGPR:$Rm)>; + +def t2PKHBT : T2ThreeReg< + (outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm, pkh_lsl_amt:$sh), + IIC_iBITsi, "pkhbt", "\t$Rd, $Rn, $Rm$sh", + [(set rGPR:$Rd, (or (and rGPR:$Rn, 0xFFFF), + (and (shl rGPR:$Rm, pkh_lsl_amt:$sh), + 0xFFFF0000)))]>, + Requires<[HasDSP, IsThumb2]>, + Sched<[WriteALUsi, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-20} = 0b01100; + let Inst{5} = 0; // BT form + let Inst{4} = 0; + + bits<5> sh; + let Inst{14-12} = sh{4-2}; + let Inst{7-6} = sh{1-0}; +} + +// Alternate cases for PKHBT where identities eliminate some nodes. +def : T2Pat<(or (and rGPR:$src1, 0xFFFF), (and rGPR:$src2, 0xFFFF0000)), + (t2PKHBT rGPR:$src1, rGPR:$src2, 0)>, + Requires<[HasDSP, IsThumb2]>; +def : T2Pat<(or (and rGPR:$src1, 0xFFFF), (shl rGPR:$src2, imm16_31:$sh)), + (t2PKHBT rGPR:$src1, rGPR:$src2, imm16_31:$sh)>, + Requires<[HasDSP, IsThumb2]>; + +// Note: Shifts of 1-15 bits will be transformed to srl instead of sra and +// will match the pattern below. +def t2PKHTB : T2ThreeReg< + (outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm, pkh_asr_amt:$sh), + IIC_iBITsi, "pkhtb", "\t$Rd, $Rn, $Rm$sh", + [(set rGPR:$Rd, (or (and rGPR:$Rn, 0xFFFF0000), + (and (sra rGPR:$Rm, pkh_asr_amt:$sh), + 0xFFFF)))]>, + Requires<[HasDSP, IsThumb2]>, + Sched<[WriteALUsi, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-20} = 0b01100; + let Inst{5} = 1; // TB form + let Inst{4} = 0; + + bits<5> sh; + let Inst{14-12} = sh{4-2}; + let Inst{7-6} = sh{1-0}; +} + +// Alternate cases for PKHTB where identities eliminate some nodes. Note that +// a shift amount of 0 is *not legal* here, it is PKHBT instead. +// We also can not replace a srl (17..31) by an arithmetic shift we would use in +// pkhtb src1, src2, asr (17..31). +def : T2Pat<(or (and rGPR:$src1, 0xFFFF0000), (srl rGPR:$src2, imm16:$sh)), + (t2PKHTB rGPR:$src1, rGPR:$src2, imm16:$sh)>, + Requires<[HasDSP, IsThumb2]>; +def : T2Pat<(or (and rGPR:$src1, 0xFFFF0000), (sra rGPR:$src2, imm16_31:$sh)), + (t2PKHTB rGPR:$src1, rGPR:$src2, imm16_31:$sh)>, + Requires<[HasDSP, IsThumb2]>; +def : T2Pat<(or (and rGPR:$src1, 0xFFFF0000), + (and (srl rGPR:$src2, imm1_15:$sh), 0xFFFF)), + (t2PKHTB rGPR:$src1, rGPR:$src2, imm1_15:$sh)>, + Requires<[HasDSP, IsThumb2]>; + +//===----------------------------------------------------------------------===// +// CRC32 Instructions +// +// Polynomials: +// + CRC32{B,H,W} 0x04C11DB7 +// + CRC32C{B,H,W} 0x1EDC6F41 +// + +class T2I_crc32<bit C, bits<2> sz, string suffix, SDPatternOperator builtin> + : T2ThreeRegNoP<(outs rGPR:$Rd), (ins rGPR:$Rn, rGPR:$Rm), NoItinerary, + !strconcat("crc32", suffix, "\t$Rd, $Rn, $Rm"), + [(set rGPR:$Rd, (builtin rGPR:$Rn, rGPR:$Rm))]>, + Requires<[IsThumb2, HasV8, HasCRC]> { + let Inst{31-27} = 0b11111; + let Inst{26-21} = 0b010110; + let Inst{20} = C; + let Inst{15-12} = 0b1111; + let Inst{7-6} = 0b10; + let Inst{5-4} = sz; +} + +def t2CRC32B : T2I_crc32<0, 0b00, "b", int_arm_crc32b>; +def t2CRC32CB : T2I_crc32<1, 0b00, "cb", int_arm_crc32cb>; +def t2CRC32H : T2I_crc32<0, 0b01, "h", int_arm_crc32h>; +def t2CRC32CH : T2I_crc32<1, 0b01, "ch", int_arm_crc32ch>; +def t2CRC32W : T2I_crc32<0, 0b10, "w", int_arm_crc32w>; +def t2CRC32CW : T2I_crc32<1, 0b10, "cw", int_arm_crc32cw>; + +//===----------------------------------------------------------------------===// +// Comparison Instructions... +// +defm t2CMP : T2I_cmp_irs<0b1101, "cmp", + IIC_iCMPi, IIC_iCMPr, IIC_iCMPsi, ARMcmp>; + +def : T2Pat<(ARMcmpZ GPRnopc:$lhs, t2_so_imm:$imm), + (t2CMPri GPRnopc:$lhs, t2_so_imm:$imm)>; +def : T2Pat<(ARMcmpZ GPRnopc:$lhs, rGPR:$rhs), + (t2CMPrr GPRnopc:$lhs, rGPR:$rhs)>; +def : T2Pat<(ARMcmpZ GPRnopc:$lhs, t2_so_reg:$rhs), + (t2CMPrs GPRnopc:$lhs, t2_so_reg:$rhs)>; + +let isCompare = 1, Defs = [CPSR] in { + // shifted imm + def t2CMNri : T2OneRegCmpImm< + (outs), (ins GPRnopc:$Rn, t2_so_imm:$imm), IIC_iCMPi, + "cmn", ".w\t$Rn, $imm", + [(ARMcmn GPRnopc:$Rn, (ineg t2_so_imm:$imm))]>, + Sched<[WriteCMP, ReadALU]> { + let Inst{31-27} = 0b11110; + let Inst{25} = 0; + let Inst{24-21} = 0b1000; + let Inst{20} = 1; // The S bit. + let Inst{15} = 0; + let Inst{11-8} = 0b1111; // Rd + } + // register + def t2CMNzrr : T2TwoRegCmp< + (outs), (ins GPRnopc:$Rn, rGPR:$Rm), IIC_iCMPr, + "cmn", ".w\t$Rn, $Rm", + [(BinOpFrag<(ARMcmpZ node:$LHS,(ineg node:$RHS))> + GPRnopc:$Rn, rGPR:$Rm)]>, Sched<[WriteCMP, ReadALU, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = 0b1000; + let Inst{20} = 1; // The S bit. + let Inst{14-12} = 0b000; // imm3 + let Inst{11-8} = 0b1111; // Rd + let Inst{7-6} = 0b00; // imm2 + let Inst{5-4} = 0b00; // type + } + // shifted register + def t2CMNzrs : T2OneRegCmpShiftedReg< + (outs), (ins GPRnopc:$Rn, t2_so_reg:$ShiftedRm), IIC_iCMPsi, + "cmn", ".w\t$Rn, $ShiftedRm", + [(BinOpFrag<(ARMcmpZ node:$LHS,(ineg node:$RHS))> + GPRnopc:$Rn, t2_so_reg:$ShiftedRm)]>, + Sched<[WriteCMPsi, ReadALU, ReadALU]> { + let Inst{31-27} = 0b11101; + let Inst{26-25} = 0b01; + let Inst{24-21} = 0b1000; + let Inst{20} = 1; // The S bit. + let Inst{11-8} = 0b1111; // Rd + } +} + +// Assembler aliases w/o the ".w" suffix. +// No alias here for 'rr' version as not all instantiations of this multiclass +// want one (CMP in particular, does not). +def : t2InstAlias<"cmn${p} $Rn, $imm", + (t2CMNri GPRnopc:$Rn, t2_so_imm:$imm, pred:$p)>; +def : t2InstAlias<"cmn${p} $Rn, $shift", + (t2CMNzrs GPRnopc:$Rn, t2_so_reg:$shift, pred:$p)>; + +def : T2Pat<(ARMcmp GPR:$src, t2_so_imm_neg:$imm), + (t2CMNri GPR:$src, t2_so_imm_neg:$imm)>; + +def : T2Pat<(ARMcmpZ GPRnopc:$src, t2_so_imm_neg:$imm), + (t2CMNri GPRnopc:$src, t2_so_imm_neg:$imm)>; + +defm t2TST : T2I_cmp_irs<0b0000, "tst", + IIC_iTSTi, IIC_iTSTr, IIC_iTSTsi, + BinOpFrag<(ARMcmpZ (and_su node:$LHS, node:$RHS), 0)>>; +defm t2TEQ : T2I_cmp_irs<0b0100, "teq", + IIC_iTSTi, IIC_iTSTr, IIC_iTSTsi, + BinOpFrag<(ARMcmpZ (xor_su node:$LHS, node:$RHS), 0)>>; + +// Conditional moves +let hasSideEffects = 0 in { + +let isCommutable = 1, isSelect = 1 in +def t2MOVCCr : t2PseudoInst<(outs rGPR:$Rd), + (ins rGPR:$false, rGPR:$Rm, cmovpred:$p), + 4, IIC_iCMOVr, + [(set rGPR:$Rd, (ARMcmov rGPR:$false, rGPR:$Rm, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + +let isMoveImm = 1 in +def t2MOVCCi + : t2PseudoInst<(outs rGPR:$Rd), + (ins rGPR:$false, t2_so_imm:$imm, cmovpred:$p), + 4, IIC_iCMOVi, + [(set rGPR:$Rd, (ARMcmov rGPR:$false,t2_so_imm:$imm, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + +let isCodeGenOnly = 1 in { +let isMoveImm = 1 in +def t2MOVCCi16 + : t2PseudoInst<(outs rGPR:$Rd), + (ins rGPR:$false, imm0_65535_expr:$imm, cmovpred:$p), + 4, IIC_iCMOVi, + [(set rGPR:$Rd, (ARMcmov rGPR:$false, imm0_65535:$imm, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + +let isMoveImm = 1 in +def t2MVNCCi + : t2PseudoInst<(outs rGPR:$Rd), + (ins rGPR:$false, t2_so_imm:$imm, cmovpred:$p), + 4, IIC_iCMOVi, + [(set rGPR:$Rd, + (ARMcmov rGPR:$false, t2_so_imm_not:$imm, + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + +class MOVCCShPseudo<SDPatternOperator opnode, Operand ty> + : t2PseudoInst<(outs rGPR:$Rd), + (ins rGPR:$false, rGPR:$Rm, i32imm:$imm, cmovpred:$p), + 4, IIC_iCMOVsi, + [(set rGPR:$Rd, (ARMcmov rGPR:$false, + (opnode rGPR:$Rm, (i32 ty:$imm)), + cmovpred:$p))]>, + RegConstraint<"$false = $Rd">, Sched<[WriteALU]>; + +def t2MOVCClsl : MOVCCShPseudo<shl, imm0_31>; +def t2MOVCClsr : MOVCCShPseudo<srl, imm_sr>; +def t2MOVCCasr : MOVCCShPseudo<sra, imm_sr>; +def t2MOVCCror : MOVCCShPseudo<rotr, imm0_31>; + +let isMoveImm = 1 in +def t2MOVCCi32imm + : t2PseudoInst<(outs rGPR:$dst), + (ins rGPR:$false, i32imm:$src, cmovpred:$p), + 8, IIC_iCMOVix2, + [(set rGPR:$dst, (ARMcmov rGPR:$false, imm:$src, + cmovpred:$p))]>, + RegConstraint<"$false = $dst">; +} // isCodeGenOnly = 1 + +} // hasSideEffects + +//===----------------------------------------------------------------------===// +// Atomic operations intrinsics +// + +// memory barriers protect the atomic sequences +let hasSideEffects = 1 in { +def t2DMB : T2I<(outs), (ins memb_opt:$opt), NoItinerary, + "dmb", "\t$opt", [(int_arm_dmb (i32 imm0_15:$opt))]>, + Requires<[IsThumb, HasDB]> { + bits<4> opt; + let Inst{31-4} = 0xf3bf8f5; + let Inst{3-0} = opt; +} + +def t2DSB : T2I<(outs), (ins memb_opt:$opt), NoItinerary, + "dsb", "\t$opt", [(int_arm_dsb (i32 imm0_15:$opt))]>, + Requires<[IsThumb, HasDB]> { + bits<4> opt; + let Inst{31-4} = 0xf3bf8f4; + let Inst{3-0} = opt; +} + +def t2ISB : T2I<(outs), (ins instsyncb_opt:$opt), NoItinerary, + "isb", "\t$opt", [(int_arm_isb (i32 imm0_15:$opt))]>, + Requires<[IsThumb, HasDB]> { + bits<4> opt; + let Inst{31-4} = 0xf3bf8f6; + let Inst{3-0} = opt; +} + +let hasNoSchedulingInfo = 1 in +def t2TSB : T2I<(outs), (ins tsb_opt:$opt), NoItinerary, + "tsb", "\t$opt", []>, Requires<[IsThumb, HasV8_4a]> { + let Inst{31-0} = 0xf3af8012; +} +} + +class T2I_ldrex<bits<4> opcod, dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, string opc, string asm, string cstr, + list<dag> pattern, bits<4> rt2 = 0b1111> + : Thumb2I<oops, iops, am, sz, itin, opc, asm, cstr, pattern> { + let Inst{31-27} = 0b11101; + let Inst{26-20} = 0b0001101; + let Inst{11-8} = rt2; + let Inst{7-4} = opcod; + let Inst{3-0} = 0b1111; + + bits<4> addr; + bits<4> Rt; + let Inst{19-16} = addr; + let Inst{15-12} = Rt; +} +class T2I_strex<bits<4> opcod, dag oops, dag iops, AddrMode am, int sz, + InstrItinClass itin, string opc, string asm, string cstr, + list<dag> pattern, bits<4> rt2 = 0b1111> + : Thumb2I<oops, iops, am, sz, itin, opc, asm, cstr, pattern> { + let Inst{31-27} = 0b11101; + let Inst{26-20} = 0b0001100; + let Inst{11-8} = rt2; + let Inst{7-4} = opcod; + + bits<4> Rd; + bits<4> addr; + bits<4> Rt; + let Inst{3-0} = Rd; + let Inst{19-16} = addr; + let Inst{15-12} = Rt; +} + +let mayLoad = 1 in { +def t2LDREXB : T2I_ldrex<0b0100, (outs rGPR:$Rt), (ins addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "ldrexb", "\t$Rt, $addr", "", + [(set rGPR:$Rt, (ldrex_1 addr_offset_none:$addr))]>, + Requires<[IsThumb, HasV8MBaseline]>; +def t2LDREXH : T2I_ldrex<0b0101, (outs rGPR:$Rt), (ins addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "ldrexh", "\t$Rt, $addr", "", + [(set rGPR:$Rt, (ldrex_2 addr_offset_none:$addr))]>, + Requires<[IsThumb, HasV8MBaseline]>; +def t2LDREX : Thumb2I<(outs rGPR:$Rt), (ins t2addrmode_imm0_1020s4:$addr), + AddrModeT2_ldrex, 4, NoItinerary, + "ldrex", "\t$Rt, $addr", "", + [(set rGPR:$Rt, (ldrex_4 t2addrmode_imm0_1020s4:$addr))]>, + Requires<[IsThumb, HasV8MBaseline]> { + bits<4> Rt; + bits<12> addr; + let Inst{31-27} = 0b11101; + let Inst{26-20} = 0b0000101; + let Inst{19-16} = addr{11-8}; + let Inst{15-12} = Rt; + let Inst{11-8} = 0b1111; + let Inst{7-0} = addr{7-0}; +} +let hasExtraDefRegAllocReq = 1 in +def t2LDREXD : T2I_ldrex<0b0111, (outs rGPR:$Rt, rGPR:$Rt2), + (ins addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "ldrexd", "\t$Rt, $Rt2, $addr", "", + [], {?, ?, ?, ?}>, + Requires<[IsThumb2, IsNotMClass]> { + bits<4> Rt2; + let Inst{11-8} = Rt2; +} +def t2LDAEXB : T2I_ldrex<0b1100, (outs rGPR:$Rt), (ins addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "ldaexb", "\t$Rt, $addr", "", + [(set rGPR:$Rt, (ldaex_1 addr_offset_none:$addr))]>, + Requires<[IsThumb, HasAcquireRelease, HasV7Clrex]>; +def t2LDAEXH : T2I_ldrex<0b1101, (outs rGPR:$Rt), (ins addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "ldaexh", "\t$Rt, $addr", "", + [(set rGPR:$Rt, (ldaex_2 addr_offset_none:$addr))]>, + Requires<[IsThumb, HasAcquireRelease, HasV7Clrex]>; +def t2LDAEX : Thumb2I<(outs rGPR:$Rt), (ins addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "ldaex", "\t$Rt, $addr", "", + [(set rGPR:$Rt, (ldaex_4 addr_offset_none:$addr))]>, + Requires<[IsThumb, HasAcquireRelease, HasV7Clrex]> { + bits<4> Rt; + bits<4> addr; + let Inst{31-27} = 0b11101; + let Inst{26-20} = 0b0001101; + let Inst{19-16} = addr; + let Inst{15-12} = Rt; + let Inst{11-8} = 0b1111; + let Inst{7-0} = 0b11101111; +} +let hasExtraDefRegAllocReq = 1 in +def t2LDAEXD : T2I_ldrex<0b1111, (outs rGPR:$Rt, rGPR:$Rt2), + (ins addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "ldaexd", "\t$Rt, $Rt2, $addr", "", + [], {?, ?, ?, ?}>, Requires<[IsThumb, + HasAcquireRelease, HasV7Clrex, IsNotMClass]> { + bits<4> Rt2; + let Inst{11-8} = Rt2; + + let Inst{7} = 1; +} +} + +let mayStore = 1, Constraints = "@earlyclobber $Rd" in { +def t2STREXB : T2I_strex<0b0100, (outs rGPR:$Rd), + (ins rGPR:$Rt, addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "strexb", "\t$Rd, $Rt, $addr", "", + [(set rGPR:$Rd, + (strex_1 rGPR:$Rt, addr_offset_none:$addr))]>, + Requires<[IsThumb, HasV8MBaseline]>; +def t2STREXH : T2I_strex<0b0101, (outs rGPR:$Rd), + (ins rGPR:$Rt, addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "strexh", "\t$Rd, $Rt, $addr", "", + [(set rGPR:$Rd, + (strex_2 rGPR:$Rt, addr_offset_none:$addr))]>, + Requires<[IsThumb, HasV8MBaseline]>; + +def t2STREX : Thumb2I<(outs rGPR:$Rd), (ins rGPR:$Rt, + t2addrmode_imm0_1020s4:$addr), + AddrModeT2_ldrex, 4, NoItinerary, + "strex", "\t$Rd, $Rt, $addr", "", + [(set rGPR:$Rd, + (strex_4 rGPR:$Rt, t2addrmode_imm0_1020s4:$addr))]>, + Requires<[IsThumb, HasV8MBaseline]> { + bits<4> Rd; + bits<4> Rt; + bits<12> addr; + let Inst{31-27} = 0b11101; + let Inst{26-20} = 0b0000100; + let Inst{19-16} = addr{11-8}; + let Inst{15-12} = Rt; + let Inst{11-8} = Rd; + let Inst{7-0} = addr{7-0}; +} +let hasExtraSrcRegAllocReq = 1 in +def t2STREXD : T2I_strex<0b0111, (outs rGPR:$Rd), + (ins rGPR:$Rt, rGPR:$Rt2, addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "strexd", "\t$Rd, $Rt, $Rt2, $addr", "", [], + {?, ?, ?, ?}>, + Requires<[IsThumb2, IsNotMClass]> { + bits<4> Rt2; + let Inst{11-8} = Rt2; +} +def t2STLEXB : T2I_strex<0b1100, (outs rGPR:$Rd), + (ins rGPR:$Rt, addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "stlexb", "\t$Rd, $Rt, $addr", "", + [(set rGPR:$Rd, + (stlex_1 rGPR:$Rt, addr_offset_none:$addr))]>, + Requires<[IsThumb, HasAcquireRelease, + HasV7Clrex]>; + +def t2STLEXH : T2I_strex<0b1101, (outs rGPR:$Rd), + (ins rGPR:$Rt, addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "stlexh", "\t$Rd, $Rt, $addr", "", + [(set rGPR:$Rd, + (stlex_2 rGPR:$Rt, addr_offset_none:$addr))]>, + Requires<[IsThumb, HasAcquireRelease, + HasV7Clrex]>; + +def t2STLEX : Thumb2I<(outs rGPR:$Rd), (ins rGPR:$Rt, + addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "stlex", "\t$Rd, $Rt, $addr", "", + [(set rGPR:$Rd, + (stlex_4 rGPR:$Rt, addr_offset_none:$addr))]>, + Requires<[IsThumb, HasAcquireRelease, HasV7Clrex]> { + bits<4> Rd; + bits<4> Rt; + bits<4> addr; + let Inst{31-27} = 0b11101; + let Inst{26-20} = 0b0001100; + let Inst{19-16} = addr; + let Inst{15-12} = Rt; + let Inst{11-4} = 0b11111110; + let Inst{3-0} = Rd; +} +let hasExtraSrcRegAllocReq = 1 in +def t2STLEXD : T2I_strex<0b1111, (outs rGPR:$Rd), + (ins rGPR:$Rt, rGPR:$Rt2, addr_offset_none:$addr), + AddrModeNone, 4, NoItinerary, + "stlexd", "\t$Rd, $Rt, $Rt2, $addr", "", [], + {?, ?, ?, ?}>, Requires<[IsThumb, HasAcquireRelease, + HasV7Clrex, IsNotMClass]> { + bits<4> Rt2; + let Inst{11-8} = Rt2; +} +} + +def t2CLREX : T2I<(outs), (ins), NoItinerary, "clrex", "", [(int_arm_clrex)]>, + Requires<[IsThumb, HasV7Clrex]> { + let Inst{31-16} = 0xf3bf; + let Inst{15-14} = 0b10; + let Inst{13} = 0; + let Inst{12} = 0; + let Inst{11-8} = 0b1111; + let Inst{7-4} = 0b0010; + let Inst{3-0} = 0b1111; +} + +def : T2Pat<(and (ldrex_1 addr_offset_none:$addr), 0xff), + (t2LDREXB addr_offset_none:$addr)>, + Requires<[IsThumb, HasV8MBaseline]>; +def : T2Pat<(and (ldrex_2 addr_offset_none:$addr), 0xffff), + (t2LDREXH addr_offset_none:$addr)>, + Requires<[IsThumb, HasV8MBaseline]>; +def : T2Pat<(strex_1 (and GPR:$Rt, 0xff), addr_offset_none:$addr), + (t2STREXB GPR:$Rt, addr_offset_none:$addr)>, + Requires<[IsThumb, HasV8MBaseline]>; +def : T2Pat<(strex_2 (and GPR:$Rt, 0xffff), addr_offset_none:$addr), + (t2STREXH GPR:$Rt, addr_offset_none:$addr)>, + Requires<[IsThumb, HasV8MBaseline]>; + +def : T2Pat<(and (ldaex_1 addr_offset_none:$addr), 0xff), + (t2LDAEXB addr_offset_none:$addr)>, + Requires<[IsThumb, HasAcquireRelease, HasV7Clrex]>; +def : T2Pat<(and (ldaex_2 addr_offset_none:$addr), 0xffff), + (t2LDAEXH addr_offset_none:$addr)>, + Requires<[IsThumb, HasAcquireRelease, HasV7Clrex]>; +def : T2Pat<(stlex_1 (and GPR:$Rt, 0xff), addr_offset_none:$addr), + (t2STLEXB GPR:$Rt, addr_offset_none:$addr)>, + Requires<[IsThumb, HasAcquireRelease, HasV7Clrex]>; +def : T2Pat<(stlex_2 (and GPR:$Rt, 0xffff), addr_offset_none:$addr), + (t2STLEXH GPR:$Rt, addr_offset_none:$addr)>, + Requires<[IsThumb, HasAcquireRelease, HasV7Clrex]>; + +//===----------------------------------------------------------------------===// +// SJLJ Exception handling intrinsics +// eh_sjlj_setjmp() is an instruction sequence to store the return +// address and save #0 in R0 for the non-longjmp case. +// Since by its nature we may be coming from some other function to get +// here, and we're using the stack frame for the containing function to +// save/restore registers, we can't keep anything live in regs across +// the eh_sjlj_setjmp(), else it will almost certainly have been tromped upon +// when we get here from a longjmp(). We force everything out of registers +// except for our own input by listing the relevant registers in Defs. By +// doing so, we also cause the prologue/epilogue code to actively preserve +// all of the callee-saved resgisters, which is exactly what we want. +// $val is a scratch register for our use. +let Defs = + [ R0, R1, R2, R3, R4, R5, R6, R7, R8, R9, R10, R11, R12, LR, CPSR, + Q0, Q1, Q2, Q3, Q8, Q9, Q10, Q11, Q12, Q13, Q14, Q15], + hasSideEffects = 1, isBarrier = 1, isCodeGenOnly = 1, + usesCustomInserter = 1 in { + def t2Int_eh_sjlj_setjmp : Thumb2XI<(outs), (ins tGPR:$src, tGPR:$val), + AddrModeNone, 0, NoItinerary, "", "", + [(set R0, (ARMeh_sjlj_setjmp tGPR:$src, tGPR:$val))]>, + Requires<[IsThumb2, HasVFP2]>; +} + +let Defs = + [ R0, R1, R2, R3, R4, R5, R6, R7, R8, R9, R10, R11, R12, LR, CPSR ], + hasSideEffects = 1, isBarrier = 1, isCodeGenOnly = 1, + usesCustomInserter = 1 in { + def t2Int_eh_sjlj_setjmp_nofp : Thumb2XI<(outs), (ins tGPR:$src, tGPR:$val), + AddrModeNone, 0, NoItinerary, "", "", + [(set R0, (ARMeh_sjlj_setjmp tGPR:$src, tGPR:$val))]>, + Requires<[IsThumb2, NoVFP]>; +} + + +//===----------------------------------------------------------------------===// +// Control-Flow Instructions +// + +// FIXME: remove when we have a way to marking a MI with these properties. +// FIXME: Should pc be an implicit operand like PICADD, etc? +let isReturn = 1, isTerminator = 1, isBarrier = 1, mayLoad = 1, + hasExtraDefRegAllocReq = 1, isCodeGenOnly = 1 in +def t2LDMIA_RET: t2PseudoExpand<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, + reglist:$regs, variable_ops), + 4, IIC_iLoad_mBr, [], + (t2LDMIA_UPD GPR:$wb, GPR:$Rn, pred:$p, reglist:$regs)>, + RegConstraint<"$Rn = $wb">; + +let isBranch = 1, isTerminator = 1, isBarrier = 1 in { +let isPredicable = 1 in +def t2B : T2I<(outs), (ins thumb_br_target:$target), IIC_Br, + "b", ".w\t$target", + [(br bb:$target)]>, Sched<[WriteBr]>, + Requires<[IsThumb, HasV8MBaseline]> { + let Inst{31-27} = 0b11110; + let Inst{15-14} = 0b10; + let Inst{12} = 1; + + bits<24> target; + let Inst{26} = target{23}; + let Inst{13} = target{22}; + let Inst{11} = target{21}; + let Inst{25-16} = target{20-11}; + let Inst{10-0} = target{10-0}; + let DecoderMethod = "DecodeT2BInstruction"; + let AsmMatchConverter = "cvtThumbBranches"; +} + +let Size = 4, isNotDuplicable = 1, isBranch = 1, isTerminator = 1, + isBarrier = 1, isIndirectBranch = 1 in { + +// available in both v8-M.Baseline and Thumb2 targets +def t2BR_JT : t2basePseudoInst<(outs), + (ins GPR:$target, GPR:$index, i32imm:$jt), + 0, IIC_Br, + [(ARMbr2jt GPR:$target, GPR:$index, tjumptable:$jt)]>, + Sched<[WriteBr]>; + +// FIXME: Add a case that can be predicated. +def t2TBB_JT : t2PseudoInst<(outs), + (ins GPR:$base, GPR:$index, i32imm:$jt, i32imm:$pclbl), 0, IIC_Br, []>, + Sched<[WriteBr]>; + +def t2TBH_JT : t2PseudoInst<(outs), + (ins GPR:$base, GPR:$index, i32imm:$jt, i32imm:$pclbl), 0, IIC_Br, []>, + Sched<[WriteBr]>; + +def t2TBB : T2I<(outs), (ins addrmode_tbb:$addr), IIC_Br, + "tbb", "\t$addr", []>, Sched<[WriteBrTbl]> { + bits<4> Rn; + bits<4> Rm; + let Inst{31-20} = 0b111010001101; + let Inst{19-16} = Rn; + let Inst{15-5} = 0b11110000000; + let Inst{4} = 0; // B form + let Inst{3-0} = Rm; + + let DecoderMethod = "DecodeThumbTableBranch"; +} + +def t2TBH : T2I<(outs), (ins addrmode_tbh:$addr), IIC_Br, + "tbh", "\t$addr", []>, Sched<[WriteBrTbl]> { + bits<4> Rn; + bits<4> Rm; + let Inst{31-20} = 0b111010001101; + let Inst{19-16} = Rn; + let Inst{15-5} = 0b11110000000; + let Inst{4} = 1; // H form + let Inst{3-0} = Rm; + + let DecoderMethod = "DecodeThumbTableBranch"; +} +} // isNotDuplicable, isIndirectBranch + +} // isBranch, isTerminator, isBarrier + +// FIXME: should be able to write a pattern for ARMBrcond, but can't use +// a two-value operand where a dag node expects ", "two operands. :( +let isBranch = 1, isTerminator = 1 in +def t2Bcc : T2I<(outs), (ins brtarget:$target), IIC_Br, + "b", ".w\t$target", + [/*(ARMbrcond bb:$target, imm:$cc)*/]>, Sched<[WriteBr]> { + let Inst{31-27} = 0b11110; + let Inst{15-14} = 0b10; + let Inst{12} = 0; + + bits<4> p; + let Inst{25-22} = p; + + bits<21> target; + let Inst{26} = target{20}; + let Inst{11} = target{19}; + let Inst{13} = target{18}; + let Inst{21-16} = target{17-12}; + let Inst{10-0} = target{11-1}; + + let DecoderMethod = "DecodeThumb2BCCInstruction"; + let AsmMatchConverter = "cvtThumbBranches"; +} + +// Tail calls. The MachO version of thumb tail calls uses a t2 branch, so +// it goes here. +let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1 in { + // IOS version. + let Uses = [SP] in + def tTAILJMPd: tPseudoExpand<(outs), + (ins thumb_br_target:$dst, pred:$p), + 4, IIC_Br, [], + (t2B thumb_br_target:$dst, pred:$p)>, + Requires<[IsThumb2, IsMachO]>, Sched<[WriteBr]>; +} + +// IT block +let Defs = [ITSTATE] in +def t2IT : Thumb2XI<(outs), (ins it_pred:$cc, it_mask:$mask), + AddrModeNone, 2, IIC_iALUx, + "it$mask\t$cc", "", []>, + ComplexDeprecationPredicate<"IT"> { + // 16-bit instruction. + let Inst{31-16} = 0x0000; + let Inst{15-8} = 0b10111111; + + bits<4> cc; + bits<4> mask; + let Inst{7-4} = cc; + let Inst{3-0} = mask; + + let DecoderMethod = "DecodeIT"; +} + +// Branch and Exchange Jazelle -- for disassembly only +// Rm = Inst{19-16} +let isBranch = 1, isTerminator = 1, isBarrier = 1, isIndirectBranch = 1 in +def t2BXJ : T2I<(outs), (ins GPRnopc:$func), NoItinerary, "bxj", "\t$func", []>, + Sched<[WriteBr]>, Requires<[IsThumb2, IsNotMClass]> { + bits<4> func; + let Inst{31-27} = 0b11110; + let Inst{26} = 0; + let Inst{25-20} = 0b111100; + let Inst{19-16} = func; + let Inst{15-0} = 0b1000111100000000; +} + +// Compare and branch on zero / non-zero +let isBranch = 1, isTerminator = 1 in { + def tCBZ : T1I<(outs), (ins tGPR:$Rn, thumb_cb_target:$target), IIC_Br, + "cbz\t$Rn, $target", []>, + T1Misc<{0,0,?,1,?,?,?}>, + Requires<[IsThumb, HasV8MBaseline]>, Sched<[WriteBr]> { + // A8.6.27 + bits<6> target; + bits<3> Rn; + let Inst{9} = target{5}; + let Inst{7-3} = target{4-0}; + let Inst{2-0} = Rn; + } + + def tCBNZ : T1I<(outs), (ins tGPR:$Rn, thumb_cb_target:$target), IIC_Br, + "cbnz\t$Rn, $target", []>, + T1Misc<{1,0,?,1,?,?,?}>, + Requires<[IsThumb, HasV8MBaseline]>, Sched<[WriteBr]> { + // A8.6.27 + bits<6> target; + bits<3> Rn; + let Inst{9} = target{5}; + let Inst{7-3} = target{4-0}; + let Inst{2-0} = Rn; + } +} + + +// Change Processor State is a system instruction. +// FIXME: Since the asm parser has currently no clean way to handle optional +// operands, create 3 versions of the same instruction. Once there's a clean +// framework to represent optional operands, change this behavior. +class t2CPS<dag iops, string asm_op> : T2XI<(outs), iops, NoItinerary, + !strconcat("cps", asm_op), []>, + Requires<[IsThumb2, IsNotMClass]> { + bits<2> imod; + bits<3> iflags; + bits<5> mode; + bit M; + + let Inst{31-11} = 0b111100111010111110000; + let Inst{10-9} = imod; + let Inst{8} = M; + let Inst{7-5} = iflags; + let Inst{4-0} = mode; + let DecoderMethod = "DecodeT2CPSInstruction"; +} + +let M = 1 in + def t2CPS3p : t2CPS<(ins imod_op:$imod, iflags_op:$iflags, i32imm:$mode), + "$imod\t$iflags, $mode">; +let mode = 0, M = 0 in + def t2CPS2p : t2CPS<(ins imod_op:$imod, iflags_op:$iflags), + "$imod.w\t$iflags">; +let imod = 0, iflags = 0, M = 1 in + def t2CPS1p : t2CPS<(ins imm0_31:$mode), "\t$mode">; + +def : t2InstAlias<"cps$imod.w $iflags, $mode", + (t2CPS3p imod_op:$imod, iflags_op:$iflags, i32imm:$mode), 0>; +def : t2InstAlias<"cps.w $mode", (t2CPS1p imm0_31:$mode), 0>; + +// A6.3.4 Branches and miscellaneous control +// Table A6-14 Change Processor State, and hint instructions +def t2HINT : T2I<(outs), (ins imm0_239:$imm), NoItinerary, "hint", ".w\t$imm", + [(int_arm_hint imm0_239:$imm)]> { + bits<8> imm; + let Inst{31-3} = 0b11110011101011111000000000000; + let Inst{7-0} = imm; +} + +def : t2InstAlias<"hint$p $imm", (t2HINT imm0_239:$imm, pred:$p), 0>; +def : t2InstAlias<"nop$p.w", (t2HINT 0, pred:$p), 1>; +def : t2InstAlias<"yield$p.w", (t2HINT 1, pred:$p), 1>; +def : t2InstAlias<"wfe$p.w", (t2HINT 2, pred:$p), 1>; +def : t2InstAlias<"wfi$p.w", (t2HINT 3, pred:$p), 1>; +def : t2InstAlias<"sev$p.w", (t2HINT 4, pred:$p), 1>; +def : t2InstAlias<"sevl$p.w", (t2HINT 5, pred:$p), 1> { + let Predicates = [IsThumb2, HasV8]; +} +def : t2InstAlias<"esb$p.w", (t2HINT 16, pred:$p), 1> { + let Predicates = [IsThumb2, HasRAS]; +} +def : t2InstAlias<"esb$p", (t2HINT 16, pred:$p), 0> { + let Predicates = [IsThumb2, HasRAS]; +} +def : t2InstAlias<"csdb$p.w", (t2HINT 20, pred:$p), 0>; +def : t2InstAlias<"csdb$p", (t2HINT 20, pred:$p), 1>; + +def t2DBG : T2I<(outs), (ins imm0_15:$opt), NoItinerary, "dbg", "\t$opt", + [(int_arm_dbg imm0_15:$opt)]> { + bits<4> opt; + let Inst{31-20} = 0b111100111010; + let Inst{19-16} = 0b1111; + let Inst{15-8} = 0b10000000; + let Inst{7-4} = 0b1111; + let Inst{3-0} = opt; +} + +// Secure Monitor Call is a system instruction. +// Option = Inst{19-16} +let isCall = 1, Uses = [SP] in +def t2SMC : T2I<(outs), (ins imm0_15:$opt), NoItinerary, "smc", "\t$opt", + []>, Requires<[IsThumb2, HasTrustZone]> { + let Inst{31-27} = 0b11110; + let Inst{26-20} = 0b1111111; + let Inst{15-12} = 0b1000; + + bits<4> opt; + let Inst{19-16} = opt; +} + +class T2DCPS<bits<2> opt, string opc> + : T2I<(outs), (ins), NoItinerary, opc, "", []>, Requires<[IsThumb2, HasV8]> { + let Inst{31-27} = 0b11110; + let Inst{26-20} = 0b1111000; + let Inst{19-16} = 0b1111; + let Inst{15-12} = 0b1000; + let Inst{11-2} = 0b0000000000; + let Inst{1-0} = opt; +} + +def t2DCPS1 : T2DCPS<0b01, "dcps1">; +def t2DCPS2 : T2DCPS<0b10, "dcps2">; +def t2DCPS3 : T2DCPS<0b11, "dcps3">; + +class T2SRS<bits<2> Op, bit W, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern>, + Requires<[IsThumb2,IsNotMClass]> { + bits<5> mode; + let Inst{31-25} = 0b1110100; + let Inst{24-23} = Op; + let Inst{22} = 0; + let Inst{21} = W; + let Inst{20-16} = 0b01101; + let Inst{15-5} = 0b11000000000; + let Inst{4-0} = mode{4-0}; +} + +// Store Return State is a system instruction. +def t2SRSDB_UPD : T2SRS<0b00, 1, (outs), (ins imm0_31:$mode), NoItinerary, + "srsdb", "\tsp!, $mode", []>; +def t2SRSDB : T2SRS<0b00, 0, (outs), (ins imm0_31:$mode), NoItinerary, + "srsdb","\tsp, $mode", []>; +def t2SRSIA_UPD : T2SRS<0b11, 1, (outs), (ins imm0_31:$mode), NoItinerary, + "srsia","\tsp!, $mode", []>; +def t2SRSIA : T2SRS<0b11, 0, (outs), (ins imm0_31:$mode), NoItinerary, + "srsia","\tsp, $mode", []>; + + +def : t2InstAlias<"srsdb${p} $mode", (t2SRSDB imm0_31:$mode, pred:$p)>; +def : t2InstAlias<"srsdb${p} $mode!", (t2SRSDB_UPD imm0_31:$mode, pred:$p)>; + +def : t2InstAlias<"srsia${p} $mode", (t2SRSIA imm0_31:$mode, pred:$p)>; +def : t2InstAlias<"srsia${p} $mode!", (t2SRSIA_UPD imm0_31:$mode, pred:$p)>; + +// Return From Exception is a system instruction. +let isReturn = 1, isBarrier = 1, isTerminator = 1, Defs = [PC] in +class T2RFE<bits<12> op31_20, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : T2I<oops, iops, itin, opc, asm, pattern>, + Requires<[IsThumb2,IsNotMClass]> { + let Inst{31-20} = op31_20{11-0}; + + bits<4> Rn; + let Inst{19-16} = Rn; + let Inst{15-0} = 0xc000; +} + +def t2RFEDBW : T2RFE<0b111010000011, + (outs), (ins GPR:$Rn), NoItinerary, "rfedb", "\t$Rn!", + [/* For disassembly only; pattern left blank */]>; +def t2RFEDB : T2RFE<0b111010000001, + (outs), (ins GPR:$Rn), NoItinerary, "rfedb", "\t$Rn", + [/* For disassembly only; pattern left blank */]>; +def t2RFEIAW : T2RFE<0b111010011011, + (outs), (ins GPR:$Rn), NoItinerary, "rfeia", "\t$Rn!", + [/* For disassembly only; pattern left blank */]>; +def t2RFEIA : T2RFE<0b111010011001, + (outs), (ins GPR:$Rn), NoItinerary, "rfeia", "\t$Rn", + [/* For disassembly only; pattern left blank */]>; + +// B9.3.19 SUBS PC, LR, #imm (Thumb2) system instruction. +// Exception return instruction is "subs pc, lr, #imm". +let isReturn = 1, isBarrier = 1, isTerminator = 1, Defs = [PC] in +def t2SUBS_PC_LR : T2I <(outs), (ins imm0_255:$imm), NoItinerary, + "subs", "\tpc, lr, $imm", + [(ARMintretflag imm0_255:$imm)]>, + Requires<[IsThumb2,IsNotMClass]> { + let Inst{31-8} = 0b111100111101111010001111; + + bits<8> imm; + let Inst{7-0} = imm; +} + +// Hypervisor Call is a system instruction. +let isCall = 1 in { +def t2HVC : T2XI <(outs), (ins imm0_65535:$imm16), IIC_Br, "hvc.w\t$imm16", []>, + Requires<[IsThumb2, HasVirtualization]>, Sched<[WriteBr]> { + bits<16> imm16; + let Inst{31-20} = 0b111101111110; + let Inst{19-16} = imm16{15-12}; + let Inst{15-12} = 0b1000; + let Inst{11-0} = imm16{11-0}; +} +} + +// Alias for HVC without the ".w" optional width specifier +def : t2InstAlias<"hvc\t$imm16", (t2HVC imm0_65535:$imm16)>; + +// ERET - Return from exception in Hypervisor mode. +// B9.3.3, B9.3.20: ERET is an alias for "SUBS PC, LR, #0" in an implementation that +// includes virtualization extensions. +def t2ERET : InstAlias<"eret${p}", (t2SUBS_PC_LR 0, pred:$p), 1>, + Requires<[IsThumb2, HasVirtualization]>; + +//===----------------------------------------------------------------------===// +// Non-Instruction Patterns +// + +// 32-bit immediate using movw + movt. +// This is a single pseudo instruction to make it re-materializable. +// FIXME: Remove this when we can do generalized remat. +let isReMaterializable = 1, isMoveImm = 1 in +def t2MOVi32imm : PseudoInst<(outs rGPR:$dst), (ins i32imm:$src), IIC_iMOVix2, + [(set rGPR:$dst, (i32 imm:$src))]>, + Requires<[IsThumb, UseMovt]>; + +// Pseudo instruction that combines movw + movt + add pc (if pic). +// It also makes it possible to rematerialize the instructions. +// FIXME: Remove this when we can do generalized remat and when machine licm +// can properly the instructions. +let isReMaterializable = 1 in { +def t2MOV_ga_pcrel : PseudoInst<(outs rGPR:$dst), (ins i32imm:$addr), + IIC_iMOVix2addpc, + [(set rGPR:$dst, (ARMWrapperPIC tglobaladdr:$addr))]>, + Requires<[IsThumb, HasV8MBaseline, UseMovtInPic]>; + +} + +def : T2Pat<(ARMWrapperPIC tglobaltlsaddr :$dst), + (t2MOV_ga_pcrel tglobaltlsaddr:$dst)>, + Requires<[IsThumb2, UseMovtInPic]>; +def : T2Pat<(ARMWrapper tglobaltlsaddr:$dst), + (t2MOVi32imm tglobaltlsaddr:$dst)>, + Requires<[IsThumb2, UseMovt]>; + +// ConstantPool, GlobalAddress, and JumpTable +def : T2Pat<(ARMWrapper tconstpool :$dst), (t2LEApcrel tconstpool :$dst)>; +def : T2Pat<(ARMWrapper texternalsym :$dst), (t2MOVi32imm texternalsym :$dst)>, + Requires<[IsThumb, HasV8MBaseline, UseMovt]>; +def : T2Pat<(ARMWrapper tglobaladdr :$dst), (t2MOVi32imm tglobaladdr :$dst)>, + Requires<[IsThumb, HasV8MBaseline, UseMovt]>; + +def : T2Pat<(ARMWrapperJT tjumptable:$dst), (t2LEApcrelJT tjumptable:$dst)>; + +// Pseudo instruction that combines ldr from constpool and add pc. This should +// be expanded into two instructions late to allow if-conversion and +// scheduling. +let canFoldAsLoad = 1, isReMaterializable = 1 in +def t2LDRpci_pic : PseudoInst<(outs rGPR:$dst), (ins i32imm:$addr, pclabel:$cp), + IIC_iLoadiALU, + [(set rGPR:$dst, (ARMpic_add (load (ARMWrapper tconstpool:$addr)), + imm:$cp))]>, + Requires<[IsThumb2]>; + +// Pseudo isntruction that combines movs + predicated rsbmi +// to implement integer ABS +let usesCustomInserter = 1, Defs = [CPSR] in { +def t2ABS : PseudoInst<(outs rGPR:$dst), (ins rGPR:$src), + NoItinerary, []>, Requires<[IsThumb2]>; +} + +//===----------------------------------------------------------------------===// +// Coprocessor load/store -- for disassembly only +// +class T2CI<bits<4> op31_28, dag oops, dag iops, string opc, string asm, list<dag> pattern> + : T2I<oops, iops, NoItinerary, opc, asm, pattern> { + let Inst{31-28} = op31_28; + let Inst{27-25} = 0b110; +} + +multiclass t2LdStCop<bits<4> op31_28, bit load, bit Dbit, string asm, list<dag> pattern> { + def _OFFSET : T2CI<op31_28, + (outs), (ins p_imm:$cop, c_imm:$CRd, addrmode5:$addr), + asm, "\t$cop, $CRd, $addr", pattern> { + bits<13> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 1; // P = 1 + let Inst{23} = addr{8}; + let Inst{22} = Dbit; + let Inst{21} = 0; // W = 0 + let Inst{20} = load; + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = addr{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _PRE : T2CI<op31_28, + (outs), (ins p_imm:$cop, c_imm:$CRd, addrmode5_pre:$addr), + asm, "\t$cop, $CRd, $addr!", []> { + bits<13> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 1; // P = 1 + let Inst{23} = addr{8}; + let Inst{22} = Dbit; + let Inst{21} = 1; // W = 1 + let Inst{20} = load; + let Inst{19-16} = addr{12-9}; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = addr{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _POST: T2CI<op31_28, + (outs), (ins p_imm:$cop, c_imm:$CRd, addr_offset_none:$addr, + postidx_imm8s4:$offset), + asm, "\t$cop, $CRd, $addr, $offset", []> { + bits<9> offset; + bits<4> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 0; // P = 0 + let Inst{23} = offset{8}; + let Inst{22} = Dbit; + let Inst{21} = 1; // W = 1 + let Inst{20} = load; + let Inst{19-16} = addr; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = offset{7-0}; + let DecoderMethod = "DecodeCopMemInstruction"; + } + def _OPTION : T2CI<op31_28, (outs), + (ins p_imm:$cop, c_imm:$CRd, addr_offset_none:$addr, + coproc_option_imm:$option), + asm, "\t$cop, $CRd, $addr, $option", []> { + bits<8> option; + bits<4> addr; + bits<4> cop; + bits<4> CRd; + let Inst{24} = 0; // P = 0 + let Inst{23} = 1; // U = 1 + let Inst{22} = Dbit; + let Inst{21} = 0; // W = 0 + let Inst{20} = load; + let Inst{19-16} = addr; + let Inst{15-12} = CRd; + let Inst{11-8} = cop; + let Inst{7-0} = option; + let DecoderMethod = "DecodeCopMemInstruction"; + } +} + +let DecoderNamespace = "Thumb2CoProc" in { +defm t2LDC : t2LdStCop<0b1110, 1, 0, "ldc", [(int_arm_ldc imm:$cop, imm:$CRd, addrmode5:$addr)]>; +defm t2LDCL : t2LdStCop<0b1110, 1, 1, "ldcl", [(int_arm_ldcl imm:$cop, imm:$CRd, addrmode5:$addr)]>; +defm t2LDC2 : t2LdStCop<0b1111, 1, 0, "ldc2", [(int_arm_ldc2 imm:$cop, imm:$CRd, addrmode5:$addr)]>, Requires<[PreV8,IsThumb2]>; +defm t2LDC2L : t2LdStCop<0b1111, 1, 1, "ldc2l", [(int_arm_ldc2l imm:$cop, imm:$CRd, addrmode5:$addr)]>, Requires<[PreV8,IsThumb2]>; + +defm t2STC : t2LdStCop<0b1110, 0, 0, "stc", [(int_arm_stc imm:$cop, imm:$CRd, addrmode5:$addr)]>; +defm t2STCL : t2LdStCop<0b1110, 0, 1, "stcl", [(int_arm_stcl imm:$cop, imm:$CRd, addrmode5:$addr)]>; +defm t2STC2 : t2LdStCop<0b1111, 0, 0, "stc2", [(int_arm_stc2 imm:$cop, imm:$CRd, addrmode5:$addr)]>, Requires<[PreV8,IsThumb2]>; +defm t2STC2L : t2LdStCop<0b1111, 0, 1, "stc2l", [(int_arm_stc2l imm:$cop, imm:$CRd, addrmode5:$addr)]>, Requires<[PreV8,IsThumb2]>; +} + + +//===----------------------------------------------------------------------===// +// Move between special register and ARM core register -- for disassembly only +// +// Move to ARM core register from Special Register + +// A/R class MRS. +// +// A/R class can only move from CPSR or SPSR. +def t2MRS_AR : T2I<(outs GPR:$Rd), (ins), NoItinerary, "mrs", "\t$Rd, apsr", + []>, Requires<[IsThumb2,IsNotMClass]> { + bits<4> Rd; + let Inst{31-12} = 0b11110011111011111000; + let Inst{11-8} = Rd; + let Inst{7-0} = 0b00000000; +} + +def : t2InstAlias<"mrs${p} $Rd, cpsr", (t2MRS_AR GPR:$Rd, pred:$p)>; + +def t2MRSsys_AR: T2I<(outs GPR:$Rd), (ins), NoItinerary, "mrs", "\t$Rd, spsr", + []>, Requires<[IsThumb2,IsNotMClass]> { + bits<4> Rd; + let Inst{31-12} = 0b11110011111111111000; + let Inst{11-8} = Rd; + let Inst{7-0} = 0b00000000; +} + +def t2MRSbanked : T2I<(outs rGPR:$Rd), (ins banked_reg:$banked), + NoItinerary, "mrs", "\t$Rd, $banked", []>, + Requires<[IsThumb, HasVirtualization]> { + bits<6> banked; + bits<4> Rd; + + let Inst{31-21} = 0b11110011111; + let Inst{20} = banked{5}; // R bit + let Inst{19-16} = banked{3-0}; + let Inst{15-12} = 0b1000; + let Inst{11-8} = Rd; + let Inst{7-5} = 0b001; + let Inst{4} = banked{4}; + let Inst{3-0} = 0b0000; +} + + +// M class MRS. +// +// This MRS has a mask field in bits 7-0 and can take more values than +// the A/R class (a full msr_mask). +def t2MRS_M : T2I<(outs rGPR:$Rd), (ins msr_mask:$SYSm), NoItinerary, + "mrs", "\t$Rd, $SYSm", []>, + Requires<[IsThumb,IsMClass]> { + bits<4> Rd; + bits<8> SYSm; + let Inst{31-12} = 0b11110011111011111000; + let Inst{11-8} = Rd; + let Inst{7-0} = SYSm; + + let Unpredictable{20-16} = 0b11111; + let Unpredictable{13} = 0b1; +} + + +// Move from ARM core register to Special Register +// +// A/R class MSR. +// +// No need to have both system and application versions, the encodings are the +// same and the assembly parser has no way to distinguish between them. The mask +// operand contains the special register (R Bit) in bit 4 and bits 3-0 contains +// the mask with the fields to be accessed in the special register. +let Defs = [CPSR] in +def t2MSR_AR : T2I<(outs), (ins msr_mask:$mask, rGPR:$Rn), + NoItinerary, "msr", "\t$mask, $Rn", []>, + Requires<[IsThumb2,IsNotMClass]> { + bits<5> mask; + bits<4> Rn; + let Inst{31-21} = 0b11110011100; + let Inst{20} = mask{4}; // R Bit + let Inst{19-16} = Rn; + let Inst{15-12} = 0b1000; + let Inst{11-8} = mask{3-0}; + let Inst{7-0} = 0; +} + +// However, the MSR (banked register) system instruction (ARMv7VE) *does* have a +// separate encoding (distinguished by bit 5. +def t2MSRbanked : T2I<(outs), (ins banked_reg:$banked, rGPR:$Rn), + NoItinerary, "msr", "\t$banked, $Rn", []>, + Requires<[IsThumb, HasVirtualization]> { + bits<6> banked; + bits<4> Rn; + + let Inst{31-21} = 0b11110011100; + let Inst{20} = banked{5}; // R bit + let Inst{19-16} = Rn; + let Inst{15-12} = 0b1000; + let Inst{11-8} = banked{3-0}; + let Inst{7-5} = 0b001; + let Inst{4} = banked{4}; + let Inst{3-0} = 0b0000; +} + + +// M class MSR. +// +// Move from ARM core register to Special Register +let Defs = [CPSR] in +def t2MSR_M : T2I<(outs), (ins msr_mask:$SYSm, rGPR:$Rn), + NoItinerary, "msr", "\t$SYSm, $Rn", []>, + Requires<[IsThumb,IsMClass]> { + bits<12> SYSm; + bits<4> Rn; + let Inst{31-21} = 0b11110011100; + let Inst{20} = 0b0; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b1000; + let Inst{11-10} = SYSm{11-10}; + let Inst{9-8} = 0b00; + let Inst{7-0} = SYSm{7-0}; + + let Unpredictable{20} = 0b1; + let Unpredictable{13} = 0b1; + let Unpredictable{9-8} = 0b11; +} + + +//===----------------------------------------------------------------------===// +// Move between coprocessor and ARM core register +// + +class t2MovRCopro<bits<4> Op, string opc, bit direction, dag oops, dag iops, + list<dag> pattern> + : T2Cop<Op, oops, iops, opc, "\t$cop, $opc1, $Rt, $CRn, $CRm, $opc2", + pattern> { + let Inst{27-24} = 0b1110; + let Inst{20} = direction; + let Inst{4} = 1; + + bits<4> Rt; + bits<4> cop; + bits<3> opc1; + bits<3> opc2; + bits<4> CRm; + bits<4> CRn; + + let Inst{15-12} = Rt; + let Inst{11-8} = cop; + let Inst{23-21} = opc1; + let Inst{7-5} = opc2; + let Inst{3-0} = CRm; + let Inst{19-16} = CRn; + + let DecoderNamespace = "Thumb2CoProc"; +} + +class t2MovRRCopro<bits<4> Op, string opc, bit direction, dag oops, dag iops, + list<dag> pattern = []> + : T2Cop<Op, oops, iops, opc, "\t$cop, $opc1, $Rt, $Rt2, $CRm", pattern> { + let Inst{27-24} = 0b1100; + let Inst{23-21} = 0b010; + let Inst{20} = direction; + + bits<4> Rt; + bits<4> Rt2; + bits<4> cop; + bits<4> opc1; + bits<4> CRm; + + let Inst{15-12} = Rt; + let Inst{19-16} = Rt2; + let Inst{11-8} = cop; + let Inst{7-4} = opc1; + let Inst{3-0} = CRm; + + let DecoderNamespace = "Thumb2CoProc"; +} + +/* from ARM core register to coprocessor */ +def t2MCR : t2MovRCopro<0b1110, "mcr", 0, + (outs), + (ins p_imm:$cop, imm0_7:$opc1, GPR:$Rt, c_imm:$CRn, + c_imm:$CRm, imm0_7:$opc2), + [(int_arm_mcr imm:$cop, imm:$opc1, GPR:$Rt, imm:$CRn, + imm:$CRm, imm:$opc2)]>, + ComplexDeprecationPredicate<"MCR">; +def : t2InstAlias<"mcr${p} $cop, $opc1, $Rt, $CRn, $CRm", + (t2MCR p_imm:$cop, imm0_7:$opc1, GPR:$Rt, c_imm:$CRn, + c_imm:$CRm, 0, pred:$p)>; +def t2MCR2 : t2MovRCopro<0b1111, "mcr2", 0, + (outs), (ins p_imm:$cop, imm0_7:$opc1, GPR:$Rt, c_imm:$CRn, + c_imm:$CRm, imm0_7:$opc2), + [(int_arm_mcr2 imm:$cop, imm:$opc1, GPR:$Rt, imm:$CRn, + imm:$CRm, imm:$opc2)]> { + let Predicates = [IsThumb2, PreV8]; +} +def : t2InstAlias<"mcr2${p} $cop, $opc1, $Rt, $CRn, $CRm", + (t2MCR2 p_imm:$cop, imm0_7:$opc1, GPR:$Rt, c_imm:$CRn, + c_imm:$CRm, 0, pred:$p)>; + +/* from coprocessor to ARM core register */ +def t2MRC : t2MovRCopro<0b1110, "mrc", 1, + (outs GPRwithAPSR:$Rt), (ins p_imm:$cop, imm0_7:$opc1, c_imm:$CRn, + c_imm:$CRm, imm0_7:$opc2), []>; +def : t2InstAlias<"mrc${p} $cop, $opc1, $Rt, $CRn, $CRm", + (t2MRC GPRwithAPSR:$Rt, p_imm:$cop, imm0_7:$opc1, c_imm:$CRn, + c_imm:$CRm, 0, pred:$p)>; + +def t2MRC2 : t2MovRCopro<0b1111, "mrc2", 1, + (outs GPRwithAPSR:$Rt), (ins p_imm:$cop, imm0_7:$opc1, c_imm:$CRn, + c_imm:$CRm, imm0_7:$opc2), []> { + let Predicates = [IsThumb2, PreV8]; +} +def : t2InstAlias<"mrc2${p} $cop, $opc1, $Rt, $CRn, $CRm", + (t2MRC2 GPRwithAPSR:$Rt, p_imm:$cop, imm0_7:$opc1, c_imm:$CRn, + c_imm:$CRm, 0, pred:$p)>; + +def : T2v6Pat<(int_arm_mrc imm:$cop, imm:$opc1, imm:$CRn, imm:$CRm, imm:$opc2), + (t2MRC imm:$cop, imm:$opc1, imm:$CRn, imm:$CRm, imm:$opc2)>; + +def : T2v6Pat<(int_arm_mrc2 imm:$cop, imm:$opc1, imm:$CRn, imm:$CRm, imm:$opc2), + (t2MRC2 imm:$cop, imm:$opc1, imm:$CRn, imm:$CRm, imm:$opc2)>; + + +/* from ARM core register to coprocessor */ +def t2MCRR : t2MovRRCopro<0b1110, "mcrr", 0, (outs), + (ins p_imm:$cop, imm0_15:$opc1, GPR:$Rt, GPR:$Rt2, + c_imm:$CRm), + [(int_arm_mcrr imm:$cop, imm:$opc1, GPR:$Rt, GPR:$Rt2, + imm:$CRm)]>; +def t2MCRR2 : t2MovRRCopro<0b1111, "mcrr2", 0, (outs), + (ins p_imm:$cop, imm0_15:$opc1, GPR:$Rt, GPR:$Rt2, + c_imm:$CRm), + [(int_arm_mcrr2 imm:$cop, imm:$opc1, GPR:$Rt, + GPR:$Rt2, imm:$CRm)]> { + let Predicates = [IsThumb2, PreV8]; +} + +/* from coprocessor to ARM core register */ +def t2MRRC : t2MovRRCopro<0b1110, "mrrc", 1, (outs GPR:$Rt, GPR:$Rt2), + (ins p_imm:$cop, imm0_15:$opc1, c_imm:$CRm)>; + +def t2MRRC2 : t2MovRRCopro<0b1111, "mrrc2", 1, (outs GPR:$Rt, GPR:$Rt2), + (ins p_imm:$cop, imm0_15:$opc1, c_imm:$CRm)> { + let Predicates = [IsThumb2, PreV8]; +} + +//===----------------------------------------------------------------------===// +// Other Coprocessor Instructions. +// + +def t2CDP : T2Cop<0b1110, (outs), (ins p_imm:$cop, imm0_15:$opc1, + c_imm:$CRd, c_imm:$CRn, c_imm:$CRm, imm0_7:$opc2), + "cdp", "\t$cop, $opc1, $CRd, $CRn, $CRm, $opc2", + [(int_arm_cdp imm:$cop, imm:$opc1, imm:$CRd, imm:$CRn, + imm:$CRm, imm:$opc2)]> { + let Inst{27-24} = 0b1110; + + bits<4> opc1; + bits<4> CRn; + bits<4> CRd; + bits<4> cop; + bits<3> opc2; + bits<4> CRm; + + let Inst{3-0} = CRm; + let Inst{4} = 0; + let Inst{7-5} = opc2; + let Inst{11-8} = cop; + let Inst{15-12} = CRd; + let Inst{19-16} = CRn; + let Inst{23-20} = opc1; + + let Predicates = [IsThumb2, PreV8]; + let DecoderNamespace = "Thumb2CoProc"; +} + +def t2CDP2 : T2Cop<0b1111, (outs), (ins p_imm:$cop, imm0_15:$opc1, + c_imm:$CRd, c_imm:$CRn, c_imm:$CRm, imm0_7:$opc2), + "cdp2", "\t$cop, $opc1, $CRd, $CRn, $CRm, $opc2", + [(int_arm_cdp2 imm:$cop, imm:$opc1, imm:$CRd, imm:$CRn, + imm:$CRm, imm:$opc2)]> { + let Inst{27-24} = 0b1110; + + bits<4> opc1; + bits<4> CRn; + bits<4> CRd; + bits<4> cop; + bits<3> opc2; + bits<4> CRm; + + let Inst{3-0} = CRm; + let Inst{4} = 0; + let Inst{7-5} = opc2; + let Inst{11-8} = cop; + let Inst{15-12} = CRd; + let Inst{19-16} = CRn; + let Inst{23-20} = opc1; + + let Predicates = [IsThumb2, PreV8]; + let DecoderNamespace = "Thumb2CoProc"; +} + + + +//===----------------------------------------------------------------------===// +// ARMv8.1 Privilege Access Never extension +// +// SETPAN #imm1 + +def t2SETPAN : T1I<(outs), (ins imm0_1:$imm), NoItinerary, "setpan\t$imm", []>, + T1Misc<0b0110000>, Requires<[IsThumb2, HasV8, HasV8_1a]> { + bits<1> imm; + + let Inst{4} = 0b1; + let Inst{3} = imm; + let Inst{2-0} = 0b000; + + let Unpredictable{4} = 0b1; + let Unpredictable{2-0} = 0b111; +} + +//===----------------------------------------------------------------------===// +// ARMv8-M Security Extensions instructions +// + +let hasSideEffects = 1 in +def t2SG : T2I<(outs), (ins), NoItinerary, "sg", "", []>, + Requires<[Has8MSecExt]> { + let Inst = 0xe97fe97f; +} + +class T2TT<bits<2> at, string asm, list<dag> pattern> + : T2I<(outs rGPR:$Rt), (ins GPRnopc:$Rn), NoItinerary, asm, "\t$Rt, $Rn", + pattern> { + bits<4> Rn; + bits<4> Rt; + + let Inst{31-20} = 0b111010000100; + let Inst{19-16} = Rn; + let Inst{15-12} = 0b1111; + let Inst{11-8} = Rt; + let Inst{7-6} = at; + let Inst{5-0} = 0b000000; + + let Unpredictable{5-0} = 0b111111; +} + +def t2TT : T2TT<0b00, "tt", []>, Requires<[IsThumb,Has8MSecExt]>; +def t2TTT : T2TT<0b01, "ttt", []>, Requires<[IsThumb,Has8MSecExt]>; +def t2TTA : T2TT<0b10, "tta", []>, Requires<[IsThumb,Has8MSecExt]>; +def t2TTAT : T2TT<0b11, "ttat", []>, Requires<[IsThumb,Has8MSecExt]>; + +//===----------------------------------------------------------------------===// +// Non-Instruction Patterns +// + +// SXT/UXT with no rotate +let AddedComplexity = 16 in { +def : T2Pat<(and rGPR:$Rm, 0x000000FF), (t2UXTB rGPR:$Rm, 0)>, + Requires<[IsThumb2]>; +def : T2Pat<(and rGPR:$Rm, 0x0000FFFF), (t2UXTH rGPR:$Rm, 0)>, + Requires<[IsThumb2]>; +def : T2Pat<(and rGPR:$Rm, 0x00FF00FF), (t2UXTB16 rGPR:$Rm, 0)>, + Requires<[HasDSP, IsThumb2]>; +def : T2Pat<(add rGPR:$Rn, (and rGPR:$Rm, 0x00FF)), + (t2UXTAB rGPR:$Rn, rGPR:$Rm, 0)>, + Requires<[HasDSP, IsThumb2]>; +def : T2Pat<(add rGPR:$Rn, (and rGPR:$Rm, 0xFFFF)), + (t2UXTAH rGPR:$Rn, rGPR:$Rm, 0)>, + Requires<[HasDSP, IsThumb2]>; +} + +def : T2Pat<(sext_inreg rGPR:$Src, i8), (t2SXTB rGPR:$Src, 0)>, + Requires<[IsThumb2]>; +def : T2Pat<(sext_inreg rGPR:$Src, i16), (t2SXTH rGPR:$Src, 0)>, + Requires<[IsThumb2]>; +def : T2Pat<(add rGPR:$Rn, (sext_inreg rGPR:$Rm, i8)), + (t2SXTAB rGPR:$Rn, rGPR:$Rm, 0)>, + Requires<[HasDSP, IsThumb2]>; +def : T2Pat<(add rGPR:$Rn, (sext_inreg rGPR:$Rm, i16)), + (t2SXTAH rGPR:$Rn, rGPR:$Rm, 0)>, + Requires<[HasDSP, IsThumb2]>; + +// Atomic load/store patterns +def : T2Pat<(atomic_load_8 t2addrmode_imm12:$addr), + (t2LDRBi12 t2addrmode_imm12:$addr)>; +def : T2Pat<(atomic_load_8 t2addrmode_negimm8:$addr), + (t2LDRBi8 t2addrmode_negimm8:$addr)>; +def : T2Pat<(atomic_load_8 t2addrmode_so_reg:$addr), + (t2LDRBs t2addrmode_so_reg:$addr)>; +def : T2Pat<(atomic_load_16 t2addrmode_imm12:$addr), + (t2LDRHi12 t2addrmode_imm12:$addr)>; +def : T2Pat<(atomic_load_16 t2addrmode_negimm8:$addr), + (t2LDRHi8 t2addrmode_negimm8:$addr)>; +def : T2Pat<(atomic_load_16 t2addrmode_so_reg:$addr), + (t2LDRHs t2addrmode_so_reg:$addr)>; +def : T2Pat<(atomic_load_32 t2addrmode_imm12:$addr), + (t2LDRi12 t2addrmode_imm12:$addr)>; +def : T2Pat<(atomic_load_32 t2addrmode_negimm8:$addr), + (t2LDRi8 t2addrmode_negimm8:$addr)>; +def : T2Pat<(atomic_load_32 t2addrmode_so_reg:$addr), + (t2LDRs t2addrmode_so_reg:$addr)>; +def : T2Pat<(atomic_store_8 t2addrmode_imm12:$addr, GPR:$val), + (t2STRBi12 GPR:$val, t2addrmode_imm12:$addr)>; +def : T2Pat<(atomic_store_8 t2addrmode_negimm8:$addr, GPR:$val), + (t2STRBi8 GPR:$val, t2addrmode_negimm8:$addr)>; +def : T2Pat<(atomic_store_8 t2addrmode_so_reg:$addr, GPR:$val), + (t2STRBs GPR:$val, t2addrmode_so_reg:$addr)>; +def : T2Pat<(atomic_store_16 t2addrmode_imm12:$addr, GPR:$val), + (t2STRHi12 GPR:$val, t2addrmode_imm12:$addr)>; +def : T2Pat<(atomic_store_16 t2addrmode_negimm8:$addr, GPR:$val), + (t2STRHi8 GPR:$val, t2addrmode_negimm8:$addr)>; +def : T2Pat<(atomic_store_16 t2addrmode_so_reg:$addr, GPR:$val), + (t2STRHs GPR:$val, t2addrmode_so_reg:$addr)>; +def : T2Pat<(atomic_store_32 t2addrmode_imm12:$addr, GPR:$val), + (t2STRi12 GPR:$val, t2addrmode_imm12:$addr)>; +def : T2Pat<(atomic_store_32 t2addrmode_negimm8:$addr, GPR:$val), + (t2STRi8 GPR:$val, t2addrmode_negimm8:$addr)>; +def : T2Pat<(atomic_store_32 t2addrmode_so_reg:$addr, GPR:$val), + (t2STRs GPR:$val, t2addrmode_so_reg:$addr)>; + +let AddedComplexity = 8 in { + def : T2Pat<(atomic_load_acquire_8 addr_offset_none:$addr), (t2LDAB addr_offset_none:$addr)>; + def : T2Pat<(atomic_load_acquire_16 addr_offset_none:$addr), (t2LDAH addr_offset_none:$addr)>; + def : T2Pat<(atomic_load_acquire_32 addr_offset_none:$addr), (t2LDA addr_offset_none:$addr)>; + def : T2Pat<(atomic_store_release_8 addr_offset_none:$addr, GPR:$val), (t2STLB GPR:$val, addr_offset_none:$addr)>; + def : T2Pat<(atomic_store_release_16 addr_offset_none:$addr, GPR:$val), (t2STLH GPR:$val, addr_offset_none:$addr)>; + def : T2Pat<(atomic_store_release_32 addr_offset_none:$addr, GPR:$val), (t2STL GPR:$val, addr_offset_none:$addr)>; +} + + +//===----------------------------------------------------------------------===// +// Assembler aliases +// + +// Aliases for ADC without the ".w" optional width specifier. +def : t2InstAlias<"adc${s}${p} $Rd, $Rn, $Rm", + (t2ADCrr rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"adc${s}${p} $Rd, $Rn, $ShiftedRm", + (t2ADCrs rGPR:$Rd, rGPR:$Rn, t2_so_reg:$ShiftedRm, + pred:$p, cc_out:$s)>; + +// Aliases for SBC without the ".w" optional width specifier. +def : t2InstAlias<"sbc${s}${p} $Rd, $Rn, $Rm", + (t2SBCrr rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"sbc${s}${p} $Rd, $Rn, $ShiftedRm", + (t2SBCrs rGPR:$Rd, rGPR:$Rn, t2_so_reg:$ShiftedRm, + pred:$p, cc_out:$s)>; + +// Aliases for ADD without the ".w" optional width specifier. +def : t2InstAlias<"add${s}${p} $Rd, $Rn, $imm", + (t2ADDri GPRnopc:$Rd, GPRnopc:$Rn, t2_so_imm:$imm, pred:$p, + cc_out:$s)>; +def : t2InstAlias<"add${p} $Rd, $Rn, $imm", + (t2ADDri12 GPRnopc:$Rd, GPR:$Rn, imm0_4095:$imm, pred:$p)>; +def : t2InstAlias<"add${s}${p} $Rd, $Rn, $Rm", + (t2ADDrr GPRnopc:$Rd, GPRnopc:$Rn, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"add${s}${p} $Rd, $Rn, $ShiftedRm", + (t2ADDrs GPRnopc:$Rd, GPRnopc:$Rn, t2_so_reg:$ShiftedRm, + pred:$p, cc_out:$s)>; +// ... and with the destination and source register combined. +def : t2InstAlias<"add${s}${p} $Rdn, $imm", + (t2ADDri GPRnopc:$Rdn, GPRnopc:$Rdn, t2_so_imm:$imm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"add${p} $Rdn, $imm", + (t2ADDri12 GPRnopc:$Rdn, GPRnopc:$Rdn, imm0_4095:$imm, pred:$p)>; +def : t2InstAlias<"add${s}${p} $Rdn, $Rm", + (t2ADDrr GPRnopc:$Rdn, GPRnopc:$Rdn, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"add${s}${p} $Rdn, $ShiftedRm", + (t2ADDrs GPRnopc:$Rdn, GPRnopc:$Rdn, t2_so_reg:$ShiftedRm, + pred:$p, cc_out:$s)>; + +// add w/ negative immediates is just a sub. +def : t2InstSubst<"add${s}${p} $Rd, $Rn, $imm", + (t2SUBri GPRnopc:$Rd, GPRnopc:$Rn, t2_so_imm_neg:$imm, pred:$p, + cc_out:$s)>; +def : t2InstSubst<"add${p} $Rd, $Rn, $imm", + (t2SUBri12 GPRnopc:$Rd, GPR:$Rn, imm0_4095_neg:$imm, pred:$p)>; +def : t2InstSubst<"add${s}${p} $Rdn, $imm", + (t2SUBri GPRnopc:$Rdn, GPRnopc:$Rdn, t2_so_imm_neg:$imm, pred:$p, + cc_out:$s)>; +def : t2InstSubst<"add${p} $Rdn, $imm", + (t2SUBri12 GPRnopc:$Rdn, GPRnopc:$Rdn, imm0_4095_neg:$imm, pred:$p)>; + +def : t2InstSubst<"add${s}${p}.w $Rd, $Rn, $imm", + (t2SUBri GPRnopc:$Rd, GPRnopc:$Rn, t2_so_imm_neg:$imm, pred:$p, + cc_out:$s)>; +def : t2InstSubst<"addw${p} $Rd, $Rn, $imm", + (t2SUBri12 GPRnopc:$Rd, GPR:$Rn, imm0_4095_neg:$imm, pred:$p)>; +def : t2InstSubst<"add${s}${p}.w $Rdn, $imm", + (t2SUBri GPRnopc:$Rdn, GPRnopc:$Rdn, t2_so_imm_neg:$imm, pred:$p, + cc_out:$s)>; +def : t2InstSubst<"addw${p} $Rdn, $imm", + (t2SUBri12 GPRnopc:$Rdn, GPRnopc:$Rdn, imm0_4095_neg:$imm, pred:$p)>; + + +// Aliases for SUB without the ".w" optional width specifier. +def : t2InstAlias<"sub${s}${p} $Rd, $Rn, $imm", + (t2SUBri GPRnopc:$Rd, GPRnopc:$Rn, t2_so_imm:$imm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"sub${p} $Rd, $Rn, $imm", + (t2SUBri12 GPRnopc:$Rd, GPR:$Rn, imm0_4095:$imm, pred:$p)>; +def : t2InstAlias<"sub${s}${p} $Rd, $Rn, $Rm", + (t2SUBrr GPRnopc:$Rd, GPRnopc:$Rn, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"sub${s}${p} $Rd, $Rn, $ShiftedRm", + (t2SUBrs GPRnopc:$Rd, GPRnopc:$Rn, t2_so_reg:$ShiftedRm, + pred:$p, cc_out:$s)>; +// ... and with the destination and source register combined. +def : t2InstAlias<"sub${s}${p} $Rdn, $imm", + (t2SUBri GPRnopc:$Rdn, GPRnopc:$Rdn, t2_so_imm:$imm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"sub${p} $Rdn, $imm", + (t2SUBri12 GPRnopc:$Rdn, GPRnopc:$Rdn, imm0_4095:$imm, pred:$p)>; +def : t2InstAlias<"sub${s}${p}.w $Rdn, $Rm", + (t2SUBrr GPRnopc:$Rdn, GPRnopc:$Rdn, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"sub${s}${p} $Rdn, $Rm", + (t2SUBrr GPRnopc:$Rdn, GPRnopc:$Rdn, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"sub${s}${p} $Rdn, $ShiftedRm", + (t2SUBrs GPRnopc:$Rdn, GPRnopc:$Rdn, t2_so_reg:$ShiftedRm, + pred:$p, cc_out:$s)>; + +// Alias for compares without the ".w" optional width specifier. +def : t2InstAlias<"cmn${p} $Rn, $Rm", + (t2CMNzrr GPRnopc:$Rn, rGPR:$Rm, pred:$p)>; +def : t2InstAlias<"teq${p} $Rn, $Rm", + (t2TEQrr GPRnopc:$Rn, rGPR:$Rm, pred:$p)>; +def : t2InstAlias<"tst${p} $Rn, $Rm", + (t2TSTrr GPRnopc:$Rn, rGPR:$Rm, pred:$p)>; + +// Memory barriers +def : InstAlias<"dmb${p}", (t2DMB 0xf, pred:$p), 0>, Requires<[HasDB]>; +def : InstAlias<"dsb${p}", (t2DSB 0xf, pred:$p), 0>, Requires<[HasDB]>; +def : InstAlias<"isb${p}", (t2ISB 0xf, pred:$p), 0>, Requires<[HasDB]>; +// Armv8-R 'Data Full Barrier' +def : InstAlias<"dfb${p}", (t2DSB 0xc, pred:$p), 1>, Requires<[HasDFB]>; + +// Alias for LDR, LDRB, LDRH, LDRSB, and LDRSH without the ".w" optional +// width specifier. +def : t2InstAlias<"ldr${p} $Rt, $addr", + (t2LDRi12 GPR:$Rt, t2addrmode_imm12:$addr, pred:$p)>; +def : t2InstAlias<"ldrb${p} $Rt, $addr", + (t2LDRBi12 rGPR:$Rt, t2addrmode_imm12:$addr, pred:$p)>; +def : t2InstAlias<"ldrh${p} $Rt, $addr", + (t2LDRHi12 rGPR:$Rt, t2addrmode_imm12:$addr, pred:$p)>; +def : t2InstAlias<"ldrsb${p} $Rt, $addr", + (t2LDRSBi12 rGPR:$Rt, t2addrmode_imm12:$addr, pred:$p)>; +def : t2InstAlias<"ldrsh${p} $Rt, $addr", + (t2LDRSHi12 rGPR:$Rt, t2addrmode_imm12:$addr, pred:$p)>; + +def : t2InstAlias<"ldr${p} $Rt, $addr", + (t2LDRs GPR:$Rt, t2addrmode_so_reg:$addr, pred:$p)>; +def : t2InstAlias<"ldrb${p} $Rt, $addr", + (t2LDRBs rGPR:$Rt, t2addrmode_so_reg:$addr, pred:$p)>; +def : t2InstAlias<"ldrh${p} $Rt, $addr", + (t2LDRHs rGPR:$Rt, t2addrmode_so_reg:$addr, pred:$p)>; +def : t2InstAlias<"ldrsb${p} $Rt, $addr", + (t2LDRSBs rGPR:$Rt, t2addrmode_so_reg:$addr, pred:$p)>; +def : t2InstAlias<"ldrsh${p} $Rt, $addr", + (t2LDRSHs rGPR:$Rt, t2addrmode_so_reg:$addr, pred:$p)>; + +def : t2InstAlias<"ldr${p} $Rt, $addr", + (t2LDRpci GPR:$Rt, t2ldrlabel:$addr, pred:$p)>; +def : t2InstAlias<"ldrb${p} $Rt, $addr", + (t2LDRBpci rGPR:$Rt, t2ldrlabel:$addr, pred:$p)>; +def : t2InstAlias<"ldrh${p} $Rt, $addr", + (t2LDRHpci rGPR:$Rt, t2ldrlabel:$addr, pred:$p)>; +def : t2InstAlias<"ldrsb${p} $Rt, $addr", + (t2LDRSBpci rGPR:$Rt, t2ldrlabel:$addr, pred:$p)>; +def : t2InstAlias<"ldrsh${p} $Rt, $addr", + (t2LDRSHpci rGPR:$Rt, t2ldrlabel:$addr, pred:$p)>; + +// Alias for MVN with(out) the ".w" optional width specifier. +def : t2InstAlias<"mvn${s}${p}.w $Rd, $imm", + (t2MVNi rGPR:$Rd, t2_so_imm:$imm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"mvn${s}${p} $Rd, $Rm", + (t2MVNr rGPR:$Rd, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"mvn${s}${p} $Rd, $ShiftedRm", + (t2MVNs rGPR:$Rd, t2_so_reg:$ShiftedRm, pred:$p, cc_out:$s)>; + +// PKHBT/PKHTB with default shift amount. PKHTB is equivalent to PKHBT with the +// input operands swapped when the shift amount is zero (i.e., unspecified). +def : InstAlias<"pkhbt${p} $Rd, $Rn, $Rm", + (t2PKHBT rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : InstAlias<"pkhtb${p} $Rd, $Rn, $Rm", + (t2PKHBT rGPR:$Rd, rGPR:$Rm, rGPR:$Rn, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; + +// PUSH/POP aliases for STM/LDM +def : t2InstAlias<"push${p}.w $regs", (t2STMDB_UPD SP, pred:$p, reglist:$regs)>; +def : t2InstAlias<"push${p} $regs", (t2STMDB_UPD SP, pred:$p, reglist:$regs)>; +def : t2InstAlias<"pop${p}.w $regs", (t2LDMIA_UPD SP, pred:$p, reglist:$regs)>; +def : t2InstAlias<"pop${p} $regs", (t2LDMIA_UPD SP, pred:$p, reglist:$regs)>; + +// STMIA/STMIA_UPD aliases w/o the optional .w suffix +def : t2InstAlias<"stm${p} $Rn, $regs", + (t2STMIA GPR:$Rn, pred:$p, reglist:$regs)>; +def : t2InstAlias<"stm${p} $Rn!, $regs", + (t2STMIA_UPD GPR:$Rn, pred:$p, reglist:$regs)>; + +// LDMIA/LDMIA_UPD aliases w/o the optional .w suffix +def : t2InstAlias<"ldm${p} $Rn, $regs", + (t2LDMIA GPR:$Rn, pred:$p, reglist:$regs)>; +def : t2InstAlias<"ldm${p} $Rn!, $regs", + (t2LDMIA_UPD GPR:$Rn, pred:$p, reglist:$regs)>; + +// STMDB/STMDB_UPD aliases w/ the optional .w suffix +def : t2InstAlias<"stmdb${p}.w $Rn, $regs", + (t2STMDB GPR:$Rn, pred:$p, reglist:$regs)>; +def : t2InstAlias<"stmdb${p}.w $Rn!, $regs", + (t2STMDB_UPD GPR:$Rn, pred:$p, reglist:$regs)>; + +// LDMDB/LDMDB_UPD aliases w/ the optional .w suffix +def : t2InstAlias<"ldmdb${p}.w $Rn, $regs", + (t2LDMDB GPR:$Rn, pred:$p, reglist:$regs)>; +def : t2InstAlias<"ldmdb${p}.w $Rn!, $regs", + (t2LDMDB_UPD GPR:$Rn, pred:$p, reglist:$regs)>; + +// Alias for REV/REV16/REVSH without the ".w" optional width specifier. +def : t2InstAlias<"rev${p} $Rd, $Rm", (t2REV rGPR:$Rd, rGPR:$Rm, pred:$p)>; +def : t2InstAlias<"rev16${p} $Rd, $Rm", (t2REV16 rGPR:$Rd, rGPR:$Rm, pred:$p)>; +def : t2InstAlias<"revsh${p} $Rd, $Rm", (t2REVSH rGPR:$Rd, rGPR:$Rm, pred:$p)>; + + +// Alias for RSB without the ".w" optional width specifier, and with optional +// implied destination register. +def : t2InstAlias<"rsb${s}${p} $Rd, $Rn, $imm", + (t2RSBri rGPR:$Rd, rGPR:$Rn, t2_so_imm:$imm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"rsb${s}${p} $Rdn, $imm", + (t2RSBri rGPR:$Rdn, rGPR:$Rdn, t2_so_imm:$imm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"rsb${s}${p} $Rdn, $Rm", + (t2RSBrr rGPR:$Rdn, rGPR:$Rdn, rGPR:$Rm, pred:$p, cc_out:$s)>; +def : t2InstAlias<"rsb${s}${p} $Rdn, $ShiftedRm", + (t2RSBrs rGPR:$Rdn, rGPR:$Rdn, t2_so_reg:$ShiftedRm, pred:$p, + cc_out:$s)>; + +// SSAT/USAT optional shift operand. +def : t2InstAlias<"ssat${p} $Rd, $sat_imm, $Rn", + (t2SSAT rGPR:$Rd, imm1_32:$sat_imm, rGPR:$Rn, 0, pred:$p)>; +def : t2InstAlias<"usat${p} $Rd, $sat_imm, $Rn", + (t2USAT rGPR:$Rd, imm0_31:$sat_imm, rGPR:$Rn, 0, pred:$p)>; + +// STM w/o the .w suffix. +def : t2InstAlias<"stm${p} $Rn, $regs", + (t2STMIA GPR:$Rn, pred:$p, reglist:$regs)>; + +// Alias for STR, STRB, and STRH without the ".w" optional +// width specifier. +def : t2InstAlias<"str${p} $Rt, $addr", + (t2STRi12 GPR:$Rt, t2addrmode_imm12:$addr, pred:$p)>; +def : t2InstAlias<"strb${p} $Rt, $addr", + (t2STRBi12 rGPR:$Rt, t2addrmode_imm12:$addr, pred:$p)>; +def : t2InstAlias<"strh${p} $Rt, $addr", + (t2STRHi12 rGPR:$Rt, t2addrmode_imm12:$addr, pred:$p)>; + +def : t2InstAlias<"str${p} $Rt, $addr", + (t2STRs GPR:$Rt, t2addrmode_so_reg:$addr, pred:$p)>; +def : t2InstAlias<"strb${p} $Rt, $addr", + (t2STRBs rGPR:$Rt, t2addrmode_so_reg:$addr, pred:$p)>; +def : t2InstAlias<"strh${p} $Rt, $addr", + (t2STRHs rGPR:$Rt, t2addrmode_so_reg:$addr, pred:$p)>; + +// Extend instruction optional rotate operand. +def : InstAlias<"sxtab${p} $Rd, $Rn, $Rm", + (t2SXTAB rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : InstAlias<"sxtah${p} $Rd, $Rn, $Rm", + (t2SXTAH rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : InstAlias<"sxtab16${p} $Rd, $Rn, $Rm", + (t2SXTAB16 rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : InstAlias<"sxtb16${p} $Rd, $Rm", + (t2SXTB16 rGPR:$Rd, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; + +def : t2InstAlias<"sxtb${p} $Rd, $Rm", + (t2SXTB rGPR:$Rd, rGPR:$Rm, 0, pred:$p)>; +def : t2InstAlias<"sxth${p} $Rd, $Rm", + (t2SXTH rGPR:$Rd, rGPR:$Rm, 0, pred:$p)>; +def : t2InstAlias<"sxtb${p}.w $Rd, $Rm", + (t2SXTB rGPR:$Rd, rGPR:$Rm, 0, pred:$p)>; +def : t2InstAlias<"sxth${p}.w $Rd, $Rm", + (t2SXTH rGPR:$Rd, rGPR:$Rm, 0, pred:$p)>; + +def : InstAlias<"uxtab${p} $Rd, $Rn, $Rm", + (t2UXTAB rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : InstAlias<"uxtah${p} $Rd, $Rn, $Rm", + (t2UXTAH rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : InstAlias<"uxtab16${p} $Rd, $Rn, $Rm", + (t2UXTAB16 rGPR:$Rd, rGPR:$Rn, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : InstAlias<"uxtb16${p} $Rd, $Rm", + (t2UXTB16 rGPR:$Rd, rGPR:$Rm, 0, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; + +def : t2InstAlias<"uxtb${p} $Rd, $Rm", + (t2UXTB rGPR:$Rd, rGPR:$Rm, 0, pred:$p)>; +def : t2InstAlias<"uxth${p} $Rd, $Rm", + (t2UXTH rGPR:$Rd, rGPR:$Rm, 0, pred:$p)>; +def : t2InstAlias<"uxtb${p}.w $Rd, $Rm", + (t2UXTB rGPR:$Rd, rGPR:$Rm, 0, pred:$p)>; +def : t2InstAlias<"uxth${p}.w $Rd, $Rm", + (t2UXTH rGPR:$Rd, rGPR:$Rm, 0, pred:$p)>; + +// Extend instruction w/o the ".w" optional width specifier. +def : t2InstAlias<"uxtb${p} $Rd, $Rm$rot", + (t2UXTB rGPR:$Rd, rGPR:$Rm, rot_imm:$rot, pred:$p)>; +def : InstAlias<"uxtb16${p} $Rd, $Rm$rot", + (t2UXTB16 rGPR:$Rd, rGPR:$Rm, rot_imm:$rot, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : t2InstAlias<"uxth${p} $Rd, $Rm$rot", + (t2UXTH rGPR:$Rd, rGPR:$Rm, rot_imm:$rot, pred:$p)>; + +def : t2InstAlias<"sxtb${p} $Rd, $Rm$rot", + (t2SXTB rGPR:$Rd, rGPR:$Rm, rot_imm:$rot, pred:$p)>; +def : InstAlias<"sxtb16${p} $Rd, $Rm$rot", + (t2SXTB16 rGPR:$Rd, rGPR:$Rm, rot_imm:$rot, pred:$p), 0>, + Requires<[HasDSP, IsThumb2]>; +def : t2InstAlias<"sxth${p} $Rd, $Rm$rot", + (t2SXTH rGPR:$Rd, rGPR:$Rm, rot_imm:$rot, pred:$p)>; + + +// "mov Rd, t2_so_imm_not" can be handled via "mvn" in assembly, just like +// for isel. +def : t2InstSubst<"mov${p} $Rd, $imm", + (t2MVNi rGPR:$Rd, t2_so_imm_not:$imm, pred:$p, zero_reg)>; +def : t2InstSubst<"mvn${s}${p} $Rd, $imm", + (t2MOVi rGPR:$Rd, t2_so_imm_not:$imm, pred:$p, s_cc_out:$s)>; +// Same for AND <--> BIC +def : t2InstSubst<"bic${s}${p} $Rd, $Rn, $imm", + (t2ANDri rGPR:$Rd, rGPR:$Rn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"bic${s}${p} $Rdn, $imm", + (t2ANDri rGPR:$Rdn, rGPR:$Rdn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"bic${s}${p}.w $Rd, $Rn, $imm", + (t2ANDri rGPR:$Rd, rGPR:$Rn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"bic${s}${p}.w $Rdn, $imm", + (t2ANDri rGPR:$Rdn, rGPR:$Rdn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"and${s}${p} $Rd, $Rn, $imm", + (t2BICri rGPR:$Rd, rGPR:$Rn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"and${s}${p} $Rdn, $imm", + (t2BICri rGPR:$Rdn, rGPR:$Rdn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"and${s}${p}.w $Rd, $Rn, $imm", + (t2BICri rGPR:$Rd, rGPR:$Rn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"and${s}${p}.w $Rdn, $imm", + (t2BICri rGPR:$Rdn, rGPR:$Rdn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +// And ORR <--> ORN +def : t2InstSubst<"orn${s}${p} $Rd, $Rn, $imm", + (t2ORRri rGPR:$Rd, rGPR:$Rn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"orn${s}${p} $Rdn, $imm", + (t2ORRri rGPR:$Rdn, rGPR:$Rdn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"orr${s}${p} $Rd, $Rn, $imm", + (t2ORNri rGPR:$Rd, rGPR:$Rn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"orr${s}${p} $Rdn, $imm", + (t2ORNri rGPR:$Rdn, rGPR:$Rdn, t2_so_imm_not:$imm, + pred:$p, cc_out:$s)>; +// Likewise, "add Rd, t2_so_imm_neg" -> sub +def : t2InstSubst<"add${s}${p} $Rd, $Rn, $imm", + (t2SUBri GPRnopc:$Rd, GPRnopc:$Rn, t2_so_imm_neg:$imm, + pred:$p, cc_out:$s)>; +def : t2InstSubst<"add${s}${p} $Rd, $imm", + (t2SUBri GPRnopc:$Rd, GPRnopc:$Rd, t2_so_imm_neg:$imm, + pred:$p, cc_out:$s)>; +// Same for CMP <--> CMN via t2_so_imm_neg +def : t2InstSubst<"cmp${p} $Rd, $imm", + (t2CMNri rGPR:$Rd, t2_so_imm_neg:$imm, pred:$p)>; +def : t2InstSubst<"cmn${p} $Rd, $imm", + (t2CMPri rGPR:$Rd, t2_so_imm_neg:$imm, pred:$p)>; + + +// Wide 'mul' encoding can be specified with only two operands. +def : t2InstAlias<"mul${p} $Rn, $Rm", + (t2MUL rGPR:$Rn, rGPR:$Rm, rGPR:$Rn, pred:$p)>; + +// "neg" is and alias for "rsb rd, rn, #0" +def : t2InstAlias<"neg${s}${p} $Rd, $Rm", + (t2RSBri rGPR:$Rd, rGPR:$Rm, 0, pred:$p, cc_out:$s)>; + +// MOV so_reg assembler pseudos. InstAlias isn't expressive enough for +// these, unfortunately. +// FIXME: LSL #0 in the shift should allow SP to be used as either the +// source or destination (but not both). +def t2MOVsi: t2AsmPseudo<"mov${p} $Rd, $shift", + (ins rGPR:$Rd, t2_so_reg:$shift, pred:$p)>; +def t2MOVSsi: t2AsmPseudo<"movs${p} $Rd, $shift", + (ins rGPR:$Rd, t2_so_reg:$shift, pred:$p)>; + +def t2MOVsr: t2AsmPseudo<"mov${p} $Rd, $shift", + (ins rGPR:$Rd, so_reg_reg:$shift, pred:$p)>; +def t2MOVSsr: t2AsmPseudo<"movs${p} $Rd, $shift", + (ins rGPR:$Rd, so_reg_reg:$shift, pred:$p)>; + +// Aliases for the above with the .w qualifier +def : t2InstAlias<"mov${p}.w $Rd, $shift", + (t2MOVsi rGPR:$Rd, t2_so_reg:$shift, pred:$p)>; +def : t2InstAlias<"movs${p}.w $Rd, $shift", + (t2MOVSsi rGPR:$Rd, t2_so_reg:$shift, pred:$p)>; +def : t2InstAlias<"mov${p}.w $Rd, $shift", + (t2MOVsr rGPR:$Rd, so_reg_reg:$shift, pred:$p)>; +def : t2InstAlias<"movs${p}.w $Rd, $shift", + (t2MOVSsr rGPR:$Rd, so_reg_reg:$shift, pred:$p)>; + +// ADR w/o the .w suffix +def : t2InstAlias<"adr${p} $Rd, $addr", + (t2ADR rGPR:$Rd, t2adrlabel:$addr, pred:$p)>; + +// LDR(literal) w/ alternate [pc, #imm] syntax. +def t2LDRpcrel : t2AsmPseudo<"ldr${p} $Rt, $addr", + (ins GPR:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; +def t2LDRBpcrel : t2AsmPseudo<"ldrb${p} $Rt, $addr", + (ins GPRnopc:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; +def t2LDRHpcrel : t2AsmPseudo<"ldrh${p} $Rt, $addr", + (ins GPRnopc:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; +def t2LDRSBpcrel : t2AsmPseudo<"ldrsb${p} $Rt, $addr", + (ins GPRnopc:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; +def t2LDRSHpcrel : t2AsmPseudo<"ldrsh${p} $Rt, $addr", + (ins GPRnopc:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; + // Version w/ the .w suffix. +def : t2InstAlias<"ldr${p}.w $Rt, $addr", + (t2LDRpcrel GPR:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p), 0>; +def : t2InstAlias<"ldrb${p}.w $Rt, $addr", + (t2LDRBpcrel GPRnopc:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; +def : t2InstAlias<"ldrh${p}.w $Rt, $addr", + (t2LDRHpcrel GPRnopc:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; +def : t2InstAlias<"ldrsb${p}.w $Rt, $addr", + (t2LDRSBpcrel GPRnopc:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; +def : t2InstAlias<"ldrsh${p}.w $Rt, $addr", + (t2LDRSHpcrel GPRnopc:$Rt, t2ldr_pcrel_imm12:$addr, pred:$p)>; + +def : t2InstAlias<"add${p} $Rd, pc, $imm", + (t2ADR rGPR:$Rd, imm0_4095:$imm, pred:$p)>; + +// Pseudo instruction ldr Rt, =immediate +def t2LDRConstPool + : t2AsmPseudo<"ldr${p} $Rt, $immediate", + (ins GPR:$Rt, const_pool_asm_imm:$immediate, pred:$p)>; +// Version w/ the .w suffix. +def : t2InstAlias<"ldr${p}.w $Rt, $immediate", + (t2LDRConstPool GPRnopc:$Rt, + const_pool_asm_imm:$immediate, pred:$p)>; + +// PLD/PLDW/PLI with alternate literal form. +def : t2InstAlias<"pld${p} $addr", + (t2PLDpci t2ldr_pcrel_imm12:$addr, pred:$p)>; +def : InstAlias<"pli${p} $addr", + (t2PLIpci t2ldr_pcrel_imm12:$addr, pred:$p), 0>, + Requires<[IsThumb2,HasV7]>; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMInstrVFP.td b/capstone/suite/synctools/tablegen/ARM/ARMInstrVFP.td new file mode 100644 index 000000000..2f14b78c9 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMInstrVFP.td @@ -0,0 +1,2482 @@ +//===-- ARMInstrVFP.td - VFP support for ARM ---------------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file describes the ARM VFP instruction set. +// +//===----------------------------------------------------------------------===// + +def SDT_CMPFP0 : SDTypeProfile<0, 2, [SDTCisFP<0>, SDTCisVT<1, i32>]>; +def SDT_VMOVDRR : SDTypeProfile<1, 2, [SDTCisVT<0, f64>, SDTCisVT<1, i32>, + SDTCisSameAs<1, 2>]>; +def SDT_VMOVRRD : SDTypeProfile<2, 1, [SDTCisVT<0, i32>, SDTCisSameAs<0, 1>, + SDTCisVT<2, f64>]>; + +def SDT_VMOVSR : SDTypeProfile<1, 1, [SDTCisVT<0, f32>, SDTCisVT<1, i32>]>; + +def arm_fmstat : SDNode<"ARMISD::FMSTAT", SDTNone, [SDNPInGlue, SDNPOutGlue]>; +def arm_cmpfp : SDNode<"ARMISD::CMPFP", SDT_ARMFCmp, [SDNPOutGlue]>; +def arm_cmpfp0 : SDNode<"ARMISD::CMPFPw0", SDT_CMPFP0, [SDNPOutGlue]>; +def arm_fmdrr : SDNode<"ARMISD::VMOVDRR", SDT_VMOVDRR>; +def arm_fmrrd : SDNode<"ARMISD::VMOVRRD", SDT_VMOVRRD>; +def arm_vmovsr : SDNode<"ARMISD::VMOVSR", SDT_VMOVSR>; + +def SDT_VMOVhr : SDTypeProfile<1, 1, [SDTCisFP<0>, SDTCisVT<1, i32>] >; +def SDT_VMOVrh : SDTypeProfile<1, 1, [SDTCisVT<0, i32>, SDTCisFP<1>] >; +def arm_vmovhr : SDNode<"ARMISD::VMOVhr", SDT_VMOVhr>; +def arm_vmovrh : SDNode<"ARMISD::VMOVrh", SDT_VMOVrh>; + +//===----------------------------------------------------------------------===// +// Operand Definitions. +// + +// 8-bit floating-point immediate encodings. +def FPImmOperand : AsmOperandClass { + let Name = "FPImm"; + let ParserMethod = "parseFPImm"; +} + +def vfp_f16imm : Operand<f16>, + PatLeaf<(f16 fpimm), [{ + return ARM_AM::getFP16Imm(N->getValueAPF()) != -1; + }], SDNodeXForm<fpimm, [{ + APFloat InVal = N->getValueAPF(); + uint32_t enc = ARM_AM::getFP16Imm(InVal); + return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32); + }]>> { + let PrintMethod = "printFPImmOperand"; + let ParserMatchClass = FPImmOperand; +} + +def vfp_f32imm : Operand<f32>, + PatLeaf<(f32 fpimm), [{ + return ARM_AM::getFP32Imm(N->getValueAPF()) != -1; + }], SDNodeXForm<fpimm, [{ + APFloat InVal = N->getValueAPF(); + uint32_t enc = ARM_AM::getFP32Imm(InVal); + return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32); + }]>> { + let PrintMethod = "printFPImmOperand"; + let ParserMatchClass = FPImmOperand; +} + +def vfp_f64imm : Operand<f64>, + PatLeaf<(f64 fpimm), [{ + return ARM_AM::getFP64Imm(N->getValueAPF()) != -1; + }], SDNodeXForm<fpimm, [{ + APFloat InVal = N->getValueAPF(); + uint32_t enc = ARM_AM::getFP64Imm(InVal); + return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32); + }]>> { + let PrintMethod = "printFPImmOperand"; + let ParserMatchClass = FPImmOperand; +} + +def alignedload16 : PatFrag<(ops node:$ptr), (load node:$ptr), [{ + return cast<LoadSDNode>(N)->getAlignment() >= 2; +}]>; + +def alignedload32 : PatFrag<(ops node:$ptr), (load node:$ptr), [{ + return cast<LoadSDNode>(N)->getAlignment() >= 4; +}]>; + +def alignedstore16 : PatFrag<(ops node:$val, node:$ptr), + (store node:$val, node:$ptr), [{ + return cast<StoreSDNode>(N)->getAlignment() >= 2; +}]>; + +def alignedstore32 : PatFrag<(ops node:$val, node:$ptr), + (store node:$val, node:$ptr), [{ + return cast<StoreSDNode>(N)->getAlignment() >= 4; +}]>; + +// The VCVT to/from fixed-point instructions encode the 'fbits' operand +// (the number of fixed bits) differently than it appears in the assembly +// source. It's encoded as "Size - fbits" where Size is the size of the +// fixed-point representation (32 or 16) and fbits is the value appearing +// in the assembly source, an integer in [0,16] or (0,32], depending on size. +def fbits32_asm_operand : AsmOperandClass { let Name = "FBits32"; } +def fbits32 : Operand<i32> { + let PrintMethod = "printFBits32"; + let ParserMatchClass = fbits32_asm_operand; +} + +def fbits16_asm_operand : AsmOperandClass { let Name = "FBits16"; } +def fbits16 : Operand<i32> { + let PrintMethod = "printFBits16"; + let ParserMatchClass = fbits16_asm_operand; +} + +//===----------------------------------------------------------------------===// +// Load / store Instructions. +// + +let canFoldAsLoad = 1, isReMaterializable = 1 in { + +def VLDRD : ADI5<0b1101, 0b01, (outs DPR:$Dd), (ins addrmode5:$addr), + IIC_fpLoad64, "vldr", "\t$Dd, $addr", + [(set DPR:$Dd, (f64 (alignedload32 addrmode5:$addr)))]>; + +def VLDRS : ASI5<0b1101, 0b01, (outs SPR:$Sd), (ins addrmode5:$addr), + IIC_fpLoad32, "vldr", "\t$Sd, $addr", + [(set SPR:$Sd, (alignedload32 addrmode5:$addr))]> { + // Some single precision VFP instructions may be executed on both NEON and VFP + // pipelines. + let D = VFPNeonDomain; +} + +def VLDRH : AHI5<0b1101, 0b01, (outs HPR:$Sd), (ins addrmode5fp16:$addr), + IIC_fpLoad16, "vldr", ".16\t$Sd, $addr", + [(set HPR:$Sd, (alignedload16 addrmode5fp16:$addr))]>, + Requires<[HasFullFP16]>; + +} // End of 'let canFoldAsLoad = 1, isReMaterializable = 1 in' + +def VSTRD : ADI5<0b1101, 0b00, (outs), (ins DPR:$Dd, addrmode5:$addr), + IIC_fpStore64, "vstr", "\t$Dd, $addr", + [(alignedstore32 (f64 DPR:$Dd), addrmode5:$addr)]>; + +def VSTRS : ASI5<0b1101, 0b00, (outs), (ins SPR:$Sd, addrmode5:$addr), + IIC_fpStore32, "vstr", "\t$Sd, $addr", + [(alignedstore32 SPR:$Sd, addrmode5:$addr)]> { + // Some single precision VFP instructions may be executed on both NEON and VFP + // pipelines. + let D = VFPNeonDomain; +} + +def VSTRH : AHI5<0b1101, 0b00, (outs), (ins HPR:$Sd, addrmode5fp16:$addr), + IIC_fpStore16, "vstr", ".16\t$Sd, $addr", + [(alignedstore16 HPR:$Sd, addrmode5fp16:$addr)]>, + Requires<[HasFullFP16]>; + +//===----------------------------------------------------------------------===// +// Load / store multiple Instructions. +// + +multiclass vfp_ldst_mult<string asm, bit L_bit, + InstrItinClass itin, InstrItinClass itin_upd> { + // Double Precision + def DIA : + AXDI4<(outs), (ins GPR:$Rn, pred:$p, dpr_reglist:$regs, variable_ops), + IndexModeNone, itin, + !strconcat(asm, "ia${p}\t$Rn, $regs"), "", []> { + let Inst{24-23} = 0b01; // Increment After + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + } + def DIA_UPD : + AXDI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, dpr_reglist:$regs, + variable_ops), + IndexModeUpd, itin_upd, + !strconcat(asm, "ia${p}\t$Rn!, $regs"), "$Rn = $wb", []> { + let Inst{24-23} = 0b01; // Increment After + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + } + def DDB_UPD : + AXDI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, dpr_reglist:$regs, + variable_ops), + IndexModeUpd, itin_upd, + !strconcat(asm, "db${p}\t$Rn!, $regs"), "$Rn = $wb", []> { + let Inst{24-23} = 0b10; // Decrement Before + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + } + + // Single Precision + def SIA : + AXSI4<(outs), (ins GPR:$Rn, pred:$p, spr_reglist:$regs, variable_ops), + IndexModeNone, itin, + !strconcat(asm, "ia${p}\t$Rn, $regs"), "", []> { + let Inst{24-23} = 0b01; // Increment After + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines. + let D = VFPNeonDomain; + } + def SIA_UPD : + AXSI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, spr_reglist:$regs, + variable_ops), + IndexModeUpd, itin_upd, + !strconcat(asm, "ia${p}\t$Rn!, $regs"), "$Rn = $wb", []> { + let Inst{24-23} = 0b01; // Increment After + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines. + let D = VFPNeonDomain; + } + def SDB_UPD : + AXSI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, spr_reglist:$regs, + variable_ops), + IndexModeUpd, itin_upd, + !strconcat(asm, "db${p}\t$Rn!, $regs"), "$Rn = $wb", []> { + let Inst{24-23} = 0b10; // Decrement Before + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines. + let D = VFPNeonDomain; + } +} + +let hasSideEffects = 0 in { + +let mayLoad = 1, hasExtraDefRegAllocReq = 1 in +defm VLDM : vfp_ldst_mult<"vldm", 1, IIC_fpLoad_m, IIC_fpLoad_mu>; + +let mayStore = 1, hasExtraSrcRegAllocReq = 1 in +defm VSTM : vfp_ldst_mult<"vstm", 0, IIC_fpStore_m, IIC_fpStore_mu>; + +} // hasSideEffects + +def : MnemonicAlias<"vldm", "vldmia">; +def : MnemonicAlias<"vstm", "vstmia">; + + +//===----------------------------------------------------------------------===// +// Lazy load / store multiple Instructions +// +let mayLoad = 1 in +def VLLDM : AXSI4<(outs), (ins GPRnopc:$Rn, pred:$p), IndexModeNone, + IIC_fpLoad_m, "vlldm${p}\t$Rn", "", []>, + Requires<[HasV8MMainline, Has8MSecExt]> { + let Inst{24-23} = 0b00; + let Inst{22} = 0; + let Inst{21} = 1; + let Inst{20} = 1; + let Inst{15-12} = 0; + let Inst{7-0} = 0; + let mayLoad = 1; +} + +let mayStore = 1 in +def VLSTM : AXSI4<(outs), (ins GPRnopc:$Rn, pred:$p), IndexModeNone, + IIC_fpStore_m, "vlstm${p}\t$Rn", "", []>, + Requires<[HasV8MMainline, Has8MSecExt]> { + let Inst{24-23} = 0b00; + let Inst{22} = 0; + let Inst{21} = 1; + let Inst{20} = 0; + let Inst{15-12} = 0; + let Inst{7-0} = 0; + let mayStore = 1; +} + +def : InstAlias<"vpush${p} $r", (VSTMDDB_UPD SP, pred:$p, dpr_reglist:$r), 0>, + Requires<[HasVFP2]>; +def : InstAlias<"vpush${p} $r", (VSTMSDB_UPD SP, pred:$p, spr_reglist:$r), 0>, + Requires<[HasVFP2]>; +def : InstAlias<"vpop${p} $r", (VLDMDIA_UPD SP, pred:$p, dpr_reglist:$r), 0>, + Requires<[HasVFP2]>; +def : InstAlias<"vpop${p} $r", (VLDMSIA_UPD SP, pred:$p, spr_reglist:$r), 0>, + Requires<[HasVFP2]>; +defm : VFPDTAnyInstAlias<"vpush${p}", "$r", + (VSTMSDB_UPD SP, pred:$p, spr_reglist:$r)>; +defm : VFPDTAnyInstAlias<"vpush${p}", "$r", + (VSTMDDB_UPD SP, pred:$p, dpr_reglist:$r)>; +defm : VFPDTAnyInstAlias<"vpop${p}", "$r", + (VLDMSIA_UPD SP, pred:$p, spr_reglist:$r)>; +defm : VFPDTAnyInstAlias<"vpop${p}", "$r", + (VLDMDIA_UPD SP, pred:$p, dpr_reglist:$r)>; + +// FLDMX, FSTMX - Load and store multiple unknown precision registers for +// pre-armv6 cores. +// These instruction are deprecated so we don't want them to get selected. +// However, there is no UAL syntax for them, so we keep them around for +// (dis)assembly only. +multiclass vfp_ldstx_mult<string asm, bit L_bit> { + // Unknown precision + def XIA : + AXXI4<(outs), (ins GPR:$Rn, pred:$p, dpr_reglist:$regs, variable_ops), + IndexModeNone, !strconcat(asm, "iax${p}\t$Rn, $regs"), "", []> { + let Inst{24-23} = 0b01; // Increment After + let Inst{21} = 0; // No writeback + let Inst{20} = L_bit; + } + def XIA_UPD : + AXXI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, dpr_reglist:$regs, variable_ops), + IndexModeUpd, !strconcat(asm, "iax${p}\t$Rn!, $regs"), "$Rn = $wb", []> { + let Inst{24-23} = 0b01; // Increment After + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + } + def XDB_UPD : + AXXI4<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, dpr_reglist:$regs, variable_ops), + IndexModeUpd, !strconcat(asm, "dbx${p}\t$Rn!, $regs"), "$Rn = $wb", []> { + let Inst{24-23} = 0b10; // Decrement Before + let Inst{21} = 1; // Writeback + let Inst{20} = L_bit; + } +} + +defm FLDM : vfp_ldstx_mult<"fldm", 1>; +defm FSTM : vfp_ldstx_mult<"fstm", 0>; + +def : VFP2MnemonicAlias<"fldmeax", "fldmdbx">; +def : VFP2MnemonicAlias<"fldmfdx", "fldmiax">; + +def : VFP2MnemonicAlias<"fstmeax", "fstmiax">; +def : VFP2MnemonicAlias<"fstmfdx", "fstmdbx">; + +//===----------------------------------------------------------------------===// +// FP Binary Operations. +// + +let TwoOperandAliasConstraint = "$Dn = $Dd" in +def VADDD : ADbI<0b11100, 0b11, 0, 0, + (outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm), + IIC_fpALU64, "vadd", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fadd DPR:$Dn, (f64 DPR:$Dm)))]>, + Sched<[WriteFPALU64]>; + +let TwoOperandAliasConstraint = "$Sn = $Sd" in +def VADDS : ASbIn<0b11100, 0b11, 0, 0, + (outs SPR:$Sd), (ins SPR:$Sn, SPR:$Sm), + IIC_fpALU32, "vadd", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fadd SPR:$Sn, SPR:$Sm))]>, + Sched<[WriteFPALU32]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +let TwoOperandAliasConstraint = "$Sn = $Sd" in +def VADDH : AHbI<0b11100, 0b11, 0, 0, + (outs HPR:$Sd), (ins HPR:$Sn, HPR:$Sm), + IIC_fpALU16, "vadd", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fadd HPR:$Sn, HPR:$Sm))]>, + Sched<[WriteFPALU32]>; + +let TwoOperandAliasConstraint = "$Dn = $Dd" in +def VSUBD : ADbI<0b11100, 0b11, 1, 0, + (outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm), + IIC_fpALU64, "vsub", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fsub DPR:$Dn, (f64 DPR:$Dm)))]>, + Sched<[WriteFPALU64]>; + +let TwoOperandAliasConstraint = "$Sn = $Sd" in +def VSUBS : ASbIn<0b11100, 0b11, 1, 0, + (outs SPR:$Sd), (ins SPR:$Sn, SPR:$Sm), + IIC_fpALU32, "vsub", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fsub SPR:$Sn, SPR:$Sm))]>, + Sched<[WriteFPALU32]>{ + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +let TwoOperandAliasConstraint = "$Sn = $Sd" in +def VSUBH : AHbI<0b11100, 0b11, 1, 0, + (outs HPR:$Sd), (ins HPR:$Sn, HPR:$Sm), + IIC_fpALU16, "vsub", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fsub HPR:$Sn, HPR:$Sm))]>, + Sched<[WriteFPALU32]>; + +let TwoOperandAliasConstraint = "$Dn = $Dd" in +def VDIVD : ADbI<0b11101, 0b00, 0, 0, + (outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm), + IIC_fpDIV64, "vdiv", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fdiv DPR:$Dn, (f64 DPR:$Dm)))]>, + Sched<[WriteFPDIV64]>; + +let TwoOperandAliasConstraint = "$Sn = $Sd" in +def VDIVS : ASbI<0b11101, 0b00, 0, 0, + (outs SPR:$Sd), (ins SPR:$Sn, SPR:$Sm), + IIC_fpDIV32, "vdiv", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fdiv SPR:$Sn, SPR:$Sm))]>, + Sched<[WriteFPDIV32]>; + +let TwoOperandAliasConstraint = "$Sn = $Sd" in +def VDIVH : AHbI<0b11101, 0b00, 0, 0, + (outs HPR:$Sd), (ins HPR:$Sn, HPR:$Sm), + IIC_fpDIV16, "vdiv", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fdiv HPR:$Sn, HPR:$Sm))]>, + Sched<[WriteFPDIV32]>; + +let TwoOperandAliasConstraint = "$Dn = $Dd" in +def VMULD : ADbI<0b11100, 0b10, 0, 0, + (outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm), + IIC_fpMUL64, "vmul", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fmul DPR:$Dn, (f64 DPR:$Dm)))]>, + Sched<[WriteFPMUL64, ReadFPMUL, ReadFPMUL]>; + +let TwoOperandAliasConstraint = "$Sn = $Sd" in +def VMULS : ASbIn<0b11100, 0b10, 0, 0, + (outs SPR:$Sd), (ins SPR:$Sn, SPR:$Sm), + IIC_fpMUL32, "vmul", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fmul SPR:$Sn, SPR:$Sm))]>, + Sched<[WriteFPMUL32, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +let TwoOperandAliasConstraint = "$Sn = $Sd" in +def VMULH : AHbI<0b11100, 0b10, 0, 0, + (outs HPR:$Sd), (ins HPR:$Sn, HPR:$Sm), + IIC_fpMUL16, "vmul", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fmul HPR:$Sn, HPR:$Sm))]>, + Sched<[WriteFPMUL32, ReadFPMUL, ReadFPMUL]>; + +def VNMULD : ADbI<0b11100, 0b10, 1, 0, + (outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm), + IIC_fpMUL64, "vnmul", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fneg (fmul DPR:$Dn, (f64 DPR:$Dm))))]>, + Sched<[WriteFPMUL64, ReadFPMUL, ReadFPMUL]>; + +def VNMULS : ASbI<0b11100, 0b10, 1, 0, + (outs SPR:$Sd), (ins SPR:$Sn, SPR:$Sm), + IIC_fpMUL32, "vnmul", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fneg (fmul SPR:$Sn, SPR:$Sm)))]>, + Sched<[WriteFPMUL32, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VNMULH : AHbI<0b11100, 0b10, 1, 0, + (outs HPR:$Sd), (ins HPR:$Sn, HPR:$Sm), + IIC_fpMUL16, "vnmul", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fneg (fmul HPR:$Sn, HPR:$Sm)))]>, + Sched<[WriteFPMUL32, ReadFPMUL, ReadFPMUL]>; + +multiclass vsel_inst<string op, bits<2> opc, int CC> { + let DecoderNamespace = "VFPV8", PostEncoderMethod = "", + Uses = [CPSR], AddedComplexity = 4 in { + def H : AHbInp<0b11100, opc, 0, + (outs HPR:$Sd), (ins HPR:$Sn, HPR:$Sm), + NoItinerary, !strconcat("vsel", op, ".f16\t$Sd, $Sn, $Sm"), + [(set HPR:$Sd, (ARMcmov HPR:$Sm, HPR:$Sn, CC))]>, + Requires<[HasFullFP16]>; + + def S : ASbInp<0b11100, opc, 0, + (outs SPR:$Sd), (ins SPR:$Sn, SPR:$Sm), + NoItinerary, !strconcat("vsel", op, ".f32\t$Sd, $Sn, $Sm"), + [(set SPR:$Sd, (ARMcmov SPR:$Sm, SPR:$Sn, CC))]>, + Requires<[HasFPARMv8]>; + + def D : ADbInp<0b11100, opc, 0, + (outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm), + NoItinerary, !strconcat("vsel", op, ".f64\t$Dd, $Dn, $Dm"), + [(set DPR:$Dd, (ARMcmov (f64 DPR:$Dm), (f64 DPR:$Dn), CC))]>, + Requires<[HasFPARMv8, HasDPVFP]>; + } +} + +// The CC constants here match ARMCC::CondCodes. +defm VSELGT : vsel_inst<"gt", 0b11, 12>; +defm VSELGE : vsel_inst<"ge", 0b10, 10>; +defm VSELEQ : vsel_inst<"eq", 0b00, 0>; +defm VSELVS : vsel_inst<"vs", 0b01, 6>; + +multiclass vmaxmin_inst<string op, bit opc, SDNode SD> { + let DecoderNamespace = "VFPV8", PostEncoderMethod = "" in { + def H : AHbInp<0b11101, 0b00, opc, + (outs HPR:$Sd), (ins HPR:$Sn, HPR:$Sm), + NoItinerary, !strconcat(op, ".f16\t$Sd, $Sn, $Sm"), + [(set HPR:$Sd, (SD HPR:$Sn, HPR:$Sm))]>, + Requires<[HasFullFP16]>; + + def S : ASbInp<0b11101, 0b00, opc, + (outs SPR:$Sd), (ins SPR:$Sn, SPR:$Sm), + NoItinerary, !strconcat(op, ".f32\t$Sd, $Sn, $Sm"), + [(set SPR:$Sd, (SD SPR:$Sn, SPR:$Sm))]>, + Requires<[HasFPARMv8]>; + + def D : ADbInp<0b11101, 0b00, opc, + (outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm), + NoItinerary, !strconcat(op, ".f64\t$Dd, $Dn, $Dm"), + [(set DPR:$Dd, (f64 (SD (f64 DPR:$Dn), (f64 DPR:$Dm))))]>, + Requires<[HasFPARMv8, HasDPVFP]>; + } +} + +defm VMAXNM : vmaxmin_inst<"vmaxnm", 0, fmaxnum>; +defm VMINNM : vmaxmin_inst<"vminnm", 1, fminnum>; + +// Match reassociated forms only if not sign dependent rounding. +def : Pat<(fmul (fneg DPR:$a), (f64 DPR:$b)), + (VNMULD DPR:$a, DPR:$b)>, + Requires<[NoHonorSignDependentRounding,HasDPVFP]>; +def : Pat<(fmul (fneg SPR:$a), SPR:$b), + (VNMULS SPR:$a, SPR:$b)>, Requires<[NoHonorSignDependentRounding]>; + +// These are encoded as unary instructions. +let Defs = [FPSCR_NZCV] in { +def VCMPED : ADuI<0b11101, 0b11, 0b0100, 0b11, 0, + (outs), (ins DPR:$Dd, DPR:$Dm), + IIC_fpCMP64, "vcmpe", ".f64\t$Dd, $Dm", + [(arm_cmpfp DPR:$Dd, (f64 DPR:$Dm), (i32 1))]>; + +def VCMPES : ASuI<0b11101, 0b11, 0b0100, 0b11, 0, + (outs), (ins SPR:$Sd, SPR:$Sm), + IIC_fpCMP32, "vcmpe", ".f32\t$Sd, $Sm", + [(arm_cmpfp SPR:$Sd, SPR:$Sm, (i32 1))]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VCMPEH : AHuI<0b11101, 0b11, 0b0100, 0b11, 0, + (outs), (ins HPR:$Sd, HPR:$Sm), + IIC_fpCMP16, "vcmpe", ".f16\t$Sd, $Sm", + [(arm_cmpfp HPR:$Sd, HPR:$Sm, (i32 1))]>; + +def VCMPD : ADuI<0b11101, 0b11, 0b0100, 0b01, 0, + (outs), (ins DPR:$Dd, DPR:$Dm), + IIC_fpCMP64, "vcmp", ".f64\t$Dd, $Dm", + [(arm_cmpfp DPR:$Dd, (f64 DPR:$Dm), (i32 0))]>; + +def VCMPS : ASuI<0b11101, 0b11, 0b0100, 0b01, 0, + (outs), (ins SPR:$Sd, SPR:$Sm), + IIC_fpCMP32, "vcmp", ".f32\t$Sd, $Sm", + [(arm_cmpfp SPR:$Sd, SPR:$Sm, (i32 0))]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VCMPH : AHuI<0b11101, 0b11, 0b0100, 0b01, 0, + (outs), (ins HPR:$Sd, HPR:$Sm), + IIC_fpCMP16, "vcmp", ".f16\t$Sd, $Sm", + [(arm_cmpfp HPR:$Sd, HPR:$Sm, (i32 0))]>; +} // Defs = [FPSCR_NZCV] + +//===----------------------------------------------------------------------===// +// FP Unary Operations. +// + +def VABSD : ADuI<0b11101, 0b11, 0b0000, 0b11, 0, + (outs DPR:$Dd), (ins DPR:$Dm), + IIC_fpUNA64, "vabs", ".f64\t$Dd, $Dm", + [(set DPR:$Dd, (fabs (f64 DPR:$Dm)))]>; + +def VABSS : ASuIn<0b11101, 0b11, 0b0000, 0b11, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpUNA32, "vabs", ".f32\t$Sd, $Sm", + [(set SPR:$Sd, (fabs SPR:$Sm))]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VABSH : AHuI<0b11101, 0b11, 0b0000, 0b11, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpUNA16, "vabs", ".f16\t$Sd, $Sm", + []>; + +let Defs = [FPSCR_NZCV] in { +def VCMPEZD : ADuI<0b11101, 0b11, 0b0101, 0b11, 0, + (outs), (ins DPR:$Dd), + IIC_fpCMP64, "vcmpe", ".f64\t$Dd, #0", + [(arm_cmpfp0 (f64 DPR:$Dd), (i32 1))]> { + let Inst{3-0} = 0b0000; + let Inst{5} = 0; +} + +def VCMPEZS : ASuI<0b11101, 0b11, 0b0101, 0b11, 0, + (outs), (ins SPR:$Sd), + IIC_fpCMP32, "vcmpe", ".f32\t$Sd, #0", + [(arm_cmpfp0 SPR:$Sd, (i32 1))]> { + let Inst{3-0} = 0b0000; + let Inst{5} = 0; + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VCMPEZH : AHuI<0b11101, 0b11, 0b0101, 0b11, 0, + (outs), (ins HPR:$Sd), + IIC_fpCMP16, "vcmpe", ".f16\t$Sd, #0", + [(arm_cmpfp0 HPR:$Sd, (i32 1))]> { + let Inst{3-0} = 0b0000; + let Inst{5} = 0; +} + +def VCMPZD : ADuI<0b11101, 0b11, 0b0101, 0b01, 0, + (outs), (ins DPR:$Dd), + IIC_fpCMP64, "vcmp", ".f64\t$Dd, #0", + [(arm_cmpfp0 (f64 DPR:$Dd), (i32 0))]> { + let Inst{3-0} = 0b0000; + let Inst{5} = 0; +} + +def VCMPZS : ASuI<0b11101, 0b11, 0b0101, 0b01, 0, + (outs), (ins SPR:$Sd), + IIC_fpCMP32, "vcmp", ".f32\t$Sd, #0", + [(arm_cmpfp0 SPR:$Sd, (i32 0))]> { + let Inst{3-0} = 0b0000; + let Inst{5} = 0; + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VCMPZH : AHuI<0b11101, 0b11, 0b0101, 0b01, 0, + (outs), (ins HPR:$Sd), + IIC_fpCMP16, "vcmp", ".f16\t$Sd, #0", + [(arm_cmpfp0 HPR:$Sd, (i32 0))]> { + let Inst{3-0} = 0b0000; + let Inst{5} = 0; +} +} // Defs = [FPSCR_NZCV] + +def VCVTDS : ASuI<0b11101, 0b11, 0b0111, 0b11, 0, + (outs DPR:$Dd), (ins SPR:$Sm), + IIC_fpCVTDS, "vcvt", ".f64.f32\t$Dd, $Sm", + [(set DPR:$Dd, (fpextend SPR:$Sm))]>, + Sched<[WriteFPCVT]> { + // Instruction operands. + bits<5> Dd; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Dd{3-0}; + let Inst{22} = Dd{4}; + + let Predicates = [HasVFP2, HasDPVFP]; +} + +// Special case encoding: bits 11-8 is 0b1011. +def VCVTSD : VFPAI<(outs SPR:$Sd), (ins DPR:$Dm), VFPUnaryFrm, + IIC_fpCVTSD, "vcvt", ".f32.f64\t$Sd, $Dm", + [(set SPR:$Sd, (fpround DPR:$Dm))]>, + Sched<[WriteFPCVT]> { + // Instruction operands. + bits<5> Sd; + bits<5> Dm; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Inst{27-23} = 0b11101; + let Inst{21-16} = 0b110111; + let Inst{11-8} = 0b1011; + let Inst{7-6} = 0b11; + let Inst{4} = 0; + + let Predicates = [HasVFP2, HasDPVFP]; +} + +// Between half, single and double-precision. +def VCVTBHS: ASuI<0b11101, 0b11, 0b0010, 0b01, 0, (outs SPR:$Sd), (ins SPR:$Sm), + /* FIXME */ IIC_fpCVTSH, "vcvtb", ".f32.f16\t$Sd, $Sm", + [/* Intentionally left blank, see patterns below */]>, + Requires<[HasFP16]>, + Sched<[WriteFPCVT]>; + +def : FullFP16Pat<(f32 (fpextend HPR:$Sm)), + (VCVTBHS (COPY_TO_REGCLASS HPR:$Sm, SPR))>; +def : FP16Pat<(f16_to_fp GPR:$a), + (VCVTBHS (COPY_TO_REGCLASS GPR:$a, SPR))>; + +def VCVTBSH: ASuI<0b11101, 0b11, 0b0011, 0b01, 0, (outs SPR:$Sd), (ins SPR:$Sm), + /* FIXME */ IIC_fpCVTHS, "vcvtb", ".f16.f32\t$Sd, $Sm", + [/* Intentionally left blank, see patterns below */]>, + Requires<[HasFP16]>, + Sched<[WriteFPCVT]>; + +def : FullFP16Pat<(f16 (fpround SPR:$Sm)), + (COPY_TO_REGCLASS (VCVTBSH SPR:$Sm), HPR)>; +def : FP16Pat<(fp_to_f16 SPR:$a), + (i32 (COPY_TO_REGCLASS (VCVTBSH SPR:$a), GPR))>; + +def VCVTTHS: ASuI<0b11101, 0b11, 0b0010, 0b11, 0, (outs SPR:$Sd), (ins SPR:$Sm), + /* FIXME */ IIC_fpCVTSH, "vcvtt", ".f32.f16\t$Sd, $Sm", + [/* For disassembly only; pattern left blank */]>, + Requires<[HasFP16]>, + Sched<[WriteFPCVT]>; + +def VCVTTSH: ASuI<0b11101, 0b11, 0b0011, 0b11, 0, (outs SPR:$Sd), (ins SPR:$Sm), + /* FIXME */ IIC_fpCVTHS, "vcvtt", ".f16.f32\t$Sd, $Sm", + [/* For disassembly only; pattern left blank */]>, + Requires<[HasFP16]>, + Sched<[WriteFPCVT]>; + +def VCVTBHD : ADuI<0b11101, 0b11, 0b0010, 0b01, 0, + (outs DPR:$Dd), (ins SPR:$Sm), + NoItinerary, "vcvtb", ".f64.f16\t$Dd, $Sm", + [/* Intentionally left blank, see patterns below */]>, + Requires<[HasFPARMv8, HasDPVFP]>, + Sched<[WriteFPCVT]> { + // Instruction operands. + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; +} + +def : FullFP16Pat<(f64 (fpextend HPR:$Sm)), + (VCVTBHD (COPY_TO_REGCLASS HPR:$Sm, SPR))>; +def : FP16Pat<(f64 (f16_to_fp GPR:$a)), + (VCVTBHD (COPY_TO_REGCLASS GPR:$a, SPR))>; + +def VCVTBDH : ADuI<0b11101, 0b11, 0b0011, 0b01, 0, + (outs SPR:$Sd), (ins DPR:$Dm), + NoItinerary, "vcvtb", ".f16.f64\t$Sd, $Dm", + [/* Intentionally left blank, see patterns below */]>, + Requires<[HasFPARMv8, HasDPVFP]> { + // Instruction operands. + bits<5> Sd; + bits<5> Dm; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; +} + +def : FullFP16Pat<(f16 (fpround DPR:$Dm)), + (COPY_TO_REGCLASS (VCVTBDH DPR:$Dm), HPR)>; +def : FP16Pat<(fp_to_f16 (f64 DPR:$a)), + (i32 (COPY_TO_REGCLASS (VCVTBDH DPR:$a), GPR))>; + +def VCVTTHD : ADuI<0b11101, 0b11, 0b0010, 0b11, 0, + (outs DPR:$Dd), (ins SPR:$Sm), + NoItinerary, "vcvtt", ".f64.f16\t$Dd, $Sm", + []>, Requires<[HasFPARMv8, HasDPVFP]> { + // Instruction operands. + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; +} + +def VCVTTDH : ADuI<0b11101, 0b11, 0b0011, 0b11, 0, + (outs SPR:$Sd), (ins DPR:$Dm), + NoItinerary, "vcvtt", ".f16.f64\t$Sd, $Dm", + []>, Requires<[HasFPARMv8, HasDPVFP]> { + // Instruction operands. + bits<5> Sd; + bits<5> Dm; + + // Encode instruction operands. + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; +} + +multiclass vcvt_inst<string opc, bits<2> rm, + SDPatternOperator node = null_frag> { + let PostEncoderMethod = "", DecoderNamespace = "VFPV8" in { + def SH : AHuInp<0b11101, 0b11, 0b1100, 0b11, 0, + (outs SPR:$Sd), (ins HPR:$Sm), + NoItinerary, !strconcat("vcvt", opc, ".s32.f16\t$Sd, $Sm"), + []>, + Requires<[HasFullFP16]> { + let Inst{17-16} = rm; + } + + def UH : AHuInp<0b11101, 0b11, 0b1100, 0b01, 0, + (outs SPR:$Sd), (ins HPR:$Sm), + NoItinerary, !strconcat("vcvt", opc, ".u32.f16\t$Sd, $Sm"), + []>, + Requires<[HasFullFP16]> { + let Inst{17-16} = rm; + } + + def SS : ASuInp<0b11101, 0b11, 0b1100, 0b11, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + NoItinerary, !strconcat("vcvt", opc, ".s32.f32\t$Sd, $Sm"), + []>, + Requires<[HasFPARMv8]> { + let Inst{17-16} = rm; + } + + def US : ASuInp<0b11101, 0b11, 0b1100, 0b01, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + NoItinerary, !strconcat("vcvt", opc, ".u32.f32\t$Sd, $Sm"), + []>, + Requires<[HasFPARMv8]> { + let Inst{17-16} = rm; + } + + def SD : ASuInp<0b11101, 0b11, 0b1100, 0b11, 0, + (outs SPR:$Sd), (ins DPR:$Dm), + NoItinerary, !strconcat("vcvt", opc, ".s32.f64\t$Sd, $Dm"), + []>, + Requires<[HasFPARMv8, HasDPVFP]> { + bits<5> Dm; + + let Inst{17-16} = rm; + + // Encode instruction operands + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{8} = 1; + } + + def UD : ASuInp<0b11101, 0b11, 0b1100, 0b01, 0, + (outs SPR:$Sd), (ins DPR:$Dm), + NoItinerary, !strconcat("vcvt", opc, ".u32.f64\t$Sd, $Dm"), + []>, + Requires<[HasFPARMv8, HasDPVFP]> { + bits<5> Dm; + + let Inst{17-16} = rm; + + // Encode instruction operands + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{8} = 1; + } + } + + let Predicates = [HasFPARMv8] in { + let Predicates = [HasFullFP16] in { + def : Pat<(i32 (fp_to_sint (node HPR:$a))), + (COPY_TO_REGCLASS + (!cast<Instruction>(NAME#"SH") HPR:$a), + GPR)>; + + def : Pat<(i32 (fp_to_uint (node HPR:$a))), + (COPY_TO_REGCLASS + (!cast<Instruction>(NAME#"UH") HPR:$a), + GPR)>; + } + def : Pat<(i32 (fp_to_sint (node SPR:$a))), + (COPY_TO_REGCLASS + (!cast<Instruction>(NAME#"SS") SPR:$a), + GPR)>; + def : Pat<(i32 (fp_to_uint (node SPR:$a))), + (COPY_TO_REGCLASS + (!cast<Instruction>(NAME#"US") SPR:$a), + GPR)>; + } + let Predicates = [HasFPARMv8, HasDPVFP] in { + def : Pat<(i32 (fp_to_sint (node (f64 DPR:$a)))), + (COPY_TO_REGCLASS + (!cast<Instruction>(NAME#"SD") DPR:$a), + GPR)>; + def : Pat<(i32 (fp_to_uint (node (f64 DPR:$a)))), + (COPY_TO_REGCLASS + (!cast<Instruction>(NAME#"UD") DPR:$a), + GPR)>; + } +} + +defm VCVTA : vcvt_inst<"a", 0b00, fround>; +defm VCVTN : vcvt_inst<"n", 0b01>; +defm VCVTP : vcvt_inst<"p", 0b10, fceil>; +defm VCVTM : vcvt_inst<"m", 0b11, ffloor>; + +def VNEGD : ADuI<0b11101, 0b11, 0b0001, 0b01, 0, + (outs DPR:$Dd), (ins DPR:$Dm), + IIC_fpUNA64, "vneg", ".f64\t$Dd, $Dm", + [(set DPR:$Dd, (fneg (f64 DPR:$Dm)))]>; + +def VNEGS : ASuIn<0b11101, 0b11, 0b0001, 0b01, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpUNA32, "vneg", ".f32\t$Sd, $Sm", + [(set SPR:$Sd, (fneg SPR:$Sm))]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VNEGH : AHuI<0b11101, 0b11, 0b0001, 0b01, 0, + (outs HPR:$Sd), (ins HPR:$Sm), + IIC_fpUNA16, "vneg", ".f16\t$Sd, $Sm", + [(set HPR:$Sd, (fneg HPR:$Sm))]>; + +multiclass vrint_inst_zrx<string opc, bit op, bit op2, SDPatternOperator node> { + def H : AHuI<0b11101, 0b11, 0b0110, 0b11, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + NoItinerary, !strconcat("vrint", opc), ".f16\t$Sd, $Sm", + []>, + Requires<[HasFullFP16]> { + let Inst{7} = op2; + let Inst{16} = op; + } + + def S : ASuI<0b11101, 0b11, 0b0110, 0b11, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + NoItinerary, !strconcat("vrint", opc), ".f32\t$Sd, $Sm", + [(set (f32 SPR:$Sd), (node (f32 SPR:$Sm)))]>, + Requires<[HasFPARMv8]> { + let Inst{7} = op2; + let Inst{16} = op; + } + def D : ADuI<0b11101, 0b11, 0b0110, 0b11, 0, + (outs DPR:$Dd), (ins DPR:$Dm), + NoItinerary, !strconcat("vrint", opc), ".f64\t$Dd, $Dm", + [(set (f64 DPR:$Dd), (node (f64 DPR:$Dm)))]>, + Requires<[HasFPARMv8, HasDPVFP]> { + let Inst{7} = op2; + let Inst{16} = op; + } + + def : InstAlias<!strconcat("vrint", opc, "$p.f16.f16\t$Sd, $Sm"), + (!cast<Instruction>(NAME#"H") SPR:$Sd, SPR:$Sm, pred:$p), 0>, + Requires<[HasFullFP16]>; + def : InstAlias<!strconcat("vrint", opc, "$p.f32.f32\t$Sd, $Sm"), + (!cast<Instruction>(NAME#"S") SPR:$Sd, SPR:$Sm, pred:$p), 0>, + Requires<[HasFPARMv8]>; + def : InstAlias<!strconcat("vrint", opc, "$p.f64.f64\t$Dd, $Dm"), + (!cast<Instruction>(NAME#"D") DPR:$Dd, DPR:$Dm, pred:$p), 0>, + Requires<[HasFPARMv8,HasDPVFP]>; +} + +defm VRINTZ : vrint_inst_zrx<"z", 0, 1, ftrunc>; +defm VRINTR : vrint_inst_zrx<"r", 0, 0, fnearbyint>; +defm VRINTX : vrint_inst_zrx<"x", 1, 0, frint>; + +multiclass vrint_inst_anpm<string opc, bits<2> rm, + SDPatternOperator node = null_frag> { + let PostEncoderMethod = "", DecoderNamespace = "VFPV8" in { + def H : AHuInp<0b11101, 0b11, 0b1000, 0b01, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + NoItinerary, !strconcat("vrint", opc, ".f16\t$Sd, $Sm"), + []>, + Requires<[HasFullFP16]> { + let Inst{17-16} = rm; + } + def S : ASuInp<0b11101, 0b11, 0b1000, 0b01, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + NoItinerary, !strconcat("vrint", opc, ".f32\t$Sd, $Sm"), + [(set (f32 SPR:$Sd), (node (f32 SPR:$Sm)))]>, + Requires<[HasFPARMv8]> { + let Inst{17-16} = rm; + } + def D : ADuInp<0b11101, 0b11, 0b1000, 0b01, 0, + (outs DPR:$Dd), (ins DPR:$Dm), + NoItinerary, !strconcat("vrint", opc, ".f64\t$Dd, $Dm"), + [(set (f64 DPR:$Dd), (node (f64 DPR:$Dm)))]>, + Requires<[HasFPARMv8, HasDPVFP]> { + let Inst{17-16} = rm; + } + } + + def : InstAlias<!strconcat("vrint", opc, ".f32.f32\t$Sd, $Sm"), + (!cast<Instruction>(NAME#"S") SPR:$Sd, SPR:$Sm), 0>, + Requires<[HasFPARMv8]>; + def : InstAlias<!strconcat("vrint", opc, ".f64.f64\t$Dd, $Dm"), + (!cast<Instruction>(NAME#"D") DPR:$Dd, DPR:$Dm), 0>, + Requires<[HasFPARMv8,HasDPVFP]>; +} + +defm VRINTA : vrint_inst_anpm<"a", 0b00, fround>; +defm VRINTN : vrint_inst_anpm<"n", 0b01, int_arm_neon_vrintn>; +defm VRINTP : vrint_inst_anpm<"p", 0b10, fceil>; +defm VRINTM : vrint_inst_anpm<"m", 0b11, ffloor>; + +def VSQRTD : ADuI<0b11101, 0b11, 0b0001, 0b11, 0, + (outs DPR:$Dd), (ins DPR:$Dm), + IIC_fpSQRT64, "vsqrt", ".f64\t$Dd, $Dm", + [(set DPR:$Dd, (fsqrt (f64 DPR:$Dm)))]>, + Sched<[WriteFPSQRT64]>; + +def VSQRTS : ASuI<0b11101, 0b11, 0b0001, 0b11, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpSQRT32, "vsqrt", ".f32\t$Sd, $Sm", + [(set SPR:$Sd, (fsqrt SPR:$Sm))]>, + Sched<[WriteFPSQRT32]>; + +def VSQRTH : AHuI<0b11101, 0b11, 0b0001, 0b11, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpSQRT16, "vsqrt", ".f16\t$Sd, $Sm", + []>; + +let hasSideEffects = 0 in { +let isMoveReg = 1 in { +def VMOVD : ADuI<0b11101, 0b11, 0b0000, 0b01, 0, + (outs DPR:$Dd), (ins DPR:$Dm), + IIC_fpUNA64, "vmov", ".f64\t$Dd, $Dm", []>; + +def VMOVS : ASuI<0b11101, 0b11, 0b0000, 0b01, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpUNA32, "vmov", ".f32\t$Sd, $Sm", []>; +} // isMoveReg + +let PostEncoderMethod = "", DecoderNamespace = "VFPV8" in { +def VMOVH : ASuInp<0b11101, 0b11, 0b0000, 0b01, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpUNA16, "vmovx.f16\t$Sd, $Sm", []>, + Requires<[HasFullFP16]>; + +def VINSH : ASuInp<0b11101, 0b11, 0b0000, 0b11, 0, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpUNA16, "vins.f16\t$Sd, $Sm", []>, + Requires<[HasFullFP16]>; +} // PostEncoderMethod +} // hasSideEffects + +//===----------------------------------------------------------------------===// +// FP <-> GPR Copies. Int <-> FP Conversions. +// + +let isMoveReg = 1 in { +def VMOVRS : AVConv2I<0b11100001, 0b1010, + (outs GPR:$Rt), (ins SPR:$Sn), + IIC_fpMOVSI, "vmov", "\t$Rt, $Sn", + [(set GPR:$Rt, (bitconvert SPR:$Sn))]>, + Sched<[WriteFPMOV]> { + // Instruction operands. + bits<4> Rt; + bits<5> Sn; + + // Encode instruction operands. + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Rt; + + let Inst{6-5} = 0b00; + let Inst{3-0} = 0b0000; + + // Some single precision VFP instructions may be executed on both NEON and VFP + // pipelines. + let D = VFPNeonDomain; +} + +// Bitcast i32 -> f32. NEON prefers to use VMOVDRR. +def VMOVSR : AVConv4I<0b11100000, 0b1010, + (outs SPR:$Sn), (ins GPR:$Rt), + IIC_fpMOVIS, "vmov", "\t$Sn, $Rt", + [(set SPR:$Sn, (bitconvert GPR:$Rt))]>, + Requires<[HasVFP2, UseVMOVSR]>, + Sched<[WriteFPMOV]> { + // Instruction operands. + bits<5> Sn; + bits<4> Rt; + + // Encode instruction operands. + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Rt; + + let Inst{6-5} = 0b00; + let Inst{3-0} = 0b0000; + + // Some single precision VFP instructions may be executed on both NEON and VFP + // pipelines. + let D = VFPNeonDomain; +} +} // isMoveReg +def : Pat<(arm_vmovsr GPR:$Rt), (VMOVSR GPR:$Rt)>, Requires<[HasVFP2, UseVMOVSR]>; + +let hasSideEffects = 0 in { +def VMOVRRD : AVConv3I<0b11000101, 0b1011, + (outs GPR:$Rt, GPR:$Rt2), (ins DPR:$Dm), + IIC_fpMOVDI, "vmov", "\t$Rt, $Rt2, $Dm", + [(set GPR:$Rt, GPR:$Rt2, (arm_fmrrd DPR:$Dm))]>, + Sched<[WriteFPMOV]> { + // Instruction operands. + bits<5> Dm; + bits<4> Rt; + bits<4> Rt2; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{15-12} = Rt; + let Inst{19-16} = Rt2; + + let Inst{7-6} = 0b00; + + // Some single precision VFP instructions may be executed on both NEON and VFP + // pipelines. + let D = VFPNeonDomain; + + // This instruction is equivalent to + // $Rt = EXTRACT_SUBREG $Dm, ssub_0 + // $Rt2 = EXTRACT_SUBREG $Dm, ssub_1 + let isExtractSubreg = 1; +} + +def VMOVRRS : AVConv3I<0b11000101, 0b1010, + (outs GPR:$Rt, GPR:$Rt2), (ins SPR:$src1, SPR:$src2), + IIC_fpMOVDI, "vmov", "\t$Rt, $Rt2, $src1, $src2", + [/* For disassembly only; pattern left blank */]>, + Sched<[WriteFPMOV]> { + bits<5> src1; + bits<4> Rt; + bits<4> Rt2; + + // Encode instruction operands. + let Inst{3-0} = src1{4-1}; + let Inst{5} = src1{0}; + let Inst{15-12} = Rt; + let Inst{19-16} = Rt2; + + let Inst{7-6} = 0b00; + + // Some single precision VFP instructions may be executed on both NEON and VFP + // pipelines. + let D = VFPNeonDomain; + let DecoderMethod = "DecodeVMOVRRS"; +} +} // hasSideEffects + +// FMDHR: GPR -> SPR +// FMDLR: GPR -> SPR + +def VMOVDRR : AVConv5I<0b11000100, 0b1011, + (outs DPR:$Dm), (ins GPR:$Rt, GPR:$Rt2), + IIC_fpMOVID, "vmov", "\t$Dm, $Rt, $Rt2", + [(set DPR:$Dm, (arm_fmdrr GPR:$Rt, GPR:$Rt2))]>, + Sched<[WriteFPMOV]> { + // Instruction operands. + bits<5> Dm; + bits<4> Rt; + bits<4> Rt2; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{15-12} = Rt; + let Inst{19-16} = Rt2; + + let Inst{7-6} = 0b00; + + // Some single precision VFP instructions may be executed on both NEON and VFP + // pipelines. + let D = VFPNeonDomain; + + // This instruction is equivalent to + // $Dm = REG_SEQUENCE $Rt, ssub_0, $Rt2, ssub_1 + let isRegSequence = 1; +} + +// Hoist an fabs or a fneg of a value coming from integer registers +// and do the fabs/fneg on the integer value. This is never a lose +// and could enable the conversion to float to be removed completely. +def : Pat<(fabs (arm_fmdrr GPR:$Rl, GPR:$Rh)), + (VMOVDRR GPR:$Rl, (BFC GPR:$Rh, (i32 0x7FFFFFFF)))>, + Requires<[IsARM, HasV6T2]>; +def : Pat<(fabs (arm_fmdrr GPR:$Rl, GPR:$Rh)), + (VMOVDRR GPR:$Rl, (t2BFC GPR:$Rh, (i32 0x7FFFFFFF)))>, + Requires<[IsThumb2, HasV6T2]>; +def : Pat<(fneg (arm_fmdrr GPR:$Rl, GPR:$Rh)), + (VMOVDRR GPR:$Rl, (EORri GPR:$Rh, (i32 0x80000000)))>, + Requires<[IsARM]>; +def : Pat<(fneg (arm_fmdrr GPR:$Rl, GPR:$Rh)), + (VMOVDRR GPR:$Rl, (t2EORri GPR:$Rh, (i32 0x80000000)))>, + Requires<[IsThumb2]>; + +let hasSideEffects = 0 in +def VMOVSRR : AVConv5I<0b11000100, 0b1010, + (outs SPR:$dst1, SPR:$dst2), (ins GPR:$src1, GPR:$src2), + IIC_fpMOVID, "vmov", "\t$dst1, $dst2, $src1, $src2", + [/* For disassembly only; pattern left blank */]>, + Sched<[WriteFPMOV]> { + // Instruction operands. + bits<5> dst1; + bits<4> src1; + bits<4> src2; + + // Encode instruction operands. + let Inst{3-0} = dst1{4-1}; + let Inst{5} = dst1{0}; + let Inst{15-12} = src1; + let Inst{19-16} = src2; + + let Inst{7-6} = 0b00; + + // Some single precision VFP instructions may be executed on both NEON and VFP + // pipelines. + let D = VFPNeonDomain; + + let DecoderMethod = "DecodeVMOVSRR"; +} + +// Move H->R, clearing top 16 bits +def VMOVRH : AVConv2I<0b11100001, 0b1001, + (outs GPR:$Rt), (ins HPR:$Sn), + IIC_fpMOVSI, "vmov", ".f16\t$Rt, $Sn", + [(set GPR:$Rt, (arm_vmovrh HPR:$Sn))]>, + Requires<[HasFullFP16]>, + Sched<[WriteFPMOV]> { + // Instruction operands. + bits<4> Rt; + bits<5> Sn; + + // Encode instruction operands. + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Rt; + + let Inst{6-5} = 0b00; + let Inst{3-0} = 0b0000; +} + +// Move R->H, clearing top 16 bits +def VMOVHR : AVConv4I<0b11100000, 0b1001, + (outs HPR:$Sn), (ins GPR:$Rt), + IIC_fpMOVIS, "vmov", ".f16\t$Sn, $Rt", + [(set HPR:$Sn, (arm_vmovhr GPR:$Rt))]>, + Requires<[HasFullFP16]>, + Sched<[WriteFPMOV]> { + // Instruction operands. + bits<5> Sn; + bits<4> Rt; + + // Encode instruction operands. + let Inst{19-16} = Sn{4-1}; + let Inst{7} = Sn{0}; + let Inst{15-12} = Rt; + + let Inst{6-5} = 0b00; + let Inst{3-0} = 0b0000; +} + +// FMRDH: SPR -> GPR +// FMRDL: SPR -> GPR +// FMRRS: SPR -> GPR +// FMRX: SPR system reg -> GPR +// FMSRR: GPR -> SPR +// FMXR: GPR -> VFP system reg + + +// Int -> FP: + +class AVConv1IDs_Encode<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, + bits<4> opcod4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : AVConv1I<opcod1, opcod2, opcod3, opcod4, oops, iops, itin, opc, asm, + pattern> { + // Instruction operands. + bits<5> Dd; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Dd{3-0}; + let Inst{22} = Dd{4}; + + let Predicates = [HasVFP2, HasDPVFP]; +} + +class AVConv1InSs_Encode<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, + bits<4> opcod4, dag oops, dag iops,InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AVConv1In<opcod1, opcod2, opcod3, opcod4, oops, iops, itin, opc, asm, + pattern> { + // Instruction operands. + bits<5> Sd; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; +} + +class AVConv1IHs_Encode<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, + bits<4> opcod4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : AVConv1I<opcod1, opcod2, opcod3, opcod4, oops, iops, itin, opc, asm, + pattern> { + // Instruction operands. + bits<5> Sd; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Predicates = [HasFullFP16]; +} + +def VSITOD : AVConv1IDs_Encode<0b11101, 0b11, 0b1000, 0b1011, + (outs DPR:$Dd), (ins SPR:$Sm), + IIC_fpCVTID, "vcvt", ".f64.s32\t$Dd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // s32 +} + +let Predicates=[HasVFP2, HasDPVFP] in { + def : VFPPat<(f64 (sint_to_fp GPR:$a)), + (VSITOD (COPY_TO_REGCLASS GPR:$a, SPR))>; + + def : VFPPat<(f64 (sint_to_fp (i32 (alignedload32 addrmode5:$a)))), + (VSITOD (VLDRS addrmode5:$a))>; +} + +def VSITOS : AVConv1InSs_Encode<0b11101, 0b11, 0b1000, 0b1010, + (outs SPR:$Sd),(ins SPR:$Sm), + IIC_fpCVTIS, "vcvt", ".f32.s32\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // s32 + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def : VFPNoNEONPat<(f32 (sint_to_fp GPR:$a)), + (VSITOS (COPY_TO_REGCLASS GPR:$a, SPR))>; + +def : VFPNoNEONPat<(f32 (sint_to_fp (i32 (alignedload32 addrmode5:$a)))), + (VSITOS (VLDRS addrmode5:$a))>; + +def VSITOH : AVConv1IHs_Encode<0b11101, 0b11, 0b1000, 0b1001, + (outs HPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTIH, "vcvt", ".f16.s32\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // s32 +} + +def : VFPNoNEONPat<(f16 (sint_to_fp GPR:$a)), + (VSITOH (COPY_TO_REGCLASS GPR:$a, SPR))>; + +def VUITOD : AVConv1IDs_Encode<0b11101, 0b11, 0b1000, 0b1011, + (outs DPR:$Dd), (ins SPR:$Sm), + IIC_fpCVTID, "vcvt", ".f64.u32\t$Dd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // u32 +} + +let Predicates=[HasVFP2, HasDPVFP] in { + def : VFPPat<(f64 (uint_to_fp GPR:$a)), + (VUITOD (COPY_TO_REGCLASS GPR:$a, SPR))>; + + def : VFPPat<(f64 (uint_to_fp (i32 (alignedload32 addrmode5:$a)))), + (VUITOD (VLDRS addrmode5:$a))>; +} + +def VUITOS : AVConv1InSs_Encode<0b11101, 0b11, 0b1000, 0b1010, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTIS, "vcvt", ".f32.u32\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // u32 + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def : VFPNoNEONPat<(f32 (uint_to_fp GPR:$a)), + (VUITOS (COPY_TO_REGCLASS GPR:$a, SPR))>; + +def : VFPNoNEONPat<(f32 (uint_to_fp (i32 (alignedload32 addrmode5:$a)))), + (VUITOS (VLDRS addrmode5:$a))>; + +def VUITOH : AVConv1IHs_Encode<0b11101, 0b11, 0b1000, 0b1001, + (outs HPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTIH, "vcvt", ".f16.u32\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // u32 +} + +def : VFPNoNEONPat<(f16 (uint_to_fp GPR:$a)), + (VUITOH (COPY_TO_REGCLASS GPR:$a, SPR))>; + +// FP -> Int: + +class AVConv1IsD_Encode<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, + bits<4> opcod4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : AVConv1I<opcod1, opcod2, opcod3, opcod4, oops, iops, itin, opc, asm, + pattern> { + // Instruction operands. + bits<5> Sd; + bits<5> Dm; + + // Encode instruction operands. + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Predicates = [HasVFP2, HasDPVFP]; +} + +class AVConv1InsS_Encode<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, + bits<4> opcod4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : AVConv1In<opcod1, opcod2, opcod3, opcod4, oops, iops, itin, opc, asm, + pattern> { + // Instruction operands. + bits<5> Sd; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; +} + +class AVConv1IsH_Encode<bits<5> opcod1, bits<2> opcod2, bits<4> opcod3, + bits<4> opcod4, dag oops, dag iops, + InstrItinClass itin, string opc, string asm, + list<dag> pattern> + : AVConv1I<opcod1, opcod2, opcod3, opcod4, oops, iops, itin, opc, asm, + pattern> { + // Instruction operands. + bits<5> Sd; + bits<5> Sm; + + // Encode instruction operands. + let Inst{3-0} = Sm{4-1}; + let Inst{5} = Sm{0}; + let Inst{15-12} = Sd{4-1}; + let Inst{22} = Sd{0}; + + let Predicates = [HasFullFP16]; +} + +// Always set Z bit in the instruction, i.e. "round towards zero" variants. +def VTOSIZD : AVConv1IsD_Encode<0b11101, 0b11, 0b1101, 0b1011, + (outs SPR:$Sd), (ins DPR:$Dm), + IIC_fpCVTDI, "vcvt", ".s32.f64\t$Sd, $Dm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // Z bit +} + +let Predicates=[HasVFP2, HasDPVFP] in { + def : VFPPat<(i32 (fp_to_sint (f64 DPR:$a))), + (COPY_TO_REGCLASS (VTOSIZD DPR:$a), GPR)>; + + def : VFPPat<(alignedstore32 (i32 (fp_to_sint (f64 DPR:$a))), addrmode5:$ptr), + (VSTRS (VTOSIZD DPR:$a), addrmode5:$ptr)>; +} + +def VTOSIZS : AVConv1InsS_Encode<0b11101, 0b11, 0b1101, 0b1010, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTSI, "vcvt", ".s32.f32\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // Z bit + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def : VFPNoNEONPat<(i32 (fp_to_sint SPR:$a)), + (COPY_TO_REGCLASS (VTOSIZS SPR:$a), GPR)>; + +def : VFPNoNEONPat<(alignedstore32 (i32 (fp_to_sint (f32 SPR:$a))), + addrmode5:$ptr), + (VSTRS (VTOSIZS SPR:$a), addrmode5:$ptr)>; + +def VTOSIZH : AVConv1IsH_Encode<0b11101, 0b11, 0b1101, 0b1001, + (outs SPR:$Sd), (ins HPR:$Sm), + IIC_fpCVTHI, "vcvt", ".s32.f16\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // Z bit +} + +def : VFPNoNEONPat<(i32 (fp_to_sint HPR:$a)), + (COPY_TO_REGCLASS (VTOSIZH HPR:$a), GPR)>; + +def VTOUIZD : AVConv1IsD_Encode<0b11101, 0b11, 0b1100, 0b1011, + (outs SPR:$Sd), (ins DPR:$Dm), + IIC_fpCVTDI, "vcvt", ".u32.f64\t$Sd, $Dm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // Z bit +} + +let Predicates=[HasVFP2, HasDPVFP] in { + def : VFPPat<(i32 (fp_to_uint (f64 DPR:$a))), + (COPY_TO_REGCLASS (VTOUIZD DPR:$a), GPR)>; + + def : VFPPat<(alignedstore32 (i32 (fp_to_uint (f64 DPR:$a))), addrmode5:$ptr), + (VSTRS (VTOUIZD DPR:$a), addrmode5:$ptr)>; +} + +def VTOUIZS : AVConv1InsS_Encode<0b11101, 0b11, 0b1100, 0b1010, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTSI, "vcvt", ".u32.f32\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // Z bit + + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def : VFPNoNEONPat<(i32 (fp_to_uint SPR:$a)), + (COPY_TO_REGCLASS (VTOUIZS SPR:$a), GPR)>; + +def : VFPNoNEONPat<(alignedstore32 (i32 (fp_to_uint (f32 SPR:$a))), + addrmode5:$ptr), + (VSTRS (VTOUIZS SPR:$a), addrmode5:$ptr)>; + +def VTOUIZH : AVConv1IsH_Encode<0b11101, 0b11, 0b1100, 0b1001, + (outs SPR:$Sd), (ins HPR:$Sm), + IIC_fpCVTHI, "vcvt", ".u32.f16\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 1; // Z bit +} + +def : VFPNoNEONPat<(i32 (fp_to_uint HPR:$a)), + (COPY_TO_REGCLASS (VTOUIZH HPR:$a), GPR)>; + +// And the Z bit '0' variants, i.e. use the rounding mode specified by FPSCR. +let Uses = [FPSCR] in { +def VTOSIRD : AVConv1IsD_Encode<0b11101, 0b11, 0b1101, 0b1011, + (outs SPR:$Sd), (ins DPR:$Dm), + IIC_fpCVTDI, "vcvtr", ".s32.f64\t$Sd, $Dm", + [(set SPR:$Sd, (int_arm_vcvtr (f64 DPR:$Dm)))]>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // Z bit +} + +def VTOSIRS : AVConv1InsS_Encode<0b11101, 0b11, 0b1101, 0b1010, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTSI, "vcvtr", ".s32.f32\t$Sd, $Sm", + [(set SPR:$Sd, (int_arm_vcvtr SPR:$Sm))]>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // Z bit +} + +def VTOSIRH : AVConv1IsH_Encode<0b11101, 0b11, 0b1101, 0b1001, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTHI, "vcvtr", ".s32.f16\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // Z bit +} + +def VTOUIRD : AVConv1IsD_Encode<0b11101, 0b11, 0b1100, 0b1011, + (outs SPR:$Sd), (ins DPR:$Dm), + IIC_fpCVTDI, "vcvtr", ".u32.f64\t$Sd, $Dm", + [(set SPR:$Sd, (int_arm_vcvtru(f64 DPR:$Dm)))]>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // Z bit +} + +def VTOUIRS : AVConv1InsS_Encode<0b11101, 0b11, 0b1100, 0b1010, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTSI, "vcvtr", ".u32.f32\t$Sd, $Sm", + [(set SPR:$Sd, (int_arm_vcvtru SPR:$Sm))]>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // Z bit +} + +def VTOUIRH : AVConv1IsH_Encode<0b11101, 0b11, 0b1100, 0b1001, + (outs SPR:$Sd), (ins SPR:$Sm), + IIC_fpCVTHI, "vcvtr", ".u32.f16\t$Sd, $Sm", + []>, + Sched<[WriteFPCVT]> { + let Inst{7} = 0; // Z bit +} +} + +// v8.3-a Javascript Convert to Signed fixed-point +def VJCVT : AVConv1IsD_Encode<0b11101, 0b11, 0b1001, 0b1011, + (outs SPR:$Sd), (ins DPR:$Dm), + IIC_fpCVTDI, "vjcvt", ".s32.f64\t$Sd, $Dm", + []>, + Requires<[HasFPARMv8, HasV8_3a]> { + let Inst{7} = 1; // Z bit +} + +// Convert between floating-point and fixed-point +// Data type for fixed-point naming convention: +// S16 (U=0, sx=0) -> SH +// U16 (U=1, sx=0) -> UH +// S32 (U=0, sx=1) -> SL +// U32 (U=1, sx=1) -> UL + +let Constraints = "$a = $dst" in { + +// FP to Fixed-Point: + +// Single Precision register +class AVConv1XInsS_Encode<bits<5> op1, bits<2> op2, bits<4> op3, bits<4> op4, + bit op5, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AVConv1XI<op1, op2, op3, op4, op5, oops, iops, itin, opc, asm, pattern> { + bits<5> dst; + // if dp_operation then UInt(D:Vd) else UInt(Vd:D); + let Inst{22} = dst{0}; + let Inst{15-12} = dst{4-1}; +} + +// Double Precision register +class AVConv1XInsD_Encode<bits<5> op1, bits<2> op2, bits<4> op3, bits<4> op4, + bit op5, dag oops, dag iops, InstrItinClass itin, + string opc, string asm, list<dag> pattern> + : AVConv1XI<op1, op2, op3, op4, op5, oops, iops, itin, opc, asm, pattern> { + bits<5> dst; + // if dp_operation then UInt(D:Vd) else UInt(Vd:D); + let Inst{22} = dst{4}; + let Inst{15-12} = dst{3-0}; + + let Predicates = [HasVFP2, HasDPVFP]; +} + +def VTOSHH : AVConv1XInsS_Encode<0b11101, 0b11, 0b1110, 0b1001, 0, + (outs SPR:$dst), (ins SPR:$a, fbits16:$fbits), + IIC_fpCVTHI, "vcvt", ".s16.f16\t$dst, $a, $fbits", []>, + Requires<[HasFullFP16]>, + Sched<[WriteFPCVT]>; + +def VTOUHH : AVConv1XInsS_Encode<0b11101, 0b11, 0b1111, 0b1001, 0, + (outs SPR:$dst), (ins SPR:$a, fbits16:$fbits), + IIC_fpCVTHI, "vcvt", ".u16.f16\t$dst, $a, $fbits", []>, + Requires<[HasFullFP16]>, + Sched<[WriteFPCVT]>; + +def VTOSLH : AVConv1XInsS_Encode<0b11101, 0b11, 0b1110, 0b1001, 1, + (outs SPR:$dst), (ins SPR:$a, fbits32:$fbits), + IIC_fpCVTHI, "vcvt", ".s32.f16\t$dst, $a, $fbits", []>, + Requires<[HasFullFP16]>, + Sched<[WriteFPCVT]>; + +def VTOULH : AVConv1XInsS_Encode<0b11101, 0b11, 0b1111, 0b1001, 1, + (outs SPR:$dst), (ins SPR:$a, fbits32:$fbits), + IIC_fpCVTHI, "vcvt", ".u32.f16\t$dst, $a, $fbits", []>, + Requires<[HasFullFP16]>, + Sched<[WriteFPCVT]>; + +def VTOSHS : AVConv1XInsS_Encode<0b11101, 0b11, 0b1110, 0b1010, 0, + (outs SPR:$dst), (ins SPR:$a, fbits16:$fbits), + IIC_fpCVTSI, "vcvt", ".s16.f32\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VTOUHS : AVConv1XInsS_Encode<0b11101, 0b11, 0b1111, 0b1010, 0, + (outs SPR:$dst), (ins SPR:$a, fbits16:$fbits), + IIC_fpCVTSI, "vcvt", ".u16.f32\t$dst, $a, $fbits", []> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VTOSLS : AVConv1XInsS_Encode<0b11101, 0b11, 0b1110, 0b1010, 1, + (outs SPR:$dst), (ins SPR:$a, fbits32:$fbits), + IIC_fpCVTSI, "vcvt", ".s32.f32\t$dst, $a, $fbits", []> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VTOULS : AVConv1XInsS_Encode<0b11101, 0b11, 0b1111, 0b1010, 1, + (outs SPR:$dst), (ins SPR:$a, fbits32:$fbits), + IIC_fpCVTSI, "vcvt", ".u32.f32\t$dst, $a, $fbits", []> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VTOSHD : AVConv1XInsD_Encode<0b11101, 0b11, 0b1110, 0b1011, 0, + (outs DPR:$dst), (ins DPR:$a, fbits16:$fbits), + IIC_fpCVTDI, "vcvt", ".s16.f64\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]>; + +def VTOUHD : AVConv1XInsD_Encode<0b11101, 0b11, 0b1111, 0b1011, 0, + (outs DPR:$dst), (ins DPR:$a, fbits16:$fbits), + IIC_fpCVTDI, "vcvt", ".u16.f64\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]>; + +def VTOSLD : AVConv1XInsD_Encode<0b11101, 0b11, 0b1110, 0b1011, 1, + (outs DPR:$dst), (ins DPR:$a, fbits32:$fbits), + IIC_fpCVTDI, "vcvt", ".s32.f64\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]>; + +def VTOULD : AVConv1XInsD_Encode<0b11101, 0b11, 0b1111, 0b1011, 1, + (outs DPR:$dst), (ins DPR:$a, fbits32:$fbits), + IIC_fpCVTDI, "vcvt", ".u32.f64\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]>; + +// Fixed-Point to FP: + +def VSHTOH : AVConv1XInsS_Encode<0b11101, 0b11, 0b1010, 0b1001, 0, + (outs SPR:$dst), (ins SPR:$a, fbits16:$fbits), + IIC_fpCVTIH, "vcvt", ".f16.s16\t$dst, $a, $fbits", []>, + Requires<[HasFullFP16]>, + Sched<[WriteFPCVT]>; + +def VUHTOH : AVConv1XInsS_Encode<0b11101, 0b11, 0b1011, 0b1001, 0, + (outs SPR:$dst), (ins SPR:$a, fbits16:$fbits), + IIC_fpCVTIH, "vcvt", ".f16.u16\t$dst, $a, $fbits", []>, + Requires<[HasFullFP16]>, + Sched<[WriteFPCVT]>; + +def VSLTOH : AVConv1XInsS_Encode<0b11101, 0b11, 0b1010, 0b1001, 1, + (outs SPR:$dst), (ins SPR:$a, fbits32:$fbits), + IIC_fpCVTIH, "vcvt", ".f16.s32\t$dst, $a, $fbits", []>, + Requires<[HasFullFP16]>, + Sched<[WriteFPCVT]>; + +def VULTOH : AVConv1XInsS_Encode<0b11101, 0b11, 0b1011, 0b1001, 1, + (outs SPR:$dst), (ins SPR:$a, fbits32:$fbits), + IIC_fpCVTIH, "vcvt", ".f16.u32\t$dst, $a, $fbits", []>, + Requires<[HasFullFP16]>, + Sched<[WriteFPCVT]>; + +def VSHTOS : AVConv1XInsS_Encode<0b11101, 0b11, 0b1010, 0b1010, 0, + (outs SPR:$dst), (ins SPR:$a, fbits16:$fbits), + IIC_fpCVTIS, "vcvt", ".f32.s16\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VUHTOS : AVConv1XInsS_Encode<0b11101, 0b11, 0b1011, 0b1010, 0, + (outs SPR:$dst), (ins SPR:$a, fbits16:$fbits), + IIC_fpCVTIS, "vcvt", ".f32.u16\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VSLTOS : AVConv1XInsS_Encode<0b11101, 0b11, 0b1010, 0b1010, 1, + (outs SPR:$dst), (ins SPR:$a, fbits32:$fbits), + IIC_fpCVTIS, "vcvt", ".f32.s32\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VULTOS : AVConv1XInsS_Encode<0b11101, 0b11, 0b1011, 0b1010, 1, + (outs SPR:$dst), (ins SPR:$a, fbits32:$fbits), + IIC_fpCVTIS, "vcvt", ".f32.u32\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VSHTOD : AVConv1XInsD_Encode<0b11101, 0b11, 0b1010, 0b1011, 0, + (outs DPR:$dst), (ins DPR:$a, fbits16:$fbits), + IIC_fpCVTID, "vcvt", ".f64.s16\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]>; + +def VUHTOD : AVConv1XInsD_Encode<0b11101, 0b11, 0b1011, 0b1011, 0, + (outs DPR:$dst), (ins DPR:$a, fbits16:$fbits), + IIC_fpCVTID, "vcvt", ".f64.u16\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]>; + +def VSLTOD : AVConv1XInsD_Encode<0b11101, 0b11, 0b1010, 0b1011, 1, + (outs DPR:$dst), (ins DPR:$a, fbits32:$fbits), + IIC_fpCVTID, "vcvt", ".f64.s32\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]>; + +def VULTOD : AVConv1XInsD_Encode<0b11101, 0b11, 0b1011, 0b1011, 1, + (outs DPR:$dst), (ins DPR:$a, fbits32:$fbits), + IIC_fpCVTID, "vcvt", ".f64.u32\t$dst, $a, $fbits", []>, + Sched<[WriteFPCVT]>; + +} // End of 'let Constraints = "$a = $dst" in' + +//===----------------------------------------------------------------------===// +// FP Multiply-Accumulate Operations. +// + +def VMLAD : ADbI<0b11100, 0b00, 0, 0, + (outs DPR:$Dd), (ins DPR:$Ddin, DPR:$Dn, DPR:$Dm), + IIC_fpMAC64, "vmla", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fadd_mlx (fmul_su DPR:$Dn, DPR:$Dm), + (f64 DPR:$Ddin)))]>, + RegConstraint<"$Ddin = $Dd">, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>, + Sched<[WriteFPMAC64, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def VMLAS : ASbIn<0b11100, 0b00, 0, 0, + (outs SPR:$Sd), (ins SPR:$Sdin, SPR:$Sn, SPR:$Sm), + IIC_fpMAC32, "vmla", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fadd_mlx (fmul_su SPR:$Sn, SPR:$Sm), + SPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasVFP2,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VMLAH : AHbI<0b11100, 0b00, 0, 0, + (outs HPR:$Sd), (ins HPR:$Sdin, HPR:$Sn, HPR:$Sm), + IIC_fpMAC16, "vmla", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fadd_mlx (fmul_su HPR:$Sn, HPR:$Sm), + HPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasFullFP16,UseFPVMLx,DontUseFusedMAC]>; + +def : Pat<(fadd_mlx DPR:$dstin, (fmul_su DPR:$a, (f64 DPR:$b))), + (VMLAD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fadd_mlx SPR:$dstin, (fmul_su SPR:$a, SPR:$b)), + (VMLAS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP2,DontUseNEONForFP, UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fadd_mlx HPR:$dstin, (fmul_su HPR:$a, HPR:$b)), + (VMLAH HPR:$dstin, HPR:$a, HPR:$b)>, + Requires<[HasFullFP16,DontUseNEONForFP, UseFPVMLx,DontUseFusedMAC]>; + + +def VMLSD : ADbI<0b11100, 0b00, 1, 0, + (outs DPR:$Dd), (ins DPR:$Ddin, DPR:$Dn, DPR:$Dm), + IIC_fpMAC64, "vmls", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fadd_mlx (fneg (fmul_su DPR:$Dn,DPR:$Dm)), + (f64 DPR:$Ddin)))]>, + RegConstraint<"$Ddin = $Dd">, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>, + Sched<[WriteFPMAC64, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def VMLSS : ASbIn<0b11100, 0b00, 1, 0, + (outs SPR:$Sd), (ins SPR:$Sdin, SPR:$Sn, SPR:$Sm), + IIC_fpMAC32, "vmls", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fadd_mlx (fneg (fmul_su SPR:$Sn, SPR:$Sm)), + SPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasVFP2,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VMLSH : AHbI<0b11100, 0b00, 1, 0, + (outs HPR:$Sd), (ins HPR:$Sdin, HPR:$Sn, HPR:$Sm), + IIC_fpMAC16, "vmls", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fadd_mlx (fneg (fmul_su HPR:$Sn, HPR:$Sm)), + HPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasFullFP16,UseFPVMLx,DontUseFusedMAC]>; + +def : Pat<(fsub_mlx DPR:$dstin, (fmul_su DPR:$a, (f64 DPR:$b))), + (VMLSD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fsub_mlx SPR:$dstin, (fmul_su SPR:$a, SPR:$b)), + (VMLSS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP2,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fsub_mlx HPR:$dstin, (fmul_su HPR:$a, HPR:$b)), + (VMLSH HPR:$dstin, HPR:$a, HPR:$b)>, + Requires<[HasFullFP16,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>; + +def VNMLAD : ADbI<0b11100, 0b01, 1, 0, + (outs DPR:$Dd), (ins DPR:$Ddin, DPR:$Dn, DPR:$Dm), + IIC_fpMAC64, "vnmla", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd,(fsub_mlx (fneg (fmul_su DPR:$Dn,DPR:$Dm)), + (f64 DPR:$Ddin)))]>, + RegConstraint<"$Ddin = $Dd">, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>, + Sched<[WriteFPMAC64, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def VNMLAS : ASbI<0b11100, 0b01, 1, 0, + (outs SPR:$Sd), (ins SPR:$Sdin, SPR:$Sn, SPR:$Sm), + IIC_fpMAC32, "vnmla", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fsub_mlx (fneg (fmul_su SPR:$Sn, SPR:$Sm)), + SPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasVFP2,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VNMLAH : AHbI<0b11100, 0b01, 1, 0, + (outs HPR:$Sd), (ins HPR:$Sdin, HPR:$Sn, HPR:$Sm), + IIC_fpMAC16, "vnmla", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fsub_mlx (fneg (fmul_su HPR:$Sn, HPR:$Sm)), + HPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasFullFP16,UseFPVMLx,DontUseFusedMAC]>; + +// (-(a * b) - dst) -> -(dst + (a * b)) +def : Pat<(fsub_mlx (fneg (fmul_su DPR:$a, (f64 DPR:$b))), DPR:$dstin), + (VNMLAD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fsub_mlx (fneg (fmul_su SPR:$a, SPR:$b)), SPR:$dstin), + (VNMLAS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP2,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fsub_mlx (fneg (fmul_su HPR:$a, HPR:$b)), HPR:$dstin), + (VNMLAH HPR:$dstin, HPR:$a, HPR:$b)>, + Requires<[HasFullFP16,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>; + +// (-dst - (a * b)) -> -(dst + (a * b)) +def : Pat<(fsub_mlx (fneg DPR:$dstin), (fmul_su DPR:$a, (f64 DPR:$b))), + (VNMLAD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fsub_mlx (fneg SPR:$dstin), (fmul_su SPR:$a, SPR:$b)), + (VNMLAS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP2,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fsub_mlx (fneg HPR:$dstin), (fmul_su HPR:$a, HPR:$b)), + (VNMLAH HPR:$dstin, HPR:$a, HPR:$b)>, + Requires<[HasFullFP16,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>; + +def VNMLSD : ADbI<0b11100, 0b01, 0, 0, + (outs DPR:$Dd), (ins DPR:$Ddin, DPR:$Dn, DPR:$Dm), + IIC_fpMAC64, "vnmls", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fsub_mlx (fmul_su DPR:$Dn, DPR:$Dm), + (f64 DPR:$Ddin)))]>, + RegConstraint<"$Ddin = $Dd">, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>, + Sched<[WriteFPMAC64, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def VNMLSS : ASbI<0b11100, 0b01, 0, 0, + (outs SPR:$Sd), (ins SPR:$Sdin, SPR:$Sn, SPR:$Sm), + IIC_fpMAC32, "vnmls", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fsub_mlx (fmul_su SPR:$Sn, SPR:$Sm), SPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasVFP2,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines on A8. + let D = VFPNeonA8Domain; +} + +def VNMLSH : AHbI<0b11100, 0b01, 0, 0, + (outs HPR:$Sd), (ins HPR:$Sdin, HPR:$Sn, HPR:$Sm), + IIC_fpMAC16, "vnmls", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fsub_mlx (fmul_su HPR:$Sn, HPR:$Sm), HPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasFullFP16,UseFPVMLx,DontUseFusedMAC]>; + +def : Pat<(fsub_mlx (fmul_su DPR:$a, (f64 DPR:$b)), DPR:$dstin), + (VNMLSD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP2,HasDPVFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fsub_mlx (fmul_su SPR:$a, SPR:$b), SPR:$dstin), + (VNMLSS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP2,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>; +def : Pat<(fsub_mlx (fmul_su HPR:$a, HPR:$b), HPR:$dstin), + (VNMLSH HPR:$dstin, HPR:$a, HPR:$b)>, + Requires<[HasFullFP16,DontUseNEONForFP,UseFPVMLx,DontUseFusedMAC]>; + +//===----------------------------------------------------------------------===// +// Fused FP Multiply-Accumulate Operations. +// +def VFMAD : ADbI<0b11101, 0b10, 0, 0, + (outs DPR:$Dd), (ins DPR:$Ddin, DPR:$Dn, DPR:$Dm), + IIC_fpFMAC64, "vfma", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fadd_mlx (fmul_su DPR:$Dn, DPR:$Dm), + (f64 DPR:$Ddin)))]>, + RegConstraint<"$Ddin = $Dd">, + Requires<[HasVFP4,HasDPVFP,UseFusedMAC]>, + Sched<[WriteFPMAC64, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def VFMAS : ASbIn<0b11101, 0b10, 0, 0, + (outs SPR:$Sd), (ins SPR:$Sdin, SPR:$Sn, SPR:$Sm), + IIC_fpFMAC32, "vfma", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fadd_mlx (fmul_su SPR:$Sn, SPR:$Sm), + SPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasVFP4,DontUseNEONForFP,UseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines. +} + +def VFMAH : AHbI<0b11101, 0b10, 0, 0, + (outs HPR:$Sd), (ins HPR:$Sdin, HPR:$Sn, HPR:$Sm), + IIC_fpFMAC16, "vfma", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fadd_mlx (fmul_su HPR:$Sn, HPR:$Sm), + HPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasFullFP16,UseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def : Pat<(fadd_mlx DPR:$dstin, (fmul_su DPR:$a, (f64 DPR:$b))), + (VFMAD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP4,HasDPVFP,UseFusedMAC]>; +def : Pat<(fadd_mlx SPR:$dstin, (fmul_su SPR:$a, SPR:$b)), + (VFMAS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP4,DontUseNEONForFP,UseFusedMAC]>; +def : Pat<(fadd_mlx HPR:$dstin, (fmul_su HPR:$a, HPR:$b)), + (VFMAH HPR:$dstin, HPR:$a, HPR:$b)>, + Requires<[HasFullFP16,DontUseNEONForFP,UseFusedMAC]>; + +// Match @llvm.fma.* intrinsics +// (fma x, y, z) -> (vfms z, x, y) +def : Pat<(f64 (fma DPR:$Dn, DPR:$Dm, DPR:$Ddin)), + (VFMAD DPR:$Ddin, DPR:$Dn, DPR:$Dm)>, + Requires<[HasVFP4,HasDPVFP]>; +def : Pat<(f32 (fma SPR:$Sn, SPR:$Sm, SPR:$Sdin)), + (VFMAS SPR:$Sdin, SPR:$Sn, SPR:$Sm)>, + Requires<[HasVFP4]>; + +def VFMSD : ADbI<0b11101, 0b10, 1, 0, + (outs DPR:$Dd), (ins DPR:$Ddin, DPR:$Dn, DPR:$Dm), + IIC_fpFMAC64, "vfms", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fadd_mlx (fneg (fmul_su DPR:$Dn,DPR:$Dm)), + (f64 DPR:$Ddin)))]>, + RegConstraint<"$Ddin = $Dd">, + Requires<[HasVFP4,HasDPVFP,UseFusedMAC]>, + Sched<[WriteFPMAC64, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def VFMSS : ASbIn<0b11101, 0b10, 1, 0, + (outs SPR:$Sd), (ins SPR:$Sdin, SPR:$Sn, SPR:$Sm), + IIC_fpFMAC32, "vfms", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fadd_mlx (fneg (fmul_su SPR:$Sn, SPR:$Sm)), + SPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasVFP4,DontUseNEONForFP,UseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines. +} + +def VFMSH : AHbI<0b11101, 0b10, 1, 0, + (outs HPR:$Sd), (ins HPR:$Sdin, HPR:$Sn, HPR:$Sm), + IIC_fpFMAC16, "vfms", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fadd_mlx (fneg (fmul_su HPR:$Sn, HPR:$Sm)), + HPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasFullFP16,UseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def : Pat<(fsub_mlx DPR:$dstin, (fmul_su DPR:$a, (f64 DPR:$b))), + (VFMSD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP4,HasDPVFP,UseFusedMAC]>; +def : Pat<(fsub_mlx SPR:$dstin, (fmul_su SPR:$a, SPR:$b)), + (VFMSS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP4,DontUseNEONForFP,UseFusedMAC]>; +def : Pat<(fsub_mlx HPR:$dstin, (fmul_su HPR:$a, HPR:$b)), + (VFMSH HPR:$dstin, HPR:$a, HPR:$b)>, + Requires<[HasFullFP16,DontUseNEONForFP,UseFusedMAC]>; + +// Match @llvm.fma.* intrinsics +// (fma (fneg x), y, z) -> (vfms z, x, y) +def : Pat<(f64 (fma (fneg DPR:$Dn), DPR:$Dm, DPR:$Ddin)), + (VFMSD DPR:$Ddin, DPR:$Dn, DPR:$Dm)>, + Requires<[HasVFP4,HasDPVFP]>; +def : Pat<(f32 (fma (fneg SPR:$Sn), SPR:$Sm, SPR:$Sdin)), + (VFMSS SPR:$Sdin, SPR:$Sn, SPR:$Sm)>, + Requires<[HasVFP4]>; +// (fma x, (fneg y), z) -> (vfms z, x, y) +def : Pat<(f64 (fma DPR:$Dn, (fneg DPR:$Dm), DPR:$Ddin)), + (VFMSD DPR:$Ddin, DPR:$Dn, DPR:$Dm)>, + Requires<[HasVFP4,HasDPVFP]>; +def : Pat<(f32 (fma SPR:$Sn, (fneg SPR:$Sm), SPR:$Sdin)), + (VFMSS SPR:$Sdin, SPR:$Sn, SPR:$Sm)>, + Requires<[HasVFP4]>; + +def VFNMAD : ADbI<0b11101, 0b01, 1, 0, + (outs DPR:$Dd), (ins DPR:$Ddin, DPR:$Dn, DPR:$Dm), + IIC_fpFMAC64, "vfnma", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd,(fsub_mlx (fneg (fmul_su DPR:$Dn,DPR:$Dm)), + (f64 DPR:$Ddin)))]>, + RegConstraint<"$Ddin = $Dd">, + Requires<[HasVFP4,HasDPVFP,UseFusedMAC]>, + Sched<[WriteFPMAC64, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def VFNMAS : ASbI<0b11101, 0b01, 1, 0, + (outs SPR:$Sd), (ins SPR:$Sdin, SPR:$Sn, SPR:$Sm), + IIC_fpFMAC32, "vfnma", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fsub_mlx (fneg (fmul_su SPR:$Sn, SPR:$Sm)), + SPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasVFP4,DontUseNEONForFP,UseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines. +} + +def VFNMAH : AHbI<0b11101, 0b01, 1, 0, + (outs HPR:$Sd), (ins HPR:$Sdin, HPR:$Sn, HPR:$Sm), + IIC_fpFMAC16, "vfnma", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fsub_mlx (fneg (fmul_su HPR:$Sn, HPR:$Sm)), + HPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasFullFP16,UseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def : Pat<(fsub_mlx (fneg (fmul_su DPR:$a, (f64 DPR:$b))), DPR:$dstin), + (VFNMAD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP4,HasDPVFP,UseFusedMAC]>; +def : Pat<(fsub_mlx (fneg (fmul_su SPR:$a, SPR:$b)), SPR:$dstin), + (VFNMAS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP4,DontUseNEONForFP,UseFusedMAC]>; + +// Match @llvm.fma.* intrinsics +// (fneg (fma x, y, z)) -> (vfnma z, x, y) +def : Pat<(fneg (fma (f64 DPR:$Dn), (f64 DPR:$Dm), (f64 DPR:$Ddin))), + (VFNMAD DPR:$Ddin, DPR:$Dn, DPR:$Dm)>, + Requires<[HasVFP4,HasDPVFP]>; +def : Pat<(fneg (fma (f32 SPR:$Sn), (f32 SPR:$Sm), (f32 SPR:$Sdin))), + (VFNMAS SPR:$Sdin, SPR:$Sn, SPR:$Sm)>, + Requires<[HasVFP4]>; +// (fma (fneg x), y, (fneg z)) -> (vfnma z, x, y) +def : Pat<(f64 (fma (fneg DPR:$Dn), DPR:$Dm, (fneg DPR:$Ddin))), + (VFNMAD DPR:$Ddin, DPR:$Dn, DPR:$Dm)>, + Requires<[HasVFP4,HasDPVFP]>; +def : Pat<(f32 (fma (fneg SPR:$Sn), SPR:$Sm, (fneg SPR:$Sdin))), + (VFNMAS SPR:$Sdin, SPR:$Sn, SPR:$Sm)>, + Requires<[HasVFP4]>; + +def VFNMSD : ADbI<0b11101, 0b01, 0, 0, + (outs DPR:$Dd), (ins DPR:$Ddin, DPR:$Dn, DPR:$Dm), + IIC_fpFMAC64, "vfnms", ".f64\t$Dd, $Dn, $Dm", + [(set DPR:$Dd, (fsub_mlx (fmul_su DPR:$Dn, DPR:$Dm), + (f64 DPR:$Ddin)))]>, + RegConstraint<"$Ddin = $Dd">, + Requires<[HasVFP4,HasDPVFP,UseFusedMAC]>, + Sched<[WriteFPMAC64, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def VFNMSS : ASbI<0b11101, 0b01, 0, 0, + (outs SPR:$Sd), (ins SPR:$Sdin, SPR:$Sn, SPR:$Sm), + IIC_fpFMAC32, "vfnms", ".f32\t$Sd, $Sn, $Sm", + [(set SPR:$Sd, (fsub_mlx (fmul_su SPR:$Sn, SPR:$Sm), SPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasVFP4,DontUseNEONForFP,UseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]> { + // Some single precision VFP instructions may be executed on both NEON and + // VFP pipelines. +} + +def VFNMSH : AHbI<0b11101, 0b01, 0, 0, + (outs HPR:$Sd), (ins HPR:$Sdin, HPR:$Sn, HPR:$Sm), + IIC_fpFMAC16, "vfnms", ".f16\t$Sd, $Sn, $Sm", + [(set HPR:$Sd, (fsub_mlx (fmul_su HPR:$Sn, HPR:$Sm), HPR:$Sdin))]>, + RegConstraint<"$Sdin = $Sd">, + Requires<[HasFullFP16,UseFusedMAC]>, + Sched<[WriteFPMAC32, ReadFPMAC, ReadFPMUL, ReadFPMUL]>; + +def : Pat<(fsub_mlx (fmul_su DPR:$a, (f64 DPR:$b)), DPR:$dstin), + (VFNMSD DPR:$dstin, DPR:$a, DPR:$b)>, + Requires<[HasVFP4,HasDPVFP,UseFusedMAC]>; +def : Pat<(fsub_mlx (fmul_su SPR:$a, SPR:$b), SPR:$dstin), + (VFNMSS SPR:$dstin, SPR:$a, SPR:$b)>, + Requires<[HasVFP4,DontUseNEONForFP,UseFusedMAC]>; + +// Match @llvm.fma.* intrinsics + +// (fma x, y, (fneg z)) -> (vfnms z, x, y)) +def : Pat<(f64 (fma DPR:$Dn, DPR:$Dm, (fneg DPR:$Ddin))), + (VFNMSD DPR:$Ddin, DPR:$Dn, DPR:$Dm)>, + Requires<[HasVFP4,HasDPVFP]>; +def : Pat<(f32 (fma SPR:$Sn, SPR:$Sm, (fneg SPR:$Sdin))), + (VFNMSS SPR:$Sdin, SPR:$Sn, SPR:$Sm)>, + Requires<[HasVFP4]>; +// (fneg (fma (fneg x), y, z)) -> (vfnms z, x, y) +def : Pat<(fneg (f64 (fma (fneg DPR:$Dn), DPR:$Dm, DPR:$Ddin))), + (VFNMSD DPR:$Ddin, DPR:$Dn, DPR:$Dm)>, + Requires<[HasVFP4,HasDPVFP]>; +def : Pat<(fneg (f32 (fma (fneg SPR:$Sn), SPR:$Sm, SPR:$Sdin))), + (VFNMSS SPR:$Sdin, SPR:$Sn, SPR:$Sm)>, + Requires<[HasVFP4]>; +// (fneg (fma x, (fneg y), z) -> (vfnms z, x, y) +def : Pat<(fneg (f64 (fma DPR:$Dn, (fneg DPR:$Dm), DPR:$Ddin))), + (VFNMSD DPR:$Ddin, DPR:$Dn, DPR:$Dm)>, + Requires<[HasVFP4,HasDPVFP]>; +def : Pat<(fneg (f32 (fma SPR:$Sn, (fneg SPR:$Sm), SPR:$Sdin))), + (VFNMSS SPR:$Sdin, SPR:$Sn, SPR:$Sm)>, + Requires<[HasVFP4]>; + +//===----------------------------------------------------------------------===// +// FP Conditional moves. +// + +let hasSideEffects = 0 in { +def VMOVDcc : PseudoInst<(outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm, cmovpred:$p), + IIC_fpUNA64, + [(set (f64 DPR:$Dd), + (ARMcmov DPR:$Dn, DPR:$Dm, cmovpred:$p))]>, + RegConstraint<"$Dn = $Dd">, Requires<[HasVFP2,HasDPVFP]>; + +def VMOVScc : PseudoInst<(outs SPR:$Sd), (ins SPR:$Sn, SPR:$Sm, cmovpred:$p), + IIC_fpUNA32, + [(set (f32 SPR:$Sd), + (ARMcmov SPR:$Sn, SPR:$Sm, cmovpred:$p))]>, + RegConstraint<"$Sn = $Sd">, Requires<[HasVFP2]>; +} // hasSideEffects + +//===----------------------------------------------------------------------===// +// Move from VFP System Register to ARM core register. +// + +class MovFromVFP<bits<4> opc19_16, dag oops, dag iops, string opc, string asm, + list<dag> pattern>: + VFPAI<oops, iops, VFPMiscFrm, IIC_fpSTAT, opc, asm, pattern> { + + // Instruction operand. + bits<4> Rt; + + let Inst{27-20} = 0b11101111; + let Inst{19-16} = opc19_16; + let Inst{15-12} = Rt; + let Inst{11-8} = 0b1010; + let Inst{7} = 0; + let Inst{6-5} = 0b00; + let Inst{4} = 1; + let Inst{3-0} = 0b0000; +} + +// APSR is the application level alias of CPSR. This FPSCR N, Z, C, V flags +// to APSR. +let Defs = [CPSR], Uses = [FPSCR_NZCV], Rt = 0b1111 /* apsr_nzcv */ in +def FMSTAT : MovFromVFP<0b0001 /* fpscr */, (outs), (ins), + "vmrs", "\tAPSR_nzcv, fpscr", [(arm_fmstat)]>; + +let DecoderMethod = "DecodeForVMRSandVMSR" in { + // Application level FPSCR -> GPR + let hasSideEffects = 1, Uses = [FPSCR] in + def VMRS : MovFromVFP<0b0001 /* fpscr */, (outs GPRnopc:$Rt), (ins), + "vmrs", "\t$Rt, fpscr", + [(set GPRnopc:$Rt, (int_arm_get_fpscr))]>; + + // System level FPEXC, FPSID -> GPR + let Uses = [FPSCR] in { + def VMRS_FPEXC : MovFromVFP<0b1000 /* fpexc */, (outs GPRnopc:$Rt), (ins), + "vmrs", "\t$Rt, fpexc", []>; + def VMRS_FPSID : MovFromVFP<0b0000 /* fpsid */, (outs GPRnopc:$Rt), (ins), + "vmrs", "\t$Rt, fpsid", []>; + def VMRS_MVFR0 : MovFromVFP<0b0111 /* mvfr0 */, (outs GPRnopc:$Rt), (ins), + "vmrs", "\t$Rt, mvfr0", []>; + def VMRS_MVFR1 : MovFromVFP<0b0110 /* mvfr1 */, (outs GPRnopc:$Rt), (ins), + "vmrs", "\t$Rt, mvfr1", []>; + let Predicates = [HasFPARMv8] in { + def VMRS_MVFR2 : MovFromVFP<0b0101 /* mvfr2 */, (outs GPRnopc:$Rt), (ins), + "vmrs", "\t$Rt, mvfr2", []>; + } + def VMRS_FPINST : MovFromVFP<0b1001 /* fpinst */, (outs GPRnopc:$Rt), (ins), + "vmrs", "\t$Rt, fpinst", []>; + def VMRS_FPINST2 : MovFromVFP<0b1010 /* fpinst2 */, (outs GPRnopc:$Rt), + (ins), "vmrs", "\t$Rt, fpinst2", []>; + } +} + +//===----------------------------------------------------------------------===// +// Move from ARM core register to VFP System Register. +// + +class MovToVFP<bits<4> opc19_16, dag oops, dag iops, string opc, string asm, + list<dag> pattern>: + VFPAI<oops, iops, VFPMiscFrm, IIC_fpSTAT, opc, asm, pattern> { + + // Instruction operand. + bits<4> src; + + // Encode instruction operand. + let Inst{15-12} = src; + + let Inst{27-20} = 0b11101110; + let Inst{19-16} = opc19_16; + let Inst{11-8} = 0b1010; + let Inst{7} = 0; + let Inst{4} = 1; +} + +let DecoderMethod = "DecodeForVMRSandVMSR" in { + let Defs = [FPSCR] in { + // Application level GPR -> FPSCR + def VMSR : MovToVFP<0b0001 /* fpscr */, (outs), (ins GPRnopc:$src), + "vmsr", "\tfpscr, $src", + [(int_arm_set_fpscr GPRnopc:$src)]>; + // System level GPR -> FPEXC + def VMSR_FPEXC : MovToVFP<0b1000 /* fpexc */, (outs), (ins GPRnopc:$src), + "vmsr", "\tfpexc, $src", []>; + // System level GPR -> FPSID + def VMSR_FPSID : MovToVFP<0b0000 /* fpsid */, (outs), (ins GPRnopc:$src), + "vmsr", "\tfpsid, $src", []>; + def VMSR_FPINST : MovToVFP<0b1001 /* fpinst */, (outs), (ins GPRnopc:$src), + "vmsr", "\tfpinst, $src", []>; + def VMSR_FPINST2 : MovToVFP<0b1010 /* fpinst2 */, (outs), (ins GPRnopc:$src), + "vmsr", "\tfpinst2, $src", []>; + } +} + +//===----------------------------------------------------------------------===// +// Misc. +// + +// Materialize FP immediates. VFP3 only. +let isReMaterializable = 1 in { +def FCONSTD : VFPAI<(outs DPR:$Dd), (ins vfp_f64imm:$imm), + VFPMiscFrm, IIC_fpUNA64, + "vmov", ".f64\t$Dd, $imm", + [(set DPR:$Dd, vfp_f64imm:$imm)]>, + Requires<[HasVFP3,HasDPVFP]> { + bits<5> Dd; + bits<8> imm; + + let Inst{27-23} = 0b11101; + let Inst{22} = Dd{4}; + let Inst{21-20} = 0b11; + let Inst{19-16} = imm{7-4}; + let Inst{15-12} = Dd{3-0}; + let Inst{11-9} = 0b101; + let Inst{8} = 1; // Double precision. + let Inst{7-4} = 0b0000; + let Inst{3-0} = imm{3-0}; +} + +def FCONSTS : VFPAI<(outs SPR:$Sd), (ins vfp_f32imm:$imm), + VFPMiscFrm, IIC_fpUNA32, + "vmov", ".f32\t$Sd, $imm", + [(set SPR:$Sd, vfp_f32imm:$imm)]>, Requires<[HasVFP3]> { + bits<5> Sd; + bits<8> imm; + + let Inst{27-23} = 0b11101; + let Inst{22} = Sd{0}; + let Inst{21-20} = 0b11; + let Inst{19-16} = imm{7-4}; + let Inst{15-12} = Sd{4-1}; + let Inst{11-9} = 0b101; + let Inst{8} = 0; // Single precision. + let Inst{7-4} = 0b0000; + let Inst{3-0} = imm{3-0}; +} + +def FCONSTH : VFPAI<(outs HPR:$Sd), (ins vfp_f16imm:$imm), + VFPMiscFrm, IIC_fpUNA16, + "vmov", ".f16\t$Sd, $imm", + [(set HPR:$Sd, vfp_f16imm:$imm)]>, + Requires<[HasFullFP16]> { + bits<5> Sd; + bits<8> imm; + + let Inst{27-23} = 0b11101; + let Inst{22} = Sd{0}; + let Inst{21-20} = 0b11; + let Inst{19-16} = imm{7-4}; + let Inst{15-12} = Sd{4-1}; + let Inst{11-8} = 0b1001; // Half precision + let Inst{7-4} = 0b0000; + let Inst{3-0} = imm{3-0}; +} +} + +//===----------------------------------------------------------------------===// +// Assembler aliases. +// +// A few mnemonic aliases for pre-unifixed syntax. We don't guarantee to +// support them all, but supporting at least some of the basics is +// good to be friendly. +def : VFP2MnemonicAlias<"flds", "vldr">; +def : VFP2MnemonicAlias<"fldd", "vldr">; +def : VFP2MnemonicAlias<"fmrs", "vmov">; +def : VFP2MnemonicAlias<"fmsr", "vmov">; +def : VFP2MnemonicAlias<"fsqrts", "vsqrt">; +def : VFP2MnemonicAlias<"fsqrtd", "vsqrt">; +def : VFP2MnemonicAlias<"fadds", "vadd.f32">; +def : VFP2MnemonicAlias<"faddd", "vadd.f64">; +def : VFP2MnemonicAlias<"fmrdd", "vmov">; +def : VFP2MnemonicAlias<"fmrds", "vmov">; +def : VFP2MnemonicAlias<"fmrrd", "vmov">; +def : VFP2MnemonicAlias<"fmdrr", "vmov">; +def : VFP2MnemonicAlias<"fmuls", "vmul.f32">; +def : VFP2MnemonicAlias<"fmuld", "vmul.f64">; +def : VFP2MnemonicAlias<"fnegs", "vneg.f32">; +def : VFP2MnemonicAlias<"fnegd", "vneg.f64">; +def : VFP2MnemonicAlias<"ftosizd", "vcvt.s32.f64">; +def : VFP2MnemonicAlias<"ftosid", "vcvtr.s32.f64">; +def : VFP2MnemonicAlias<"ftosizs", "vcvt.s32.f32">; +def : VFP2MnemonicAlias<"ftosis", "vcvtr.s32.f32">; +def : VFP2MnemonicAlias<"ftouizd", "vcvt.u32.f64">; +def : VFP2MnemonicAlias<"ftouid", "vcvtr.u32.f64">; +def : VFP2MnemonicAlias<"ftouizs", "vcvt.u32.f32">; +def : VFP2MnemonicAlias<"ftouis", "vcvtr.u32.f32">; +def : VFP2MnemonicAlias<"fsitod", "vcvt.f64.s32">; +def : VFP2MnemonicAlias<"fsitos", "vcvt.f32.s32">; +def : VFP2MnemonicAlias<"fuitod", "vcvt.f64.u32">; +def : VFP2MnemonicAlias<"fuitos", "vcvt.f32.u32">; +def : VFP2MnemonicAlias<"fsts", "vstr">; +def : VFP2MnemonicAlias<"fstd", "vstr">; +def : VFP2MnemonicAlias<"fmacd", "vmla.f64">; +def : VFP2MnemonicAlias<"fmacs", "vmla.f32">; +def : VFP2MnemonicAlias<"fcpys", "vmov.f32">; +def : VFP2MnemonicAlias<"fcpyd", "vmov.f64">; +def : VFP2MnemonicAlias<"fcmps", "vcmp.f32">; +def : VFP2MnemonicAlias<"fcmpd", "vcmp.f64">; +def : VFP2MnemonicAlias<"fdivs", "vdiv.f32">; +def : VFP2MnemonicAlias<"fdivd", "vdiv.f64">; +def : VFP2MnemonicAlias<"fmrx", "vmrs">; +def : VFP2MnemonicAlias<"fmxr", "vmsr">; + +// Be friendly and accept the old form of zero-compare +def : VFP2DPInstAlias<"fcmpzd${p} $val", (VCMPZD DPR:$val, pred:$p)>; +def : VFP2InstAlias<"fcmpzs${p} $val", (VCMPZS SPR:$val, pred:$p)>; + + +def : VFP2InstAlias<"fmstat${p}", (FMSTAT pred:$p)>; +def : VFP2InstAlias<"fadds${p} $Sd, $Sn, $Sm", + (VADDS SPR:$Sd, SPR:$Sn, SPR:$Sm, pred:$p)>; +def : VFP2DPInstAlias<"faddd${p} $Dd, $Dn, $Dm", + (VADDD DPR:$Dd, DPR:$Dn, DPR:$Dm, pred:$p)>; +def : VFP2InstAlias<"fsubs${p} $Sd, $Sn, $Sm", + (VSUBS SPR:$Sd, SPR:$Sn, SPR:$Sm, pred:$p)>; +def : VFP2DPInstAlias<"fsubd${p} $Dd, $Dn, $Dm", + (VSUBD DPR:$Dd, DPR:$Dn, DPR:$Dm, pred:$p)>; + +// No need for the size suffix on VSQRT. It's implied by the register classes. +def : VFP2InstAlias<"vsqrt${p} $Sd, $Sm", (VSQRTS SPR:$Sd, SPR:$Sm, pred:$p)>; +def : VFP2DPInstAlias<"vsqrt${p} $Dd, $Dm", (VSQRTD DPR:$Dd, DPR:$Dm, pred:$p)>; + +// VLDR/VSTR accept an optional type suffix. +def : VFP2InstAlias<"vldr${p}.32 $Sd, $addr", + (VLDRS SPR:$Sd, addrmode5:$addr, pred:$p)>; +def : VFP2InstAlias<"vstr${p}.32 $Sd, $addr", + (VSTRS SPR:$Sd, addrmode5:$addr, pred:$p)>; +def : VFP2InstAlias<"vldr${p}.64 $Dd, $addr", + (VLDRD DPR:$Dd, addrmode5:$addr, pred:$p)>; +def : VFP2InstAlias<"vstr${p}.64 $Dd, $addr", + (VSTRD DPR:$Dd, addrmode5:$addr, pred:$p)>; + +// VMOV can accept optional 32-bit or less data type suffix suffix. +def : VFP2InstAlias<"vmov${p}.8 $Rt, $Sn", + (VMOVRS GPR:$Rt, SPR:$Sn, pred:$p)>; +def : VFP2InstAlias<"vmov${p}.16 $Rt, $Sn", + (VMOVRS GPR:$Rt, SPR:$Sn, pred:$p)>; +def : VFP2InstAlias<"vmov${p}.32 $Rt, $Sn", + (VMOVRS GPR:$Rt, SPR:$Sn, pred:$p)>; +def : VFP2InstAlias<"vmov${p}.8 $Sn, $Rt", + (VMOVSR SPR:$Sn, GPR:$Rt, pred:$p)>; +def : VFP2InstAlias<"vmov${p}.16 $Sn, $Rt", + (VMOVSR SPR:$Sn, GPR:$Rt, pred:$p)>; +def : VFP2InstAlias<"vmov${p}.32 $Sn, $Rt", + (VMOVSR SPR:$Sn, GPR:$Rt, pred:$p)>; + +def : VFP2InstAlias<"vmov${p}.f64 $Rt, $Rt2, $Dn", + (VMOVRRD GPR:$Rt, GPR:$Rt2, DPR:$Dn, pred:$p)>; +def : VFP2InstAlias<"vmov${p}.f64 $Dn, $Rt, $Rt2", + (VMOVDRR DPR:$Dn, GPR:$Rt, GPR:$Rt2, pred:$p)>; + +// VMOVS doesn't need the .f32 to disambiguate from the NEON encoding the way +// VMOVD does. +def : VFP2InstAlias<"vmov${p} $Sd, $Sm", + (VMOVS SPR:$Sd, SPR:$Sm, pred:$p)>; + +// FCONSTD/FCONSTS alias for vmov.f64/vmov.f32 +// These aliases provide added functionality over vmov.f instructions by +// allowing users to write assembly containing encoded floating point constants +// (e.g. #0x70 vs #1.0). Without these alises there is no way for the +// assembler to accept encoded fp constants (but the equivalent fp-literal is +// accepted directly by vmovf). +def : VFP3InstAlias<"fconstd${p} $Dd, $val", + (FCONSTD DPR:$Dd, vfp_f64imm:$val, pred:$p)>; +def : VFP3InstAlias<"fconsts${p} $Sd, $val", + (FCONSTS SPR:$Sd, vfp_f32imm:$val, pred:$p)>; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMRegisterBanks.td b/capstone/suite/synctools/tablegen/ARM/ARMRegisterBanks.td new file mode 100644 index 000000000..6e3834da3 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMRegisterBanks.td @@ -0,0 +1,14 @@ +//=- ARMRegisterBank.td - Describe the AArch64 Banks ---------*- tablegen -*-=// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// +//===----------------------------------------------------------------------===// + +def GPRRegBank : RegisterBank<"GPRB", [GPR, GPRwithAPSR]>; +def FPRRegBank : RegisterBank<"FPRB", [HPR, SPR, DPR, QPR]>; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMRegisterInfo-digit.td b/capstone/suite/synctools/tablegen/ARM/ARMRegisterInfo-digit.td new file mode 100644 index 000000000..3076bfc81 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMRegisterInfo-digit.td @@ -0,0 +1,481 @@ +//===-- ARMRegisterInfo.td - ARM Register defs -------------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +include "ARMSystemRegister.td" + +//===----------------------------------------------------------------------===// +// Declarations that describe the ARM register file +//===----------------------------------------------------------------------===// + +// Registers are identified with 4-bit ID numbers. +class ARMReg<bits<16> Enc, string n, list<Register> subregs = []> : Register<n> { + let HWEncoding = Enc; + let Namespace = "ARM"; + let SubRegs = subregs; + // All bits of ARM registers with sub-registers are covered by sub-registers. + let CoveredBySubRegs = 1; +} + +class ARMFReg<bits<16> Enc, string n> : Register<n> { + let HWEncoding = Enc; + let Namespace = "ARM"; +} + +// Subregister indices. +let Namespace = "ARM" in { +def qqsub_0 : SubRegIndex<256>; +def qqsub_1 : SubRegIndex<256, 256>; + +// Note: Code depends on these having consecutive numbers. +def qsub_0 : SubRegIndex<128>; +def qsub_1 : SubRegIndex<128, 128>; +def qsub_2 : ComposedSubRegIndex<qqsub_1, qsub_0>; +def qsub_3 : ComposedSubRegIndex<qqsub_1, qsub_1>; + +def dsub_0 : SubRegIndex<64>; +def dsub_1 : SubRegIndex<64, 64>; +def dsub_2 : ComposedSubRegIndex<qsub_1, dsub_0>; +def dsub_3 : ComposedSubRegIndex<qsub_1, dsub_1>; +def dsub_4 : ComposedSubRegIndex<qsub_2, dsub_0>; +def dsub_5 : ComposedSubRegIndex<qsub_2, dsub_1>; +def dsub_6 : ComposedSubRegIndex<qsub_3, dsub_0>; +def dsub_7 : ComposedSubRegIndex<qsub_3, dsub_1>; + +def ssub_0 : SubRegIndex<32>; +def ssub_1 : SubRegIndex<32, 32>; +def ssub_2 : ComposedSubRegIndex<dsub_1, ssub_0>; +def ssub_3 : ComposedSubRegIndex<dsub_1, ssub_1>; +def ssub_4 : ComposedSubRegIndex<dsub_2, ssub_0>; +def ssub_5 : ComposedSubRegIndex<dsub_2, ssub_1>; +def ssub_6 : ComposedSubRegIndex<dsub_3, ssub_0>; +def ssub_7 : ComposedSubRegIndex<dsub_3, ssub_1>; +def ssub_8 : ComposedSubRegIndex<dsub_4, ssub_0>; +def ssub_9 : ComposedSubRegIndex<dsub_4, ssub_1>; +def ssub_10 : ComposedSubRegIndex<dsub_5, ssub_0>; +def ssub_11 : ComposedSubRegIndex<dsub_5, ssub_1>; +def ssub_12 : ComposedSubRegIndex<dsub_6, ssub_0>; +def ssub_13 : ComposedSubRegIndex<dsub_6, ssub_1>; + +def gsub_0 : SubRegIndex<32>; +def gsub_1 : SubRegIndex<32, 32>; +// Let TableGen synthesize the remaining 12 ssub_* indices. +// We don't need to name them. +} + +// Integer registers +def R0 : ARMReg< 0, "r0">, DwarfRegNum<[0]>; +def R1 : ARMReg< 1, "r1">, DwarfRegNum<[1]>; +def R2 : ARMReg< 2, "r2">, DwarfRegNum<[2]>; +def R3 : ARMReg< 3, "r3">, DwarfRegNum<[3]>; +def R4 : ARMReg< 4, "r4">, DwarfRegNum<[4]>; +def R5 : ARMReg< 5, "r5">, DwarfRegNum<[5]>; +def R6 : ARMReg< 6, "r6">, DwarfRegNum<[6]>; +def R7 : ARMReg< 7, "r7">, DwarfRegNum<[7]>; +// These require 32-bit instructions. +let CostPerUse = 1 in { +def R8 : ARMReg< 8, "r8">, DwarfRegNum<[8]>; +def R9 : ARMReg< 9, "r9">, DwarfRegNum<[9]>; +def R10 : ARMReg<10, "r10">, DwarfRegNum<[10]>; +def R11 : ARMReg<11, "r11">, DwarfRegNum<[11]>; +def R12 : ARMReg<12, "r12">, DwarfRegNum<[12]>; +def SP : ARMReg<13, "r13">, DwarfRegNum<[13]>; +def LR : ARMReg<14, "r14">, DwarfRegNum<[14]>; +def PC : ARMReg<15, "pc">, DwarfRegNum<[15]>; +} + +// Float registers +def S0 : ARMFReg< 0, "s0">; def S1 : ARMFReg< 1, "s1">; +def S2 : ARMFReg< 2, "s2">; def S3 : ARMFReg< 3, "s3">; +def S4 : ARMFReg< 4, "s4">; def S5 : ARMFReg< 5, "s5">; +def S6 : ARMFReg< 6, "s6">; def S7 : ARMFReg< 7, "s7">; +def S8 : ARMFReg< 8, "s8">; def S9 : ARMFReg< 9, "s9">; +def S10 : ARMFReg<10, "s10">; def S11 : ARMFReg<11, "s11">; +def S12 : ARMFReg<12, "s12">; def S13 : ARMFReg<13, "s13">; +def S14 : ARMFReg<14, "s14">; def S15 : ARMFReg<15, "s15">; +def S16 : ARMFReg<16, "s16">; def S17 : ARMFReg<17, "s17">; +def S18 : ARMFReg<18, "s18">; def S19 : ARMFReg<19, "s19">; +def S20 : ARMFReg<20, "s20">; def S21 : ARMFReg<21, "s21">; +def S22 : ARMFReg<22, "s22">; def S23 : ARMFReg<23, "s23">; +def S24 : ARMFReg<24, "s24">; def S25 : ARMFReg<25, "s25">; +def S26 : ARMFReg<26, "s26">; def S27 : ARMFReg<27, "s27">; +def S28 : ARMFReg<28, "s28">; def S29 : ARMFReg<29, "s29">; +def S30 : ARMFReg<30, "s30">; def S31 : ARMFReg<31, "s31">; + +// Aliases of the F* registers used to hold 64-bit fp values (doubles) +let SubRegIndices = [ssub_0, ssub_1] in { +def D0 : ARMReg< 0, "d0", [S0, S1]>, DwarfRegNum<[256]>; +def D1 : ARMReg< 1, "d1", [S2, S3]>, DwarfRegNum<[257]>; +def D2 : ARMReg< 2, "d2", [S4, S5]>, DwarfRegNum<[258]>; +def D3 : ARMReg< 3, "d3", [S6, S7]>, DwarfRegNum<[259]>; +def D4 : ARMReg< 4, "d4", [S8, S9]>, DwarfRegNum<[260]>; +def D5 : ARMReg< 5, "d5", [S10, S11]>, DwarfRegNum<[261]>; +def D6 : ARMReg< 6, "d6", [S12, S13]>, DwarfRegNum<[262]>; +def D7 : ARMReg< 7, "d7", [S14, S15]>, DwarfRegNum<[263]>; +def D8 : ARMReg< 8, "d8", [S16, S17]>, DwarfRegNum<[264]>; +def D9 : ARMReg< 9, "d9", [S18, S19]>, DwarfRegNum<[265]>; +def D10 : ARMReg<10, "d10", [S20, S21]>, DwarfRegNum<[266]>; +def D11 : ARMReg<11, "d11", [S22, S23]>, DwarfRegNum<[267]>; +def D12 : ARMReg<12, "d12", [S24, S25]>, DwarfRegNum<[268]>; +def D13 : ARMReg<13, "d13", [S26, S27]>, DwarfRegNum<[269]>; +def D14 : ARMReg<14, "d14", [S28, S29]>, DwarfRegNum<[270]>; +def D15 : ARMReg<15, "d15", [S30, S31]>, DwarfRegNum<[271]>; +} + +// VFP3 defines 16 additional double registers +def D16 : ARMFReg<16, "d16">, DwarfRegNum<[272]>; +def D17 : ARMFReg<17, "d17">, DwarfRegNum<[273]>; +def D18 : ARMFReg<18, "d18">, DwarfRegNum<[274]>; +def D19 : ARMFReg<19, "d19">, DwarfRegNum<[275]>; +def D20 : ARMFReg<20, "d20">, DwarfRegNum<[276]>; +def D21 : ARMFReg<21, "d21">, DwarfRegNum<[277]>; +def D22 : ARMFReg<22, "d22">, DwarfRegNum<[278]>; +def D23 : ARMFReg<23, "d23">, DwarfRegNum<[279]>; +def D24 : ARMFReg<24, "d24">, DwarfRegNum<[280]>; +def D25 : ARMFReg<25, "d25">, DwarfRegNum<[281]>; +def D26 : ARMFReg<26, "d26">, DwarfRegNum<[282]>; +def D27 : ARMFReg<27, "d27">, DwarfRegNum<[283]>; +def D28 : ARMFReg<28, "d28">, DwarfRegNum<[284]>; +def D29 : ARMFReg<29, "d29">, DwarfRegNum<[285]>; +def D30 : ARMFReg<30, "d30">, DwarfRegNum<[286]>; +def D31 : ARMFReg<31, "d31">, DwarfRegNum<[287]>; + +// Advanced SIMD (NEON) defines 16 quad-word aliases +let SubRegIndices = [dsub_0, dsub_1] in { +def Q0 : ARMReg< 0, "q0", [D0, D1]>; +def Q1 : ARMReg< 1, "q1", [D2, D3]>; +def Q2 : ARMReg< 2, "q2", [D4, D5]>; +def Q3 : ARMReg< 3, "q3", [D6, D7]>; +def Q4 : ARMReg< 4, "q4", [D8, D9]>; +def Q5 : ARMReg< 5, "q5", [D10, D11]>; +def Q6 : ARMReg< 6, "q6", [D12, D13]>; +def Q7 : ARMReg< 7, "q7", [D14, D15]>; +} +let SubRegIndices = [dsub_0, dsub_1] in { +def Q8 : ARMReg< 8, "q8", [D16, D17]>; +def Q9 : ARMReg< 9, "q9", [D18, D19]>; +def Q10 : ARMReg<10, "q10", [D20, D21]>; +def Q11 : ARMReg<11, "q11", [D22, D23]>; +def Q12 : ARMReg<12, "q12", [D24, D25]>; +def Q13 : ARMReg<13, "q13", [D26, D27]>; +def Q14 : ARMReg<14, "q14", [D28, D29]>; +def Q15 : ARMReg<15, "q15", [D30, D31]>; +} + +// Current Program Status Register. +// We model fpscr with two registers: FPSCR models the control bits and will be +// reserved. FPSCR_NZCV models the flag bits and will be unreserved. APSR_NZCV +// models the APSR when it's accessed by some special instructions. In such cases +// it has the same encoding as PC. +def CPSR : ARMReg<0, "cpsr">; +def APSR : ARMReg<1, "apsr">; +def APSR_NZCV : ARMReg<15, "apsr_nzcv">; +def SPSR : ARMReg<2, "spsr">; +def FPSCR : ARMReg<3, "fpscr">; +def FPSCR_NZCV : ARMReg<3, "fpscr_nzcv"> { + let Aliases = [FPSCR]; +} +def ITSTATE : ARMReg<4, "itstate">; + +// Special Registers - only available in privileged mode. +def FPSID : ARMReg<0, "fpsid">; +def MVFR2 : ARMReg<5, "mvfr2">; +def MVFR1 : ARMReg<6, "mvfr1">; +def MVFR0 : ARMReg<7, "mvfr0">; +def FPEXC : ARMReg<8, "fpexc">; +def FPINST : ARMReg<9, "fpinst">; +def FPINST2 : ARMReg<10, "fpinst2">; + +// Register classes. +// +// pc == Program Counter +// lr == Link Register +// sp == Stack Pointer +// r12 == ip (scratch) +// r7 == Frame Pointer (thumb-style backtraces) +// r9 == May be reserved as Thread Register +// r11 == Frame Pointer (arm-style backtraces) +// r10 == Stack Limit +// +def GPR : RegisterClass<"ARM", [i32], 32, (add (sequence "R%u", 0, 12), + SP, LR, PC)> { + // Allocate LR as the first CSR since it is always saved anyway. + // For Thumb1 mode, we don't want to allocate hi regs at all, as we don't + // know how to spill them. If we make our prologue/epilogue code smarter at + // some point, we can go back to using the above allocation orders for the + // Thumb1 instructions that know how to use hi regs. + let AltOrders = [(add LR, GPR), (trunc GPR, 8)]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; + let DiagnosticString = "operand must be a register in range [r0, r15]"; +} + +// GPRs without the PC. Some ARM instructions do not allow the PC in +// certain operand slots, particularly as the destination. Primarily +// useful for disassembly. +def GPRnopc : RegisterClass<"ARM", [i32], 32, (sub GPR, PC)> { + let AltOrders = [(add LR, GPRnopc), (trunc GPRnopc, 8)]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; + let DiagnosticString = "operand must be a register in range [r0, r14]"; +} + +// GPRs without the PC but with APSR. Some instructions allow accessing the +// APSR, while actually encoding PC in the register field. This is useful +// for assembly and disassembly only. +def GPRwithAPSR : RegisterClass<"ARM", [i32], 32, (add (sub GPR, PC), APSR_NZCV)> { + let AltOrders = [(add LR, GPRnopc), (trunc GPRnopc, 8)]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; + let DiagnosticString = "operand must be a register in range [r0, r14] or apsr_nzcv"; +} + +// GPRsp - Only the SP is legal. Used by Thumb1 instructions that want the +// implied SP argument list. +// FIXME: It would be better to not use this at all and refactor the +// instructions to not have SP an an explicit argument. That makes +// frame index resolution a bit trickier, though. +def GPRsp : RegisterClass<"ARM", [i32], 32, (add SP)> { + let DiagnosticString = "operand must be a register sp"; +} + +// restricted GPR register class. Many Thumb2 instructions allow the full +// register range for operands, but have undefined behaviours when PC +// or SP (R13 or R15) are used. The ARM ISA refers to these operands +// via the BadReg() pseudo-code description. +def rGPR : RegisterClass<"ARM", [i32], 32, (sub GPR, SP, PC)> { + let AltOrders = [(add LR, rGPR), (trunc rGPR, 8)]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; + let DiagnosticType = "rGPR"; +} + +// Thumb registers are R0-R7 normally. Some instructions can still use +// the general GPR register class above (MOV, e.g.) +def tGPR : RegisterClass<"ARM", [i32], 32, (trunc GPR, 8)> { + let DiagnosticString = "operand must be a register in range [r0, r7]"; +} + +// Thumb registers R0-R7 and the PC. Some instructions like TBB or THH allow +// the PC to be used as a destination operand as well. +def tGPRwithpc : RegisterClass<"ARM", [i32], 32, (add tGPR, PC)>; + +// The high registers in thumb mode, R8-R15. +def hGPR : RegisterClass<"ARM", [i32], 32, (sub GPR, tGPR)> { + let DiagnosticString = "operand must be a register in range [r8, r15]"; +} + +// For tail calls, we can't use callee-saved registers, as they are restored +// to the saved value before the tail call, which would clobber a call address. +// Note, getMinimalPhysRegClass(R0) returns tGPR because of the names of +// this class and the preceding one(!) This is what we want. +def tcGPR : RegisterClass<"ARM", [i32], 32, (add R0, R1, R2, R3, R12)> { + let AltOrders = [(and tcGPR, tGPR)]; + let AltOrderSelect = [{ + return MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; +} + +// Condition code registers. +def CCR : RegisterClass<"ARM", [i32], 32, (add CPSR)> { + let CopyCost = -1; // Don't allow copying of status registers. + let isAllocatable = 0; +} + +// Scalar single precision floating point register class.. +// FIXME: Allocation order changed to s0, s2, ... or s0, s4, ... as a quick hack +// to avoid partial-write dependencies on D or Q (depending on platform) +// registers (S registers are renamed as portions of D/Q registers). +def SPR : RegisterClass<"ARM", [f32], 32, (sequence "S%u", 0, 31)> { + let AltOrders = [(add (decimate SPR, 2), SPR), + (add (decimate SPR, 4), + (decimate SPR, 2), + (decimate (rotl SPR, 1), 4), + (decimate (rotl SPR, 1), 2))]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().useStride4VFPs(MF); + }]; + let DiagnosticString = "operand must be a register in range [s0, s31]"; +} + +def HPR : RegisterClass<"ARM", [f16], 32, (sequence "S%u", 0, 31)> { + let AltOrders = [(add (decimate HPR, 2), SPR), + (add (decimate HPR, 4), + (decimate HPR, 2), + (decimate (rotl HPR, 1), 4), + (decimate (rotl HPR, 1), 2))]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().useStride4VFPs(MF); + }]; + let DiagnosticString = "operand must be a register in range [s0, s31]"; +} + +// Subset of SPR which can be used as a source of NEON scalars for 16-bit +// operations +def SPR_8 : RegisterClass<"ARM", [f32], 32, (sequence "S%u", 0, 15)> { + let DiagnosticString = "operand must be a register in range [s0, s15]"; +} + +// Scalar double precision floating point / generic 64-bit vector register +// class. +// ARM requires only word alignment for double. It's more performant if it +// is double-word alignment though. +def DPR : RegisterClass<"ARM", [f64, v8i8, v4i16, v2i32, v1i64, v2f32, v4f16], 64, + (sequence "D%u", 0, 31)> { + // Allocate non-VFP2 registers D16-D31 first, and prefer even registers on + // Darwin platforms. + let AltOrders = [(rotl DPR, 16), + (add (decimate (rotl DPR, 16), 2), (rotl DPR, 16))]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().useStride4VFPs(MF); + }]; + let DiagnosticType = "DPR"; +} + +// Subset of DPR that are accessible with VFP2 (and so that also have +// 32-bit SPR subregs). +def DPR_VFP2 : RegisterClass<"ARM", [f64, v8i8, v4i16, v2i32, v1i64, v2f32, v4f16], 64, + (trunc DPR, 16)> { + let DiagnosticString = "operand must be a register in range [d0, d15]"; +} + +// Subset of DPR which can be used as a source of NEON scalars for 16-bit +// operations +def DPR_8 : RegisterClass<"ARM", [f64, v8i8, v4i16, v2i32, v1i64, v2f32, v4f16], 64, + (trunc DPR, 8)> { + let DiagnosticString = "operand must be a register in range [d0, d7]"; +} + +// Generic 128-bit vector register class. +def QPR : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64, v8f16], 128, + (sequence "Q%u", 0, 15)> { + // Allocate non-VFP2 aliases Q8-Q15 first. + let AltOrders = [(rotl QPR, 8)]; + let AltOrderSelect = [{ return 1; }]; + let DiagnosticString = "operand must be a register in range [q0, q15]"; +} + +// Subset of QPR that have 32-bit SPR subregs. +def QPR_VFP2 : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], + 128, (trunc QPR, 8)> { + let DiagnosticString = "operand must be a register in range [q0, q7]"; +} + +// Subset of QPR that have DPR_8 and SPR_8 subregs. +def QPR_8 : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], + 128, (trunc QPR, 4)> { + let DiagnosticString = "operand must be a register in range [q0, q3]"; +} + +// Pseudo-registers representing odd-even pairs of D registers. The even-odd +// pairs are already represented by the Q registers. +// These are needed by NEON instructions requiring two consecutive D registers. +// There is no D31_D0 register as that is always an UNPREDICTABLE encoding. +def TuplesOE2D : RegisterTuples<[dsub_0, dsub_1], + [(decimate (shl DPR, 1), 2), + (decimate (shl DPR, 2), 2)]>; + +// Register class representing a pair of consecutive D registers. +// Use the Q registers for the even-odd pairs. +def DPair : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], + 128, (interleave QPR, TuplesOE2D)> { + // Allocate starting at non-VFP2 registers D16-D31 first. + // Prefer even-odd pairs as they are easier to copy. + let AltOrders = [(add (rotl QPR, 8), (rotl DPair, 16))]; + let AltOrderSelect = [{ return 1; }]; +} + +// Pseudo-registers representing even-odd pairs of GPRs from R1 to R13/SP. +// These are needed by instructions (e.g. ldrexd/strexd) requiring even-odd GPRs. +def Tuples2R : RegisterTuples<[gsub_0, gsub_1], + [(add R0, R2, R4, R6, R8, R10, R12), + (add R1, R3, R5, R7, R9, R11, SP)]>; + +// Register class representing a pair of even-odd GPRs. +def GPRPair : RegisterClass<"ARM", [untyped], 64, (add Tuples2R)> { + let Size = 64; // 2 x 32 bits, we have no predefined type of that size. +} + +// Pseudo-registers representing 3 consecutive D registers. +def Tuples3D : RegisterTuples<[dsub_0, dsub_1, dsub_2], + [(shl DPR, 0), + (shl DPR, 1), + (shl DPR, 2)]>; + +// 3 consecutive D registers. +def DTriple : RegisterClass<"ARM", [untyped], 64, (add Tuples3D)> { + let Size = 192; // 3 x 64 bits, we have no predefined type of that size. +} + +// Pseudo 256-bit registers to represent pairs of Q registers. These should +// never be present in the emitted code. +// These are used for NEON load / store instructions, e.g., vld4, vst3. +def Tuples2Q : RegisterTuples<[qsub_0, qsub_1], [(shl QPR, 0), (shl QPR, 1)]>; + +// Pseudo 256-bit vector register class to model pairs of Q registers +// (4 consecutive D registers). +def QQPR : RegisterClass<"ARM", [v4i64], 256, (add Tuples2Q)> { + // Allocate non-VFP2 aliases first. + let AltOrders = [(rotl QQPR, 8)]; + let AltOrderSelect = [{ return 1; }]; +} + +// Tuples of 4 D regs that isn't also a pair of Q regs. +def TuplesOE4D : RegisterTuples<[dsub_0, dsub_1, dsub_2, dsub_3], + [(decimate (shl DPR, 1), 2), + (decimate (shl DPR, 2), 2), + (decimate (shl DPR, 3), 2), + (decimate (shl DPR, 4), 2)]>; + +// 4 consecutive D registers. +def DQuad : RegisterClass<"ARM", [v4i64], 256, + (interleave Tuples2Q, TuplesOE4D)>; + +// Pseudo 512-bit registers to represent four consecutive Q registers. +def Tuples2QQ : RegisterTuples<[qqsub_0, qqsub_1], + [(shl QQPR, 0), (shl QQPR, 2)]>; + +// Pseudo 512-bit vector register class to model 4 consecutive Q registers +// (8 consecutive D registers). +def QQQQPR : RegisterClass<"ARM", [v8i64], 256, (add Tuples2QQ)> { + // Allocate non-VFP2 aliases first. + let AltOrders = [(rotl QQQQPR, 8)]; + let AltOrderSelect = [{ return 1; }]; +} + + +// Pseudo-registers representing 2-spaced consecutive D registers. +def Tuples2DSpc : RegisterTuples<[dsub_0, dsub_2], + [(shl DPR, 0), + (shl DPR, 2)]>; + +// Spaced pairs of D registers. +def DPairSpc : RegisterClass<"ARM", [v2i64], 64, (add Tuples2DSpc)>; + +def Tuples3DSpc : RegisterTuples<[dsub_0, dsub_2, dsub_4], + [(shl DPR, 0), + (shl DPR, 2), + (shl DPR, 4)]>; + +// Spaced triples of D registers. +def DTripleSpc : RegisterClass<"ARM", [untyped], 64, (add Tuples3DSpc)> { + let Size = 192; // 3 x 64 bits, we have no predefined type of that size. +} + +def Tuples4DSpc : RegisterTuples<[dsub_0, dsub_2, dsub_4, dsub_6], + [(shl DPR, 0), + (shl DPR, 2), + (shl DPR, 4), + (shl DPR, 6)]>; + +// Spaced quads of D registers. +def DQuadSpc : RegisterClass<"ARM", [v4i64], 64, (add Tuples3DSpc)>; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMRegisterInfo.td b/capstone/suite/synctools/tablegen/ARM/ARMRegisterInfo.td new file mode 100644 index 000000000..0f7bbdf6c --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMRegisterInfo.td @@ -0,0 +1,481 @@ +//===-- ARMRegisterInfo.td - ARM Register defs -------------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +include "ARMSystemRegister.td" + +//===----------------------------------------------------------------------===// +// Declarations that describe the ARM register file +//===----------------------------------------------------------------------===// + +// Registers are identified with 4-bit ID numbers. +class ARMReg<bits<16> Enc, string n, list<Register> subregs = []> : Register<n> { + let HWEncoding = Enc; + let Namespace = "ARM"; + let SubRegs = subregs; + // All bits of ARM registers with sub-registers are covered by sub-registers. + let CoveredBySubRegs = 1; +} + +class ARMFReg<bits<16> Enc, string n> : Register<n> { + let HWEncoding = Enc; + let Namespace = "ARM"; +} + +// Subregister indices. +let Namespace = "ARM" in { +def qqsub_0 : SubRegIndex<256>; +def qqsub_1 : SubRegIndex<256, 256>; + +// Note: Code depends on these having consecutive numbers. +def qsub_0 : SubRegIndex<128>; +def qsub_1 : SubRegIndex<128, 128>; +def qsub_2 : ComposedSubRegIndex<qqsub_1, qsub_0>; +def qsub_3 : ComposedSubRegIndex<qqsub_1, qsub_1>; + +def dsub_0 : SubRegIndex<64>; +def dsub_1 : SubRegIndex<64, 64>; +def dsub_2 : ComposedSubRegIndex<qsub_1, dsub_0>; +def dsub_3 : ComposedSubRegIndex<qsub_1, dsub_1>; +def dsub_4 : ComposedSubRegIndex<qsub_2, dsub_0>; +def dsub_5 : ComposedSubRegIndex<qsub_2, dsub_1>; +def dsub_6 : ComposedSubRegIndex<qsub_3, dsub_0>; +def dsub_7 : ComposedSubRegIndex<qsub_3, dsub_1>; + +def ssub_0 : SubRegIndex<32>; +def ssub_1 : SubRegIndex<32, 32>; +def ssub_2 : ComposedSubRegIndex<dsub_1, ssub_0>; +def ssub_3 : ComposedSubRegIndex<dsub_1, ssub_1>; +def ssub_4 : ComposedSubRegIndex<dsub_2, ssub_0>; +def ssub_5 : ComposedSubRegIndex<dsub_2, ssub_1>; +def ssub_6 : ComposedSubRegIndex<dsub_3, ssub_0>; +def ssub_7 : ComposedSubRegIndex<dsub_3, ssub_1>; +def ssub_8 : ComposedSubRegIndex<dsub_4, ssub_0>; +def ssub_9 : ComposedSubRegIndex<dsub_4, ssub_1>; +def ssub_10 : ComposedSubRegIndex<dsub_5, ssub_0>; +def ssub_11 : ComposedSubRegIndex<dsub_5, ssub_1>; +def ssub_12 : ComposedSubRegIndex<dsub_6, ssub_0>; +def ssub_13 : ComposedSubRegIndex<dsub_6, ssub_1>; + +def gsub_0 : SubRegIndex<32>; +def gsub_1 : SubRegIndex<32, 32>; +// Let TableGen synthesize the remaining 12 ssub_* indices. +// We don't need to name them. +} + +// Integer registers +def R0 : ARMReg< 0, "r0">, DwarfRegNum<[0]>; +def R1 : ARMReg< 1, "r1">, DwarfRegNum<[1]>; +def R2 : ARMReg< 2, "r2">, DwarfRegNum<[2]>; +def R3 : ARMReg< 3, "r3">, DwarfRegNum<[3]>; +def R4 : ARMReg< 4, "r4">, DwarfRegNum<[4]>; +def R5 : ARMReg< 5, "r5">, DwarfRegNum<[5]>; +def R6 : ARMReg< 6, "r6">, DwarfRegNum<[6]>; +def R7 : ARMReg< 7, "r7">, DwarfRegNum<[7]>; +// These require 32-bit instructions. +let CostPerUse = 1 in { +def R8 : ARMReg< 8, "r8">, DwarfRegNum<[8]>; +def R9 : ARMReg< 9, "sb">, DwarfRegNum<[9]>; +def R10 : ARMReg<10, "sl">, DwarfRegNum<[10]>; +def R11 : ARMReg<11, "fp">, DwarfRegNum<[11]>; +def R12 : ARMReg<12, "ip">, DwarfRegNum<[12]>; +def SP : ARMReg<13, "sp">, DwarfRegNum<[13]>; +def LR : ARMReg<14, "lr">, DwarfRegNum<[14]>; +def PC : ARMReg<15, "pc">, DwarfRegNum<[15]>; +} + +// Float registers +def S0 : ARMFReg< 0, "s0">; def S1 : ARMFReg< 1, "s1">; +def S2 : ARMFReg< 2, "s2">; def S3 : ARMFReg< 3, "s3">; +def S4 : ARMFReg< 4, "s4">; def S5 : ARMFReg< 5, "s5">; +def S6 : ARMFReg< 6, "s6">; def S7 : ARMFReg< 7, "s7">; +def S8 : ARMFReg< 8, "s8">; def S9 : ARMFReg< 9, "s9">; +def S10 : ARMFReg<10, "s10">; def S11 : ARMFReg<11, "s11">; +def S12 : ARMFReg<12, "s12">; def S13 : ARMFReg<13, "s13">; +def S14 : ARMFReg<14, "s14">; def S15 : ARMFReg<15, "s15">; +def S16 : ARMFReg<16, "s16">; def S17 : ARMFReg<17, "s17">; +def S18 : ARMFReg<18, "s18">; def S19 : ARMFReg<19, "s19">; +def S20 : ARMFReg<20, "s20">; def S21 : ARMFReg<21, "s21">; +def S22 : ARMFReg<22, "s22">; def S23 : ARMFReg<23, "s23">; +def S24 : ARMFReg<24, "s24">; def S25 : ARMFReg<25, "s25">; +def S26 : ARMFReg<26, "s26">; def S27 : ARMFReg<27, "s27">; +def S28 : ARMFReg<28, "s28">; def S29 : ARMFReg<29, "s29">; +def S30 : ARMFReg<30, "s30">; def S31 : ARMFReg<31, "s31">; + +// Aliases of the F* registers used to hold 64-bit fp values (doubles) +let SubRegIndices = [ssub_0, ssub_1] in { +def D0 : ARMReg< 0, "d0", [S0, S1]>, DwarfRegNum<[256]>; +def D1 : ARMReg< 1, "d1", [S2, S3]>, DwarfRegNum<[257]>; +def D2 : ARMReg< 2, "d2", [S4, S5]>, DwarfRegNum<[258]>; +def D3 : ARMReg< 3, "d3", [S6, S7]>, DwarfRegNum<[259]>; +def D4 : ARMReg< 4, "d4", [S8, S9]>, DwarfRegNum<[260]>; +def D5 : ARMReg< 5, "d5", [S10, S11]>, DwarfRegNum<[261]>; +def D6 : ARMReg< 6, "d6", [S12, S13]>, DwarfRegNum<[262]>; +def D7 : ARMReg< 7, "d7", [S14, S15]>, DwarfRegNum<[263]>; +def D8 : ARMReg< 8, "d8", [S16, S17]>, DwarfRegNum<[264]>; +def D9 : ARMReg< 9, "d9", [S18, S19]>, DwarfRegNum<[265]>; +def D10 : ARMReg<10, "d10", [S20, S21]>, DwarfRegNum<[266]>; +def D11 : ARMReg<11, "d11", [S22, S23]>, DwarfRegNum<[267]>; +def D12 : ARMReg<12, "d12", [S24, S25]>, DwarfRegNum<[268]>; +def D13 : ARMReg<13, "d13", [S26, S27]>, DwarfRegNum<[269]>; +def D14 : ARMReg<14, "d14", [S28, S29]>, DwarfRegNum<[270]>; +def D15 : ARMReg<15, "d15", [S30, S31]>, DwarfRegNum<[271]>; +} + +// VFP3 defines 16 additional double registers +def D16 : ARMFReg<16, "d16">, DwarfRegNum<[272]>; +def D17 : ARMFReg<17, "d17">, DwarfRegNum<[273]>; +def D18 : ARMFReg<18, "d18">, DwarfRegNum<[274]>; +def D19 : ARMFReg<19, "d19">, DwarfRegNum<[275]>; +def D20 : ARMFReg<20, "d20">, DwarfRegNum<[276]>; +def D21 : ARMFReg<21, "d21">, DwarfRegNum<[277]>; +def D22 : ARMFReg<22, "d22">, DwarfRegNum<[278]>; +def D23 : ARMFReg<23, "d23">, DwarfRegNum<[279]>; +def D24 : ARMFReg<24, "d24">, DwarfRegNum<[280]>; +def D25 : ARMFReg<25, "d25">, DwarfRegNum<[281]>; +def D26 : ARMFReg<26, "d26">, DwarfRegNum<[282]>; +def D27 : ARMFReg<27, "d27">, DwarfRegNum<[283]>; +def D28 : ARMFReg<28, "d28">, DwarfRegNum<[284]>; +def D29 : ARMFReg<29, "d29">, DwarfRegNum<[285]>; +def D30 : ARMFReg<30, "d30">, DwarfRegNum<[286]>; +def D31 : ARMFReg<31, "d31">, DwarfRegNum<[287]>; + +// Advanced SIMD (NEON) defines 16 quad-word aliases +let SubRegIndices = [dsub_0, dsub_1] in { +def Q0 : ARMReg< 0, "q0", [D0, D1]>; +def Q1 : ARMReg< 1, "q1", [D2, D3]>; +def Q2 : ARMReg< 2, "q2", [D4, D5]>; +def Q3 : ARMReg< 3, "q3", [D6, D7]>; +def Q4 : ARMReg< 4, "q4", [D8, D9]>; +def Q5 : ARMReg< 5, "q5", [D10, D11]>; +def Q6 : ARMReg< 6, "q6", [D12, D13]>; +def Q7 : ARMReg< 7, "q7", [D14, D15]>; +} +let SubRegIndices = [dsub_0, dsub_1] in { +def Q8 : ARMReg< 8, "q8", [D16, D17]>; +def Q9 : ARMReg< 9, "q9", [D18, D19]>; +def Q10 : ARMReg<10, "q10", [D20, D21]>; +def Q11 : ARMReg<11, "q11", [D22, D23]>; +def Q12 : ARMReg<12, "q12", [D24, D25]>; +def Q13 : ARMReg<13, "q13", [D26, D27]>; +def Q14 : ARMReg<14, "q14", [D28, D29]>; +def Q15 : ARMReg<15, "q15", [D30, D31]>; +} + +// Current Program Status Register. +// We model fpscr with two registers: FPSCR models the control bits and will be +// reserved. FPSCR_NZCV models the flag bits and will be unreserved. APSR_NZCV +// models the APSR when it's accessed by some special instructions. In such cases +// it has the same encoding as PC. +def CPSR : ARMReg<0, "cpsr">; +def APSR : ARMReg<1, "apsr">; +def APSR_NZCV : ARMReg<15, "apsr_nzcv">; +def SPSR : ARMReg<2, "spsr">; +def FPSCR : ARMReg<3, "fpscr">; +def FPSCR_NZCV : ARMReg<3, "fpscr_nzcv"> { + let Aliases = [FPSCR]; +} +def ITSTATE : ARMReg<4, "itstate">; + +// Special Registers - only available in privileged mode. +def FPSID : ARMReg<0, "fpsid">; +def MVFR2 : ARMReg<5, "mvfr2">; +def MVFR1 : ARMReg<6, "mvfr1">; +def MVFR0 : ARMReg<7, "mvfr0">; +def FPEXC : ARMReg<8, "fpexc">; +def FPINST : ARMReg<9, "fpinst">; +def FPINST2 : ARMReg<10, "fpinst2">; + +// Register classes. +// +// pc == Program Counter +// lr == Link Register +// sp == Stack Pointer +// r12 == ip (scratch) +// r7 == Frame Pointer (thumb-style backtraces) +// r9 == May be reserved as Thread Register +// r11 == Frame Pointer (arm-style backtraces) +// r10 == Stack Limit +// +def GPR : RegisterClass<"ARM", [i32], 32, (add (sequence "R%u", 0, 12), + SP, LR, PC)> { + // Allocate LR as the first CSR since it is always saved anyway. + // For Thumb1 mode, we don't want to allocate hi regs at all, as we don't + // know how to spill them. If we make our prologue/epilogue code smarter at + // some point, we can go back to using the above allocation orders for the + // Thumb1 instructions that know how to use hi regs. + let AltOrders = [(add LR, GPR), (trunc GPR, 8)]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; + let DiagnosticString = "operand must be a register in range [r0, r15]"; +} + +// GPRs without the PC. Some ARM instructions do not allow the PC in +// certain operand slots, particularly as the destination. Primarily +// useful for disassembly. +def GPRnopc : RegisterClass<"ARM", [i32], 32, (sub GPR, PC)> { + let AltOrders = [(add LR, GPRnopc), (trunc GPRnopc, 8)]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; + let DiagnosticString = "operand must be a register in range [r0, r14]"; +} + +// GPRs without the PC but with APSR. Some instructions allow accessing the +// APSR, while actually encoding PC in the register field. This is useful +// for assembly and disassembly only. +def GPRwithAPSR : RegisterClass<"ARM", [i32], 32, (add (sub GPR, PC), APSR_NZCV)> { + let AltOrders = [(add LR, GPRnopc), (trunc GPRnopc, 8)]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; + let DiagnosticString = "operand must be a register in range [r0, r14] or apsr_nzcv"; +} + +// GPRsp - Only the SP is legal. Used by Thumb1 instructions that want the +// implied SP argument list. +// FIXME: It would be better to not use this at all and refactor the +// instructions to not have SP an an explicit argument. That makes +// frame index resolution a bit trickier, though. +def GPRsp : RegisterClass<"ARM", [i32], 32, (add SP)> { + let DiagnosticString = "operand must be a register sp"; +} + +// restricted GPR register class. Many Thumb2 instructions allow the full +// register range for operands, but have undefined behaviours when PC +// or SP (R13 or R15) are used. The ARM ISA refers to these operands +// via the BadReg() pseudo-code description. +def rGPR : RegisterClass<"ARM", [i32], 32, (sub GPR, SP, PC)> { + let AltOrders = [(add LR, rGPR), (trunc rGPR, 8)]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; + let DiagnosticType = "rGPR"; +} + +// Thumb registers are R0-R7 normally. Some instructions can still use +// the general GPR register class above (MOV, e.g.) +def tGPR : RegisterClass<"ARM", [i32], 32, (trunc GPR, 8)> { + let DiagnosticString = "operand must be a register in range [r0, r7]"; +} + +// Thumb registers R0-R7 and the PC. Some instructions like TBB or THH allow +// the PC to be used as a destination operand as well. +def tGPRwithpc : RegisterClass<"ARM", [i32], 32, (add tGPR, PC)>; + +// The high registers in thumb mode, R8-R15. +def hGPR : RegisterClass<"ARM", [i32], 32, (sub GPR, tGPR)> { + let DiagnosticString = "operand must be a register in range [r8, r15]"; +} + +// For tail calls, we can't use callee-saved registers, as they are restored +// to the saved value before the tail call, which would clobber a call address. +// Note, getMinimalPhysRegClass(R0) returns tGPR because of the names of +// this class and the preceding one(!) This is what we want. +def tcGPR : RegisterClass<"ARM", [i32], 32, (add R0, R1, R2, R3, R12)> { + let AltOrders = [(and tcGPR, tGPR)]; + let AltOrderSelect = [{ + return MF.getSubtarget<ARMSubtarget>().isThumb1Only(); + }]; +} + +// Condition code registers. +def CCR : RegisterClass<"ARM", [i32], 32, (add CPSR)> { + let CopyCost = -1; // Don't allow copying of status registers. + let isAllocatable = 0; +} + +// Scalar single precision floating point register class.. +// FIXME: Allocation order changed to s0, s2, ... or s0, s4, ... as a quick hack +// to avoid partial-write dependencies on D or Q (depending on platform) +// registers (S registers are renamed as portions of D/Q registers). +def SPR : RegisterClass<"ARM", [f32], 32, (sequence "S%u", 0, 31)> { + let AltOrders = [(add (decimate SPR, 2), SPR), + (add (decimate SPR, 4), + (decimate SPR, 2), + (decimate (rotl SPR, 1), 4), + (decimate (rotl SPR, 1), 2))]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().useStride4VFPs(MF); + }]; + let DiagnosticString = "operand must be a register in range [s0, s31]"; +} + +def HPR : RegisterClass<"ARM", [f16], 32, (sequence "S%u", 0, 31)> { + let AltOrders = [(add (decimate HPR, 2), SPR), + (add (decimate HPR, 4), + (decimate HPR, 2), + (decimate (rotl HPR, 1), 4), + (decimate (rotl HPR, 1), 2))]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().useStride4VFPs(MF); + }]; + let DiagnosticString = "operand must be a register in range [s0, s31]"; +} + +// Subset of SPR which can be used as a source of NEON scalars for 16-bit +// operations +def SPR_8 : RegisterClass<"ARM", [f32], 32, (sequence "S%u", 0, 15)> { + let DiagnosticString = "operand must be a register in range [s0, s15]"; +} + +// Scalar double precision floating point / generic 64-bit vector register +// class. +// ARM requires only word alignment for double. It's more performant if it +// is double-word alignment though. +def DPR : RegisterClass<"ARM", [f64, v8i8, v4i16, v2i32, v1i64, v2f32, v4f16], 64, + (sequence "D%u", 0, 31)> { + // Allocate non-VFP2 registers D16-D31 first, and prefer even registers on + // Darwin platforms. + let AltOrders = [(rotl DPR, 16), + (add (decimate (rotl DPR, 16), 2), (rotl DPR, 16))]; + let AltOrderSelect = [{ + return 1 + MF.getSubtarget<ARMSubtarget>().useStride4VFPs(MF); + }]; + let DiagnosticType = "DPR"; +} + +// Subset of DPR that are accessible with VFP2 (and so that also have +// 32-bit SPR subregs). +def DPR_VFP2 : RegisterClass<"ARM", [f64, v8i8, v4i16, v2i32, v1i64, v2f32, v4f16], 64, + (trunc DPR, 16)> { + let DiagnosticString = "operand must be a register in range [d0, d15]"; +} + +// Subset of DPR which can be used as a source of NEON scalars for 16-bit +// operations +def DPR_8 : RegisterClass<"ARM", [f64, v8i8, v4i16, v2i32, v1i64, v2f32, v4f16], 64, + (trunc DPR, 8)> { + let DiagnosticString = "operand must be a register in range [d0, d7]"; +} + +// Generic 128-bit vector register class. +def QPR : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64, v8f16], 128, + (sequence "Q%u", 0, 15)> { + // Allocate non-VFP2 aliases Q8-Q15 first. + let AltOrders = [(rotl QPR, 8)]; + let AltOrderSelect = [{ return 1; }]; + let DiagnosticString = "operand must be a register in range [q0, q15]"; +} + +// Subset of QPR that have 32-bit SPR subregs. +def QPR_VFP2 : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], + 128, (trunc QPR, 8)> { + let DiagnosticString = "operand must be a register in range [q0, q7]"; +} + +// Subset of QPR that have DPR_8 and SPR_8 subregs. +def QPR_8 : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], + 128, (trunc QPR, 4)> { + let DiagnosticString = "operand must be a register in range [q0, q3]"; +} + +// Pseudo-registers representing odd-even pairs of D registers. The even-odd +// pairs are already represented by the Q registers. +// These are needed by NEON instructions requiring two consecutive D registers. +// There is no D31_D0 register as that is always an UNPREDICTABLE encoding. +def TuplesOE2D : RegisterTuples<[dsub_0, dsub_1], + [(decimate (shl DPR, 1), 2), + (decimate (shl DPR, 2), 2)]>; + +// Register class representing a pair of consecutive D registers. +// Use the Q registers for the even-odd pairs. +def DPair : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], + 128, (interleave QPR, TuplesOE2D)> { + // Allocate starting at non-VFP2 registers D16-D31 first. + // Prefer even-odd pairs as they are easier to copy. + let AltOrders = [(add (rotl QPR, 8), (rotl DPair, 16))]; + let AltOrderSelect = [{ return 1; }]; +} + +// Pseudo-registers representing even-odd pairs of GPRs from R1 to R13/SP. +// These are needed by instructions (e.g. ldrexd/strexd) requiring even-odd GPRs. +def Tuples2R : RegisterTuples<[gsub_0, gsub_1], + [(add R0, R2, R4, R6, R8, R10, R12), + (add R1, R3, R5, R7, R9, R11, SP)]>; + +// Register class representing a pair of even-odd GPRs. +def GPRPair : RegisterClass<"ARM", [untyped], 64, (add Tuples2R)> { + let Size = 64; // 2 x 32 bits, we have no predefined type of that size. +} + +// Pseudo-registers representing 3 consecutive D registers. +def Tuples3D : RegisterTuples<[dsub_0, dsub_1, dsub_2], + [(shl DPR, 0), + (shl DPR, 1), + (shl DPR, 2)]>; + +// 3 consecutive D registers. +def DTriple : RegisterClass<"ARM", [untyped], 64, (add Tuples3D)> { + let Size = 192; // 3 x 64 bits, we have no predefined type of that size. +} + +// Pseudo 256-bit registers to represent pairs of Q registers. These should +// never be present in the emitted code. +// These are used for NEON load / store instructions, e.g., vld4, vst3. +def Tuples2Q : RegisterTuples<[qsub_0, qsub_1], [(shl QPR, 0), (shl QPR, 1)]>; + +// Pseudo 256-bit vector register class to model pairs of Q registers +// (4 consecutive D registers). +def QQPR : RegisterClass<"ARM", [v4i64], 256, (add Tuples2Q)> { + // Allocate non-VFP2 aliases first. + let AltOrders = [(rotl QQPR, 8)]; + let AltOrderSelect = [{ return 1; }]; +} + +// Tuples of 4 D regs that isn't also a pair of Q regs. +def TuplesOE4D : RegisterTuples<[dsub_0, dsub_1, dsub_2, dsub_3], + [(decimate (shl DPR, 1), 2), + (decimate (shl DPR, 2), 2), + (decimate (shl DPR, 3), 2), + (decimate (shl DPR, 4), 2)]>; + +// 4 consecutive D registers. +def DQuad : RegisterClass<"ARM", [v4i64], 256, + (interleave Tuples2Q, TuplesOE4D)>; + +// Pseudo 512-bit registers to represent four consecutive Q registers. +def Tuples2QQ : RegisterTuples<[qqsub_0, qqsub_1], + [(shl QQPR, 0), (shl QQPR, 2)]>; + +// Pseudo 512-bit vector register class to model 4 consecutive Q registers +// (8 consecutive D registers). +def QQQQPR : RegisterClass<"ARM", [v8i64], 256, (add Tuples2QQ)> { + // Allocate non-VFP2 aliases first. + let AltOrders = [(rotl QQQQPR, 8)]; + let AltOrderSelect = [{ return 1; }]; +} + + +// Pseudo-registers representing 2-spaced consecutive D registers. +def Tuples2DSpc : RegisterTuples<[dsub_0, dsub_2], + [(shl DPR, 0), + (shl DPR, 2)]>; + +// Spaced pairs of D registers. +def DPairSpc : RegisterClass<"ARM", [v2i64], 64, (add Tuples2DSpc)>; + +def Tuples3DSpc : RegisterTuples<[dsub_0, dsub_2, dsub_4], + [(shl DPR, 0), + (shl DPR, 2), + (shl DPR, 4)]>; + +// Spaced triples of D registers. +def DTripleSpc : RegisterClass<"ARM", [untyped], 64, (add Tuples3DSpc)> { + let Size = 192; // 3 x 64 bits, we have no predefined type of that size. +} + +def Tuples4DSpc : RegisterTuples<[dsub_0, dsub_2, dsub_4, dsub_6], + [(shl DPR, 0), + (shl DPR, 2), + (shl DPR, 4), + (shl DPR, 6)]>; + +// Spaced quads of D registers. +def DQuadSpc : RegisterClass<"ARM", [v4i64], 64, (add Tuples3DSpc)>; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMSchedule.td b/capstone/suite/synctools/tablegen/ARM/ARMSchedule.td new file mode 100644 index 000000000..ed5a3a7bb --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMSchedule.td @@ -0,0 +1,428 @@ +//===-- ARMSchedule.td - ARM Scheduling Definitions --------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +//===----------------------------------------------------------------------===// +// Instruction scheduling annotations for in-order and out-of-order CPUs. +// These annotations are independent of the itinerary class defined below. +// Here we define the subtarget independent read/write per-operand resources. +// The subtarget schedule definitions will then map these to the subtarget's +// resource usages. +// For example: +// The instruction cycle timings table might contain an entry for an operation +// like the following: +// Rd <- ADD Rn, Rm, <shift> Rs +// Uops | Latency from register | Uops - resource requirements - latency +// 2 | Rn: 1 Rm: 4 Rs: 4 | uop T0, Rm, Rs - P01 - 3 +// | | uopc Rd, Rn, T0 - P01 - 1 +// This is telling us that the result will be available in destination register +// Rd after a minimum of three cycles after the result in Rm and Rs is available +// and one cycle after the result in Rn is available. The micro-ops can execute +// on resource P01. +// To model this, we need to express that we need to dispatch two micro-ops, +// that the resource P01 is needed and that the latency to Rn is different than +// the latency to Rm and Rs. The scheduler can decrease Rn's producer latency by +// two. +// We will do this by assigning (abstract) resources to register defs/uses. +// ARMSchedule.td: +// def WriteALUsr : SchedWrite; +// def ReadAdvanceALUsr : ScheRead; +// +// ARMInstrInfo.td: +// def ADDrs : I<>, Sched<[WriteALUsr, ReadAdvanceALUsr, ReadDefault, +// ReadDefault]> { ...} +// ReadAdvance read resources allow us to define "pipeline by-passes" or +// shorter latencies to certain registers as needed in the example above. +// The "ReadDefault" can be omitted. +// Next, the subtarget td file assigns resources to the abstract resources +// defined here. +// ARMScheduleSubtarget.td: +// // Resources. +// def P01 : ProcResource<3>; // ALU unit (3 of it). +// ... +// // Resource usages. +// def : WriteRes<WriteALUsr, [P01, P01]> { +// Latency = 4; // Latency of 4. +// NumMicroOps = 2; // Dispatch 2 micro-ops. +// // The two instances of resource P01 are occupied for one cycle. It is one +// // cycle because these resources happen to be pipelined. +// ResourceCycles = [1, 1]; +// } +// def : ReadAdvance<ReadAdvanceALUsr, 3>; + +//===----------------------------------------------------------------------===// +// Sched definitions for integer pipeline instructions +// +// Basic ALU operation. +def WriteALU : SchedWrite; +def ReadALU : SchedRead; + +// Basic ALU with shifts. +def WriteALUsi : SchedWrite; // Shift by immediate. +def WriteALUsr : SchedWrite; // Shift by register. +def WriteALUSsr : SchedWrite; // Shift by register (flag setting). +def ReadALUsr : SchedRead; // Some operands are read later. + +// Compares. +def WriteCMP : SchedWrite; +def WriteCMPsi : SchedWrite; +def WriteCMPsr : SchedWrite; + +// Multiplys. +def WriteMUL16 : SchedWrite; // 16-bit multiply. +def WriteMUL32 : SchedWrite; // 32-bit multiply. +def WriteMUL64Lo : SchedWrite; // 64-bit result. Low reg. +def WriteMUL64Hi : SchedWrite; // 64-bit result. High reg. +def ReadMUL : SchedRead; + +// Multiply-accumulates. +def WriteMAC16 : SchedWrite; // 16-bit mac. +def WriteMAC32 : SchedWrite; // 32-bit mac. +def WriteMAC64Lo : SchedWrite; // 64-bit mac. Low reg. +def WriteMAC64Hi : SchedWrite; // 64-bit mac. High reg. +def ReadMAC : SchedRead; + +// Divisions. +def WriteDIV : SchedWrite; + +// Loads/Stores. +def WriteLd : SchedWrite; +def WritePreLd : SchedWrite; +def WriteST : SchedWrite; + +// Branches. +def WriteBr : SchedWrite; +def WriteBrL : SchedWrite; +def WriteBrTbl : SchedWrite; + +// Noop. +def WriteNoop : SchedWrite; + +//===----------------------------------------------------------------------===// +// Sched definitions for floating-point and neon instructions +// +// Floating point conversions +def WriteFPCVT : SchedWrite; +def WriteFPMOV : SchedWrite; // FP -> GPR and vice-versa + +// ALU operations (32/64-bit) +def WriteFPALU32 : SchedWrite; +def WriteFPALU64 : SchedWrite; + +// Multiplication +def WriteFPMUL32 : SchedWrite; +def WriteFPMUL64 : SchedWrite; +def ReadFPMUL : SchedRead; // multiplier read +def ReadFPMAC : SchedRead; // accumulator read + +// Multiply-accumulate +def WriteFPMAC32 : SchedWrite; +def WriteFPMAC64 : SchedWrite; + +// Division +def WriteFPDIV32 : SchedWrite; +def WriteFPDIV64 : SchedWrite; + +// Square-root +def WriteFPSQRT32 : SchedWrite; +def WriteFPSQRT64 : SchedWrite; + +// Vector load and stores +def WriteVLD1 : SchedWrite; +def WriteVLD2 : SchedWrite; +def WriteVLD3 : SchedWrite; +def WriteVLD4 : SchedWrite; +def WriteVST1 : SchedWrite; +def WriteVST2 : SchedWrite; +def WriteVST3 : SchedWrite; +def WriteVST4 : SchedWrite; + + +// Define TII for use in SchedVariant Predicates. +def : PredicateProlog<[{ + const ARMBaseInstrInfo *TII = + static_cast<const ARMBaseInstrInfo*>(SchedModel->getInstrInfo()); + (void)TII; + const ARMSubtarget *STI = + static_cast<const ARMSubtarget*>(SchedModel->getSubtargetInfo()); + (void)STI; +}]>; + +def IsPredicatedPred : SchedPredicate<[{TII->isPredicated(*MI)}]>; + +//===----------------------------------------------------------------------===// +// Instruction Itinerary classes used for ARM +// +def IIC_iALUx : InstrItinClass; +def IIC_iALUi : InstrItinClass; +def IIC_iALUr : InstrItinClass; +def IIC_iALUsi : InstrItinClass; +def IIC_iALUsir : InstrItinClass; +def IIC_iALUsr : InstrItinClass; +def IIC_iBITi : InstrItinClass; +def IIC_iBITr : InstrItinClass; +def IIC_iBITsi : InstrItinClass; +def IIC_iBITsr : InstrItinClass; +def IIC_iUNAr : InstrItinClass; +def IIC_iUNAsi : InstrItinClass; +def IIC_iEXTr : InstrItinClass; +def IIC_iEXTAr : InstrItinClass; +def IIC_iEXTAsr : InstrItinClass; +def IIC_iCMPi : InstrItinClass; +def IIC_iCMPr : InstrItinClass; +def IIC_iCMPsi : InstrItinClass; +def IIC_iCMPsr : InstrItinClass; +def IIC_iTSTi : InstrItinClass; +def IIC_iTSTr : InstrItinClass; +def IIC_iTSTsi : InstrItinClass; +def IIC_iTSTsr : InstrItinClass; +def IIC_iMOVi : InstrItinClass; +def IIC_iMOVr : InstrItinClass; +def IIC_iMOVsi : InstrItinClass; +def IIC_iMOVsr : InstrItinClass; +def IIC_iMOVix2 : InstrItinClass; +def IIC_iMOVix2addpc : InstrItinClass; +def IIC_iMOVix2ld : InstrItinClass; +def IIC_iMVNi : InstrItinClass; +def IIC_iMVNr : InstrItinClass; +def IIC_iMVNsi : InstrItinClass; +def IIC_iMVNsr : InstrItinClass; +def IIC_iCMOVi : InstrItinClass; +def IIC_iCMOVr : InstrItinClass; +def IIC_iCMOVsi : InstrItinClass; +def IIC_iCMOVsr : InstrItinClass; +def IIC_iCMOVix2 : InstrItinClass; +def IIC_iMUL16 : InstrItinClass; +def IIC_iMAC16 : InstrItinClass; +def IIC_iMUL32 : InstrItinClass; +def IIC_iMAC32 : InstrItinClass; +def IIC_iMUL64 : InstrItinClass; +def IIC_iMAC64 : InstrItinClass; +def IIC_iDIV : InstrItinClass; +def IIC_iLoad_i : InstrItinClass; +def IIC_iLoad_r : InstrItinClass; +def IIC_iLoad_si : InstrItinClass; +def IIC_iLoad_iu : InstrItinClass; +def IIC_iLoad_ru : InstrItinClass; +def IIC_iLoad_siu : InstrItinClass; +def IIC_iLoad_bh_i : InstrItinClass; +def IIC_iLoad_bh_r : InstrItinClass; +def IIC_iLoad_bh_si : InstrItinClass; +def IIC_iLoad_bh_iu : InstrItinClass; +def IIC_iLoad_bh_ru : InstrItinClass; +def IIC_iLoad_bh_siu : InstrItinClass; +def IIC_iLoad_d_i : InstrItinClass; +def IIC_iLoad_d_r : InstrItinClass; +def IIC_iLoad_d_ru : InstrItinClass; +def IIC_iLoad_m : InstrItinClass; +def IIC_iLoad_mu : InstrItinClass; +def IIC_iLoad_mBr : InstrItinClass; +def IIC_iPop : InstrItinClass; +def IIC_iPop_Br : InstrItinClass; +def IIC_iLoadiALU : InstrItinClass; +def IIC_iStore_i : InstrItinClass; +def IIC_iStore_r : InstrItinClass; +def IIC_iStore_si : InstrItinClass; +def IIC_iStore_iu : InstrItinClass; +def IIC_iStore_ru : InstrItinClass; +def IIC_iStore_siu : InstrItinClass; +def IIC_iStore_bh_i : InstrItinClass; +def IIC_iStore_bh_r : InstrItinClass; +def IIC_iStore_bh_si : InstrItinClass; +def IIC_iStore_bh_iu : InstrItinClass; +def IIC_iStore_bh_ru : InstrItinClass; +def IIC_iStore_bh_siu : InstrItinClass; +def IIC_iStore_d_i : InstrItinClass; +def IIC_iStore_d_r : InstrItinClass; +def IIC_iStore_d_ru : InstrItinClass; +def IIC_iStore_m : InstrItinClass; +def IIC_iStore_mu : InstrItinClass; +def IIC_Preload : InstrItinClass; +def IIC_Br : InstrItinClass; +def IIC_fpSTAT : InstrItinClass; +def IIC_fpUNA16 : InstrItinClass; +def IIC_fpUNA32 : InstrItinClass; +def IIC_fpUNA64 : InstrItinClass; +def IIC_fpCMP16 : InstrItinClass; +def IIC_fpCMP32 : InstrItinClass; +def IIC_fpCMP64 : InstrItinClass; +def IIC_fpCVTSD : InstrItinClass; +def IIC_fpCVTDS : InstrItinClass; +def IIC_fpCVTSH : InstrItinClass; +def IIC_fpCVTHS : InstrItinClass; +def IIC_fpCVTIH : InstrItinClass; +def IIC_fpCVTIS : InstrItinClass; +def IIC_fpCVTID : InstrItinClass; +def IIC_fpCVTHI : InstrItinClass; +def IIC_fpCVTSI : InstrItinClass; +def IIC_fpCVTDI : InstrItinClass; +def IIC_fpMOVIS : InstrItinClass; +def IIC_fpMOVID : InstrItinClass; +def IIC_fpMOVSI : InstrItinClass; +def IIC_fpMOVDI : InstrItinClass; +def IIC_fpALU16 : InstrItinClass; +def IIC_fpALU32 : InstrItinClass; +def IIC_fpALU64 : InstrItinClass; +def IIC_fpMUL16 : InstrItinClass; +def IIC_fpMUL32 : InstrItinClass; +def IIC_fpMUL64 : InstrItinClass; +def IIC_fpMAC16 : InstrItinClass; +def IIC_fpMAC32 : InstrItinClass; +def IIC_fpMAC64 : InstrItinClass; +def IIC_fpFMAC16 : InstrItinClass; +def IIC_fpFMAC32 : InstrItinClass; +def IIC_fpFMAC64 : InstrItinClass; +def IIC_fpDIV16 : InstrItinClass; +def IIC_fpDIV32 : InstrItinClass; +def IIC_fpDIV64 : InstrItinClass; +def IIC_fpSQRT16 : InstrItinClass; +def IIC_fpSQRT32 : InstrItinClass; +def IIC_fpSQRT64 : InstrItinClass; +def IIC_fpLoad16 : InstrItinClass; +def IIC_fpLoad32 : InstrItinClass; +def IIC_fpLoad64 : InstrItinClass; +def IIC_fpLoad_m : InstrItinClass; +def IIC_fpLoad_mu : InstrItinClass; +def IIC_fpStore16 : InstrItinClass; +def IIC_fpStore32 : InstrItinClass; +def IIC_fpStore64 : InstrItinClass; +def IIC_fpStore_m : InstrItinClass; +def IIC_fpStore_mu : InstrItinClass; +def IIC_VLD1 : InstrItinClass; +def IIC_VLD1x2 : InstrItinClass; +def IIC_VLD1x3 : InstrItinClass; +def IIC_VLD1x4 : InstrItinClass; +def IIC_VLD1u : InstrItinClass; +def IIC_VLD1x2u : InstrItinClass; +def IIC_VLD1x3u : InstrItinClass; +def IIC_VLD1x4u : InstrItinClass; +def IIC_VLD1ln : InstrItinClass; +def IIC_VLD1lnu : InstrItinClass; +def IIC_VLD1dup : InstrItinClass; +def IIC_VLD1dupu : InstrItinClass; +def IIC_VLD2 : InstrItinClass; +def IIC_VLD2x2 : InstrItinClass; +def IIC_VLD2u : InstrItinClass; +def IIC_VLD2x2u : InstrItinClass; +def IIC_VLD2ln : InstrItinClass; +def IIC_VLD2lnu : InstrItinClass; +def IIC_VLD2dup : InstrItinClass; +def IIC_VLD2dupu : InstrItinClass; +def IIC_VLD3 : InstrItinClass; +def IIC_VLD3ln : InstrItinClass; +def IIC_VLD3u : InstrItinClass; +def IIC_VLD3lnu : InstrItinClass; +def IIC_VLD3dup : InstrItinClass; +def IIC_VLD3dupu : InstrItinClass; +def IIC_VLD4 : InstrItinClass; +def IIC_VLD4ln : InstrItinClass; +def IIC_VLD4u : InstrItinClass; +def IIC_VLD4lnu : InstrItinClass; +def IIC_VLD4dup : InstrItinClass; +def IIC_VLD4dupu : InstrItinClass; +def IIC_VST1 : InstrItinClass; +def IIC_VST1x2 : InstrItinClass; +def IIC_VST1x3 : InstrItinClass; +def IIC_VST1x4 : InstrItinClass; +def IIC_VST1u : InstrItinClass; +def IIC_VST1x2u : InstrItinClass; +def IIC_VST1x3u : InstrItinClass; +def IIC_VST1x4u : InstrItinClass; +def IIC_VST1ln : InstrItinClass; +def IIC_VST1lnu : InstrItinClass; +def IIC_VST2 : InstrItinClass; +def IIC_VST2x2 : InstrItinClass; +def IIC_VST2u : InstrItinClass; +def IIC_VST2x2u : InstrItinClass; +def IIC_VST2ln : InstrItinClass; +def IIC_VST2lnu : InstrItinClass; +def IIC_VST3 : InstrItinClass; +def IIC_VST3u : InstrItinClass; +def IIC_VST3ln : InstrItinClass; +def IIC_VST3lnu : InstrItinClass; +def IIC_VST4 : InstrItinClass; +def IIC_VST4u : InstrItinClass; +def IIC_VST4ln : InstrItinClass; +def IIC_VST4lnu : InstrItinClass; +def IIC_VUNAD : InstrItinClass; +def IIC_VUNAQ : InstrItinClass; +def IIC_VBIND : InstrItinClass; +def IIC_VBINQ : InstrItinClass; +def IIC_VPBIND : InstrItinClass; +def IIC_VFMULD : InstrItinClass; +def IIC_VFMULQ : InstrItinClass; +def IIC_VMOV : InstrItinClass; +def IIC_VMOVImm : InstrItinClass; +def IIC_VMOVD : InstrItinClass; +def IIC_VMOVQ : InstrItinClass; +def IIC_VMOVIS : InstrItinClass; +def IIC_VMOVID : InstrItinClass; +def IIC_VMOVISL : InstrItinClass; +def IIC_VMOVSI : InstrItinClass; +def IIC_VMOVDI : InstrItinClass; +def IIC_VMOVN : InstrItinClass; +def IIC_VPERMD : InstrItinClass; +def IIC_VPERMQ : InstrItinClass; +def IIC_VPERMQ3 : InstrItinClass; +def IIC_VMACD : InstrItinClass; +def IIC_VMACQ : InstrItinClass; +def IIC_VFMACD : InstrItinClass; +def IIC_VFMACQ : InstrItinClass; +def IIC_VRECSD : InstrItinClass; +def IIC_VRECSQ : InstrItinClass; +def IIC_VCNTiD : InstrItinClass; +def IIC_VCNTiQ : InstrItinClass; +def IIC_VUNAiD : InstrItinClass; +def IIC_VUNAiQ : InstrItinClass; +def IIC_VQUNAiD : InstrItinClass; +def IIC_VQUNAiQ : InstrItinClass; +def IIC_VBINiD : InstrItinClass; +def IIC_VBINiQ : InstrItinClass; +def IIC_VSUBiD : InstrItinClass; +def IIC_VSUBiQ : InstrItinClass; +def IIC_VBINi4D : InstrItinClass; +def IIC_VBINi4Q : InstrItinClass; +def IIC_VSUBi4D : InstrItinClass; +def IIC_VSUBi4Q : InstrItinClass; +def IIC_VABAD : InstrItinClass; +def IIC_VABAQ : InstrItinClass; +def IIC_VSHLiD : InstrItinClass; +def IIC_VSHLiQ : InstrItinClass; +def IIC_VSHLi4D : InstrItinClass; +def IIC_VSHLi4Q : InstrItinClass; +def IIC_VPALiD : InstrItinClass; +def IIC_VPALiQ : InstrItinClass; +def IIC_VMULi16D : InstrItinClass; +def IIC_VMULi32D : InstrItinClass; +def IIC_VMULi16Q : InstrItinClass; +def IIC_VMULi32Q : InstrItinClass; +def IIC_VMACi16D : InstrItinClass; +def IIC_VMACi32D : InstrItinClass; +def IIC_VMACi16Q : InstrItinClass; +def IIC_VMACi32Q : InstrItinClass; +def IIC_VEXTD : InstrItinClass; +def IIC_VEXTQ : InstrItinClass; +def IIC_VTB1 : InstrItinClass; +def IIC_VTB2 : InstrItinClass; +def IIC_VTB3 : InstrItinClass; +def IIC_VTB4 : InstrItinClass; +def IIC_VTBX1 : InstrItinClass; +def IIC_VTBX2 : InstrItinClass; +def IIC_VTBX3 : InstrItinClass; +def IIC_VTBX4 : InstrItinClass; +def IIC_VDOTPROD : InstrItinClass; + +//===----------------------------------------------------------------------===// +// Processor instruction itineraries. + +include "ARMScheduleV6.td" +include "ARMScheduleA8.td" +include "ARMScheduleA9.td" +include "ARMScheduleSwift.td" +include "ARMScheduleR52.td" +include "ARMScheduleA57.td" +include "ARMScheduleM3.td" diff --git a/capstone/suite/synctools/tablegen/ARM/ARMScheduleA57.td b/capstone/suite/synctools/tablegen/ARM/ARMScheduleA57.td new file mode 100644 index 000000000..63f975ba6 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMScheduleA57.td @@ -0,0 +1,1502 @@ +//=- ARMScheduleA57.td - ARM Cortex-A57 Scheduling Defs -----*- tablegen -*-=// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the machine model for ARM Cortex-A57 to support +// instruction scheduling and other instruction cost heuristics. +// +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// *** Common description and scheduling model parameters taken from AArch64 *** +// The Cortex-A57 is a traditional superscalar microprocessor with a +// conservative 3-wide in-order stage for decode and dispatch. Combined with the +// much wider out-of-order issue stage, this produced a need to carefully +// schedule micro-ops so that all three decoded each cycle are successfully +// issued as the reservation station(s) simply don't stay occupied for long. +// Therefore, IssueWidth is set to the narrower of the two at three, while still +// modeling the machine as out-of-order. + +def IsCPSRDefinedPred : SchedPredicate<[{TII->isCPSRDefined(*MI)}]>; +def IsCPSRDefinedAndPredicatedPred : + SchedPredicate<[{TII->isCPSRDefined(*MI) && TII->isPredicated(*MI)}]>; + +// Cortex A57 rev. r1p0 or later (false = r0px) +def IsR1P0AndLaterPred : SchedPredicate<[{false}]>; + +// If Addrmode3 contains register offset (not immediate) +def IsLdrAm3RegOffPred : + SchedPredicate<[{!TII->isAddrMode3OpImm(*MI, 1)}]>; +// The same predicate with operand offset 2 and 3: +def IsLdrAm3RegOffPredX2 : + SchedPredicate<[{!TII->isAddrMode3OpImm(*MI, 2)}]>; +def IsLdrAm3RegOffPredX3 : + SchedPredicate<[{!TII->isAddrMode3OpImm(*MI, 3)}]>; + +// If Addrmode3 contains "minus register" +def IsLdrAm3NegRegOffPred : + SchedPredicate<[{TII->isAddrMode3OpMinusReg(*MI, 1)}]>; +// The same predicate with operand offset 2 and 3: +def IsLdrAm3NegRegOffPredX2 : + SchedPredicate<[{TII->isAddrMode3OpMinusReg(*MI, 2)}]>; +def IsLdrAm3NegRegOffPredX3 : + SchedPredicate<[{TII->isAddrMode3OpMinusReg(*MI, 3)}]>; + +// Load, scaled register offset, not plus LSL2 +def IsLdstsoScaledNotOptimalPredX0 : + SchedPredicate<[{TII->isLdstScaledRegNotPlusLsl2(*MI, 0)}]>; +def IsLdstsoScaledNotOptimalPred : + SchedPredicate<[{TII->isLdstScaledRegNotPlusLsl2(*MI, 1)}]>; +def IsLdstsoScaledNotOptimalPredX2 : + SchedPredicate<[{TII->isLdstScaledRegNotPlusLsl2(*MI, 2)}]>; + +// Load, scaled register offset +def IsLdstsoScaledPred : + SchedPredicate<[{TII->isLdstScaledReg(*MI, 1)}]>; +def IsLdstsoScaledPredX2 : + SchedPredicate<[{TII->isLdstScaledReg(*MI, 2)}]>; + +def IsLdstsoMinusRegPredX0 : + SchedPredicate<[{TII->isLdstSoMinusReg(*MI, 0)}]>; +def IsLdstsoMinusRegPred : + SchedPredicate<[{TII->isLdstSoMinusReg(*MI, 1)}]>; +def IsLdstsoMinusRegPredX2 : + SchedPredicate<[{TII->isLdstSoMinusReg(*MI, 2)}]>; + +// Load, scaled register offset +def IsLdrAm2ScaledPred : + SchedPredicate<[{TII->isAm2ScaledReg(*MI, 1)}]>; + +// LDM, base reg in list +def IsLdmBaseRegInList : + SchedPredicate<[{TII->isLDMBaseRegInList(*MI)}]>; + +class A57WriteLMOpsListType<list<SchedWriteRes> writes> { + list <SchedWriteRes> Writes = writes; + SchedMachineModel SchedModel = ?; +} + +// *** Common description and scheduling model parameters taken from AArch64 *** +// (AArch64SchedA57.td) +def CortexA57Model : SchedMachineModel { + let IssueWidth = 3; // 3-way decode and dispatch + let MicroOpBufferSize = 128; // 128 micro-op re-order buffer + let LoadLatency = 4; // Optimistic load latency + let MispredictPenalty = 16; // Fetch + Decode/Rename/Dispatch + Branch + + // Enable partial & runtime unrolling. + let LoopMicroOpBufferSize = 16; + let CompleteModel = 1; + + // FIXME: Remove when all errors have been fixed. + let FullInstRWOverlapCheck = 0; +} + +//===----------------------------------------------------------------------===// +// Define each kind of processor resource and number available on Cortex-A57. +// Cortex A-57 has 8 pipelines that each has its own 8-entry queue where +// micro-ops wait for their operands and then issue out-of-order. + +def A57UnitB : ProcResource<1>; // Type B micro-ops +def A57UnitI : ProcResource<2>; // Type I micro-ops +def A57UnitM : ProcResource<1>; // Type M micro-ops +def A57UnitL : ProcResource<1>; // Type L micro-ops +def A57UnitS : ProcResource<1>; // Type S micro-ops + +def A57UnitX : ProcResource<1>; // Type X micro-ops (F1) +def A57UnitW : ProcResource<1>; // Type W micro-ops (F0) + +let SchedModel = CortexA57Model in { + def A57UnitV : ProcResGroup<[A57UnitX, A57UnitW]>; // Type V micro-ops +} + +let SchedModel = CortexA57Model in { + +//===----------------------------------------------------------------------===// +// Define customized scheduler read/write types specific to the Cortex-A57. + +include "ARMScheduleA57WriteRes.td" + +// To have "CompleteModel = 1", support of pseudos and special instructions +def : InstRW<[WriteNoop], (instregex "(t)?BKPT$", "(t2)?CDP(2)?$", + "(t2)?CLREX$", "CONSTPOOL_ENTRY$", "COPY_STRUCT_BYVAL_I32$", + "(t2)?CPS[123]p$", "(t2)?DBG$", "(t2)?DMB$", "(t2)?DSB$", "ERET$", + "(t2|t)?HINT$", "(t)?HLT$", "(t2)?HVC$", "(t2)?ISB$", "ITasm$", + "(t2)?RFE(DA|DB|IA|IB)", "(t)?SETEND", "(t2)?SETPAN", "(t2)?SMC", "SPACE", + "(t2)?SRS(DA|DB|IA|IB)", "SWP(B)?", "t?TRAP", "(t2|t)?UDF$", "t2DCPS", "t2SG", + "t2TT", "tCPS", "CMP_SWAP", "t?SVC", "t2IT", "CompilerBarrier", + "t__brkdiv0")>; + +def : InstRW<[WriteNoop], (instregex "VMRS", "VMSR", "FMSTAT")>; + +// Specific memory instrs +def : InstRW<[WriteNoop, WriteNoop], (instregex "(t2)?LDA", "(t2)?LDC", "(t2)?STC", + "(t2)?STL", "(t2)?LDREX", "(t2)?STREX", "MEMCPY")>; + +// coprocessor moves +def : InstRW<[WriteNoop, WriteNoop], (instregex + "(t2)?MCR(2|R|R2)?$", "(t2)?MRC(2)?$", + "(t2)?MRRC(2)?$", "(t2)?MRS(banked|sys|_AR|_M|sys_AR)?$", + "(t2)?MSR(banked|i|_AR|_M)?$")>; + +// Deprecated instructions +def : InstRW<[WriteNoop], (instregex "FLDM", "FSTM")>; + +// Pseudos +def : InstRW<[WriteNoop], (instregex "(t2)?ABS$", + "(t)?ADJCALLSTACKDOWN$", "(t)?ADJCALLSTACKUP$", "(t2|t)?Int_eh_sjlj", + "tLDRpci_pic", "(t2)?SUBS_PC_LR", + "JUMPTABLE", "tInt_WIN_eh_sjlj_longjmp", + "VLD(1|2)LN(d|q)(WB_fixed_|WB_register_)?Asm", + "VLD(3|4)(DUP|LN)?(d|q)(WB_fixed_|WB_register_)?Asm", + "VST(1|2)LN(d|q)(WB_fixed_|WB_register_)?Asm", + "VST(3|4)(DUP|LN)?(d|q)(WB_fixed_|WB_register_)?Asm", + "WIN__CHKSTK", "WIN__DBZCHK")>; + +// Miscellaneous +// ----------------------------------------------------------------------------- + +def : InstRW<[A57Write_1cyc_1I], (instrs COPY)>; + +// --- 3.2 Branch Instructions --- +// B, BX, BL, BLX (imm, reg != LR, reg == LR), CBZ, CBNZ + +def : InstRW<[A57Write_1cyc_1B], (instregex "(t2|t)?B$", "t?BX", "(t2|t)?Bcc$", + "t?TAILJMP(d|r)", "TCRETURN(d|r)i", "tBfar", "tCBN?Z")>; +def : InstRW<[A57Write_1cyc_1B_1I], + (instregex "t?BL$", "BL_pred$", "t?BLXi", "t?TPsoft")>; +def : InstRW<[A57Write_2cyc_1B_1I], (instregex "BLX", "tBLX(NS)?r")>; +// Pseudos +def : InstRW<[A57Write_2cyc_1B_1I], (instregex "BCCi64", "BCCZi64")>; +def : InstRW<[A57Write_3cyc_1B_1I], (instregex "BR_JTadd", "t?BR_JTr", + "t2BR_JT", "t2BXJ", "(t2)?TB(B|H)(_JT)?$", "tBRIND")>; +def : InstRW<[A57Write_6cyc_1B_1L], (instregex "BR_JTm")>; + +// --- 3.3 Arithmetic and Logical Instructions --- +// ADD{S}, ADC{S}, ADR, AND{S}, BIC{S}, CMN, CMP, EOR{S}, ORN{S}, ORR{S}, +// RSB{S}, RSC{S}, SUB{S}, SBC{S}, TEQ, TST + +def : InstRW<[A57Write_1cyc_1I], (instregex "tADDframe")>; + +// shift by register, conditional or unconditional +// TODO: according to the doc, conditional uses I0/I1, unconditional uses M +// Why more complex instruction uses more simple pipeline? +// May be an error in doc. +def A57WriteALUsi : SchedWriteVariant<[ + // lsl #2, lsl #1, or lsr #1. + SchedVar<IsPredicatedPred, [A57Write_2cyc_1M]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1M]> +]>; +def A57WriteALUsr : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57Write_2cyc_1I]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1M]> +]>; +def A57WriteALUSsr : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57Write_2cyc_1I]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1M]> +]>; +def A57ReadALUsr : SchedReadVariant<[ + SchedVar<IsPredicatedPred, [ReadDefault]>, + SchedVar<NoSchedPred, [ReadDefault]> +]>; +def : SchedAlias<WriteALUsi, A57WriteALUsi>; +def : SchedAlias<WriteALUsr, A57WriteALUsr>; +def : SchedAlias<WriteALUSsr, A57WriteALUSsr>; +def : SchedAlias<ReadALUsr, A57ReadALUsr>; + +def A57WriteCMPsr : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57Write_2cyc_1I]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1M]> +]>; +def : SchedAlias<WriteCMP, A57Write_1cyc_1I>; +def : SchedAlias<WriteCMPsi, A57Write_2cyc_1M>; +def : SchedAlias<WriteCMPsr, A57WriteCMPsr>; + +// --- 3.4 Move and Shift Instructions --- +// Move, basic +// MOV{S}, MOVW, MVN{S} +def : InstRW<[A57Write_1cyc_1I], (instregex "MOV(r|i|i16|r_TC)", + "(t2)?MVN(CC)?(r|i)", "BMOVPCB_CALL", "BMOVPCRX_CALL", + "MOVCC(r|i|i16|i32imm)", "tMOV", "tMVN")>; + +// Move, shift by immed, setflags/no setflags +// (ASR, LSL, LSR, ROR, RRX)=MOVsi, MVN +// setflags = isCPSRDefined +def A57WriteMOVsi : SchedWriteVariant<[ + SchedVar<IsCPSRDefinedPred, [A57Write_2cyc_1M]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1I]> +]>; +def : InstRW<[A57WriteMOVsi], (instregex "MOV(CC)?si", "MVNsi", + "ASRi", "(t2|t)ASRri", "LSRi", "(t2|t)LSRri", "LSLi", "(t2|t)LSLri", "RORi", + "(t2|t)RORri", "(t2)?RRX", "t2MOV", "tROR")>; + +// shift by register, conditional or unconditional, setflags/no setflags +def A57WriteMOVsr : SchedWriteVariant<[ + SchedVar<IsCPSRDefinedAndPredicatedPred, [A57Write_2cyc_1I]>, + SchedVar<IsCPSRDefinedPred, [A57Write_2cyc_1M]>, + SchedVar<IsPredicatedPred, [A57Write_2cyc_1I]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1I]> +]>; +def : InstRW<[A57WriteMOVsr], (instregex "MOV(CC)?sr", "MVNsr", "t2MVNs", + "ASRr", "(t2|t)ASRrr", "LSRr", "(t2|t)LSRrr", "LSLr", "(t2|t)?LSLrr", "RORr", + "(t2|t)RORrr")>; + +// Move, top +// MOVT - A57Write_2cyc_1M for r0px, A57Write_1cyc_1I for r1p0 and later +def A57WriteMOVT : SchedWriteVariant<[ + SchedVar<IsR1P0AndLaterPred, [A57Write_1cyc_1I]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1M]> +]>; +def : InstRW<[A57WriteMOVT], (instregex "MOVTi16")>; + +def A57WriteI2pc : + WriteSequence<[A57Write_1cyc_1I, A57Write_1cyc_1I, A57Write_1cyc_1I]>; +def A57WriteI2ld : + WriteSequence<[A57Write_1cyc_1I, A57Write_1cyc_1I, A57Write_4cyc_1L]>; +def : InstRW< [A57WriteI2pc], (instregex "MOV_ga_pcrel")>; +def : InstRW< [A57WriteI2ld], (instregex "MOV_ga_pcrel_ldr")>; + +// +2cyc for branch forms +def : InstRW<[A57Write_3cyc_1I], (instregex "MOVPC(LR|RX)")>; + +// --- 3.5 Divide and Multiply Instructions --- +// Divide: SDIV, UDIV +// latency from documentration: 4 ‐ 20, maximum taken +def : SchedAlias<WriteDIV, A57Write_20cyc_1M>; +// Multiply: tMul not bound to common WriteRes types +def : InstRW<[A57Write_3cyc_1M], (instregex "tMUL")>; +def : SchedAlias<WriteMUL16, A57Write_3cyc_1M>; +def : SchedAlias<WriteMUL32, A57Write_3cyc_1M>; +def : ReadAdvance<ReadMUL, 0>; + +// Multiply accumulate: MLA, MLS, SMLABB, SMLABT, SMLATB, SMLATT, SMLAWB, +// SMLAWT, SMLAD{X}, SMLSD{X}, SMMLA{R}, SMMLS{R} +// Multiply-accumulate pipelines support late-forwarding of accumulate operands +// from similar μops, allowing a typical sequence of multiply-accumulate μops +// to issue one every 1 cycle (sched advance = 2). +def A57WriteMLA : SchedWriteRes<[A57UnitM]> { let Latency = 3; } +def A57WriteMLAL : SchedWriteRes<[A57UnitM]> { let Latency = 4; } +def A57ReadMLA : SchedReadAdvance<2, [A57WriteMLA, A57WriteMLAL]>; + +def : InstRW<[A57WriteMLA], + (instregex "t2SMLAD", "t2SMLADX", "t2SMLSD", "t2SMLSDX")>; + +def : SchedAlias<WriteMAC16, A57WriteMLA>; +def : SchedAlias<WriteMAC32, A57WriteMLA>; +def : SchedAlias<ReadMAC, A57ReadMLA>; + +def : SchedAlias<WriteMAC64Lo, A57WriteMLAL>; +def : SchedAlias<WriteMAC64Hi, A57WriteMLAL>; + +// Multiply long: SMULL, UMULL +def : SchedAlias<WriteMUL64Lo, A57Write_4cyc_1M>; +def : SchedAlias<WriteMUL64Hi, A57Write_4cyc_1M>; + +// --- 3.6 Saturating and Parallel Arithmetic Instructions --- +// Parallel arith +// SADD16, SADD8, SSUB16, SSUB8, UADD16, UADD8, USUB16, USUB8 +// Conditional GE-setting instructions require three extra μops +// and two additional cycles to conditionally update the GE field. +def A57WriteParArith : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57Write_4cyc_1I_1M]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1I_1M]> +]>; +def : InstRW< [A57WriteParArith], (instregex + "(t2)?SADD(16|8)", "(t2)?SSUB(16|8)", + "(t2)?UADD(16|8)", "(t2)?USUB(16|8)")>; + +// Parallel arith with exchange: SASX, SSAX, UASX, USAX +def A57WriteParArithExch : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57Write_5cyc_1I_1M]>, + SchedVar<NoSchedPred, [A57Write_3cyc_1I_1M]> +]>; +def : InstRW<[A57WriteParArithExch], + (instregex "(t2)?SASX", "(t2)?SSAX", "(t2)?UASX", "(t2)?USAX")>; + +// Parallel halving arith +// SHADD16, SHADD8, SHSUB16, SHSUB8, UHADD16, UHADD8, UHSUB16, UHSUB8 +def : InstRW<[A57Write_2cyc_1M], (instregex + "(t2)?SHADD(16|8)", "(t2)?SHSUB(16|8)", + "(t2)?UHADD(16|8)", "(t2)?UHSUB(16|8)")>; + +// Parallel halving arith with exchange +// SHASX, SHSAX, UHASX, UHSAX +def : InstRW<[A57Write_3cyc_1I_1M], (instregex "(t2)?SHASX", "(t2)?SHSAX", + "(t2)?UHASX", "(t2)?UHSAX")>; + +// Parallel saturating arith +// QADD16, QADD8, QSUB16, QSUB8, UQADD16, UQADD8, UQSUB16, UQSUB8 +def : InstRW<[A57Write_2cyc_1M], (instregex "QADD(16|8)", "QSUB(16|8)", + "UQADD(16|8)", "UQSUB(16|8)", "t2(U?)QADD", "t2(U?)QSUB")>; + +// Parallel saturating arith with exchange +// QASX, QSAX, UQASX, UQSAX +def : InstRW<[A57Write_3cyc_1I_1M], (instregex "(t2)?QASX", "(t2)?QSAX", + "(t2)?UQASX", "(t2)?UQSAX")>; + +// Saturate: SSAT, SSAT16, USAT, USAT16 +def : InstRW<[A57Write_2cyc_1M], + (instregex "(t2)?SSAT(16)?", "(t2)?USAT(16)?")>; + +// Saturating arith: QADD, QSUB +def : InstRW<[A57Write_2cyc_1M], (instregex "QADD$", "QSUB$")>; + +// Saturating doubling arith: QDADD, QDSUB +def : InstRW<[A57Write_3cyc_1I_1M], (instregex "(t2)?QDADD", "(t2)?QDSUB")>; + +// --- 3.7 Miscellaneous Data-Processing Instructions --- +// Bit field extract: SBFX, UBFX +def : InstRW<[A57Write_1cyc_1I], (instregex "(t2)?SBFX", "(t2)?UBFX")>; + +// Bit field insert/clear: BFI, BFC +def : InstRW<[A57Write_2cyc_1M], (instregex "(t2)?BFI", "(t2)?BFC")>; + +// Select bytes, conditional/unconditional +def A57WriteSEL : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57Write_2cyc_1I]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1I]> +]>; +def : InstRW<[A57WriteSEL], (instregex "(t2)?SEL")>; + +// Sign/zero extend, normal: SXTB, SXTH, UXTB, UXTH +def : InstRW<[A57Write_1cyc_1I], + (instregex "(t2|t)?SXT(B|H)$", "(t2|t)?UXT(B|H)$")>; + +// Sign/zero extend and add, normal: SXTAB, SXTAH, UXTAB, UXTAH +def : InstRW<[A57Write_2cyc_1M], + (instregex "(t2)?SXTA(B|H)$", "(t2)?UXTA(B|H)$")>; + +// Sign/zero extend and add, parallel: SXTAB16, UXTAB16 +def : InstRW<[A57Write_4cyc_1M], (instregex "(t2)?SXTAB16", "(t2)?UXTAB16")>; + +// Sum of absolute differences: USAD8, USADA8 +def : InstRW<[A57Write_3cyc_1M], (instregex "(t2)?USAD8", "(t2)?USADA8")>; + +// --- 3.8 Load Instructions --- + +// Load, immed offset +// LDR and LDRB have LDRi12 and LDRBi12 forms for immediate +def : InstRW<[A57Write_4cyc_1L], (instregex "LDRi12", "LDRBi12", + "LDRcp", "(t2|t)?LDRConstPool", "LDRLIT_ga_(pcrel|abs)", + "PICLDR", "tLDR")>; + +def : InstRW<[A57Write_4cyc_1L], + (instregex "t2LDRS?(B|H)?(pcrel|T|i8|i12|pci|pci_pic|s)?$")>; + +// For "Load, register offset, minus" we need +1cyc, +1I +def A57WriteLdrAm3 : SchedWriteVariant<[ + SchedVar<IsLdrAm3NegRegOffPred, [A57Write_5cyc_1I_1L]>, + SchedVar<NoSchedPred, [A57Write_4cyc_1L]> +]>; +def : InstRW<[A57WriteLdrAm3], (instregex "LDR(H|SH|SB)$")>; +def A57WriteLdrAm3X2 : SchedWriteVariant<[ + SchedVar<IsLdrAm3NegRegOffPredX2, [A57Write_5cyc_1I_1L]>, + SchedVar<NoSchedPred, [A57Write_4cyc_1L]> +]>; +def : InstRW<[A57WriteLdrAm3X2, A57WriteLdrAm3X2], (instregex "LDRD$")>; +def : InstRW<[A57Write_4cyc_1L, A57Write_4cyc_1L], (instregex "t2LDRDi8")>; + +def A57WriteLdrAmLDSTSO : SchedWriteVariant<[ + SchedVar<IsLdstsoScaledNotOptimalPred, [A57Write_5cyc_1I_1L]>, + SchedVar<IsLdstsoMinusRegPred, [A57Write_5cyc_1I_1L]>, + SchedVar<NoSchedPred, [A57Write_4cyc_1L]> +]>; +def : InstRW<[A57WriteLdrAmLDSTSO], (instregex "LDRrs", "LDRBrs")>; + +def A57WrBackOne : SchedWriteRes<[]> { + let Latency = 1; + let NumMicroOps = 0; +} +def A57WrBackTwo : SchedWriteRes<[]> { + let Latency = 2; + let NumMicroOps = 0; +} +def A57WrBackThree : SchedWriteRes<[]> { + let Latency = 3; + let NumMicroOps = 0; +} + +// --- LDR pre-indexed --- +// Load, immed pre-indexed (4 cyc for load result, 1 cyc for Base update) +def : InstRW<[A57Write_4cyc_1L_1I, A57WrBackOne], (instregex "LDR_PRE_IMM", + "LDRB_PRE_IMM", "t2LDRB_PRE")>; + +// Load, register pre-indexed (4 cyc for load result, 2 cyc for Base update) +// (5 cyc load result for not-lsl2 scaled) +def A57WriteLdrAmLDSTSOPre : SchedWriteVariant<[ + SchedVar<IsLdstsoScaledNotOptimalPredX2, [A57Write_5cyc_1I_1L]>, + SchedVar<NoSchedPred, [A57Write_4cyc_1L_1I]> +]>; +def : InstRW<[A57WriteLdrAmLDSTSOPre, A57WrBackTwo], + (instregex "LDR_PRE_REG", "LDRB_PRE_REG")>; + +def A57WriteLdrAm3PreWrBack : SchedWriteVariant<[ + SchedVar<IsLdrAm3RegOffPredX2, [A57WrBackTwo]>, + SchedVar<NoSchedPred, [A57WrBackOne]> +]>; +def : InstRW<[A57Write_4cyc_1L, A57WriteLdrAm3PreWrBack], + (instregex "LDR(H|SH|SB)_PRE")>; +def : InstRW<[A57Write_4cyc_1L, A57WrBackOne], + (instregex "t2LDR(H|SH|SB)?_PRE")>; + +// LDRD pre-indexed: 5(2) cyc for reg, 4(1) cyc for imm. +def A57WriteLdrDAm3Pre : SchedWriteVariant<[ + SchedVar<IsLdrAm3RegOffPredX3, [A57Write_5cyc_1I_1L]>, + SchedVar<NoSchedPred, [A57Write_4cyc_1L_1I]> +]>; +def A57WriteLdrDAm3PreWrBack : SchedWriteVariant<[ + SchedVar<IsLdrAm3RegOffPredX3, [A57WrBackTwo]>, + SchedVar<NoSchedPred, [A57WrBackOne]> +]>; +def : InstRW<[A57WriteLdrDAm3Pre, A57WriteLdrDAm3Pre, A57WriteLdrDAm3PreWrBack], + (instregex "LDRD_PRE")>; +def : InstRW<[A57Write_4cyc_1L_1I, A57Write_4cyc_1L_1I, A57WrBackOne], + (instregex "t2LDRD_PRE")>; + +// --- LDR post-indexed --- +def : InstRW<[A57Write_4cyc_1L_1I, A57WrBackOne], (instregex "LDR(T?)_POST_IMM", + "LDRB(T?)_POST_IMM", "LDR(SB|H|SH)Ti", "t2LDRB_POST")>; + +def A57WriteLdrAm3PostWrBack : SchedWriteVariant<[ + SchedVar<IsLdrAm3RegOffPred, [A57WrBackTwo]>, + SchedVar<NoSchedPred, [A57WrBackOne]> +]>; +def : InstRW<[A57Write_4cyc_1L_1I, A57WriteLdrAm3PostWrBack], + (instregex "LDR(H|SH|SB)_POST")>; +def : InstRW<[A57Write_4cyc_1L, A57WrBackOne], + (instregex "t2LDR(H|SH|SB)?_POST")>; + +def : InstRW<[A57Write_4cyc_1L_1I, A57WrBackTwo], (instregex "LDR_POST_REG", + "LDRB_POST_REG", "LDR(B?)T_POST$")>; + +def A57WriteLdrTRegPost : SchedWriteVariant<[ + SchedVar<IsLdrAm2ScaledPred, [A57Write_4cyc_1I_1L_1M]>, + SchedVar<NoSchedPred, [A57Write_4cyc_1L_1I]> +]>; +def A57WriteLdrTRegPostWrBack : SchedWriteVariant<[ + SchedVar<IsLdrAm2ScaledPred, [A57WrBackThree]>, + SchedVar<NoSchedPred, [A57WrBackTwo]> +]>; +// 4(3) "I0/I1,L,M" for scaled register, otherwise 4(2) "I0/I1,L" +def : InstRW<[A57WriteLdrTRegPost, A57WriteLdrTRegPostWrBack], + (instregex "LDRT_POST_REG", "LDRBT_POST_REG")>; + +def : InstRW<[A57Write_4cyc_1L_1I, A57WrBackTwo], (instregex "LDR(SB|H|SH)Tr")>; + +def A57WriteLdrAm3PostWrBackX3 : SchedWriteVariant<[ + SchedVar<IsLdrAm3RegOffPredX3, [A57WrBackTwo]>, + SchedVar<NoSchedPred, [A57WrBackOne]> +]>; +// LDRD post-indexed: 4(2) cyc for reg, 4(1) cyc for imm. +def : InstRW<[A57Write_4cyc_1L_1I, A57Write_4cyc_1L_1I, + A57WriteLdrAm3PostWrBackX3], (instregex "LDRD_POST")>; +def : InstRW<[A57Write_4cyc_1L_1I, A57Write_4cyc_1L_1I, A57WrBackOne], + (instregex "t2LDRD_POST")>; + +// --- Preload instructions --- +// Preload, immed offset +def : InstRW<[A57Write_4cyc_1L], (instregex "(t2)?PLDi12", "(t2)?PLDWi12", + "t2PLDW?(i8|pci|s)", "(t2)?PLI")>; + +// Preload, register offset, +// 5cyc "I0/I1,L" for minus reg or scaled not plus lsl2 +// otherwise 4cyc "L" +def A57WritePLD : SchedWriteVariant<[ + SchedVar<IsLdstsoScaledNotOptimalPredX0, [A57Write_5cyc_1I_1L]>, + SchedVar<IsLdstsoMinusRegPredX0, [A57Write_5cyc_1I_1L]>, + SchedVar<NoSchedPred, [A57Write_4cyc_1L]> +]>; +def : InstRW<[A57WritePLD], (instregex "PLDrs", "PLDWrs")>; + +// --- Load multiple instructions --- +foreach NumAddr = 1-8 in { + def A57LMAddrPred#NumAddr : + SchedPredicate<"(TII->getLDMVariableDefsSize(*MI)+1)/2 == "#NumAddr>; +} + +def A57LDMOpsListNoregin : A57WriteLMOpsListType< + [A57Write_3cyc_1L, A57Write_3cyc_1L, + A57Write_4cyc_1L, A57Write_4cyc_1L, + A57Write_5cyc_1L, A57Write_5cyc_1L, + A57Write_6cyc_1L, A57Write_6cyc_1L, + A57Write_7cyc_1L, A57Write_7cyc_1L, + A57Write_8cyc_1L, A57Write_8cyc_1L, + A57Write_9cyc_1L, A57Write_9cyc_1L, + A57Write_10cyc_1L, A57Write_10cyc_1L]>; +def A57WriteLDMnoreginlist : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, A57LDMOpsListNoregin.Writes[0-1]>, + SchedVar<A57LMAddrPred2, A57LDMOpsListNoregin.Writes[0-3]>, + SchedVar<A57LMAddrPred3, A57LDMOpsListNoregin.Writes[0-5]>, + SchedVar<A57LMAddrPred4, A57LDMOpsListNoregin.Writes[0-7]>, + SchedVar<A57LMAddrPred5, A57LDMOpsListNoregin.Writes[0-9]>, + SchedVar<A57LMAddrPred6, A57LDMOpsListNoregin.Writes[0-11]>, + SchedVar<A57LMAddrPred7, A57LDMOpsListNoregin.Writes[0-13]>, + SchedVar<A57LMAddrPred8, A57LDMOpsListNoregin.Writes[0-15]>, + SchedVar<NoSchedPred, A57LDMOpsListNoregin.Writes[0-15]> +]> { let Variadic=1; } + +def A57LDMOpsListRegin : A57WriteLMOpsListType< + [A57Write_4cyc_1L_1I, A57Write_4cyc_1L_1I, + A57Write_5cyc_1L_1I, A57Write_5cyc_1L_1I, + A57Write_6cyc_1L_1I, A57Write_6cyc_1L_1I, + A57Write_7cyc_1L_1I, A57Write_7cyc_1L_1I, + A57Write_8cyc_1L_1I, A57Write_8cyc_1L_1I, + A57Write_9cyc_1L_1I, A57Write_9cyc_1L_1I, + A57Write_10cyc_1L_1I, A57Write_10cyc_1L_1I, + A57Write_11cyc_1L_1I, A57Write_11cyc_1L_1I]>; +def A57WriteLDMreginlist : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, A57LDMOpsListRegin.Writes[0-1]>, + SchedVar<A57LMAddrPred2, A57LDMOpsListRegin.Writes[0-3]>, + SchedVar<A57LMAddrPred3, A57LDMOpsListRegin.Writes[0-5]>, + SchedVar<A57LMAddrPred4, A57LDMOpsListRegin.Writes[0-7]>, + SchedVar<A57LMAddrPred5, A57LDMOpsListRegin.Writes[0-9]>, + SchedVar<A57LMAddrPred6, A57LDMOpsListRegin.Writes[0-11]>, + SchedVar<A57LMAddrPred7, A57LDMOpsListRegin.Writes[0-13]>, + SchedVar<A57LMAddrPred8, A57LDMOpsListRegin.Writes[0-15]>, + SchedVar<NoSchedPred, A57LDMOpsListRegin.Writes[0-15]> +]> { let Variadic=1; } + +def A57LDMOpsList_Upd : A57WriteLMOpsListType< + [A57WrBackOne, + A57Write_3cyc_1L_1I, A57Write_3cyc_1L_1I, + A57Write_4cyc_1L_1I, A57Write_4cyc_1L_1I, + A57Write_5cyc_1L_1I, A57Write_5cyc_1L_1I, + A57Write_6cyc_1L_1I, A57Write_6cyc_1L_1I, + A57Write_7cyc_1L_1I, A57Write_7cyc_1L_1I, + A57Write_8cyc_1L_1I, A57Write_8cyc_1L_1I, + A57Write_9cyc_1L_1I, A57Write_9cyc_1L_1I, + A57Write_10cyc_1L_1I, A57Write_10cyc_1L_1I]>; +def A57WriteLDM_Upd : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, A57LDMOpsList_Upd.Writes[0-2]>, + SchedVar<A57LMAddrPred2, A57LDMOpsList_Upd.Writes[0-4]>, + SchedVar<A57LMAddrPred3, A57LDMOpsList_Upd.Writes[0-6]>, + SchedVar<A57LMAddrPred4, A57LDMOpsList_Upd.Writes[0-8]>, + SchedVar<A57LMAddrPred5, A57LDMOpsList_Upd.Writes[0-10]>, + SchedVar<A57LMAddrPred6, A57LDMOpsList_Upd.Writes[0-12]>, + SchedVar<A57LMAddrPred7, A57LDMOpsList_Upd.Writes[0-14]>, + SchedVar<A57LMAddrPred8, A57LDMOpsList_Upd.Writes[0-16]>, + SchedVar<NoSchedPred, A57LDMOpsList_Upd.Writes[0-16]> +]> { let Variadic=1; } + +def A57WriteLDM : SchedWriteVariant<[ + SchedVar<IsLdmBaseRegInList, [A57WriteLDMreginlist]>, + SchedVar<NoSchedPred, [A57WriteLDMnoreginlist]> +]> { let Variadic=1; } + +def : InstRW<[A57WriteLDM], (instregex "(t|t2|sys)?LDM(IA|DA|DB|IB)$")>; + +// TODO: no writeback latency defined in documentation (implemented as 1 cyc) +def : InstRW<[A57WriteLDM_Upd], + (instregex "(t|t2|sys)?LDM(IA_UPD|DA_UPD|DB_UPD|IB_UPD|IA_RET)", "tPOP")>; + +def : InstRW<[A57Write_5cyc_1L], (instregex "VLLDM")>; + +// --- 3.9 Store Instructions --- + +// Store, immed offset +def : InstRW<[A57Write_1cyc_1S], (instregex "STRi12", "STRBi12", "PICSTR", + "t2STR(B?)(T|i12|i8|s)", "t2STRDi8", "t2STRH(i12|i8|s)", "tSTR")>; + +// Store, register offset +// For minus or for not plus lsl2 scaled we need 3cyc "I0/I1, S", +// otherwise 1cyc S. +def A57WriteStrAmLDSTSO : SchedWriteVariant<[ + SchedVar<IsLdstsoScaledNotOptimalPred, [A57Write_3cyc_1I_1S]>, + SchedVar<IsLdstsoMinusRegPred, [A57Write_3cyc_1I_1S]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1S]> +]>; +def : InstRW<[A57WriteStrAmLDSTSO], (instregex "STRrs", "STRBrs")>; + +// STRH,STRD: 3cyc "I0/I1, S" for minus reg, 1cyc S for imm or for plus reg. +def A57WriteStrAm3 : SchedWriteVariant<[ + SchedVar<IsLdrAm3NegRegOffPred, [A57Write_3cyc_1I_1S]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1S]> +]>; +def : InstRW<[A57WriteStrAm3], (instregex "STRH$")>; +def A57WriteStrAm3X2 : SchedWriteVariant<[ + SchedVar<IsLdrAm3NegRegOffPredX2, [A57Write_3cyc_1I_1S]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1S]> +]>; +def : InstRW<[A57WriteStrAm3X2], (instregex "STRD$")>; + +// Store, immed pre-indexed (1cyc "S, I0/I1", 1cyc writeback) +def : InstRW<[A57WrBackOne, A57Write_1cyc_1S_1I], (instregex "STR_PRE_IMM", + "STRB_PRE_IMM", "STR(B)?(r|i)_preidx", "(t2)?STRH_(preidx|PRE)", + "t2STR(B?)_(PRE|preidx)", "t2STRD_PRE")>; + +// Store, register pre-indexed: +// 1(1) "S, I0/I1" for plus reg +// 3(2) "I0/I1, S" for minus reg +// 1(2) "S, M" for scaled plus lsl2 +// 3(2) "I0/I1, S" for other scaled +def A57WriteStrAmLDSTSOPre : SchedWriteVariant<[ + SchedVar<IsLdstsoScaledNotOptimalPredX2, [A57Write_3cyc_1I_1S]>, + SchedVar<IsLdstsoMinusRegPredX2, [A57Write_3cyc_1I_1S]>, + SchedVar<IsLdstsoScaledPredX2, [A57Write_1cyc_1S_1M]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1S_1I]> +]>; +def A57WriteStrAmLDSTSOPreWrBack : SchedWriteVariant<[ + SchedVar<IsLdstsoScaledPredX2, [A57WrBackTwo]>, + SchedVar<IsLdstsoMinusRegPredX2, [A57WrBackTwo]>, + SchedVar<NoSchedPred, [A57WrBackOne]> +]>; +def : InstRW<[A57WriteStrAmLDSTSOPreWrBack, A57WriteStrAmLDSTSOPre], + (instregex "STR_PRE_REG", "STRB_PRE_REG")>; + +// pre-indexed STRH/STRD (STRH_PRE, STRD_PRE) +// 1(1) "S, I0/I1" for imm or reg plus +// 3(2) "I0/I1, S" for reg minus +def A57WriteStrAm3PreX2 : SchedWriteVariant<[ + SchedVar<IsLdrAm3NegRegOffPredX2, [A57Write_3cyc_1I_1S]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1S_1I]> +]>; +def A57WriteStrAm3PreWrBackX2 : SchedWriteVariant<[ + SchedVar<IsLdrAm3NegRegOffPredX2, [A57WrBackTwo]>, + SchedVar<NoSchedPred, [A57WrBackOne]> +]>; +def : InstRW<[A57WriteStrAm3PreWrBackX2, A57WriteStrAm3PreX2], + (instregex "STRH_PRE")>; + +def A57WriteStrAm3PreX3 : SchedWriteVariant<[ + SchedVar<IsLdrAm3NegRegOffPredX3, [A57Write_3cyc_1I_1S]>, + SchedVar<NoSchedPred, [A57Write_1cyc_1S_1I]> +]>; +def A57WriteStrAm3PreWrBackX3 : SchedWriteVariant<[ + SchedVar<IsLdrAm3NegRegOffPredX3, [A57WrBackTwo]>, + SchedVar<NoSchedPred, [A57WrBackOne]> +]>; +def : InstRW<[A57WriteStrAm3PreWrBackX3, A57WriteStrAm3PreX3], + (instregex "STRD_PRE")>; + +def : InstRW<[A57WrBackOne, A57Write_1cyc_1S_1I], (instregex "STR(T?)_POST_IMM", + "STRB(T?)_POST_IMM", "t2STR(B?)_POST")>; + +// 1(2) "S, M" for STR/STRB register post-indexed (both scaled or not) +def : InstRW<[A57WrBackTwo, A57Write_1cyc_1S_1M], (instregex "STR(T?)_POST_REG", + "STRB(T?)_POST_REG", "STR(B?)T_POST$")>; + +// post-indexed STRH/STRD(STRH_POST, STRD_POST), STRHTi, STRHTr +// 1(1) "S, I0/I1" both for reg or imm +def : InstRW<[A57WrBackOne, A57Write_1cyc_1S_1I], + (instregex "(t2)?STR(H|D)_POST", "STRHT(i|r)", "t2STRHT")>; + +// --- Store multiple instructions --- +// TODO: no writeback latency defined in documentation +def A57WriteSTM : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, [A57Write_1cyc_1S]>, + SchedVar<A57LMAddrPred2, [A57Write_2cyc_1S]>, + SchedVar<A57LMAddrPred3, [A57Write_3cyc_1S]>, + SchedVar<A57LMAddrPred4, [A57Write_4cyc_1S]>, + SchedVar<A57LMAddrPred5, [A57Write_5cyc_1S]>, + SchedVar<A57LMAddrPred6, [A57Write_6cyc_1S]>, + SchedVar<A57LMAddrPred7, [A57Write_7cyc_1S]>, + SchedVar<A57LMAddrPred8, [A57Write_8cyc_1S]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1S]> +]>; +def A57WriteSTM_Upd : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, [A57Write_1cyc_1S_1I]>, + SchedVar<A57LMAddrPred2, [A57Write_2cyc_1S_1I]>, + SchedVar<A57LMAddrPred3, [A57Write_3cyc_1S_1I]>, + SchedVar<A57LMAddrPred4, [A57Write_4cyc_1S_1I]>, + SchedVar<A57LMAddrPred5, [A57Write_5cyc_1S_1I]>, + SchedVar<A57LMAddrPred6, [A57Write_6cyc_1S_1I]>, + SchedVar<A57LMAddrPred7, [A57Write_7cyc_1S_1I]>, + SchedVar<A57LMAddrPred8, [A57Write_8cyc_1S_1I]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1S_1I]> +]>; + +def : InstRW<[A57WriteSTM], (instregex "(t2|sys|t)?STM(IA|DA|DB|IB)$")>; +def : InstRW<[A57WrBackOne, A57WriteSTM_Upd], + (instregex "(t2|sys|t)?STM(IA_UPD|DA_UPD|DB_UPD|IB_UPD)", "tPUSH")>; + +def : InstRW<[A57Write_5cyc_1S], (instregex "VLSTM")>; + +// --- 3.10 FP Data Processing Instructions --- +def : SchedAlias<WriteFPALU32, A57Write_5cyc_1V>; +def : SchedAlias<WriteFPALU64, A57Write_5cyc_1V>; + +def : InstRW<[A57Write_3cyc_1V], (instregex "VABS(S|D|H)")>; + +// fp compare - 3cyc F1 for unconditional, 6cyc "F0/F1, F1" for conditional +def A57WriteVcmp : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57Write_6cyc_1V_1X]>, + SchedVar<NoSchedPred, [A57Write_3cyc_1X]> +]>; +def : InstRW<[A57WriteVcmp], + (instregex "VCMP(D|S|H|ZD|ZS|ZH)$", "VCMPE(D|S|H|ZD|ZS|ZH)")>; + +// fp convert +def : InstRW<[A57Write_5cyc_1V], (instregex + "VCVT(A|N|P|M)(SH|UH|SS|US|SD|UD)", "VCVT(BDH|THD|TDH)")>; +def : InstRW<[A57Write_5cyc_1V], (instregex "VTOSLS", "VTOUHS", "VTOULS")>; +def : SchedAlias<WriteFPCVT, A57Write_5cyc_1V>; + +def : InstRW<[A57Write_5cyc_1V], (instregex "VJCVT")>; + +// FP round to integral +def : InstRW<[A57Write_5cyc_1V], (instregex "VRINT(A|N|P|M|Z|R|X)(H|S|D)$")>; + +// FP divide, FP square root +def : SchedAlias<WriteFPDIV32, A57Write_17cyc_1W>; +def : SchedAlias<WriteFPDIV64, A57Write_32cyc_1W>; +def : SchedAlias<WriteFPSQRT32, A57Write_17cyc_1W>; +def : SchedAlias<WriteFPSQRT64, A57Write_32cyc_1W>; + +def : InstRW<[A57Write_17cyc_1W], (instregex "VSQRTH")>; + +// FP max/min +def : InstRW<[A57Write_5cyc_1V], (instregex "VMAX", "VMIN")>; + +// FP multiply-accumulate pipelines support late forwarding of the result +// from FP multiply μops to the accumulate operands of an +// FP multiply-accumulate μop. The latter can potentially be issued 1 cycle +// after the FP multiply μop has been issued +// FP multiply, FZ +def A57WriteVMUL : SchedWriteRes<[A57UnitV]> { let Latency = 5; } + +def : SchedAlias<WriteFPMUL32, A57WriteVMUL>; +def : SchedAlias<WriteFPMUL64, A57WriteVMUL>; +def : ReadAdvance<ReadFPMUL, 0>; + +// FP multiply accumulate, FZ: 9cyc "F0/F1" or 4 cyc for sequenced accumulate +// VFMA, VFMS, VFNMA, VFNMS, VMLA, VMLS, VNMLA, VNMLS +def A57WriteVFMA : SchedWriteRes<[A57UnitV]> { let Latency = 9; } + +// VFMA takes 9 cyc for common case and 4 cyc for VFMA->VFMA chain (5 read adv.) +// VMUL takes 5 cyc for common case and 1 cyc for VMUL->VFMA chain (4 read adv.) +// Currently, there is no way to define different read advances for VFMA operand +// from VFMA or from VMUL, so there will be 5 read advance. +// Zero latency (instead of one) for VMUL->VFMA shouldn't break something. +// The same situation with ASIMD VMUL/VFMA instructions +// def A57ReadVFMA : SchedRead; +// def : ReadAdvance<A57ReadVFMA, 5, [A57WriteVFMA]>; +// def : ReadAdvance<A57ReadVFMA, 4, [A57WriteVMUL]>; +def A57ReadVFMA5 : SchedReadAdvance<5, [A57WriteVFMA, A57WriteVMUL]>; + +def : SchedAlias<WriteFPMAC32, A57WriteVFMA>; +def : SchedAlias<WriteFPMAC64, A57WriteVFMA>; +def : SchedAlias<ReadFPMAC, A57ReadVFMA5>; + +// VMLAH/VMLSH are not binded to scheduling classes by default, so here custom: +def : InstRW<[A57WriteVFMA, A57ReadVFMA5, ReadFPMUL, ReadFPMUL], + (instregex "VMLAH", "VMLSH", "VNMLAH", "VNMLSH")>; + +def : InstRW<[A57WriteVMUL], + (instregex "VUDOTD", "VSDOTD", "VUDOTQ", "VSDOTQ")>; + +def : InstRW<[A57Write_3cyc_1V], (instregex "VNEG")>; +def : InstRW<[A57Write_3cyc_1V], (instregex "VSEL")>; + +// --- 3.11 FP Miscellaneous Instructions --- +// VMOV: 3cyc "F0/F1" for imm/reg +def : InstRW<[A57Write_3cyc_1V], (instregex "FCONST(D|S|H)")>; +def : InstRW<[A57Write_3cyc_1V], (instregex "VMOV(D|S|H)(cc)?$")>; + +def : InstRW<[A57Write_3cyc_1V], (instregex "VINSH")>; + +// 5cyc L for FP transfer, vfp to core reg, +// 5cyc L for FP transfer, core reg to vfp +def : SchedAlias<WriteFPMOV, A57Write_5cyc_1L>; +// VMOVRRS/VMOVRRD in common code declared with one WriteFPMOV (instead of 2). +def : InstRW<[A57Write_5cyc_1L, A57Write_5cyc_1L], (instregex "VMOV(RRS|RRD)")>; + +// 8cyc "L,F0/F1" for FP transfer, core reg to upper or lower half of vfp D-reg +def : InstRW<[A57Write_8cyc_1L_1I], (instregex "VMOVDRR")>; + +// --- 3.12 FP Load Instructions --- +def : InstRW<[A57Write_5cyc_1L], (instregex "VLDR(D|S|H)")>; + +def : InstRW<[A57Write_5cyc_1L], (instregex "VLDMQIA$")>; + +// FP load multiple (VLDM) + +def A57VLDMOpsListUncond : A57WriteLMOpsListType< + [A57Write_5cyc_1L, A57Write_5cyc_1L, + A57Write_6cyc_1L, A57Write_6cyc_1L, + A57Write_7cyc_1L, A57Write_7cyc_1L, + A57Write_8cyc_1L, A57Write_8cyc_1L, + A57Write_9cyc_1L, A57Write_9cyc_1L, + A57Write_10cyc_1L, A57Write_10cyc_1L, + A57Write_11cyc_1L, A57Write_11cyc_1L, + A57Write_12cyc_1L, A57Write_12cyc_1L]>; +def A57WriteVLDMuncond : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, A57VLDMOpsListUncond.Writes[0-1]>, + SchedVar<A57LMAddrPred2, A57VLDMOpsListUncond.Writes[0-3]>, + SchedVar<A57LMAddrPred3, A57VLDMOpsListUncond.Writes[0-5]>, + SchedVar<A57LMAddrPred4, A57VLDMOpsListUncond.Writes[0-7]>, + SchedVar<A57LMAddrPred5, A57VLDMOpsListUncond.Writes[0-9]>, + SchedVar<A57LMAddrPred6, A57VLDMOpsListUncond.Writes[0-11]>, + SchedVar<A57LMAddrPred7, A57VLDMOpsListUncond.Writes[0-13]>, + SchedVar<A57LMAddrPred8, A57VLDMOpsListUncond.Writes[0-15]>, + SchedVar<NoSchedPred, A57VLDMOpsListUncond.Writes[0-15]> +]> { let Variadic=1; } + +def A57VLDMOpsListCond : A57WriteLMOpsListType< + [A57Write_5cyc_1L, A57Write_6cyc_1L, + A57Write_7cyc_1L, A57Write_8cyc_1L, + A57Write_9cyc_1L, A57Write_10cyc_1L, + A57Write_11cyc_1L, A57Write_12cyc_1L, + A57Write_13cyc_1L, A57Write_14cyc_1L, + A57Write_15cyc_1L, A57Write_16cyc_1L, + A57Write_17cyc_1L, A57Write_18cyc_1L, + A57Write_19cyc_1L, A57Write_20cyc_1L]>; +def A57WriteVLDMcond : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, A57VLDMOpsListCond.Writes[0-1]>, + SchedVar<A57LMAddrPred2, A57VLDMOpsListCond.Writes[0-3]>, + SchedVar<A57LMAddrPred3, A57VLDMOpsListCond.Writes[0-5]>, + SchedVar<A57LMAddrPred4, A57VLDMOpsListCond.Writes[0-7]>, + SchedVar<A57LMAddrPred5, A57VLDMOpsListCond.Writes[0-9]>, + SchedVar<A57LMAddrPred6, A57VLDMOpsListCond.Writes[0-11]>, + SchedVar<A57LMAddrPred7, A57VLDMOpsListCond.Writes[0-13]>, + SchedVar<A57LMAddrPred8, A57VLDMOpsListCond.Writes[0-15]>, + SchedVar<NoSchedPred, A57VLDMOpsListCond.Writes[0-15]> +]> { let Variadic=1; } + +def A57WriteVLDM : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57WriteVLDMcond]>, + SchedVar<NoSchedPred, [A57WriteVLDMuncond]> +]> { let Variadic=1; } + +def : InstRW<[A57WriteVLDM], (instregex "VLDM(DIA|SIA)$")>; + +def A57VLDMOpsListUncond_Upd : A57WriteLMOpsListType< + [A57Write_5cyc_1L_1I, A57Write_5cyc_1L_1I, + A57Write_6cyc_1L_1I, A57Write_6cyc_1L_1I, + A57Write_7cyc_1L_1I, A57Write_7cyc_1L_1I, + A57Write_8cyc_1L_1I, A57Write_8cyc_1L_1I, + A57Write_9cyc_1L_1I, A57Write_9cyc_1L_1I, + A57Write_10cyc_1L_1I, A57Write_10cyc_1L_1I, + A57Write_11cyc_1L_1I, A57Write_11cyc_1L_1I, + A57Write_12cyc_1L_1I, A57Write_12cyc_1L_1I]>; +def A57WriteVLDMuncond_UPD : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, A57VLDMOpsListUncond_Upd.Writes[0-1]>, + SchedVar<A57LMAddrPred2, A57VLDMOpsListUncond_Upd.Writes[0-3]>, + SchedVar<A57LMAddrPred3, A57VLDMOpsListUncond_Upd.Writes[0-5]>, + SchedVar<A57LMAddrPred4, A57VLDMOpsListUncond_Upd.Writes[0-7]>, + SchedVar<A57LMAddrPred5, A57VLDMOpsListUncond_Upd.Writes[0-9]>, + SchedVar<A57LMAddrPred6, A57VLDMOpsListUncond_Upd.Writes[0-11]>, + SchedVar<A57LMAddrPred7, A57VLDMOpsListUncond_Upd.Writes[0-13]>, + SchedVar<A57LMAddrPred8, A57VLDMOpsListUncond_Upd.Writes[0-15]>, + SchedVar<NoSchedPred, A57VLDMOpsListUncond_Upd.Writes[0-15]> +]> { let Variadic=1; } + +def A57VLDMOpsListCond_Upd : A57WriteLMOpsListType< + [A57Write_5cyc_1L_1I, A57Write_6cyc_1L_1I, + A57Write_7cyc_1L_1I, A57Write_8cyc_1L_1I, + A57Write_9cyc_1L_1I, A57Write_10cyc_1L_1I, + A57Write_11cyc_1L_1I, A57Write_12cyc_1L_1I, + A57Write_13cyc_1L_1I, A57Write_14cyc_1L_1I, + A57Write_15cyc_1L_1I, A57Write_16cyc_1L_1I, + A57Write_17cyc_1L_1I, A57Write_18cyc_1L_1I, + A57Write_19cyc_1L_1I, A57Write_20cyc_1L_1I]>; +def A57WriteVLDMcond_UPD : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, A57VLDMOpsListCond_Upd.Writes[0-1]>, + SchedVar<A57LMAddrPred2, A57VLDMOpsListCond_Upd.Writes[0-3]>, + SchedVar<A57LMAddrPred3, A57VLDMOpsListCond_Upd.Writes[0-5]>, + SchedVar<A57LMAddrPred4, A57VLDMOpsListCond_Upd.Writes[0-7]>, + SchedVar<A57LMAddrPred5, A57VLDMOpsListCond_Upd.Writes[0-9]>, + SchedVar<A57LMAddrPred6, A57VLDMOpsListCond_Upd.Writes[0-11]>, + SchedVar<A57LMAddrPred7, A57VLDMOpsListCond_Upd.Writes[0-13]>, + SchedVar<A57LMAddrPred8, A57VLDMOpsListCond_Upd.Writes[0-15]>, + SchedVar<NoSchedPred, A57VLDMOpsListCond_Upd.Writes[0-15]> +]> { let Variadic=1; } + +def A57WriteVLDM_UPD : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [A57WriteVLDMcond_UPD]>, + SchedVar<NoSchedPred, [A57WriteVLDMuncond_UPD]> +]> { let Variadic=1; } + +def : InstRW<[A57WrBackOne, A57WriteVLDM_UPD], + (instregex "VLDM(DIA_UPD|DDB_UPD|SIA_UPD|SDB_UPD)")>; + +// --- 3.13 FP Store Instructions --- +def : InstRW<[A57Write_1cyc_1S], (instregex "VSTR(D|S|H)")>; + +def : InstRW<[A57Write_2cyc_1S], (instregex "VSTMQIA$")>; + +def A57WriteVSTMs : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, [A57Write_1cyc_1S]>, + SchedVar<A57LMAddrPred2, [A57Write_2cyc_1S]>, + SchedVar<A57LMAddrPred3, [A57Write_3cyc_1S]>, + SchedVar<A57LMAddrPred4, [A57Write_4cyc_1S]>, + SchedVar<A57LMAddrPred5, [A57Write_5cyc_1S]>, + SchedVar<A57LMAddrPred6, [A57Write_6cyc_1S]>, + SchedVar<A57LMAddrPred7, [A57Write_7cyc_1S]>, + SchedVar<A57LMAddrPred8, [A57Write_8cyc_1S]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1S]> +]>; +def A57WriteVSTMd : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, [A57Write_2cyc_1S]>, + SchedVar<A57LMAddrPred2, [A57Write_4cyc_1S]>, + SchedVar<A57LMAddrPred3, [A57Write_6cyc_1S]>, + SchedVar<A57LMAddrPred4, [A57Write_8cyc_1S]>, + SchedVar<A57LMAddrPred5, [A57Write_10cyc_1S]>, + SchedVar<A57LMAddrPred6, [A57Write_12cyc_1S]>, + SchedVar<A57LMAddrPred7, [A57Write_14cyc_1S]>, + SchedVar<A57LMAddrPred8, [A57Write_16cyc_1S]>, + SchedVar<NoSchedPred, [A57Write_4cyc_1S]> +]>; +def A57WriteVSTMs_Upd : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, [A57Write_1cyc_1S_1I]>, + SchedVar<A57LMAddrPred2, [A57Write_2cyc_1S_1I]>, + SchedVar<A57LMAddrPred3, [A57Write_3cyc_1S_1I]>, + SchedVar<A57LMAddrPred4, [A57Write_4cyc_1S_1I]>, + SchedVar<A57LMAddrPred5, [A57Write_5cyc_1S_1I]>, + SchedVar<A57LMAddrPred6, [A57Write_6cyc_1S_1I]>, + SchedVar<A57LMAddrPred7, [A57Write_7cyc_1S_1I]>, + SchedVar<A57LMAddrPred8, [A57Write_8cyc_1S_1I]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1S_1I]> +]>; +def A57WriteVSTMd_Upd : SchedWriteVariant<[ + SchedVar<A57LMAddrPred1, [A57Write_2cyc_1S_1I]>, + SchedVar<A57LMAddrPred2, [A57Write_4cyc_1S_1I]>, + SchedVar<A57LMAddrPred3, [A57Write_6cyc_1S_1I]>, + SchedVar<A57LMAddrPred4, [A57Write_8cyc_1S_1I]>, + SchedVar<A57LMAddrPred5, [A57Write_10cyc_1S_1I]>, + SchedVar<A57LMAddrPred6, [A57Write_12cyc_1S_1I]>, + SchedVar<A57LMAddrPred7, [A57Write_14cyc_1S_1I]>, + SchedVar<A57LMAddrPred8, [A57Write_16cyc_1S_1I]>, + SchedVar<NoSchedPred, [A57Write_2cyc_1S_1I]> +]>; + +def : InstRW<[A57WriteVSTMs], (instregex "VSTMSIA$")>; +def : InstRW<[A57WriteVSTMd], (instregex "VSTMDIA$")>; +def : InstRW<[A57WrBackOne, A57WriteVSTMs_Upd], + (instregex "VSTM(SIA_UPD|SDB_UPD)")>; +def : InstRW<[A57WrBackOne, A57WriteVSTMd_Upd], + (instregex "VSTM(DIA_UPD|DDB_UPD)")>; + +// --- 3.14 ASIMD Integer Instructions --- + +// ASIMD absolute diff, 3cyc F0/F1 for integer VABD +def : InstRW<[A57Write_3cyc_1V], (instregex "VABD(s|u)")>; + +// ASIMD absolute diff accum: 4(1) F1 for D-form, 5(2) F1 for Q-form +def A57WriteVABAD : SchedWriteRes<[A57UnitX]> { let Latency = 4; } +def A57ReadVABAD : SchedReadAdvance<3, [A57WriteVABAD]>; +def : InstRW<[A57WriteVABAD, A57ReadVABAD], + (instregex "VABA(s|u)(v8i8|v4i16|v2i32)")>; +def A57WriteVABAQ : SchedWriteRes<[A57UnitX]> { let Latency = 5; } +def A57ReadVABAQ : SchedReadAdvance<3, [A57WriteVABAQ]>; +def : InstRW<[A57WriteVABAQ, A57ReadVABAQ], + (instregex "VABA(s|u)(v16i8|v8i16|v4i32)")>; + +// ASIMD absolute diff accum long: 4(1) F1 for VABAL +def A57WriteVABAL : SchedWriteRes<[A57UnitX]> { let Latency = 4; } +def A57ReadVABAL : SchedReadAdvance<3, [A57WriteVABAL]>; +def : InstRW<[A57WriteVABAL, A57ReadVABAL], (instregex "VABAL(s|u)")>; + +// ASIMD absolute diff long: 3cyc F0/F1 for VABDL +def : InstRW<[A57Write_3cyc_1V], (instregex "VABDL(s|u)")>; + +// ASIMD arith, basic +def : InstRW<[A57Write_3cyc_1V], (instregex "VADDv", "VADDL", "VADDW", + "VNEG(s8d|s16d|s32d|s8q|s16q|s32q|d|q)", + "VPADDi", "VPADDL", "VSUBv", "VSUBL", "VSUBW")>; + +// ASIMD arith, complex +def : InstRW<[A57Write_3cyc_1V], (instregex "VABS", "VADDHN", "VHADD", "VHSUB", + "VQABS", "VQADD", "VQNEG", "VQSUB", + "VRADDHN", "VRHADD", "VRSUBHN", "VSUBHN")>; + +// ASIMD compare +def : InstRW<[A57Write_3cyc_1V], + (instregex "VCEQ", "VCGE", "VCGT", "VCLE", "VTST", "VCLT")>; + +// ASIMD logical +def : InstRW<[A57Write_3cyc_1V], + (instregex "VAND", "VBIC", "VMVN", "VORR", "VORN", "VEOR")>; + +// ASIMD max/min +def : InstRW<[A57Write_3cyc_1V], + (instregex "(VMAX|VMIN)(s|u)", "(VPMAX|VPMIN)(s8|s16|s32|u8|u16|u32)")>; + +// ASIMD multiply, D-form: 5cyc F0 for r0px, 4cyc F0 for r1p0 and later +// Cortex-A57 r1p0 and later reduce the latency of ASIMD multiply +// and multiply-with-accumulate instructions relative to r0pX. +def A57WriteVMULD_VecInt : SchedWriteVariant<[ + SchedVar<IsR1P0AndLaterPred, [A57Write_4cyc_1W]>, + SchedVar<NoSchedPred, [A57Write_5cyc_1W]>]>; +def : InstRW<[A57WriteVMULD_VecInt], (instregex + "VMUL(v8i8|v4i16|v2i32|pd)", "VMULsl(v4i16|v2i32)", + "VQDMULH(sl)?(v4i16|v2i32)", "VQRDMULH(sl)?(v4i16|v2i32)")>; + +// ASIMD multiply, Q-form: 6cyc F0 for r0px, 5cyc F0 for r1p0 and later +def A57WriteVMULQ_VecInt : SchedWriteVariant<[ + SchedVar<IsR1P0AndLaterPred, [A57Write_5cyc_1W]>, + SchedVar<NoSchedPred, [A57Write_6cyc_1W]>]>; +def : InstRW<[A57WriteVMULQ_VecInt], (instregex + "VMUL(v16i8|v8i16|v4i32|pq)", "VMULsl(v8i16|v4i32)", + "VQDMULH(sl)?(v8i16|v4i32)", "VQRDMULH(sl)?(v8i16|v4i32)")>; + +// ASIMD multiply accumulate, D-form +// 5cyc F0 for r0px, 4cyc F0 for r1p0 and later, 1cyc for accumulate sequence +// (4 or 3 ReadAdvance) +def A57WriteVMLAD_VecInt : SchedWriteVariant<[ + SchedVar<IsR1P0AndLaterPred, [A57Write_4cyc_1W]>, + SchedVar<NoSchedPred, [A57Write_5cyc_1W]>]>; +def A57ReadVMLAD_VecInt : SchedReadVariant<[ + SchedVar<IsR1P0AndLaterPred, [SchedReadAdvance<3, [A57WriteVMLAD_VecInt]>]>, + SchedVar<NoSchedPred, [SchedReadAdvance<4, [A57WriteVMLAD_VecInt]>]> +]>; +def : InstRW<[A57WriteVMLAD_VecInt, A57ReadVMLAD_VecInt], + (instregex "VMLA(sl)?(v8i8|v4i16|v2i32)", "VMLS(sl)?(v8i8|v4i16|v2i32)")>; + +// ASIMD multiply accumulate, Q-form +// 6cyc F0 for r0px, 5cyc F0 for r1p0 and later, 2cyc for accumulate sequence +// (4 or 3 ReadAdvance) +def A57WriteVMLAQ_VecInt : SchedWriteVariant<[ + SchedVar<IsR1P0AndLaterPred, [A57Write_5cyc_1W]>, + SchedVar<NoSchedPred, [A57Write_6cyc_1W]>]>; +def A57ReadVMLAQ_VecInt : SchedReadVariant<[ + SchedVar<IsR1P0AndLaterPred, [SchedReadAdvance<3, [A57WriteVMLAQ_VecInt]>]>, + SchedVar<NoSchedPred, [SchedReadAdvance<4, [A57WriteVMLAQ_VecInt]>]> +]>; +def : InstRW<[A57WriteVMLAQ_VecInt, A57ReadVMLAQ_VecInt], + (instregex "VMLA(sl)?(v16i8|v8i16|v4i32)", "VMLS(sl)?(v16i8|v8i16|v4i32)")>; + +// ASIMD multiply accumulate long +// 5cyc F0 for r0px, 4cyc F0 for r1p0 and later, 1cyc for accumulate sequence +// (4 or 3 ReadAdvance) +def A57WriteVMLAL_VecInt : SchedWriteVariant<[ + SchedVar<IsR1P0AndLaterPred, [A57Write_4cyc_1W]>, + SchedVar<NoSchedPred, [A57Write_5cyc_1W]>]>; +def A57ReadVMLAL_VecInt : SchedReadVariant<[ + SchedVar<IsR1P0AndLaterPred, [SchedReadAdvance<3, [A57WriteVMLAL_VecInt]>]>, + SchedVar<NoSchedPred, [SchedReadAdvance<4, [A57WriteVMLAL_VecInt]>]> +]>; +def : InstRW<[A57WriteVMLAL_VecInt, A57ReadVMLAL_VecInt], + (instregex "VMLAL(s|u)", "VMLSL(s|u)")>; + +// ASIMD multiply accumulate saturating long +// 5cyc F0 for r0px, 4cyc F0 for r1p0 and later, 2cyc for accumulate sequence +// (3 or 2 ReadAdvance) +def A57WriteVQDMLAL_VecInt : SchedWriteVariant<[ + SchedVar<IsR1P0AndLaterPred, [A57Write_4cyc_1W]>, + SchedVar<NoSchedPred, [A57Write_5cyc_1W]>]>; +def A57ReadVQDMLAL_VecInt : SchedReadVariant<[ + SchedVar<IsR1P0AndLaterPred, [SchedReadAdvance<2, [A57WriteVQDMLAL_VecInt]>]>, + SchedVar<NoSchedPred, [SchedReadAdvance<3, [A57WriteVQDMLAL_VecInt]>]> +]>; +def : InstRW<[A57WriteVQDMLAL_VecInt, A57ReadVQDMLAL_VecInt], + (instregex "VQDMLAL", "VQDMLSL")>; + +// Vector Saturating Rounding Doubling Multiply Accumulate/Subtract Long +// Scheduling info from VQDMLAL/VQDMLSL +def : InstRW<[A57WriteVQDMLAL_VecInt, A57ReadVQDMLAL_VecInt], + (instregex "VQRDMLAH", "VQRDMLSH")>; + +// ASIMD multiply long +// 5cyc F0 for r0px, 4cyc F0 for r1p0 and later +def A57WriteVMULL_VecInt : SchedWriteVariant<[ + SchedVar<IsR1P0AndLaterPred, [A57Write_4cyc_1W]>, + SchedVar<NoSchedPred, [A57Write_5cyc_1W]>]>; +def : InstRW<[A57WriteVMULL_VecInt], + (instregex "VMULL(s|u|p8|sls|slu)", "VQDMULL")>; + +// ASIMD pairwise add and accumulate +// 4cyc F1, 1cyc for accumulate sequence (3cyc ReadAdvance) +def A57WriteVPADAL : SchedWriteRes<[A57UnitX]> { let Latency = 4; } +def A57ReadVPADAL : SchedReadAdvance<3, [A57WriteVPADAL]>; +def : InstRW<[A57WriteVPADAL, A57ReadVPADAL], (instregex "VPADAL(s|u)")>; + +// ASIMD shift accumulate +// 4cyc F1, 1cyc for accumulate sequence (3cyc ReadAdvance) +def A57WriteVSRA : SchedWriteRes<[A57UnitX]> { let Latency = 4; } +def A57ReadVSRA : SchedReadAdvance<3, [A57WriteVSRA]>; +def : InstRW<[A57WriteVSRA, A57ReadVSRA], (instregex "VSRA", "VRSRA")>; + +// ASIMD shift by immed, basic +def : InstRW<[A57Write_3cyc_1X], + (instregex "VMOVL", "VSHLi", "VSHLL", "VSHR(s|u)", "VSHRN")>; + +// ASIMD shift by immed, complex +def : InstRW<[A57Write_4cyc_1X], (instregex + "VQRSHRN", "VQRSHRUN", "VQSHL(si|ui|su)", "VQSHRN", "VQSHRUN", "VRSHR(s|u)", + "VRSHRN")>; + +// ASIMD shift by immed and insert, basic, D-form +def : InstRW<[A57Write_4cyc_1X], (instregex + "VSLI(v8i8|v4i16|v2i32|v1i64)", "VSRI(v8i8|v4i16|v2i32|v1i64)")>; + +// ASIMD shift by immed and insert, basic, Q-form +def : InstRW<[A57Write_5cyc_1X], (instregex + "VSLI(v16i8|v8i16|v4i32|v2i64)", "VSRI(v16i8|v8i16|v4i32|v2i64)")>; + +// ASIMD shift by register, basic, D-form +def : InstRW<[A57Write_3cyc_1X], (instregex + "VSHL(s|u)(v8i8|v4i16|v2i32|v1i64)")>; + +// ASIMD shift by register, basic, Q-form +def : InstRW<[A57Write_4cyc_1X], (instregex + "VSHL(s|u)(v16i8|v8i16|v4i32|v2i64)")>; + +// ASIMD shift by register, complex, D-form +// VQRSHL, VQSHL, VRSHL +def : InstRW<[A57Write_4cyc_1X], (instregex + "VQRSHL(s|u)(v8i8|v4i16|v2i32|v1i64)", "VQSHL(s|u)(v8i8|v4i16|v2i32|v1i64)", + "VRSHL(s|u)(v8i8|v4i16|v2i32|v1i64)")>; + +// ASIMD shift by register, complex, Q-form +def : InstRW<[A57Write_5cyc_1X], (instregex + "VQRSHL(s|u)(v16i8|v8i16|v4i32|v2i64)", "VQSHL(s|u)(v16i8|v8i16|v4i32|v2i64)", + "VRSHL(s|u)(v16i8|v8i16|v4i32|v2i64)")>; + +// --- 3.15 ASIMD Floating-Point Instructions --- +// ASIMD FP absolute value +def : InstRW<[A57Write_3cyc_1V], (instregex "VABS(fd|fq|hd|hq)")>; + +// ASIMD FP arith +def : InstRW<[A57Write_5cyc_1V], (instregex "VABD(fd|fq|hd|hq)", + "VADD(fd|fq|hd|hq)", "VPADD(f|h)", "VSUB(fd|fq|hd|hq)")>; + +def : InstRW<[A57Write_5cyc_1V], (instregex "VCADD", "VCMLA")>; + +// ASIMD FP compare +def : InstRW<[A57Write_5cyc_1V], (instregex "VAC(GE|GT|LE|LT)", + "VC(EQ|GE|GT|LE)(fd|fq|hd|hq)")>; + +// ASIMD FP convert, integer +def : InstRW<[A57Write_5cyc_1V], (instregex + "VCVT(f2sd|f2ud|s2fd|u2fd|f2sq|f2uq|s2fq|u2fq|f2xsd|f2xud|xs2fd|xu2fd)", + "VCVT(f2xsq|f2xuq|xs2fq|xu2fq)", + "VCVT(AN|MN|NN|PN)(SDf|SQf|UDf|UQf|SDh|SQh|UDh|UQh)")>; + +// ASIMD FP convert, half-precision: 8cyc F0/F1 +def : InstRW<[A57Write_8cyc_1V], (instregex + "VCVT(h2sd|h2ud|s2hd|u2hd|h2sq|h2uq|s2hq|u2hq|h2xsd|h2xud|xs2hd|xu2hd)", + "VCVT(h2xsq|h2xuq|xs2hq|xu2hq)", + "VCVT(f2h|h2f)")>; + +// ASIMD FP max/min +def : InstRW<[A57Write_5cyc_1V], (instregex + "(VMAX|VMIN)(fd|fq|hd|hq)", "(VPMAX|VPMIN)(f|h)", "VMAXNM", "VMINNM")>; + +// ASIMD FP multiply +def A57WriteVMUL_VecFP : SchedWriteRes<[A57UnitV]> { let Latency = 5; } +def : InstRW<[A57WriteVMUL_VecFP], (instregex "VMUL(sl)?(fd|fq|hd|hq)")>; + +// ASIMD FP multiply accumulate: 9cyc F0/F1, 4cyc for accumulate sequence +def A57WriteVMLA_VecFP : SchedWriteRes<[A57UnitV]> { let Latency = 9; } +def A57ReadVMLA_VecFP : + SchedReadAdvance<5, [A57WriteVMLA_VecFP, A57WriteVMUL_VecFP]>; +def : InstRW<[A57WriteVMLA_VecFP, A57ReadVMLA_VecFP], + (instregex "(VMLA|VMLS)(sl)?(fd|fq|hd|hq)", "(VFMA|VFMS)(fd|fq|hd|hq)")>; + +// ASIMD FP negate +def : InstRW<[A57Write_3cyc_1V], (instregex "VNEG(fd|f32q|hd|hq)")>; + +// ASIMD FP round to integral +def : InstRW<[A57Write_5cyc_1V], (instregex + "VRINT(AN|MN|NN|PN|XN|ZN)(Df|Qf|Dh|Qh)")>; + +// --- 3.16 ASIMD Miscellaneous Instructions --- + +// ASIMD bitwise insert +def : InstRW<[A57Write_3cyc_1V], (instregex "VBIF", "VBIT", "VBSL")>; + +// ASIMD count +def : InstRW<[A57Write_3cyc_1V], (instregex "VCLS", "VCLZ", "VCNT")>; + +// ASIMD duplicate, core reg: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V], (instregex "VDUP(8|16|32)(d|q)")>; + +// ASIMD duplicate, scalar: 3cyc "F0/F1" +def : InstRW<[A57Write_3cyc_1V], (instregex "VDUPLN(8|16|32)(d|q)")>; + +// ASIMD extract +def : InstRW<[A57Write_3cyc_1V], (instregex "VEXT(d|q)(8|16|32|64)")>; + +// ASIMD move, immed +def : InstRW<[A57Write_3cyc_1V], (instregex + "VMOV(v8i8|v16i8|v4i16|v8i16|v2i32|v4i32|v1i64|v2i64|v2f32|v4f32)", + "VMOVD0", "VMOVQ0")>; + +// ASIMD move, narrowing +def : InstRW<[A57Write_3cyc_1V], (instregex "VMOVN")>; + +// ASIMD move, saturating +def : InstRW<[A57Write_4cyc_1X], (instregex "VQMOVN")>; + +// ASIMD reciprocal estimate +def : InstRW<[A57Write_5cyc_1V], (instregex "VRECPE", "VRSQRTE")>; + +// ASIMD reciprocal step, FZ +def : InstRW<[A57Write_9cyc_1V], (instregex "VRECPS", "VRSQRTS")>; + +// ASIMD reverse, swap, table lookup (1-2 reg) +def : InstRW<[A57Write_3cyc_1V], (instregex "VREV", "VSWP", "VTB(L|X)(1|2)")>; + +// ASIMD table lookup (3-4 reg) +def : InstRW<[A57Write_6cyc_1V], (instregex "VTBL(3|4)", "VTBX(3|4)")>; + +// ASIMD transfer, scalar to core reg: 6cyc "L, I0/I1" +def : InstRW<[A57Write_6cyc_1L_1I], (instregex "VGETLN")>; + +// ASIMD transfer, core reg to scalar: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V], (instregex "VSETLN")>; + +// ASIMD transpose +def : InstRW<[A57Write_3cyc_1V, A57Write_3cyc_1V], (instregex "VTRN")>; + +// ASIMD unzip/zip, D-form +def : InstRW<[A57Write_3cyc_1V, A57Write_3cyc_1V], + (instregex "VUZPd", "VZIPd")>; + +// ASIMD unzip/zip, Q-form +def : InstRW<[A57Write_6cyc_1V, A57Write_6cyc_1V], + (instregex "VUZPq", "VZIPq")>; + +// --- 3.17 ASIMD Load Instructions --- + +// Overriden via InstRW for this processor. +def : WriteRes<WriteVLD1, []>; +def : WriteRes<WriteVLD2, []>; +def : WriteRes<WriteVLD3, []>; +def : WriteRes<WriteVLD4, []>; +def : WriteRes<WriteVST1, []>; +def : WriteRes<WriteVST2, []>; +def : WriteRes<WriteVST3, []>; +def : WriteRes<WriteVST4, []>; + +// 1-2 reg: 5cyc L, +I for writeback, 1 cyc wb latency +def : InstRW<[A57Write_5cyc_1L], (instregex "VLD1(d|q)(8|16|32|64)$")>; +def : InstRW<[A57Write_5cyc_1L_1I, A57WrBackOne], + (instregex "VLD1(d|q)(8|16|32|64)wb")>; + +// 3-4 reg: 6cyc L, +I for writeback, 1 cyc wb latency +def : InstRW<[A57Write_6cyc_1L], + (instregex "VLD1(d|q)(8|16|32|64)(T|Q)$", "VLD1d64(T|Q)Pseudo")>; + +def : InstRW<[A57Write_6cyc_1L_1I, A57WrBackOne], + (instregex "VLD1(d|q)(8|16|32|64)(T|Q)wb")>; + +// ASIMD load, 1 element, one lane and all lanes: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V], (instregex + "VLD1(LN|DUP)(d|q)(8|16|32)$", "VLD1(LN|DUP)(d|q)(8|16|32)Pseudo$")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57WrBackOne], (instregex + "VLD1(LN|DUP)(d|q)(8|16|32)(wb|_UPD)", "VLD1LNq(8|16|32)Pseudo_UPD")>; + +// ASIMD load, 2 element, multiple, 2 reg: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V], + (instregex "VLD2(d|q)(8|16|32)$", "VLD2q(8|16|32)Pseudo$")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD2(d|q)(8|16|32)wb", "VLD2q(8|16|32)PseudoWB")>; + +// ASIMD load, 2 element, multiple, 4 reg: 9cyc "L, F0/F1" +def : InstRW<[A57Write_9cyc_1L_1V], (instregex "VLD2b(8|16|32)$")>; +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD2b(8|16|32)wb")>; + +// ASIMD load, 2 element, one lane and all lanes: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V], + (instregex "VLD2(DUP|LN)(d|q)(8|16|32|8x2|16x2|32x2)$", + "VLD2LN(d|q)(8|16|32)Pseudo$")>; +// 2 results + wb result +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57Write_8cyc_1L_1V, A57WrBackOne], + (instregex "VLD2LN(d|q)(8|16|32)_UPD$")>; +// 1 result + wb result +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD2DUPd(8|16|32|8x2|16x2|32x2)wb", + "VLD2LN(d|q)(8|16|32)Pseudo_UPD")>; + +// ASIMD load, 3 element, multiple, 3 reg: 9cyc "L, F0/F1" +// 3 results +def : InstRW<[A57Write_9cyc_1L_1V, A57Write_9cyc_1L_1V, A57Write_9cyc_1L_1V], + (instregex "VLD3(d|q)(8|16|32)$")>; +// 1 result +def : InstRW<[A57Write_9cyc_1L_1V], + (instregex "VLD3(d|q)(8|16|32)(oddP|P)seudo$")>; +// 3 results + wb +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57Write_9cyc_1L_1V_1I, + A57Write_9cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD3(d|q)(8|16|32)_UPD$")>; +// 1 result + wb +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD3(d|q)(8|16|32)(oddP|P)seudo_UPD")>; + +// ASIMD load, 3 element, one lane, size 32: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V], + (instregex "VLD3LN(d|q)32$", + "VLD3LN(d|q)32Pseudo$")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57Write_8cyc_1L_1V_1I, + A57Write_8cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD3LN(d|q)32_UPD")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD3LN(d|q)32Pseudo_UPD")>; + +// ASIMD load, 3 element, one lane, size 8/16: 9cyc "L, F0/F1" +def : InstRW<[A57Write_9cyc_1L_1V, A57Write_9cyc_1L_1V, A57Write_9cyc_1L_1V], + (instregex "VLD3LN(d|q)(8|16)$", + "VLD3LN(d|q)(8|16)Pseudo$")>; +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57Write_9cyc_1L_1V_1I, + A57Write_9cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD3LN(d|q)(8|16)_UPD")>; +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD3LN(d|q)(8|16)Pseudo_UPD")>; + +// ASIMD load, 3 element, all lanes: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V], + (instregex "VLD3DUP(d|q)(8|16|32)$", + "VLD3DUP(d|q)(8|16|32)Pseudo$")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57Write_8cyc_1L_1V_1I, + A57Write_8cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD3DUP(d|q)(8|16|32)_UPD")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD3DUP(d|q)(8|16|32)Pseudo_UPD")>; + +// ASIMD load, 4 element, multiple, 4 reg: 9cyc "L, F0/F1" +def : InstRW<[A57Write_9cyc_1L_1V, A57Write_9cyc_1L_1V, A57Write_9cyc_1L_1V, + A57Write_9cyc_1L_1V], + (instregex "VLD4(d|q)(8|16|32)$")>; +def : InstRW<[A57Write_9cyc_1L_1V], + (instregex "VLD4(d|q)(8|16|32)(oddP|P)seudo$")>; +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57Write_9cyc_1L_1V_1I, + A57Write_9cyc_1L_1V_1I, A57Write_9cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD4(d|q)(8|16|32)_UPD")>; +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD4(d|q)(8|16|32)(oddP|P)seudo_UPD")>; + +// ASIMD load, 4 element, one lane, size 32: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V, + A57Write_8cyc_1L_1V], + (instregex "VLD4LN(d|q)32$", + "VLD4LN(d|q)32Pseudo$")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57Write_8cyc_1L_1V_1I, + A57Write_8cyc_1L_1V_1I, A57Write_8cyc_1L_1V_1I, + A57WrBackOne], + (instregex "VLD4LN(d|q)32_UPD")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD4LN(d|q)32Pseudo_UPD")>; + +// ASIMD load, 4 element, one lane, size 8/16: 9cyc "L, F0/F1" +def : InstRW<[A57Write_9cyc_1L_1V, A57Write_9cyc_1L_1V, A57Write_9cyc_1L_1V, + A57Write_9cyc_1L_1V], + (instregex "VLD4LN(d|q)(8|16)$", + "VLD4LN(d|q)(8|16)Pseudo$")>; +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57Write_9cyc_1L_1V_1I, + A57Write_9cyc_1L_1V_1I, A57Write_9cyc_1L_1V_1I, + A57WrBackOne], + (instregex "VLD4LN(d|q)(8|16)_UPD")>; +def : InstRW<[A57Write_9cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD4LN(d|q)(8|16)Pseudo_UPD")>; + +// ASIMD load, 4 element, all lanes: 8cyc "L, F0/F1" +def : InstRW<[A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V, A57Write_8cyc_1L_1V, + A57Write_8cyc_1L_1V], + (instregex "VLD4DUP(d|q)(8|16|32)$", + "VLD4DUP(d|q)(8|16|32)Pseudo$")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57Write_8cyc_1L_1V_1I, + A57Write_8cyc_1L_1V_1I, A57Write_8cyc_1L_1V_1I, + A57WrBackOne], + (instregex "VLD4DUP(d|q)(8|16|32)_UPD")>; +def : InstRW<[A57Write_8cyc_1L_1V_1I, A57WrBackOne], + (instregex "VLD4DUP(d|q)(8|16|32)Pseudo_UPD")>; + +// --- 3.18 ASIMD Store Instructions --- + +// ASIMD store, 1 element, multiple, 1 reg: 1cyc S +def : InstRW<[A57Write_1cyc_1S], (instregex "VST1d(8|16|32|64)$")>; +def : InstRW<[A57WrBackOne, A57Write_1cyc_1S_1I], + (instregex "VST1d(8|16|32|64)wb")>; +// ASIMD store, 1 element, multiple, 2 reg: 2cyc S +def : InstRW<[A57Write_2cyc_1S], (instregex "VST1q(8|16|32|64)$")>; +def : InstRW<[A57WrBackOne, A57Write_2cyc_1S_1I], + (instregex "VST1q(8|16|32|64)wb")>; +// ASIMD store, 1 element, multiple, 3 reg: 3cyc S +def : InstRW<[A57Write_3cyc_1S], + (instregex "VST1d(8|16|32|64)T$", "VST1d64TPseudo$")>; +def : InstRW<[A57WrBackOne, A57Write_3cyc_1S_1I], + (instregex "VST1d(8|16|32|64)Twb", "VST1d64TPseudoWB")>; +// ASIMD store, 1 element, multiple, 4 reg: 4cyc S +def : InstRW<[A57Write_4cyc_1S], + (instregex "VST1d(8|16|32|64)(Q|QPseudo)$")>; +def : InstRW<[A57WrBackOne, A57Write_4cyc_1S_1I], + (instregex "VST1d(8|16|32|64)(Qwb|QPseudoWB)")>; +// ASIMD store, 1 element, one lane: 3cyc "F0/F1, S" +def : InstRW<[A57Write_3cyc_1S_1V], + (instregex "VST1LNd(8|16|32)$", "VST1LNq(8|16|32)Pseudo$")>; +def : InstRW<[A57WrBackOne, A57Write_3cyc_1S_1V_1I], + (instregex "VST1LNd(8|16|32)_UPD", "VST1LNq(8|16|32)Pseudo_UPD")>; +// ASIMD store, 2 element, multiple, 2 reg: 3cyc "F0/F1, S" +def : InstRW<[A57Write_3cyc_1S_1V], + (instregex "VST2(d|b)(8|16|32)$")>; +def : InstRW<[A57WrBackOne, A57Write_3cyc_1S_1V_1I], + (instregex "VST2(b|d)(8|16|32)wb")>; +// ASIMD store, 2 element, multiple, 4 reg: 4cyc "F0/F1, S" +def : InstRW<[A57Write_4cyc_1S_1V], + (instregex "VST2q(8|16|32)$", "VST2q(8|16|32)Pseudo$")>; +def : InstRW<[A57WrBackOne, A57Write_4cyc_1S_1V_1I], + (instregex "VST2q(8|16|32)wb", "VST2q(8|16|32)PseudoWB")>; +// ASIMD store, 2 element, one lane: 3cyc "F0/F1, S" +def : InstRW<[A57Write_3cyc_1S_1V], + (instregex "VST2LN(d|q)(8|16|32)$", "VST2LN(d|q)(8|16|32)Pseudo$")>; +def : InstRW<[A57WrBackOne, A57Write_3cyc_1S_1V_1I], + (instregex "VST2LN(d|q)(8|16|32)_UPD", + "VST2LN(d|q)(8|16|32)Pseudo_UPD")>; +// ASIMD store, 3 element, multiple, 3 reg +def : InstRW<[A57Write_3cyc_1S_1V], + (instregex "VST3(d|q)(8|16|32)$", "VST3(d|q)(8|16|32)(oddP|P)seudo$")>; +def : InstRW<[A57WrBackOne, A57Write_3cyc_1S_1V_1I], + (instregex "VST3(d|q)(8|16|32)_UPD", + "VST3(d|q)(8|16|32)(oddP|P)seudo_UPD$")>; +// ASIMD store, 3 element, one lane +def : InstRW<[A57Write_3cyc_1S_1V], + (instregex "VST3LN(d|q)(8|16|32)$", "VST3LN(d|q)(8|16|32)Pseudo$")>; +def : InstRW<[A57WrBackOne, A57Write_3cyc_1S_1V_1I], + (instregex "VST3LN(d|q)(8|16|32)_UPD", + "VST3LN(d|q)(8|16|32)Pseudo_UPD")>; +// ASIMD store, 4 element, multiple, 4 reg +def : InstRW<[A57Write_4cyc_1S_1V], + (instregex "VST4(d|q)(8|16|32)$", "VST4(d|q)(8|16|32)(oddP|P)seudo$")>; +def : InstRW<[A57WrBackOne, A57Write_4cyc_1S_1V_1I], + (instregex "VST4(d|q)(8|16|32)_UPD", + "VST4(d|q)(8|16|32)(oddP|P)seudo_UPD$")>; +// ASIMD store, 4 element, one lane +def : InstRW<[A57Write_3cyc_1S_1V], + (instregex "VST4LN(d|q)(8|16|32)$", "VST4LN(d|q)(8|16|32)Pseudo$")>; +def : InstRW<[A57WrBackOne, A57Write_3cyc_1S_1V_1I], + (instregex "VST4LN(d|q)(8|16|32)_UPD", + "VST4LN(d|q)(8|16|32)Pseudo_UPD")>; + +// --- 3.19 Cryptography Extensions --- +// Crypto AES ops +// AESD, AESE, AESIMC, AESMC: 3cyc F0 +def : InstRW<[A57Write_3cyc_1W], (instregex "^AES")>; +// Crypto polynomial (64x64) multiply long (VMULL.P64): 3cyc F0 +def : InstRW<[A57Write_3cyc_1W], (instregex "^VMULLp64")>; +// Crypto SHA1 xor ops: 6cyc F0/F1 +def : InstRW<[A57Write_6cyc_2V], (instregex "^SHA1SU0")>; +// Crypto SHA1 fast ops: 3cyc F0 +def : InstRW<[A57Write_3cyc_1W], (instregex "^SHA1(H|SU1)")>; +// Crypto SHA1 slow ops: 6cyc F0 +def : InstRW<[A57Write_6cyc_2W], (instregex "^SHA1[CMP]")>; +// Crypto SHA256 fast ops: 3cyc F0 +def : InstRW<[A57Write_3cyc_1W], (instregex "^SHA256SU0")>; +// Crypto SHA256 slow ops: 6cyc F0 +def : InstRW<[A57Write_6cyc_2W], (instregex "^SHA256(H|H2|SU1)")>; + +// --- 3.20 CRC --- +def : InstRW<[A57Write_3cyc_1W], (instregex "^(t2)?CRC32")>; + +// ----------------------------------------------------------------------------- +// Common definitions +def : WriteRes<WriteNoop, []> { let Latency = 0; let NumMicroOps = 0; } +def : SchedAlias<WriteALU, A57Write_1cyc_1I>; + +def : SchedAlias<WriteBr, A57Write_1cyc_1B>; +def : SchedAlias<WriteBrL, A57Write_1cyc_1B_1I>; +def : SchedAlias<WriteBrTbl, A57Write_1cyc_1B_1I>; +def : SchedAlias<WritePreLd, A57Write_4cyc_1L>; + +def : SchedAlias<WriteLd, A57Write_4cyc_1L>; +def : SchedAlias<WriteST, A57Write_1cyc_1S>; +def : ReadAdvance<ReadALU, 0>; + +} // SchedModel = CortexA57Model + diff --git a/capstone/suite/synctools/tablegen/ARM/ARMScheduleA57WriteRes.td b/capstone/suite/synctools/tablegen/ARM/ARMScheduleA57WriteRes.td new file mode 100644 index 000000000..670717dc7 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMScheduleA57WriteRes.td @@ -0,0 +1,323 @@ +//=- ARMScheduleA57WriteRes.td - ARM Cortex-A57 Write Res ---*- tablegen -*-=// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// Contains all of the Cortex-A57 specific SchedWriteRes types. The approach +// below is to define a generic SchedWriteRes for every combination of +// latency and microOps. The naming conventions is to use a prefix, one field +// for latency, and one or more microOp count/type designators. +// Prefix: A57Write +// Latency: #cyc +// MicroOp Count/Types: #(B|I|M|L|S|X|W|V) +// +// e.g. A57Write_6cyc_1I_6S_4V means the total latency is 6 and there are +// 11 micro-ops to be issued as follows: one to I pipe, six to S pipes and +// four to V pipes. +// +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// Define Generic 1 micro-op types + +def A57Write_5cyc_1M : SchedWriteRes<[A57UnitM]> { let Latency = 5; } +def A57Write_5cyc_1V : SchedWriteRes<[A57UnitV]> { let Latency = 5; } +def A57Write_5cyc_1W : SchedWriteRes<[A57UnitW]> { let Latency = 5; } +def A57Write_10cyc_1V : SchedWriteRes<[A57UnitV]> { let Latency = 10; } +def A57Write_17cyc_1W : SchedWriteRes<[A57UnitW]> { let Latency = 17; + let ResourceCycles = [17]; } +def A57Write_18cyc_1X : SchedWriteRes<[A57UnitX]> { let Latency = 18; + let ResourceCycles = [18]; } +def A57Write_19cyc_1M : SchedWriteRes<[A57UnitM]> { let Latency = 19; + let ResourceCycles = [19]; } +def A57Write_20cyc_1M : SchedWriteRes<[A57UnitM]> { let Latency = 20; + let ResourceCycles = [20]; } +def A57Write_1cyc_1B : SchedWriteRes<[A57UnitB]> { let Latency = 1; } +def A57Write_1cyc_1I : SchedWriteRes<[A57UnitI]> { let Latency = 1; } +def A57Write_2cyc_1I : SchedWriteRes<[A57UnitI]> { let Latency = 2; } +def A57Write_3cyc_1I : SchedWriteRes<[A57UnitI]> { let Latency = 3; } +def A57Write_1cyc_1S : SchedWriteRes<[A57UnitS]> { let Latency = 1; } +def A57Write_2cyc_1S : SchedWriteRes<[A57UnitS]> { let Latency = 2; } +def A57Write_3cyc_1S : SchedWriteRes<[A57UnitS]> { let Latency = 3; } +def A57Write_2cyc_1M : SchedWriteRes<[A57UnitM]> { let Latency = 2; } +def A57Write_32cyc_1W : SchedWriteRes<[A57UnitW]> { let Latency = 32; + let ResourceCycles = [32]; } +def A57Write_32cyc_1X : SchedWriteRes<[A57UnitX]> { let Latency = 32; + let ResourceCycles = [32]; } +def A57Write_35cyc_1M : SchedWriteRes<[A57UnitM]> { let Latency = 35; + let ResourceCycles = [35]; } +def A57Write_3cyc_1M : SchedWriteRes<[A57UnitM]> { let Latency = 3; } +def A57Write_3cyc_1V : SchedWriteRes<[A57UnitV]> { let Latency = 3; } +def A57Write_3cyc_1W : SchedWriteRes<[A57UnitW]> { let Latency = 3; } +def A57Write_3cyc_1X : SchedWriteRes<[A57UnitX]> { let Latency = 3; } + +// A57Write_3cyc_1L - A57Write_20cyc_1L +foreach Lat = 3-20 in { + def A57Write_#Lat#cyc_1L : SchedWriteRes<[A57UnitL]> { + let Latency = Lat; + } +} + +// A57Write_4cyc_1S - A57Write_16cyc_1S +foreach Lat = 4-16 in { + def A57Write_#Lat#cyc_1S : SchedWriteRes<[A57UnitS]> { + let Latency = Lat; + } +} + +def A57Write_4cyc_1M : SchedWriteRes<[A57UnitL]> { let Latency = 4; } +def A57Write_4cyc_1X : SchedWriteRes<[A57UnitX]> { let Latency = 4; } +def A57Write_4cyc_1W : SchedWriteRes<[A57UnitW]> { let Latency = 4; } +def A57Write_5cyc_1X : SchedWriteRes<[A57UnitX]> { let Latency = 5; } +def A57Write_6cyc_1X : SchedWriteRes<[A57UnitX]> { let Latency = 6; } +def A57Write_6cyc_1W : SchedWriteRes<[A57UnitW]> { let Latency = 6; } +def A57Write_8cyc_1V : SchedWriteRes<[A57UnitV]> { let Latency = 8; } +def A57Write_9cyc_1V : SchedWriteRes<[A57UnitV]> { let Latency = 9; } +def A57Write_6cyc_1M : SchedWriteRes<[A57UnitM]> { let Latency = 6; } +def A57Write_6cyc_1V : SchedWriteRes<[A57UnitV]> { let Latency = 6; } + + +//===----------------------------------------------------------------------===// +// Define Generic 2 micro-op types + +def A57Write_64cyc_2X : SchedWriteRes<[A57UnitX, A57UnitX]> { + let Latency = 64; + let NumMicroOps = 2; + let ResourceCycles = [32, 32]; +} +def A57Write_6cyc_1I_1L : SchedWriteRes<[A57UnitI, + A57UnitL]> { + let Latency = 6; + let NumMicroOps = 2; +} +def A57Write_6cyc_1V_1X : SchedWriteRes<[A57UnitV, + A57UnitX]> { + let Latency = 6; + let NumMicroOps = 2; +} +def A57Write_7cyc_1V_1X : SchedWriteRes<[A57UnitV, + A57UnitX]> { + let Latency = 7; + let NumMicroOps = 2; +} +def A57Write_8cyc_1L_1V : SchedWriteRes<[A57UnitL, + A57UnitV]> { + let Latency = 8; + let NumMicroOps = 2; +} +def A57Write_9cyc_1L_1V : SchedWriteRes<[A57UnitL, + A57UnitV]> { + let Latency = 9; + let NumMicroOps = 2; +} +def A57Write_9cyc_2V : SchedWriteRes<[A57UnitV, A57UnitV]> { + let Latency = 9; + let NumMicroOps = 2; +} +def A57Write_8cyc_2X : SchedWriteRes<[A57UnitX, A57UnitX]> { + let Latency = 8; + let NumMicroOps = 2; +} +def A57Write_6cyc_2L : SchedWriteRes<[A57UnitL, A57UnitL]> { + let Latency = 6; + let NumMicroOps = 2; +} +def A57Write_6cyc_2V : SchedWriteRes<[A57UnitV, A57UnitV]> { + let Latency = 6; + let NumMicroOps = 2; +} +def A57Write_6cyc_2W : SchedWriteRes<[A57UnitW, A57UnitW]> { + let Latency = 6; + let NumMicroOps = 2; +} +def A57Write_5cyc_1I_1L : SchedWriteRes<[A57UnitI, + A57UnitL]> { + let Latency = 5; + let NumMicroOps = 2; +} +def A57Write_5cyc_1I_1M : SchedWriteRes<[A57UnitI, + A57UnitM]> { + let Latency = 5; + let NumMicroOps = 2; +} +def A57Write_5cyc_2V : SchedWriteRes<[A57UnitV, A57UnitV]> { + let Latency = 5; + let NumMicroOps = 2; +} +def A57Write_5cyc_2X : SchedWriteRes<[A57UnitX, A57UnitX]> { + let Latency = 5; + let NumMicroOps = 2; +} +def A57Write_10cyc_1L_1V : SchedWriteRes<[A57UnitL, + A57UnitV]> { + let Latency = 10; + let NumMicroOps = 2; +} +def A57Write_10cyc_2V : SchedWriteRes<[A57UnitV, A57UnitV]> { + let Latency = 10; + let NumMicroOps = 2; +} +def A57Write_1cyc_1B_1I : SchedWriteRes<[A57UnitB, + A57UnitI]> { + let Latency = 1; + let NumMicroOps = 2; +} +def A57Write_1cyc_1I_1S : SchedWriteRes<[A57UnitI, + A57UnitS]> { + let Latency = 1; + let NumMicroOps = 2; +} +def A57Write_1cyc_1S_1I : SchedWriteRes<[A57UnitS, + A57UnitI]> { + let Latency = 1; + let NumMicroOps = 2; +} +def A57Write_2cyc_1S_1I : SchedWriteRes<[A57UnitS, + A57UnitI]> { + let Latency = 2; + let NumMicroOps = 2; +} +def A57Write_3cyc_1S_1I : SchedWriteRes<[A57UnitS, + A57UnitI]> { + let Latency = 3; + let NumMicroOps = 2; +} +def A57Write_1cyc_1S_1M : SchedWriteRes<[A57UnitS, + A57UnitM]> { + let Latency = 1; + let NumMicroOps = 2; +} +def A57Write_2cyc_1B_1I : SchedWriteRes<[A57UnitB, + A57UnitI]> { + let Latency = 2; + let NumMicroOps = 2; +} +def A57Write_3cyc_1B_1I : SchedWriteRes<[A57UnitB, + A57UnitI]> { + let Latency = 3; + let NumMicroOps = 2; +} +def A57Write_6cyc_1B_1L : SchedWriteRes<[A57UnitB, + A57UnitI]> { + let Latency = 6; + let NumMicroOps = 2; +} +def A57Write_2cyc_1I_1M : SchedWriteRes<[A57UnitI, + A57UnitM]> { + let Latency = 2; + let NumMicroOps = 2; +} +def A57Write_2cyc_2S : SchedWriteRes<[A57UnitS, A57UnitS]> { + let Latency = 2; + let NumMicroOps = 2; +} +def A57Write_2cyc_2V : SchedWriteRes<[A57UnitV, A57UnitV]> { + let Latency = 2; + let NumMicroOps = 2; +} +def A57Write_36cyc_2X : SchedWriteRes<[A57UnitX, A57UnitX]> { + let Latency = 36; + let NumMicroOps = 2; + let ResourceCycles = [18, 18]; +} +def A57Write_3cyc_1I_1M : SchedWriteRes<[A57UnitI, + A57UnitM]> { + let Latency = 3; + let NumMicroOps = 2; +} +def A57Write_4cyc_1I_1M : SchedWriteRes<[A57UnitI, + A57UnitM]> { + let Latency = 4; + let NumMicroOps = 2; +} + +// A57Write_3cyc_1L_1I - A57Write_20cyc_1L_1I +foreach Lat = 3-20 in { + def A57Write_#Lat#cyc_1L_1I : SchedWriteRes<[A57UnitL, A57UnitI]> { + let Latency = Lat; let NumMicroOps = 2; + } +} + +def A57Write_3cyc_1I_1S : SchedWriteRes<[A57UnitI, + A57UnitS]> { + let Latency = 3; + let NumMicroOps = 2; +} +def A57Write_3cyc_1S_1V : SchedWriteRes<[A57UnitS, + A57UnitV]> { + let Latency = 3; + let NumMicroOps = 2; +} +def A57Write_4cyc_1S_1V : SchedWriteRes<[A57UnitS, + A57UnitV]> { + let Latency = 4; + let NumMicroOps = 2; +} +def A57Write_3cyc_2V : SchedWriteRes<[A57UnitV, A57UnitV]> { + let Latency = 3; + let NumMicroOps = 2; +} + +// A57Write_4cyc_1S_1I - A57Write_16cyc_1S_1I +foreach Lat = 4-16 in { + def A57Write_#Lat#cyc_1S_1I : SchedWriteRes<[A57UnitS, A57UnitI]> { + let Latency = Lat; let NumMicroOps = 2; + } +} + +def A57Write_4cyc_2X : SchedWriteRes<[A57UnitX, A57UnitX]> { + let Latency = 4; + let NumMicroOps = 2; +} + + +//===----------------------------------------------------------------------===// +// Define Generic 3 micro-op types + +def A57Write_10cyc_3V : SchedWriteRes<[A57UnitV, A57UnitV, A57UnitV]> { + let Latency = 10; + let NumMicroOps = 3; +} +def A57Write_2cyc_1I_2S : SchedWriteRes<[A57UnitI, + A57UnitS, A57UnitS]> { + let Latency = 2; + let NumMicroOps = 3; +} +def A57Write_3cyc_1I_1S_1V : SchedWriteRes<[A57UnitI, + A57UnitS, + A57UnitV]> { + let Latency = 3; + let NumMicroOps = 3; +} +def A57Write_3cyc_1S_1V_1I : SchedWriteRes<[A57UnitS, + A57UnitV, + A57UnitI]> { + let Latency = 3; + let NumMicroOps = 3; +} +def A57Write_4cyc_1S_1V_1I : SchedWriteRes<[A57UnitS, + A57UnitV, + A57UnitI]> { + let Latency = 4; + let NumMicroOps = 3; +} +def A57Write_4cyc_1I_1L_1M : SchedWriteRes<[A57UnitI, A57UnitL, A57UnitM]> { + let Latency = 4; + let NumMicroOps = 3; +} +def A57Write_8cyc_1L_1V_1I : SchedWriteRes<[A57UnitL, + A57UnitV, + A57UnitI]> { + let Latency = 8; + let NumMicroOps = 3; +} +def A57Write_9cyc_1L_1V_1I : SchedWriteRes<[A57UnitL, + A57UnitV, + A57UnitI]> { + let Latency = 9; + let NumMicroOps = 3; +} diff --git a/capstone/suite/synctools/tablegen/ARM/ARMScheduleA8.td b/capstone/suite/synctools/tablegen/ARM/ARMScheduleA8.td new file mode 100644 index 000000000..ba380cba1 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMScheduleA8.td @@ -0,0 +1,1075 @@ +//=- ARMScheduleA8.td - ARM Cortex-A8 Scheduling Definitions -*- tablegen -*-=// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the itinerary class data for the ARM Cortex A8 processors. +// +//===----------------------------------------------------------------------===// + +// +// Scheduling information derived from "Cortex-A8 Technical Reference Manual". +// Functional Units. +def A8_Pipe0 : FuncUnit; // pipeline 0 +def A8_Pipe1 : FuncUnit; // pipeline 1 +def A8_LSPipe : FuncUnit; // Load / store pipeline +def A8_NPipe : FuncUnit; // NEON ALU/MUL pipe +def A8_NLSPipe : FuncUnit; // NEON LS pipe +// +// Dual issue pipeline represented by A8_Pipe0 | A8_Pipe1 +// +def CortexA8Itineraries : ProcessorItineraries< + [A8_Pipe0, A8_Pipe1, A8_LSPipe, A8_NPipe, A8_NLSPipe], + [], [ + // Two fully-pipelined integer ALU pipelines + // + // No operand cycles + InstrItinData<IIC_iALUx , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>]>, + // + // Binary Instructions that produce a result + InstrItinData<IIC_iALUi ,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2]>, + InstrItinData<IIC_iALUr ,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2, 2]>, + InstrItinData<IIC_iALUsi,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2, 1]>, + InstrItinData<IIC_iALUsir,[InstrStage<1,[A8_Pipe0, A8_Pipe1]>], [2, 1, 2]>, + InstrItinData<IIC_iALUsr,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2, 1, 1]>, + // + // Bitwise Instructions that produce a result + InstrItinData<IIC_iBITi ,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2]>, + InstrItinData<IIC_iBITr ,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2, 2]>, + InstrItinData<IIC_iBITsi,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2, 1]>, + InstrItinData<IIC_iBITsr,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2, 1, 1]>, + // + // Unary Instructions that produce a result + InstrItinData<IIC_iUNAr , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2]>, + InstrItinData<IIC_iUNAsi, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 1]>, + // + // Zero and sign extension instructions + InstrItinData<IIC_iEXTr , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1, 1]>, + InstrItinData<IIC_iEXTAr, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2, 1]>, + InstrItinData<IIC_iEXTAsr,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>],[2, 2, 1, 1]>, + // + // Compare instructions + InstrItinData<IIC_iCMPi , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2]>, + InstrItinData<IIC_iCMPr , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2]>, + InstrItinData<IIC_iCMPsi, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 1]>, + InstrItinData<IIC_iCMPsr, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 1, 1]>, + // + // Test instructions + InstrItinData<IIC_iTSTi , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2]>, + InstrItinData<IIC_iTSTr , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2]>, + InstrItinData<IIC_iTSTsi, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 1]>, + InstrItinData<IIC_iTSTsr, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 1, 1]>, + // + // Move instructions, unconditional + InstrItinData<IIC_iMOVi , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1]>, + InstrItinData<IIC_iMOVr , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1, 1]>, + InstrItinData<IIC_iMOVsi, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1, 1]>, + InstrItinData<IIC_iMOVsr, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1, 1, 1]>, + InstrItinData<IIC_iMOVix2,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2]>, + InstrItinData<IIC_iMOVix2addpc,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [3]>, + InstrItinData<IIC_iMOVix2ld,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<1, [A8_LSPipe]>], [5]>, + // + // Move instructions, conditional + InstrItinData<IIC_iCMOVi , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2]>, + InstrItinData<IIC_iCMOVr , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 1]>, + InstrItinData<IIC_iCMOVsi, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 1]>, + InstrItinData<IIC_iCMOVsr, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 1, 1]>, + InstrItinData<IIC_iCMOVix2,[InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [3, 1]>, + // + // MVN instructions + InstrItinData<IIC_iMVNi , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1]>, + InstrItinData<IIC_iMVNr , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1, 1]>, + InstrItinData<IIC_iMVNsi, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1, 1]>, + InstrItinData<IIC_iMVNsr, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [1, 1, 1]>, + + // Integer multiply pipeline + // Result written in E5, but that is relative to the last cycle of multicycle, + // so we use 6 for those cases + // + InstrItinData<IIC_iMUL16 , [InstrStage<1, [A8_Pipe0]>], [5, 1, 1]>, + InstrItinData<IIC_iMAC16 , [InstrStage<2, [A8_Pipe0]>], [6, 1, 1, 4]>, + InstrItinData<IIC_iMUL32 , [InstrStage<2, [A8_Pipe0]>], [6, 1, 1]>, + InstrItinData<IIC_iMAC32 , [InstrStage<2, [A8_Pipe0]>], [6, 1, 1, 4]>, + InstrItinData<IIC_iMUL64 , [InstrStage<3, [A8_Pipe0]>], [6, 6, 1, 1]>, + InstrItinData<IIC_iMAC64 , [InstrStage<3, [A8_Pipe0]>], [6, 6, 1, 1]>, + + // Integer load pipeline + // + // Immediate offset + InstrItinData<IIC_iLoad_i , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1]>, + InstrItinData<IIC_iLoad_bh_i, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1]>, + InstrItinData<IIC_iLoad_d_i, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1]>, + // + // Register offset + InstrItinData<IIC_iLoad_r , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1, 1]>, + InstrItinData<IIC_iLoad_bh_r, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1, 1]>, + InstrItinData<IIC_iLoad_d_r , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1, 1]>, + // + // Scaled register offset, issues over 2 cycles + // FIXME: lsl by 2 takes 1 cycle. + InstrItinData<IIC_iLoad_si , [InstrStage<2, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [4, 1, 1]>, + InstrItinData<IIC_iLoad_bh_si,[InstrStage<2, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [4, 1, 1]>, + // + // Immediate offset with update + InstrItinData<IIC_iLoad_iu , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 2, 1]>, + InstrItinData<IIC_iLoad_bh_iu,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 2, 1]>, + // + // Register offset with update + InstrItinData<IIC_iLoad_ru , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 2, 1, 1]>, + InstrItinData<IIC_iLoad_bh_ru,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 2, 1, 1]>, + InstrItinData<IIC_iLoad_d_ru, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 2, 1, 1]>, + // + // Scaled register offset with update, issues over 2 cycles + InstrItinData<IIC_iLoad_siu , [InstrStage<2, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_LSPipe]>], [4, 3, 1, 1]>, + InstrItinData<IIC_iLoad_bh_siu,[InstrStage<2, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_LSPipe]>], [4, 3, 1, 1]>, + // + // Load multiple, def is the 5th operand. Pipeline 0 only. + // FIXME: A8_LSPipe cycle time is dynamic, this assumes 3 to 4 registers. + InstrItinData<IIC_iLoad_m , [InstrStage<2, [A8_Pipe0], 0>, + InstrStage<2, [A8_LSPipe]>], + [1, 1, 1, 1, 3], [], -1>, // dynamic uops + // + // Load multiple + update, defs are the 1st and 5th operands. + InstrItinData<IIC_iLoad_mu , [InstrStage<3, [A8_Pipe0], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 1, 1, 1, 3], [], -1>, // dynamic uops + // + // Load multiple plus branch + InstrItinData<IIC_iLoad_mBr, [InstrStage<3, [A8_Pipe0], 0>, + InstrStage<3, [A8_LSPipe]>, + InstrStage<1, [A8_Pipe0, A8_Pipe1]>], + [1, 2, 1, 1, 3], [], -1>, // dynamic uops + // + // Pop, def is the 3rd operand. + InstrItinData<IIC_iPop , [InstrStage<3, [A8_Pipe0], 0>, + InstrStage<3, [A8_LSPipe]>], + [1, 1, 3], [], -1>, // dynamic uops + // + // Push, def is the 3th operand. + InstrItinData<IIC_iPop_Br, [InstrStage<3, [A8_Pipe0], 0>, + InstrStage<3, [A8_LSPipe]>, + InstrStage<1, [A8_Pipe0, A8_Pipe1]>], + [1, 1, 3], [], -1>, // dynamic uops + // + // iLoadi + iALUr for t2LDRpci_pic. + InstrItinData<IIC_iLoadiALU, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>, + InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [4, 1]>, + + + // Integer store pipeline + // + // Immediate offset + InstrItinData<IIC_iStore_i , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1]>, + InstrItinData<IIC_iStore_bh_i,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1]>, + InstrItinData<IIC_iStore_d_i, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1]>, + // + // Register offset + InstrItinData<IIC_iStore_r , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1, 1]>, + InstrItinData<IIC_iStore_bh_r,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1, 1]>, + InstrItinData<IIC_iStore_d_r, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [3, 1, 1]>, + // + // Scaled register offset, issues over 2 cycles + InstrItinData<IIC_iStore_si , [InstrStage<2, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_LSPipe]>], [3, 1, 1]>, + InstrItinData<IIC_iStore_bh_si,[InstrStage<2, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_LSPipe]>], [3, 1, 1]>, + // + // Immediate offset with update + InstrItinData<IIC_iStore_iu , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [2, 3, 1]>, + InstrItinData<IIC_iStore_bh_iu,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [2, 3, 1]>, + // + // Register offset with update + InstrItinData<IIC_iStore_ru , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [2, 3, 1, 1]>, + InstrItinData<IIC_iStore_bh_ru,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [2, 3, 1, 1]>, + InstrItinData<IIC_iStore_d_ru, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_LSPipe]>], [2, 3, 1, 1]>, + // + // Scaled register offset with update, issues over 2 cycles + InstrItinData<IIC_iStore_siu, [InstrStage<2, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_LSPipe]>], [3, 3, 1, 1]>, + InstrItinData<IIC_iStore_bh_siu,[InstrStage<2, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_LSPipe]>], [3, 3, 1, 1]>, + // + // Store multiple. Pipeline 0 only. + // FIXME: A8_LSPipe cycle time is dynamic, this assumes 3 to 4 registers. + InstrItinData<IIC_iStore_m , [InstrStage<2, [A8_Pipe0], 0>, + InstrStage<2, [A8_LSPipe]>], + [], [], -1>, // dynamic uops + // + // Store multiple + update + InstrItinData<IIC_iStore_mu, [InstrStage<2, [A8_Pipe0], 0>, + InstrStage<2, [A8_LSPipe]>], + [2], [], -1>, // dynamic uops + // + // Preload + InstrItinData<IIC_Preload, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>], [2, 2]>, + + // Branch + // + // no delay slots, so the latency of a branch is unimportant + InstrItinData<IIC_Br , [InstrStage<1, [A8_Pipe0, A8_Pipe1]>]>, + + // VFP + // Issue through integer pipeline, and execute in NEON unit. We assume + // RunFast mode so that NFP pipeline is used for single-precision when + // possible. + // + // FP Special Register to Integer Register File Move + InstrItinData<IIC_fpSTAT , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>], [20]>, + // + // Single-precision FP Unary + InstrItinData<IIC_fpUNA32 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [7, 1]>, + // + // Double-precision FP Unary + InstrItinData<IIC_fpUNA64 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NPipe], 0>, + InstrStage<4, [A8_NLSPipe]>], [4, 1]>, + // + // Single-precision FP Compare + InstrItinData<IIC_fpCMP32 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [1, 1]>, + // + // Double-precision FP Compare + InstrItinData<IIC_fpCMP64 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NPipe], 0>, + InstrStage<4, [A8_NLSPipe]>], [4, 1]>, + // + // Single to Double FP Convert + InstrItinData<IIC_fpCVTSD , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<7, [A8_NPipe], 0>, + InstrStage<7, [A8_NLSPipe]>], [7, 1]>, + // + // Double to Single FP Convert + InstrItinData<IIC_fpCVTDS , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<5, [A8_NPipe], 0>, + InstrStage<5, [A8_NLSPipe]>], [5, 1]>, + // + // Single-Precision FP to Integer Convert + InstrItinData<IIC_fpCVTSI , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [7, 1]>, + // + // Double-Precision FP to Integer Convert + InstrItinData<IIC_fpCVTDI , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<8, [A8_NPipe], 0>, + InstrStage<8, [A8_NLSPipe]>], [8, 1]>, + // + // Integer to Single-Precision FP Convert + InstrItinData<IIC_fpCVTIS , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [7, 1]>, + // + // Integer to Double-Precision FP Convert + InstrItinData<IIC_fpCVTID , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<8, [A8_NPipe], 0>, + InstrStage<8, [A8_NLSPipe]>], [8, 1]>, + // + // Single-precision FP ALU + InstrItinData<IIC_fpALU32 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [7, 1, 1]>, + // + // Double-precision FP ALU + InstrItinData<IIC_fpALU64 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<9, [A8_NPipe], 0>, + InstrStage<9, [A8_NLSPipe]>], [9, 1, 1]>, + // + // Single-precision FP Multiply + InstrItinData<IIC_fpMUL32 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [7, 1, 1]>, + // + // Double-precision FP Multiply + InstrItinData<IIC_fpMUL64 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<11, [A8_NPipe], 0>, + InstrStage<11, [A8_NLSPipe]>], [11, 1, 1]>, + // + // Single-precision FP MAC + InstrItinData<IIC_fpMAC32 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [7, 2, 1, 1]>, + // + // Double-precision FP MAC + InstrItinData<IIC_fpMAC64 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<19, [A8_NPipe], 0>, + InstrStage<19, [A8_NLSPipe]>], [19, 2, 1, 1]>, + // + // Single-precision Fused FP MAC + InstrItinData<IIC_fpFMAC32, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [7, 2, 1, 1]>, + // + // Double-precision Fused FP MAC + InstrItinData<IIC_fpFMAC64, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<19, [A8_NPipe], 0>, + InstrStage<19, [A8_NLSPipe]>], [19, 2, 1, 1]>, + // + // Single-precision FP DIV + InstrItinData<IIC_fpDIV32 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<20, [A8_NPipe], 0>, + InstrStage<20, [A8_NLSPipe]>], [20, 1, 1]>, + // + // Double-precision FP DIV + InstrItinData<IIC_fpDIV64 , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<29, [A8_NPipe], 0>, + InstrStage<29, [A8_NLSPipe]>], [29, 1, 1]>, + // + // Single-precision FP SQRT + InstrItinData<IIC_fpSQRT32, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<19, [A8_NPipe], 0>, + InstrStage<19, [A8_NLSPipe]>], [19, 1]>, + // + // Double-precision FP SQRT + InstrItinData<IIC_fpSQRT64, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<29, [A8_NPipe], 0>, + InstrStage<29, [A8_NLSPipe]>], [29, 1]>, + + // + // Integer to Single-precision Move + InstrItinData<IIC_fpMOVIS, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], + [2, 1]>, + // + // Integer to Double-precision Move + InstrItinData<IIC_fpMOVID, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], + [2, 1, 1]>, + // + // Single-precision to Integer Move + InstrItinData<IIC_fpMOVSI, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], + [20, 1]>, + // + // Double-precision to Integer Move + InstrItinData<IIC_fpMOVDI, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], + [20, 20, 1]>, + + // + // Single-precision FP Load + InstrItinData<IIC_fpLoad32, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>], + [2, 1]>, + // + // Double-precision FP Load + InstrItinData<IIC_fpLoad64, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>], + [2, 1]>, + // + // FP Load Multiple + // FIXME: A8_LSPipe cycle time is dynamic, this assumes 3 to 4 registers. + InstrItinData<IIC_fpLoad_m, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>], + [1, 1, 1, 2], [], -1>, // dynamic uops + // + // FP Load Multiple + update + InstrItinData<IIC_fpLoad_mu,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>], + [2, 1, 1, 1, 2], [], -1>, // dynamic uops + // + // Single-precision FP Store + InstrItinData<IIC_fpStore32,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>], + [1, 1]>, + // + // Double-precision FP Store + InstrItinData<IIC_fpStore64,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>], + [1, 1]>, + // + // FP Store Multiple + InstrItinData<IIC_fpStore_m,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>], + [1, 1, 1, 1], [], -1>, // dynamic uops + // + // FP Store Multiple + update + InstrItinData<IIC_fpStore_mu,[InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>, + InstrStage<1, [A8_NLSPipe], 0>, + InstrStage<1, [A8_LSPipe]>], + [2, 1, 1, 1, 1], [], -1>, // dynamic uops + // NEON + // Issue through integer pipeline, and execute in NEON unit. + // + // VLD1 + InstrItinData<IIC_VLD1, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 1]>, + // VLD1x2 + InstrItinData<IIC_VLD1x2, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 2, 1]>, + // + // VLD1x3 + InstrItinData<IIC_VLD1x3, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 1]>, + // + // VLD1x4 + InstrItinData<IIC_VLD1x4, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 3, 1]>, + // + // VLD1u + InstrItinData<IIC_VLD1u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 2, 1]>, + // + // VLD1x2u + InstrItinData<IIC_VLD1x2u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 2, 2, 1]>, + // + // VLD1x3u + InstrItinData<IIC_VLD1x3u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 2, 1]>, + // + // VLD1x4u + InstrItinData<IIC_VLD1x4u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 3, 2, 1]>, + // + // VLD1ln + InstrItinData<IIC_VLD1ln, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [3, 1, 1, 1]>, + // + // VLD1lnu + InstrItinData<IIC_VLD1lnu, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [3, 2, 1, 1, 1, 1]>, + // + // VLD1dup + InstrItinData<IIC_VLD1dup, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 1]>, + // + // VLD1dupu + InstrItinData<IIC_VLD1dupu, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 2, 1, 1]>, + // + // VLD2 + InstrItinData<IIC_VLD2, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 2, 1]>, + // + // VLD2x2 + InstrItinData<IIC_VLD2x2, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 3, 1]>, + // + // VLD2ln + InstrItinData<IIC_VLD2ln, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [3, 3, 1, 1, 1, 1]>, + // + // VLD2u + InstrItinData<IIC_VLD2u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 2, 2, 1, 1, 1]>, + // + // VLD2x2u + InstrItinData<IIC_VLD2x2u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 3, 2, 1]>, + // + // VLD2lnu + InstrItinData<IIC_VLD2lnu, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [3, 3, 2, 1, 1, 1, 1, 1]>, + // + // VLD2dup + InstrItinData<IIC_VLD2dup, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 2, 1]>, + // + // VLD2dupu + InstrItinData<IIC_VLD2dupu, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 2, 2, 1, 1]>, + // + // VLD3 + InstrItinData<IIC_VLD3, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [3, 3, 4, 1]>, + // + // VLD3ln + InstrItinData<IIC_VLD3ln, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<5, [A8_NLSPipe], 0>, + InstrStage<5, [A8_LSPipe]>], + [4, 4, 5, 1, 1, 1, 1, 2]>, + // + // VLD3u + InstrItinData<IIC_VLD3u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [3, 3, 4, 2, 1]>, + // + // VLD3lnu + InstrItinData<IIC_VLD3lnu, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<5, [A8_NLSPipe], 0>, + InstrStage<5, [A8_LSPipe]>], + [4, 4, 5, 2, 1, 1, 1, 1, 1, 2]>, + // + // VLD3dup + InstrItinData<IIC_VLD3dup, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 1]>, + // + // VLD3dupu + InstrItinData<IIC_VLD3dupu, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 2, 1, 1]>, + // + // VLD4 + InstrItinData<IIC_VLD4, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [3, 3, 4, 4, 1]>, + // + // VLD4ln + InstrItinData<IIC_VLD4ln, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<5, [A8_NLSPipe], 0>, + InstrStage<5, [A8_LSPipe]>], + [4, 4, 5, 5, 1, 1, 1, 1, 2, 2]>, + // + // VLD4u + InstrItinData<IIC_VLD4u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [3, 3, 4, 4, 2, 1]>, + // + // VLD4lnu + InstrItinData<IIC_VLD4lnu, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<5, [A8_NLSPipe], 0>, + InstrStage<5, [A8_LSPipe]>], + [4, 4, 5, 5, 2, 1, 1, 1, 1, 1, 2, 2]>, + // + // VLD4dup + InstrItinData<IIC_VLD4dup, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 3, 1]>, + // + // VLD4dupu + InstrItinData<IIC_VLD4dupu, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 2, 3, 3, 2, 1, 1]>, + // + // VST1 + InstrItinData<IIC_VST1, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [1, 1, 1]>, + // + // VST1x2 + InstrItinData<IIC_VST1x2, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [1, 1, 1, 1]>, + // + // VST1x3 + InstrItinData<IIC_VST1x3, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [1, 1, 1, 1, 2]>, + // + // VST1x4 + InstrItinData<IIC_VST1x4, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [1, 1, 1, 1, 2, 2]>, + // + // VST1u + InstrItinData<IIC_VST1u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 1, 1, 1, 1]>, + // + // VST1x2u + InstrItinData<IIC_VST1x2u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1]>, + // + // VST1x3u + InstrItinData<IIC_VST1x3u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1, 2]>, + // + // VST1x4u + InstrItinData<IIC_VST1x4u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1, 2, 2]>, + // + // VST1ln + InstrItinData<IIC_VST1ln, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [1, 1, 1]>, + // + // VST1lnu + InstrItinData<IIC_VST1lnu, [InstrStage<1, [A8_Pipe0, A8_Pipe1]>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 1, 1, 1, 1]>, + // + // VST2 + InstrItinData<IIC_VST2, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [1, 1, 1, 1]>, + // + // VST2x2 + InstrItinData<IIC_VST2x2, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [1, 1, 1, 1, 2, 2]>, + // + // VST2u + InstrItinData<IIC_VST2u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1]>, + // + // VST2x2u + InstrItinData<IIC_VST2x2u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1, 2, 2]>, + // + // VST2ln + InstrItinData<IIC_VST2ln, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [1, 1, 1, 1]>, + // + // VST2lnu + InstrItinData<IIC_VST2lnu, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<2, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1]>, + // + // VST3 + InstrItinData<IIC_VST3, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [1, 1, 1, 1, 2]>, + // + // VST3u + InstrItinData<IIC_VST3u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1, 2]>, + // + // VST3ln + InstrItinData<IIC_VST3ln, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [1, 1, 1, 1, 2]>, + // + // VST3lnu + InstrItinData<IIC_VST3lnu, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<3, [A8_NLSPipe], 0>, + InstrStage<3, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1, 2]>, + // + // VST4 + InstrItinData<IIC_VST4, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [1, 1, 1, 1, 2, 2]>, + // + // VST4u + InstrItinData<IIC_VST4u, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1, 2, 2]>, + // + // VST4ln + InstrItinData<IIC_VST4ln, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [1, 1, 1, 1, 2, 2]>, + // + // VST4lnu + InstrItinData<IIC_VST4lnu, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<4, [A8_NLSPipe], 0>, + InstrStage<4, [A8_LSPipe]>], + [2, 1, 1, 1, 1, 1, 2, 2]>, + // + // Double-register FP Unary + InstrItinData<IIC_VUNAD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [5, 2]>, + // + // Quad-register FP Unary + // Result written in N5, but that is relative to the last cycle of multicycle, + // so we use 6 for those cases + InstrItinData<IIC_VUNAQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [6, 2]>, + // + // Double-register FP Binary + InstrItinData<IIC_VBIND, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [5, 2, 2]>, + // + // VPADD, etc. + InstrItinData<IIC_VPBIND, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [5, 2, 2]>, + // + // Double-register FP VMUL + InstrItinData<IIC_VFMULD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [5, 2, 1]>, + + // + // Quad-register FP Binary + // Result written in N5, but that is relative to the last cycle of multicycle, + // so we use 6 for those cases + InstrItinData<IIC_VBINQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [6, 2, 2]>, + // + // Quad-register FP VMUL + InstrItinData<IIC_VFMULQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [6, 2, 1]>, + // + // Move + InstrItinData<IIC_VMOV, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [1, 1]>, + // + // Move Immediate + InstrItinData<IIC_VMOVImm, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [3]>, + // + // Double-register Permute Move + InstrItinData<IIC_VMOVD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>], [2, 1]>, + // + // Quad-register Permute Move + // Result written in N2, but that is relative to the last cycle of multicycle, + // so we use 3 for those cases + InstrItinData<IIC_VMOVQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe]>], [3, 1]>, + // + // Integer to Single-precision Move + InstrItinData<IIC_VMOVIS , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>], [2, 1]>, + // + // Integer to Double-precision Move + InstrItinData<IIC_VMOVID , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>], [2, 1, 1]>, + // + // Single-precision to Integer Move + InstrItinData<IIC_VMOVSI , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>], [20, 1]>, + // + // Double-precision to Integer Move + InstrItinData<IIC_VMOVDI , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>], [20, 20, 1]>, + // + // Integer to Lane Move + InstrItinData<IIC_VMOVISL , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe]>], [3, 1, 1]>, + // + // Vector narrow move + InstrItinData<IIC_VMOVN , [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [2, 1]>, + // + // Double-register Permute + InstrItinData<IIC_VPERMD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>], [2, 2, 1, 1]>, + // + // Quad-register Permute + // Result written in N2, but that is relative to the last cycle of multicycle, + // so we use 3 for those cases + InstrItinData<IIC_VPERMQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe]>], [3, 3, 1, 1]>, + // + // Quad-register Permute (3 cycle issue) + // Result written in N2, but that is relative to the last cycle of multicycle, + // so we use 4 for those cases + InstrItinData<IIC_VPERMQ3, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>, + InstrStage<1, [A8_NPipe], 0>, + InstrStage<2, [A8_NLSPipe]>], [4, 4, 1, 1]>, + // + // Double-register FP Multiple-Accumulate + InstrItinData<IIC_VMACD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [9, 3, 2, 2]>, + // + // Quad-register FP Multiple-Accumulate + // Result written in N9, but that is relative to the last cycle of multicycle, + // so we use 10 for those cases + InstrItinData<IIC_VMACQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [10, 3, 2, 2]>, + // + // Double-register Fused FP Multiple-Accumulate + InstrItinData<IIC_VFMACD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [9, 3, 2, 2]>, + // + // Quad-register Fused FP Multiple-Accumulate + // Result written in N9, but that is relative to the last cycle of multicycle, + // so we use 10 for those cases + InstrItinData<IIC_VFMACQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [10, 3, 2, 2]>, + // + // Double-register Reciprical Step + InstrItinData<IIC_VRECSD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [9, 2, 2]>, + // + // Quad-register Reciprical Step + InstrItinData<IIC_VRECSQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [10, 2, 2]>, + // + // Double-register Integer Count + InstrItinData<IIC_VCNTiD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [3, 2, 2]>, + // + // Quad-register Integer Count + // Result written in N3, but that is relative to the last cycle of multicycle, + // so we use 4 for those cases + InstrItinData<IIC_VCNTiQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [4, 2, 2]>, + // + // Double-register Integer Unary + InstrItinData<IIC_VUNAiD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 2]>, + // + // Quad-register Integer Unary + InstrItinData<IIC_VUNAiQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 2]>, + // + // Double-register Integer Q-Unary + InstrItinData<IIC_VQUNAiD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 1]>, + // + // Quad-register Integer CountQ-Unary + InstrItinData<IIC_VQUNAiQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 1]>, + // + // Double-register Integer Binary + InstrItinData<IIC_VBINiD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [3, 2, 2]>, + // + // Quad-register Integer Binary + InstrItinData<IIC_VBINiQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [3, 2, 2]>, + // + // Double-register Integer Binary (4 cycle) + InstrItinData<IIC_VBINi4D, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 2, 1]>, + // + // Quad-register Integer Binary (4 cycle) + InstrItinData<IIC_VBINi4Q, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 2, 1]>, + + // + // Double-register Integer Subtract + InstrItinData<IIC_VSUBiD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [3, 2, 1]>, + // + // Quad-register Integer Subtract + InstrItinData<IIC_VSUBiQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [3, 2, 1]>, + // + // Double-register Integer Subtract + InstrItinData<IIC_VSUBi4D, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 2, 1]>, + // + // Quad-register Integer Subtract + InstrItinData<IIC_VSUBi4Q, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 2, 1]>, + // + // Double-register Integer Shift + InstrItinData<IIC_VSHLiD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [3, 1, 1]>, + // + // Quad-register Integer Shift + InstrItinData<IIC_VSHLiQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [4, 1, 1]>, + // + // Double-register Integer Shift (4 cycle) + InstrItinData<IIC_VSHLi4D, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [4, 1, 1]>, + // + // Quad-register Integer Shift (4 cycle) + InstrItinData<IIC_VSHLi4Q, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [5, 1, 1]>, + // + // Double-register Integer Pair Add Long + InstrItinData<IIC_VPALiD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [6, 3, 1]>, + // + // Quad-register Integer Pair Add Long + InstrItinData<IIC_VPALiQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [7, 3, 1]>, + // + // Double-register Absolute Difference and Accumulate + InstrItinData<IIC_VABAD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [6, 3, 2, 1]>, + // + // Quad-register Absolute Difference and Accumulate + InstrItinData<IIC_VABAQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [6, 3, 2, 1]>, + + // + // Double-register Integer Multiply (.8, .16) + InstrItinData<IIC_VMULi16D, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [6, 2, 2]>, + // + // Double-register Integer Multiply (.32) + InstrItinData<IIC_VMULi32D, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [7, 2, 1]>, + // + // Quad-register Integer Multiply (.8, .16) + InstrItinData<IIC_VMULi16Q, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [7, 2, 2]>, + // + // Quad-register Integer Multiply (.32) + InstrItinData<IIC_VMULi32Q, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<3, [A8_NPipe]>], [9, 2, 1]>, + // + // Double-register Integer Multiply-Accumulate (.8, .16) + InstrItinData<IIC_VMACi16D, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>], [6, 3, 2, 2]>, + // + // Double-register Integer Multiply-Accumulate (.32) + InstrItinData<IIC_VMACi32D, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [7, 3, 2, 1]>, + // + // Quad-register Integer Multiply-Accumulate (.8, .16) + InstrItinData<IIC_VMACi16Q, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NPipe]>], [7, 3, 2, 2]>, + // + // Quad-register Integer Multiply-Accumulate (.32) + InstrItinData<IIC_VMACi32Q, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NPipe]>, + InstrStage<2, [A8_NLSPipe], 0>, + InstrStage<3, [A8_NPipe]>], [9, 3, 2, 1]>, + // + // Double-register VEXT + InstrItinData<IIC_VEXTD, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>], [2, 1, 1]>, + // + // Quad-register VEXT + InstrItinData<IIC_VEXTQ, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe]>], [3, 1, 1]>, + // + // VTB + InstrItinData<IIC_VTB1, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe]>], [3, 2, 1]>, + InstrItinData<IIC_VTB2, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe]>], [3, 2, 2, 1]>, + InstrItinData<IIC_VTB3, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>, + InstrStage<1, [A8_NPipe], 0>, + InstrStage<2, [A8_NLSPipe]>], [4, 2, 2, 3, 1]>, + InstrItinData<IIC_VTB4, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>, + InstrStage<1, [A8_NPipe], 0>, + InstrStage<2, [A8_NLSPipe]>],[4, 2, 2, 3, 3, 1]>, + // + // VTBX + InstrItinData<IIC_VTBX1, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe]>], [3, 1, 2, 1]>, + InstrItinData<IIC_VTBX2, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<2, [A8_NLSPipe]>], [3, 1, 2, 2, 1]>, + InstrItinData<IIC_VTBX3, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>, + InstrStage<1, [A8_NPipe], 0>, + InstrStage<2, [A8_NLSPipe]>],[4, 1, 2, 2, 3, 1]>, + InstrItinData<IIC_VTBX4, [InstrStage<1, [A8_Pipe0, A8_Pipe1], 0>, + InstrStage<1, [A8_NLSPipe]>, + InstrStage<1, [A8_NPipe], 0>, + InstrStage<2, [A8_NLSPipe]>], [4, 1, 2, 2, 3, 3, 1]> +]>; + +// ===---------------------------------------------------------------------===// +// This following definitions describe the simple machine model which +// will replace itineraries. + +// Cortex-A8 machine model for scheduling and other instruction cost heuristics. +def CortexA8Model : SchedMachineModel { + let IssueWidth = 2; // 2 micro-ops are dispatched per cycle. + let LoadLatency = 2; // Optimistic load latency assuming bypass. + // This is overriden by OperandCycles if the + // Itineraries are queried instead. + let MispredictPenalty = 13; // Based on estimate of pipeline depth. + let CompleteModel = 0; + + let Itineraries = CortexA8Itineraries; +} diff --git a/capstone/suite/synctools/tablegen/ARM/ARMScheduleA9.td b/capstone/suite/synctools/tablegen/ARM/ARMScheduleA9.td new file mode 100644 index 000000000..fc301c589 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMScheduleA9.td @@ -0,0 +1,2579 @@ +//=- ARMScheduleA9.td - ARM Cortex-A9 Scheduling Definitions -*- tablegen -*-=// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the itinerary class data for the ARM Cortex A9 processors. +// +//===----------------------------------------------------------------------===// + +// ===---------------------------------------------------------------------===// +// This section contains legacy support for itineraries. This is +// required until SD and PostRA schedulers are replaced by MachineScheduler. + +// +// Ad-hoc scheduling information derived from pretty vague "Cortex-A9 Technical +// Reference Manual". +// +// Functional units +def A9_Issue0 : FuncUnit; // Issue 0 +def A9_Issue1 : FuncUnit; // Issue 1 +def A9_Branch : FuncUnit; // Branch +def A9_ALU0 : FuncUnit; // ALU / MUL pipeline 0 +def A9_ALU1 : FuncUnit; // ALU pipeline 1 +def A9_AGU : FuncUnit; // Address generation unit for ld / st +def A9_NPipe : FuncUnit; // NEON pipeline +def A9_MUX0 : FuncUnit; // AGU + NEON/FPU multiplexer +def A9_LSUnit : FuncUnit; // L/S Unit +def A9_DRegsVFP: FuncUnit; // FP register set, VFP side +def A9_DRegsN : FuncUnit; // FP register set, NEON side + +// Bypasses +def A9_LdBypass : Bypass; + +def CortexA9Itineraries : ProcessorItineraries< + [A9_Issue0, A9_Issue1, A9_Branch, A9_ALU0, A9_ALU1, A9_AGU, A9_NPipe, A9_MUX0, + A9_LSUnit, A9_DRegsVFP, A9_DRegsN], + [A9_LdBypass], [ + // Two fully-pipelined integer ALU pipelines + + // + // Move instructions, unconditional + InstrItinData<IIC_iMOVi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1]>, + InstrItinData<IIC_iMOVr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1, 1]>, + InstrItinData<IIC_iMOVsi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1, 1]>, + InstrItinData<IIC_iMOVsr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], [2, 1, 1]>, + InstrItinData<IIC_iMOVix2 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [2]>, + InstrItinData<IIC_iMOVix2addpc,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>, + InstrStage<1, [A9_ALU0, A9_ALU1]>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [3]>, + InstrItinData<IIC_iMOVix2ld,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>, + InstrStage<1, [A9_ALU0, A9_ALU1]>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], [5]>, + // + // MVN instructions + InstrItinData<IIC_iMVNi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], + [1]>, + InstrItinData<IIC_iMVNr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], + [1, 1], [NoBypass, A9_LdBypass]>, + InstrItinData<IIC_iMVNsi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], + [2, 1]>, + InstrItinData<IIC_iMVNsr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<3, [A9_ALU0, A9_ALU1]>], + [3, 1, 1]>, + // + // No operand cycles + InstrItinData<IIC_iALUx , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>]>, + // + // Binary Instructions that produce a result + InstrItinData<IIC_iALUi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], + [1, 1], [NoBypass, A9_LdBypass]>, + InstrItinData<IIC_iALUr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], + [1, 1, 1], [NoBypass, A9_LdBypass, A9_LdBypass]>, + InstrItinData<IIC_iALUsi, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], + [2, 1, 1], [NoBypass, A9_LdBypass, NoBypass]>, + InstrItinData<IIC_iALUsir,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], + [2, 1, 1], [NoBypass, NoBypass, A9_LdBypass]>, + InstrItinData<IIC_iALUsr, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<3, [A9_ALU0, A9_ALU1]>], + [3, 1, 1, 1], + [NoBypass, A9_LdBypass, NoBypass, NoBypass]>, + // + // Bitwise Instructions that produce a result + InstrItinData<IIC_iBITi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1, 1]>, + InstrItinData<IIC_iBITr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1, 1, 1]>, + InstrItinData<IIC_iBITsi, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], [2, 1, 1]>, + InstrItinData<IIC_iBITsr, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<3, [A9_ALU0, A9_ALU1]>], [3, 1, 1, 1]>, + // + // Unary Instructions that produce a result + + // CLZ, RBIT, etc. + InstrItinData<IIC_iUNAr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1, 1]>, + + // BFC, BFI, UBFX, SBFX + InstrItinData<IIC_iUNAsi, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], [2, 1]>, + + // + // Zero and sign extension instructions + InstrItinData<IIC_iEXTr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [2, 1]>, + InstrItinData<IIC_iEXTAr, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], [3, 1, 1]>, + InstrItinData<IIC_iEXTAsr,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<3, [A9_ALU0, A9_ALU1]>], [3, 1, 1, 1]>, + // + // Compare instructions + InstrItinData<IIC_iCMPi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], + [1], [A9_LdBypass]>, + InstrItinData<IIC_iCMPr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], + [1, 1], [A9_LdBypass, A9_LdBypass]>, + InstrItinData<IIC_iCMPsi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], + [1, 1], [A9_LdBypass, NoBypass]>, + InstrItinData<IIC_iCMPsr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<3, [A9_ALU0, A9_ALU1]>], + [1, 1, 1], [A9_LdBypass, NoBypass, NoBypass]>, + // + // Test instructions + InstrItinData<IIC_iTSTi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1]>, + InstrItinData<IIC_iTSTr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1, 1]>, + InstrItinData<IIC_iTSTsi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], [1, 1]>, + InstrItinData<IIC_iTSTsr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<3, [A9_ALU0, A9_ALU1]>], [1, 1, 1]>, + // + // Move instructions, conditional + // FIXME: Correctly model the extra input dep on the destination. + InstrItinData<IIC_iCMOVi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1]>, + InstrItinData<IIC_iCMOVr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1, 1]>, + InstrItinData<IIC_iCMOVsi , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [1, 1]>, + InstrItinData<IIC_iCMOVsr , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0, A9_ALU1]>], [2, 1, 1]>, + InstrItinData<IIC_iCMOVix2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>, + InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], [2]>, + + // Integer multiply pipeline + // + InstrItinData<IIC_iMUL16 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0]>], [3, 1, 1]>, + InstrItinData<IIC_iMAC16 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0]>], + [3, 1, 1, 1]>, + InstrItinData<IIC_iMUL32 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0]>], [4, 1, 1]>, + InstrItinData<IIC_iMAC32 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<2, [A9_ALU0]>], + [4, 1, 1, 1]>, + InstrItinData<IIC_iMUL64 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<3, [A9_ALU0]>], [4, 5, 1, 1]>, + InstrItinData<IIC_iMAC64 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<3, [A9_ALU0]>], + [4, 5, 1, 1]>, + // Integer load pipeline + // FIXME: The timings are some rough approximations + // + // Immediate offset + InstrItinData<IIC_iLoad_i , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [3, 1], [A9_LdBypass]>, + InstrItinData<IIC_iLoad_bh_i, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [4, 1], [A9_LdBypass]>, + // FIXME: If address is 64-bit aligned, AGU cycles is 1. + InstrItinData<IIC_iLoad_d_i , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [3, 3, 1], [A9_LdBypass]>, + // + // Register offset + InstrItinData<IIC_iLoad_r , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [3, 1, 1], [A9_LdBypass]>, + InstrItinData<IIC_iLoad_bh_r, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [4, 1, 1], [A9_LdBypass]>, + InstrItinData<IIC_iLoad_d_r , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [3, 3, 1, 1], [A9_LdBypass]>, + // + // Scaled register offset + InstrItinData<IIC_iLoad_si , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit], 0>], + [4, 1, 1], [A9_LdBypass]>, + InstrItinData<IIC_iLoad_bh_si,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [5, 1, 1], [A9_LdBypass]>, + // + // Immediate offset with update + InstrItinData<IIC_iLoad_iu , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [3, 2, 1], [A9_LdBypass]>, + InstrItinData<IIC_iLoad_bh_iu,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [4, 3, 1], [A9_LdBypass]>, + // + // Register offset with update + InstrItinData<IIC_iLoad_ru , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [3, 2, 1, 1], [A9_LdBypass]>, + InstrItinData<IIC_iLoad_bh_ru,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [4, 3, 1, 1], [A9_LdBypass]>, + InstrItinData<IIC_iLoad_d_ru, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [3, 3, 1, 1], [A9_LdBypass]>, + // + // Scaled register offset with update + InstrItinData<IIC_iLoad_siu , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [4, 3, 1, 1], [A9_LdBypass]>, + InstrItinData<IIC_iLoad_bh_siu,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [5, 4, 1, 1], [A9_LdBypass]>, + // + // Load multiple, def is the 5th operand. + // FIXME: This assumes 3 to 4 registers. + InstrItinData<IIC_iLoad_m , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 1, 1, 3], + [NoBypass, NoBypass, NoBypass, NoBypass, A9_LdBypass], + -1>, // dynamic uops + // + // Load multiple + update, defs are the 1st and 5th operands. + InstrItinData<IIC_iLoad_mu , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<2, [A9_LSUnit]>], + [2, 1, 1, 1, 3], + [NoBypass, NoBypass, NoBypass, NoBypass, A9_LdBypass], + -1>, // dynamic uops + // + // Load multiple plus branch + InstrItinData<IIC_iLoad_mBr, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 1>, + InstrStage<2, [A9_LSUnit]>, + InstrStage<1, [A9_Branch]>], + [1, 2, 1, 1, 3], + [NoBypass, NoBypass, NoBypass, NoBypass, A9_LdBypass], + -1>, // dynamic uops + // + // Pop, def is the 3rd operand. + InstrItinData<IIC_iPop , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 3], + [NoBypass, NoBypass, A9_LdBypass], + -1>, // dynamic uops + // + // Pop + branch, def is the 3rd operand. + InstrItinData<IIC_iPop_Br, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<2, [A9_LSUnit]>, + InstrStage<1, [A9_Branch]>], + [1, 1, 3], + [NoBypass, NoBypass, A9_LdBypass], + -1>, // dynamic uops + // + // iLoadi + iALUr for t2LDRpci_pic. + InstrItinData<IIC_iLoadiALU, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>, + InstrStage<1, [A9_ALU0, A9_ALU1]>], + [2, 1]>, + + // Integer store pipeline + /// + // Immediate offset + InstrItinData<IIC_iStore_i , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], [1, 1]>, + InstrItinData<IIC_iStore_bh_i,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], [1, 1]>, + // FIXME: If address is 64-bit aligned, AGU cycles is 1. + InstrItinData<IIC_iStore_d_i, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], [1, 1]>, + // + // Register offset + InstrItinData<IIC_iStore_r , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], [1, 1, 1]>, + InstrItinData<IIC_iStore_bh_r,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], [1, 1, 1]>, + InstrItinData<IIC_iStore_d_r, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], [1, 1, 1]>, + // + // Scaled register offset + InstrItinData<IIC_iStore_si , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], [1, 1, 1]>, + InstrItinData<IIC_iStore_bh_si,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], [1, 1, 1]>, + // + // Immediate offset with update + InstrItinData<IIC_iStore_iu , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], [2, 1, 1]>, + InstrItinData<IIC_iStore_bh_iu,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], [3, 1, 1]>, + // + // Register offset with update + InstrItinData<IIC_iStore_ru , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1, 1, 1]>, + InstrItinData<IIC_iStore_bh_ru,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], + [3, 1, 1, 1]>, + InstrItinData<IIC_iStore_d_ru, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], + [3, 1, 1, 1]>, + // + // Scaled register offset with update + InstrItinData<IIC_iStore_siu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1, 1, 1]>, + InstrItinData<IIC_iStore_bh_siu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_AGU], 1>, + InstrStage<1, [A9_LSUnit]>], + [3, 1, 1, 1]>, + // + // Store multiple + InstrItinData<IIC_iStore_m , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<2, [A9_LSUnit]>], + [], [], -1>, // dynamic uops + // + // Store multiple + update + InstrItinData<IIC_iStore_mu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_AGU], 0>, + InstrStage<2, [A9_LSUnit]>], + [2], [], -1>, // dynamic uops + // + // Preload + InstrItinData<IIC_Preload, [InstrStage<1, [A9_Issue0, A9_Issue1]>], [1, 1]>, + + // Branch + // + // no delay slots, so the latency of a branch is unimportant + InstrItinData<IIC_Br , [InstrStage<1, [A9_Issue0], 0>, + InstrStage<1, [A9_Issue1], 0>, + InstrStage<1, [A9_Branch]>]>, + + // VFP and NEON shares the same register file. This means that every VFP + // instruction should wait for full completion of the consecutive NEON + // instruction and vice-versa. We model this behavior with two artificial FUs: + // DRegsVFP and DRegsVFP. + // + // Every VFP instruction: + // - Acquires DRegsVFP resource for 1 cycle + // - Reserves DRegsN resource for the whole duration (including time to + // register file writeback!). + // Every NEON instruction does the same but with FUs swapped. + // + // Since the reserved FU cannot be acquired, this models precisely + // "cross-domain" stalls. + + // VFP + // Issue through integer pipeline, and execute in NEON unit. + + // FP Special Register to Integer Register File Move + InstrItinData<IIC_fpSTAT , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1]>, + // + // Single-precision FP Unary + InstrItinData<IIC_fpUNA32 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + // Extra latency cycles since wbck is 2 cycles + InstrStage<3, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1, 1]>, + // + // Double-precision FP Unary + InstrItinData<IIC_fpUNA64 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + // Extra latency cycles since wbck is 2 cycles + InstrStage<3, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1, 1]>, + + // + // Single-precision FP Compare + InstrItinData<IIC_fpCMP32 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + // Extra latency cycles since wbck is 4 cycles + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1, 1]>, + // + // Double-precision FP Compare + InstrItinData<IIC_fpCMP64 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + // Extra latency cycles since wbck is 4 cycles + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1, 1]>, + // + // Single to Double FP Convert + InstrItinData<IIC_fpCVTSD , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + // + // Double to Single FP Convert + InstrItinData<IIC_fpCVTDS , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + + // + // Single to Half FP Convert + InstrItinData<IIC_fpCVTSH , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + // + // Half to Single FP Convert + InstrItinData<IIC_fpCVTHS , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<3, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [2, 1]>, + + // + // Single-Precision FP to Integer Convert + InstrItinData<IIC_fpCVTSI , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + // + // Double-Precision FP to Integer Convert + InstrItinData<IIC_fpCVTDI , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + // + // Integer to Single-Precision FP Convert + InstrItinData<IIC_fpCVTIS , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + // + // Integer to Double-Precision FP Convert + InstrItinData<IIC_fpCVTID , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + // + // Single-precision FP ALU + InstrItinData<IIC_fpALU32 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1, 1]>, + // + // Double-precision FP ALU + InstrItinData<IIC_fpALU64 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<5, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1, 1]>, + // + // Single-precision FP Multiply + InstrItinData<IIC_fpMUL32 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<6, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [5, 1, 1]>, + // + // Double-precision FP Multiply + InstrItinData<IIC_fpMUL64 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<7, [A9_DRegsN], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [6, 1, 1]>, + // + // Single-precision FP MAC + InstrItinData<IIC_fpMAC32 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<9, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [8, 1, 1, 1]>, + // + // Double-precision FP MAC + InstrItinData<IIC_fpMAC64 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<10, [A9_DRegsN], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [9, 1, 1, 1]>, + // + // Single-precision Fused FP MAC + InstrItinData<IIC_fpFMAC32, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<9, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [8, 1, 1, 1]>, + // + // Double-precision Fused FP MAC + InstrItinData<IIC_fpFMAC64, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<10, [A9_DRegsN], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [9, 1, 1, 1]>, + // + // Single-precision FP DIV + InstrItinData<IIC_fpDIV32 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<16, [A9_DRegsN], 0, Reserved>, + InstrStage<10, [A9_NPipe]>], + [15, 1, 1]>, + // + // Double-precision FP DIV + InstrItinData<IIC_fpDIV64 , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<26, [A9_DRegsN], 0, Reserved>, + InstrStage<20, [A9_NPipe]>], + [25, 1, 1]>, + // + // Single-precision FP SQRT + InstrItinData<IIC_fpSQRT32, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<18, [A9_DRegsN], 0, Reserved>, + InstrStage<13, [A9_NPipe]>], + [17, 1]>, + // + // Double-precision FP SQRT + InstrItinData<IIC_fpSQRT64, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<33, [A9_DRegsN], 0, Reserved>, + InstrStage<28, [A9_NPipe]>], + [32, 1]>, + + // + // Integer to Single-precision Move + InstrItinData<IIC_fpMOVIS, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + // Extra 1 latency cycle since wbck is 2 cycles + InstrStage<3, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1, 1]>, + // + // Integer to Double-precision Move + InstrItinData<IIC_fpMOVID, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + // Extra 1 latency cycle since wbck is 2 cycles + InstrStage<3, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1, 1, 1]>, + // + // Single-precision to Integer Move + // + // On A9 move-from-VFP is free to issue with no stall if other VFP + // operations are in flight. I assume it still can't dual-issue though. + InstrItinData<IIC_fpMOVSI, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>], + [2, 1]>, + // + // Double-precision to Integer Move + // + // On A9 move-from-VFP is free to issue with no stall if other VFP + // operations are in flight. I assume it still can't dual-issue though. + InstrItinData<IIC_fpMOVDI, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>], + [2, 1, 1]>, + // + // Single-precision FP Load + InstrItinData<IIC_fpLoad32, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1]>, + // + // Double-precision FP Load + // FIXME: Result latency is 1 if address is 64-bit aligned. + InstrItinData<IIC_fpLoad64, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1]>, + // + // FP Load Multiple + // FIXME: assumes 2 doubles which requires 2 LS cycles. + InstrItinData<IIC_fpLoad_m, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 1, 1], [], -1>, // dynamic uops + // + // FP Load Multiple + update + // FIXME: assumes 2 doubles which requires 2 LS cycles. + InstrItinData<IIC_fpLoad_mu,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 1, 1, 1], [], -1>, // dynamic uops + // + // Single-precision FP Store + InstrItinData<IIC_fpStore32,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1]>, + // + // Double-precision FP Store + InstrItinData<IIC_fpStore64,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1]>, + // + // FP Store Multiple + // FIXME: assumes 2 doubles which requires 2 LS cycles. + InstrItinData<IIC_fpStore_m,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 1, 1], [], -1>, // dynamic uops + // + // FP Store Multiple + update + // FIXME: assumes 2 doubles which requires 2 LS cycles. + InstrItinData<IIC_fpStore_mu,[InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsVFP], 0, Required>, + InstrStage<2, [A9_DRegsN], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 1, 1, 1], [], -1>, // dynamic uops + // NEON + // VLD1 + InstrItinData<IIC_VLD1, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1]>, + // VLD1x2 + InstrItinData<IIC_VLD1x2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1, 1]>, + // VLD1x3 + InstrItinData<IIC_VLD1x3, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 2, 1]>, + // VLD1x4 + InstrItinData<IIC_VLD1x4, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 2, 2, 1]>, + // VLD1u + InstrItinData<IIC_VLD1u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 2, 1]>, + // VLD1x2u + InstrItinData<IIC_VLD1x2u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1, 2, 1]>, + // VLD1x3u + InstrItinData<IIC_VLD1x3u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 2, 2, 1]>, + // VLD1x4u + InstrItinData<IIC_VLD1x4u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 2, 2, 2, 1]>, + // + // VLD1ln + InstrItinData<IIC_VLD1ln, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [3, 1, 1, 1]>, + // + // VLD1lnu + InstrItinData<IIC_VLD1lnu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [3, 2, 1, 1, 1, 1]>, + // + // VLD1dup + InstrItinData<IIC_VLD1dup, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1]>, + // + // VLD1dupu + InstrItinData<IIC_VLD1dupu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 2, 1, 1]>, + // + // VLD2 + InstrItinData<IIC_VLD2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 2, 1]>, + // + // VLD2x2 + InstrItinData<IIC_VLD2x2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 3, 2, 3, 1]>, + // + // VLD2ln + InstrItinData<IIC_VLD2ln, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [3, 3, 1, 1, 1, 1]>, + // + // VLD2u + InstrItinData<IIC_VLD2u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 2, 2, 1, 1, 1]>, + // + // VLD2x2u + InstrItinData<IIC_VLD2x2u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 3, 2, 3, 2, 1]>, + // + // VLD2lnu + InstrItinData<IIC_VLD2lnu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [3, 3, 2, 1, 1, 1, 1, 1]>, + // + // VLD2dup + InstrItinData<IIC_VLD2dup, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 2, 1]>, + // + // VLD2dupu + InstrItinData<IIC_VLD2dupu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 2, 2, 1, 1]>, + // + // VLD3 + InstrItinData<IIC_VLD3, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<9,[A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [3, 3, 4, 1]>, + // + // VLD3ln + InstrItinData<IIC_VLD3ln, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<11,[A9_DRegsVFP], 0, Reserved>, + InstrStage<5, [A9_NPipe], 0>, + InstrStage<5, [A9_LSUnit]>], + [5, 5, 6, 1, 1, 1, 1, 2]>, + // + // VLD3u + InstrItinData<IIC_VLD3u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<9,[A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [3, 3, 4, 2, 1]>, + // + // VLD3lnu + InstrItinData<IIC_VLD3lnu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<11,[A9_DRegsVFP], 0, Reserved>, + InstrStage<5, [A9_NPipe], 0>, + InstrStage<5, [A9_LSUnit]>], + [5, 5, 6, 2, 1, 1, 1, 1, 1, 2]>, + // + // VLD3dup + InstrItinData<IIC_VLD3dup, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<9, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [3, 3, 4, 1]>, + // + // VLD3dupu + InstrItinData<IIC_VLD3dupu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<9, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [3, 3, 4, 2, 1, 1]>, + // + // VLD4 + InstrItinData<IIC_VLD4, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<9,[A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [3, 3, 4, 4, 1]>, + // + // VLD4ln + InstrItinData<IIC_VLD4ln, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<10,[A9_DRegsVFP], 0, Reserved>, + InstrStage<4, [A9_NPipe], 0>, + InstrStage<4, [A9_LSUnit]>], + [4, 4, 5, 5, 1, 1, 1, 1, 2, 2]>, + // + // VLD4u + InstrItinData<IIC_VLD4u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<9,[A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [3, 3, 4, 4, 2, 1]>, + // + // VLD4lnu + InstrItinData<IIC_VLD4lnu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<10,[A9_DRegsVFP], 0, Reserved>, + InstrStage<4, [A9_NPipe], 0>, + InstrStage<4, [A9_LSUnit]>], + [4, 4, 5, 5, 2, 1, 1, 1, 1, 1, 2, 2]>, + // + // VLD4dup + InstrItinData<IIC_VLD4dup, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 2, 3, 3, 1]>, + // + // VLD4dupu + InstrItinData<IIC_VLD4dupu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 2, 3, 3, 2, 1, 1]>, + // + // VST1 + InstrItinData<IIC_VST1, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1, 1]>, + // + // VST1x2 + InstrItinData<IIC_VST1x2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1, 1, 1]>, + // + // VST1x3 + InstrItinData<IIC_VST1x3, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 1, 1, 2]>, + // + // VST1x4 + InstrItinData<IIC_VST1x4, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 1, 1, 2, 2]>, + // + // VST1u + InstrItinData<IIC_VST1u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1, 1, 1, 1]>, + // + // VST1x2u + InstrItinData<IIC_VST1x2u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1]>, + // + // VST1x3u + InstrItinData<IIC_VST1x3u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1, 2]>, + // + // VST1x4u + InstrItinData<IIC_VST1x4u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1, 2, 2]>, + // + // VST1ln + InstrItinData<IIC_VST1ln, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1, 1]>, + // + // VST1lnu + InstrItinData<IIC_VST1lnu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1, 1, 1, 1]>, + // + // VST2 + InstrItinData<IIC_VST2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1, 1, 1]>, + // + // VST2x2 + InstrItinData<IIC_VST2x2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<3, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [1, 1, 1, 1, 2, 2]>, + // + // VST2u + InstrItinData<IIC_VST2u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1]>, + // + // VST2x2u + InstrItinData<IIC_VST2x2u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<3, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1, 2, 2]>, + // + // VST2ln + InstrItinData<IIC_VST2ln, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [1, 1, 1, 1]>, + // + // VST2lnu + InstrItinData<IIC_VST2lnu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe], 0>, + InstrStage<1, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1]>, + // + // VST3 + InstrItinData<IIC_VST3, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 1, 1, 2]>, + // + // VST3u + InstrItinData<IIC_VST3u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1, 2]>, + // + // VST3ln + InstrItinData<IIC_VST3ln, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<3, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [1, 1, 1, 1, 2]>, + // + // VST3lnu + InstrItinData<IIC_VST3lnu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<3, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe], 0>, + InstrStage<3, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1, 2]>, + // + // VST4 + InstrItinData<IIC_VST4, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 1, 1, 2, 2]>, + // + // VST4u + InstrItinData<IIC_VST4u, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1, 2, 2]>, + // + // VST4ln + InstrItinData<IIC_VST4ln, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [1, 1, 1, 1, 2, 2]>, + // + // VST4lnu + InstrItinData<IIC_VST4lnu, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<2, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe], 0>, + InstrStage<2, [A9_LSUnit]>], + [2, 1, 1, 1, 1, 1, 2, 2]>, + + // + // Double-register Integer Unary + InstrItinData<IIC_VUNAiD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 2]>, + // + // Quad-register Integer Unary + InstrItinData<IIC_VUNAiQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 2]>, + // + // Double-register Integer Q-Unary + InstrItinData<IIC_VQUNAiD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + // + // Quad-register Integer CountQ-Unary + InstrItinData<IIC_VQUNAiQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1]>, + // + // Double-register Integer Binary + InstrItinData<IIC_VBINiD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3, 2, 2]>, + // + // Quad-register Integer Binary + InstrItinData<IIC_VBINiQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3, 2, 2]>, + // + // Double-register Integer Subtract + InstrItinData<IIC_VSUBiD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3, 2, 1]>, + // + // Quad-register Integer Subtract + InstrItinData<IIC_VSUBiQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3, 2, 1]>, + // + // Double-register Integer Shift + InstrItinData<IIC_VSHLiD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3, 1, 1]>, + // + // Quad-register Integer Shift + InstrItinData<IIC_VSHLiQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3, 1, 1]>, + // + // Double-register Integer Shift (4 cycle) + InstrItinData<IIC_VSHLi4D, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1, 1]>, + // + // Quad-register Integer Shift (4 cycle) + InstrItinData<IIC_VSHLi4Q, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 1, 1]>, + // + // Double-register Integer Binary (4 cycle) + InstrItinData<IIC_VBINi4D, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 2, 2]>, + // + // Quad-register Integer Binary (4 cycle) + InstrItinData<IIC_VBINi4Q, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 2, 2]>, + // + // Double-register Integer Subtract (4 cycle) + InstrItinData<IIC_VSUBi4D, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 2, 1]>, + // + // Quad-register Integer Subtract (4 cycle) + InstrItinData<IIC_VSUBi4Q, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [4, 2, 1]>, + + // + // Double-register Integer Count + InstrItinData<IIC_VCNTiD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3, 2, 2]>, + // + // Quad-register Integer Count + // Result written in N3, but that is relative to the last cycle of multicycle, + // so we use 4 for those cases + InstrItinData<IIC_VCNTiQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [4, 2, 2]>, + // + // Double-register Absolute Difference and Accumulate + InstrItinData<IIC_VABAD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [6, 3, 2, 1]>, + // + // Quad-register Absolute Difference and Accumulate + InstrItinData<IIC_VABAQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [6, 3, 2, 1]>, + // + // Double-register Integer Pair Add Long + InstrItinData<IIC_VPALiD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [6, 3, 1]>, + // + // Quad-register Integer Pair Add Long + InstrItinData<IIC_VPALiQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [6, 3, 1]>, + + // + // Double-register Integer Multiply (.8, .16) + InstrItinData<IIC_VMULi16D, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [6, 2, 2]>, + // + // Quad-register Integer Multiply (.8, .16) + InstrItinData<IIC_VMULi16Q, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [7, 2, 2]>, + + // + // Double-register Integer Multiply (.32) + InstrItinData<IIC_VMULi32D, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [7, 2, 1]>, + // + // Quad-register Integer Multiply (.32) + InstrItinData<IIC_VMULi32Q, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 9 cycles + InstrStage<10, [A9_DRegsVFP], 0, Reserved>, + InstrStage<4, [A9_NPipe]>], + [9, 2, 1]>, + // + // Double-register Integer Multiply-Accumulate (.8, .16) + InstrItinData<IIC_VMACi16D, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [6, 3, 2, 2]>, + // + // Double-register Integer Multiply-Accumulate (.32) + InstrItinData<IIC_VMACi32D, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [7, 3, 2, 1]>, + // + // Quad-register Integer Multiply-Accumulate (.8, .16) + InstrItinData<IIC_VMACi16Q, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [7, 3, 2, 2]>, + // + // Quad-register Integer Multiply-Accumulate (.32) + InstrItinData<IIC_VMACi32Q, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 9 cycles + InstrStage<10, [A9_DRegsVFP], 0, Reserved>, + InstrStage<4, [A9_NPipe]>], + [9, 3, 2, 1]>, + + // + // Move + InstrItinData<IIC_VMOV, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<1, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1,1]>, + // + // Move Immediate + InstrItinData<IIC_VMOVImm, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3]>, + // + // Double-register Permute Move + InstrItinData<IIC_VMOVD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [2, 1]>, + // + // Quad-register Permute Move + InstrItinData<IIC_VMOVQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [2, 1]>, + // + // Integer to Single-precision Move + InstrItinData<IIC_VMOVIS , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<3, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1, 1]>, + // + // Integer to Double-precision Move + InstrItinData<IIC_VMOVID , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<3, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [1, 1, 1]>, + // + // Single-precision to Integer Move + InstrItinData<IIC_VMOVSI , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<3, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [2, 1]>, + // + // Double-precision to Integer Move + InstrItinData<IIC_VMOVDI , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<3, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [2, 2, 1]>, + // + // Integer to Lane Move + InstrItinData<IIC_VMOVISL , [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + InstrStage<4, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [3, 1, 1]>, + + // + // Vector narrow move + InstrItinData<IIC_VMOVN, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [3, 1]>, + // + // Double-register FP Unary + InstrItinData<IIC_VUNAD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [5, 2]>, + // + // Quad-register FP Unary + // Result written in N5, but that is relative to the last cycle of multicycle, + // so we use 6 for those cases + InstrItinData<IIC_VUNAQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [6, 2]>, + // + // Double-register FP Binary + // FIXME: We're using this itin for many instructions and [2, 2] here is too + // optimistic. + InstrItinData<IIC_VBIND, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [5, 2, 2]>, + + // + // VPADD, etc. + InstrItinData<IIC_VPBIND, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [5, 1, 1]>, + // + // Double-register FP VMUL + InstrItinData<IIC_VFMULD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [5, 2, 1]>, + // + // Quad-register FP Binary + // Result written in N5, but that is relative to the last cycle of multicycle, + // so we use 6 for those cases + // FIXME: We're using this itin for many instructions and [2, 2] here is too + // optimistic. + InstrItinData<IIC_VBINQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [6, 2, 2]>, + // + // Quad-register FP VMUL + InstrItinData<IIC_VFMULQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [6, 2, 1]>, + // + // Double-register FP Multiple-Accumulate + InstrItinData<IIC_VMACD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [6, 3, 2, 1]>, + // + // Quad-register FP Multiple-Accumulate + // Result written in N9, but that is relative to the last cycle of multicycle, + // so we use 10 for those cases + InstrItinData<IIC_VMACQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 9 cycles + InstrStage<10, [A9_DRegsVFP], 0, Reserved>, + InstrStage<4, [A9_NPipe]>], + [8, 4, 2, 1]>, + // + // Double-register Fused FP Multiple-Accumulate + InstrItinData<IIC_VFMACD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [6, 3, 2, 1]>, + // + // Quad-register Fused FP Multiple-Accumulate + // Result written in N9, but that is relative to the last cycle of multicycle, + // so we use 10 for those cases + InstrItinData<IIC_VFMACQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 9 cycles + InstrStage<10, [A9_DRegsVFP], 0, Reserved>, + InstrStage<4, [A9_NPipe]>], + [8, 4, 2, 1]>, + // + // Double-register Reciprical Step + InstrItinData<IIC_VRECSD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 10 cycles + InstrStage<11, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [9, 2, 2]>, + // + // Quad-register Reciprical Step + InstrItinData<IIC_VRECSQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 11 cycles + InstrStage<12, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [10, 2, 2]>, + // + // Double-register Permute + InstrItinData<IIC_VPERMD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [2, 2, 1, 1]>, + // + // Quad-register Permute + // Result written in N2, but that is relative to the last cycle of multicycle, + // so we use 3 for those cases + InstrItinData<IIC_VPERMQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [3, 3, 1, 1]>, + // + // Quad-register Permute (3 cycle issue) + // Result written in N2, but that is relative to the last cycle of multicycle, + // so we use 4 for those cases + InstrItinData<IIC_VPERMQ3, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 8 cycles + InstrStage<9, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe]>], + [4, 4, 1, 1]>, + + // + // Double-register VEXT + InstrItinData<IIC_VEXTD, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 6 cycles + InstrStage<7, [A9_DRegsVFP], 0, Reserved>, + InstrStage<1, [A9_NPipe]>], + [2, 1, 1]>, + // + // Quad-register VEXT + InstrItinData<IIC_VEXTQ, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [3, 1, 2]>, + // + // VTB + InstrItinData<IIC_VTB1, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [3, 2, 1]>, + InstrItinData<IIC_VTB2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [3, 2, 2, 1]>, + InstrItinData<IIC_VTB3, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<2, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 8 cycles + InstrStage<9, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe]>], + [4, 2, 2, 3, 1]>, + InstrItinData<IIC_VTB4, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 8 cycles + InstrStage<9, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe]>], + [4, 2, 2, 3, 3, 1]>, + // + // VTBX + InstrItinData<IIC_VTBX1, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [3, 1, 2, 1]>, + InstrItinData<IIC_VTBX2, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 7 cycles + InstrStage<8, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [3, 1, 2, 2, 1]>, + InstrItinData<IIC_VTBX3, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 8 cycles + InstrStage<9, [A9_DRegsVFP], 0, Reserved>, + InstrStage<3, [A9_NPipe]>], + [4, 1, 2, 2, 3, 1]>, + InstrItinData<IIC_VTBX4, [InstrStage<1, [A9_Issue0, A9_Issue1], 0>, + InstrStage<1, [A9_MUX0], 0>, + InstrStage<1, [A9_DRegsN], 0, Required>, + // Extra latency cycles since wbck is 8 cycles + InstrStage<9, [A9_DRegsVFP], 0, Reserved>, + InstrStage<2, [A9_NPipe]>], + [4, 1, 2, 2, 3, 3, 1]> +]>; + +// ===---------------------------------------------------------------------===// +// The following definitions describe the simpler per-operand machine model. +// This works with MachineScheduler and will eventually replace itineraries. + +class A9WriteLMOpsListType<list<WriteSequence> writes> { + list <WriteSequence> Writes = writes; + SchedMachineModel SchedModel = ?; +} + +// Cortex-A9 machine model for scheduling and other instruction cost heuristics. +def CortexA9Model : SchedMachineModel { + let IssueWidth = 2; // 2 micro-ops are dispatched per cycle. + let MicroOpBufferSize = 56; // Based on available renamed registers. + let LoadLatency = 2; // Optimistic load latency assuming bypass. + // This is overriden by OperandCycles if the + // Itineraries are queried instead. + let MispredictPenalty = 8; // Based on estimate of pipeline depth. + + let Itineraries = CortexA9Itineraries; + + // FIXME: Many vector operations were never given an itinerary. We + // haven't mapped these to the new model either. + let CompleteModel = 0; + + // FIXME: Remove when all errors have been fixed. + let FullInstRWOverlapCheck = 0; +} + +//===----------------------------------------------------------------------===// +// Define each kind of processor resource and number available. +// +// The AGU unit has BufferSize=1 so that the latency between operations +// that use it are considered to stall other operations. +// +// The FP unit has BufferSize=0 so that it is a hard dispatch +// hazard. No instruction may be dispatched while the unit is reserved. + +let SchedModel = CortexA9Model in { + +def A9UnitALU : ProcResource<2>; +def A9UnitMul : ProcResource<1> { let Super = A9UnitALU; } +def A9UnitAGU : ProcResource<1> { let BufferSize = 1; } +def A9UnitLS : ProcResource<1>; +def A9UnitFP : ProcResource<1> { let BufferSize = 0; } +def A9UnitB : ProcResource<1>; + +//===----------------------------------------------------------------------===// +// Define scheduler read/write types with their resources and latency on A9. + +// Consume an issue slot, but no processor resources. This is useful when all +// other writes associated with the operand have NumMicroOps = 0. +def A9WriteIssue : SchedWriteRes<[]> { let Latency = 0; } + +// Write an integer register. +def A9WriteI : SchedWriteRes<[A9UnitALU]>; +// Write an integer shifted-by register +def A9WriteIsr : SchedWriteRes<[A9UnitALU]> { let Latency = 2; } + +// Basic ALU. +def A9WriteALU : SchedWriteRes<[A9UnitALU]>; +// ALU with operand shifted by immediate. +def : WriteRes<WriteALUsi, [A9UnitALU]> { let Latency = 2; } +// ALU with operand shifted by register. +def A9WriteALUsr : SchedWriteRes<[A9UnitALU]> { let Latency = 3; } + +// Multiplication +def A9WriteM : SchedWriteRes<[A9UnitMul, A9UnitMul]> { let Latency = 4; } +def A9WriteMHi : SchedWriteRes<[A9UnitMul]> { let Latency = 5; + let NumMicroOps = 0; } +def A9WriteM16 : SchedWriteRes<[A9UnitMul]> { let Latency = 3; } +def A9WriteM16Hi : SchedWriteRes<[A9UnitMul]> { let Latency = 4; + let NumMicroOps = 0; } +def : SchedAlias<WriteMUL16, A9WriteM16>; +def : SchedAlias<WriteMUL32, A9WriteM>; +def : SchedAlias<WriteMUL64Lo, A9WriteM>; +def : SchedAlias<WriteMUL64Hi, A9WriteMHi>; +def : SchedAlias<WriteMAC16, A9WriteM16>; +def : SchedAlias<WriteMAC32, A9WriteM>; +def : SchedAlias<WriteMAC64Lo, A9WriteM>; +def : SchedAlias<WriteMAC64Hi, A9WriteMHi>; +def : ReadAdvance<ReadMUL, 0>; +def : ReadAdvance<ReadMAC, 0>; + +// Floating-point +// Only one FP or AGU instruction may issue per cycle. We model this +// by having FP instructions consume the AGU resource. +def A9WriteF : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 4; } +def A9WriteFMov : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 1; } +def A9WriteFMulS : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 5; } +def A9WriteFMulD : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 6; } +def A9WriteFMAS : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 8; } + +def A9WriteFMAD : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 9; } +def A9WriteFDivS : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 15; } +def A9WriteFDivD : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 25; } +def A9WriteFSqrtS : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 17; } +def A9WriteFSqrtD : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 32; } + +// NEON has an odd mix of latencies. Simply name the write types by latency. +def A9WriteV1 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 1; } +def A9WriteV2 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 2; } +def A9WriteV3 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 3; } +def A9WriteV4 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 4; } +def A9WriteV5 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 5; } +def A9WriteV6 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 6; } +def A9WriteV7 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 7; } +def A9WriteV9 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 9; } +def A9WriteV10 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { let Latency = 10; } + +def : WriteRes<WriteVLD1, []>; +def : WriteRes<WriteVLD2, []>; +def : WriteRes<WriteVLD3, []>; +def : WriteRes<WriteVLD4, []>; +def : WriteRes<WriteVST1, []>; +def : WriteRes<WriteVST2, []>; +def : WriteRes<WriteVST3, []>; +def : WriteRes<WriteVST4, []>; + +// Reserve A9UnitFP for 2 consecutive cycles. +def A9Write2V4 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { + let Latency = 4; + let ResourceCycles = [2, 1]; +} +def A9Write2V7 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { + let Latency = 7; + let ResourceCycles = [2, 1]; +} +def A9Write2V9 : SchedWriteRes<[A9UnitFP, A9UnitAGU]> { + let Latency = 9; + let ResourceCycles = [2, 1]; +} + +// Branches don't have a def operand but still consume resources. +def A9WriteB : SchedWriteRes<[A9UnitB]>; + +// Address generation. +def A9WriteAdr : SchedWriteRes<[A9UnitAGU]> { let NumMicroOps = 0; } + +// Load Integer. +def A9WriteL : SchedWriteRes<[A9UnitLS]> { let Latency = 3; } +def : SchedAlias<WriteLd, A9WriteL>; +// Load the upper 32-bits using the same micro-op. +def A9WriteLHi : SchedWriteRes<[]> { let Latency = 3; + let NumMicroOps = 0; } +// Offset shifted by register. +def A9WriteLsi : SchedWriteRes<[A9UnitLS]> { let Latency = 4; } +// Load (and zero extend) a byte. +def A9WriteLb : SchedWriteRes<[A9UnitLS]> { let Latency = 4; } +def A9WriteLbsi : SchedWriteRes<[A9UnitLS]> { let Latency = 5; } + +// Load or Store Float, aligned. +def A9WriteLSfp : SchedWriteRes<[A9UnitLS, A9UnitFP]> { let Latency = 1; } + +// Store Integer. +def A9WriteS : SchedWriteRes<[A9UnitLS]>; + +//===----------------------------------------------------------------------===// +// Define resources dynamically for load multiple variants. + +// Define helpers for extra latency without consuming resources. +def A9WriteCycle1 : SchedWriteRes<[]> { let Latency = 1; let NumMicroOps = 0; } +foreach NumCycles = 2-8 in { +def A9WriteCycle#NumCycles : WriteSequence<[A9WriteCycle1], NumCycles>; +} // foreach NumCycles + +// Define address generation sequences and predicates for 8 flavors of LDMs. +foreach NumAddr = 1-8 in { + +// Define A9WriteAdr1-8 as a sequence of A9WriteAdr with additive +// latency for instructions that generate multiple loads or stores. +def A9WriteAdr#NumAddr : WriteSequence<[A9WriteAdr], NumAddr>; + +// Define a predicate to select the LDM based on number of memory addresses. +def A9LMAdr#NumAddr#Pred : + SchedPredicate<"(TII->getNumLDMAddresses(*MI)+1)/2 == "#NumAddr>; + +} // foreach NumAddr + +// Fall-back for unknown LDMs. +def A9LMUnknownPred : SchedPredicate<"TII->getNumLDMAddresses(*MI) == 0">; + +// LDM/VLDM/VLDn address generation latency & resources. +// Dynamically select the A9WriteAdrN sequence using a predicate. +def A9WriteLMAdr : SchedWriteVariant<[ + SchedVar<A9LMAdr1Pred, [A9WriteAdr1]>, + SchedVar<A9LMAdr2Pred, [A9WriteAdr2]>, + SchedVar<A9LMAdr3Pred, [A9WriteAdr3]>, + SchedVar<A9LMAdr4Pred, [A9WriteAdr4]>, + SchedVar<A9LMAdr5Pred, [A9WriteAdr5]>, + SchedVar<A9LMAdr6Pred, [A9WriteAdr6]>, + SchedVar<A9LMAdr7Pred, [A9WriteAdr7]>, + SchedVar<A9LMAdr8Pred, [A9WriteAdr8]>, + // For unknown LDM/VLDM/VSTM, assume 2 32-bit registers. + SchedVar<A9LMUnknownPred, [A9WriteAdr2]>]>; + +// Define LDM Resources. +// These take no issue resource, so they can be combined with other +// writes like WriteB. +// A9WriteLMLo takes a single LS resource and 2 cycles. +def A9WriteLMLo : SchedWriteRes<[A9UnitLS]> { let Latency = 2; + let NumMicroOps = 0; } +// Assuming aligned access, the upper half of each pair is free with +// the same latency. +def A9WriteLMHi : SchedWriteRes<[]> { let Latency = 2; + let NumMicroOps = 0; } +// Each A9WriteL#N variant adds N cycles of latency without consuming +// additional resources. +foreach NumAddr = 1-8 in { +def A9WriteL#NumAddr : WriteSequence< + [A9WriteLMLo, !cast<SchedWrite>("A9WriteCycle"#NumAddr)]>; +def A9WriteL#NumAddr#Hi : WriteSequence< + [A9WriteLMHi, !cast<SchedWrite>("A9WriteCycle"#NumAddr)]>; +} + +//===----------------------------------------------------------------------===// +// LDM: Load multiple into 32-bit integer registers. + +def A9WriteLMOpsList : A9WriteLMOpsListType< + [A9WriteL1, A9WriteL1Hi, + A9WriteL2, A9WriteL2Hi, + A9WriteL3, A9WriteL3Hi, + A9WriteL4, A9WriteL4Hi, + A9WriteL5, A9WriteL5Hi, + A9WriteL6, A9WriteL6Hi, + A9WriteL7, A9WriteL7Hi, + A9WriteL8, A9WriteL8Hi]>; + +// A9WriteLM variants expand into a pair of writes for each 64-bit +// value loaded. When the number of registers is odd, the last +// A9WriteLnHi is naturally ignored because the instruction has no +// following def operands. These variants take no issue resource, so +// they may need to be part of a WriteSequence that includes A9WriteIssue. +def A9WriteLM : SchedWriteVariant<[ + SchedVar<A9LMAdr1Pred, A9WriteLMOpsList.Writes[0-1]>, + SchedVar<A9LMAdr2Pred, A9WriteLMOpsList.Writes[0-3]>, + SchedVar<A9LMAdr3Pred, A9WriteLMOpsList.Writes[0-5]>, + SchedVar<A9LMAdr4Pred, A9WriteLMOpsList.Writes[0-7]>, + SchedVar<A9LMAdr5Pred, A9WriteLMOpsList.Writes[0-9]>, + SchedVar<A9LMAdr6Pred, A9WriteLMOpsList.Writes[0-11]>, + SchedVar<A9LMAdr7Pred, A9WriteLMOpsList.Writes[0-13]>, + SchedVar<A9LMAdr8Pred, A9WriteLMOpsList.Writes[0-15]>, + // For unknown LDMs, define the maximum number of writes, but only + // make the first two consume resources. + SchedVar<A9LMUnknownPred, [A9WriteL1, A9WriteL1Hi, + A9WriteL2, A9WriteL2Hi, + A9WriteL3Hi, A9WriteL3Hi, + A9WriteL4Hi, A9WriteL4Hi, + A9WriteL5Hi, A9WriteL5Hi, + A9WriteL6Hi, A9WriteL6Hi, + A9WriteL7Hi, A9WriteL7Hi, + A9WriteL8Hi, A9WriteL8Hi]>]> { + let Variadic = 1; +} + +//===----------------------------------------------------------------------===// +// VFP Load/Store Multiple Variants, and NEON VLDn/VSTn support. + +// A9WriteLfpOp is the same as A9WriteLSfp but takes no issue resources +// so can be used in WriteSequences for in single-issue instructions that +// encapsulate multiple loads. +def A9WriteLfpOp : SchedWriteRes<[A9UnitLS, A9UnitFP]> { + let Latency = 1; + let NumMicroOps = 0; +} + +foreach NumAddr = 1-8 in { + +// Helper for A9WriteLfp1-8: A sequence of fp loads with no micro-ops. +def A9WriteLfp#NumAddr#Seq : WriteSequence<[A9WriteLfpOp], NumAddr>; + +// A9WriteLfp1-8 definitions are statically expanded into a sequence of +// A9WriteLfpOps with additive latency that takes a single issue slot. +// Used directly to describe NEON VLDn. +def A9WriteLfp#NumAddr : WriteSequence< + [A9WriteIssue, !cast<SchedWrite>("A9WriteLfp"#NumAddr#Seq)]>; + +// A9WriteLfp1-8Mov adds a cycle of latency and FP resource for +// permuting loaded values. +def A9WriteLfp#NumAddr#Mov : WriteSequence< + [A9WriteF, !cast<SchedWrite>("A9WriteLfp"#NumAddr#Seq)]>; + +} // foreach NumAddr + +// Define VLDM/VSTM PreRA resources. +// A9WriteLMfpPreRA are dynamically expanded into the correct +// A9WriteLfp1-8 sequence based on a predicate. This supports the +// preRA VLDM variants in which all 64-bit loads are written to the +// same tuple of either single or double precision registers. +def A9WriteLMfpPreRA : SchedWriteVariant<[ + SchedVar<A9LMAdr1Pred, [A9WriteLfp1]>, + SchedVar<A9LMAdr2Pred, [A9WriteLfp2]>, + SchedVar<A9LMAdr3Pred, [A9WriteLfp3]>, + SchedVar<A9LMAdr4Pred, [A9WriteLfp4]>, + SchedVar<A9LMAdr5Pred, [A9WriteLfp5]>, + SchedVar<A9LMAdr6Pred, [A9WriteLfp6]>, + SchedVar<A9LMAdr7Pred, [A9WriteLfp7]>, + SchedVar<A9LMAdr8Pred, [A9WriteLfp8]>, + // For unknown VLDM/VSTM PreRA, assume 2xS registers. + SchedVar<A9LMUnknownPred, [A9WriteLfp2]>]>; + +// Define VLDM/VSTM PostRA Resources. +// A9WriteLMfpLo takes a LS and FP resource and one issue slot but no latency. +def A9WriteLMfpLo : SchedWriteRes<[A9UnitLS, A9UnitFP]> { let Latency = 0; } + +foreach NumAddr = 1-8 in { + +// Each A9WriteL#N variant adds N cycles of latency without consuming +// additional resources. +def A9WriteLMfp#NumAddr : WriteSequence< + [A9WriteLMfpLo, !cast<SchedWrite>("A9WriteCycle"#NumAddr)]>; + +// Assuming aligned access, the upper half of each pair is free with +// the same latency. +def A9WriteLMfp#NumAddr#Hi : WriteSequence< + [A9WriteLMHi, !cast<SchedWrite>("A9WriteCycle"#NumAddr)]>; + +} // foreach NumAddr + +// VLDM PostRA Variants. These variants expand A9WriteLMfpPostRA into a +// pair of writes for each 64-bit data loaded. When the number of +// registers is odd, the last WriteLMfpnHi is naturally ignored because +// the instruction has no following def operands. + +def A9WriteLMfpPostRAOpsList : A9WriteLMOpsListType< + [A9WriteLMfp1, A9WriteLMfp2, // 0-1 + A9WriteLMfp3, A9WriteLMfp4, // 2-3 + A9WriteLMfp5, A9WriteLMfp6, // 4-5 + A9WriteLMfp7, A9WriteLMfp8, // 6-7 + A9WriteLMfp1Hi, // 8-8 + A9WriteLMfp2Hi, A9WriteLMfp2Hi, // 9-10 + A9WriteLMfp3Hi, A9WriteLMfp3Hi, // 11-12 + A9WriteLMfp4Hi, A9WriteLMfp4Hi, // 13-14 + A9WriteLMfp5Hi, A9WriteLMfp5Hi, // 15-16 + A9WriteLMfp6Hi, A9WriteLMfp6Hi, // 17-18 + A9WriteLMfp7Hi, A9WriteLMfp7Hi, // 19-20 + A9WriteLMfp8Hi, A9WriteLMfp8Hi]>; // 21-22 + +def A9WriteLMfpPostRA : SchedWriteVariant<[ + SchedVar<A9LMAdr1Pred, A9WriteLMfpPostRAOpsList.Writes[0-0, 8-8]>, + SchedVar<A9LMAdr2Pred, A9WriteLMfpPostRAOpsList.Writes[0-1, 9-10]>, + SchedVar<A9LMAdr3Pred, A9WriteLMfpPostRAOpsList.Writes[0-2, 10-12]>, + SchedVar<A9LMAdr4Pred, A9WriteLMfpPostRAOpsList.Writes[0-3, 11-14]>, + SchedVar<A9LMAdr5Pred, A9WriteLMfpPostRAOpsList.Writes[0-4, 12-16]>, + SchedVar<A9LMAdr6Pred, A9WriteLMfpPostRAOpsList.Writes[0-5, 13-18]>, + SchedVar<A9LMAdr7Pred, A9WriteLMfpPostRAOpsList.Writes[0-6, 14-20]>, + SchedVar<A9LMAdr8Pred, A9WriteLMfpPostRAOpsList.Writes[0-7, 15-22]>, + // For unknown LDMs, define the maximum number of writes, but only + // make the first two consume resources. We are optimizing for the case + // where the operands are DPRs, and this determines the first eight + // types. The remaining eight types are filled to cover the case + // where the operands are SPRs. + SchedVar<A9LMUnknownPred, [A9WriteLMfp1, A9WriteLMfp2, + A9WriteLMfp3Hi, A9WriteLMfp4Hi, + A9WriteLMfp5Hi, A9WriteLMfp6Hi, + A9WriteLMfp7Hi, A9WriteLMfp8Hi, + A9WriteLMfp5Hi, A9WriteLMfp5Hi, + A9WriteLMfp6Hi, A9WriteLMfp6Hi, + A9WriteLMfp7Hi, A9WriteLMfp7Hi, + A9WriteLMfp8Hi, A9WriteLMfp8Hi]>]> { + let Variadic = 1; +} + +// Distinguish between our multiple MI-level forms of the same +// VLDM/VSTM instructions. +def A9PreRA : SchedPredicate< + "TargetRegisterInfo::isVirtualRegister(MI->getOperand(0).getReg())">; +def A9PostRA : SchedPredicate< + "TargetRegisterInfo::isPhysicalRegister(MI->getOperand(0).getReg())">; + +// VLDM represents all destination registers as a single register +// tuple, unlike LDM. So the number of write operands is not variadic. +def A9WriteLMfp : SchedWriteVariant<[ + SchedVar<A9PreRA, [A9WriteLMfpPreRA]>, + SchedVar<A9PostRA, [A9WriteLMfpPostRA]>]>; + +//===----------------------------------------------------------------------===// +// Resources for other (non-LDM/VLDM) Variants. + +// These mov immediate writers are unconditionally expanded with +// additive latency. +def A9WriteI2 : WriteSequence<[A9WriteI, A9WriteI]>; +def A9WriteI2pc : WriteSequence<[A9WriteI, A9WriteI, WriteALU]>; +def A9WriteI2ld : WriteSequence<[A9WriteI, A9WriteI, A9WriteL]>; + +// Some ALU operations can read loaded integer values one cycle early. +def A9ReadALU : SchedReadAdvance<1, + [A9WriteL, A9WriteLHi, A9WriteLsi, A9WriteLb, A9WriteLbsi, + A9WriteL1, A9WriteL2, A9WriteL3, A9WriteL4, + A9WriteL5, A9WriteL6, A9WriteL7, A9WriteL8, + A9WriteL1Hi, A9WriteL2Hi, A9WriteL3Hi, A9WriteL4Hi, + A9WriteL5Hi, A9WriteL6Hi, A9WriteL7Hi, A9WriteL8Hi]>; + +// Read types for operands that are unconditionally read in cycle N +// after the instruction issues, decreases producer latency by N-1. +def A9Read2 : SchedReadAdvance<1>; +def A9Read3 : SchedReadAdvance<2>; +def A9Read4 : SchedReadAdvance<3>; + +//===----------------------------------------------------------------------===// +// Map itinerary classes to scheduler read/write resources per operand. +// +// For ARM, we piggyback scheduler resources on the Itinerary classes +// to avoid perturbing the existing instruction definitions. + +// This table follows the ARM Cortex-A9 Technical Reference Manuals, +// mostly in order. + +def :ItinRW<[WriteALU], [IIC_iMOVi,IIC_iMOVr,IIC_iMOVsi, + IIC_iMVNi,IIC_iMVNsi, + IIC_iCMOVi,IIC_iCMOVr,IIC_iCMOVsi]>; +def :ItinRW<[WriteALU, A9ReadALU],[IIC_iMVNr]>; +def :ItinRW<[A9WriteIsr], [IIC_iMOVsr,IIC_iMVNsr,IIC_iCMOVsr]>; + +def :ItinRW<[A9WriteI2], [IIC_iMOVix2,IIC_iCMOVix2]>; +def :ItinRW<[A9WriteI2pc], [IIC_iMOVix2addpc]>; +def :ItinRW<[A9WriteI2ld], [IIC_iMOVix2ld]>; + +def :ItinRW<[WriteALU], [IIC_iBITi,IIC_iBITr,IIC_iUNAr,IIC_iTSTi,IIC_iTSTr]>; +def :ItinRW<[WriteALU, A9ReadALU], [IIC_iALUi, IIC_iCMPi, IIC_iCMPsi]>; +def :ItinRW<[WriteALU, A9ReadALU, A9ReadALU],[IIC_iALUr,IIC_iCMPr]>; +def :ItinRW<[WriteALUsi], [IIC_iBITsi,IIC_iUNAsi,IIC_iEXTr,IIC_iTSTsi]>; +def :ItinRW<[WriteALUsi, A9ReadALU], [IIC_iALUsi]>; +def :ItinRW<[WriteALUsi, ReadDefault, A9ReadALU], [IIC_iALUsir]>; // RSB +def :ItinRW<[A9WriteALUsr], [IIC_iBITsr,IIC_iTSTsr,IIC_iEXTAr,IIC_iEXTAsr]>; +def :ItinRW<[A9WriteALUsr, A9ReadALU], [IIC_iALUsr,IIC_iCMPsr]>; + +// A9WriteHi ignored for MUL32. +def :ItinRW<[A9WriteM, A9WriteMHi], [IIC_iMUL32,IIC_iMAC32, + IIC_iMUL64,IIC_iMAC64]>; +// FIXME: SMLALxx needs itin classes +def :ItinRW<[A9WriteM16, A9WriteM16Hi], [IIC_iMUL16,IIC_iMAC16]>; + +// TODO: For floating-point ops, we model the pipeline forwarding +// latencies here. WAW latencies are sometimes longer. + +def :ItinRW<[A9WriteFMov], [IIC_fpSTAT, IIC_fpMOVIS, IIC_fpMOVID, IIC_fpMOVSI, + IIC_fpUNA32, IIC_fpUNA64, + IIC_fpCMP32, IIC_fpCMP64]>; +def :ItinRW<[A9WriteFMov, A9WriteFMov], [IIC_fpMOVDI]>; +def :ItinRW<[A9WriteF], [IIC_fpCVTSD, IIC_fpCVTDS, IIC_fpCVTSH, IIC_fpCVTHS, + IIC_fpCVTIS, IIC_fpCVTID, IIC_fpCVTSI, IIC_fpCVTDI, + IIC_fpALU32, IIC_fpALU64]>; +def :ItinRW<[A9WriteFMulS], [IIC_fpMUL32]>; +def :ItinRW<[A9WriteFMulD], [IIC_fpMUL64]>; +def :ItinRW<[A9WriteFMAS], [IIC_fpMAC32]>; +def :ItinRW<[A9WriteFMAD], [IIC_fpMAC64]>; +def :ItinRW<[A9WriteFDivS], [IIC_fpDIV32]>; +def :ItinRW<[A9WriteFDivD], [IIC_fpDIV64]>; +def :ItinRW<[A9WriteFSqrtS], [IIC_fpSQRT32]>; +def :ItinRW<[A9WriteFSqrtD], [IIC_fpSQRT64]>; + +def :ItinRW<[A9WriteB], [IIC_Br]>; + +// A9 PLD is processed in a dedicated unit. +def :ItinRW<[], [IIC_Preload]>; + +// Note: We must assume that loads are aligned, since the machine +// model cannot know this statically and A9 ignores alignment hints. + +// A9WriteAdr consumes AGU regardless address writeback. But it's +// latency is only relevant for users of an updated address. +def :ItinRW<[A9WriteL, A9WriteAdr], [IIC_iLoad_i,IIC_iLoad_r, + IIC_iLoad_iu,IIC_iLoad_ru]>; +def :ItinRW<[A9WriteLsi, A9WriteAdr], [IIC_iLoad_si,IIC_iLoad_siu]>; +def :ItinRW<[A9WriteLb, A9WriteAdr2], [IIC_iLoad_bh_i,IIC_iLoad_bh_r, + IIC_iLoad_bh_iu,IIC_iLoad_bh_ru]>; +def :ItinRW<[A9WriteLbsi, A9WriteAdr2], [IIC_iLoad_bh_si,IIC_iLoad_bh_siu]>; +def :ItinRW<[A9WriteL, A9WriteLHi, A9WriteAdr], [IIC_iLoad_d_i,IIC_iLoad_d_r, + IIC_iLoad_d_ru]>; +// Store either has no def operands, or the one def for address writeback. +def :ItinRW<[A9WriteAdr, A9WriteS], [IIC_iStore_i, IIC_iStore_r, + IIC_iStore_iu, IIC_iStore_ru, + IIC_iStore_d_i, IIC_iStore_d_r, + IIC_iStore_d_ru]>; +def :ItinRW<[A9WriteAdr2, A9WriteS], [IIC_iStore_si, IIC_iStore_siu, + IIC_iStore_bh_i, IIC_iStore_bh_r, + IIC_iStore_bh_iu, IIC_iStore_bh_ru]>; +def :ItinRW<[A9WriteAdr3, A9WriteS], [IIC_iStore_bh_si, IIC_iStore_bh_siu]>; + +// A9WriteML will be expanded into a separate write for each def +// operand. Address generation consumes resources, but A9WriteLMAdr +// is listed after all def operands, so has no effective latency. +// +// Note: A9WriteLM expands into an even number of def operands. The +// actual number of def operands may be less by one. +def :ItinRW<[A9WriteLM, A9WriteLMAdr, A9WriteIssue], [IIC_iLoad_m, IIC_iPop]>; + +// Load multiple with address writeback has an extra def operand in +// front of the loaded registers. +// +// Reuse the load-multiple variants for store-multiple because the +// resources are identical, For stores only the address writeback +// has a def operand so the WriteL latencies are unused. +def :ItinRW<[A9WriteLMAdr, A9WriteLM, A9WriteIssue], [IIC_iLoad_mu, + IIC_iStore_m, + IIC_iStore_mu]>; +def :ItinRW<[A9WriteLM, A9WriteLMAdr, A9WriteB], [IIC_iLoad_mBr, IIC_iPop_Br]>; +def :ItinRW<[A9WriteL, A9WriteAdr, WriteALU], [IIC_iLoadiALU]>; + +def :ItinRW<[A9WriteLSfp, A9WriteAdr], [IIC_fpLoad32, IIC_fpLoad64]>; + +def :ItinRW<[A9WriteLMfp, A9WriteLMAdr], [IIC_fpLoad_m]>; +def :ItinRW<[A9WriteLMAdr, A9WriteLMfp], [IIC_fpLoad_mu]>; +def :ItinRW<[A9WriteAdr, A9WriteLSfp], [IIC_fpStore32, IIC_fpStore64, + IIC_fpStore_m, IIC_fpStore_mu]>; + +// Note: Unlike VLDM, VLD1 expects the writeback operand after the +// normal writes. +def :ItinRW<[A9WriteLfp1, A9WriteAdr1], [IIC_VLD1, IIC_VLD1u, + IIC_VLD1x2, IIC_VLD1x2u]>; +def :ItinRW<[A9WriteLfp2, A9WriteAdr2], [IIC_VLD1x3, IIC_VLD1x3u, + IIC_VLD1x4, IIC_VLD1x4u, + IIC_VLD4dup, IIC_VLD4dupu]>; +def :ItinRW<[A9WriteLfp1Mov, A9WriteAdr1], [IIC_VLD1dup, IIC_VLD1dupu, + IIC_VLD2, IIC_VLD2u, + IIC_VLD2dup, IIC_VLD2dupu]>; +def :ItinRW<[A9WriteLfp2Mov, A9WriteAdr1], [IIC_VLD1ln, IIC_VLD1lnu, + IIC_VLD2x2, IIC_VLD2x2u, + IIC_VLD2ln, IIC_VLD2lnu]>; +def :ItinRW<[A9WriteLfp3Mov, A9WriteAdr3], [IIC_VLD3, IIC_VLD3u, + IIC_VLD3dup, IIC_VLD3dupu]>; +def :ItinRW<[A9WriteLfp4Mov, A9WriteAdr4], [IIC_VLD4, IIC_VLD4u, + IIC_VLD4ln, IIC_VLD4lnu]>; +def :ItinRW<[A9WriteLfp5Mov, A9WriteAdr5], [IIC_VLD3ln, IIC_VLD3lnu]>; + +// Vector stores use similar resources to vector loads, so use the +// same write types. The address write must be first for stores with +// address writeback. +def :ItinRW<[A9WriteAdr1, A9WriteLfp1], [IIC_VST1, IIC_VST1u, + IIC_VST1x2, IIC_VST1x2u, + IIC_VST1ln, IIC_VST1lnu, + IIC_VST2, IIC_VST2u, + IIC_VST2x2, IIC_VST2x2u, + IIC_VST2ln, IIC_VST2lnu]>; +def :ItinRW<[A9WriteAdr2, A9WriteLfp2], [IIC_VST1x3, IIC_VST1x3u, + IIC_VST1x4, IIC_VST1x4u, + IIC_VST3, IIC_VST3u, + IIC_VST3ln, IIC_VST3lnu, + IIC_VST4, IIC_VST4u, + IIC_VST4ln, IIC_VST4lnu]>; + +// NEON moves. +def :ItinRW<[A9WriteV2], [IIC_VMOVSI, IIC_VMOVDI, IIC_VMOVD, IIC_VMOVQ]>; +def :ItinRW<[A9WriteV1], [IIC_VMOV, IIC_VMOVIS, IIC_VMOVID]>; +def :ItinRW<[A9WriteV3], [IIC_VMOVISL, IIC_VMOVN]>; + +// NEON integer arithmetic +// +// VADD/VAND/VORR/VEOR/VBIC/VORN/VBIT/VBIF/VBSL +def :ItinRW<[A9WriteV3, A9Read2, A9Read2], [IIC_VBINiD, IIC_VBINiQ]>; +// VSUB/VMVN/VCLSD/VCLZD/VCNTD +def :ItinRW<[A9WriteV3, A9Read2], [IIC_VSUBiD, IIC_VSUBiQ, IIC_VCNTiD]>; +// VADDL/VSUBL/VNEG are mapped later under IIC_SHLi. +// ... +// VHADD/VRHADD/VQADD/VTST/VADH/VRADH +def :ItinRW<[A9WriteV4, A9Read2, A9Read2], [IIC_VBINi4D, IIC_VBINi4Q]>; + +// VSBH/VRSBH/VHSUB/VQSUB/VABD/VCEQ/VCGE/VCGT/VMAX/VMIN/VPMAX/VPMIN/VABDL +def :ItinRW<[A9WriteV4, A9Read2], [IIC_VSUBi4D, IIC_VSUBi4Q]>; +// VQNEG/VQABS +def :ItinRW<[A9WriteV4], [IIC_VQUNAiD, IIC_VQUNAiQ]>; +// VABS +def :ItinRW<[A9WriteV4, A9Read2], [IIC_VUNAiD, IIC_VUNAiQ]>; +// VPADD/VPADDL are mapped later under IIC_SHLi. +// ... +// VCLSQ/VCLZQ/VCNTQ, takes two cycles. +def :ItinRW<[A9Write2V4, A9Read3], [IIC_VCNTiQ]>; +// VMOVimm/VMVNimm/VORRimm/VBICimm +def :ItinRW<[A9WriteV3], [IIC_VMOVImm]>; +def :ItinRW<[A9WriteV6, A9Read3, A9Read2], [IIC_VABAD, IIC_VABAQ]>; +def :ItinRW<[A9WriteV6, A9Read3], [IIC_VPALiD, IIC_VPALiQ]>; + +// NEON integer multiply +// +// Note: these don't quite match the timing docs, but they do match +// the original A9 itinerary. +def :ItinRW<[A9WriteV6, A9Read2, A9Read2], [IIC_VMULi16D]>; +def :ItinRW<[A9WriteV7, A9Read2, A9Read2], [IIC_VMULi16Q]>; +def :ItinRW<[A9Write2V7, A9Read2], [IIC_VMULi32D]>; +def :ItinRW<[A9Write2V9, A9Read2], [IIC_VMULi32Q]>; +def :ItinRW<[A9WriteV6, A9Read3, A9Read2, A9Read2], [IIC_VMACi16D]>; +def :ItinRW<[A9WriteV7, A9Read3, A9Read2, A9Read2], [IIC_VMACi16Q]>; +def :ItinRW<[A9Write2V7, A9Read3, A9Read2], [IIC_VMACi32D]>; +def :ItinRW<[A9Write2V9, A9Read3, A9Read2], [IIC_VMACi32Q]>; + +// NEON integer shift +// TODO: Q,Q,Q shifts should actually reserve FP for 2 cycles. +def :ItinRW<[A9WriteV3], [IIC_VSHLiD, IIC_VSHLiQ]>; +def :ItinRW<[A9WriteV4], [IIC_VSHLi4D, IIC_VSHLi4Q]>; + +// NEON permute +def :ItinRW<[A9WriteV2, A9WriteV2], [IIC_VPERMD, IIC_VPERMQ, IIC_VEXTD]>; +def :ItinRW<[A9WriteV3, A9WriteV4, ReadDefault, A9Read2], + [IIC_VPERMQ3, IIC_VEXTQ]>; +def :ItinRW<[A9WriteV3, A9Read2], [IIC_VTB1]>; +def :ItinRW<[A9WriteV3, A9Read2, A9Read2], [IIC_VTB2]>; +def :ItinRW<[A9WriteV4, A9Read2, A9Read2, A9Read3], [IIC_VTB3]>; +def :ItinRW<[A9WriteV4, A9Read2, A9Read2, A9Read3, A9Read3], [IIC_VTB4]>; +def :ItinRW<[A9WriteV3, ReadDefault, A9Read2], [IIC_VTBX1]>; +def :ItinRW<[A9WriteV3, ReadDefault, A9Read2, A9Read2], [IIC_VTBX2]>; +def :ItinRW<[A9WriteV4, ReadDefault, A9Read2, A9Read2, A9Read3], [IIC_VTBX3]>; +def :ItinRW<[A9WriteV4, ReadDefault, A9Read2, A9Read2, A9Read3, A9Read3], + [IIC_VTBX4]>; + +// NEON floating-point +def :ItinRW<[A9WriteV5, A9Read2, A9Read2], [IIC_VBIND]>; +def :ItinRW<[A9WriteV6, A9Read2, A9Read2], [IIC_VBINQ]>; +def :ItinRW<[A9WriteV5, A9Read2], [IIC_VUNAD, IIC_VFMULD]>; +def :ItinRW<[A9WriteV6, A9Read2], [IIC_VUNAQ, IIC_VFMULQ]>; +def :ItinRW<[A9WriteV9, A9Read3, A9Read2], [IIC_VMACD, IIC_VFMACD]>; +def :ItinRW<[A9WriteV10, A9Read3, A9Read2], [IIC_VMACQ, IIC_VFMACQ]>; +def :ItinRW<[A9WriteV9, A9Read2, A9Read2], [IIC_VRECSD]>; +def :ItinRW<[A9WriteV10, A9Read2, A9Read2], [IIC_VRECSQ]>; + +// Map SchedRWs that are identical for cortexa9 to existing resources. +def : SchedAlias<WriteALU, A9WriteALU>; +def : SchedAlias<WriteALUsr, A9WriteALUsr>; +def : SchedAlias<WriteALUSsr, A9WriteALUsr>; +def : SchedAlias<ReadALU, A9ReadALU>; +def : SchedAlias<ReadALUsr, A9ReadALU>; +def : SchedAlias<WriteST, A9WriteS>; + +// ===---------------------------------------------------------------------===// +// Floating-point. Map target defined SchedReadWrite to processor specific ones +// +def : WriteRes<WriteFPCVT, [A9UnitFP, A9UnitAGU]> { let Latency = 4; } +def : SchedAlias<WriteFPMOV, A9WriteFMov>; + +def : SchedAlias<WriteFPALU32, A9WriteF>; +def : SchedAlias<WriteFPALU64, A9WriteF>; + +def : SchedAlias<WriteFPMUL32, A9WriteFMulS>; +def : SchedAlias<WriteFPMUL64, A9WriteFMulD>; + +def : SchedAlias<WriteFPMAC32, A9WriteFMAS>; +def : SchedAlias<WriteFPMAC64, A9WriteFMAD>; + +def : SchedAlias<WriteFPDIV32, A9WriteFDivS>; +def : SchedAlias<WriteFPDIV64, A9WriteFDivD>; +def : SchedAlias<WriteFPSQRT32, A9WriteFSqrtS>; +def : SchedAlias<WriteFPSQRT64, A9WriteFSqrtD>; + +def : ReadAdvance<ReadFPMUL, 0>; +def : ReadAdvance<ReadFPMAC, 0>; + +// ===---------------------------------------------------------------------===// +// Subtarget-specific overrides. Map opcodes to list of SchedReadWrite types. +// +def : InstRW< [WriteALU], + (instregex "ANDri", "ORRri", "EORri", "BICri", "ANDrr", "ORRrr", "EORrr", + "BICrr")>; +def : InstRW< [WriteALUsi], (instregex "ANDrsi", "ORRrsi", "EORrsi", "BICrsi")>; +def : InstRW< [WriteALUsr], (instregex "ANDrsr", "ORRrsr", "EORrsr", "BICrsr")>; + + +def : SchedAlias<WriteCMP, A9WriteALU>; +def : SchedAlias<WriteCMPsi, A9WriteALU>; +def : SchedAlias<WriteCMPsr, A9WriteALU>; + +def : InstRW< [A9WriteIsr], (instregex "MOVsr", "MOVsi", "MVNsr", "MOVCCsi", + "MOVCCsr")>; +def : InstRW< [WriteALU, A9ReadALU], (instregex "MVNr")>; +def : InstRW< [A9WriteI2], (instregex "MOVCCi32imm", "MOVi32imm")>; +def : InstRW< [A9WriteI2pc], (instregex "MOV_ga_pcrel")>; +def : InstRW< [A9WriteI2ld], (instregex "MOV_ga_pcrel_ldr")>; + +def : InstRW< [WriteALU], (instregex "SEL")>; + +def : InstRW< [WriteALUsi], (instregex "BFC", "BFI", "UBFX", "SBFX")>; + +def : InstRW< [A9WriteM], + (instregex "MUL", "MULv5", "SMMUL", "SMMULR", "MLA", "MLAv5", "MLS", + "SMMLA", "SMMLAR", "SMMLS", "SMMLSR")>; +def : InstRW< [A9WriteM, A9WriteMHi], + (instregex "SMULL", "SMULLv5", "UMULL", "UMULLv5", "SMLAL$", "UMLAL", + "UMAAL", "SMLALv5", "UMLALv5", "SMLALBB", "SMLALBT", "SMLALTB", + "SMLALTT")>; +// FIXME: These instructions used to have NoItinerary. Just copied the one from above. +def : InstRW< [A9WriteM, A9WriteMHi], + (instregex "SMLAD", "SMLADX", "SMLALD", "SMLALDX", "SMLSD", "SMLSDX", + "SMLSLD", "SMLSLDX", "SMUAD", "SMUADX", "SMUSD", "SMUSDX")>; + +def : InstRW<[A9WriteM16, A9WriteM16Hi], + (instregex "SMULBB", "SMULBT", "SMULTB", "SMULTT", "SMULWB", "SMULWT")>; +def : InstRW<[A9WriteM16, A9WriteM16Hi], + (instregex "SMLABB", "SMLABT", "SMLATB", "SMLATT", "SMLAWB", "SMLAWT")>; + +def : InstRW<[A9WriteL], (instregex "LDRi12", "PICLDR$")>; +def : InstRW<[A9WriteLsi], (instregex "LDRrs")>; +def : InstRW<[A9WriteLb], + (instregex "LDRBi12", "PICLDRH", "PICLDRB", "PICLDRSH", "PICLDRSB", + "LDRH", "LDRSH", "LDRSB")>; +def : InstRW<[A9WriteLbsi], (instregex "LDRrs")>; + +def : WriteRes<WriteDIV, []> { let Latency = 0; } + +def : WriteRes<WriteBr, [A9UnitB]>; +def : WriteRes<WriteBrL, [A9UnitB]>; +def : WriteRes<WriteBrTbl, [A9UnitB]>; +def : WriteRes<WritePreLd, []>; +def : WriteRes<WriteNoop, []> { let Latency = 0; let NumMicroOps = 0; } +} // SchedModel = CortexA9Model diff --git a/capstone/suite/synctools/tablegen/ARM/ARMScheduleM3.td b/capstone/suite/synctools/tablegen/ARM/ARMScheduleM3.td new file mode 100644 index 000000000..93f8299f9 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMScheduleM3.td @@ -0,0 +1,21 @@ +//=- ARMScheduleM3.td - ARM Cortex-M3 Scheduling Definitions -*- tablegen -*-=// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the machine model for the ARM Cortex-M3 processor. +// +//===----------------------------------------------------------------------===// + +def CortexM3Model : SchedMachineModel { + let IssueWidth = 1; // Only IT can be dual-issued, so assume single-issue + let MicroOpBufferSize = 0; // In-order + let LoadLatency = 2; // Latency when not pipelined, not pc-relative + let MispredictPenalty = 2; // Best case branch taken cost + + let CompleteModel = 0; +} diff --git a/capstone/suite/synctools/tablegen/ARM/ARMScheduleR52.td b/capstone/suite/synctools/tablegen/ARM/ARMScheduleR52.td new file mode 100644 index 000000000..11bce4516 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMScheduleR52.td @@ -0,0 +1,928 @@ +//==- ARMScheduleR52.td - Cortex-R52 Scheduling Definitions -*- tablegen -*-=// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the SchedRead/Write data for the ARM Cortex-R52 processor. +// +//===----------------------------------------------------------------------===// + +// ===---------------------------------------------------------------------===// +// The Cortex-R52 is an in-order pipelined superscalar microprocessor with +// a 8 stage pipeline. It can issue maximum two instructions in each cycle. +// There are two ALUs, one LDST, one MUL and a non-pipelined integer DIV. +// A number of forwarding paths enable results of computations to be input +// to subsequent operations before they are written to registers. +// This scheduler is a MachineScheduler. See TargetSchedule.td for details. + +def CortexR52Model : SchedMachineModel { + let MicroOpBufferSize = 0; // R52 is in-order processor + let IssueWidth = 2; // 2 micro-ops dispatched per cycle + let LoadLatency = 1; // Optimistic, assuming no misses + let MispredictPenalty = 8; // A branch direction mispredict, including PFU + let CompleteModel = 0; // Covers instructions applicable to cortex-r52. +} + + +//===----------------------------------------------------------------------===// +// Define each kind of processor resource and number available. + +// Modeling each pipeline as a ProcResource using the BufferSize = 0 since +// Cortex-R52 is an in-order processor. + +def R52UnitALU : ProcResource<2> { let BufferSize = 0; } // Int ALU +def R52UnitMAC : ProcResource<1> { let BufferSize = 0; } // Int MAC +def R52UnitDiv : ProcResource<1> { let BufferSize = 0; } // Int Division +def R52UnitLd : ProcResource<1> { let BufferSize = 0; } // Load/Store +def R52UnitB : ProcResource<1> { let BufferSize = 0; } // Branch +def R52UnitFPALU : ProcResource<2> { let BufferSize = 0; } // FP ALU +def R52UnitFPMUL : ProcResource<2> { let BufferSize = 0; } // FP MUL +def R52UnitFPDIV : ProcResource<1> { let BufferSize = 0; } // FP DIV + +// Cortex-R52 specific SchedReads +def R52Read_ISS : SchedRead; +def R52Read_EX1 : SchedRead; +def R52Read_EX2 : SchedRead; +def R52Read_WRI : SchedRead; +def R52Read_F0 : SchedRead; // F0 maps to ISS stage of integer pipe +def R52Read_F1 : SchedRead; +def R52Read_F2 : SchedRead; + + +//===----------------------------------------------------------------------===// +// Subtarget-specific SchedWrite types which map ProcResources and set latency. + +let SchedModel = CortexR52Model in { + +// ALU - Write occurs in Late EX2 (independent of whether shift was required) +def : WriteRes<WriteALU, [R52UnitALU]> { let Latency = 3; } +def : WriteRes<WriteALUsi, [R52UnitALU]> { let Latency = 3; } +def : WriteRes<WriteALUsr, [R52UnitALU]> { let Latency = 3; } +def : WriteRes<WriteALUSsr, [R52UnitALU]> { let Latency = 3; } + +// Compares +def : WriteRes<WriteCMP, [R52UnitALU]> { let Latency = 0; } +def : WriteRes<WriteCMPsi, [R52UnitALU]> { let Latency = 0; } +def : WriteRes<WriteCMPsr, [R52UnitALU]> { let Latency = 0; } + +// Multiply - aliased to sub-target specific later + +// Div - may stall 0-9 cycles depending on input (i.e. WRI+(0-9)/2) +def : WriteRes<WriteDIV, [R52UnitDiv]> { + let Latency = 8; let ResourceCycles = [8]; // non-pipelined +} + +// Branches - LR written in Late EX2 +def : WriteRes<WriteBr, [R52UnitB]> { let Latency = 0; } +def : WriteRes<WriteBrL, [R52UnitB]> { let Latency = 0; } +def : WriteRes<WriteBrTbl, [R52UnitALU]> { let Latency = 0; } + +// Misc +def : WriteRes<WriteNoop, []> { let Latency = 0; let NumMicroOps = 0; } + +// Integer pipeline by-passes +def : ReadAdvance<ReadALU, 1>; // Operand needed in EX1 stage +def : ReadAdvance<ReadALUsr, 0>; // Shift operands needed in ISS +def : ReadAdvance<ReadMUL, 0>; +def : ReadAdvance<ReadMAC, 0>; + +// Floating-point. Map target-defined SchedReadWrites to subtarget +def : WriteRes<WriteFPMUL32, [R52UnitFPMUL]> { let Latency = 6; } + +def : WriteRes<WriteFPMUL64, [R52UnitFPMUL, R52UnitFPMUL]> { + let Latency = 6; +} + +def : WriteRes<WriteFPMAC32, [R52UnitFPMUL, R52UnitFPALU]> { + let Latency = 11; // as it is internally two insns (MUL then ADD) +} + +def : WriteRes<WriteFPMAC64, [R52UnitFPMUL, R52UnitFPMUL, + R52UnitFPALU, R52UnitFPALU]> { + let Latency = 11; +} + +def : WriteRes<WriteFPDIV32, [R52UnitDiv]> { + let Latency = 7; // FP div takes fixed #cycles + let ResourceCycles = [7]; // is not pipelined +} + +def : WriteRes<WriteFPDIV64, [R52UnitDiv]> { + let Latency = 17; + let ResourceCycles = [17]; +} + +def : WriteRes<WriteFPSQRT32, [R52UnitDiv]> { let Latency = 7; } +def : WriteRes<WriteFPSQRT64, [R52UnitDiv]> { let Latency = 17; } + +// Overriden via InstRW for this processor. +def : WriteRes<WriteVST1, []>; +def : WriteRes<WriteVST2, []>; +def : WriteRes<WriteVST3, []>; +def : WriteRes<WriteVST4, []>; + +def : ReadAdvance<ReadFPMUL, 1>; // mul operand read in F1 +def : ReadAdvance<ReadFPMAC, 1>; // fp-mac operand read in F1 + +//===----------------------------------------------------------------------===// +// Subtarget-specific SchedReadWrites. + +// Forwarding information - based on when an operand is read +def : ReadAdvance<R52Read_ISS, 0>; +def : ReadAdvance<R52Read_EX1, 1>; +def : ReadAdvance<R52Read_EX2, 2>; +def : ReadAdvance<R52Read_F0, 0>; +def : ReadAdvance<R52Read_F1, 1>; +def : ReadAdvance<R52Read_F2, 2>; + + +// Cortex-R52 specific SchedWrites for use with InstRW +def R52WriteMAC : SchedWriteRes<[R52UnitMAC]> { let Latency = 4; } +def R52WriteMACHi : SchedWriteRes<[R52UnitMAC]> { + let Latency = 4; let NumMicroOps = 0; +} +def R52WriteDIV : SchedWriteRes<[R52UnitDiv]> { + let Latency = 8; let ResourceCycles = [8]; // not pipelined +} +def R52WriteLd : SchedWriteRes<[R52UnitLd]> { let Latency = 4; } +def R52WriteST : SchedWriteRes<[R52UnitLd]> { let Latency = 4; } +def R52WriteAdr : SchedWriteRes<[]> { let Latency = 0; } +def R52WriteCC : SchedWriteRes<[]> { let Latency = 0; } +def R52WriteALU_EX1 : SchedWriteRes<[R52UnitALU]> { let Latency = 2; } +def R52WriteALU_EX2 : SchedWriteRes<[R52UnitALU]> { let Latency = 3; } +def R52WriteALU_WRI : SchedWriteRes<[R52UnitALU]> { let Latency = 4; } + +def R52WriteNoRSRC_EX2 : SchedWriteRes<[]> { let Latency = 3; } +def R52WriteNoRSRC_WRI : SchedWriteRes<[]> { let Latency = 4; } + +// Alias generics to sub-target specific +def : SchedAlias<WriteMUL16, R52WriteMAC>; +def : SchedAlias<WriteMUL32, R52WriteMAC>; +def : SchedAlias<WriteMUL64Lo, R52WriteMAC>; +def : SchedAlias<WriteMUL64Hi, R52WriteMACHi>; +def : SchedAlias<WriteMAC16, R52WriteMAC>; +def : SchedAlias<WriteMAC32, R52WriteMAC>; +def : SchedAlias<WriteMAC64Lo, R52WriteMAC>; +def : SchedAlias<WriteMAC64Hi, R52WriteMACHi>; +def : SchedAlias<WritePreLd, R52WriteLd>; +def : SchedAlias<WriteLd, R52WriteLd>; +def : SchedAlias<WriteST, R52WriteST>; + +def R52WriteFPALU_F3 : SchedWriteRes<[R52UnitFPALU]> { let Latency = 4; } +def R52Write2FPALU_F3 : SchedWriteRes<[R52UnitFPALU, R52UnitFPALU]> { + let Latency = 4; +} +def R52WriteFPALU_F4 : SchedWriteRes<[R52UnitFPALU]> { let Latency = 5; } +def R52Write2FPALU_F4 : SchedWriteRes<[R52UnitFPALU, R52UnitFPALU]> { + let Latency = 5; +} +def R52WriteFPALU_F5 : SchedWriteRes<[R52UnitFPALU]> { let Latency = 6; } +def R52Write2FPALU_F5 : SchedWriteRes<[R52UnitFPALU, R52UnitFPALU]> { + let Latency = 6; +} +def R52WriteFPMUL_F5 : SchedWriteRes<[R52UnitFPMUL]> { let Latency = 6; } +def R52Write2FPMUL_F5 : SchedWriteRes<[R52UnitFPMUL, R52UnitFPMUL]> { + let Latency = 6; +} +def R52WriteFPMAC_F5 : SchedWriteRes<[R52UnitFPMUL, R52UnitFPALU]> { + let Latency = 11; // as it is internally two insns (MUL then ADD) +} +def R52Write2FPMAC_F5 : SchedWriteRes<[R52UnitFPMUL, R52UnitFPMUL, + R52UnitFPALU, R52UnitFPALU]> { + let Latency = 11; +} + +def R52WriteFPLd_F4 : SchedWriteRes<[R52UnitLd]> { let Latency = 5; } +def R52WriteFPST_F4 : SchedWriteRes<[R52UnitLd]> { let Latency = 5; } + +//===----------------------------------------------------------------------===// +// Floating-point. Map target defined SchedReadWrites to processor specific ones +// +def : SchedAlias<WriteFPCVT, R52WriteFPALU_F5>; +def : SchedAlias<WriteFPMOV, R52WriteFPALU_F3>; +def : SchedAlias<WriteFPALU32, R52WriteFPALU_F5>; +def : SchedAlias<WriteFPALU64, R52WriteFPALU_F5>; + +//===----------------------------------------------------------------------===// +// Subtarget-specific overrides. Map opcodes to list of SchedReadWrites types. +// +def : InstRW<[WriteALU], (instrs COPY)>; + +def : InstRW<[R52WriteALU_EX2, R52Read_EX1, R52Read_ISS], + (instregex "SXTB", "SXTH", "SXTB16", "UXTB", "UXTH", "UXTB16", + "t2SXTB", "t2SXTH", "t2SXTB16", "t2UXTB", "t2UXTH", "t2UXTB16")>; + +def : InstRW<[R52WriteALU_EX1, R52Read_ISS], + (instregex "MOVCCi32imm", "MOVi32imm", "t2MOVCCi", "t2MOVi")>; +def : InstRW<[R52WriteALU_EX2, R52Read_EX1], + (instregex "MOV_ga_pcrel$")>; +def : InstRW<[R52WriteLd,R52Read_ISS], + (instregex "MOV_ga_pcrel_ldr")>; + +def : InstRW<[R52WriteALU_EX2, R52Read_EX1, R52Read_EX1], (instregex "SEL", "t2SEL")>; + +def : InstRW< [R52WriteALU_EX2, R52Read_ISS, R52Read_ISS], + (instregex "BFC", "BFI", "UBFX", "SBFX", "(t|t2)BFC", "(t|t2)BFI", + "(t|t2)UBFX", "(t|t2)SBFX")>; + +// Saturating arithmetic +def : InstRW< [R52WriteALU_WRI, R52Read_EX1, R52Read_EX1], + (instregex "QADD", "QSUB", "QDADD", "QDSUB", "SSAT", "SSAT16", "USAT", + "QADD8", "QADD16", "QSUB8", "QSUB16", "QASX", "QSAX", + "UQADD8", "UQADD16","UQSUB8","UQSUB16","UQASX","UQSAX", "t2QADD", + "t2QSUB", "t2QDADD", "t2QDSUB", "t2SSAT", "t2SSAT16", "t2USAT", + "t2QADD8", "t2QADD16", "t2QSUB8", "t2QSUB16", "t2QASX", "t2QSAX", + "t2UQADD8", "t2UQADD16","t2UQSUB8","t2UQSUB16","t2UQASX","t2UQSAX","t2ABS")>; + +// Parallel arithmetic +def : InstRW< [R52WriteALU_EX2, R52Read_EX1, R52Read_EX1], + (instregex "SADD8", "SADD16", "SSUB8", "SSUB16", "SASX", "SSAX", + "UADD8", "UADD16", "USUB8", "USUB16", "UASX", "USAX", "t2SADD8", + "t2SADD16", "t2SSUB8", "t2SSUB16", "t2SASX", "t2SSAX", "t2UADD8", + "t2UADD16", "t2USUB8", "t2USUB16", "t2UASX", "t2USAX")>; + +// Flag setting. +def : InstRW< [R52WriteALU_EX2, R52Read_EX1, R52Read_EX1], + (instregex "SHADD8", "SHADD16", "SHSUB8", "SHSUB16", "SHASX", "SHSAX", + "SXTAB", "SXTAB16", "SXTAH", "UHADD8", "UHADD16", "UHSUB8", "UHSUB16", + "UHASX", "UHSAX", "UXTAB", "UXTAB16", "UXTAH", "t2SHADD8", "t2SHADD16", + "t2SHSUB8", "t2SHSUB16", "t2SHASX", "t2SHSAX", "t2SXTAB", "t2SXTAB16", + "t2SXTAH", "t2UHADD8", "t2UHADD16", "t2UHSUB8", "t2UHSUB16", "t2UHASX", + "t2UHSAX", "t2UXTAB", "t2UXTAB16", "t2UXTAH")>; + +// Sum of Absolute Difference +def : InstRW< [R52WriteALU_WRI, R52Read_ISS, R52Read_ISS, R52Read_ISS], + (instregex "USAD8", "t2USAD8", "USADA8", "t2USADA8") >; + +// Integer Multiply +def : InstRW<[R52WriteMAC, R52Read_ISS, R52Read_ISS], + (instregex "MUL", "SMMUL", "SMMULR", "SMULBB", "SMULBT", + "SMULTB", "SMULTT", "SMULWB", "SMULWT", "SMUSD", "SMUSDX", "t2MUL", + "t2SMMUL", "t2SMMULR", "t2SMULBB", "t2SMULBT", "t2SMULTB", "t2SMULTT", + "t2SMULWB", "t2SMULWT", "t2SMUSD")>; + +// Multiply Accumulate +// Even for 64-bit accumulation (or Long), the single MAC is used (not ALUs). +// The store pipeline is used partly for 64-bit operations. +def : InstRW<[R52WriteMAC, R52Read_ISS, R52Read_ISS, R52Read_ISS], + (instregex "MLA", "MLS", "SMMLA", "SMMLAR", "SMMLS", "SMMLSR", + "t2MLA", "t2MLS", "t2SMMLA", "t2SMMLAR", "t2SMMLS", "t2SMMLSR", + "SMUAD", "SMUADX", "t2SMUAD", "t2SMUADX", + "SMLABB", "SMLABT", "SMLATB", "SMLATT", "SMLSD", "SMLSDX", + "SMLAWB", "SMLAWT", "t2SMLABB", "t2SMLABT", "t2SMLATB", "t2SMLATT", + "t2SMLSD", "t2SMLSDX", "t2SMLAWB", "t2SMLAWT", + "SMLAD", "SMLADX", "t2SMLAD", "t2SMLADX", + "SMULL$", "UMULL$", "t2SMULL$", "t2UMULL$", + "SMLAL", "UMLAL", "SMLALBT", + "SMLALTB", "SMLALTT", "SMLALD", "SMLALDX", "SMLSLD", "SMLSLDX", + "UMAAL", "t2SMLAL", "t2UMLAL", + "t2SMLALBT", "t2SMLALTB", "t2SMLALTT", "t2SMLALD", "t2SMLALDX", + "t2SMLSLD", "t2SMLSLDX", "t2UMAAL")>; + +def : InstRW <[R52WriteDIV, R52Read_ISS, R52Read_ISS], + (instregex "t2SDIV", "t2UDIV")>; + +// Loads (except POST) with SHL > 2, or ror, require 2 extra cycles. +// However, that's non-trivial to specify, so we keep it uniform +def : InstRW<[R52WriteLd, R52Read_ISS, R52Read_ISS], + (instregex "LDR(i12|rs)$", "LDRB(i12|rs)$", "t2LDR(i8|i12|s|pci)", + "t2LDR(H|B)(i8|i12|s|pci)", "LDREX", "t2LDREX", + "tLDR[BH](r|i|spi|pci|pciASM)", "tLDR(r|i|spi|pci|pciASM)", + "LDRH$", "PICLDR$", "PICLDR(H|B)$", "LDRcp$", + "PICLDRS(H|B)$", "t2LDRS(H|B)(i|r|p|s)", "LDRS(H|B)$", + "t2LDRpci_pic", "tLDRS(B|H)", "t2LDRDi8", "LDRD$", "LDA", "t2LDA")>; +def : InstRW<[R52WriteLd, R52WriteAdr, R52Read_ISS, R52Read_ISS], + (instregex "LD(RB|R)(_|T_)(POST|PRE)_(IMM|REG)", "LDRH(_PRE|_POST)", + "LDRBT_POST$", "LDR(T|BT)_POST_(REG|IMM)", "LDRHT(i|r)", + "t2LD(R|RB|RH)_(PRE|POST)", "t2LD(R|RB|RH)T", + "LDR(SH|SB)(_POST|_PRE)", "t2LDR(SH|SB)(_POST|_PRE)", + "LDRS(B|H)T(i|r)", "t2LDRS(B|H)T(i|r)?", + "LDRD_(POST|PRE)", "t2LDRD_(POST|PRE)")>; + +def : InstRW<[R52WriteALU_EX2, R52Read_EX1], (instregex "MOVS?sr", "t2MOVS?sr")>; +def : InstRW<[R52WriteALU_WRI, R52Read_EX2], (instregex "MOVT", "t2MOVT")>; + +def : InstRW<[R52WriteALU_EX2, R52Read_EX1], (instregex "AD(C|D)S?ri", "ANDS?ri", + "BICS?ri", "CLZ", "EORri", "MVNS?r", "ORRri", "RSBS?ri", "RSCri", "SBCri", + "t2AD(C|D)S?ri", "t2ANDS?ri", "t2BICS?ri","t2CLZ", "t2EORri", "t2MVN", + "t2ORRri", "t2RSBS?ri", "t2SBCri")>; + +def : InstRW<[R52WriteALU_EX2, R52Read_EX1, R52Read_EX1], (instregex "AD(C|D)S?rr", + "ANDS?rr", "BICS?rr", "CRC", "EORrr", "ORRrr", "RSBrr", "RSCrr", "SBCrr", + "t2AD(C|D)S?rr", "t2ANDS?rr", "t2BICS?rr", "t2CRC", "t2EORrr", "t2SBCrr")>; + +def : InstRW<[R52WriteALU_EX2, R52Read_EX1, R52Read_ISS], (instregex "AD(C|D)S?rsi", + "ANDS?rsi", "BICS?rsi", "EORrsi", "ORRrsi", "RSBrsi", "RSCrsi", "SBCrsi", + "t2AD(C|D)S?rs", "t2ANDS?rs", "t2BICS?rs", "t2EORrs", "t2ORRrs", "t2RSBrs", "t2SBCrs")>; + +def : InstRW<[R52WriteALU_EX2, R52Read_EX1, R52Read_ISS, R52Read_ISS], + (instregex "AD(C|D)S?rsr", "ANDS?rsr", "BICS?rsr", "EORrsr", "MVNS?sr", + "ORRrsr", "RSBrsr", "RSCrsr", "SBCrsr")>; + +def : InstRW<[R52WriteALU_EX1], + (instregex "ADR", "MOVsi", "MVNS?s?i", "t2MOVS?si")>; + +def : InstRW<[R52WriteALU_EX1, R52Read_ISS], (instregex "ASRi", "RORS?i")>; +def : InstRW<[R52WriteALU_EX1, R52Read_ISS, R52Read_ISS], + (instregex "ASRr", "RORS?r", "LSR", "LSL")>; + +def : InstRW<[R52WriteCC, R52Read_EX1], (instregex "CMPri", "CMNri")>; +def : InstRW<[R52WriteCC, R52Read_EX1, R52Read_EX1], (instregex "CMPrr", "CMNzrr")>; +def : InstRW<[R52WriteCC, R52Read_EX1, R52Read_ISS], (instregex "CMPrsi", "CMNzrsi")>; +def : InstRW<[R52WriteCC, R52Read_EX1, R52Read_ISS, R52Read_ISS], (instregex "CMPrsr", "CMNzrsr")>; + +def : InstRW<[R52WriteALU_EX2, R52Read_ISS], + (instregex "t2LDC", "RBIT", "REV", "REV16", "REVSH", "RRX")>; + +def : InstRW<[R52WriteCC, R52Read_ISS], (instregex "TST")>; + +def : InstRW<[R52WriteLd], (instregex "MRS", "MRSbanked")>; +def : InstRW<[R52WriteLd, R52Read_EX1], (instregex "MSR", "MSRbanked")>; + +// Integer Load, Multiple. +foreach Lat = 3-25 in { + def R52WriteILDM#Lat#Cy : SchedWriteRes<[R52UnitLd]> { + let Latency = Lat; + } + def R52WriteILDM#Lat#CyNo : SchedWriteRes<[]> { + let Latency = Lat; + let NumMicroOps = 0; + } +} +foreach NAddr = 1-16 in { + def R52ILDMAddr#NAddr#Pred : SchedPredicate<"TII->getNumLDMAddresses(*MI) == "#NAddr>; +} +def R52WriteILDMAddrNoWB : SchedWriteRes<[R52UnitLd]> { let Latency = 0; } +def R52WriteILDMAddrWB : SchedWriteRes<[R52UnitLd]>; +def R52WriteILDM : SchedWriteVariant<[ + SchedVar<R52ILDMAddr2Pred, [R52WriteILDM4Cy, R52WriteILDM5Cy]>, + + SchedVar<R52ILDMAddr3Pred, [R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy]>, + SchedVar<R52ILDMAddr4Pred, [R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy]>, + + SchedVar<R52ILDMAddr5Pred, [R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy]>, + SchedVar<R52ILDMAddr6Pred, [R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy]>, + + SchedVar<R52ILDMAddr7Pred, [R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy]>, + SchedVar<R52ILDMAddr8Pred, [R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy]>, + + SchedVar<R52ILDMAddr9Pred, [R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy, + R52WriteILDM12Cy]>, + SchedVar<R52ILDMAddr10Pred,[R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy, + R52WriteILDM12Cy, R52WriteILDM13Cy]>, + + SchedVar<R52ILDMAddr11Pred,[R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy, + R52WriteILDM12Cy, R52WriteILDM13Cy, + R52WriteILDM14Cy]>, + SchedVar<R52ILDMAddr12Pred,[R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy, + R52WriteILDM12Cy, R52WriteILDM13Cy, + R52WriteILDM14Cy, R52WriteILDM15Cy]>, + + SchedVar<R52ILDMAddr13Pred,[R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy, + R52WriteILDM12Cy, R52WriteILDM13Cy, + R52WriteILDM14Cy, R52WriteILDM15Cy, + R52WriteILDM16Cy]>, + SchedVar<R52ILDMAddr14Pred,[R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy, + R52WriteILDM12Cy, R52WriteILDM13Cy, + R52WriteILDM14Cy, R52WriteILDM15Cy, + R52WriteILDM16Cy, R52WriteILDM17Cy]>, + + SchedVar<R52ILDMAddr15Pred,[R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy, + R52WriteILDM12Cy, R52WriteILDM13Cy, + R52WriteILDM14Cy, R52WriteILDM15Cy, + R52WriteILDM16Cy, R52WriteILDM17Cy, + R52WriteILDM18Cy]>, + SchedVar<R52ILDMAddr15Pred,[R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6Cy, R52WriteILDM7Cy, + R52WriteILDM8Cy, R52WriteILDM9Cy, + R52WriteILDM10Cy, R52WriteILDM11Cy, + R52WriteILDM12Cy, R52WriteILDM13Cy, + R52WriteILDM14Cy, R52WriteILDM15Cy, + R52WriteILDM16Cy, R52WriteILDM17Cy, + R52WriteILDM18Cy, R52WriteILDM19Cy]>, + +// Unknown number of registers, just use resources for two registers. + SchedVar<NoSchedPred, [R52WriteILDM4Cy, R52WriteILDM5Cy, + R52WriteILDM6CyNo, R52WriteILDM7CyNo, + R52WriteILDM8CyNo, R52WriteILDM9CyNo, + R52WriteILDM10CyNo, R52WriteILDM11CyNo, + R52WriteILDM12CyNo, R52WriteILDM13CyNo, + R52WriteILDM14CyNo, R52WriteILDM15CyNo, + R52WriteILDM16CyNo, R52WriteILDM17CyNo, + R52WriteILDM18Cy, R52WriteILDM19Cy]> +]> { let Variadic=1; } + +// Integer Store, Multiple +def R52WriteIStIncAddr : SchedWriteRes<[R52UnitLd]> { + let Latency = 4; + let NumMicroOps = 2; +} +foreach NumAddr = 1-16 in { + def R52WriteISTM#NumAddr : WriteSequence<[R52WriteIStIncAddr], NumAddr>; +} +def R52WriteISTM : SchedWriteVariant<[ + SchedVar<R52ILDMAddr2Pred, [R52WriteISTM2]>, + SchedVar<R52ILDMAddr3Pred, [R52WriteISTM3]>, + SchedVar<R52ILDMAddr4Pred, [R52WriteISTM4]>, + SchedVar<R52ILDMAddr5Pred, [R52WriteISTM5]>, + SchedVar<R52ILDMAddr6Pred, [R52WriteISTM6]>, + SchedVar<R52ILDMAddr7Pred, [R52WriteISTM7]>, + SchedVar<R52ILDMAddr8Pred, [R52WriteISTM8]>, + SchedVar<R52ILDMAddr9Pred, [R52WriteISTM9]>, + SchedVar<R52ILDMAddr10Pred,[R52WriteISTM10]>, + SchedVar<R52ILDMAddr11Pred,[R52WriteISTM11]>, + SchedVar<R52ILDMAddr12Pred,[R52WriteISTM12]>, + SchedVar<R52ILDMAddr13Pred,[R52WriteISTM13]>, + SchedVar<R52ILDMAddr14Pred,[R52WriteISTM14]>, + SchedVar<R52ILDMAddr15Pred,[R52WriteISTM15]>, + SchedVar<R52ILDMAddr16Pred,[R52WriteISTM16]>, + // Unknow number of registers, just use resources for two registers. + SchedVar<NoSchedPred, [R52WriteISTM2]> +]>; + +def : InstRW<[R52WriteILDM, R52Read_ISS], + (instregex "LDM(IA|DA|DB|IB)$", "t2LDM(IA|DA|DB|IB)$", + "(t|sys)LDM(IA|DA|DB|IB)$")>; +def : InstRW<[R52WriteILDM, R52WriteAdr, R52Read_ISS], + (instregex "LDM(IA|DA|DB|IB)_UPD", "(t2|sys|t)LDM(IA|DA|DB|IB)_UPD")>; +def : InstRW<[R52WriteILDM, R52WriteAdr, R52Read_ISS], + (instregex "LDMIA_RET", "(t|t2)LDMIA_RET", "tPOP")>; + +// Integer Store, Single Element +def : InstRW<[R52WriteLd, R52Read_ISS, R52Read_EX2], + (instregex "PICSTR", "STR(i12|rs)", "STRB(i12|rs)", "STRH$", "STREX", "SRS", "t2SRS", + "t2SRSDB", "t2STREX", "t2STREXB", "t2STREXD", "t2STREXH", "t2STR(i12|i8|s)$", + "RFE", "t2RFE", "t2STR[BH](i12|i8|s)$", "tSTR[BH](i|r)", "tSTR(i|r)", "tSTRspi")>; + +def : InstRW<[R52WriteLd, R52WriteAdr, R52Read_ISS, R52Read_EX2], + (instregex "STR(B_|_|BT_|T_)(PRE_IMM|PRE_REG|POST_REG|POST_IMM)", + "STR(i|r)_preidx", "STRB(i|r)_preidx", "STRH_preidx", "STR(H_|HT_)(PRE|POST)", + "STR(BT|HT|T)", "t2STR_(PRE|POST)", "t2STR[BH]_(PRE|POST)", + "t2STR_preidx", "t2STR[BH]_preidx", "t2ST(RB|RH|R)T")>; + +// Integer Store, Dual +def : InstRW<[R52WriteLd, R52Read_ISS, R52Read_EX2], + (instregex "STRD$", "t2STRDi8", "STL", "t2STL")>; +def : InstRW<[R52WriteLd, R52WriteAdr, R52Read_ISS, R52Read_EX2], + (instregex "(t2|t)STRD_(POST|PRE)", "STRD_(POST|PRE)")>; + +def : InstRW<[R52WriteISTM, R52Read_ISS, R52Read_EX2], + (instregex "STM(IB|IA|DB|DA)$", "(t2|sys|t)STM(IB|IA|DB|DA)$")>; +def : InstRW<[R52WriteISTM, R52WriteAdr, R52Read_ISS, R52Read_EX2], + (instregex "STM(IB|IA|DB|DA)_UPD", "(t2|sys|t)STM(IB|IA|DB|DA)_UPD", + "tPUSH")>; + +// LDRLIT pseudo instructions, they expand to LDR + PICADD +def : InstRW<[R52WriteLd], + (instregex "t?LDRLIT_ga_abs", "t?LDRLIT_ga_pcrel$")>; +// LDRLIT_ga_pcrel_ldr expands to LDR + PICLDR +def : InstRW<[R52WriteLd], (instregex "LDRLIT_ga_pcrel_ldr")>; + + + +//===----------------------------------------------------------------------===// +// VFP, Floating Point Support +def : InstRW<[R52WriteFPALU_F5, R52Read_F1, R52Read_F1], (instregex "VABD(fd|hd)")>; +def : InstRW<[R52Write2FPALU_F5, R52Read_F1, R52Read_F1], (instregex "VABD(fq|hq)")>; + +def : InstRW<[R52WriteFPALU_F5, R52Read_F1], (instregex "VABS(D|S|H)")>; +def : InstRW<[R52WriteFPALU_F5, R52Read_F1], (instregex "VABS(fd|hd)")>; +def : InstRW<[R52Write2FPALU_F5, R52Read_F1], (instregex "VABS(fq|hq)")>; + +def : InstRW<[R52WriteFPALU_F3, R52Read_F1, R52Read_F1], (instregex "(VACGE|VACGT)(fd|hd)")>; +def : InstRW<[R52Write2FPALU_F3, R52Read_F1, R52Read_F1], (instregex "(VACGE|VACGT)(fq|hq)")>; + +def : InstRW<[R52WriteFPALU_F5, R52Read_F1, R52Read_F1], (instregex "(VADD|VSUB)(D|S|H|fd|hd)$")>; +def : InstRW<[R52Write2FPALU_F5, R52Read_F1, R52Read_F1], (instregex "(VADD|VSUB)(fq|hq)")>; + +def : InstRW<[R52WriteFPLd_F4, R52Read_ISS, R52Read_F1], (instregex "VLDR")>; +def : InstRW<[R52WriteFPST_F4, R52Read_ISS, R52Read_F1], (instregex "VSTR")>; + + +//===----------------------------------------------------------------------===// +// Neon Support + +// vector multiple load stores +foreach NumAddr = 1-16 in { + def R52LMAddrPred#NumAddr : + SchedPredicate<"MI->getNumOperands() == "#NumAddr>; +} +foreach Lat = 1-32 in { + def R52WriteLM#Lat#Cy : SchedWriteRes<[]> { + let Latency = Lat; + } +} +foreach Num = 1-32 in { // reserve LdSt resource, no dual-issue + def R52ReserveLd#Num#Cy : SchedWriteRes<[R52UnitLd]> { + let Latency = 0; + let NumMicroOps = Num; + let ResourceCycles = [Num]; + } +} +def R52WriteVLDM : SchedWriteVariant<[ + // 1 D reg + SchedVar<R52LMAddrPred1, [R52WriteLM5Cy, + R52ReserveLd5Cy]>, + SchedVar<R52LMAddrPred2, [R52WriteLM5Cy, + R52ReserveLd5Cy]>, + + // 2 D reg + SchedVar<R52LMAddrPred3, [R52WriteLM5Cy, R52WriteLM6Cy, + R52ReserveLd6Cy]>, + SchedVar<R52LMAddrPred4, [R52WriteLM5Cy, R52WriteLM6Cy, + R52ReserveLd6Cy]>, + + // 3 D reg + SchedVar<R52LMAddrPred5, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, + R52ReserveLd4Cy]>, + SchedVar<R52LMAddrPred6, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, + R52ReserveLd7Cy]>, + + // 4 D reg + SchedVar<R52LMAddrPred7, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52ReserveLd8Cy]>, + SchedVar<R52LMAddrPred8, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52ReserveLd8Cy]>, + + // 5 D reg + SchedVar<R52LMAddrPred9, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, + R52ReserveLd9Cy]>, + SchedVar<R52LMAddrPred10, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, + R52ReserveLd9Cy]>, + + // 6 D reg + SchedVar<R52LMAddrPred11, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, R52WriteLM10Cy, + R52ReserveLd10Cy]>, + SchedVar<R52LMAddrPred12, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, R52WriteLM10Cy, + R52ReserveLd10Cy]>, + + // 7 D reg + SchedVar<R52LMAddrPred13, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, R52WriteLM10Cy, + R52WriteLM11Cy, + R52ReserveLd11Cy]>, + SchedVar<R52LMAddrPred14, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, R52WriteLM10Cy, + R52WriteLM11Cy, + R52ReserveLd11Cy]>, + + // 8 D reg + SchedVar<R52LMAddrPred14, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, R52WriteLM10Cy, + R52WriteLM11Cy, R52WriteLM12Cy, + R52ReserveLd12Cy]>, + SchedVar<R52LMAddrPred15, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, R52WriteLM10Cy, + R52WriteLM11Cy, R52WriteLM12Cy, + R52ReserveLd12Cy]>, + // unknown number of reg. + SchedVar<NoSchedPred, [R52WriteLM5Cy, R52WriteLM6Cy, + R52WriteLM7Cy, R52WriteLM8Cy, + R52WriteLM9Cy, R52WriteLM10Cy, + R52WriteLM11Cy, R52WriteLM12Cy, + R52ReserveLd5Cy]> +]> { let Variadic=1;} + +// variable stores. Cannot dual-issue +def R52WriteSTM5 : SchedWriteRes<[R52UnitLd]> { + let Latency = 5; + let NumMicroOps = 2; + let ResourceCycles = [1]; +} +def R52WriteSTM6 : SchedWriteRes<[R52UnitLd]> { + let Latency = 6; + let NumMicroOps = 4; + let ResourceCycles = [2]; +} +def R52WriteSTM7 : SchedWriteRes<[R52UnitLd]> { + let Latency = 7; + let NumMicroOps = 6; + let ResourceCycles = [3]; +} +def R52WriteSTM8 : SchedWriteRes<[R52UnitLd]> { + let Latency = 8; + let NumMicroOps = 8; + let ResourceCycles = [4]; +} +def R52WriteSTM9 : SchedWriteRes<[R52UnitLd]> { + let Latency = 9; + let NumMicroOps = 10; + let ResourceCycles = [5]; +} +def R52WriteSTM10 : SchedWriteRes<[R52UnitLd]> { + let Latency = 10; + let NumMicroOps = 12; + let ResourceCycles = [6]; +} +def R52WriteSTM11 : SchedWriteRes<[R52UnitLd]> { + let Latency = 11; + let NumMicroOps = 14; + let ResourceCycles = [7]; +} +def R52WriteSTM12 : SchedWriteRes<[R52UnitLd]> { + let Latency = 12; + let NumMicroOps = 16; + let ResourceCycles = [8]; +} +def R52WriteSTM13 : SchedWriteRes<[R52UnitLd]> { + let Latency = 13; + let NumMicroOps = 18; + let ResourceCycles = [9]; +} +def R52WriteSTM14 : SchedWriteRes<[R52UnitLd]> { + let Latency = 14; + let NumMicroOps = 20; + let ResourceCycles = [10]; +} +def R52WriteSTM15 : SchedWriteRes<[R52UnitLd]> { + let Latency = 15; + let NumMicroOps = 22; + let ResourceCycles = [11]; +} + +def R52WriteSTM : SchedWriteVariant<[ + SchedVar<R52LMAddrPred1, [R52WriteSTM5]>, + SchedVar<R52LMAddrPred2, [R52WriteSTM5]>, + SchedVar<R52LMAddrPred3, [R52WriteSTM6]>, + SchedVar<R52LMAddrPred4, [R52WriteSTM6]>, + SchedVar<R52LMAddrPred5, [R52WriteSTM7]>, + SchedVar<R52LMAddrPred6, [R52WriteSTM7]>, + SchedVar<R52LMAddrPred7, [R52WriteSTM8]>, + SchedVar<R52LMAddrPred8, [R52WriteSTM8]>, + SchedVar<R52LMAddrPred9, [R52WriteSTM9]>, + SchedVar<R52LMAddrPred10, [R52WriteSTM9]>, + SchedVar<R52LMAddrPred11, [R52WriteSTM10]>, + SchedVar<R52LMAddrPred12, [R52WriteSTM10]>, + SchedVar<R52LMAddrPred13, [R52WriteSTM11]>, + SchedVar<R52LMAddrPred14, [R52WriteSTM11]>, + SchedVar<R52LMAddrPred15, [R52WriteSTM12]>, + SchedVar<R52LMAddrPred16, [R52WriteSTM12]>, + // unknown number of registers, just use resources for two + SchedVar<NoSchedPred, [R52WriteSTM6]> +]>; + +// Vector Load/Stores. Can issue only in slot-0. Can dual-issue with +// another instruction in slot-1, but only in the last issue. +def : WriteRes<WriteVLD1, [R52UnitLd]> { let Latency = 5;} +def : WriteRes<WriteVLD2, [R52UnitLd]> { + let Latency = 6; + let NumMicroOps = 3; + let ResourceCycles = [2]; + let SingleIssue = 1; +} +def : WriteRes<WriteVLD3, [R52UnitLd]> { + let Latency = 7; + let NumMicroOps = 5; + let ResourceCycles = [3]; + let SingleIssue = 1; +} +def : WriteRes<WriteVLD4, [R52UnitLd]> { + let Latency = 8; + let NumMicroOps = 7; + let ResourceCycles = [4]; + let SingleIssue = 1; +} +def R52WriteVST1Mem : SchedWriteRes<[R52UnitLd]> { + let Latency = 5; + let NumMicroOps = 1; + let ResourceCycles = [1]; +} +def R52WriteVST2Mem : SchedWriteRes<[R52UnitLd]> { + let Latency = 6; + let NumMicroOps = 3; + let ResourceCycles = [2]; +} +def R52WriteVST3Mem : SchedWriteRes<[R52UnitLd]> { + let Latency = 7; + let NumMicroOps = 5; + let ResourceCycles = [3]; +} +def R52WriteVST4Mem : SchedWriteRes<[R52UnitLd]> { + let Latency = 8; + let NumMicroOps = 7; + let ResourceCycles = [4]; +} +def R52WriteVST5Mem : SchedWriteRes<[R52UnitLd]> { + let Latency = 9; + let NumMicroOps = 9; + let ResourceCycles = [5]; +} + + +def : InstRW<[R52WriteFPALU_F5, R52Read_F1, R52Read_F1, R52Read_F1], (instregex "VABA(u|s)(v8i8|v4i16|v2i32)")>; +def : InstRW<[R52Write2FPALU_F5, R52Read_F1, R52Read_F1, R52Read_F1], (instregex "VABA(u|s)(v16i8|v8i16|v4i32)")>; +def : InstRW<[R52Write2FPALU_F5, R52Read_F1, R52Read_F1, R52Read_F1], (instregex "VABAL(u|s)(v8i16|v4i32|v2i64)")>; + +def : InstRW<[R52WriteFPALU_F4, R52Read_F1, R52Read_F1], (instregex "VABD(u|s)(v8i8|v4i16|v2i32)")>; +def : InstRW<[R52Write2FPALU_F4, R52Read_F1, R52Read_F1], (instregex "VABD(u|s)(v16i8|v8i16|v4i32)")>; +def : InstRW<[R52Write2FPALU_F4, R52Read_F1, R52Read_F1], (instregex "VABDL(u|s)(v16i8|v8i16|v4i32)")>; + +def : InstRW<[R52Write2FPALU_F4, R52Read_F1], (instregex "VABS(v16i8|v8i16|v4i32)")>; + +def : InstRW<[R52WriteFPALU_F4, R52Read_F2, R52Read_F2], + (instregex "(VADD|VSUB)(v8i8|v4i16|v2i32|v1i64)")>; +def : InstRW<[R52Write2FPALU_F4, R52Read_F2, R52Read_F2], + (instregex "(VADD|VSUB)(v16i8|v8i16|v4i32|v2i64)")>; +def : InstRW<[R52Write2FPALU_F5, R52Read_F2, R52Read_F2], + (instregex "(VADDHN|VRADDHN|VSUBHN|VRSUBHN)(v8i8|v4i16|v2i32)")>; + +def : InstRW<[R52Write2FPALU_F4, R52Read_F1, R52Read_F1], + (instregex "VADDL", "VADDW", "VSUBL", "VSUBW")>; + +def : InstRW<[R52WriteFPALU_F3, R52Read_F2, R52Read_F2], (instregex "(VAND|VBIC|VEOR)d")>; +def : InstRW<[R52Write2FPALU_F3, R52Read_F2, R52Read_F2], (instregex "(VAND|VBIC|VEOR)q")>; + +def : InstRW<[R52WriteFPALU_F3, R52Read_F2], (instregex "VBICi(v4i16|v2i32)")>; +def : InstRW<[R52Write2FPALU_F3, R52Read_F2], (instregex "VBICi(v8i16|v4i32)")>; + +def : InstRW<[R52WriteFPALU_F3, R52Read_F1, R52Read_F2, R52Read_F2], (instregex "(VBIF|VBIT|VBSL)d")>; +def : InstRW<[R52Write2FPALU_F3, R52Read_F1, R52Read_F2, R52Read_F2], (instregex "(VBIF|VBIT|VBSL)q")>; + +def : InstRW<[R52WriteFPALU_F3, R52Read_F1, R52Read_F1], + (instregex "(VCEQ|VCGE|VCGT|VCLE|VCLT|VCLZ|VCMP|VCMPE|VCNT)")>; +def : InstRW<[R52WriteFPALU_F5, R52Read_F1, R52Read_F1], + (instregex "VCVT", "VSITO", "VUITO", "VTO")>; + +def : InstRW<[R52WriteFPALU_F3, R52Read_ISS], (instregex "VDUP(8|16|32)d")>; +def : InstRW<[R52Write2FPALU_F3, R52Read_ISS], (instregex "VDUP(8|16|32)q")>; +def : InstRW<[R52WriteFPALU_F3, R52Read_F1], (instregex "VDUPLN(8|16|32)d")>; +def : InstRW<[R52Write2FPALU_F3, R52Read_F1], (instregex "VDUPLN(8|16|32)q")>; + +def : InstRW<[R52WriteFPALU_F3, R52Read_F1, R52Read_F1], (instregex "VEXTd(8|16|32)", "VSEL")>; +def : InstRW<[R52Write2FPALU_F3, R52Read_F1, R52Read_F1], (instregex "VEXTq(8|16|32|64)")>; + +def : InstRW<[R52WriteFPMAC_F5, R52Read_F1, R52Read_F1, R52Read_F1], (instregex "(VFMA|VFMS)(f|h)d")>; +def : InstRW<[R52Write2FPMAC_F5, R52Read_F1, R52Read_F1, R52Read_F1], (instregex "(VFMA|VFMS)(f|h)q")>; + +def : InstRW<[R52WriteFPALU_F4, R52Read_F2, R52Read_F2], (instregex "(VHADD|VHSUB)(u|s)(v8i8|v4i16|v2i32)")>; +def : InstRW<[R52Write2FPALU_F4, R52Read_F2, R52Read_F2], (instregex "(VHADD|VHSUB)(u|s)(v16i8|v8i16|v4i32)")>; + +def : InstRW<[R52WriteVLDM], (instregex "VLDM[SD](IA|DB)$")>; +def : InstRW<[R52WriteFPALU_F4, R52Read_F1, R52Read_F1], (instregex "VMAX", "VMIN", "VPMAX", "VPMIN")>; +def : InstRW<[R52WriteFPALU_F3, R52Read_F1, R52Read_F1], (instregex "VORR", "VORN", "VREV")>; +def : InstRW<[R52WriteNoRSRC_WRI], (instregex "VMRS")>; +def : InstRW<[R52WriteFPALU_F5, R52Read_F1], (instregex "VNEG")>; +def : InstRW<[R52WriteFPALU_F4, R52Read_F1, R52Read_F1], (instregex "VPADDi")>; +def : InstRW<[R52Write2FPALU_F4, R52Read_F1, R52Read_F1], (instregex "VPADAL", "VPADDL")>; +def : InstRW<[R52WriteFPALU_F5, R52Read_F1], (instregex "VQABS(v8i8|v4i16|v2i32|v1i64)")>; +def : InstRW<[R52Write2FPALU_F5, R52Read_F1], (instregex "VQABS(v16i8|v8i16|v4i32|v2i64)")>; +def : InstRW<[R52WriteFPALU_F5, R52Read_F2, R52Read_F2], + (instregex "(VQADD|VQSUB)(u|s)(v8i8|v4i16|v2i32|v1i64)")>; +def : InstRW<[R52Write2FPALU_F5, R52Read_F2, R52Read_F2], + (instregex "(VQADD|VQSUB)(u|s)(v16i8|v8i16|v4i32|v2i64)")>; +def : InstRW<[R52Write2FPMAC_F5, R52Read_F1, R52Read_F1, R52Read_F1], (instregex "VQDMLAL", "VQDMLSL")>; +def : InstRW<[R52WriteFPMUL_F5, R52Read_F1, R52Read_F1, R52Read_F1], (instregex "VQDMUL","VQRDMUL")>; +def : InstRW<[R52WriteFPALU_F5, R52Read_F1, R52Read_F1], + (instregex "VQMOVN", "VQNEG", "VQSHL", "VQSHRN")>; +def : InstRW<[R52WriteFPALU_F4, R52Read_F1, R52Read_F1], (instregex "VRSHL", "VRSHR", "VRSHRN", "VTB")>; +def : InstRW<[R52WriteFPALU_F3, R52Read_F1, R52Read_F1], (instregex "VSWP", "VTRN", "VUZP", "VZIP")>; + +//--- +// VSTx. Vector Stores +//--- +// 1-element structure store +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST1d(8|16|32|64)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST1q(8|16|32|64)$")>; +def : InstRW<[R52WriteVST3Mem, R52Read_ISS, R52Read_F2], (instregex "VST1d(8|16|32|64)T$")>; +def : InstRW<[R52WriteVST4Mem, R52Read_ISS, R52Read_F2], (instregex "VST1d(8|16|32|64)Q$")>; +def : InstRW<[R52WriteVST3Mem, R52Read_ISS, R52Read_F2], (instregex "VST1d64TPseudo$")>; +def : InstRW<[R52WriteVST4Mem, R52Read_ISS, R52Read_F2], (instregex "VST1d64QPseudo$")>; + +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST1LNd(8|16|32)$")>; +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST1LNdAsm_(8|16|32)$")>; +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST1LNq(8|16|32)Pseudo$")>; + +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1d(8|16|32|64)wb")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1q(8|16|32|64)wb")>; +def : InstRW<[R52WriteVST3Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1d(8|16|32|64)Twb")>; +def : InstRW<[R52WriteVST4Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1d(8|16|32|64)Qwb")>; +def : InstRW<[R52WriteVST3Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1d64TPseudoWB")>; +def : InstRW<[R52WriteVST4Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1d64QPseudoWB")>; + +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1LNd(8|16|32)_UPD")>; +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1LNdWB_(fixed|register)_Asm_(8|16|32)")>; +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST1LNq(8|16|32)Pseudo_UPD")>; + +// 2-element structure store +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST2(d|b)(8|16|32)$")>; +def : InstRW<[R52WriteVST4Mem, R52Read_ISS, R52Read_F2], (instregex "VST2q(8|16|32)$")>; +def : InstRW<[R52WriteVST4Mem, R52Read_ISS, R52Read_F2], (instregex "VST2q(8|16|32)Pseudo$")>; + +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST2LNd(8|16|32)$")>; +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST2LNdAsm_(8|16|32)$")>; +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST2LNd(8|16|32)Pseudo$")>; +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST2LNq(16|32)$")>; +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST2LNqAsm_(16|32)$")>; +def : InstRW<[R52WriteVST1Mem, R52Read_ISS, R52Read_F2], (instregex "VST2LNq(16|32)Pseudo$")>; + +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2(d|b)(8|16|32)wb")>; +def : InstRW<[R52WriteVST4Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2q(8|16|32)wb")>; +def : InstRW<[R52WriteVST4Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2q(8|16|32)PseudoWB")>; + +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2LNd(8|16|32)_UPD")>; +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2LNdWB_(fixed|register)_Asm_(8|16|32)")>; +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2LNd(8|16|32)Pseudo_UPD")>; +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2LNq(16|32)_UPD")>; +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2LNqWB_(fixed|register)_Asm_(16|32)")>; +def : InstRW<[R52WriteVST1Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST2LNq(16|32)Pseudo_UPD")>; + +// 3-element structure store +def : InstRW<[R52WriteVST4Mem, R52Read_ISS, R52Read_F2], (instregex "VST3(d|q)(8|16|32)$")>; +def : InstRW<[R52WriteVST4Mem, R52Read_ISS, R52Read_F2], (instregex "VST3(d|q)Asm_(8|16|32)$")>; +def : InstRW<[R52WriteVST4Mem, R52Read_ISS, R52Read_F2], (instregex "VST3d(8|16|32)(oddP|P)seudo$")>; + +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST3LNd(8|16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST3LNdAsm_(8|16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST3LNd(8|16|32)Pseudo$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST3LNq(16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST3LNqAsm_(16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST3LNq(16|32)Pseudo$")>; + +def : InstRW<[R52WriteVST4Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3(d|q)(8|16|32)_UPD$")>; +def : InstRW<[R52WriteVST4Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3(d|q)WB_(fixed|register)_Asm_(8|16|32)$")>; +def : InstRW<[R52WriteVST4Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3(d|q)(8|16|32)(oddP|P)seudo_UPD$")>; + +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3LNd(8|16|32)_UPD$")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3LNdWB_(fixed|register)_Asm_(8|16|32)")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3LNd(8|16|32)Pseudo_UPD$")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3LNq(16|32)_UPD$")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3LNqWB_(fixed|register)_Asm_(16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST3LNq(16|32)Pseudo_UPD$")>; + +// 4-element structure store +def : InstRW<[R52WriteVST5Mem, R52Read_ISS, R52Read_F2], (instregex "VST4(d|q)(8|16|32)$")>; +def : InstRW<[R52WriteVST5Mem, R52Read_ISS, R52Read_F2], (instregex "VST4(d|q)Asm_(8|16|32)$")>; +def : InstRW<[R52WriteVST5Mem, R52Read_ISS, R52Read_F2], (instregex "VST4d(8|16|32)Pseudo$")>; + +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST4LNd(8|16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST4LNdAsm_(8|16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST4LNd(8|16|32)Pseudo$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST4LNq(16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST4LNqAsm_(16|32)$")>; +def : InstRW<[R52WriteVST2Mem, R52Read_ISS, R52Read_F2], (instregex "VST4LNq(16|32)Pseudo$")>; + +def : InstRW<[R52WriteVST5Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4(d|q)(8|16|32)_UPD")>; +def : InstRW<[R52WriteVST5Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4(d|q)WB_(fixed|register)_Asm_(8|16|32)")>; +def : InstRW<[R52WriteVST5Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4(d|q)(8|16|32)(oddP|P)seudo_UPD")>; + +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4LNd(8|16|32)_UPD")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4LNdWB_(fixed|register)_Asm_(8|16|32)")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4LNd(8|16|32)Pseudo_UPD")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4LNq(16|32)_UPD")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4LNqWB_(fixed|register)_Asm_(16|32)")>; +def : InstRW<[R52WriteVST2Mem, R52WriteAdr, R52Read_ISS, R52Read_F2], (instregex "VST4LNq(16|32)Pseudo_UPD")>; + +} // R52 SchedModel diff --git a/capstone/suite/synctools/tablegen/ARM/ARMScheduleSwift.td b/capstone/suite/synctools/tablegen/ARM/ARMScheduleSwift.td new file mode 100644 index 000000000..879846481 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMScheduleSwift.td @@ -0,0 +1,1093 @@ +//=- ARMScheduleSwift.td - Swift Scheduling Definitions -*- tablegen -*----===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the itinerary class data for the Swift processor.. +// +//===----------------------------------------------------------------------===// + +// ===---------------------------------------------------------------------===// +// This section contains legacy support for itineraries. This is +// required until SD and PostRA schedulers are replaced by MachineScheduler. + +def SW_DIS0 : FuncUnit; +def SW_DIS1 : FuncUnit; +def SW_DIS2 : FuncUnit; + +def SW_ALU0 : FuncUnit; +def SW_ALU1 : FuncUnit; +def SW_LS : FuncUnit; +def SW_IDIV : FuncUnit; +def SW_FDIV : FuncUnit; + +// FIXME: Need bypasses. +// FIXME: Model the multiple stages of IIC_iMOVix2, IIC_iMOVix2addpc, and +// IIC_iMOVix2ld better. +// FIXME: Model the special immediate shifts that are not microcoded. +// FIXME: Do we need to model the fact that uses of r15 in a micro-op force it +// to issue on pipe 1? +// FIXME: Model the pipelined behavior of CMP / TST instructions. +// FIXME: Better model the microcode stages of multiply instructions, especially +// conditional variants. +// FIXME: Add preload instruction when it is documented. +// FIXME: Model non-pipelined nature of FP div / sqrt unit. + +// Swift machine model for scheduling and other instruction cost heuristics. +def SwiftModel : SchedMachineModel { + let IssueWidth = 3; // 3 micro-ops are dispatched per cycle. + let MicroOpBufferSize = 45; // Based on NEON renamed registers. + let LoadLatency = 3; + let MispredictPenalty = 14; // A branch direction mispredict. + let CompleteModel = 0; // FIXME: Remove if all instructions are covered. + + // FIXME: Remove when all errors have been fixed. + let FullInstRWOverlapCheck = 0; +} + +// Swift predicates. +def IsFastImmShiftSwiftPred : SchedPredicate<[{TII->isSwiftFastImmShift(MI)}]>; + +// Swift resource mapping. +let SchedModel = SwiftModel in { + // Processor resources. + def SwiftUnitP01 : ProcResource<2>; // ALU unit. + def SwiftUnitP0 : ProcResource<1> { let Super = SwiftUnitP01; } // Mul unit. + def SwiftUnitP1 : ProcResource<1> { let Super = SwiftUnitP01; } // Br unit. + def SwiftUnitP2 : ProcResource<1>; // LS unit. + def SwiftUnitDiv : ProcResource<1>; + + // Generic resource requirements. + def SwiftWriteP0OneCycle : SchedWriteRes<[SwiftUnitP0]>; + def SwiftWriteP0TwoCycle : SchedWriteRes<[SwiftUnitP0]> { let Latency = 2; } + def SwiftWriteP0FourCycle : SchedWriteRes<[SwiftUnitP0]> { let Latency = 4; } + def SwiftWriteP0SixCycle : SchedWriteRes<[SwiftUnitP0]> { let Latency = 6; } + def SwiftWriteP0P1FourCycle : SchedWriteRes<[SwiftUnitP0, SwiftUnitP1]> { + let Latency = 4; + } + def SwiftWriteP0P1SixCycle : SchedWriteRes<[SwiftUnitP0, SwiftUnitP1]> { + let Latency = 6; + } + def SwiftWriteP01OneCycle : SchedWriteRes<[SwiftUnitP01]>; + def SwiftWriteP1TwoCycle : SchedWriteRes<[SwiftUnitP1]> { let Latency = 2; } + def SwiftWriteP1FourCycle : SchedWriteRes<[SwiftUnitP1]> { let Latency = 4; } + def SwiftWriteP1SixCycle : SchedWriteRes<[SwiftUnitP1]> { let Latency = 6; } + def SwiftWriteP1EightCycle : SchedWriteRes<[SwiftUnitP1]> { let Latency = 8; } + def SwiftWriteP1TwelveCyc : SchedWriteRes<[SwiftUnitP1]> { let Latency = 12; } + def SwiftWriteP01OneCycle2x : WriteSequence<[SwiftWriteP01OneCycle], 2>; + def SwiftWriteP01OneCycle3x : WriteSequence<[SwiftWriteP01OneCycle], 3>; + def SwiftWriteP01TwoCycle : SchedWriteRes<[SwiftUnitP01]> { let Latency = 2; } + def SwiftWriteP01ThreeCycleTwoUops : SchedWriteRes<[SwiftUnitP01, + SwiftUnitP01]> { + let Latency = 3; + let NumMicroOps = 2; + } + def SwiftWriteP0ThreeCycleThreeUops : SchedWriteRes<[SwiftUnitP0]> { + let Latency = 3; + let NumMicroOps = 3; + let ResourceCycles = [3]; + } + // Plain load without writeback. + def SwiftWriteP2ThreeCycle : SchedWriteRes<[SwiftUnitP2]> { + let Latency = 3; + } + def SwiftWriteP2FourCycle : SchedWriteRes<[SwiftUnitP2]> { + let Latency = 4; + } + // A store does not write to a register. + def SwiftWriteP2 : SchedWriteRes<[SwiftUnitP2]> { + let Latency = 0; + } + foreach Num = 1-4 in { + def SwiftWrite#Num#xP2 : WriteSequence<[SwiftWriteP2], Num>; + } + def SwiftWriteP01OneCycle2x_load : WriteSequence<[SwiftWriteP01OneCycle, + SwiftWriteP01OneCycle, + SwiftWriteP2ThreeCycle]>; + // 4.2.4 Arithmetic and Logical. + // ALU operation register shifted by immediate variant. + def SwiftWriteALUsi : SchedWriteVariant<[ + // lsl #2, lsl #1, or lsr #1. + SchedVar<IsFastImmShiftSwiftPred, [SwiftWriteP01TwoCycle]>, + SchedVar<NoSchedPred, [WriteALU]> + ]>; + def SwiftWriteALUsr : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [SwiftWriteP01ThreeCycleTwoUops]>, + SchedVar<NoSchedPred, [SwiftWriteP01TwoCycle]> + ]>; + def SwiftWriteALUSsr : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [SwiftWriteP0ThreeCycleThreeUops]>, + SchedVar<NoSchedPred, [SwiftWriteP01TwoCycle]> + ]>; + def SwiftReadAdvanceALUsr : SchedReadVariant<[ + SchedVar<IsPredicatedPred, [SchedReadAdvance<2>]>, + SchedVar<NoSchedPred, [NoReadAdvance]> + ]>; + // ADC,ADD,NEG,RSB,RSC,SBC,SUB,ADR + // AND,BIC,EOR,ORN,ORR + // CLZ,RBIT,REV,REV16,REVSH,PKH + def : WriteRes<WriteALU, [SwiftUnitP01]>; + def : SchedAlias<WriteALUsi, SwiftWriteALUsi>; + def : SchedAlias<WriteALUsr, SwiftWriteALUsr>; + def : SchedAlias<WriteALUSsr, SwiftWriteALUSsr>; + def : ReadAdvance<ReadALU, 0>; + def : SchedAlias<ReadALUsr, SwiftReadAdvanceALUsr>; + def : SchedAlias<WriteLd, SwiftWriteP2ThreeCycle>; + def : SchedAlias<WriteST, SwiftWriteP2>; + + + def SwiftChooseShiftKindP01OneOrTwoCycle : SchedWriteVariant<[ + SchedVar<IsFastImmShiftSwiftPred, [SwiftWriteP01OneCycle]>, + SchedVar<NoSchedPred, [SwiftWriteP01TwoCycle]> + ]>; + + // 4.2.5 Integer comparison + def : WriteRes<WriteCMP, [SwiftUnitP01]>; + def : SchedAlias<WriteCMPsi, SwiftChooseShiftKindP01OneOrTwoCycle>; + def : SchedAlias<WriteCMPsr, SwiftWriteP01TwoCycle>; + + // 4.2.6 Shift, Move + // Shift + // ASR,LSL,ROR,RRX + // MOV(register-shiftedregister) MVN(register-shiftedregister) + // Move + // MOV,MVN + // MOVT + // Sign/Zero extension + def : InstRW<[SwiftWriteP01OneCycle], + (instregex "SXTB", "SXTH", "SXTB16", "UXTB", "UXTH", "UXTB16", + "t2SXTB", "t2SXTH", "t2SXTB16", "t2UXTB", "t2UXTH", + "t2UXTB16")>; + // Pseudo instructions. + def : InstRW<[SwiftWriteP01OneCycle2x], + (instregex "MOVCCi32imm", "MOVi32imm", "t2MOVCCi32imm", + "t2MOVi32imm")>; + def : InstRW<[SwiftWriteP01OneCycle3x], + (instregex "MOV_ga_pcrel", "t2MOV_ga_pcrel", "t2MOVi16_ga_pcrel")>; + def : InstRW<[SwiftWriteP01OneCycle2x_load], + (instregex "MOV_ga_pcrel_ldr")>; + + def SwiftWriteP0TwoCycleTwoUops : WriteSequence<[SwiftWriteP0OneCycle], 2>; + + def SwiftPredP0OneOrTwoCycle : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [ SwiftWriteP0TwoCycleTwoUops ]>, + SchedVar<NoSchedPred, [ SwiftWriteP0OneCycle ]> + ]>; + + // 4.2.7 Select + // SEL + def : InstRW<[SwiftPredP0OneOrTwoCycle], (instregex "SEL", "t2SEL")>; + + // 4.2.8 Bitfield + // BFI,BFC, SBFX,UBFX + def : InstRW< [SwiftWriteP01TwoCycle], + (instregex "BFC", "BFI", "UBFX", "SBFX", "(t|t2)BFC", "(t|t2)BFI", + "(t|t2)UBFX", "(t|t2)SBFX")>; + + // 4.2.9 Saturating arithmetic + def : InstRW< [SwiftWriteP01TwoCycle], + (instregex "QADD", "QSUB", "QDADD", "QDSUB", "SSAT", "SSAT16", "USAT", + "USAT16", "QADD8", "QADD16", "QSUB8", "QSUB16", "QASX", "QSAX", + "UQADD8", "UQADD16","UQSUB8","UQSUB16","UQASX","UQSAX", "t2QADD", + "t2QSUB", "t2QDADD", "t2QDSUB", "t2SSAT", "t2SSAT16", "t2USAT", + "t2QADD8", "t2QADD16", "t2QSUB8", "t2QSUB16", "t2QASX", "t2QSAX", + "t2UQADD8", "t2UQADD16","t2UQSUB8","t2UQSUB16","t2UQASX","t2UQSAX")>; + + // 4.2.10 Parallel Arithmetic + // Not flag setting. + def : InstRW< [SwiftWriteALUsr], + (instregex "SADD8", "SADD16", "SSUB8", "SSUB16", "SASX", "SSAX", + "UADD8", "UADD16", "USUB8", "USUB16", "UASX", "USAX", "t2SADD8", + "t2SADD16", "t2SSUB8", "t2SSUB16", "t2SASX", "t2SSAX", "t2UADD8", + "t2UADD16", "t2USUB8", "t2USUB16", "t2UASX", "t2USAX")>; + // Flag setting. + def : InstRW< [SwiftWriteP01TwoCycle], + (instregex "SHADD8", "SHADD16", "SHSUB8", "SHSUB16", "SHASX", "SHSAX", + "SXTAB", "SXTAB16", "SXTAH", "UHADD8", "UHADD16", "UHSUB8", "UHSUB16", + "UHASX", "UHSAX", "UXTAB", "UXTAB16", "UXTAH", "t2SHADD8", "t2SHADD16", + "t2SHSUB8", "t2SHSUB16", "t2SHASX", "t2SHSAX", "t2SXTAB", "t2SXTAB16", + "t2SXTAH", "t2UHADD8", "t2UHADD16", "t2UHSUB8", "t2UHSUB16", "t2UHASX", + "t2UHSAX", "t2UXTAB", "t2UXTAB16", "t2UXTAH")>; + + // 4.2.11 Sum of Absolute Difference + def : InstRW< [SwiftWriteP0P1FourCycle], (instregex "USAD8") >; + def : InstRW<[SwiftWriteP0P1FourCycle, ReadALU, ReadALU, SchedReadAdvance<2>], + (instregex "USADA8")>; + + // 4.2.12 Integer Multiply (32-bit result) + // Two sources. + def : InstRW< [SwiftWriteP0FourCycle], + (instregex "MUL", "SMMUL", "SMMULR", "SMULBB", "SMULBT", + "SMULTB", "SMULTT", "SMULWB", "SMULWT", "SMUSD", "SMUSDX", "t2MUL", + "t2SMMUL", "t2SMMULR", "t2SMULBB", "t2SMULBT", "t2SMULTB", "t2SMULTT", + "t2SMULWB", "t2SMULWT", "t2SMUSD")>; + + def SwiftWriteP0P01FiveCycleTwoUops : + SchedWriteRes<[SwiftUnitP0, SwiftUnitP01]> { + let Latency = 5; + } + + def SwiftPredP0P01FourFiveCycle : SchedWriteVariant<[ + SchedVar<IsPredicatedPred, [ SwiftWriteP0P01FiveCycleTwoUops ]>, + SchedVar<NoSchedPred, [ SwiftWriteP0FourCycle ]> + ]>; + + def SwiftReadAdvanceFourCyclesPred : SchedReadVariant<[ + SchedVar<IsPredicatedPred, [SchedReadAdvance<4>]>, + SchedVar<NoSchedPred, [ReadALU]> + ]>; + + // Multiply accumulate, three sources + def : InstRW< [SwiftPredP0P01FourFiveCycle, ReadALU, ReadALU, + SwiftReadAdvanceFourCyclesPred], + (instregex "MLA", "MLS", "SMMLA", "SMMLAR", "SMMLS", "SMMLSR", + "t2MLA", "t2MLS", "t2SMMLA", "t2SMMLAR", "t2SMMLS", + "t2SMMLSR")>; + + // 4.2.13 Integer Multiply (32-bit result, Q flag) + def : InstRW< [SwiftWriteP0FourCycle], + (instregex "SMUAD", "SMUADX", "t2SMUAD", "t2SMUADX")>; + def : InstRW< [SwiftPredP0P01FourFiveCycle, ReadALU, ReadALU, + SwiftReadAdvanceFourCyclesPred], + (instregex "SMLABB", "SMLABT", "SMLATB", "SMLATT", "SMLSD", "SMLSDX", + "SMLAWB", "SMLAWT", "t2SMLABB", "t2SMLABT", "t2SMLATB", "t2SMLATT", + "t2SMLSD", "t2SMLSDX", "t2SMLAWB", "t2SMLAWT")>; + def : InstRW< [SwiftPredP0P01FourFiveCycle], + (instregex "SMLAD", "SMLADX", "t2SMLAD", "t2SMLADX")>; + + def SwiftP0P0P01FiveCycle : SchedWriteRes<[SwiftUnitP0, SwiftUnitP01]> { + let Latency = 5; + let NumMicroOps = 3; + let ResourceCycles = [2, 1]; + } + def SwiftWrite1Cycle : SchedWriteRes<[]> { + let Latency = 1; + let NumMicroOps = 0; + } + def SwiftWrite5Cycle : SchedWriteRes<[]> { + let Latency = 5; + let NumMicroOps = 0; + } + def SwiftWrite6Cycle : SchedWriteRes<[]> { + let Latency = 6; + let NumMicroOps = 0; + } + + // 4.2.14 Integer Multiply, Long + def : InstRW< [SwiftP0P0P01FiveCycle, SwiftWrite5Cycle], + (instregex "SMULL$", "UMULL$", "t2SMULL$", "t2UMULL$")>; + + def Swift2P03P01FiveCycle : SchedWriteRes<[SwiftUnitP0, SwiftUnitP01]> { + let Latency = 7; + let NumMicroOps = 5; + let ResourceCycles = [2, 3]; + } + + // Aliasing sub-target specific WriteRes to generic ones + def : SchedAlias<WriteMUL16, SwiftWriteP0FourCycle>; + def : SchedAlias<WriteMUL32, SwiftWriteP0FourCycle>; + def : SchedAlias<WriteMUL64Lo, SwiftP0P0P01FiveCycle>; + def : SchedAlias<WriteMUL64Hi, SwiftWrite5Cycle>; + def : SchedAlias<WriteMAC16, SwiftPredP0P01FourFiveCycle>; + def : SchedAlias<WriteMAC32, SwiftPredP0P01FourFiveCycle>; + def : SchedAlias<WriteMAC64Lo, SwiftWrite5Cycle>; + def : SchedAlias<WriteMAC64Hi, Swift2P03P01FiveCycle>; + def : ReadAdvance<ReadMUL, 0>; + def : SchedAlias<ReadMAC, SwiftReadAdvanceFourCyclesPred>; + + // 4.2.15 Integer Multiply Accumulate, Long + // 4.2.16 Integer Multiply Accumulate, Dual + // 4.2.17 Integer Multiply Accumulate Accumulate, Long + // We are being a bit inaccurate here. + def : InstRW< [SwiftWrite5Cycle, Swift2P03P01FiveCycle, ReadALU, ReadALU, + SchedReadAdvance<4>, SchedReadAdvance<3>], + (instregex "SMLAL", "UMLAL", "SMLALBT", + "SMLALTB", "SMLALTT", "SMLALD", "SMLALDX", "SMLSLD", "SMLSLDX", + "UMAAL", "t2SMLAL", "t2UMLAL", "t2SMLALBB", "t2SMLALBT", + "t2SMLALTB", "t2SMLALTT", "t2SMLALD", "t2SMLALDX", "t2SMLSLD", "t2SMLSLDX", + "t2UMAAL")>; + + def SwiftDiv : SchedWriteRes<[SwiftUnitP0, SwiftUnitDiv]> { + let NumMicroOps = 1; + let Latency = 14; + let ResourceCycles = [1, 14]; + } + // 4.2.18 Integer Divide + def : WriteRes<WriteDIV, [SwiftUnitDiv]>; // Workaround. + def : InstRW <[SwiftDiv], + (instregex "SDIV", "UDIV", "t2SDIV", "t2UDIV")>; + + // 4.2.19 Integer Load Single Element + // 4.2.20 Integer Load Signextended + def SwiftWriteP2P01ThreeCycle : SchedWriteRes<[SwiftUnitP2, SwiftUnitP01]> { + let Latency = 3; + let NumMicroOps = 2; + } + def SwiftWriteP2P01FourCycle : SchedWriteRes<[SwiftUnitP2, SwiftUnitP01]> { + let Latency = 4; + let NumMicroOps = 2; + } + def SwiftWriteP2P01P01FourCycle : SchedWriteRes<[SwiftUnitP2, SwiftUnitP01, + SwiftUnitP01]> { + let Latency = 4; + let NumMicroOps = 3; + } + def SwiftWriteP2P2ThreeCycle : SchedWriteRes<[SwiftUnitP2, SwiftUnitP2]> { + let Latency = 3; + let NumMicroOps = 2; + } + def SwiftWriteP2P2P01ThreeCycle : SchedWriteRes<[SwiftUnitP2, SwiftUnitP2, + SwiftUnitP01]> { + let Latency = 3; + let NumMicroOps = 3; + } + def SwiftWrBackOne : SchedWriteRes<[]> { + let Latency = 1; + let NumMicroOps = 0; + } + def SwiftWriteLdFour : SchedWriteRes<[]> { + let Latency = 4; + let NumMicroOps = 0; + } + // Not accurate. + def : InstRW<[SwiftWriteP2ThreeCycle], + (instregex "LDR(i12|rs)$", "LDRB(i12|rs)$", "t2LDR(i8|i12|s|pci)", + "t2LDR(H|B)(i8|i12|s|pci)", "LDREX", "tLDR[BH](r|i|spi|pci|pciASM)", + "tLDR(r|i|spi|pci|pciASM)")>; + def : InstRW<[SwiftWriteP2ThreeCycle], + (instregex "LDRH$", "PICLDR$", "PICLDR(H|B)$", "LDRcp$")>; + def : InstRW<[SwiftWriteP2P01FourCycle], + (instregex "PICLDRS(H|B)$", "t2LDRS(H|B)(i|r|p|s)", "LDRS(H|B)$", + "t2LDRpci_pic", "tLDRS(B|H)")>; + def : InstRW<[SwiftWriteP2P01ThreeCycle, SwiftWrBackOne], + (instregex "LD(RB|R)(_|T_)(POST|PRE)_(IMM|REG)", "LDRH(_PRE|_POST)", + "LDR(T|BT)_POST_(REG|IMM)", "LDRHT(i|r)", + "t2LD(R|RB|RH)_(PRE|POST)", "t2LD(R|RB|RH)T")>; + def : InstRW<[SwiftWriteP2P01P01FourCycle, SwiftWrBackOne], + (instregex "LDR(SH|SB)(_POST|_PRE)", "t2LDR(SH|SB)(_POST|_PRE)", + "LDRS(B|H)T(i|r)", "t2LDRS(B|H)T(i|r)?")>; + + // 4.2.21 Integer Dual Load + // Not accurate. + def : InstRW<[SwiftWriteP2P2ThreeCycle, SwiftWriteLdFour], + (instregex "t2LDRDi8", "LDRD$")>; + def : InstRW<[SwiftWriteP2P2P01ThreeCycle, SwiftWriteLdFour, SwiftWrBackOne], + (instregex "LDRD_(POST|PRE)", "t2LDRD_(POST|PRE)")>; + + // 4.2.22 Integer Load, Multiple + // NumReg = 1 .. 16 + foreach Lat = 3-25 in { + def SwiftWriteLM#Lat#Cy : SchedWriteRes<[SwiftUnitP2]> { + let Latency = Lat; + } + def SwiftWriteLM#Lat#CyNo : SchedWriteRes<[]> { + let Latency = Lat; + let NumMicroOps = 0; + } + } + // Predicate. + foreach NumAddr = 1-16 in { + def SwiftLMAddr#NumAddr#Pred : SchedPredicate<"TII->getNumLDMAddresses(*MI) == "#NumAddr>; + } + def SwiftWriteLDMAddrNoWB : SchedWriteRes<[SwiftUnitP01]> { let Latency = 0; } + def SwiftWriteLDMAddrWB : SchedWriteRes<[SwiftUnitP01, SwiftUnitP01]>; + def SwiftWriteLM : SchedWriteVariant<[ + SchedVar<SwiftLMAddr2Pred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy]>, + SchedVar<SwiftLMAddr3Pred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy]>, + SchedVar<SwiftLMAddr4Pred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy]>, + SchedVar<SwiftLMAddr5Pred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy]>, + SchedVar<SwiftLMAddr6Pred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy]>, + SchedVar<SwiftLMAddr7Pred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy]>, + SchedVar<SwiftLMAddr8Pred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy]>, + SchedVar<SwiftLMAddr9Pred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy]>, + SchedVar<SwiftLMAddr10Pred,[SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy, SwiftWriteLM12Cy]>, + SchedVar<SwiftLMAddr11Pred,[SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy, SwiftWriteLM12Cy, + SwiftWriteLM13Cy]>, + SchedVar<SwiftLMAddr12Pred,[SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy, SwiftWriteLM12Cy, + SwiftWriteLM13Cy, SwiftWriteLM14Cy]>, + SchedVar<SwiftLMAddr13Pred,[SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy, SwiftWriteLM12Cy, + SwiftWriteLM13Cy, SwiftWriteLM14Cy, + SwiftWriteLM15Cy]>, + SchedVar<SwiftLMAddr14Pred,[SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy, SwiftWriteLM12Cy, + SwiftWriteLM13Cy, SwiftWriteLM14Cy, + SwiftWriteLM15Cy, SwiftWriteLM16Cy]>, + SchedVar<SwiftLMAddr15Pred,[SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy, SwiftWriteLM12Cy, + SwiftWriteLM13Cy, SwiftWriteLM14Cy, + SwiftWriteLM15Cy, SwiftWriteLM16Cy, + SwiftWriteLM17Cy]>, + SchedVar<SwiftLMAddr16Pred,[SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5Cy, SwiftWriteLM6Cy, + SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy, SwiftWriteLM12Cy, + SwiftWriteLM13Cy, SwiftWriteLM14Cy, + SwiftWriteLM15Cy, SwiftWriteLM16Cy, + SwiftWriteLM17Cy, SwiftWriteLM18Cy]>, + // Unknow number of registers, just use resources for two registers. + SchedVar<NoSchedPred, [SwiftWriteLM3Cy, SwiftWriteLM4Cy, + SwiftWriteLM5CyNo, SwiftWriteLM6CyNo, + SwiftWriteLM7CyNo, SwiftWriteLM8CyNo, + SwiftWriteLM9CyNo, SwiftWriteLM10CyNo, + SwiftWriteLM11CyNo, SwiftWriteLM12CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM14CyNo, + SwiftWriteLM15CyNo, SwiftWriteLM16CyNo, + SwiftWriteLM17CyNo, SwiftWriteLM18CyNo]> + + ]> { let Variadic=1; } + + def : InstRW<[SwiftWriteLM, SwiftWriteLDMAddrNoWB], + (instregex "LDM(IA|DA|DB|IB)$", "t2LDM(IA|DA|DB|IB)$", + "(t|sys)LDM(IA|DA|DB|IB)$")>; + def : InstRW<[SwiftWriteLDMAddrWB, SwiftWriteLM], + (instregex /*"t2LDMIA_RET", "tLDMIA_RET", "LDMIA_RET",*/ + "LDM(IA|DA|DB|IB)_UPD", "(t2|sys|t)LDM(IA|DA|DB|IB)_UPD")>; + def : InstRW<[SwiftWriteLDMAddrWB, SwiftWriteLM, SwiftWriteP1TwoCycle], + (instregex "LDMIA_RET", "(t|t2)LDMIA_RET", "tPOP")>; + // 4.2.23 Integer Store, Single Element + def : InstRW<[SwiftWriteP2], + (instregex "PICSTR", "STR(i12|rs)", "STRB(i12|rs)", "STRH$", "STREX", + "t2STR(i12|i8|s)$", "t2STR[BH](i12|i8|s)$", "tSTR[BH](i|r)", "tSTR(i|r)", "tSTRspi")>; + + def : InstRW<[SwiftWriteP01OneCycle, SwiftWriteP2], + (instregex "STR(B_|_|BT_|T_)(PRE_IMM|PRE_REG|POST_REG|POST_IMM)", + "STR(i|r)_preidx", "STRB(i|r)_preidx", "STRH_preidx", "STR(H_|HT_)(PRE|POST)", + "STR(BT|HT|T)", "t2STR_(PRE|POST)", "t2STR[BH]_(PRE|POST)", + "t2STR_preidx", "t2STR[BH]_preidx", "t2ST(RB|RH|R)T")>; + + // 4.2.24 Integer Store, Dual + def : InstRW<[SwiftWriteP2, SwiftWriteP2, SwiftWriteP01OneCycle], + (instregex "STRD$", "t2STRDi8")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWriteP2, SwiftWriteP2, + SwiftWriteP01OneCycle], + (instregex "(t2|t)STRD_(POST|PRE)", "STRD_(POST|PRE)")>; + + // 4.2.25 Integer Store, Multiple + def SwiftWriteStIncAddr : SchedWriteRes<[SwiftUnitP2, SwiftUnitP01]> { + let Latency = 0; + let NumMicroOps = 2; + } + foreach NumAddr = 1-16 in { + def SwiftWriteSTM#NumAddr : WriteSequence<[SwiftWriteStIncAddr], NumAddr>; + } + def SwiftWriteSTM : SchedWriteVariant<[ + SchedVar<SwiftLMAddr2Pred, [SwiftWriteSTM2]>, + SchedVar<SwiftLMAddr3Pred, [SwiftWriteSTM3]>, + SchedVar<SwiftLMAddr4Pred, [SwiftWriteSTM4]>, + SchedVar<SwiftLMAddr5Pred, [SwiftWriteSTM5]>, + SchedVar<SwiftLMAddr6Pred, [SwiftWriteSTM6]>, + SchedVar<SwiftLMAddr7Pred, [SwiftWriteSTM7]>, + SchedVar<SwiftLMAddr8Pred, [SwiftWriteSTM8]>, + SchedVar<SwiftLMAddr9Pred, [SwiftWriteSTM9]>, + SchedVar<SwiftLMAddr10Pred,[SwiftWriteSTM10]>, + SchedVar<SwiftLMAddr11Pred,[SwiftWriteSTM11]>, + SchedVar<SwiftLMAddr12Pred,[SwiftWriteSTM12]>, + SchedVar<SwiftLMAddr13Pred,[SwiftWriteSTM13]>, + SchedVar<SwiftLMAddr14Pred,[SwiftWriteSTM14]>, + SchedVar<SwiftLMAddr15Pred,[SwiftWriteSTM15]>, + SchedVar<SwiftLMAddr16Pred,[SwiftWriteSTM16]>, + // Unknow number of registers, just use resources for two registers. + SchedVar<NoSchedPred, [SwiftWriteSTM2]> + ]>; + def : InstRW<[SwiftWriteSTM], + (instregex "STM(IB|IA|DB|DA)$", "(t2|sys|t)STM(IB|IA|DB|DA)$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWriteSTM], + (instregex "STM(IB|IA|DB|DA)_UPD", "(t2|sys|t)STM(IB|IA|DB|DA)_UPD", + "tPUSH")>; + + // LDRLIT pseudo instructions, they expand to LDR + PICADD + def : InstRW<[SwiftWriteP2ThreeCycle, WriteALU], + (instregex "t?LDRLIT_ga_abs", "t?LDRLIT_ga_pcrel")>; + // LDRLIT_ga_pcrel_ldr expands to LDR + PICLDR + def : InstRW<[SwiftWriteP2ThreeCycle, SwiftWriteP2ThreeCycle], + (instregex "LDRLIT_ga_pcrel_ldr")>; + + // 4.2.26 Branch + def : WriteRes<WriteBr, [SwiftUnitP1]> { let Latency = 0; } + def : WriteRes<WriteBrL, [SwiftUnitP1]> { let Latency = 2; } + def : WriteRes<WriteBrTbl, [SwiftUnitP1, SwiftUnitP2]> { let Latency = 0; } + + // 4.2.27 Not issued + def : WriteRes<WriteNoop, []> { let Latency = 0; let NumMicroOps = 0; } + def : InstRW<[WriteNoop], (instregex "t2IT", "IT")>; + + // 4.2.28 Advanced SIMD, Integer, 2 cycle + def : InstRW<[SwiftWriteP0TwoCycle], + (instregex "VADDv", "VSUBv", "VNEG(s|f|v)", "VADDL", "VSUBL", + "VADDW", "VSUBW", "VHADD", "VHSUB", "VRHADD", "VPADDi", + "VPADDL", "VAND", "VBIC", "VEOR", "VORN", "VORR", "VTST", + "VSHL", "VSHR(s|u)", "VSHLL", "VQSHL(s|u)", "VBIF", + "VBIT", "VBSL", "VSLI", "VSRI", "VCLS", "VCLZ", "VCNT")>; + + def : InstRW<[SwiftWriteP1TwoCycle], + (instregex "VEXT", "VREV16", "VREV32", "VREV64")>; + + // 4.2.29 Advanced SIMD, Integer, 4 cycle + // 4.2.30 Advanced SIMD, Integer with Accumulate + def : InstRW<[SwiftWriteP0FourCycle], + (instregex "VABA", "VABAL", "VPADAL", "VRSRA", "VSRA", "VACGE", "VACGT", + "VCEQ", "VCGE", "VCGT", "VCLE", "VCLT", "VRSHL", + "VQRSHL", "VRSHR(u|s)", "VABS(f|v)", "VQABS", "VQNEG", "VQADD", + "VQSUB")>; + def : InstRW<[SwiftWriteP1FourCycle], + (instregex "VRECPE", "VRSQRTE")>; + + // 4.2.31 Advanced SIMD, Add and Shift with Narrow + def : InstRW<[SwiftWriteP0P1FourCycle], + (instregex "VADDHN", "VSUBHN", "VSHRN")>; + def : InstRW<[SwiftWriteP0P1SixCycle], + (instregex "VRADDHN", "VRSUBHN", "VRSHRN", "VQSHRN", "VQSHRUN", + "VQRSHRN", "VQRSHRUN")>; + + // 4.2.32 Advanced SIMD, Vector Table Lookup + foreach Num = 1-4 in { + def SwiftWrite#Num#xP1TwoCycle : WriteSequence<[SwiftWriteP1TwoCycle], Num>; + } + def : InstRW<[SwiftWrite1xP1TwoCycle], + (instregex "VTB(L|X)1")>; + def : InstRW<[SwiftWrite2xP1TwoCycle], + (instregex "VTB(L|X)2")>; + def : InstRW<[SwiftWrite3xP1TwoCycle], + (instregex "VTB(L|X)3")>; + def : InstRW<[SwiftWrite4xP1TwoCycle], + (instregex "VTB(L|X)4")>; + + // 4.2.33 Advanced SIMD, Transpose + def : InstRW<[SwiftWriteP1FourCycle, SwiftWriteP1FourCycle, + SwiftWriteP1TwoCycle/*RsrcOnly*/, SchedReadAdvance<2>], + (instregex "VSWP", "VTRN", "VUZP", "VZIP")>; + + // 4.2.34 Advanced SIMD and VFP, Floating Point + def : InstRW<[SwiftWriteP0TwoCycle], (instregex "VABS(S|D)$", "VNEG(S|D)$")>; + def : InstRW<[SwiftWriteP0FourCycle], + (instregex "VCMP(D|S|ZD|ZS)$", "VCMPE(D|S|ZD|ZS)")>; + def : InstRW<[SwiftWriteP0FourCycle], + (instregex "VADD(S|f)", "VSUB(S|f)", "VABD", "VPADDf", "VMAX", "VMIN", "VPMAX", + "VPMIN")>; + def : InstRW<[SwiftWriteP0SixCycle], (instregex "VADDD$", "VSUBD$")>; + def : InstRW<[SwiftWriteP1EightCycle], (instregex "VRECPS", "VRSQRTS")>; + + // 4.2.35 Advanced SIMD and VFP, Multiply + def : InstRW<[SwiftWriteP1FourCycle], + (instregex "VMUL(S|v|p|f|s)", "VNMULS", "VQDMULH", "VQRDMULH", + "VMULL", "VQDMULL")>; + def : InstRW<[SwiftWriteP1FourCycle], + (instregex "VMLA", "VMLS", "VNMLA", "VNMLS", "VFMA(S|D)", "VFMS(S|D)", + "VFNMA", "VFNMS", "VMLAL", "VMLSL","VQDMLAL", "VQDMLSL")>; + def : InstRW<[SwiftWriteP1EightCycle], (instregex "VFMAfd", "VFMSfd")>; + def : InstRW<[SwiftWriteP1TwelveCyc], (instregex "VFMAfq", "VFMSfq")>; + + // 4.2.36 Advanced SIMD and VFP, Convert + def : InstRW<[SwiftWriteP1FourCycle], (instregex "VCVT", "V(S|U)IT", "VTO(S|U)")>; + + // 4.2.37 Advanced SIMD and VFP, Move + def : InstRW<[SwiftWriteP0TwoCycle], + (instregex "VMOVv", "VMOV(S|D)$", "VMOV(S|D)cc", + "VMVNv", "VMVN(d|q)", + "FCONST(D|S)")>; + def : InstRW<[SwiftWriteP1TwoCycle], (instregex "VMOVN", "VMOVL")>; + def : InstRW<[WriteSequence<[SwiftWriteP0FourCycle, SwiftWriteP1TwoCycle]>], + (instregex "VQMOVN")>; + def : InstRW<[SwiftWriteP1TwoCycle], (instregex "VDUPLN")>; + def : InstRW<[WriteSequence<[SwiftWriteP2FourCycle, SwiftWriteP1TwoCycle]>], + (instregex "VDUP(8|16|32)")>; + def : InstRW<[SwiftWriteP2ThreeCycle], (instregex "VMOVRS$")>; + def : InstRW<[WriteSequence<[SwiftWriteP2FourCycle, SwiftWriteP0TwoCycle]>], + (instregex "VMOVSR$", "VSETLN")>; + def : InstRW<[SwiftWriteP2ThreeCycle, SwiftWriteP2FourCycle], + (instregex "VMOVRR(D|S)$")>; + def : InstRW<[SwiftWriteP2FourCycle], (instregex "VMOVDRR$")>; + def : InstRW<[WriteSequence<[SwiftWriteP2FourCycle, SwiftWriteP1TwoCycle]>, + WriteSequence<[SwiftWrite1Cycle, SwiftWriteP2FourCycle, + SwiftWriteP1TwoCycle]>], + (instregex "VMOVSRR$")>; + def : InstRW<[WriteSequence<[SwiftWriteP1TwoCycle, SwiftWriteP2ThreeCycle]>], + (instregex "VGETLN(u|i)")>; + def : InstRW<[WriteSequence<[SwiftWriteP1TwoCycle, SwiftWriteP2ThreeCycle, + SwiftWriteP01OneCycle]>], + (instregex "VGETLNs")>; + + // 4.2.38 Advanced SIMD and VFP, Move FPSCR + // Serializing instructions. + def SwiftWaitP0For15Cy : SchedWriteRes<[SwiftUnitP0]> { + let Latency = 15; + let ResourceCycles = [15]; + } + def SwiftWaitP1For15Cy : SchedWriteRes<[SwiftUnitP1]> { + let Latency = 15; + let ResourceCycles = [15]; + } + def SwiftWaitP2For15Cy : SchedWriteRes<[SwiftUnitP2]> { + let Latency = 15; + let ResourceCycles = [15]; + } + def : InstRW<[SwiftWaitP0For15Cy, SwiftWaitP1For15Cy, SwiftWaitP2For15Cy], + (instregex "VMRS")>; + def : InstRW<[SwiftWaitP0For15Cy, SwiftWaitP1For15Cy, SwiftWaitP2For15Cy], + (instregex "VMSR")>; + // Not serializing. + def : InstRW<[SwiftWriteP0TwoCycle], (instregex "FMSTAT")>; + + // 4.2.39 Advanced SIMD and VFP, Load Single Element + def : InstRW<[SwiftWriteLM4Cy], (instregex "VLDRD$", "VLDRS$")>; + + // 4.2.40 Advanced SIMD and VFP, Store Single Element + def : InstRW<[SwiftWriteLM4Cy], (instregex "VSTRD$", "VSTRS$")>; + + // 4.2.41 Advanced SIMD and VFP, Load Multiple + // 4.2.42 Advanced SIMD and VFP, Store Multiple + + // Resource requirement for permuting, just reserves the resources. + foreach Num = 1-28 in { + def SwiftVLDMPerm#Num : SchedWriteRes<[SwiftUnitP1]> { + let Latency = 0; + let NumMicroOps = Num; + let ResourceCycles = [Num]; + } + } + + // Pre RA pseudos - load/store to a Q register as a D register pair. + def : InstRW<[SwiftWriteLM4Cy], (instregex "VLDMQIA$", "VSTMQIA$")>; + + // Post RA not modelled accurately. We assume that register use of width 64 + // bit maps to a D register, 128 maps to a Q register. Not all different kinds + // are accurately represented. + def SwiftWriteVLDM : SchedWriteVariant<[ + // Load of one S register. + SchedVar<SwiftLMAddr1Pred, [SwiftWriteLM4Cy]>, + // Load of one D register. + SchedVar<SwiftLMAddr2Pred, [SwiftWriteLM4Cy, SwiftWriteLM4CyNo]>, + // Load of 3 S register. + SchedVar<SwiftLMAddr3Pred, [SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM13CyNo, SwiftWriteP01OneCycle, + SwiftVLDMPerm3]>, + // Load of a Q register (not necessarily true). We should not be mapping to + // 4 S registers, either. + SchedVar<SwiftLMAddr4Pred, [SwiftWriteLM4Cy, SwiftWriteLM4CyNo, + SwiftWriteLM4CyNo, SwiftWriteLM4CyNo]>, + // Load of 5 S registers. + SchedVar<SwiftLMAddr5Pred, [SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM13CyNo, SwiftWriteLM14CyNo, + SwiftWriteLM17CyNo, SwiftWriteP01OneCycle, + SwiftVLDMPerm5]>, + // Load of 3 D registers. (Must also be able to handle s register list - + // though, not accurate) + SchedVar<SwiftLMAddr6Pred, [SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM10Cy, SwiftWriteLM14CyNo, + SwiftWriteLM14CyNo, SwiftWriteLM14CyNo, + SwiftWriteP01OneCycle, SwiftVLDMPerm5]>, + // Load of 7 S registers. + SchedVar<SwiftLMAddr7Pred, [SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM13Cy, SwiftWriteLM14CyNo, + SwiftWriteLM17CyNo, SwiftWriteLM18CyNo, + SwiftWriteLM21CyNo, SwiftWriteP01OneCycle, + SwiftVLDMPerm7]>, + // Load of two Q registers. + SchedVar<SwiftLMAddr8Pred, [SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM13Cy, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteP01OneCycle, SwiftVLDMPerm2]>, + // Load of 9 S registers. + SchedVar<SwiftLMAddr9Pred, [SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM13Cy, SwiftWriteLM14CyNo, + SwiftWriteLM17CyNo, SwiftWriteLM18CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM25CyNo, SwiftWriteP01OneCycle, + SwiftVLDMPerm9]>, + // Load of 5 D registers. + SchedVar<SwiftLMAddr10Pred,[SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM10Cy, SwiftWriteLM14Cy, + SwiftWriteLM14CyNo, SwiftWriteLM14CyNo, + SwiftWriteLM14CyNo, SwiftWriteLM14CyNo, + SwiftWriteLM14CyNo, SwiftWriteLM14CyNo, + SwiftWriteP01OneCycle, SwiftVLDMPerm5]>, + // Inaccurate: reuse describtion from 9 S registers. + SchedVar<SwiftLMAddr11Pred,[SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM13Cy, SwiftWriteLM14CyNo, + SwiftWriteLM17CyNo, SwiftWriteLM18CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM25CyNo, SwiftWriteP01OneCycle, + SwiftVLDMPerm9]>, + // Load of three Q registers. + SchedVar<SwiftLMAddr12Pred,[SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM11Cy, SwiftWriteLM11Cy, + SwiftWriteLM11CyNo, SwiftWriteLM11CyNo, + SwiftWriteLM11CyNo, SwiftWriteLM11CyNo, + SwiftWriteLM11CyNo, SwiftWriteLM11CyNo, + SwiftWriteLM11CyNo, SwiftWriteLM11CyNo, + SwiftWriteP01OneCycle, SwiftVLDMPerm3]>, + // Inaccurate: reuse describtion from 9 S registers. + SchedVar<SwiftLMAddr13Pred, [SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM13Cy, SwiftWriteLM14CyNo, + SwiftWriteLM17CyNo, SwiftWriteLM18CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM25CyNo, SwiftWriteP01OneCycle, + SwiftVLDMPerm9]>, + // Load of 7 D registers inaccurate. + SchedVar<SwiftLMAddr14Pred,[SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM10Cy, SwiftWriteLM14Cy, + SwiftWriteLM14Cy, SwiftWriteLM14CyNo, + SwiftWriteLM14CyNo, SwiftWriteLM14CyNo, + SwiftWriteLM14CyNo, SwiftWriteLM14CyNo, + SwiftWriteLM14CyNo, SwiftWriteLM14CyNo, + SwiftWriteP01OneCycle, SwiftVLDMPerm7]>, + SchedVar<SwiftLMAddr15Pred,[SwiftWriteLM9Cy, SwiftWriteLM10Cy, + SwiftWriteLM13Cy, SwiftWriteLM14Cy, + SwiftWriteLM17Cy, SwiftWriteLM18CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM21CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM25CyNo, SwiftWriteP01OneCycle, + SwiftVLDMPerm9]>, + // Load of 4 Q registers. + SchedVar<SwiftLMAddr16Pred,[SwiftWriteLM7Cy, SwiftWriteLM10Cy, + SwiftWriteLM11Cy, SwiftWriteLM14Cy, + SwiftWriteLM15Cy, SwiftWriteLM18CyNo, + SwiftWriteLM19CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM19CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM19CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM19CyNo, SwiftWriteLM22CyNo, + SwiftWriteLM19CyNo, SwiftWriteLM22CyNo, + SwiftWriteP01OneCycle, SwiftVLDMPerm4]>, + // Unknow number of registers, just use resources for two registers. + SchedVar<NoSchedPred, [SwiftWriteLM7Cy, SwiftWriteLM8Cy, + SwiftWriteLM13Cy, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteLM13CyNo, SwiftWriteLM13CyNo, + SwiftWriteP01OneCycle, SwiftVLDMPerm2]> + ]> { let Variadic = 1; } + + def : InstRW<[SwiftWriteVLDM], (instregex "VLDM[SD](IA|DB)$")>; + + def : InstRW<[SwiftWriteP01OneCycle2x, SwiftWriteVLDM], + (instregex "VLDM[SD](IA|DB)_UPD$")>; + + def SwiftWriteVSTM : SchedWriteVariant<[ + // One S register. + SchedVar<SwiftLMAddr1Pred, [SwiftWriteSTM1]>, + // One D register. + SchedVar<SwiftLMAddr2Pred, [SwiftWriteSTM1]>, + // Three S registers. + SchedVar<SwiftLMAddr3Pred, [SwiftWriteSTM4]>, + // Assume one Q register. + SchedVar<SwiftLMAddr4Pred, [SwiftWriteSTM1]>, + SchedVar<SwiftLMAddr5Pred, [SwiftWriteSTM6]>, + // Assume three D registers. + SchedVar<SwiftLMAddr6Pred, [SwiftWriteSTM4]>, + SchedVar<SwiftLMAddr7Pred, [SwiftWriteSTM8]>, + // Assume two Q registers. + SchedVar<SwiftLMAddr8Pred, [SwiftWriteSTM3]>, + SchedVar<SwiftLMAddr9Pred, [SwiftWriteSTM10]>, + // Assume 5 D registers. + SchedVar<SwiftLMAddr10Pred, [SwiftWriteSTM6]>, + SchedVar<SwiftLMAddr11Pred, [SwiftWriteSTM12]>, + // Assume three Q registers. + SchedVar<SwiftLMAddr12Pred, [SwiftWriteSTM4]>, + SchedVar<SwiftLMAddr13Pred, [SwiftWriteSTM14]>, + // Assume 7 D registers. + SchedVar<SwiftLMAddr14Pred, [SwiftWriteSTM8]>, + SchedVar<SwiftLMAddr15Pred, [SwiftWriteSTM16]>, + // Assume four Q registers. + SchedVar<SwiftLMAddr16Pred, [SwiftWriteSTM5]>, + // Asumme two Q registers. + SchedVar<NoSchedPred, [SwiftWriteSTM3]> + ]> { let Variadic = 1; } + + def : InstRW<[SwiftWriteVSTM], (instregex "VSTM[SD](IA|DB)$")>; + + def : InstRW<[SwiftWriteP01OneCycle2x, SwiftWriteVSTM], + (instregex "VSTM[SD](IA|DB)_UPD")>; + + // 4.2.43 Advanced SIMD, Element or Structure Load and Store + def SwiftWrite2xP2FourCy : SchedWriteRes<[SwiftUnitP2]> { + let Latency = 4; + let ResourceCycles = [2]; + } + def SwiftWrite3xP2FourCy : SchedWriteRes<[SwiftUnitP2]> { + let Latency = 4; + let ResourceCycles = [3]; + } + foreach Num = 1-2 in { + def SwiftExt#Num#xP0 : SchedWriteRes<[SwiftUnitP0]> { + let Latency = 0; + let NumMicroOps = Num; + let ResourceCycles = [Num]; + } + } + // VLDx + // Multiple structures. + // Single element structure loads. + // We assume aligned. + // Single/two register. + def : InstRW<[SwiftWriteLM4Cy], (instregex "VLD1(d|q)(8|16|32|64)$")>; + def : InstRW<[SwiftWriteLM4Cy, SwiftWriteP01OneCycle], + (instregex "VLD1(d|q)(8|16|32|64)wb")>; + // Three register. + def : InstRW<[SwiftWrite3xP2FourCy], + (instregex "VLD1(d|q)(8|16|32|64)T$", "VLD1d64TPseudo")>; + def : InstRW<[SwiftWrite3xP2FourCy, SwiftWriteP01OneCycle], + (instregex "VLD1(d|q)(8|16|32|64)Twb")>; + /// Four Register. + def : InstRW<[SwiftWrite2xP2FourCy], + (instregex "VLD1(d|q)(8|16|32|64)Q$", "VLD1d64QPseudo")>; + def : InstRW<[SwiftWrite2xP2FourCy, SwiftWriteP01OneCycle], + (instregex "VLD1(d|q)(8|16|32|64)Qwb")>; + // Two element structure loads. + // Two/four register. + def : InstRW<[SwiftWriteLM9Cy, SwiftExt2xP0, SwiftVLDMPerm2], + (instregex "VLD2(d|q|b)(8|16|32)$", "VLD2q(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteLM9Cy, SwiftWriteP01OneCycle, SwiftExt2xP0, + SwiftVLDMPerm2], + (instregex "VLD2(d|q|b)(8|16|32)wb", "VLD2q(8|16|32)PseudoWB")>; + // Three element structure. + def : InstRW<[SwiftWriteLM9Cy, SwiftWriteLM9CyNo, SwiftWriteLM9CyNo, + SwiftVLDMPerm3, SwiftWrite3xP2FourCy], + (instregex "VLD3(d|q)(8|16|32)$")>; + def : InstRW<[SwiftWriteLM9Cy, SwiftVLDMPerm3, SwiftWrite3xP2FourCy], + (instregex "VLD3(d|q)(8|16|32)(oddP|P)seudo$")>; + + def : InstRW<[SwiftWriteLM9Cy, SwiftWriteLM9CyNo, SwiftWriteLM9CyNo, + SwiftWriteP01OneCycle, SwiftVLDMPerm3, SwiftWrite3xP2FourCy], + (instregex "VLD3(d|q)(8|16|32)_UPD$")>; + def : InstRW<[SwiftWriteLM9Cy, SwiftWriteP01OneCycle, SwiftVLDMPerm3, + SwiftWrite3xP2FourCy], + (instregex "VLD3(d|q)(8|16|32)(oddP|P)seudo_UPD")>; + // Four element structure loads. + def : InstRW<[SwiftWriteLM11Cy, SwiftWriteLM11Cy, SwiftWriteLM11Cy, + SwiftWriteLM11Cy, SwiftExt2xP0, SwiftVLDMPerm4, + SwiftWrite3xP2FourCy], + (instregex "VLD4(d|q)(8|16|32)$")>; + def : InstRW<[SwiftWriteLM11Cy, SwiftExt2xP0, SwiftVLDMPerm4, + SwiftWrite3xP2FourCy], + (instregex "VLD4(d|q)(8|16|32)(oddP|P)seudo$")>; + def : InstRW<[SwiftWriteLM11Cy, SwiftWriteLM11Cy, SwiftWriteLM11Cy, + SwiftWriteLM11Cy, SwiftWriteP01OneCycle, SwiftExt2xP0, + SwiftVLDMPerm4, SwiftWrite3xP2FourCy], + (instregex "VLD4(d|q)(8|16|32)_UPD")>; + def : InstRW<[SwiftWriteLM11Cy, SwiftWriteP01OneCycle, SwiftExt2xP0, + SwiftVLDMPerm4, SwiftWrite3xP2FourCy], + (instregex "VLD4(d|q)(8|16|32)(oddP|P)seudo_UPD")>; + + // Single all/lane loads. + // One element structure. + def : InstRW<[SwiftWriteLM6Cy, SwiftVLDMPerm2], + (instregex "VLD1(LN|DUP)(d|q)(8|16|32)$", "VLD1(LN|DUP)(d|q)(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteLM6Cy, SwiftWriteP01OneCycle, SwiftVLDMPerm2], + (instregex "VLD1(LN|DUP)(d|q)(8|16|32)(wb|_UPD)", + "VLD1LNq(8|16|32)Pseudo_UPD")>; + // Two element structure. + def : InstRW<[SwiftWriteLM6Cy, SwiftWriteLM6Cy, SwiftExt1xP0, SwiftVLDMPerm2], + (instregex "VLD2(DUP|LN)(d|q)(8|16|32|8x2|16x2|32x2)$", + "VLD2LN(d|q)(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteLM6Cy, SwiftWriteLM6Cy, SwiftWriteP01OneCycle, + SwiftExt1xP0, SwiftVLDMPerm2], + (instregex "VLD2LN(d|q)(8|16|32)_UPD$")>; + def : InstRW<[SwiftWriteLM6Cy, SwiftWriteP01OneCycle, SwiftWriteLM6Cy, + SwiftExt1xP0, SwiftVLDMPerm2], + (instregex "VLD2DUPd(8|16|32|8x2|16x2|32x2)wb")>; + def : InstRW<[SwiftWriteLM6Cy, SwiftWriteP01OneCycle, SwiftWriteLM6Cy, + SwiftExt1xP0, SwiftVLDMPerm2], + (instregex "VLD2LN(d|q)(8|16|32)Pseudo_UPD")>; + // Three element structure. + def : InstRW<[SwiftWriteLM7Cy, SwiftWriteLM8Cy, SwiftWriteLM8Cy, SwiftExt1xP0, + SwiftVLDMPerm3], + (instregex "VLD3(DUP|LN)(d|q)(8|16|32)$", + "VLD3(LN|DUP)(d|q)(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteLM7Cy, SwiftWriteLM8Cy, SwiftWriteLM8Cy, + SwiftWriteP01OneCycle, SwiftExt1xP0, SwiftVLDMPerm3], + (instregex "VLD3(LN|DUP)(d|q)(8|16|32)_UPD")>; + def : InstRW<[SwiftWriteLM7Cy, SwiftWriteP01OneCycle, SwiftWriteLM8Cy, + SwiftWriteLM8Cy, SwiftExt1xP0, SwiftVLDMPerm3], + (instregex "VLD3(LN|DUP)(d|q)(8|16|32)Pseudo_UPD")>; + // Four element struture. + def : InstRW<[SwiftWriteLM8Cy, SwiftWriteLM9Cy, SwiftWriteLM10CyNo, + SwiftWriteLM10CyNo, SwiftExt1xP0, SwiftVLDMPerm5], + (instregex "VLD4(LN|DUP)(d|q)(8|16|32)$", + "VLD4(LN|DUP)(d|q)(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteLM8Cy, SwiftWriteLM9Cy, SwiftWriteLM10CyNo, + SwiftWriteLM10CyNo, SwiftWriteP01OneCycle, SwiftExt1xP0, + SwiftVLDMPerm5], + (instregex "VLD4(DUP|LN)(d|q)(8|16|32)_UPD")>; + def : InstRW<[SwiftWriteLM8Cy, SwiftWriteP01OneCycle, SwiftWriteLM9Cy, + SwiftWriteLM10CyNo, SwiftWriteLM10CyNo, SwiftExt1xP0, + SwiftVLDMPerm5], + (instregex "VLD4(DUP|LN)(d|q)(8|16|32)Pseudo_UPD")>; + // VSTx + // Multiple structures. + // Single element structure store. + def : InstRW<[SwiftWrite1xP2], (instregex "VST1d(8|16|32|64)$")>; + def : InstRW<[SwiftWrite2xP2], (instregex "VST1q(8|16|32|64)$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite1xP2], + (instregex "VST1d(8|16|32|64)wb")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite2xP2], + (instregex "VST1q(8|16|32|64)wb")>; + def : InstRW<[SwiftWrite3xP2], + (instregex "VST1d(8|16|32|64)T$", "VST1d64TPseudo$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite3xP2], + (instregex "VST1d(8|16|32|64)Twb", "VST1d64TPseudoWB")>; + def : InstRW<[SwiftWrite4xP2], + (instregex "VST1d(8|16|32|64)(Q|QPseudo)$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite4xP2], + (instregex "VST1d(8|16|32|64)(Qwb|QPseudoWB)")>; + // Two element structure store. + def : InstRW<[SwiftWrite1xP2, SwiftVLDMPerm1], + (instregex "VST2(d|b)(8|16|32)$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite1xP2, SwiftVLDMPerm1], + (instregex "VST2(b|d)(8|16|32)wb")>; + def : InstRW<[SwiftWrite2xP2, SwiftVLDMPerm2], + (instregex "VST2q(8|16|32)$", "VST2q(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWrite2xP2, SwiftVLDMPerm2], + (instregex "VST2q(8|16|32)wb", "VST2q(8|16|32)PseudoWB")>; + // Three element structure store. + def : InstRW<[SwiftWrite4xP2, SwiftVLDMPerm2], + (instregex "VST3(d|q)(8|16|32)$", "VST3(d|q)(8|16|32)(oddP|P)seudo$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite4xP2, SwiftVLDMPerm2], + (instregex "VST3(d|q)(8|16|32)_UPD", + "VST3(d|q)(8|16|32)(oddP|P)seudo_UPD$")>; + // Four element structure store. + def : InstRW<[SwiftWrite4xP2, SwiftVLDMPerm2], + (instregex "VST4(d|q)(8|16|32)$", "VST4(d|q)(8|16|32)(oddP|P)seudo$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite4xP2, SwiftVLDMPerm4], + (instregex "VST4(d|q)(8|16|32)_UPD", + "VST4(d|q)(8|16|32)(oddP|P)seudo_UPD$")>; + // Single/all lane store. + // One element structure. + def : InstRW<[SwiftWrite1xP2, SwiftVLDMPerm1], + (instregex "VST1LNd(8|16|32)$", "VST1LNq(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite1xP2, SwiftVLDMPerm1], + (instregex "VST1LNd(8|16|32)_UPD", "VST1LNq(8|16|32)Pseudo_UPD")>; + // Two element structure. + def : InstRW<[SwiftWrite1xP2, SwiftVLDMPerm2], + (instregex "VST2LN(d|q)(8|16|32)$", "VST2LN(d|q)(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite1xP2, SwiftVLDMPerm2], + (instregex "VST2LN(d|q)(8|16|32)_UPD", + "VST2LN(d|q)(8|16|32)Pseudo_UPD")>; + // Three element structure. + def : InstRW<[SwiftWrite4xP2, SwiftVLDMPerm2], + (instregex "VST3LN(d|q)(8|16|32)$", "VST3LN(d|q)(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite4xP2, SwiftVLDMPerm2], + (instregex "VST3LN(d|q)(8|16|32)_UPD", + "VST3LN(d|q)(8|16|32)Pseudo_UPD")>; + // Four element structure. + def : InstRW<[SwiftWrite2xP2, SwiftVLDMPerm2], + (instregex "VST4LN(d|q)(8|16|32)$", "VST4LN(d|q)(8|16|32)Pseudo$")>; + def : InstRW<[SwiftWriteP01OneCycle, SwiftWrite2xP2, SwiftVLDMPerm2], + (instregex "VST4LN(d|q)(8|16|32)_UPD", + "VST4LN(d|q)(8|16|32)Pseudo_UPD")>; + + // 4.2.44 VFP, Divide and Square Root + def SwiftDiv17 : SchedWriteRes<[SwiftUnitP0, SwiftUnitDiv]> { + let NumMicroOps = 1; + let Latency = 17; + let ResourceCycles = [1, 15]; + } + def SwiftDiv32 : SchedWriteRes<[SwiftUnitP0, SwiftUnitDiv]> { + let NumMicroOps = 1; + let Latency = 32; + let ResourceCycles = [1, 30]; + } + def : InstRW<[SwiftDiv17], (instregex "VDIVS", "VSQRTS")>; + def : InstRW<[SwiftDiv32], (instregex "VDIVD", "VSQRTD")>; + + // ===---------------------------------------------------------------------===// + // Floating-point. Map target defined SchedReadWrite to processor specific ones + // + def : SchedAlias<WriteFPCVT, SwiftWriteP1FourCycle>; + def : SchedAlias<WriteFPMOV, SwiftWriteP2ThreeCycle>; + + def : SchedAlias<WriteFPALU32, SwiftWriteP0FourCycle>; + def : SchedAlias<WriteFPALU64, SwiftWriteP0SixCycle>; + + def : SchedAlias<WriteFPMUL32, SwiftWriteP1FourCycle>; + def : SchedAlias<WriteFPMUL64, SwiftWriteP1SixCycle>; + + def : SchedAlias<WriteFPMAC32, SwiftWriteP1FourCycle>; + def : SchedAlias<WriteFPMAC64, SwiftWriteP1FourCycle>; + + def : SchedAlias<WriteFPDIV32, SwiftDiv17>; + def : SchedAlias<WriteFPSQRT32, SwiftDiv17>; + + def : SchedAlias<WriteFPDIV64, SwiftDiv32>; + def : SchedAlias<WriteFPSQRT64, SwiftDiv32>; + + def : ReadAdvance<ReadFPMUL, 0>; + def : ReadAdvance<ReadFPMAC, 0>; + + // Overriden via InstRW for this processor. + def : WriteRes<WriteVLD1, []>; + def : WriteRes<WriteVLD2, []>; + def : WriteRes<WriteVLD3, []>; + def : WriteRes<WriteVLD4, []>; + def : WriteRes<WriteVST1, []>; + def : WriteRes<WriteVST2, []>; + def : WriteRes<WriteVST3, []>; + def : WriteRes<WriteVST4, []>; + + // Not specified. + def : InstRW<[SwiftWriteP01OneCycle2x], (instregex "ABS")>; + // Preload. + def : WriteRes<WritePreLd, [SwiftUnitP2]> { let Latency = 0; + let ResourceCycles = [0]; + } + +} diff --git a/capstone/suite/synctools/tablegen/ARM/ARMScheduleV6.td b/capstone/suite/synctools/tablegen/ARM/ARMScheduleV6.td new file mode 100644 index 000000000..57d0bfb65 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMScheduleV6.td @@ -0,0 +1,300 @@ +//===-- ARMScheduleV6.td - ARM v6 Scheduling Definitions ---*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the itinerary class data for the ARM v6 processors. +// +//===----------------------------------------------------------------------===// + +// Model based on ARM1176 +// +// Functional Units +def V6_Pipe : FuncUnit; // pipeline + +// Scheduling information derived from "ARM1176JZF-S Technical Reference Manual" +// +def ARMV6Itineraries : ProcessorItineraries< + [V6_Pipe], [], [ + // + // No operand cycles + InstrItinData<IIC_iALUx , [InstrStage<1, [V6_Pipe]>]>, + // + // Binary Instructions that produce a result + InstrItinData<IIC_iALUi , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + InstrItinData<IIC_iALUr , [InstrStage<1, [V6_Pipe]>], [2, 2, 2]>, + InstrItinData<IIC_iALUsi , [InstrStage<1, [V6_Pipe]>], [2, 2, 1]>, + InstrItinData<IIC_iALUsr , [InstrStage<2, [V6_Pipe]>], [3, 3, 2, 1]>, + // + // Bitwise Instructions that produce a result + InstrItinData<IIC_iBITi , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + InstrItinData<IIC_iBITr , [InstrStage<1, [V6_Pipe]>], [2, 2, 2]>, + InstrItinData<IIC_iBITsi , [InstrStage<1, [V6_Pipe]>], [2, 2, 1]>, + InstrItinData<IIC_iBITsr , [InstrStage<2, [V6_Pipe]>], [3, 3, 2, 1]>, + // + // Unary Instructions that produce a result + InstrItinData<IIC_iUNAr , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + InstrItinData<IIC_iUNAsi , [InstrStage<1, [V6_Pipe]>], [2, 1]>, + // + // Zero and sign extension instructions + InstrItinData<IIC_iEXTr , [InstrStage<1, [V6_Pipe]>], [1, 1]>, + InstrItinData<IIC_iEXTAr , [InstrStage<1, [V6_Pipe]>], [2, 2, 1]>, + InstrItinData<IIC_iEXTAsr , [InstrStage<2, [V6_Pipe]>], [3, 3, 2, 1]>, + // + // Compare instructions + InstrItinData<IIC_iCMPi , [InstrStage<1, [V6_Pipe]>], [2]>, + InstrItinData<IIC_iCMPr , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + InstrItinData<IIC_iCMPsi , [InstrStage<1, [V6_Pipe]>], [2, 1]>, + InstrItinData<IIC_iCMPsr , [InstrStage<2, [V6_Pipe]>], [3, 2, 1]>, + // + // Test instructions + InstrItinData<IIC_iTSTi , [InstrStage<1, [V6_Pipe]>], [2]>, + InstrItinData<IIC_iTSTr , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + InstrItinData<IIC_iTSTsi , [InstrStage<1, [V6_Pipe]>], [2, 1]>, + InstrItinData<IIC_iTSTsr , [InstrStage<2, [V6_Pipe]>], [3, 2, 1]>, + // + // Move instructions, unconditional + InstrItinData<IIC_iMOVi , [InstrStage<1, [V6_Pipe]>], [2]>, + InstrItinData<IIC_iMOVr , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + InstrItinData<IIC_iMOVsi , [InstrStage<1, [V6_Pipe]>], [2, 1]>, + InstrItinData<IIC_iMOVsr , [InstrStage<2, [V6_Pipe]>], [3, 2, 1]>, + InstrItinData<IIC_iMOVix2 , [InstrStage<1, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>], [2]>, + InstrItinData<IIC_iMOVix2addpc,[InstrStage<1, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>], [3]>, + InstrItinData<IIC_iMOVix2ld , [InstrStage<1, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>], [5]>, + // + // Move instructions, conditional + InstrItinData<IIC_iCMOVi , [InstrStage<1, [V6_Pipe]>], [3]>, + InstrItinData<IIC_iCMOVr , [InstrStage<1, [V6_Pipe]>], [3, 2]>, + InstrItinData<IIC_iCMOVsi , [InstrStage<1, [V6_Pipe]>], [3, 1]>, + InstrItinData<IIC_iCMOVsr , [InstrStage<1, [V6_Pipe]>], [4, 2, 1]>, + InstrItinData<IIC_iCMOVix2 , [InstrStage<1, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>], [4]>, + // + // MVN instructions + InstrItinData<IIC_iMVNi , [InstrStage<1, [V6_Pipe]>], [2]>, + InstrItinData<IIC_iMVNr , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + InstrItinData<IIC_iMVNsi , [InstrStage<1, [V6_Pipe]>], [2, 1]>, + InstrItinData<IIC_iMVNsr , [InstrStage<2, [V6_Pipe]>], [3, 2, 1]>, + + // Integer multiply pipeline + // + InstrItinData<IIC_iMUL16 , [InstrStage<1, [V6_Pipe]>], [4, 1, 1]>, + InstrItinData<IIC_iMAC16 , [InstrStage<1, [V6_Pipe]>], [4, 1, 1, 2]>, + InstrItinData<IIC_iMUL32 , [InstrStage<2, [V6_Pipe]>], [5, 1, 1]>, + InstrItinData<IIC_iMAC32 , [InstrStage<2, [V6_Pipe]>], [5, 1, 1, 2]>, + InstrItinData<IIC_iMUL64 , [InstrStage<3, [V6_Pipe]>], [6, 1, 1]>, + InstrItinData<IIC_iMAC64 , [InstrStage<3, [V6_Pipe]>], [6, 1, 1, 2]>, + + // Integer load pipeline + // + // Immediate offset + InstrItinData<IIC_iLoad_i , [InstrStage<1, [V6_Pipe]>], [4, 1]>, + InstrItinData<IIC_iLoad_bh_i, [InstrStage<1, [V6_Pipe]>], [4, 1]>, + InstrItinData<IIC_iLoad_d_i , [InstrStage<1, [V6_Pipe]>], [4, 1]>, + // + // Register offset + InstrItinData<IIC_iLoad_r , [InstrStage<1, [V6_Pipe]>], [4, 1, 1]>, + InstrItinData<IIC_iLoad_bh_r, [InstrStage<1, [V6_Pipe]>], [4, 1, 1]>, + InstrItinData<IIC_iLoad_d_r , [InstrStage<1, [V6_Pipe]>], [4, 1, 1]>, + // + // Scaled register offset, issues over 2 cycles + InstrItinData<IIC_iLoad_si , [InstrStage<2, [V6_Pipe]>], [5, 2, 1]>, + InstrItinData<IIC_iLoad_bh_si, [InstrStage<2, [V6_Pipe]>], [5, 2, 1]>, + // + // Immediate offset with update + InstrItinData<IIC_iLoad_iu , [InstrStage<1, [V6_Pipe]>], [4, 2, 1]>, + InstrItinData<IIC_iLoad_bh_iu, [InstrStage<1, [V6_Pipe]>], [4, 2, 1]>, + // + // Register offset with update + InstrItinData<IIC_iLoad_ru , [InstrStage<1, [V6_Pipe]>], [4, 2, 1, 1]>, + InstrItinData<IIC_iLoad_bh_ru, [InstrStage<1, [V6_Pipe]>], [4, 2, 1, 1]>, + InstrItinData<IIC_iLoad_d_ru , [InstrStage<1, [V6_Pipe]>], [4, 2, 1, 1]>, + // + // Scaled register offset with update, issues over 2 cycles + InstrItinData<IIC_iLoad_siu, [InstrStage<2, [V6_Pipe]>], [5, 2, 2, 1]>, + InstrItinData<IIC_iLoad_bh_siu,[InstrStage<2, [V6_Pipe]>], [5, 2, 2, 1]>, + + // + // Load multiple, def is the 5th operand. + InstrItinData<IIC_iLoad_m , [InstrStage<3, [V6_Pipe]>], [1, 1, 1, 1, 4]>, + // + // Load multiple + update, defs are the 1st and 5th operands. + InstrItinData<IIC_iLoad_mu , [InstrStage<3, [V6_Pipe]>], [2, 1, 1, 1, 4]>, + // + // Load multiple plus branch + InstrItinData<IIC_iLoad_mBr, [InstrStage<3, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>], [1, 2, 1, 1, 4]>, + + // + // iLoadi + iALUr for t2LDRpci_pic. + InstrItinData<IIC_iLoadiALU, [InstrStage<1, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>], [3, 1]>, + + // + // Pop, def is the 3rd operand. + InstrItinData<IIC_iPop , [InstrStage<3, [V6_Pipe]>], [1, 1, 4]>, + // + // Pop + branch, def is the 3rd operand. + InstrItinData<IIC_iPop_Br, [InstrStage<3, [V6_Pipe]>, + InstrStage<1, [V6_Pipe]>], [1, 2, 4]>, + + // Integer store pipeline + // + // Immediate offset + InstrItinData<IIC_iStore_i , [InstrStage<1, [V6_Pipe]>], [2, 1]>, + InstrItinData<IIC_iStore_bh_i, [InstrStage<1, [V6_Pipe]>], [2, 1]>, + InstrItinData<IIC_iStore_d_i , [InstrStage<1, [V6_Pipe]>], [2, 1]>, + // + // Register offset + InstrItinData<IIC_iStore_r , [InstrStage<1, [V6_Pipe]>], [2, 1, 1]>, + InstrItinData<IIC_iStore_bh_r, [InstrStage<1, [V6_Pipe]>], [2, 1, 1]>, + InstrItinData<IIC_iStore_d_r , [InstrStage<1, [V6_Pipe]>], [2, 1, 1]>, + // + // Scaled register offset, issues over 2 cycles + InstrItinData<IIC_iStore_si , [InstrStage<2, [V6_Pipe]>], [2, 2, 1]>, + InstrItinData<IIC_iStore_bh_si, [InstrStage<2, [V6_Pipe]>], [2, 2, 1]>, + // + // Immediate offset with update + InstrItinData<IIC_iStore_iu , [InstrStage<1, [V6_Pipe]>], [2, 2, 1]>, + InstrItinData<IIC_iStore_bh_iu, [InstrStage<1, [V6_Pipe]>], [2, 2, 1]>, + // + // Register offset with update + InstrItinData<IIC_iStore_ru, [InstrStage<1, [V6_Pipe]>], [2, 2, 1, 1]>, + InstrItinData<IIC_iStore_bh_ru,[InstrStage<1, [V6_Pipe]>], [2, 2, 1, 1]>, + InstrItinData<IIC_iStore_d_ru, [InstrStage<1, [V6_Pipe]>], [2, 2, 1, 1]>, + // + // Scaled register offset with update, issues over 2 cycles + InstrItinData<IIC_iStore_siu, [InstrStage<2, [V6_Pipe]>], [2, 2, 2, 1]>, + InstrItinData<IIC_iStore_bh_siu,[InstrStage<2, [V6_Pipe]>], [2, 2, 2, 1]>, + // + // Store multiple + InstrItinData<IIC_iStore_m , [InstrStage<3, [V6_Pipe]>]>, + // + // Store multiple + update + InstrItinData<IIC_iStore_mu , [InstrStage<3, [V6_Pipe]>], [2]>, + + // Branch + // + // no delay slots, so the latency of a branch is unimportant + InstrItinData<IIC_Br , [InstrStage<1, [V6_Pipe]>]>, + + // VFP + // Issue through integer pipeline, and execute in NEON unit. We assume + // RunFast mode so that NFP pipeline is used for single-precision when + // possible. + // + // FP Special Register to Integer Register File Move + InstrItinData<IIC_fpSTAT , [InstrStage<1, [V6_Pipe]>], [3]>, + // + // Single-precision FP Unary + InstrItinData<IIC_fpUNA32 , [InstrStage<1, [V6_Pipe]>], [5, 2]>, + // + // Double-precision FP Unary + InstrItinData<IIC_fpUNA64 , [InstrStage<1, [V6_Pipe]>], [5, 2]>, + // + // Single-precision FP Compare + InstrItinData<IIC_fpCMP32 , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + // + // Double-precision FP Compare + InstrItinData<IIC_fpCMP64 , [InstrStage<1, [V6_Pipe]>], [2, 2]>, + // + // Single to Double FP Convert + InstrItinData<IIC_fpCVTSD , [InstrStage<1, [V6_Pipe]>], [5, 2]>, + // + // Double to Single FP Convert + InstrItinData<IIC_fpCVTDS , [InstrStage<1, [V6_Pipe]>], [5, 2]>, + // + // Single-Precision FP to Integer Convert + InstrItinData<IIC_fpCVTSI , [InstrStage<1, [V6_Pipe]>], [9, 2]>, + // + // Double-Precision FP to Integer Convert + InstrItinData<IIC_fpCVTDI , [InstrStage<1, [V6_Pipe]>], [9, 2]>, + // + // Integer to Single-Precision FP Convert + InstrItinData<IIC_fpCVTIS , [InstrStage<1, [V6_Pipe]>], [9, 2]>, + // + // Integer to Double-Precision FP Convert + InstrItinData<IIC_fpCVTID , [InstrStage<1, [V6_Pipe]>], [9, 2]>, + // + // Single-precision FP ALU + InstrItinData<IIC_fpALU32 , [InstrStage<1, [V6_Pipe]>], [9, 2, 2]>, + // + // Double-precision FP ALU + InstrItinData<IIC_fpALU64 , [InstrStage<1, [V6_Pipe]>], [9, 2, 2]>, + // + // Single-precision FP Multiply + InstrItinData<IIC_fpMUL32 , [InstrStage<1, [V6_Pipe]>], [9, 2, 2]>, + // + // Double-precision FP Multiply + InstrItinData<IIC_fpMUL64 , [InstrStage<2, [V6_Pipe]>], [9, 2, 2]>, + // + // Single-precision FP MAC + InstrItinData<IIC_fpMAC32 , [InstrStage<1, [V6_Pipe]>], [9, 2, 2, 2]>, + // + // Double-precision FP MAC + InstrItinData<IIC_fpMAC64 , [InstrStage<2, [V6_Pipe]>], [9, 2, 2, 2]>, + // + // Single-precision Fused FP MAC + InstrItinData<IIC_fpFMAC32, [InstrStage<1, [V6_Pipe]>], [9, 2, 2, 2]>, + // + // Double-precision Fused FP MAC + InstrItinData<IIC_fpFMAC64, [InstrStage<2, [V6_Pipe]>], [9, 2, 2, 2]>, + // + // Single-precision FP DIV + InstrItinData<IIC_fpDIV32 , [InstrStage<15, [V6_Pipe]>], [20, 2, 2]>, + // + // Double-precision FP DIV + InstrItinData<IIC_fpDIV64 , [InstrStage<29, [V6_Pipe]>], [34, 2, 2]>, + // + // Single-precision FP SQRT + InstrItinData<IIC_fpSQRT32 , [InstrStage<15, [V6_Pipe]>], [20, 2, 2]>, + // + // Double-precision FP SQRT + InstrItinData<IIC_fpSQRT64 , [InstrStage<29, [V6_Pipe]>], [34, 2, 2]>, + // + // Integer to Single-precision Move + InstrItinData<IIC_fpMOVIS, [InstrStage<1, [V6_Pipe]>], [10, 1]>, + // + // Integer to Double-precision Move + InstrItinData<IIC_fpMOVID, [InstrStage<1, [V6_Pipe]>], [10, 1, 1]>, + // + // Single-precision to Integer Move + InstrItinData<IIC_fpMOVSI, [InstrStage<1, [V6_Pipe]>], [10, 1]>, + // + // Double-precision to Integer Move + InstrItinData<IIC_fpMOVDI, [InstrStage<1, [V6_Pipe]>], [10, 10, 1]>, + // + // Single-precision FP Load + InstrItinData<IIC_fpLoad32 , [InstrStage<1, [V6_Pipe]>], [5, 2, 2]>, + // + // Double-precision FP Load + InstrItinData<IIC_fpLoad64 , [InstrStage<1, [V6_Pipe]>], [5, 2, 2]>, + // + // FP Load Multiple + InstrItinData<IIC_fpLoad_m , [InstrStage<3, [V6_Pipe]>], [2, 1, 1, 5]>, + // + // FP Load Multiple + update + InstrItinData<IIC_fpLoad_mu, [InstrStage<3, [V6_Pipe]>], [3, 2, 1, 1, 5]>, + // + // Single-precision FP Store + InstrItinData<IIC_fpStore32 , [InstrStage<1, [V6_Pipe]>], [2, 2, 2]>, + // + // Double-precision FP Store + // use FU_Issue to enforce the 1 load/store per cycle limit + InstrItinData<IIC_fpStore64 , [InstrStage<1, [V6_Pipe]>], [2, 2, 2]>, + // + // FP Store Multiple + InstrItinData<IIC_fpStore_m, [InstrStage<3, [V6_Pipe]>], [2, 2, 2, 2]>, + // + // FP Store Multiple + update + InstrItinData<IIC_fpStore_mu,[InstrStage<3, [V6_Pipe]>], [3, 2, 2, 2, 2]> +]>; diff --git a/capstone/suite/synctools/tablegen/ARM/ARMSystemRegister.td b/capstone/suite/synctools/tablegen/ARM/ARMSystemRegister.td new file mode 100644 index 000000000..ad1d37168 --- /dev/null +++ b/capstone/suite/synctools/tablegen/ARM/ARMSystemRegister.td @@ -0,0 +1,156 @@ +//===-- ARMSystemRegister.td - ARM Register defs -------------*- tablegen -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +include "llvm/TableGen/SearchableTable.td" + +//===----------------------------------------------------------------------===// +// Declarations that describe the ARM system-registers +//===----------------------------------------------------------------------===// + +// M-Class System Registers. +// 'Mask' bits create unique keys for searches. +// +class MClassSysReg<bits<1> UniqMask1, + bits<1> UniqMask2, + bits<1> UniqMask3, + bits<12> Enc12, + string name> : SearchableTable { + let SearchableFields = ["Name", "M1Encoding12", "M2M3Encoding8", "Encoding"]; + string Name; + bits<13> M1Encoding12; + bits<10> M2M3Encoding8; + bits<12> Encoding; + + let Name = name; + let EnumValueField = "M1Encoding12"; + let EnumValueField = "M2M3Encoding8"; + let EnumValueField = "Encoding"; + + let M1Encoding12{12} = UniqMask1; + let M1Encoding12{11-00} = Enc12; + let Encoding = Enc12; + + let M2M3Encoding8{9} = UniqMask2; + let M2M3Encoding8{8} = UniqMask3; + let M2M3Encoding8{7-0} = Enc12{7-0}; + code Requires = [{ {} }]; +} + +// [|i|e|x]apsr_nzcvq has alias [|i|e|x]apsr. +// Mask1 Mask2 Mask3 Enc12, Name +let Requires = [{ {ARM::FeatureDSP} }] in { +def : MClassSysReg<0, 0, 0, 0x400, "apsr_g">; +def : MClassSysReg<0, 1, 1, 0xc00, "apsr_nzcvqg">; +def : MClassSysReg<0, 0, 0, 0x401, "iapsr_g">; +def : MClassSysReg<0, 1, 1, 0xc01, "iapsr_nzcvqg">; +def : MClassSysReg<0, 0, 0, 0x402, "eapsr_g">; +def : MClassSysReg<0, 1, 1, 0xc02, "eapsr_nzcvqg">; +def : MClassSysReg<0, 0, 0, 0x403, "xpsr_g">; +def : MClassSysReg<0, 1, 1, 0xc03, "xpsr_nzcvqg">; +} + +def : MClassSysReg<0, 0, 1, 0x800, "apsr">; +def : MClassSysReg<1, 1, 0, 0x800, "apsr_nzcvq">; +def : MClassSysReg<0, 0, 1, 0x801, "iapsr">; +def : MClassSysReg<1, 1, 0, 0x801, "iapsr_nzcvq">; +def : MClassSysReg<0, 0, 1, 0x802, "eapsr">; +def : MClassSysReg<1, 1, 0, 0x802, "eapsr_nzcvq">; +def : MClassSysReg<0, 0, 1, 0x803, "xpsr">; +def : MClassSysReg<1, 1, 0, 0x803, "xpsr_nzcvq">; + +def : MClassSysReg<0, 0, 1, 0x805, "ipsr">; +def : MClassSysReg<0, 0, 1, 0x806, "epsr">; +def : MClassSysReg<0, 0, 1, 0x807, "iepsr">; +def : MClassSysReg<0, 0, 1, 0x808, "msp">; +def : MClassSysReg<0, 0, 1, 0x809, "psp">; + +let Requires = [{ {ARM::HasV8MBaselineOps} }] in { +def : MClassSysReg<0, 0, 1, 0x80a, "msplim">; +def : MClassSysReg<0, 0, 1, 0x80b, "psplim">; +} + +def : MClassSysReg<0, 0, 1, 0x810, "primask">; + +let Requires = [{ {ARM::HasV7Ops} }] in { +def : MClassSysReg<0, 0, 1, 0x811, "basepri">; +def : MClassSysReg<0, 0, 1, 0x812, "basepri_max">; +def : MClassSysReg<0, 0, 1, 0x813, "faultmask">; +} + +def : MClassSysReg<0, 0, 1, 0x814, "control">; + +let Requires = [{ {ARM::Feature8MSecExt} }] in { +def : MClassSysReg<0, 0, 1, 0x888, "msp_ns">; +def : MClassSysReg<0, 0, 1, 0x889, "psp_ns">; +} + +let Requires = [{ {ARM::Feature8MSecExt, ARM::HasV8MBaselineOps} }] in { +def : MClassSysReg<0, 0, 1, 0x88a, "msplim_ns">; +def : MClassSysReg<0, 0, 1, 0x88b, "psplim_ns">; +} + +def : MClassSysReg<0, 0, 1, 0x890, "primask_ns">; + +let Requires = [{ {ARM::Feature8MSecExt, ARM::HasV7Ops} }] in { +def : MClassSysReg<0, 0, 1, 0x891, "basepri_ns">; +def : MClassSysReg<0, 0, 1, 0x893, "faultmask_ns">; +} + +let Requires = [{ {ARM::Feature8MSecExt} }] in { +def : MClassSysReg<0, 0, 1, 0x894, "control_ns">; +def : MClassSysReg<0, 0, 1, 0x898, "sp_ns">; +} + + +// Banked Registers +// +class BankedReg<string name, bits<8> enc> + : SearchableTable { + string Name; + bits<8> Encoding; + let Name = name; + let Encoding = enc; + let SearchableFields = ["Name", "Encoding"]; +} + +// The values here come from B9.2.3 of the ARM ARM, where bits 4-0 are SysM +// and bit 5 is R. +def : BankedReg<"r8_usr", 0x00>; +def : BankedReg<"r9_usr", 0x01>; +def : BankedReg<"r10_usr", 0x02>; +def : BankedReg<"r11_usr", 0x03>; +def : BankedReg<"r12_usr", 0x04>; +def : BankedReg<"sp_usr", 0x05>; +def : BankedReg<"lr_usr", 0x06>; +def : BankedReg<"r8_fiq", 0x08>; +def : BankedReg<"r9_fiq", 0x09>; +def : BankedReg<"r10_fiq", 0x0a>; +def : BankedReg<"r11_fiq", 0x0b>; +def : BankedReg<"r12_fiq", 0x0c>; +def : BankedReg<"sp_fiq", 0x0d>; +def : BankedReg<"lr_fiq", 0x0e>; +def : BankedReg<"lr_irq", 0x10>; +def : BankedReg<"sp_irq", 0x11>; +def : BankedReg<"lr_svc", 0x12>; +def : BankedReg<"sp_svc", 0x13>; +def : BankedReg<"lr_abt", 0x14>; +def : BankedReg<"sp_abt", 0x15>; +def : BankedReg<"lr_und", 0x16>; +def : BankedReg<"sp_und", 0x17>; +def : BankedReg<"lr_mon", 0x1c>; +def : BankedReg<"sp_mon", 0x1d>; +def : BankedReg<"elr_hyp", 0x1e>; +def : BankedReg<"sp_hyp", 0x1f>; +def : BankedReg<"spsr_fiq", 0x2e>; +def : BankedReg<"spsr_irq", 0x30>; +def : BankedReg<"spsr_svc", 0x32>; +def : BankedReg<"spsr_abt", 0x34>; +def : BankedReg<"spsr_und", 0x36>; +def : BankedReg<"spsr_mon", 0x3c>; +def : BankedReg<"spsr_hyp", 0x3e>; |