* [PATCH] drm: i915: Wait for fences on new fb, not old
@ 2016-10-21 14:44 Daniel Stone
2016-10-21 16:21 ` ✗ Fi.CI.BAT: failure for " Patchwork
2016-10-24 7:03 ` [Intel-gfx] [PATCH] " Daniel Vetter
0 siblings, 2 replies; 6+ messages in thread
From: Daniel Stone @ 2016-10-21 14:44 UTC (permalink / raw)
To: intel-gfx; +Cc: ville.syrjala, stable, Daniel Vetter, Maarten Lankhorst
The previous code would wait for fences on the framebuffer from the old
plane state to complete, rather than the new, so you would see tearing
everywhere. Fix this to wait on the new state before we make it active.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Fixes: 94f050246b42 ("drm/i915: nonblocking commit")
Cc: stable@vger.kernel.org
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
---
drivers/gpu/drm/i915/intel_display.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index e5ad101..1bdef9e 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -14292,7 +14292,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state)
for_each_plane_in_state(state, plane, plane_state, i) {
struct intel_plane_state *intel_plane_state =
- to_intel_plane_state(plane_state);
+ to_intel_plane_state(plane->state);
if (!intel_plane_state->wait_req)
continue;
--
2.9.3
^ permalink raw reply related [flat|nested] 6+ messages in thread
* ✗ Fi.CI.BAT: failure for drm: i915: Wait for fences on new fb, not old
2016-10-21 14:44 [PATCH] drm: i915: Wait for fences on new fb, not old Daniel Stone
@ 2016-10-21 16:21 ` Patchwork
2016-10-21 18:49 ` Daniel Stone
2016-10-24 7:03 ` [Intel-gfx] [PATCH] " Daniel Vetter
1 sibling, 1 reply; 6+ messages in thread
From: Patchwork @ 2016-10-21 16:21 UTC (permalink / raw)
To: Daniel Stone; +Cc: intel-gfx
== Series Details ==
Series: drm: i915: Wait for fences on new fb, not old
URL : https://patchwork.freedesktop.org/series/14168/
State : failure
== Summary ==
Series 14168v1 drm: i915: Wait for fences on new fb, not old
https://patchwork.freedesktop.org/api/1.0/series/14168/revisions/1/mbox/
Test drv_module_reload_basic:
dmesg-warn -> PASS (fi-skl-6770hq)
Test gem_ringfill:
Subgroup basic-default-hang:
pass -> TIMEOUT (fi-hsw-4770r)
Subgroup basic-default-interruptible:
pass -> INCOMPLETE (fi-hsw-4770r)
Test gem_storedw_loop:
Subgroup basic-blt:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-bsd:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-bsd1:
skip -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-bsd2:
skip -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-default:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-render:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-vebox:
pass -> INCOMPLETE (fi-hsw-4770r)
Test gem_sync:
Subgroup basic-all:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-each:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-many-each:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-store-all:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-store-each:
pass -> INCOMPLETE (fi-hsw-4770r)
Test gem_tiled_blits:
Subgroup basic:
pass -> INCOMPLETE (fi-hsw-4770r)
Test gem_tiled_fence_blits:
Subgroup basic:
pass -> INCOMPLETE (fi-hsw-4770r)
Test gem_tiled_pread_basic:
pass -> INCOMPLETE (fi-hsw-4770r)
Test gem_workarounds:
Subgroup basic-read:
pass -> INCOMPLETE (fi-hsw-4770r)
Test gvt_basic:
Subgroup invalid-placeholder-test:
skip -> INCOMPLETE (fi-hsw-4770r)
Test kms_addfb_basic:
Subgroup addfb25-bad-modifier:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup addfb25-framebuffer-vs-set-tiling:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup addfb25-modifier-no-flag:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup addfb25-x-tiled:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup addfb25-x-tiled-mismatch:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup addfb25-y-tiled:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup addfb25-y-tiled-small:
skip -> INCOMPLETE (fi-hsw-4770r)
Subgroup addfb25-yf-tiled:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bad-pitch-0:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bad-pitch-1024:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bad-pitch-128:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bad-pitch-256:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bad-pitch-32:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bad-pitch-63:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bad-pitch-65536:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bad-pitch-999:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-x-tiled:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup basic-y-tiled:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bo-too-small:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup bo-too-small-due-to-tiling:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup clobberred-modifier:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup framebuffer-vs-set-tiling:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup invalid-get-prop:
pass -> INCOMPLETE (fi-hsw-4770r)
Subgroup invalid-get-prop-any:
WARNING: Long output truncated
Results at /archive/results/CI_IGT_test/Patchwork_2790/
45ce7992e2d3211009c035db41745e1f12bbf6e8 drm-intel-nightly: 2016y-10m-21d-14h-52m-49s UTC integration manifest
a8cbad1 drm: i915: Wait for fences on new fb, not old
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: ✗ Fi.CI.BAT: failure for drm: i915: Wait for fences on new fb, not old
2016-10-21 16:21 ` ✗ Fi.CI.BAT: failure for " Patchwork
@ 2016-10-21 18:49 ` Daniel Stone
2016-10-21 19:36 ` Saarinen, Jani
2016-10-21 19:48 ` Chris Wilson
0 siblings, 2 replies; 6+ messages in thread
From: Daniel Stone @ 2016-10-21 18:49 UTC (permalink / raw)
To: intel-gfx; +Cc: Daniel Stone
[-- Attachment #1.1: Type: text/plain, Size: 651 bytes --]
Hi,
On Friday, 21 October 2016, Patchwork <patchwork@emeril.freedesktop.org>
wrote:
>
> Test gem_ringfill:
> Subgroup basic-default-hang:
> pass -> TIMEOUT (fi-hsw-4770r)
> Subgroup basic-default-interruptible:
> pass -> INCOMPLETE (fi-hsw-4770r)
> Test gem_storedw_loop:
> Subgroup basic-blt:
> pass -> INCOMPLETE (fi-hsw-4770r)
> Subgroup basic-bsd:
> pass -> INCOMPLETE (fi-hsw-4770r)
Pretty sure this isn't my doing, but lacking access to the full logs I
can't say for sure what's actually happened here.
Cheers,
Daniel
[-- Attachment #1.2: Type: text/html, Size: 1008 bytes --]
[-- Attachment #2: Type: text/plain, Size: 160 bytes --]
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: ✗ Fi.CI.BAT: failure for drm: i915: Wait for fences on new fb, not old
2016-10-21 18:49 ` Daniel Stone
@ 2016-10-21 19:36 ` Saarinen, Jani
2016-10-21 19:48 ` Chris Wilson
1 sibling, 0 replies; 6+ messages in thread
From: Saarinen, Jani @ 2016-10-21 19:36 UTC (permalink / raw)
To: Daniel Stone, intel-gfx; +Cc: Daniel Stone
> Subject: Re: [Intel-gfx] ✗ Fi.CI.BAT: failure for drm: i915: Wait for fences on new fb, not old
>
> On Friday, 21 October 2016, Patchwork <patchwork@emeril.freedesktop.org
> <mailto:patchwork@emeril.freedesktop.org> > wrote:
>
> Test gem_ringfill:
> Subgroup basic-default-hang:
> pass -> TIMEOUT (fi-hsw-4770r)
> Subgroup basic-default-interruptible:
> pass -> INCOMPLETE (fi-hsw-4770r)
> Test gem_storedw_loop:
> Subgroup basic-blt:
> pass -> INCOMPLETE (fi-hsw-4770r)
> Subgroup basic-bsd:
> pass -> INCOMPLETE (fi-hsw-4770r)
>
Might be:
Results for igt@gem_ringfill@basic-default-hang
Result: timeout
Stdout
An internal exception that should have been handled was not:
Test run time exceeded timeout value (600 seconds)
And rest are just incomplete because of that one.
>
> Pretty sure this isn't my doing, but lacking access to the full logs I can't say for
Yep, soon should be also externally visible.
