* [PATCH] drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 9:39 ` Chris Wilson
0 siblings, 0 replies; 10+ messages in thread
From: Chris Wilson @ 2019-11-29 9:39 UTC (permalink / raw)
To: intel-gfx
As the i915_active.retire() may be running on another CPU as we detect
that the i915_active is idle, we may not wait for the retirement itself.
Wait for the remote callback by waiting for the retirement worker.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112424
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/i915_active.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
index 479195ecbc6c..e8630ee33336 100644
--- a/drivers/gpu/drm/i915/i915_active.c
+++ b/drivers/gpu/drm/i915/i915_active.c
@@ -469,6 +469,7 @@ int i915_active_wait(struct i915_active *ref)
if (wait_var_event_interruptible(ref, i915_active_is_idle(ref)))
return -EINTR;
+ flush_work(&ref->work);
return 0;
}
--
2.24.0
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [Intel-gfx] [PATCH] drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 9:39 ` Chris Wilson
0 siblings, 0 replies; 10+ messages in thread
From: Chris Wilson @ 2019-11-29 9:39 UTC (permalink / raw)
To: intel-gfx
As the i915_active.retire() may be running on another CPU as we detect
that the i915_active is idle, we may not wait for the retirement itself.
Wait for the remote callback by waiting for the retirement worker.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112424
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/i915_active.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
index 479195ecbc6c..e8630ee33336 100644
--- a/drivers/gpu/drm/i915/i915_active.c
+++ b/drivers/gpu/drm/i915/i915_active.c
@@ -469,6 +469,7 @@ int i915_active_wait(struct i915_active *ref)
if (wait_var_event_interruptible(ref, i915_active_is_idle(ref)))
return -EINTR;
+ flush_work(&ref->work);
return 0;
}
--
2.24.0
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 10+ messages in thread
* ✗ Fi.CI.BAT: failure for drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 10:42 ` Patchwork
0 siblings, 0 replies; 10+ messages in thread
From: Patchwork @ 2019-11-29 10:42 UTC (permalink / raw)
To: Chris Wilson; +Cc: intel-gfx
== Series Details ==
Series: drm/i915: Serialise i915_active_wait() with its retirement
URL : https://patchwork.freedesktop.org/series/70195/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_7440 -> Patchwork_15505
====================================================
Summary
-------
**FAILURE**
Serious unknown changes coming with Patchwork_15505 absolutely need to be
verified manually.
If you think the reported changes have nothing to do with the changes
introduced in Patchwork_15505, please notify your bug team to allow them
to document this new failure mode, which will reduce false positives in CI.
External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/index.html
Possible new issues
-------------------
Here are the unknown changes that may have been introduced in Patchwork_15505:
### IGT changes ###
#### Possible regressions ####
* igt@i915_pm_rpm@basic-pci-d3-state:
- fi-icl-u2: [PASS][1] -> [DMESG-WARN][2]
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-icl-u2/igt@i915_pm_rpm@basic-pci-d3-state.html
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-icl-u2/igt@i915_pm_rpm@basic-pci-d3-state.html
- fi-icl-u3: [PASS][3] -> [DMESG-WARN][4]
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-icl-u3/igt@i915_pm_rpm@basic-pci-d3-state.html
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-icl-u3/igt@i915_pm_rpm@basic-pci-d3-state.html
Known issues
------------
Here are the changes found in Patchwork_15505 that come from known issues:
### IGT changes ###
#### Issues hit ####
* igt@i915_pm_rpm@module-reload:
- fi-kbl-guc: [PASS][5] -> [FAIL][6] ([fdo#107707])
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-guc/igt@i915_pm_rpm@module-reload.html
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-guc/igt@i915_pm_rpm@module-reload.html
#### Possible fixes ####
* igt@i915_pm_rpm@module-reload:
- fi-skl-6770hq: [FAIL][7] ([fdo#108511]) -> [PASS][8]
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-skl-6770hq/igt@i915_pm_rpm@module-reload.html
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-skl-6770hq/igt@i915_pm_rpm@module-reload.html
* igt@i915_selftest@live_gt_heartbeat:
- fi-kbl-guc: [DMESG-FAIL][9] ([fdo#112405]) -> [PASS][10]
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-guc/igt@i915_selftest@live_gt_heartbeat.html
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-guc/igt@i915_selftest@live_gt_heartbeat.html
* igt@kms_chamelium@hdmi-hpd-fast:
- fi-icl-u2: [FAIL][11] ([fdo#109483]) -> [PASS][12]
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-icl-u2/igt@kms_chamelium@hdmi-hpd-fast.html
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-icl-u2/igt@kms_chamelium@hdmi-hpd-fast.html
- fi-kbl-7500u: [FAIL][13] ([fdo#111045] / [fdo#111096]) -> [PASS][14]
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-7500u/igt@kms_chamelium@hdmi-hpd-fast.html
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-7500u/igt@kms_chamelium@hdmi-hpd-fast.html
* igt@kms_frontbuffer_tracking@basic:
- fi-tgl-y: [FAIL][15] ([fdo# 112163] / [fdo#103167]) -> [PASS][16]
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-tgl-y/igt@kms_frontbuffer_tracking@basic.html
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-tgl-y/igt@kms_frontbuffer_tracking@basic.html
#### Warnings ####
* igt@kms_cursor_legacy@basic-busy-flip-before-cursor-legacy:
- fi-kbl-x1275: [DMESG-WARN][17] ([fdo#103558] / [fdo#105602]) -> [DMESG-WARN][18] ([fdo#103558] / [fdo#105602] / [fdo#105763]) +4 similar issues
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-x1275/igt@kms_cursor_legacy@basic-busy-flip-before-cursor-legacy.html
[18]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-x1275/igt@kms_cursor_legacy@basic-busy-flip-before-cursor-legacy.html
* igt@kms_flip@basic-flip-vs-modeset:
- fi-kbl-x1275: [DMESG-WARN][19] ([fdo#103558] / [fdo#105602] / [fdo#105763]) -> [DMESG-WARN][20] ([fdo#103558] / [fdo#105602]) +9 similar issues
[19]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-x1275/igt@kms_flip@basic-flip-vs-modeset.html
[20]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-x1275/igt@kms_flip@basic-flip-vs-modeset.html
{name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).
[fdo# 112163]: https://bugs.freedesktop.org/show_bug.cgi?id= 112163
[fdo#103167]: https://bugs.freedesktop.org/show_bug.cgi?id=103167
[fdo#103558]: https://bugs.freedesktop.org/show_bug.cgi?id=103558
[fdo#105602]: https://bugs.freedesktop.org/show_bug.cgi?id=105602
[fdo#105763]: https://bugs.freedesktop.org/show_bug.cgi?id=105763
[fdo#107707]: https://bugs.freedesktop.org/show_bug.cgi?id=107707
[fdo#108511]: https://bugs.freedesktop.org/show_bug.cgi?id=108511
[fdo#109483]: https://bugs.freedesktop.org/show_bug.cgi?id=109483
[fdo#109964]: https://bugs.freedesktop.org/show_bug.cgi?id=109964
[fdo#111045]: https://bugs.freedesktop.org/show_bug.cgi?id=111045
[fdo#111096]: https://bugs.freedesktop.org/show_bug.cgi?id=111096
[fdo#112298]: https://bugs.freedesktop.org/show_bug.cgi?id=112298
[fdo#112405]: https://bugs.freedesktop.org/show_bug.cgi?id=112405
Participating hosts (52 -> 46)
------------------------------
Missing (6): fi-ilk-m540 fi-hsw-4200u fi-bsw-cyan fi-ctg-p8600 fi-byt-clapper fi-bdw-samus
Build changes
-------------
* CI: CI-20190529 -> None
* Linux: CI_DRM_7440 -> Patchwork_15505
CI-20190529: 20190529
CI_DRM_7440: 7b08e6efdddfcde8ea60a96d5a818af032b52b4d @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_5315: 58705eb1bd29414244f4d0cfa08a9f0ce42d6545 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
Patchwork_15505: 1d4aaedc3c5188faa858c90145565737af08b153 @ git://anongit.freedesktop.org/gfx-ci/linux
== Linux commits ==
1d4aaedc3c51 drm/i915: Serialise i915_active_wait() with its retirement
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/index.html
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 10+ messages in thread
* [Intel-gfx] ✗ Fi.CI.BAT: failure for drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 10:42 ` Patchwork
0 siblings, 0 replies; 10+ messages in thread
From: Patchwork @ 2019-11-29 10:42 UTC (permalink / raw)
To: Chris Wilson; +Cc: intel-gfx
== Series Details ==
Series: drm/i915: Serialise i915_active_wait() with its retirement
URL : https://patchwork.freedesktop.org/series/70195/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_7440 -> Patchwork_15505
====================================================
Summary
-------
**FAILURE**
Serious unknown changes coming with Patchwork_15505 absolutely need to be
verified manually.
If you think the reported changes have nothing to do with the changes
introduced in Patchwork_15505, please notify your bug team to allow them
to document this new failure mode, which will reduce false positives in CI.
External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/index.html
Possible new issues
-------------------
Here are the unknown changes that may have been introduced in Patchwork_15505:
### IGT changes ###
#### Possible regressions ####
* igt@i915_pm_rpm@basic-pci-d3-state:
- fi-icl-u2: [PASS][1] -> [DMESG-WARN][2]
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-icl-u2/igt@i915_pm_rpm@basic-pci-d3-state.html
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-icl-u2/igt@i915_pm_rpm@basic-pci-d3-state.html
- fi-icl-u3: [PASS][3] -> [DMESG-WARN][4]
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-icl-u3/igt@i915_pm_rpm@basic-pci-d3-state.html
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-icl-u3/igt@i915_pm_rpm@basic-pci-d3-state.html
Known issues
------------
Here are the changes found in Patchwork_15505 that come from known issues:
### IGT changes ###
#### Issues hit ####
* igt@i915_pm_rpm@module-reload:
- fi-kbl-guc: [PASS][5] -> [FAIL][6] ([fdo#107707])
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-guc/igt@i915_pm_rpm@module-reload.html
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-guc/igt@i915_pm_rpm@module-reload.html
#### Possible fixes ####
* igt@i915_pm_rpm@module-reload:
- fi-skl-6770hq: [FAIL][7] ([fdo#108511]) -> [PASS][8]
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-skl-6770hq/igt@i915_pm_rpm@module-reload.html
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-skl-6770hq/igt@i915_pm_rpm@module-reload.html
* igt@i915_selftest@live_gt_heartbeat:
- fi-kbl-guc: [DMESG-FAIL][9] ([fdo#112405]) -> [PASS][10]
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-guc/igt@i915_selftest@live_gt_heartbeat.html
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-guc/igt@i915_selftest@live_gt_heartbeat.html
* igt@kms_chamelium@hdmi-hpd-fast:
- fi-icl-u2: [FAIL][11] ([fdo#109483]) -> [PASS][12]
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-icl-u2/igt@kms_chamelium@hdmi-hpd-fast.html
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-icl-u2/igt@kms_chamelium@hdmi-hpd-fast.html
- fi-kbl-7500u: [FAIL][13] ([fdo#111045] / [fdo#111096]) -> [PASS][14]
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-7500u/igt@kms_chamelium@hdmi-hpd-fast.html
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-7500u/igt@kms_chamelium@hdmi-hpd-fast.html
* igt@kms_frontbuffer_tracking@basic:
- fi-tgl-y: [FAIL][15] ([fdo# 112163] / [fdo#103167]) -> [PASS][16]
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-tgl-y/igt@kms_frontbuffer_tracking@basic.html
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-tgl-y/igt@kms_frontbuffer_tracking@basic.html
#### Warnings ####
* igt@kms_cursor_legacy@basic-busy-flip-before-cursor-legacy:
- fi-kbl-x1275: [DMESG-WARN][17] ([fdo#103558] / [fdo#105602]) -> [DMESG-WARN][18] ([fdo#103558] / [fdo#105602] / [fdo#105763]) +4 similar issues
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-x1275/igt@kms_cursor_legacy@basic-busy-flip-before-cursor-legacy.html
[18]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-x1275/igt@kms_cursor_legacy@basic-busy-flip-before-cursor-legacy.html
* igt@kms_flip@basic-flip-vs-modeset:
- fi-kbl-x1275: [DMESG-WARN][19] ([fdo#103558] / [fdo#105602] / [fdo#105763]) -> [DMESG-WARN][20] ([fdo#103558] / [fdo#105602]) +9 similar issues
[19]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7440/fi-kbl-x1275/igt@kms_flip@basic-flip-vs-modeset.html
[20]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/fi-kbl-x1275/igt@kms_flip@basic-flip-vs-modeset.html
{name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).
[fdo# 112163]: https://bugs.freedesktop.org/show_bug.cgi?id= 112163
[fdo#103167]: https://bugs.freedesktop.org/show_bug.cgi?id=103167
[fdo#103558]: https://bugs.freedesktop.org/show_bug.cgi?id=103558
[fdo#105602]: https://bugs.freedesktop.org/show_bug.cgi?id=105602
[fdo#105763]: https://bugs.freedesktop.org/show_bug.cgi?id=105763
[fdo#107707]: https://bugs.freedesktop.org/show_bug.cgi?id=107707
[fdo#108511]: https://bugs.freedesktop.org/show_bug.cgi?id=108511
[fdo#109483]: https://bugs.freedesktop.org/show_bug.cgi?id=109483
[fdo#109964]: https://bugs.freedesktop.org/show_bug.cgi?id=109964
[fdo#111045]: https://bugs.freedesktop.org/show_bug.cgi?id=111045
[fdo#111096]: https://bugs.freedesktop.org/show_bug.cgi?id=111096
[fdo#112298]: https://bugs.freedesktop.org/show_bug.cgi?id=112298
[fdo#112405]: https://bugs.freedesktop.org/show_bug.cgi?id=112405
Participating hosts (52 -> 46)
------------------------------
Missing (6): fi-ilk-m540 fi-hsw-4200u fi-bsw-cyan fi-ctg-p8600 fi-byt-clapper fi-bdw-samus
Build changes
-------------
* CI: CI-20190529 -> None
* Linux: CI_DRM_7440 -> Patchwork_15505
CI-20190529: 20190529
CI_DRM_7440: 7b08e6efdddfcde8ea60a96d5a818af032b52b4d @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_5315: 58705eb1bd29414244f4d0cfa08a9f0ce42d6545 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
Patchwork_15505: 1d4aaedc3c5188faa858c90145565737af08b153 @ git://anongit.freedesktop.org/gfx-ci/linux
== Linux commits ==
1d4aaedc3c51 drm/i915: Serialise i915_active_wait() with its retirement
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_15505/index.html
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 11:15 ` Tvrtko Ursulin
0 siblings, 0 replies; 10+ messages in thread
From: Tvrtko Ursulin @ 2019-11-29 11:15 UTC (permalink / raw)
To: Chris Wilson, intel-gfx
On 29/11/2019 09:39, Chris Wilson wrote:
> As the i915_active.retire() may be running on another CPU as we detect
> that the i915_active is idle, we may not wait for the retirement itself.
> Wait for the remote callback by waiting for the retirement worker.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112424
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> ---
> drivers/gpu/drm/i915/i915_active.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
> index 479195ecbc6c..e8630ee33336 100644
> --- a/drivers/gpu/drm/i915/i915_active.c
> +++ b/drivers/gpu/drm/i915/i915_active.c
> @@ -469,6 +469,7 @@ int i915_active_wait(struct i915_active *ref)
> if (wait_var_event_interruptible(ref, i915_active_is_idle(ref)))
> return -EINTR;
>
> + flush_work(&ref->work);
> return 0;
> }
>
>
Hm, but wake_up_war is in the worker so how does wait_var_event wake the
waiter up before it has been retired?
Regards,
Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [Intel-gfx] [PATCH] drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 11:15 ` Tvrtko Ursulin
0 siblings, 0 replies; 10+ messages in thread
From: Tvrtko Ursulin @ 2019-11-29 11:15 UTC (permalink / raw)
To: Chris Wilson, intel-gfx
On 29/11/2019 09:39, Chris Wilson wrote:
> As the i915_active.retire() may be running on another CPU as we detect
> that the i915_active is idle, we may not wait for the retirement itself.
