RE: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-10 Thread Liu, Monk
op.org Subject: Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2) You still avoid my question: what’s the theoretical backend you that you think check once instead of twice or even more is good*before* hw_job_reset() ? Because stopping all the scheduler threads takes a moment and it is

RE: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-10 Thread Liu, Monk
oun...@lists.freedesktop.org] On Behalf Of Christian König Sent: Wednesday, May 10, 2017 6:26 PM To: Liu, Monk <monk@amd.com>; Koenig, Christian <christian.koe...@amd.com>; amd-gfx@lists.freedesktop.org Subject: Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2) Am 10.0

Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-10 Thread Christian König
lt;mailto:amd-gfx@lists.freedesktop.org> *Subject:* Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2) [ML] if the job complete, the job’s sched fence callback will take this spin_lock and remove itself from mirror_list, so we are still safe to call amd_sched_job

RE: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-10 Thread Liu, Monk
: Wednesday, May 10, 2017 3:52 PM To: Liu, Monk <monk@amd.com>; Koenig, Christian <christian.koe...@amd.com>; amd-gfx@lists.freedesktop.org Subject: Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2) [ML]: One thing I agree to change with your way: in gpu_reset() we should

RE: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-10 Thread Liu, Monk
e.de>; amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org> Subject: Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2) [ML] if the job complete, the job’s sched fence callback will take this spin_lock and remove itself from mirror_list, so we are still safe

Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-09 Thread Christian König
<monk@amd.com>; Christian König <deathsim...@vodafone.de>; amd-gfx@lists.freedesktop.org *Subject:* Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2) [ML] Really not necessary, we have spin_lock to protect the mirror-list, nothing will be messed up ... You

RE: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-09 Thread Liu, Monk
, Monk <monk@amd.com>; Christian König <deathsim...@vodafone.de>; amd-gfx@lists.freedesktop.org Subject: Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2) [ML] Really not necessary, we have spin_lock to protect the mirror-list, nothing will be messed up ... You

Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-09 Thread Christian König
nnocent job and hook The scheduler fence with new hw fence. That way only the guilty job is dropped forever. -Original Message- From: Christian König [mailto:deathsim...@vodafone.de] Sent: Monday, May 08, 2017 9:12 PM To: Liu, Monk <monk@amd.com>; amd-gfx@lists.freedesktop.or

Re: [PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-08 Thread Christian König
Am 08.05.2017 um 09:01 schrieb Liu, Monk: @Christian This one is changed to guilty job scheme accordingly with your response BR Monk -Original Message- From: Monk Liu [mailto:monk@amd.com] Sent: Monday, May 08, 2017 3:00 PM To: amd-gfx@lists.freedesktop.org Cc: Liu, Monk

[PATCH 4/4] drm/amdgpu/SRIOV:implement guilty job TDR (V2)

2017-05-08 Thread Liu, Monk
@Christian This one is changed to guilty job scheme accordingly with your response BR Monk -Original Message- From: Monk Liu [mailto:monk@amd.com] Sent: Monday, May 08, 2017 3:00 PM To: amd-gfx@lists.freedesktop.org Cc: Liu, Monk Subject: [PATCH]