Skip to contentSkip to conversation

drm/test: add a test suite for GEM objects backed by shmem

PATCH Applied Marco Pagani 3 participants 7 messages Nov 30, 2023 – Feb 22, 2024

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH v5 1/1 drm/test: add a test suite for GEM objects backed by shmem

This patch introduces an initial KUnit test suite for GEM objects backed by shmem buffers.

v5:
- using __drm_kunit_helper_alloc_drm_device() to avoid local struct v4:
- Add missing MMU dependency for DRM_GEM_SHMEM_HELPER (kernel test robot) v3:
- Explicitly cast pointers in the helpers
- Removed unused pointer to parent dev in struct fake_dev
- Test entries reordering in Kconfig and Makefile sent as a separate patch v2:
- Improved description of test cases
- Cleaner error handling using KUnit actions
- Alphabetical order in Kconfig and Makefile

---
 drivers/gpu/drm/Kconfig                    |   3 +-
 drivers/gpu/drm/tests/Makefile             |   1 +
 drivers/gpu/drm/tests/drm_gem_shmem_test.c | 383 +++++++++++++++++++++
 3 files changed, 386 insertions(+), 1 deletion(-)
 create mode 100644 drivers/gpu/drm/tests/drm_gem_shmem_test.c
drivers/gpu/drm/Kconfig +2 -1
index 740c1c0bd068..b7abd436455f 100644
@@ -74,12 +74,13 @@ config DRM_KUNIT_TEST_HELPERS
 
 config DRM_KUNIT_TEST
 	tristate "KUnit tests for DRM" if !KUNIT_ALL_TESTS
-	depends on DRM && KUNIT
+	depends on DRM && KUNIT && MMU
 	select DRM_BUDDY
 	select DRM_DISPLAY_DP_HELPER
 	select DRM_DISPLAY_HELPER
 	select DRM_EXEC
 	select DRM_EXPORT_FOR_TESTS if m
+	select DRM_GEM_SHMEM_HELPER
 	select DRM_KMS_HELPER
 	select DRM_KUNIT_TEST_HELPERS
 	select DRM_LIB_RANDOM
drivers/gpu/drm/tests/Makefile +1 -0
index 2645af241ff0..d6183b3d7688 100644
@@ -13,6 +13,7 @@ obj-$(CONFIG_DRM_KUNIT_TEST) += \
 	drm_format_helper_test.o \
 	drm_format_test.o \
 	drm_framebuffer_test.o \
+	drm_gem_shmem_test.o \
 	drm_managed_test.o \
 	drm_mm_test.o \
 	drm_modes_test.o \
drivers/gpu/drm/tests/drm_gem_shmem_test.c new file +383 -0

383 lines not on this page · raw message

base-commit: a13fee31f56449fc600d9e064c7b32302f92dcef

Signature · 1 line
2.43.0
maintainer Applied
#1 lore raw
Quoted 6 lines · Marco Pagani

This patch introduces an initial KUnit test suite for GEM objects backed by shmem buffers.

Quoted 12 lines

v5:
- using __drm_kunit_helper_alloc_drm_device() to avoid local struct v4:
- Add missing MMU dependency for DRM_GEM_SHMEM_HELPER (kernel test robot) v3:
- Explicitly cast pointers in the helpers
- Removed unused pointer to parent dev in struct fake_dev
- Test entries reordering in Kconfig and Makefile sent as a separate patch v2:
- Improved description of test cases
- Cleaner error handling using KUnit actions
- Alphabetical order in Kconfig and Makefile

Applied to drm-misc-next, thanks!
Maxime

#2 lore raw
In reply to Marco Pagani

Hi,

Quoted 5 lines · Marco Pagani

This patch introduces an initial KUnit test suite for GEM objects backed by shmem buffers.

When running this in qemu, I get lots of warnings backtraces in the drm core.

WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:327
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:173
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:385
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:211
WARNING: CPU: 0 PID: 1345 at kernel/dma/mapping.c:194
WARNING: CPU: 0 PID: 1347 at drivers/gpu/drm/drm_gem_shmem_helper.c:429
WARNING: CPU: 0 PID: 1349 at drivers/gpu/drm/drm_gem_shmem_helper.c:445 

It looks like dma_resv_assert_held() asserts each time it is executed. The backtrace in kernel/dma/mapping.c is triggered by

	if (WARN_ON_ONCE(!dev->dma_mask))
		return 0;

in __dma_map_sg_attrs().

Is this a possible problem in the test code, or can it be caused by some limitations or bugs in the qemu emulation ? If so, do you have any thoughts or ideas what those limitations / bugs might be ?

Thanks,
Guenter

#3 lore raw
Quoted 22 lines · Guenter Roeck

Hi,

Quoted 5 lines · Marco Pagani

This patch introduces an initial KUnit test suite for GEM objects backed by shmem buffers.

When running this in qemu, I get lots of warnings backtraces in the drm core.

WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:327
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:173
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:385
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:211
WARNING: CPU: 0 PID: 1345 at kernel/dma/mapping.c:194
WARNING: CPU: 0 PID: 1347 at drivers/gpu/drm/drm_gem_shmem_helper.c:429
WARNING: CPU: 0 PID: 1349 at drivers/gpu/drm/drm_gem_shmem_helper.c:445 

It looks like dma_resv_assert_held() asserts each time it is executed. The backtrace in kernel/dma/mapping.c is triggered by

Quoted 7 lines
	if (WARN_ON_ONCE(!dev->dma_mask))
		return 0;

in __dma_map_sg_attrs().

Is this a possible problem in the test code, or can it be caused by some limitations or bugs in the qemu emulation ? If so, do you have any thoughts or ideas what those limitations / bugs might be ?

Hi Guenter,

Thanks for reporting this issue. As you correctly noted, the warnings appear to be caused by the dma_mask in the mock device being uninitialized. I'll send a patch to fix it soon.

Marco

#4 lore raw

Hi Marco,

Quoted 33 lines · Marco Pagani
Quoted 22 lines · Guenter Roeck

Hi,

Quoted 5 lines · Marco Pagani

This patch introduces an initial KUnit test suite for GEM objects backed by shmem buffers.

When running this in qemu, I get lots of warnings backtraces in the drm core.

WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:327
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:173
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:385
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:211
WARNING: CPU: 0 PID: 1345 at kernel/dma/mapping.c:194
WARNING: CPU: 0 PID: 1347 at drivers/gpu/drm/drm_gem_shmem_helper.c:429
WARNING: CPU: 0 PID: 1349 at drivers/gpu/drm/drm_gem_shmem_helper.c:445

It looks like dma_resv_assert_held() asserts each time it is executed. The backtrace in kernel/dma/mapping.c is triggered by

Quoted 7 lines
	if (WARN_ON_ONCE(!dev->dma_mask))
		return 0;

in __dma_map_sg_attrs().

Is this a possible problem in the test code, or can it be caused by some limitations or bugs in the qemu emulation ? If so, do you have any thoughts or ideas what those limitations / bugs might be ?

Quoted 6 lines

Hi Guenter,

Thanks for reporting this issue. As you correctly noted, the warnings appear to be caused by the dma_mask in the mock device being uninitialized. I'll send a patch to fix it soon.

Thanks a lot for the update.

In this context, the TTM unit tests fail as well in qemu, with worse result:
It seems there is some bad cleanup after a failed test case, causing list corruptions in the drm core and ultimately a crash. I don't know if this is also caused by the missing dma_mask initialization.

Thanks,
Guenter

#5 lore raw
Quoted 43 lines · Guenter Roeck

Hi Marco,

Quoted 33 lines · Marco Pagani
Quoted 26 lines · Guenter Roeck

Hi,

Quoted 5 lines · Marco Pagani

This patch introduces an initial KUnit test suite for GEM objects backed by shmem buffers.

When running this in qemu, I get lots of warnings backtraces in the drm core.

WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:327
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:173
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:385
WARNING: CPU: 0 PID: 1341 at drivers/gpu/drm/drm_gem_shmem_helper.c:211
WARNING: CPU: 0 PID: 1345 at kernel/dma/mapping.c:194
WARNING: CPU: 0 PID: 1347 at drivers/gpu/drm/drm_gem_shmem_helper.c:429
WARNING: CPU: 0 PID: 1349 at drivers/gpu/drm/drm_gem_shmem_helper.c:445

It looks like dma_resv_assert_held() asserts each time it is executed. The backtrace in kernel/dma/mapping.c is triggered by
    if (WARN_ON_ONCE(!dev->dma_mask))
        return 0;
in __dma_map_sg_attrs().

Quoted 3 lines

Is this a possible problem in the test code, or can it be caused by some limitations or bugs in the qemu emulation ? If so, do you have any thoughts or ideas what those limitations / bugs might be ?

Quoted 6 lines

Hi Guenter,

Thanks for reporting this issue. As you correctly noted, the warnings appear to be caused by the dma_mask in the mock device being uninitialized. I'll send a patch to fix it soon.

Quoted 7 lines

Thanks a lot for the update.

In this context, the TTM unit tests fail as well in qemu, with worse result:
It seems there is some bad cleanup after a failed test case, causing list corruptions in the drm core and ultimately a crash. I don't know if this is also caused by the missing dma_mask initialization.

That's interesting. Which --arch argument are you using to run the tests with QEMU?
Thanks,
Marco

#6 lore raw
Quoted 9 lines · Marco Pagani
Quoted 6 lines

In this context, the TTM unit tests fail as well in qemu, with worse result:
It seems there is some bad cleanup after a failed test case, causing list corruptions in the drm core and ultimately a crash. I don't know if this is also caused by the missing dma_mask initialization.

That's interesting. Which --arch argument are you using to run the tests with QEMU?

Example (I am not sure if any of those parameters matters; it is just one of my tests):

qemu-system-x86_64 -kernel arch/x86/boot/bzImage -M q35 -cpu IvyBridge \
	-no-reboot -snapshot -smp 2 \
	-device e1000,netdev=net0 -netdev user,id=net0 -m 512 \
	-drive file=rootfs.ext2,format=raw,if=ide \
	--append "earlycon=uart8250,io,0x3f8,9600n8 root=/dev/sda1 console=ttyS0" \
	-d unimp,guest_errors -nographic -monitor none

This results in:

[ ... ]

[    5.989496]     KTAP version 1
[    5.989639]     # Subtest: ttm_device
[    5.989711]     # module: ttm_device_test
[    5.989760]     1..5
[    6.002044]     ok 1 ttm_device_init_basic
[    6.013557]     ok 2 ttm_device_init_multiple
ILLOPC: ffffffffb8ac9350: 0f 0b
[    6.022481]     ok 3 ttm_device_fini_basic
[    6.026172] ------------[ cut here ]------------
[    6.026315] WARNING: CPU: 1 PID: 1575 at drivers/gpu/drm/ttm/ttm_device.c:206 ttm_device_init+0x170/0x190
...
[    6.135016]         ok 3 Above the allocation limit
[    6.138759] ------------[ cut here ]------------
[    6.138925] WARNING: CPU: 1 PID: 1595 at kernel/dma/mapping.c:503 dma_alloc_attrs+0xf6/0x100
...
[    6.143850]     # ttm_pool_alloc_basic: ASSERTION FAILED at drivers/gpu/drm/ttm/tests/ttm_pool_test.c:162
[    6.143850]     Expected err == 0, but
[    6.143850]         err == -12 (0xfffffffffffffff4)
[    6.148824]         not ok 4 One page, with coherent DMA mappings enabled

From there things go downhill.

[    6.152821] list_add corruption. prev->next should be next (ffffffffbbd53950), but was 0000000000000000. (prev=ffff8af1c38f9e20).

and so on until the emulation crashes.

Guenter