My patch is not needed.
Thanks,
Jiadong
-Original Message-
From: Grodzovsky, Andrey
Sent: Friday, July 15, 2022 11:43 PM
To: Zhu, Jiadong ; Christian König
; amd-gfx@lists.freedesktop.org
Cc: Huang, Ray ; Liu, Aaron
Subject: Re: [PATCH 3/3] drm/amdgpu: skip put fence if signal fails
On 202
: RE: [PATCH 3/3] drm/amdgpu: skip put fence if signal fails
Hi Christian,
The resubmitted job in function amdgpu_ib_preempt_job_recovery returns the same
hw fence because of this commit:
static void amdgpu_ib_preempt_job_recovery(struct drm_gpu_scheduler *sched) {
struct drm_sched_job
ig
Sent: Friday, July 15, 2022 4:48 PM
To: Zhu, Jiadong ; amd-gfx@lists.freedesktop.org;
Grodzovsky, Andrey
Cc: Huang, Ray ; Liu, Aaron
Subject: Re: [PATCH 3/3] drm/amdgpu: skip put fence if signal fails
[CAUTION: External Email]
Am 15.07.22 um 10:43 schrieb jiadong@amd.
gfx@lists.freedesktop.org;
Grodzovsky, Andrey
Cc: Huang, Ray ; Liu, Aaron
Subject: Re: [PATCH 3/3] drm/amdgpu: skip put fence if signal fails
[CAUTION: External Email]
Am 15.07.22 um 10:43 schrieb jiadong@amd.com:
> From: "Jiadong.Zhu"
>
> Dma_fence_signal returning non
Am 15.07.22 um 10:43 schrieb jiadong@amd.com:
From: "Jiadong.Zhu"
Dma_fence_signal returning non-zero indicates
that the fence is signaled and put somewhere else.
Skip dma_fence_put to make the fence refcount correct.
Well quite a big NAK on this.
Reference counting should be completely