1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
|
From 98081ba7f59438b1c8f9176b5467b924b9b392f5 Mon Sep 17 00:00:00 2001
From: Nikita Yushchenko <nikita.yoush@cogentembedded.com>
Date: Sun, 24 Jun 2018 18:04:23 +0300
Subject: [PATCH 006/122] arm64: dma: check parent bus restrictions in
dma_capable()
It may happen that although hardware device is capable of full 64-bit
addressing in DMA operations, way device is connected to the system
imposes addressing limitations. Example: NVMe device connected to PCIe
host sitting on AXI bus with 32-bit addressing (and no iommu configured).
There is no agreement on setting dma_mask in this case. For low level,
dma_mask must be 32-bit - that's what hardware provides for software.
But for higher level, dma_mask must be 64-bit - that's what platform
(hardware + swiotlb) provides for higher levels.
dma_capable() is swiotlb's check if current address is DMAable by device
directly or swiotlb should provide bounce buffer for it. Using dma_mask
in dma_capable() assumes that dma_mask is set for low level - which does
not always match reality.
This patch changes dma_capable() to check additional restrictions.
It is a workaround against ambiguous dma_mask semantics.
Signed-off-by: Nikita Yushchenko <nikita.yoush@cogentembedded.com>
---
arch/arm64/include/asm/device.h | 1 +
arch/arm64/include/asm/dma-mapping.h | 9 ++++++++-
arch/arm64/mm/dma-mapping.c | 13 ++++++++++++-
3 files changed, 21 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/include/asm/device.h b/arch/arm64/include/asm/device.h
index 5a5fa47..ac3eb0c 100644
--- a/arch/arm64/include/asm/device.h
+++ b/arch/arm64/include/asm/device.h
@@ -24,6 +24,7 @@ struct dev_archdata {
const struct dma_map_ops *dev_dma_ops;
#endif
bool dma_coherent;
+ u64 parent_dma_mask;
};
struct pdev_archdata {
diff --git a/arch/arm64/include/asm/dma-mapping.h b/arch/arm64/include/asm/dma-mapping.h
index 0df756b..15edb0c 100644
--- a/arch/arm64/include/asm/dma-mapping.h
+++ b/arch/arm64/include/asm/dma-mapping.h
@@ -66,10 +66,17 @@ static inline phys_addr_t dma_to_phys(struct device *dev, dma_addr_t dev_addr)
static inline bool dma_capable(struct device *dev, dma_addr_t addr, size_t size)
{
+ u64 mask;
+
if (!dev->dma_mask)
return false;
- return addr + size - 1 <= *dev->dma_mask;
+ mask = *dev->dma_mask;
+ if (dev->archdata.parent_dma_mask &&
+ mask > dev->archdata.parent_dma_mask)
+ mask = dev->archdata.parent_dma_mask;
+
+ return addr + size - 1 <= mask;
}
static inline void dma_mark_clean(void *addr, size_t size)
diff --git a/arch/arm64/mm/dma-mapping.c b/arch/arm64/mm/dma-mapping.c
index 58470b1..c9c1892 100644
--- a/arch/arm64/mm/dma-mapping.c
+++ b/arch/arm64/mm/dma-mapping.c
@@ -928,7 +928,18 @@ void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
dev->dma_ops = &swiotlb_dma_ops;
dev->archdata.dma_coherent = coherent;
- __iommu_setup_dma_ops(dev, dma_base, size, iommu);
+
+ if (!iommu) {
+ /*
+ * we don't yet support buses that have a non-zero mapping.
+ * Let's hope we won't need it
+ */
+ WARN_ON(dma_base != 0);
+
+ /* save parent_dma_mask for swiotlb's dma_capable() */
+ dev->archdata.parent_dma_mask = size - 1;
+ } else
+ __iommu_setup_dma_ops(dev, dma_base, size, iommu);
#ifdef CONFIG_XEN
if (xen_initial_domain()) {
--
2.7.4
|