> Wait for the remote callback by waiting for the retirement worker.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112424
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> ---
> drivers/gpu/drm/i915/i915_active.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
> index 479195ecbc6c..e8630ee33336 100644
> --- a/drivers/gpu/drm/i915/i915_active.c
> +++ b/drivers/gpu/drm/i915/i915_active.c
> @@ -469,6 +469,7 @@ int i915_active_wait(struct i915_active *ref)
> if (wait_var_event_interruptible(ref, i915_active_is_idle(ref)))
> return -EINTR;
>
> + flush_work(&ref->work);
> return 0;
> }
>
>
Hm, but wake_up_war is in the worker so how does wait_var_event wake the
waiter up before it has been retired?
Regards,
Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 11:28 ` Chris Wilson
0 siblings, 0 replies; 10+ messages in thread
From: Chris Wilson @ 2019-11-29 11:28 UTC (permalink / raw)
To: Tvrtko Ursulin, intel-gfx
Quoting Tvrtko Ursulin (2019-11-29 11:15:46)
>
> On 29/11/2019 09:39, Chris Wilson wrote:
> > As the i915_active.retire() may be running on another CPU as we detect
> > that the i915_active is idle, we may not wait for the retirement itself.
> > Wait for the remote callback by waiting for the retirement worker.
> >
> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112424
> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > ---
> > drivers/gpu/drm/i915/i915_active.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
> > index 479195ecbc6c..e8630ee33336 100644
> > --- a/drivers/gpu/drm/i915/i915_active.c
> > +++ b/drivers/gpu/drm/i915/i915_active.c
> > @@ -469,6 +469,7 @@ int i915_active_wait(struct i915_active *ref)
> > if (wait_var_event_interruptible(ref, i915_active_is_idle(ref)))
> > return -EINTR;
> >
> > + flush_work(&ref->work);
> > return 0;
> > }
> >
> >
>
> Hm, but wake_up_war is in the worker so how does wait_var_event wake the
> waiter up before it has been retired?
Remember the wait_event pattern is to skip the wait if COND is already
met. So since the first thing the retirement does is the
atomic_dec_and_test(), we can see ref->count == 0 very early, long
before ref->retire() is called. Our selftest is checking that if
i915_active_wait() reports completion, the callback has also run and
that the i915_active can then be destroyed.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [Intel-gfx] [PATCH] drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 11:28 ` Chris Wilson
0 siblings, 0 replies; 10+ messages in thread
From: Chris Wilson @ 2019-11-29 11:28 UTC (permalink / raw)
To: Tvrtko Ursulin, intel-gfx
Quoting Tvrtko Ursulin (2019-11-29 11:15:46)
>
> On 29/11/2019 09:39, Chris Wilson wrote:
> > As the i915_active.retire() may be running on another CPU as we detect
> > that the i915_active is idle, we may not wait for the retirement itself.
> > Wait for the remote callback by waiting for the retirement worker.
> >
> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112424
> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > ---
> > drivers/gpu/drm/i915/i915_active.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
> > index 479195ecbc6c..e8630ee33336 100644
> > --- a/drivers/gpu/drm/i915/i915_active.c
> > +++ b/drivers/gpu/drm/i915/i915_active.c
> > @@ -469,6 +469,7 @@ int i915_active_wait(struct i915_active *ref)
> > if (wait_var_event_interruptible(ref, i915_active_is_idle(ref)))
> > return -EINTR;
> >
> > + flush_work(&ref->work);
> > return 0;
> > }
> >
> >
>
> Hm, but wake_up_war is in the worker so how does wait_var_event wake the
> waiter up before it has been retired?
Remember the wait_event pattern is to skip the wait if COND is already
met. So since the first thing the retirement does is the
atomic_dec_and_test(), we can see ref->count == 0 very early, long
before ref->retire() is called. Our selftest is checking that if
i915_active_wait() reports completion, the callback has also run and
that the i915_active can then be destroyed.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 11:38 ` Tvrtko Ursulin
0 siblings, 0 replies; 10+ messages in thread
From: Tvrtko Ursulin @ 2019-11-29 11:38 UTC (permalink / raw)
To: Chris Wilson, intel-gfx
On 29/11/2019 11:28, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2019-11-29 11:15:46)
>>
>> On 29/11/2019 09:39, Chris Wilson wrote:
>>> As the i915_active.retire() may be running on another CPU as we detect
>>> that the i915_active is idle, we may not wait for the retirement itself.
>>> Wait for the remote callback by waiting for the retirement worker.
>>>
>>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112424
>>> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
>>> ---
>>> drivers/gpu/drm/i915/i915_active.c | 1 +
>>> 1 file changed, 1 insertion(+)
>>>
>>> diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
>>> index 479195ecbc6c..e8630ee33336 100644
>>> --- a/drivers/gpu/drm/i915/i915_active.c
>>> +++ b/drivers/gpu/drm/i915/i915_active.c
>>> @@ -469,6 +469,7 @@ int i915_active_wait(struct i915_active *ref)
>>> if (wait_var_event_interruptible(ref, i915_active_is_idle(ref)))
>>> return -EINTR;
>>>
>>> + flush_work(&ref->work);
>>> return 0;
>>> }
>>>
>>>
>>
>> Hm, but wake_up_war is in the worker so how does wait_var_event wake the
>> waiter up before it has been retired?
>
> Remember the wait_event pattern is to skip the wait if COND is already
> met. So since the first thing the retirement does is the
> atomic_dec_and_test(), we can see ref->count == 0 very early, long
> before ref->retire() is called. Our selftest is checking that if
> i915_active_wait() reports completion, the callback has also run and
> that the i915_active can then be destroyed.
True!
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Regards,
Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [Intel-gfx] [PATCH] drm/i915: Serialise i915_active_wait() with its retirement
@ 2019-11-29 11:38 ` Tvrtko Ursulin
0 siblings, 0 replies; 10+ messages in thread
From: Tvrtko Ursulin @ 2019-11-29 11:38 UTC (permalink / raw)
To: Chris Wilson, intel-gfx
On 29/11/2019 11:28, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2019-11-29 11:15:46)
>>
>> On 29/11/2019 09:39, Chris Wilson wrote:
>>> As the i915_active.retire() may be running on another CPU as we detect
>>> that the i915_active is idle, we may not wait for the retirement itself.
>>> Wait for the remote callback by waiting for the retirement worker.
>>>
>>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112424
>>> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
>>> ---
>>> drivers/gpu/drm/i915/i915_active.c | 1 +
>>> 1 file changed, 1 insertion(+)
>>>
>>> diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
>>> index 479195ecbc6c..e8630ee33336 100644
>>> --- a/drivers/gpu/drm/i915/i915_active.c
>>> +++ b/drivers/gpu/drm/i915/i915_active.c
>>> @@ -469,6 +469,7 @@ int i915_active_wait(struct i915_active *ref)
>>> if (wait_var_event_interruptible(ref, i915_active_is_idle(ref)))
>>> return -EINTR;
>>>
>>> + flush_work(&ref->work);
>>> return 0;
>>> }
>>>
>>>
>>
>> Hm, but wake_up_war is in the worker so how does wait_var_event wake the
>> waiter up before it has been retired?
>
> Remember the wait_event pattern is to skip the wait if COND is already
> met. So since the first thing the retirement does is the
> atomic_dec_and_test(), we can see ref->count == 0 very early, long
> before ref->retire() is called. Our selftest is checking that if
> i915_active_wait() reports completion, the callback has also run and
> that the i915_active can then be destroyed.
True!
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Regards,
Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2019-11-29 11:38 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-11-29 9:39 [PATCH] drm/i915: Serialise i915_active_wait() with its retirement Chris Wilson
2019-11-29 9:39 ` [Intel-gfx] " Chris Wilson
2019-11-29 10:42 ` ✗ Fi.CI.BAT: failure for " Patchwork
2019-11-29 10:42 ` [Intel-gfx] " Patchwork
2019-11-29 11:15 ` [PATCH] " Tvrtko Ursulin
2019-11-29 11:15 ` [Intel-gfx] " Tvrtko Ursulin
2019-11-29 11:28 ` Chris Wilson
2019-11-29 11:28 ` [Intel-gfx] " Chris Wilson
2019-11-29 11:38 ` Tvrtko Ursulin
2019-11-29 11:38 ` [Intel-gfx] " Tvrtko Ursulin
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.