mirror of
https://github.com/cirosantilli/linux-kernel-module-cheat.git
synced 2026-01-23 02:05:57 +01:00
gem5 graphic mode: checkout just the linux kernel
This commit is contained in:
Ciro Santilli
132
README.adoc
132
README.adoc
@@ -451,14 +451,17 @@ so maybe it is not possible?
|
||||
|
||||
==== Graphic mode gem5
|
||||
|
||||
https://stackoverflow.com/questions/50364863/how-to-get-graphical-gui-output-and-user-touch-keyboard-mouse-input-in-a-ful/50364864#50364864
|
||||
TODO could not get it working on `x86_64`, only ARM.
|
||||
|
||||
Overview: https://stackoverflow.com/questions/50364863/how-to-get-graphical-gui-output-and-user-touch-keyboard-mouse-input-in-a-ful/50364864#50364864
|
||||
|
||||
More concretely:
|
||||
|
||||
....
|
||||
git checkout gem5-arm
|
||||
./build -aa -lg -K linux/arch/arm/configs/gem5_defconfig
|
||||
./run -aa -gu
|
||||
git -C linux checkout gem5/v4.15
|
||||
./build -gl -aa -K linux/arch/arm/configs/gem5_defconfig -L gem5-v4.15
|
||||
git -C linux checkout -
|
||||
./run -aa -g -L gem5-v4.15
|
||||
....
|
||||
|
||||
and then on another shell:
|
||||
@@ -467,47 +470,101 @@ and then on another shell:
|
||||
vinagre localhost:5900
|
||||
....
|
||||
|
||||
The port `5900` is incremented by one if you already have something running on that port: `gem5` stdout tells us the right port on stdout as:
|
||||
The <<config_logo>> penguin only appears after several seconds, together with kernel messages of type:
|
||||
|
||||
....
|
||||
[ 0.152755] [drm] found ARM HDLCD version r0p0
|
||||
[ 0.152790] hdlcd 2b000000.hdlcd: bound virt-encoder (ops 0x80935f94)
|
||||
[ 0.152795] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
|
||||
[ 0.152799] [drm] No driver support for vblank timestamp query.
|
||||
[ 0.215179] Console: switching to colour frame buffer device 240x67
|
||||
[ 0.230389] hdlcd 2b000000.hdlcd: fb0: frame buffer device
|
||||
[ 0.230509] [drm] Initialized hdlcd 1.0.0 20151021 for 2b000000.hdlcd on minor 0
|
||||
....
|
||||
|
||||
The port `5900` is incremented by one if you already have something running on that port, `gem5` stdout tells us the right port on stdout as:
|
||||
|
||||
....
|
||||
system.vncserver: Listening for connections on port 5900
|
||||
....
|
||||
|
||||
aarch64:
|
||||
and when we connect it shows a message:
|
||||
|
||||
....
|
||||
./build -aA -lg -K linux/arch/arm64/configs/gem5_defconfig
|
||||
./run -aA -gu
|
||||
info: VNC client attached
|
||||
....
|
||||
|
||||
gem5 implements the HDLCD ARM hardware, the required kernel configs are:
|
||||
For `aarch64` we also need `-c kernel_config_fragment/display`:
|
||||
|
||||
....
|
||||
CONFIG_DRM=y
|
||||
CONFIG_DRM_HDLCD=y
|
||||
CONFIG_DRM_VIRT_ENCODER=y
|
||||
git -C linux checkout gem5/v4.15
|
||||
./build -gl -aA \
|
||||
-c kernel_config_fragment/display \
|
||||
-K linux/arch/arm64/configs/gem5_defconfig \
|
||||
-L gem5-v4.15 \
|
||||
;
|
||||
git -C linux checkout -
|
||||
./run -aA -gu -L gem5-v4.15
|
||||
....
|
||||
|
||||
We need the extra config fragment because the gem5 aarch64 defconfig does not enable HDLCD like the 32 bit one for some reason.
|
||||
This is because the gem5 `aarch64` defconfig does not enable HDLCD like the 32 bit one `arm` one for some reason.
|
||||
|
||||
TODO: without the `-K`:
|
||||
We cannot use mainline Linux because the <<gem5-arm-Linux-kernel-patches>> are required at least to provide the `CONFIG_DRM_VIRT_ENCODER` option.
|
||||
|
||||
gem5 emulates the link:http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0541c/CHDBAIDI.html[HDLCD] ARM Holdings hardware for `arm` and `aarch64`.
|
||||
|
||||
The kernel uses HDLCD to implement the <<drm>> interface, the required kernel config options are present at: link:kernel_config_fragment/display[].
|
||||
|
||||
TODO <<kmscube>> failed on `aarch64` with:
|
||||
|
||||
....
|
||||
git checkout gem5-arm
|
||||
./build -aa -lg -c kernel_config_fragment/gem5-aarch64-hdlcd
|
||||
./run -aa -gu
|
||||
kmscube[706]: unhandled level 2 translation fault (11) at 0x00000000, esr 0x92000006, in libgbm.so.1.0.0[7fbf6a6000+e000]
|
||||
....
|
||||
|
||||
it does not work with a failing dmesg:
|
||||
TODO: minimize out the `-K`. If we just remove it on `arm`: it does not work with a failing dmesg:
|
||||
|
||||
....
|
||||
[ 0.121078] [drm] found ARM HDLCD version r0p0
|
||||
[ 0.121117] hdlcd 2b000000.hdlcd: bound virt-encoder (ops drm_vencoder_ops)
|
||||
[ 0.121122] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
|
||||
[ 0.121126] [drm] No driver support for vblank timestamp query.
|
||||
[ 0.121156] hdlcd 2b000000.hdlcd: Failed to set initial hw configuration.
|
||||
[ 0.121190] hdlcd 2b000000.hdlcd: master bind failed: -12
|
||||
[ 0.121197] hdlcd: probe of 2b000000.hdlcd failed with error -12
|
||||
[ 0.066208] [drm] found ARM HDLCD version r0p0
|
||||
[ 0.066241] hdlcd 2b000000.hdlcd: bound virt-encoder (ops drm_vencoder_ops)
|
||||
[ 0.066247] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
|
||||
[ 0.066252] [drm] No driver support for vblank timestamp query.
|
||||
[ 0.066276] hdlcd 2b000000.hdlcd: Cannot do DMA to address 0x0000000000000000
|
||||
[ 0.066281] swiotlb: coherent allocation failed for device 2b000000.hdlcd size=8294400
|
||||
[ 0.066288] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.15.0 #1
|
||||
[ 0.066293] Hardware name: V2P-AARCH64 (DT)
|
||||
[ 0.066296] Call trace:
|
||||
[ 0.066301] dump_backtrace+0x0/0x1b0
|
||||
[ 0.066306] show_stack+0x24/0x30
|
||||
[ 0.066311] dump_stack+0xb8/0xf0
|
||||
[ 0.066316] swiotlb_alloc_coherent+0x17c/0x190
|
||||
[ 0.066321] __dma_alloc+0x68/0x160
|
||||
[ 0.066325] drm_gem_cma_create+0x98/0x120
|
||||
[ 0.066330] drm_fbdev_cma_create+0x74/0x2e0
|
||||
[ 0.066335] __drm_fb_helper_initial_config_and_unlock+0x1d8/0x3a0
|
||||
[ 0.066341] drm_fb_helper_initial_config+0x4c/0x58
|
||||
[ 0.066347] drm_fbdev_cma_init_with_funcs+0x98/0x148
|
||||
[ 0.066352] drm_fbdev_cma_init+0x40/0x50
|
||||
[ 0.066357] hdlcd_drm_bind+0x220/0x428
|
||||
[ 0.066362] try_to_bring_up_master+0x21c/0x2b8
|
||||
[ 0.066367] component_master_add_with_match+0xa8/0xf0
|
||||
[ 0.066372] hdlcd_probe+0x60/0x78
|
||||
[ 0.066377] platform_drv_probe+0x60/0xc8
|
||||
[ 0.066382] driver_probe_device+0x30c/0x478
|
||||
[ 0.066388] __driver_attach+0x10c/0x128
|
||||
[ 0.066393] bus_for_each_dev+0x70/0xb0
|
||||
[ 0.066398] driver_attach+0x30/0x40
|
||||
[ 0.066402] bus_add_driver+0x1d0/0x298
|
||||
[ 0.066408] driver_register+0x68/0x100
|
||||
[ 0.066413] __platform_driver_register+0x54/0x60
|
||||
[ 0.066418] hdlcd_platform_driver_init+0x20/0x28
|
||||
[ 0.066424] do_one_initcall+0x44/0x130
|
||||
[ 0.066428] kernel_init_freeable+0x13c/0x1d8
|
||||
[ 0.066433] kernel_init+0x18/0x108
|
||||
[ 0.066438] ret_from_fork+0x10/0x1c
|
||||
[ 0.066444] hdlcd 2b000000.hdlcd: Failed to set initial hw configuration.
