* [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov
@ 2019-01-23 9:50 wentalou
[not found] ` <1548237051-31881-1-git-send-email-Wentao.Lou-5C7GfCeVMHo@public.gmane.org>
0 siblings, 1 reply; 6+ messages in thread
From: wentalou @ 2019-01-23 9:50 UTC (permalink / raw)
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW; +Cc: wentalou
sriov's gpu_recover inside xgpu_ai_mailbox_flr_work would cause duplicate recover in TDR.
TDR's gpu_recover would be triggered by amdgpu_job_timedout,
that could avoid vk-cts failure by unexpected recover.
Change-Id: Ifcba4ac43a0229ae19061aad3b0ddc96957ff9c6
Signed-off-by: wentalou <Wentao.Lou@amd.com>
---
drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
index b11a1c17..f227633 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
@@ -266,7 +266,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
}
/* Trigger recovery for world switch failure if no TDR */
- if (amdgpu_device_should_recover_gpu(adev))
+ if (amdgpu_device_should_recover_gpu(adev) && amdgpu_lockup_timeout == 0)
amdgpu_device_gpu_recover(adev, NULL);
}
--
2.7.4
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply related [flat|nested] 6+ messages in thread[parent not found: <1548237051-31881-1-git-send-email-Wentao.Lou-5C7GfCeVMHo@public.gmane.org>]
* Re: [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov [not found] ` <1548237051-31881-1-git-send-email-Wentao.Lou-5C7GfCeVMHo@public.gmane.org> @ 2019-01-23 16:26 ` Grodzovsky, Andrey [not found] ` <609eec03-4d51-c2e2-8e1f-aefb996c5efa-5C7GfCeVMHo@public.gmane.org> 0 siblings, 1 reply; 6+ messages in thread From: Grodzovsky, Andrey @ 2019-01-23 16:26 UTC (permalink / raw) To: Lou, Wentao, amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org On 01/23/2019 04:50 AM, wentalou wrote: > sriov's gpu_recover inside xgpu_ai_mailbox_flr_work would cause duplicate recover in TDR. > TDR's gpu_recover would be triggered by amdgpu_job_timedout, > that could avoid vk-cts failure by unexpected recover. > > Change-Id: Ifcba4ac43a0229ae19061aad3b0ddc96957ff9c6 > Signed-off-by: wentalou <Wentao.Lou@amd.com> > --- > drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > index b11a1c17..f227633 100644 > --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > @@ -266,7 +266,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) > } > > /* Trigger recovery for world switch failure if no TDR */ > - if (amdgpu_device_should_recover_gpu(adev)) > + if (amdgpu_device_should_recover_gpu(adev) && amdgpu_lockup_timeout == 0) Not sure I fully understand the intent here but amdgpu_lockup_timeout == 0 is forced to 10000 in amdgpu_device_check_arguments and hence not an indication of disabled gpu recover - it will still happen. Andrey > amdgpu_device_gpu_recover(adev, NULL); > } > _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 6+ messages in thread
[parent not found: <609eec03-4d51-c2e2-8e1f-aefb996c5efa-5C7GfCeVMHo@public.gmane.org>]
* RE: [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov [not found] ` <609eec03-4d51-c2e2-8e1f-aefb996c5efa-5C7GfCeVMHo@public.gmane.org> @ 2019-01-24 7:18 ` Lou, Wentao [not found] ` <BYAPR12MB2742ED87810B57383393AACF839A0-ZGDeBxoHBPmbrehcvEBedAdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org> 0 siblings, 1 reply; 6+ messages in thread From: Lou, Wentao @ 2019-01-24 7:18 UTC (permalink / raw) To: Grodzovsky, Andrey, amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org Cc: Liu, Monk ++Monk Yes amdgpu_lockup_timeout was forced to 10000 in current code, that means amdgpu_device_gpu_recover was not triggered inside xgpu_ai_mailbox_flr_work. For sriov, amdgpu_device_gpu_recover was already called by amdgpu_job_timedout, unexpected amdgpu_device_gpu_recover called by xgpu_ai_mailbox_flr_work would make vk-cts failed. -----Original Message----- From: Grodzovsky, Andrey <Andrey.Grodzovsky@amd.com> Sent: Thursday, January 24, 2019 12:27 AM To: Lou, Wentao <Wentao.Lou@amd.com>; amd-gfx@lists.freedesktop.org Subject: Re: [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov On 01/23/2019 04:50 AM, wentalou wrote: > sriov's gpu_recover inside xgpu_ai_mailbox_flr_work would cause duplicate recover in