The legacy page_flip driver entry point is the only one left which requires drivers to update plane->fb themselves. All the other entry hooks will patch things up for the driver as needed since no one seems to reliable get this right, see e.g. drm_mode_set_config_internal or the plane->fb/old_fb handling in drm_mode_atomic_ioctl.
Therefore unify things, which allows us to ditch a TODO from drm_atomic_helper_page_flip.
This should also help the atomic transition in i915 since we keep a bit of legacy cruft only around because of this special behaviour in ->page_flip.
Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com Signed-off-by: Daniel Vetter daniel.vetter@intel.com --- drivers/gpu/drm/drm_atomic_helper.c | 4 ---- drivers/gpu/drm/drm_crtc.c | 8 +------- 2 files changed, 1 insertion(+), 11 deletions(-)
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 5b59d5ad7d1c..0898afbc9e23 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -1915,10 +1915,6 @@ retry: if (ret != 0) goto fail;
- /* TODO: ->page_flip is the only driver callback where the core - * doesn't update plane->fb. For now patch it up here. */ - plane->fb = plane->state->fb; - /* Driver takes ownership of state on successful async commit. */ return 0; fail: diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index d15a96fdff43..ee0b3bf166b6 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -5345,13 +5345,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, /* Keep the old fb, don't unref it. */ crtc->primary->old_fb = NULL; } else { - /* - * Warn if the driver hasn't properly updated the crtc->fb - * field to reflect that the new framebuffer is now used. - * Failing to do so will screw with the reference counting - * on framebuffers. - */ - WARN_ON(crtc->primary->fb != fb); + crtc->primary->fb = fb; /* Unref only the old framebuffer. */ fb = NULL; }
On Tue, Jul 07, 2015 at 08:43:03AM +0200, Daniel Vetter wrote:
The legacy page_flip driver entry point is the only one left which requires drivers to update plane->fb themselves. All the other entry hooks will patch things up for the driver as needed since no one seems to reliable get this right, see e.g. drm_mode_set_config_internal or the plane->fb/old_fb handling in drm_mode_atomic_ioctl.
Therefore unify things, which allows us to ditch a TODO from drm_atomic_helper_page_flip.
This should also help the atomic transition in i915 since we keep a bit of legacy cruft only around because of this special behaviour in ->page_flip.
Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com Signed-off-by: Daniel Vetter daniel.vetter@intel.com
Applied with Maarten's irc r-b to drm-misc. -Daniel
drivers/gpu/drm/drm_atomic_helper.c | 4 ---- drivers/gpu/drm/drm_crtc.c | 8 +------- 2 files changed, 1 insertion(+), 11 deletions(-)
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 5b59d5ad7d1c..0898afbc9e23 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -1915,10 +1915,6 @@ retry: if (ret != 0) goto fail;
- /* TODO: ->page_flip is the only driver callback where the core
* doesn't update plane->fb. For now patch it up here. */- plane->fb = plane->state->fb;
- /* Driver takes ownership of state on successful async commit. */ return 0;
fail: diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index d15a96fdff43..ee0b3bf166b6 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -5345,13 +5345,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, /* Keep the old fb, don't unref it. */ crtc->primary->old_fb = NULL; } else {
/** Warn if the driver hasn't properly updated the crtc->fb* field to reflect that the new framebuffer is now used.* Failing to do so will screw with the reference counting* on framebuffers.*/WARN_ON(crtc->primary->fb != fb);
crtc->primary->fb = fb;-- 2.1.4
dri-devel@lists.freedesktop.org
-
Daniel Vetter -
Daniel Vetter