> sure what's actually happened here.
>
> Cheers,
> Daniel
Jani Saarinen
Intel Finland Oy - BIC 0357606-4 - Westendinkatu 7, 02160 Espoo
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: ✗ Fi.CI.BAT: failure for drm: i915: Wait for fences on new fb, not old
2016-10-21 18:49 ` Daniel Stone
2016-10-21 19:36 ` Saarinen, Jani
@ 2016-10-21 19:48 ` Chris Wilson
1 sibling, 0 replies; 6+ messages in thread
From: Chris Wilson @ 2016-10-21 19:48 UTC (permalink / raw)
To: Daniel Stone; +Cc: intel-gfx, Daniel Stone
On Fri, Oct 21, 2016 at 07:49:34PM +0100, Daniel Stone wrote:
> Hi,
>
> On Friday, 21 October 2016, Patchwork
> <[1]patchwork@emeril.freedesktop.org> wrote:
>
> Test gem_ringfill:
> Subgroup basic-default-hang:
> pass -> TIMEOUT (fi-hsw-4770r)
> Subgroup basic-default-interruptible:
> pass -> INCOMPLETE (fi-hsw-4770r)
> Test gem_storedw_loop:
> Subgroup basic-blt:
> pass -> INCOMPLETE (fi-hsw-4770r)
> Subgroup basic-bsd:
> pass -> INCOMPLETE (fi-hsw-4770r)
>
> Pretty sure this isn't my doing, but lacking access to the full logs I
> can't say for sure what's actually happened here.
It's not. It smells like someone used the wrong spin_lock_irq within an
irq context handler (or spin_lock_irq within spin_lock_irqsave). Haven't
spotted the culprit yet and I may very well be barking up the wrong
tree.
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Intel-gfx] [PATCH] drm: i915: Wait for fences on new fb, not old
2016-10-21 14:44 [PATCH] drm: i915: Wait for fences on new fb, not old Daniel Stone
2016-10-21 16:21 ` ✗ Fi.CI.BAT: failure for " Patchwork
@ 2016-10-24 7:03 ` Daniel Vetter
1 sibling, 0 replies; 6+ messages in thread
From: Daniel Vetter @ 2016-10-24 7:03 UTC (permalink / raw)
To: Daniel Stone; +Cc: intel-gfx, Daniel Vetter, ville.syrjala, stable
On Fri, Oct 21, 2016 at 03:44:54PM +0100, Daniel Stone wrote:
> The previous code would wait for fences on the framebuffer from the old
> plane state to complete, rather than the new, so you would see tearing
> everywhere. Fix this to wait on the new state before we make it active.
>
> Signed-off-by: Daniel Stone <daniels@collabora.com>
> Fixes: 94f050246b42 ("drm/i915: nonblocking commit")
> Cc: stable@vger.kernel.org
> Cc: Daniel Vetter <daniel.vetter@intel.com>
> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Queued for -next, thanks for the patch.
-Daniel
> ---
> drivers/gpu/drm/i915/intel_display.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> index e5ad101..1bdef9e 100644
> --- a/drivers/gpu/drm/i915/intel_display.c
> +++ b/drivers/gpu/drm/i915/intel_display.c
> @@ -14292,7 +14292,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state)
>
> for_each_plane_in_state(state, plane, plane_state, i) {
> struct intel_plane_state *intel_plane_state =
> - to_intel_plane_state(plane_state);
> + to_intel_plane_state(plane->state);
>
> if (!intel_plane_state->wait_req)
> continue;
> --
> 2.9.3
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2016-10-24 7:03 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-10-21 14:44 [PATCH] drm: i915: Wait for fences on new fb, not old Daniel Stone
2016-10-21 16:21 ` ✗ Fi.CI.BAT: failure for " Patchwork
2016-10-21 18:49 ` Daniel Stone
2016-10-21 19:36 ` Saarinen, Jani
2016-10-21 19:48 ` Chris Wilson
2016-10-24 7:03 ` [Intel-gfx] [PATCH] " Daniel Vetter
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.