TDR. > TDR's gpu_recover would be triggered by amdgpu_job_timedout, that > could avoid vk-cts failure by unexpected recover. > > Change-Id: Ifcba4ac43a0229ae19061aad3b0ddc96957ff9c6 > Signed-off-by: wentalou <Wentao.Lou@amd.com> > --- > drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > index b11a1c17..f227633 100644 > --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > @@ -266,7 +266,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) > } > > /* Trigger recovery for world switch failure if no TDR */ > - if (amdgpu_device_should_recover_gpu(adev)) > + if (amdgpu_device_should_recover_gpu(adev) && amdgpu_lockup_timeout > +== 0) Not sure I fully understand the intent here but amdgpu_lockup_timeout == 0 is forced to 10000 in amdgpu_device_check_arguments and hence not an indication of disabled gpu recover - it will still happen. Andrey > amdgpu_device_gpu_recover(adev, NULL); > } > _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 6+ messages in thread
[parent not found: <BYAPR12MB2742ED87810B57383393AACF839A0-ZGDeBxoHBPmbrehcvEBedAdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>]
* Re: [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov [not found] ` <BYAPR12MB2742ED87810B57383393AACF839A0-ZGDeBxoHBPmbrehcvEBedAdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org> @ 2019-01-24 15:48 ` Grodzovsky, Andrey 0 siblings, 0 replies; 6+ messages in thread From: Grodzovsky, Andrey @ 2019-01-24 15:48 UTC (permalink / raw) To: Lou, Wentao, amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org Cc: Liu, Monk On 01/24/2019 02:18 AM, Lou, Wentao wrote: > ++Monk > > Yes amdgpu_lockup_timeout was forced to 10000 in current code, that means amdgpu_device_gpu_recover was not triggered inside xgpu_ai_mailbox_flr_work. > For sriov, amdgpu_device_gpu_recover was already called by amdgpu_job_timedout, unexpected amdgpu_device_gpu_recover called by xgpu_ai_mailbox_flr_work would make vk-cts failed. So if I understand correctly your goal is to avoid the gpu recover triggered from xgpu_ai_mailbox_flr_work, bot I don't understand how adding 'amdgpu_lockup_timeout == 0' achieves that ? Andrey > > -----Original Message----- > From: Grodzovsky, Andrey <Andrey.Grodzovsky@amd.com> > Sent: Thursday, January 24, 2019 12:27 AM > To: Lou, Wentao <Wentao.Lou@amd.com>; amd-gfx@lists.freedesktop.org > Subject: Re: [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov > > > > On 01/23/2019 04:50 AM, wentalou wrote: >> sriov's gpu_recover inside xgpu_ai_mailbox_flr_work would cause duplicate recover in TDR. >> TDR's gpu_recover would be triggered by amdgpu_job_timedout, that >> could avoid vk-cts failure by unexpected recover. >> >> Change-Id: Ifcba4ac43a0229ae19061aad3b0ddc96957ff9c6 >> Signed-off-by: wentalou <Wentao.Lou@amd.com> >> --- >> drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c >> b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c >> index b11a1c17..f227633 100644 >> --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c >> +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c >> @@ -266,7 +266,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) >> } >> >> /* Trigger recovery for world switch failure if no TDR */ >> - if (amdgpu_device_should_recover_gpu(adev)) >> + if (amdgpu_device_should_recover_gpu(adev) && amdgpu_lockup_timeout >> +== 0) > Not sure I fully understand the intent here but amdgpu_lockup_timeout == 0 is forced to 10000 in amdgpu_device_check_arguments and hence not an indication of disabled gpu recover - it will still happen. > > Andrey > >> amdgpu_device_gpu_recover(adev, NULL); >> } >> > _______________________________________________ > amd-gfx mailing list > amd-gfx@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/amd-gfx _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov
@ 2019-01-30 3:16 wentalou
[not found] ` <1548818214-18078-1-git-send-email-Wentao.Lou-5C7GfCeVMHo@public.gmane.org>
0 siblings, 1 reply; 6+ messages in thread
From: wentalou @ 2019-01-30 3:16 UTC (permalink / raw)
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW; +Cc: wentalou
sriov's gpu_recover inside xgpu_ai_mailbox_flr_work would cause duplicate recover in TDR.
TDR's gpu_recover would be triggered by amdgpu_job_timedout,
that could avoid vk-cts failure by unexpected recover.
Change-Id: I840dfc145e4e1be9ece6eac8d9f3501da9b28ebf
Signed-off-by: wentalou <Wentao.Lou@amd.com>
---
drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
index b11a1c17..73851eb 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
@@ -266,7 +266,8 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
}
/* Trigger recovery for world switch failure if no TDR */
- if (amdgpu_device_should_recover_gpu(adev))
+ if (amdgpu_device_should_recover_gpu(adev)
+ && amdgpu_lockup_timeout == MAX_SCHEDULE_TIMEOUT)
amdgpu_device_gpu_recover(adev, NULL);
}
--
2.7.4
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply related [flat|nested] 6+ messages in thread[parent not found: <1548818214-18078-1-git-send-email-Wentao.Lou-5C7GfCeVMHo@public.gmane.org>]
* Re: [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov [not found] ` <1548818214-18078-1-git-send-email-Wentao.Lou-5C7GfCeVMHo@public.gmane.org> @ 2019-01-30 15:19 ` Grodzovsky, Andrey 0 siblings, 0 replies; 6+ messages in thread From: Grodzovsky, Andrey @ 2019-01-30 15:19 UTC (permalink / raw) To: Lou, Wentao, amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org Acked-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com> Andrey On 01/29/2019 10:16 PM, wentalou wrote: > sriov's gpu_recover inside xgpu_ai_mailbox_flr_work would cause duplicate recover in TDR. > TDR's gpu_recover would be triggered by amdgpu_job_timedout, > that could avoid vk-cts failure by unexpected recover. > > Change-Id: I840dfc145e4e1be9ece6eac8d9f3501da9b28ebf > Signed-off-by: wentalou <Wentao.Lou@amd.com> > --- > drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > index b11a1c17..73851eb 100644 > --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c > @@ -266,7 +266,8 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) > } > > /* Trigger recovery for world switch failure if no TDR */ > - if (amdgpu_device_should_recover_gpu(adev)) > + if (amdgpu_device_should_recover_gpu(adev) > + && amdgpu_lockup_timeout == MAX_SCHEDULE_TIMEOUT) > amdgpu_device_gpu_recover(adev, NULL); > } > _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2019-01-30 15:19 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-01-23 9:50 [PATCH] drm/amdgpu: tighten gpu_recover in mailbox_flr to avoid duplicate recover in sriov wentalou
[not found] ` <1548237051-31881-1-git-send-email-Wentao.Lou-5C7GfCeVMHo@public.gmane.org>
2019-01-23 16:26 ` Grodzovsky, Andrey
[not found] ` <609eec03-4d51-c2e2-8e1f-aefb996c5efa-5C7GfCeVMHo@public.gmane.org>
2019-01-24 7:18 ` Lou, Wentao
[not found] ` <BYAPR12MB2742ED87810B57383393AACF839A0-ZGDeBxoHBPmbrehcvEBedAdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>
2019-01-24 15:48 ` Grodzovsky, Andrey
-- strict thread matches above, loose matches on Subject: below --
2019-01-30 3:16 wentalou
[not found] ` <1548818214-18078-1-git-send-email-Wentao.Lou-5C7GfCeVMHo@public.gmane.org>
2019-01-30 15:19 ` Grodzovsky, Andrey
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox