mbox series

[RFC,0/5] DRM scheduler kunit tests

Message ID 20250203153007.63400-1-tvrtko.ursulin@igalia.com (mailing list archive)
Headers show
Series DRM scheduler kunit tests | expand

Message

Tvrtko Ursulin Feb. 3, 2025, 3:30 p.m. UTC 
There has repeatedly been quite a bit of apprehension when any change to the DRM
scheduler is proposed, with two main reasons being code base is considered
fragile, not well understood and not very well documented, and secondly the lack
of systematic testing outside the vendor specific tests suites and/or test
farms.

This RFC is an attempt to dislodge this status quo by adding some unit tests
using the kunit framework.

General approach is that there is a mock "hardware" backend which can be
controlled from tests, which in turn allows exercising various scheduler code
paths.

I am sending this early as an RFC so we can get alignment on the general idea
and high level approach. Once past that point we can polish and then hopefully
people can go crazy and start adding more advanced tests, testing corner cases
and what not.

Only some simple basic tests get added in the series and hopefully it is easy to
understand what tests are doing.

My usage of the kunit framework may not be perfect though, since this series is
my first attempt to use it.

An obligatory "screenshot" for reference:

[15:16:33] ============ drm_sched_basic_tests (6 subtests) ============
[15:16:33] [PASSED] drm_sched_basic_submit
[15:16:33] [PASSED] drm_sched_basic_queue
[15:16:34] [PASSED] drm_sched_basic_chain
[15:16:34] [PASSED] drm_sched_basic_entities
[15:16:34] [PASSED] drm_sched_basic_entities_chain
[15:16:34] [PASSED] drm_sched_basic_entity_cleanup
[15:16:34] ============== [PASSED] drm_sched_basic_tests ==============
[15:16:34] ========== drm_sched_basic_tdr_tests (1 subtest) ===========
[15:16:36] [PASSED] drm_sched_basic_tdr
[15:16:36] ============ [PASSED] drm_sched_basic_tdr_tests ============
[15:16:36] ======= drm_sched_basic_priority_tests (2 subtests) ========
[15:16:37] [PASSED] drm_sched_priorities
[15:16:37] [PASSED] drm_sched_change_priority
[15:16:37] ========= [PASSED] drm_sched_basic_priority_tests ==========
[15:16:37] ====== drm_sched_basic_modify_sched_tests (1 subtest) ======
[15:16:39] [PASSED] drm_sched_test_modify_sched
[15:16:39] ======= [PASSED] drm_sched_basic_modify_sched_tests ========
[15:16:39] ============================================================
[15:16:39] Testing complete. Ran 10 tests: passed: 10
[15:16:39] Elapsed time: 13.442s total, 0.001s configuring, 4.117s building, 9.248s running

Cc: Christian König <christian.koenig@amd.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Philipp Stanner <phasta@kernel.org>

Tvrtko Ursulin (5):
  drm: Move some options to separate new Kconfig.debug
  drm/scheduler: Add scheduler unit testing infrastructure and some
    basic tests
  drm/scheduler: Add a simple TDR test
  drm/scheduler: Add basic priority tests
  drm/scheduler: Add a basic test for modifying entities scheduler list

 drivers/gpu/drm/Kconfig                       |  98 +---
 drivers/gpu/drm/Kconfig.debug                 | 104 ++++
 drivers/gpu/drm/scheduler/.kunitconfig        |  12 +
 drivers/gpu/drm/scheduler/Makefile            |   1 +
 drivers/gpu/drm/scheduler/tests/Makefile      |   4 +
 .../gpu/drm/scheduler/tests/drm_mock_entity.c |  29 ++
 .../gpu/drm/scheduler/tests/drm_mock_job.c    |   3 +
 .../drm/scheduler/tests/drm_mock_scheduler.c  | 260 ++++++++++
 .../gpu/drm/scheduler/tests/drm_sched_tests.h | 128 +++++
 .../scheduler/tests/drm_sched_tests_basic.c   | 464 ++++++++++++++++++
 10 files changed, 1010 insertions(+), 93 deletions(-)
 create mode 100644 drivers/gpu/drm/Kconfig.debug
 create mode 100644 drivers/gpu/drm/scheduler/.kunitconfig
 create mode 100644 drivers/gpu/drm/scheduler/tests/Makefile
 create mode 100644 drivers/gpu/drm/scheduler/tests/drm_mock_entity.c
 create mode 100644 drivers/gpu/drm/scheduler/tests/drm_mock_job.c
 create mode 100644 drivers/gpu/drm/scheduler/tests/drm_mock_scheduler.c
 create mode 100644 drivers/gpu/drm/scheduler/tests/drm_sched_tests.h
 create mode 100644 drivers/gpu/drm/scheduler/tests/drm_sched_tests_basic.c

Comments

Philipp Stanner Feb. 6, 2025, 9:12 a.m. UTC | #1 
On Mon, 2025-02-03 at 15:30 +0000, Tvrtko Ursulin wrote:
> There has repeatedly been quite a bit of apprehension when any change
> to the DRM
> scheduler is proposed, with two main reasons being code base is
> considered
> fragile, not well understood and not very well documented, and
> secondly the lack
> of systematic testing outside the vendor specific tests suites and/or
> test
> farms.
> 
> This RFC is an attempt to dislodge this status quo by adding some
> unit tests
> using the kunit framework.
> 
> General approach is that there is a mock "hardware" backend which can
> be
> controlled from tests, which in turn allows exercising various
> scheduler code
> paths.
> 
> I am sending this early as an RFC so we can get alignment on the
> general idea
> and high level approach. Once past that point we can polish and then
> hopefully
> people can go crazy and start adding more advanced tests, testing
> corner cases
> and what not.
> 
> Only some simple basic tests get added in the series and hopefully it
> is easy to
> understand what tests are doing.
> 
> My usage of the kunit framework may not be perfect though, since this
> series is
> my first attempt to use it.
> 
> An obligatory "screenshot" for reference:
> 
> [15:16:33] ============ drm_sched_basic_tests (6 subtests)
> ============
> [15:16:33] [PASSED] drm_sched_basic_submit
> [15:16:33] [PASSED] drm_sched_basic_queue
> [15:16:34] [PASSED] drm_sched_basic_chain
> [15:16:34] [PASSED] drm_sched_basic_entities
> [15:16:34] [PASSED] drm_sched_basic_entities_chain
> [15:16:34] [PASSED] drm_sched_basic_entity_cleanup
> [15:16:34] ============== [PASSED] drm_sched_basic_tests
> ==============
> [15:16:34] ========== drm_sched_basic_tdr_tests (1 subtest)
> ===========
> [15:16:36] [PASSED] drm_sched_basic_tdr
> [15:16:36] ============ [PASSED] drm_sched_basic_tdr_tests
> ============
> [15:16:36] ======= drm_sched_basic_priority_tests (2 subtests)
> ========
> [15:16:37] [PASSED] drm_sched_priorities
> [15:16:37] [PASSED] drm_sched_change_priority
> [15:16:37] ========= [PASSED] drm_sched_basic_priority_tests
> ==========
> [15:16:37] ====== drm_sched_basic_modify_sched_tests (1 subtest)
> ======
> [15:16:39] [PASSED] drm_sched_test_modify_sched
> [15:16:39] ======= [PASSED] drm_sched_basic_modify_sched_tests
> ========
> [15:16:39]
> ============================================================
> [15:16:39] Testing complete. Ran 10 tests: passed: 10
> [15:16:39] Elapsed time: 13.442s total, 0.001s configuring, 4.117s
> building, 9.248s running

Thanks for working on this – as Christian said, this is really needed.
I took a first look and don't see any big obstacles right now. I'll
apply some comments at the places so you have some early feedback
already


Thx,
P.



> 
> Cc: Christian König <christian.koenig@amd.com>
> Cc: Danilo Krummrich <dakr@kernel.org>
> Cc: Matthew Brost <matthew.brost@intel.com>
> Cc: Philipp Stanner <phasta@kernel.org>
> 
> Tvrtko Ursulin (5):
>   drm: Move some options to separate new Kconfig.debug
>   drm/scheduler: Add scheduler unit testing infrastructure and some
>     basic tests
>   drm/scheduler: Add a simple TDR test
>   drm/scheduler: Add basic priority tests
>   drm/scheduler: Add a basic test for modifying entities scheduler
> list
> 
>  drivers/gpu/drm/Kconfig                       |  98 +---
>  drivers/gpu/drm/Kconfig.debug                 | 104 ++++
>  drivers/gpu/drm/scheduler/.kunitconfig        |  12 +
>  drivers/gpu/drm/scheduler/Makefile            |   1 +
>  drivers/gpu/drm/scheduler/tests/Makefile      |   4 +
>  .../gpu/drm/scheduler/tests/drm_mock_entity.c |  29 ++
>  .../gpu/drm/scheduler/tests/drm_mock_job.c    |   3 +
>  .../drm/scheduler/tests/drm_mock_scheduler.c  | 260 ++++++++++
>  .../gpu/drm/scheduler/tests/drm_sched_tests.h | 128 +++++
>  .../scheduler/tests/drm_sched_tests_basic.c   | 464
> ++++++++++++++++++
>  10 files changed, 1010 insertions(+), 93 deletions(-)
>  create mode 100644 drivers/gpu/drm/Kconfig.debug
>  create mode 100644 drivers/gpu/drm/scheduler/.kunitconfig
>  create mode 100644 drivers/gpu/drm/scheduler/tests/Makefile
>  create mode 100644 drivers/gpu/drm/scheduler/tests/drm_mock_entity.c
>  create mode 100644 drivers/gpu/drm/scheduler/tests/drm_mock_job.c
>  create mode 100644
> drivers/gpu/drm/scheduler/tests/drm_mock_scheduler.c
>  create mode 100644 drivers/gpu/drm/scheduler/tests/drm_sched_tests.h
>  create mode 100644
> drivers/gpu/drm/scheduler/tests/drm_sched_tests_basic.c
>