Commit 184d3cf4 authored Jul 03, 2018 by Boris Brezillon

drm/vc4: Use wait_for_flip_done() instead of wait_for_vblanks()



drm_atomic_helper_wait_for_vblanks() assumes the CRTC will continuously
generate VBLANK events and the vblank counter will keep increasing.
While this work for a regular pipeline, it doesn't when you have the
CRTC is feeding the transposer block, because this block works in
oneshot mode, and, by the time we reach
drm_atomic_helper_wait_for_vblanks() the only VBLANK event might have
already been sent and the VBLANK counter will stay unchanged, thus
triggering a timeout.

Luckily, we can replace the drm_atomic_helper_wait_for_vblanks() call
by drm_atomic_helper_wait_for_flip_done() because the only thing we
want to check when calling drm_atomic_helper_wait_for_vblanks() from
vc4_atomic_complete_commit() is that new FBs are in use and the old
ones can be safely released.

Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>
Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20180703075022.15138-5-boris.brezillon@bootlin.com

parent 814bde99

drivers/gpu/drm/vc4/vc4_kms.c

+1 −10

Original line number	Diff line number	Diff line
		@@ -153,18 +153,9 @@ vc4_atomic_complete_commit(struct drm_atomic_state *state)

		drm_atomic_helper_commit_modeset_enables(dev, state);

		/* Make sure that drm_atomic_helper_wait_for_vblanks()
		* actually waits for vblank. If we're doing a full atomic
		* modeset (as opposed to a vc4_update_plane() short circuit),
		* then we need to wait for scanout to be done with our
		* display lists before we free it and potentially reallocate
		* and overwrite the dlist memory with a new modeset.
		*/
		state->legacy_cursor_update = false;

		drm_atomic_helper_commit_hw_done(state);

		drm_atomic_helper_wait_for_vblanks(dev, state);
		drm_atomic_helper_wait_for_flip_done(dev, state);

		drm_atomic_helper_cleanup_planes(dev, state);