Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider: - [Medium] Comparing plane `src` coordinates before `atomic_check` clips them causes unintended full frame updates for partially off-screen planes. - [Low] Removal of unit tests for `src` coordinate changes without replacing them causes a loss of critical test coverage. --
commit 70283d9d64d3987145c43d7d6e7df2c84b0de6ae Author: Thomas Zimmermann <[email protected]> drm/damage-helper: Test src coord in drm_atomic_helper_check_plane_damage() This commit evaluates plane source coordinate changes during the atomic check phase instead of the atomic commit phase. It sets the ignore_damage_clips flag in the plane state if coordinates change and removes related unit tests. > diff --git a/drivers/gpu/drm/drm_atomic_helper.c > b/drivers/gpu/drm/drm_atomic_helper.c > --- a/drivers/gpu/drm/drm_atomic_helper.c > +++ b/drivers/gpu/drm/drm_atomic_helper.c > @@ -1034,7 +1034,7 @@ drm_atomic_helper_check_planes(struct drm_device *dev, > > drm_atomic_helper_plane_changed(state, old_plane_state, > new_plane_state, plane); > > - drm_atomic_helper_check_plane_damage(state, new_plane_state); > + drm_atomic_helper_check_plane_damage(state, old_plane_state, > new_plane_state); [Severity: Medium] Is there a risk of unintended full frame updates for partially off-screen planes here? When userspace requests an update, the DRM core provides unclipped coordinates into new_plane_state->src. The old_plane_state->src retains the coordinates from the previous commit, which were already clipped during that commit's atomic check phase via drm_atomic_helper_check_plane_state(). Because drm_atomic_helper_check_plane_damage() now performs the drm_rect_equals check before atomic_check scales and clips the new coordinates, does it erroneously detect a change in src and permanently latch ignore_damage_clips = true? This could cause any plane positioned partially off-screen to continuously ignore damage clips and perform a full update on every single frame. Note: This regression is fixed later in the patch series by commit "drm/atomic_helper: Do not evaluate plane damage before atomic_check". > > if (!funcs || !funcs->atomic_check) > continue; [ ... ] > diff --git a/drivers/gpu/drm/tests/drm_damage_helper_test.c > b/drivers/gpu/drm/tests/drm_damage_helper_test.c > --- a/drivers/gpu/drm/tests/drm_damage_helper_test.c > +++ b/drivers/gpu/drm/tests/drm_damage_helper_test.c [ ... ] > -static void drm_test_damage_iter_no_damage_src_moved(struct kunit *test) > -{ > - struct drm_damage_mock *mock = test->priv; > - struct drm_atomic_helper_damage_iter iter; > - struct drm_rect clip; > - u32 num_hits = 0; > - > - /* Plane src moved since old plane state. */ > - set_plane_src(&mock->old_state, 0, 0, 1024 << 16, 768 << 16); > - set_plane_src(&mock->state, 10 << 16, 10 << 16, > - (10 + 1024) << 16, (10 + 768) << 16); > - drm_atomic_helper_damage_iter_init(&iter, &mock->old_state, > &mock->state); > - drm_atomic_for_each_plane_damage(&iter, &clip) > - num_hits++; > - > - KUNIT_EXPECT_EQ_MSG(test, num_hits, 1, "Should return plane src as > damage."); > - check_damage_clip(test, &clip, 10, 10, 1034, 778); > -} [Severity: Low] Does the removal of these unit tests cause a loss of valuable test coverage? The commit message mentions there is a separate unit test for when ignore_damage_clips is true, but without these removed tests, there are no longer any tests ensuring that moving the src coordinates actually correctly sets ignore_damage_clips = true via drm_atomic_helper_check_plane_damage(). Instead of completely deleting these tests, could they have been adapted to call drm_atomic_helper_check_plane_damage() prior to initializing the iterator, preserving the validation of the source coordinate tracking logic? -- Sashiko AI review ยท https://sashiko.dev/#/patchset/[email protected]?part=4
