All of lore.kernel.org
 help / color / mirror / Atom feed
* [igt-dev] [PATCH i-g-t v2] tests/kms_atomic_transition: Separate commits for pipes with shared planes
@ 2022-04-13  0:22 Jessica Zhang
  2022-04-13 11:30 ` [igt-dev] ✗ GitLab.Pipeline: warning for " Patchwork
                   ` (6 more replies)
  0 siblings, 7 replies; 10+ messages in thread
From: Jessica Zhang @ 2022-04-13  0:22 UTC (permalink / raw)
  To: igt-dev; +Cc: quic_aravindh, petri.latvala

Currently the helper method set_combinations() will set the properties
for each pipe/plane then push everything as a single commit. This will
cause issues for drivers which do not dedicate planes to pipes and allow
planes to be shared, as it causes the new pipe to overwrite the plane
properties that have been set by the old pipe. In addition, the DRM
framework forbids directly switching the assigned pipe for a plane within
the same commit [1].

To avoid this, we can first disable any pipes that currently are
handling a shared plane before setting the fb to be committed.

Note: This patch only fixes the issue for single display cases. Fixes
for multi-display cases will be addressed in a separate patch

Changes since V1:
- Removed shared plane bitmask
- Instead of immediately doing a NULL commit to disable the current pipe
  when primary plane is shared, disable the previous pipe using the
  primary plane before preparing the framebuffers for commit
- Removed commit for non-NULL framebuffers

[1]
https://gitlab.freedesktop.org/drm/msm/-/blob/msm-next/drivers/gpu/drm/drm_atomic.c#L695

Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
---
 tests/kms_atomic_transition.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/tests/kms_atomic_transition.c b/tests/kms_atomic_transition.c
index 4be24f9e3091..d8462bfcd71a 100644
--- a/tests/kms_atomic_transition.c
+++ b/tests/kms_atomic_transition.c
@@ -716,6 +716,25 @@ static unsigned set_combinations(data_t *data, unsigned mask, struct igt_fb *fb)
 	for (i = 0; i < data->display.n_outputs; i++)
 		igt_output_set_pipe(&data->display.outputs[i], PIPE_NONE);
 
+	for_each_pipe(&data->display, pipe) {
+		igt_plane_t *plane = igt_pipe_get_plane_type(&data->display.pipes[pipe],
+			DRM_PLANE_TYPE_PRIMARY);
+
+		enum pipe old_pipe = plane->ref->pipe->pipe;
+
+		/*
+		 * If a plane is being shared by multiple pipes, we must disable the pipe that
+		 * currently is holding the plane
+		 */
+		if (old_pipe != pipe) {
+			igt_plane_t *old_plane = igt_pipe_get_plane_type(&data->display.pipes[old_pipe],
+				DRM_PLANE_TYPE_PRIMARY);
+
+			igt_plane_set_fb(old_plane, NULL);
+			igt_display_commit2(&data->display, COMMIT_ATOMIC);
+		}
+	}
+
 	for_each_pipe(&data->display, pipe) {
 		igt_plane_t *plane = igt_pipe_get_plane_type(&data->display.pipes[pipe],
 			DRM_PLANE_TYPE_PRIMARY);
-- 
2.31.0

^ permalink raw reply related	[flat|nested] 10+ messages in thread
end of thread, other threads:[~2022-04-20  8:28 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-04-13  0:22 [igt-dev] [PATCH i-g-t v2] tests/kms_atomic_transition: Separate commits for pipes with shared planes Jessica Zhang
2022-04-13 11:30 ` [igt-dev] ✗ GitLab.Pipeline: warning for " Patchwork
2022-04-13 11:54 ` [igt-dev] ✓ Fi.CI.BAT: success " Patchwork
2022-04-13 13:59 ` [igt-dev] ✗ Fi.CI.IGT: failure " Patchwork
2022-04-13 17:39 ` [igt-dev] ✓ Fi.CI.BAT: success for tests/kms_atomic_transition: Separate commits for pipes with shared planes (rev2) Patchwork
2022-04-13 18:48 ` [igt-dev] ✗ Fi.CI.IGT: failure " Patchwork
2022-04-13 20:57   ` Jessica Zhang
2022-04-14  3:38     ` Vudum, Lakshminarayana
2022-04-14  3:23 ` [igt-dev] ✓ Fi.CI.IGT: success " Patchwork
2022-04-20  8:28 ` [igt-dev] [PATCH i-g-t v2] tests/kms_atomic_transition: Separate commits for pipes with shared planes Petri Latvala

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.