RE: [PATCH] drm/amdgpu: For sriov multiple VF, set compute timeout to 10s

["Deng, Emily" <Emily.Deng@amd.com>]

[AMD Official Use Only - Internal Distribution Only]

>-----Original Message-----
>From: Paul Menzel <pmenzel@molgen.mpg.de>
>Sent: Wednesday, January 6, 2021 8:54 PM
>To: Deng, Emily <Emily.Deng@amd.com>
>Cc: amd-gfx@lists.freedesktop.org
>Subject: Re: [PATCH] drm/amdgpu: For sriov multiple VF, set compute timeout
>to 10s
>
>Dear Emily,
>
>
>Am 06.01.21 um 12:41 schrieb Emily.Deng:
>
>Could you please remove the dot your name in your git configuration?
>
>     git config --global user.name "Emily Deng"
Ok, will do this.
>
>For the summary, maybe amend it to:
>
>     Decrease compute timeout to 10 s for sriov multiple VF
Ok, thanks, good suggestion.
>
>> For multiple VF, after engine hang,as host driver will first
>
>Nit: Please add a space after the comma.
>
>> encounter FLR, so has no meanning to set compute to 60s.
>
>meaning
>
>How can this be tested?
Setup the environment for sriov.
>
>> Signed-off-by: Emily.Deng <Emily.Deng@amd.com>
>> ---
>>   drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> index b69c34074d8d..ed36bf97df29 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> @@ -3117,8 +3117,10 @@ static int
>amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
>>    */
>>   adev->gfx_timeout = msecs_to_jiffies(10000);
>>   adev->sdma_timeout = adev->video_timeout = adev->gfx_timeout;
>> -if (amdgpu_sriov_vf(adev) || amdgpu_passthrough(adev))
>> +if ((amdgpu_sriov_vf(adev) && amdgpu_sriov_is_pp_one_vf(adev)) ||
>> +amdgpu_passthrough(adev))
>>   adev->compute_timeout =  msecs_to_jiffies(60000);
>> +else if (amdgpu_sriov_vf(adev))
>> +adev->compute_timeout =  msecs_to_jiffies(10000);
>
>Maybe split up the first if condition to group the condition and not he timeout
>values. At least for me that would be less confusing:
>
>     if (amdgpu_sriov_vf(adev))
>     adev->compute_timeout = amdgpu_sriov_is_pp_one_vf(adev) ?
>msecs_to_jiffies(60000) : msecs_to_jiffies(10000)
>     else if (amdgpu_passthrough(adev))
>     adev->compute_timeout =  msecs_to_jiffies(60000);
>
>>   else
>>   adev->compute_timeout = MAX_SCHEDULE_TIMEOUT;
>
Good suggestion, will send out v2 patch
>
>Kind regards,
>
>Paul
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx