* [PATCH v2 1/2] drm/fence: enable drm_atomic_helper_wait_for_fences() to interrupt
@ 2016-08-15 23:05 Gustavo Padovan
2016-08-15 23:05 ` [PATCH v2 2/2] drm/fence: allow fence waiting to be interrupted by userspace Gustavo Padovan
0 siblings, 1 reply; 3+ messages in thread
From: Gustavo Padovan @ 2016-08-15 23:05 UTC (permalink / raw)
To: dri-devel
Cc: linux-kernel, Daniel Stone, Daniel Vetter, Rob Clark,
Greg Hackmann, John Harrison, laurent.pinchart, seanpaul,
marcheu, m.chehab, Sumit Semwal, Maarten Lankhorst,
Gustavo Padovan
From: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Add an extra arg to drm_atomic_helper_wait_for_fences() to inform
if fence_wait() should be called interruptible or uninterruptible.
Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
---
drivers/gpu/drm/drm_atomic_helper.c | 19 ++++++++++++++-----
drivers/gpu/drm/msm/msm_atomic.c | 2 +-
include/drm/drm_atomic_helper.h | 5 +++--
3 files changed, 18 insertions(+), 8 deletions(-)
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
index e1f5de2..bf1bc43 100644
--- a/drivers/gpu/drm/drm_atomic_helper.c
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -1007,18 +1007,22 @@ EXPORT_SYMBOL(drm_atomic_helper_commit_modeset_enables);
* drm_atomic_helper_wait_for_fences - wait for fences stashed in plane state
* @dev: DRM device
* @state: atomic state object with old state structures
+ * @intr: if true, do an interruptible wait
*
* For implicit sync, driver should fish the exclusive fence out from the
* incoming fb's and stash it in the drm_plane_state. This is called after
* drm_atomic_helper_swap_state() so it uses the current plane state (and
* just uses the atomic state to find the changed planes)
+ *
+ * Returns zero if sucess or < 0 if fence_wait() fails.
*/
-void drm_atomic_helper_wait_for_fences(struct drm_device *dev,
- struct drm_atomic_state *state)
+int drm_atomic_helper_wait_for_fences(struct drm_device *dev,
+ struct drm_atomic_state *state,
+ bool intr)
{
struct drm_plane *plane;
struct drm_plane_state *plane_state;
- int i;
+ int i, ret;
for_each_plane_in_state(state, plane, plane_state, i) {
if (!plane->state->fence)
@@ -1026,10 +1030,15 @@ void drm_atomic_helper_wait_for_fences(struct drm_device *dev,
WARN_ON(!plane->state->fb);
- fence_wait(plane->state->fence, false);
+ ret = fence_wait(plane->state->fence, intr);
+ if (ret)
+ return ret;
+
fence_put(plane->state->fence);
plane->state->fence = NULL;
}
+
+ return 0;
}
EXPORT_SYMBOL(drm_atomic_helper_wait_for_fences);
@@ -1176,7 +1185,7 @@ static void commit_tail(struct drm_atomic_state *state)
funcs = dev->mode_config.helper_private;
- drm_atomic_helper_wait_for_fences(dev, state);
+ drm_atomic_helper_wait_for_fences(dev, state, false);
drm_atomic_helper_wait_for_dependencies(state);
diff --git a/drivers/gpu/drm/msm/msm_atomic.c b/drivers/gpu/drm/msm/msm_atomic.c
index 4a8a6f1..9518e43 100644
--- a/drivers/gpu/drm/msm/msm_atomic.c
+++ b/drivers/gpu/drm/msm/msm_atomic.c
@@ -112,7 +112,7 @@ static void complete_commit(struct msm_commit *c, bool async)
struct msm_drm_private *priv = dev->dev_private;
struct msm_kms *kms = priv->kms;
- drm_atomic_helper_wait_for_fences(dev, state);
+ drm_atomic_helper_wait_for_fences(dev, state, false);
kms->funcs->prepare_commit(kms, state);
diff --git a/include/drm/drm_atomic_helper.h b/include/drm/drm_atomic_helper.h
index d86ae5d..a42c34b 100644
--- a/include/drm/drm_atomic_helper.h
+++ b/include/drm/drm_atomic_helper.h
@@ -43,8 +43,9 @@ int drm_atomic_helper_commit(struct drm_device *dev,
struct drm_atomic_state *state,
bool nonblock);
-void drm_atomic_helper_wait_for_fences(struct drm_device *dev,
- struct drm_atomic_state *state);
+int drm_atomic_helper_wait_for_fences(struct drm_device *dev,
+ struct drm_atomic_state *state,
+ bool intr);
bool drm_atomic_helper_framebuffer_changed(struct drm_device *dev,
struct drm_atomic_state *old_state,
struct drm_crtc *crtc);
--
2.5.5
^ permalink raw reply [flat|nested] 3+ messages in thread* [PATCH v2 2/2] drm/fence: allow fence waiting to be interrupted by userspace
2016-08-15 23:05 [PATCH v2 1/2] drm/fence: enable drm_atomic_helper_wait_for_fences() to interrupt Gustavo Padovan
@ 2016-08-15 23:05 ` Gustavo Padovan
2016-08-18 8:28 ` Maarten Lankhorst
0 siblings, 1 reply; 3+ messages in thread
From: Gustavo Padovan @ 2016-08-15 23:05 UTC (permalink / raw)
To: dri-devel
Cc: linux-kernel, Daniel Stone, Daniel Vetter, Rob Clark,
Greg Hackmann, John Harrison, laurent.pinchart, seanpaul,
marcheu, m.chehab, Sumit Semwal, Maarten Lankhorst,
Gustavo Padovan
From: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
If userspace is running an synchronously atomic commit and interrupts the
atomic operation during fence_wait() it will hang until the timer expires,
so here we change the wait to be interruptible so it stop immediately when
userspace wants to quit.
Also adds the necessary error checking for fence_wait().
v2: Comment by Daniel Vetter
- Add error checking for fence_wait()
v3: Rebase on top of new atomic noblocking support
v4: Comment by Maarten Lankhorst
- remove 'swapped' bitfield as it was duplicating information
Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
---
drivers/gpu/drm/drm_atomic_helper.c | 33 ++++++++++++++++++++++++++++-----
1 file changed, 28 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
index bf1bc43..292d15b 100644
--- a/drivers/gpu/drm/drm_atomic_helper.c
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -1022,20 +1022,37 @@ int drm_atomic_helper_wait_for_fences(struct drm_device *dev,
{
struct drm_plane *plane;
struct drm_plane_state *plane_state;
+ struct fence *fence;
int i, ret;
for_each_plane_in_state(state, plane, plane_state, i) {
- if (!plane->state->fence)
+ if (!plane->state->fence && !plane_state->fence)
continue;
- WARN_ON(!plane->state->fb);
+ /*
+ * If the caller asks for an interruptible wait it means
+ * that the state were not swapped yet and the operation
+ * can still be interrupted by userspace, so we need
+ * to look to plane_state instead.
+ */
+ if (intr) {
+ fence = plane_state->fence;
+ WARN_ON(!plane_state->fb);
+ } else {
+ fence = plane->state->fence;
+ WARN_ON(!plane->state->fb);
+ }
- ret = fence_wait(plane->state->fence, intr);
+ ret = fence_wait(fence, intr);
if (ret)
return ret;
- fence_put(plane->state->fence);
- plane->state->fence = NULL;
+ fence_put(fence);
+
+ if (intr)
+ plane_state->fence = NULL;
+ else
+ plane->state->fence = NULL;
}
return 0;
@@ -1244,6 +1261,12 @@ int drm_atomic_helper_commit(struct drm_device *dev,
if (ret)
return ret;
+ if (!nonblock) {
+ ret = drm_atomic_helper_wait_for_fences(dev, state, true);
+ if (ret)
+ return ret;
+ }
+
/*
* This is the point of no return - everything below never fails except
* when the hw goes bonghits. Which means we can commit the new state on
--
2.5.5
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH v2 2/2] drm/fence: allow fence waiting to be interrupted by userspace
2016-08-15 23:05 ` [PATCH v2 2/2] drm/fence: allow fence waiting to be interrupted by userspace Gustavo Padovan
@ 2016-08-18 8:28 ` Maarten Lankhorst
0 siblings, 0 replies; 3+ messages in thread
From: Maarten Lankhorst @ 2016-08-18 8:28 UTC (permalink / raw)
To: Gustavo Padovan, dri-devel
Cc: linux-kernel, Daniel Stone, Daniel Vetter, Rob Clark,
Greg Hackmann, John Harrison, laurent.pinchart, seanpaul,
marcheu, m.chehab, Sumit Semwal, Gustavo Padovan, Daniel Vetter
Op 16-08-16 om 01:05 schreef Gustavo Padovan:
> From: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
>
> If userspace is running an synchronously atomic commit and interrupts the
> atomic operation during fence_wait() it will hang until the timer expires,
> so here we change the wait to be interruptible so it stop immediately when
> userspace wants to quit.
>
> Also adds the necessary error checking for fence_wait().
>
> v2: Comment by Daniel Vetter
> - Add error checking for fence_wait()
>
> v3: Rebase on top of new atomic noblocking support
>
> v4: Comment by Maarten Lankhorst
> - remove 'swapped' bitfield as it was duplicating information
>
> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
I think it would make sense to squash this patch with 1/2.
> ---
> drivers/gpu/drm/drm_atomic_helper.c | 33 ++++++++++++++++++++++++++++-----
> 1 file changed, 28 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
> index bf1bc43..292d15b 100644
> --- a/drivers/gpu/drm/drm_atomic_helper.c
> +++ b/drivers/gpu/drm/drm_atomic_helper.c
> @@ -1022,20 +1022,37 @@ int drm_atomic_helper_wait_for_fences(struct drm_device *dev,
> {
> struct drm_plane *plane;
> struct drm_plane_state *plane_state;
> + struct fence *fence;
> int i, ret;
>
> for_each_plane_in_state(state, plane, plane_state, i) {
> - if (!plane->state->fence)
> + if (!plane->state->fence && !plane_state->fence)
> continue;
Move this check to below, or put if (!intr) plane_state = plane->state on top, and only use plane_state.
Daniel, I keep thinking more and more that a for_each_XXX_(old,new,both)_state would be useful,
could we add those perhaps?
> - WARN_ON(!plane->state->fb);
> + /*
> + * If the caller asks for an interruptible wait it means
> + * that the state were not swapped yet and the operation
> + * can still be interrupted by userspace, so we need
> + * to look to plane_state instead.
> + */
> + if (intr) {
> + fence = plane_state->fence;
> + WARN_ON(!plane_state->fb);
> + } else {
> + fence = plane->state->fence;
> + WARN_ON(!plane->state->fb);
> + }
^ if (!fence) continue;
> - ret = fence_wait(plane->state->fence, intr);
> + ret = fence_wait(fence, intr);
> if (ret)
> return ret;
>
> - fence_put(plane->state->fence);
> - plane->state->fence = NULL;
> + fence_put(fence);
> +
> + if (intr)
> + plane_state->fence = NULL;
> + else
> + plane->state->fence = NULL;
> }
>
> return 0;
> @@ -1244,6 +1261,12 @@ int drm_atomic_helper_commit(struct drm_device *dev,
> if (ret)
> return ret;
>
> + if (!nonblock) {
> + ret = drm_atomic_helper_wait_for_fences(dev, state, true);
> + if (ret)
> + return ret;
> + }
> +
> /*
> * This is the point of no return - everything below never fails except
> * when the hw goes bonghits. Which means we can commit the new state on
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2016-08-18 8:28 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-08-15 23:05 [PATCH v2 1/2] drm/fence: enable drm_atomic_helper_wait_for_fences() to interrupt Gustavo Padovan
2016-08-15 23:05 ` [PATCH v2 2/2] drm/fence: allow fence waiting to be interrupted by userspace Gustavo Padovan
2016-08-18 8:28 ` Maarten Lankhorst
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®