|
||||
[ 0.066470] hdlcd 2b000000.hdlcd: master bind failed: -12
|
||||
[ 0.066477] hdlcd: probe of 2b000000.hdlcd failed with error -12
|
||||
[
|
||||
....
|
||||
|
||||
So what other options are missing from `gem5_defconfig`? It would be cool to minimize it out to better understand the options.
|
||||
@@ -5169,11 +5226,24 @@ gem5 is a system simulator, much <<gem5-vs-qemu,like QEMU>>: http://gem5.org/
|
||||
For the most part, just add the `-g` option to all commands and everything should magically work:
|
||||
|
||||
....
|
||||
git -C linux checkout v4.16
|
||||
./configure -g && ./build -a aarch64 -g && ./run -a aarch64 -g
|
||||
./configure -g && ./build -a arm -g && ./run -a arm -g
|
||||
....
|
||||
|
||||
On another shell:
|
||||
TODO `aarch64` boot is failing on kernel v4.17 with:
|
||||
|
||||
....
|
||||
panic: Tried to write Gic cpu at offset 0xd0
|
||||
....
|
||||
|
||||
Work around it for now by using v4.16:
|
||||
|
||||
....
|
||||
git -C linux checkout v4.16
|
||||
./configure -g && ./build -a arm -g -L v4.16 && ./run -a aarch64 -g -L v4.16
|
||||
git -C linux checkout -
|
||||
....
|
||||
|
||||
To get a terminal, open a new shell and run:
|
||||
|
||||
....
|
||||
./gem5-shell
|
||||
@@ -5187,12 +5257,6 @@ Tested architectures:
|
||||
|
||||
Like QEMU, gem5 also has a syscall emulation mode (SE), but in this tutorial we focus on the full system emulation mode (FS). For a comparison see: https://stackoverflow.com/questions/48986597/when-should-you-use-full-system-fs-vs-syscall-emulation-se-with-userland-program
|
||||
|
||||
TODO `aarch64` boot failed on v4.17 with:
|
||||
|
||||
....
|
||||
panic: Tried to write Gic cpu at offset 0xd0
|
||||
....
|
||||
|
||||
=== gem5 vs QEMU
|
||||
|
||||
* advantages of gem5:
|
||||
|
||||
2
build
2
build
@@ -155,7 +155,7 @@ BR2_ROOTFS_POST_SCRIPT_ARGS=\"${post_script_args}\"
|
||||
fi
|
||||
default_config_fragments=
|
||||
else
|
||||
default_config_fragments="${kernel_config_fragment_dir}/min ${kernel_config_fragment_dir}/default"
|
||||
default_config_fragments="${kernel_config_fragment_dir}/min ${kernel_config_fragment_dir}/default ${kernel_config_fragment_dir}/display"
|
||||
fi
|
||||
printf "BR2_LINUX_KERNEL_CONFIG_FRAGMENT_FILES=\"${default_config_fragments} ${kernel_config_fragments} ${kernel_config_fragment_cli_file}\"\n" >> "$config_file"
|
||||
if "${linux_reconfigure}"; then
|
||||
|
||||
@@ -82,13 +82,6 @@ CONFIG_FRAMEBUFFER_CONSOLE_ROTATION=y
|
||||
CONFIG_FB=y
|
||||
CONFIG_LOGO=y
|
||||
|
||||
## Display
|
||||
|
||||
CONFIG_DRM=y
|
||||
CONFIG_DRM_HDLCD=y
|
||||
CONFIG_DRM_VIRT_ENCODER=y
|
||||
CONFIG_DRM_VIRTIO_GPU=y
|
||||
|
||||
## Networking
|
||||
|
||||
# Will everything blow up?
|
||||
|
||||
9
kernel_config_fragment/display
Normal file
9
kernel_config_fragment/display
Normal file
@@ -0,0 +1,9 @@
|
||||
CONFIG_DRM=y
|
||||
|
||||
## QEMU
|
||||
CONFIG_DRM_VIRTIO_GPU=y
|
||||
|
||||
## gem5
|
||||
CONFIG_DRM_HDLCD=y
|
||||
# Only present in the gem5 fork.
|
||||
CONFIG_DRM_VIRT_ENCODER=y
|
||||
Reference in New Issue
Block a user