[RFC v6 0/3] Introduce KUnit tests for TTM subsystem

Karolina Stolarek karolina.stolarek at intel.com
Thu Aug 10 09:52:47 UTC 2023


On 10.08.2023 09:19, Christian König wrote:
> Am 08.08.23 um 11:51 schrieb Karolina Stolarek:
>> This series introduces KUnit[1] tests for TTM (Translation Table Manager)
>> subsystem, a memory manager used by graphics drivers to create and manage
>> memory buffers across different memory domains, such as system memory
>> or VRAM.
>>
>> Unit tests implemented here cover two data structures:
>>    - ttm_device -- referred as a buffer object device, which stores
>>      resource managers and page pools
>>    - ttm_pool -- a struct of pools (ttm_pool_type) of different page
>>      orders and caching attributes, with pages that can be reused on
>>      the next buffer allocation
>>
>> Use kunit_tool script to manually run the tests:
>>
>> $ ./tools/testing/kunit/kunit.py run 
>> --kunitconfig=drivers/gpu/drm/ttm/tests
>>
>> To build a kernel with TTM KUnit tests, first enable CONFIG_KUNIT, and 
>> then
>> CONFIG_DRM_TTM_KUNIT_TEST.
>>
>> As for now, tests are architecture-agnostic (i.e. KUnit runner uses UML
>> kernel), which means that we have limited coverage in some places. For
>> example, we can't fully test the initialization of global page pools,
>> such as global_write_combined. It is to be decided if we want to stick
>> to UML or use CONFIG_X86 (at least to some extent).
>>
>> These patches are just a beginning of the work to improve the test
>> coverage of TTM. Feel free to suggest changes, test cases or priorities.
> 
> Couldn't find time for further testing, so I just reviewed and pushed 
> this to drm-misc-next.
> 
> Thanks,
> Christian.

Thanks a lot for taking care of this, fingers crossed it's all working.

I'll rebase the upcoming tests on the top of drm-misc-next then.

All the best,
Karolina

> 
>>
>> Many thanks,
>> Karolina
>>
>> v6:
>>    - Rebase the series on the top of drm-misc-next (Christian)
>>    - Remove drm_dev_put() call from ttm_test_devices_put, the drm 
>> device is
>>      already freed in drm_kunit_helper_free_device()
>>    - Remove an unnecessary priv assignment in ttm_test_devices_all()
>>    - Delete ttm_bo_put() from ttm_pool_test_fini() (as for now, we 
>> don't count
>>      krefs for dummy BOs)
>>
>> v5:
>>    - Drop unnecessary brackets in 2/3
>>    - Rebase KConfig file on the top of drm-tip
>>
>> v4:
>>    - Test helpers have been changed to make the creation of init/fini
>>      functions for each test suite easier:
>>      + Decouple device creation from test initialization by adding
>>        helpers that initialize ttm_test_devices, a struct which stores
>>        DRM/TTM devices, and can be used in test-specific init/finis
>>        (see ttm_pool_tests.c for an example)
>>      + Introduce generic init/fini functions for tests that only need
>>        devices
>>      + Add ttm_device field to ttm_test_devices (previously
>>        ttm_test_devices_priv)
>>    - Make TTM buffer object outlive its TT (Christian)
>>    - Add a dedicated struct for ttm_pool_test (struct ttm_pool_test_priv)
>>    - Rename functions and structs:
>>      + struct ttm_test_devices_priv   --> struct ttm_test_devices
>>      + ttm_kunit_helper_init_device() --> ttm_device_kunit_init()
>>      + ttm_kunit_helper_ttm_bo_init() --> ttm_bo_kunit_init()
>>    - Split ttm_kunit_helper_init() into full config (with ttm_device
>>      init) and basic (init only with device/drm_device) initialization
>>      functions
>>
>> v3:
>>    - Rename ttm_kunit_helper_alloc_device() to 
>> ttm_kunit_helper_init_device()
>>      (Christian)
>>    - Don't leak a full-blown drm_gem_object in 
>> ttm_kunit_helper_ttm_bo_init().
>>      (Christian). Create a small mock object just to get 
>> ttm_tt_init_fields()
>>      to init the right number of pages
>>    - As a follow up to the change above, delete 
>> ttm_kunit_helper_ttm_bo_fini()
>>      and just use ttm_bo_put()
>>
>> v2:
>>    - Add missing symbol exports in ttm_kunit_helpers.c
>>    - Update helpers include to fix compilation issues (didn't catch it as
>>      KUnit tests weren't enabled in the kernel I tested, an oversight
>>      on my part)
>>    - Add checks for ttm_pool fields in ttm_pool_alloc_basic(), 
>> including the
>>      one for NUMA node id
>>    - Rebase the changes on the top of drm-tip
>>
>> --------------------------------
>> [1] - https://www.kernel.org/doc/html/latest/dev-tools/kunit/index.html
>>
>> Karolina Stolarek (3):
>>    drm/ttm: Introduce KUnit test
>>    drm/ttm/tests: Add tests for ttm_device
>>    drm/ttm/tests: Add tests for ttm_pool
>>
>>   drivers/gpu/drm/Kconfig                       |  15 +
>>   drivers/gpu/drm/ttm/Makefile                  |   1 +
>>   drivers/gpu/drm/ttm/tests/.kunitconfig        |   4 +
>>   drivers/gpu/drm/ttm/tests/Makefile            |   6 +
>>   drivers/gpu/drm/ttm/tests/ttm_device_test.c   | 212 +++++++++
>>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c | 113 +++++
>>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h |  41 ++
>>   drivers/gpu/drm/ttm/tests/ttm_pool_test.c     | 437 ++++++++++++++++++
>>   8 files changed, 829 insertions(+)
>>   create mode 100644 drivers/gpu/drm/ttm/tests/.kunitconfig
>>   create mode 100644 drivers/gpu/drm/ttm/tests/Makefile
>>   create mode 100644 drivers/gpu/drm/ttm/tests/ttm_device_test.c
>>   create mode 100644 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>>   create mode 100644 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>>   create mode 100644 drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>>
> 


More information about the dri-devel mailing list