Thank you for your contribution! Sashiko AI review found 2 potential issue(s) 
to consider:

New issues:
- [High] Newly added mock device is freed on test abort while scheduler 
workqueues are still running.

Pre-existing issues:
- [High] NULL pointer dereference in test suite exit functions if 
drm_mock_sched_new() aborts during initialization.
--

commit 87ec4ea74ea7485f78bc8e603289f73cd4461eb5
Author: Shixiong Ou <[email protected]>

drm/sched: Create a fake device for KUnit tests

This commit creates a dummy device using kunit_device_register() for the
mock scheduler in KUnit tests. This prevents a NULL pointer dereference
in the drm_sched_job trace event which relies on sched->dev.

> diff --git a/drivers/gpu/drm/scheduler/tests/mock_scheduler.c 
> b/drivers/gpu/drm/scheduler/tests/mock_scheduler.c
> index 8e9ae7d980eb2..2dfa3efef2100 100644
> --- a/drivers/gpu/drm/scheduler/tests/mock_scheduler.c
> +++ b/drivers/gpu/drm/scheduler/tests/mock_scheduler.c
[ ... ]
> @@ -297,11 +300,19 @@ struct drm_mock_scheduler *drm_mock_sched_new(struct 
> kunit *test, long timeout)
>               .name           = "drm-mock-scheduler",
>       };
>       struct drm_mock_scheduler *sched;
> +     struct device *dev;
> +     char name[64];
>       int ret;
>  
>       sched = kunit_kzalloc(test, sizeof(*sched), GFP_KERNEL);
>       KUNIT_ASSERT_NOT_NULL(test, sched);
>  
> +     snprintf(name, sizeof(name), "%s-%u", args.name, instance++);
> +     dev = kunit_device_register(test, name);
> +     KUNIT_ASSERT_NOT_ERR_OR_NULL(test, dev);

[Severity: High]
This is a pre-existing issue, but does this assertion expose a crash in
the test suite's exit functions?

If kunit_device_register() fails in drm_mock_sched_new(), the
KUNIT_ASSERT_NOT_ERR_OR_NULL macro aborts the initialization sequence.
Control returns to the KUnit framework without test->priv being initialized.

When the framework subsequently runs the suite's exit function
(drm_sched_basic_exit()), it extracts test->priv and passes it to
drm_mock_sched_fini(), which unconditionally dereferences the NULL pointer.

Could this lead to a kernel panic during test failure rather than cleanly
reporting the error?

> +
> +     args.dev = dev;
> +
>       ret = drm_sched_init(&sched->base, &args);
>       KUNIT_ASSERT_EQ(test, ret, 0);

[Severity: High]
Could tying the mock device to the KUnit test lifecycle cause a
use-after-free if the test aborts?

The device allocated with kunit_device_register() is automatically freed
when the KUnit test thread exits. However, drm_sched_init() spawns
background tasks (scheduler workqueues) that hold references to this
device.

If a subsequent test assertion fails (e.g., KUNIT_ASSERT_FALSE in
drm_sched_test_credits()), the test aborts, skipping the manual
drm_mock_sched_fini() cleanup. KUnit will automatically free the mock
device, but the background tasks remain active and may attempt to access
the freed device memory.

Should the scheduler cleanup be registered as a deferred KUnit action to
ensure the workqueues are stopped before the device is freed?

-- 
Sashiko AI review ยท 
https://sashiko.dev/#/patchset/[email protected]?part=1

Reply via email to