Async plane update is supposed to work only when updating the FB or FB position of an already enabled plane. That does not apply to requests where the plane was previously disabled or assigned to a different CTRC.
Check old_plane_state->crtc value to make sure async plane update is allowed.
Fixes: fef9df8b5945 ("drm/atomic: initial support for asynchronous plane update") Signed-off-by: Boris Brezillon boris.brezillon@bootlin.com --- Hello,
As discussed on IRC, I'm not sure this "plane should already be enabled and assigned to the same CRTC to allow async updates" limitation applies to all drivers, but it's required for VC4 to work properly.
Just let me know if you think I should move this check to vc4_plane_atomic_async_check().
Thanks,
Boris --- drivers/gpu/drm/drm_atomic_helper.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 866a2cc72ef6..f7ccfebd3ca8 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -1555,7 +1555,8 @@ int drm_atomic_helper_async_check(struct drm_device *dev, if (n_planes != 1) return -EINVAL;
- if (!new_plane_state->crtc) + if (!new_plane_state->crtc || + old_plane_state->crtc != new_plane_state->crtc) return -EINVAL;
funcs = plane->helper_private;
On Mon, Jul 23, 2018 at 03:43:54PM +0200, Boris Brezillon wrote:
Async plane update is supposed to work only when updating the FB or FB position of an already enabled plane. That does not apply to requests where the plane was previously disabled or assigned to a different CTRC.
Check old_plane_state->crtc value to make sure async plane update is allowed.
Fixes: fef9df8b5945 ("drm/atomic: initial support for asynchronous plane update") Signed-off-by: Boris Brezillon boris.brezillon@bootlin.com
Hello,
As discussed on IRC, I'm not sure this "plane should already be enabled and assigned to the same CRTC to allow async updates" limitation applies to all drivers, but it's required for VC4 to work properly.
Just let me know if you think I should move this check to vc4_plane_atomic_async_check().
I don't _think_ there's any hardware that can switch planes without first disabling the plane for a frame and then re-enabling it, so this passes the smell test for me. I'd be very curious to be proven wrong, though :)
Sean
Thanks,
Boris
drivers/gpu/drm/drm_atomic_helper.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 866a2cc72ef6..f7ccfebd3ca8 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -1555,7 +1555,8 @@ int drm_atomic_helper_async_check(struct drm_device *dev, if (n_planes != 1) return -EINVAL;
- if (!new_plane_state->crtc)
if (!new_plane_state->crtc ||
old_plane_state->crtc != new_plane_state->crtc)return -EINVAL;
funcs = plane->helper_private;
-- 2.14.1
dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
Boris Brezillon boris.brezillon@bootlin.com writes:
Async plane update is supposed to work only when updating the FB or FB position of an already enabled plane. That does not apply to requests where the plane was previously disabled or assigned to a different CTRC.
Check old_plane_state->crtc value to make sure async plane update is allowed.
Fixes: fef9df8b5945 ("drm/atomic: initial support for asynchronous plane update") Signed-off-by: Boris Brezillon boris.brezillon@bootlin.com
I don't think any driver I've worked on would want an async flip of a plane between CRTCs.
Reviewed-by: Eric Anholt eric@anholt.net
dri-devel@lists.freedesktop.org
-
Boris Brezillon -
Eric Anholt -
Sean Paul