[PATCH] drm/amdgpu: validate shadow before restoring from it

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



-----Original Message-----
From: Christian König [mailto:deathsimple@xxxxxxxxxxx] 
Sent: Tuesday, April 25, 2017 4:10 PM
To: Zhou, David(ChunMing); He, Hongbo; amd-gfx at lists.freedesktop.org
Subject: Re: [PATCH] drm/amdgpu: validate shadow before restoring from it

Am 25.04.2017 um 05:16 schrieb zhoucm1:
>
>
> On 2017å¹´04æ??25æ?¥ 11:14, Roger.He wrote:
>> Change-Id: Id925f4e241c4192127880d2017fbf2979aa09fc7
>> Signed-off-by: Roger.He <Hongbo.He at amd.com>
>> ---
>>   drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 33
>> ++++++++++++++++++++++++++++++
>>   1 file changed, 33 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> index f74149c..cebd546 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> @@ -2468,6 +2468,27 @@ bool amdgpu_need_backup(struct amdgpu_device
>> *adev)
>>       return amdgpu_lockup_timeout > 0 ? true : false;
>>   }
>>   +static int amdgpu_bo_validate(struct amdgpu_bo *bo)
>> +{
>> +    uint32_t domain;
>> +    int r;
>> +
>> +    if (bo->pin_count)
>> +        return 0;
>> +
>> +    domain = bo->prefered_domains;
>> +
>> +retry:
>> +    amdgpu_ttm_placement_from_domain(bo, domain);
>> +    r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
>> +    if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) {
>> +        domain = bo->allowed_domains;
>> +        goto retry;
>> +    }
>> +
>> +    return r;
>> +}
> you can move this function to amdgpu_object.c, with that fix, it looks 
> ok to me, Reviewed-by: Chunming Zhou <david1.zhou at amd.com>

Yeah, agree. Might even be a good idea to use this helper in amdgpu_cs.c as well. IIRC we had something very similar in there as well.

Yes, I have noticed  they have some similar logic code. Later if have time,  I will try to do that.  Thanks for your suggestion!

Thanks
Roger(Hongbo.He)

But that should be a separate patch, with the function moved to amdgpu_object.c this patch is Reviewed-by: Christian König <christian.koenig at amd.com> as well.

Regards,
Christian.

>
>> +
>>   static int amdgpu_recover_vram_from_shadow(struct amdgpu_device *adev,
>>                          struct amdgpu_ring *ring,
>>                          struct amdgpu_bo *bo, @@ -2485,6 +2506,18 @@ 
>> static int amdgpu_recover_vram_from_shadow(struct amdgpu_device 
>> *adev,
>>       domain = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type);
>>       /* if bo has been evicted, then no need to recover */
>>       if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
>> +        r = amdgpu_bo_validate(bo->shadow);
>> +        if (r) {
>> +            DRM_ERROR("bo validate failed!\n");
>> +            goto err;
>> +        }
>> +
>> +        r = amdgpu_ttm_bind(&bo->shadow->tbo, &bo->shadow->tbo.mem);
>> +        if (r) {
>> +            DRM_ERROR("%p bind failed\n", bo->shadow);
>> +            goto err;
>> +        }
>> +
>>           r = amdgpu_bo_restore_from_shadow(adev, ring, bo,
>>                            NULL, fence, true);
>>           if (r) {
>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx




[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux