?README-AGL.md: The Automotive Grade Linux Distribution ====================================================== AGL is creating an automotive specific Linux distribution (AGL UCB) that unifies the software that has been written in a number of places already, such as GENIVI and Tizen IVI. The layer 'meta-agl' provides a minimal set of software to boot system of AGL Distribution. 'meta-agl' is the minimal core which is used build AGL profiles on top of it. The reference UI is part of 'meta-agl-demo'. Additional components like the security framework are part of 'meta-agl-extra'. The AGL community appreciates feedback, ideas, suggestion, bugs and documentation just as much as code. Please join the irc conversation at the #automotive channel on irc.freenode.net and our mailing list. For infomation for subscribing to the mailing list [automotive-discussions](http://lists.linuxfoundation.org/mailman/listinfo/automotive-discussions) For information about AGL Distribution, see the [AGL Distribution](https://wiki.automotivelinux.org/agl-distro) For information abount Getting started with AGL [here](https://wiki.automotivelinux.org/start/getting-started) For information about contributing to the AGL Distro [here](https://wiki.automotivelinux.org/agl-distro/contributing) Quick start guide ----------------- (latest version here: https://wiki.automotivelinux.org/agl-distro/source-code) To build an image from 'meta-agl' 1. Prepare repo: > $ mkdir ~/bin > $ export PATH=~/bin:$PATH > $ curl https://storage.googleapis.com/git-repo-downloads/repo > ~/bin/repo > $ chmod a+x ~/bin/repo 2. Get all layers: > $ repo init -u https://gerrit.automotivelinux.org/gerrit/AGL/AGL-repo > $ repo sync 3. Check available targets and extensions, select target machine: > $ source meta-agl/scripts/aglsetup.sh -h > $ source meta-agl/scripts/aglsetup.sh -m porter 4. Build agl-image-ivi > $ bitbake agl-image-ivi To build the agl-demo-platform (as demo'ed @ALS) replace 3. and 4. with: 3. Check available targets and extensions, select target machine: > $ source meta-agl/scripts/aglsetup.sh -h > $ source meta-agl/scripts/aglsetup.sh -m porter agl-demo [agl-appfw-smack] [agl-devel] [agl-netboot] 4. Build agl-demo-platform > $ bitbake agl-demo-platform Note: this uses more layers and not only meta-agl AGL Layers and dependencies: ---------------------------- Check the README.md of the respective layers: - meta-agl (= core, no UI) - meta-agl-demo (= reference UI) - meta-agl-extra (= extra components) In addition, the reference hardware (Renesas R-Car Gen2 / porter) has this dependency: URI: https://gerrit.automotivelinux.org/gerrit/AGL/meta-renesas > branch: agl-1.0-bsp-1.8.0 > tested revision: 82611ccadef36ab0b8a6fd6fb1cf055e115f1ef5 Supported Machines ------------------ Reference hardware: * QEMU (x86-64) - emulated machine: qemux86-64 * Renesas R-Car Gen2 (R-Car M2) - machine: porter Community contributed hardware: * See: https://wiki.automotivelinux.org/agl-distro#supported_hardware Supported Target of bitbake --------------------------- meta-agl: * `agl-image-ivi` The baseline image of AGL Distributions (console only) * `agl-image-minimal` For internal use to develop distribution (experimental) * `agl-image-weston` For internal use to develop distribution (experimental) meta-agl-demo: * `agl-demo-platform` The demo/reference image (with graphical UI) Run this command for a full list of machines, features and targets: > $ source ./meta-agl/scripts/aglsetup.sh -h Supposed Directory Tree of Layers to build ------------------------------------------- > $ repo init -u https://gerrit.automotivelinux.org/gerrit/AGL/AGL-repo > $ repo sync > $ tree -L 1 . |-- meta-agl |-- meta-agl-demo |-- meta-agl-devel |-- meta-agl-extra |-- meta-amb |-- meta-fsl-arm |-- meta-fsl-arm-extra |-- meta-intel |-- meta-intel-iot-security |-- meta-openembedded |-- meta-qcom |-- meta-qt5 |-- meta-raspberrypi |-- meta-renesas |-- meta-rust |-- meta-security-isafw |-- meta-ti `-- poky 18 directories, 0 files Build a QEMU image ------------------ You can build a QEMU image using the following steps: (latest version here: https://wiki.automotivelinux.org/agl-distro/source-code) To build an image from 'meta-agl' 1. Prepare repo: > $ mkdir ~/bin > $ export PATH=~/bin:$PATH > $ curl https://storage.googleapis.com/git-repo-downloads/repo > ~/bin/repo > $ chmod a+x ~/bin/repo 2. Get all layers: > $ repo init -u https://gerrit.automotivelinux.org/gerrit/AGL/AGL-repo > $ repo sync 3. Check available targets and extensions, select target machine: > $ source meta-agl/scripts/aglsetup.sh -h > $ source meta-agl/scripts/aglsetup.sh -m qemux86-64 Alternatively, to enable the reference UI feature ('agl-demo'): > $ source meta-agl/scripts/aglsetup.sh -m qemux86-64 agl-demo Alternatively, to also enable also the security framework ('agl-appfw-smack'): > $ source meta-agl/scripts/aglsetup.sh -m qemux86-64 agl-demo agl agl-appfw-smack Alternatively, to also enable debugging tools ('agl-devel') and the network-boot (nbd) capabilities ('agl-netboot'): > $ source meta-agl/scripts/aglsetup.sh -m qemux86-64 agl-demo agl agl-appfw-smack agl-devel agl-netboot 4. Build agl-image-ivi > $ bitbake agl-image-ivi or alternatively, if you enabled the 'agl-demo' feature: > $ bitbake agl-demo-platform - If you want to run QEMU directly as VM in Virtual Box or your other favorite VM software then add this line to your "conf/local.conf" file. > IMAGE_FSTYPES += "vmdk" 5. Run in the emulator > $ runqemu agl-image-ivi qemux86-64 > or > $ runqemu agl-demo-platform qemux86-64 For a large screen: > $ runqemu agl-image-ivi qemux86-64 bootparams="uvesafb.mode_option=1280x720-32" To extend the amount of memory, add to runqemu: qemuparams="-m 512" > $ runqemu agl-image-ivi qemux86-64 qemuparame="-m 512" bootparams="uvesafb.mode_option=1280x720-32" To separate console from VGA screen (to avoid corrupt screen while booting), add to runqemu: serial > $ runqemu agl-image-ivi qemux86-64 bootparams="uvesafb.mode_option=1280x720-32" serial **Or use the virtual disk in Virtual Box from this location:** > tmp/deploy/images/qemux86-64/agl-demo-platform-qemux86-64.vmdk 6. Some weston samples are available from weston terminal (click top left icon). Check the folder `/opt/AGL/ALS2016`. Build a R-Car M2 (porter) image ------------------------------- ### Software setup NOTE: You will need to download the matching binary driver package from renesas. As of this writing, the version from 20151228. The main URL to retrieve these is: - https://www.renesas.com/en-eu/solutions/automotive/rcar-demoboard.html As of this writing, the necessary libary and driver packages linked to by above site are: - The subpage for the library is: - https://www.renesas.com/en-eu/software/D3017410.html (registration/login required) - The sub-page for the driver is: - https://www.renesas.com/en-eu/media/secret/solutions/automotive/rcar-demoboard/R-Car_Series_Evaluation_Software_Package_of_Linux_Drivers-20151228.zip !!! The files need to be in `${HOME}/Downloads` ( `$XDG_DOWNLOAD_DIR` ) !!! #### Getting Source Code and Build image (latest version here: https://wiki.automotivelinux.org/agl-distro/source-code) To build an image from 'meta-agl' only: 1. Prepare repo: > $ mkdir ~/bin > $ export PATH=~/bin:$PATH > $ curl https://storage.googleapis.com/git-repo-downloads/repo > ~/bin/repo > $ chmod a+x ~/bin/repo 2. Get all layers: > $ repo init -u https://gerrit.automotivelinux.org/gerrit/AGL/AGL-repo > $ repo sync 3. Check available targets and extensions, select target machine: > $ source meta-agl/scripts/aglsetup.sh -h > $ source meta-agl/scripts/aglsetup.sh -m porter Alternatively, to enable the reference UI: > $ source meta-agl/scripts/aglsetup.sh -m porter agl-demo Alternatively, to also enable also the security framework: > $ source meta-agl/scripts/aglsetup.sh -m porter agl-demo agl agl-appfw-smack Alternatively, to also enable debugging tools and the network-boot (nbd) capabilities: > $ source meta-agl/scripts/aglsetup.sh -m porter agl-demo agl agl-appfw-smack agl-devel agl-netboot 4. (optional) edit conf/local.conf )* see below > $ vi conf/local.conf 5. Build agl-image-ivi > $ bitbake agl-image-ivi or alternatively, if you enabled the 'agl-demo' feature: > $ bitbake agl-demo-platform )* : 4 . Optional edits to conf/local.conf * If you want to use multimedia accelerations, uncomment manually 4 `IMAGE_INSTALL_append_porter` in conf/local.conf. > #IMAGE_INSTALL_append_porter = " \ > # gstreamer1.0-plugins-bad-waylandsink \ > # " > #IMAGE_INSTALL_append_porter = " \ > # gstreamer1.0-plugins-base-videorate \ > ... > #" > #IMAGE_INSTALL_append_porter = " \ > # libegl libegl-dev libgbm-dev \ > ... > # " > #IMAGE_INSTALL_append_porter = " \ > # packagegroup-rcar-gen2-multimedia \ > ... > # " * Also it is needed to uncomment this: > #MACHINE_FEATURES_append = " multimedia" This `multimedia` MACHINE_FEATURES enables the meta-renesas's specific multimedia configuration. The version of GStreamer1.0 which AGL distro will use is changed to 1.2.3 (meta-renesas prefers it) over 1.4.1 (meta-agl default) by this switch. * If you want to install various Qt5 examples, add below configuration to your local.conf. > IMAGE_INSTALL_append = " \ > packagegroup-agl-demo-qt-examples \ > " > PACKAGECONFIG_append_pn-qtbase = " examples" > > IMPORTANT NOTE: > To run examples with wayland-egl plugin, > use ``LD_PRELOAD=/usr/lib/libEGL.so ``. > If not, programs should not launch by error, > 'EGL not available'. ### Deployment (SDCARD) #### Instructions on the host 1. Format SD-Card and then, create single EXT4 partition on it. 2. Mount the SD-Card, for example `/media/$SDCARD_LABEL`. 3. Copy AGL root file system onto the SD-Card 1. Go to build directory > $ cd $AGL_TOP/build/tmp/deploy/images/porter 2. Extract the root file system into the SD-Card > $ sudo tar --extract --numeric-owner --preserve-permissions --preserve-order \ > --totals --directory=/media/$SDCARD_LABEL --file=agl-image-ivi-porter.tar.bz2 3. Copy kernel and DTB into the `/boot` of the SD-Card > $ sudo cp uImage uImage-r8a7791-porter.dtb /media/$SDCARD_LABEL/boot 4. After the copy finished, unmount SD-Card and insert it into the SD-Card slot of the porter board. #### Instructions on the target NOTE: There is details about porter board [here](http://elinux.org/R-Car/Boards/Porter). NOTE: To boot weston on porter board, we need keyboard and mouse. (USB2.0 can be use for this) ##### Change U-Boot parameters to boot from SD card 1. Power up the board and, using your preferred terminal emulator, stop the board's autoboot by hitting any key. ``Debug serial settings are 38400 8N1. Any standard terminal emulator program can be used.** 2. Set the follow environment variables and save them > => setenv bootargs_console console=ttySC6,${baudrate} > => setenv bootargs_video vmalloc=384M video=HDMI-A-1:1024x768-32@60 > => setenv bootcmd_sd 'ext4load mmc 0:1 0x40007fc0 boot/uImage;ext4load mmc 0:1 0x40f00000 boot/uImage-r8a7791-porter.dtb' > => setenv bootcmd 'setenv bootargs ${bootargs_console} ${bootargs_video} root=/dev/mmcblk0p1 rw rootfstype=ext3;run bootcmd_sd;bootm 0x40007fc0 - 0x40f00000' > => saveenv ##### Boot from SD card 1. After board reset, U-Boot is started and after a countdown, ... Linux boot message should be displayed. Please wait a moment. 2. Then weston is booted automatically, and weston-terminal appears. 3. Have fun! :) 4. (Optional) This is how to test and play multimedia contents with acceleration. 1. Boot porter without mouse and keyboard, it avoid to boot weston automatically. For now, when running weston, V4L2 deosn't work correctly, so we have to stop weston first (GST plugin `waylandsink` also doesn't work correctly for now). 2. Execute these instructions # Set the mixer > $ export LD_LIBRARY_PATH="/lib:/usr/lib:/usr/local/lib:"^ # Set the mixer > $ amixer set "LINEOUT Mixer DACL" on > $ amixer set "DVC Out" 10 > > $ modprobe -a mmngr mmngrbuf s3ctl uvcs_cmn vspm fdpm > > $ media-ctl -d /dev/media0 -r > $ media-ctl -d /dev/media0 -l '"vsp1.2 rpf.0":1 -> "vsp1.2 uds.0":0 [1]' > $ media-ctl -d /dev/media0 -l '"vsp1.2 uds.0":1 -> "vsp1.2 wpf.0":0 [1]' > $ media-ctl -d /dev/media0 -l '"vsp1.2 wpf.0":1 -> "vsp1.2 lif":0 [1]' > $ media-ctl -d /dev/media0 -V '"vsp1.2 rpf.0":0 [fmt:AYUV32/1920x1080]' > $ media-ctl -d /dev/media0 -V '"vsp1.2 rpf.0":1 [fmt:AYUV32/1920x1080]' > $ media-ctl -d /dev/media0 -V '"vsp1.2 uds.0":0 [fmt:AYUV32/1920x1080]' > $ media-ctl -d /dev/media0 -V '"vsp1.2 uds.0":1 [fmt:AYUV32/640x480]' > $ media-ctl -d /dev/media0 -V '"vsp1.2 wpf.0":0 [fmt:AYUV32/640x480]' > $ media-ctl -d /dev/media0 -V '"vsp1.2 wpf.0":1 [fmt:ARGB32/640x480]' > $ media-ctl -d /dev/media0 -V '"vsp1.2 lif":0 [fmt:ARGB32/640x480]' > > # in case R-Car M2 (HDMI - DU1 - vspd0) > $ modetest -M rcar-du -s 10@8:1280x720@AR24 -d -P '8@19:640x480+100+200@XR24' & After these command, Test pattern will show on display connected to porter's HDMI port. Then, you can play H264(MP4) movie like this: > $ gst-launch-1.0 filesrc location=./sample.mp4 ! qtdemux name=d d. ! \ > queue ! omxh264dec no-copy=true ! v4l2sink device=/dev/video1 \ > io-mode=userptr d. ! queue ! faad ! alsasink device=hw:0,0 ### Deployment (TFTP/NFS) NOTE: These instructions are based on Embedded Linux Wiki, [here](http://www.elinux.org/R-Car/Boards/Yocto#Loading_kernel_via_TFTP_and_rootfs_via_NFS). And a Debian (wheezy, ip: 192.168.30.70) is used as the host for this instructions. #### Instructions on the host 1. Setup a TFTP server 1. Install necessary packages > $ sudo apt-get install tftp tftpd-hpa 2. Go to build directory, and copy kernel and DTB into TFTP server root (default server dir: /srv/tftp) > $cd $AGL_TOP/build/tmp/deploy/images/porter > $ sudo cp uImage uImage-r8a7791-porter.dtb /srv/tftp 3. Verify TFTP server is working > $ ls uImage ls: cannot access uImage: No such file or directory > $ cd /tmp/ > $ tftp 192.168.30.70 > tftp> get uImage > Received 3583604 bytes in 0.2 seconds > tftp> q > $ ls uImage > uImage 2. set NFS server 1. Install necessary packages > $ sudo apt-get install nfs-kernel-server nfs-common 2. Go to build directory, and extract the root file system into a dedicated directory (here we use /nfs/porter) > $ cd $AGL_TOP/build/tmp/deploy/images/porter > $ sudo mkdir -p /nfs/porter > $ sudo tar --extract --numeric-owner --preserve-permissions --preserve-order \ > --totals --directory=/nfs/porter --file=agl-demo-platform-porter.tar.bz2 3. Edit /etc/exports > $ sudo vi /etc/exports Add > > /nfs/porter *(rw,no_subtree_check,sync,no_root_squash,no_all_squash) > Save the file and exit. 4. Restart nfs service > $ sudo service nfs-kernel-server restart 5. Verify NFS server is working > $ sudo mount -t nfs 192.168.30.70:/nfs/porter /tmp/ > $ ls /tmp > bin boot dev etc home lib media mnt proc run sbin sys tmp usr var #### Instructions on the target board NOTE: There is details about porter board [here](http://elinux.org/R-Car/Boards/Porter). NOTE: To boot weston on porter board, we need keyboard and mouse. (USB2.0 can be use for this) ##### Change U-Boot parameters to boot from TFTP/NFS 1. Power up the board and, using your preferred terminal emulator, stop the board's autoboot by hitting any key. > Debug serial settings are 38400 8N1. Any standard terminal emulator program can be used. 2. Set the follow environment variables and save them > => setenv ipaddr > => setenv serverip > => setenv bootargs_console console=ttySC6,${baudrate} > => setenv bootargs_video vmalloc=384M video=HDMI-A-1:1024x768-32@60 > => setenv bootcmd_net 'tftp 0x40007fc0 uImage; tftp 0x40f00000 uImage-r8a7791-porter.dtb' > => setenv bootcmd 'setenv bootargs ${bootargs_console} ${bootargs_video} ip=${ipaddr} root=/dev/nfs nfsroot=${serverip}:/nfs/porter,vers=3;run bootcmd_net;bootm 0x40007fc0 - 0x40f00000' > => saveenv Replace with a proper IP address for the board, like 192.168.30.60. Replace with the IP address of the host, here we use 192.168.30.70. ##### Boot from TFTP/NFS 1. After board reset, U-Boot is started and after a countdown, ... Linux boot message should be displayed. Please wait a moment. 2. Then weston is booted automatically, and weston-terminal appears. 3. Have fun! :)