[PATCH v2 1/1] drm/amdgpu: clean up bo in vce and vcn test

Christian König christian.koenig at amd.com
Wed Dec 9 12:56:09 UTC 2020


Hi Nirmoy,

oh, ok sure. I was thinking to merge it through amd-staging-drm-next 
instead, but going through drm-misc-next is even better.

Regards,
Christian.

Am 09.12.20 um 12:57 schrieb Nirmoy:
> Hi Christian,
>
>
> Can you please pick this up for drm-misc-next?
>
>
> Thanks,
>
> Nirmoy
>
> On 12/9/20 12:51 PM, Nirmoy Das wrote:
>> BO created with amdgpu_bo_create_reserved() wasn't clean
>> properly before, which causes:
>>
>> [   21.056218] WARNING: CPU: 0 PID: 7 at 
>> drivers/gpu/drm/ttm/ttm_bo.c:518 ttm_bo_release+0x2bf/0x310 [ttm]
>> <snip>
>> [   21.056430] Call Trace:
>> [   21.056525]  amdgpu_bo_unref+0x1a/0x30 [amdgpu]
>> [   21.056635]  amdgpu_vcn_dec_send_msg+0x1b2/0x270 [amdgpu]
>> [   21.056740] amdgpu_vcn_dec_get_create_msg.constprop.0+0xd8/0x100 
>> [amdgpu]
>> [   21.056843]  amdgpu_vcn_dec_ring_test_ib+0x27/0x180 [amdgpu]
>> [   21.056936]  amdgpu_ib_ring_tests+0xf1/0x150 [amdgpu]
>> [   21.057024] amdgpu_device_delayed_init_work_handler+0x11/0x30 
>> [amdgpu]
>> [   21.057030]  process_one_work+0x1df/0x370
>> [   21.057033]  worker_thread+0x46/0x340
>> [   21.057034]  ? process_one_work+0x370/0x370
>> [   21.057037]  kthread+0x11b/0x140
>> [   21.057039]  ? __kthread_bind_mask+0x60/0x60
>> [   21.057043]  ret_from_fork+0x22/0x30
>>
>> Signed-off-by: Nirmoy Das <nirmoy.das at amd.com>
>> Reviewed-by: Christian König <christian.koenig at amd.com>
>> ---
>>   drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c | 2 +-
>>   drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 9 ++++++---
>>   2 files changed, 7 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c 
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
>> index ecaa2d7483b2..78a4dd9bf11f 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
>> @@ -1151,6 +1151,6 @@ int amdgpu_vce_ring_test_ib(struct amdgpu_ring 
>> *ring, long timeout)
>>   error:
>>       dma_fence_put(fence);
>>       amdgpu_bo_unreserve(bo);
>> -    amdgpu_bo_unref(&bo);
>> +    amdgpu_bo_free_kernel(&bo, NULL, NULL);
>>       return r;
>>   }
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c 
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
>> index 7e19a6656715..921b81054c1b 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
>> @@ -465,6 +465,7 @@ static int amdgpu_vcn_dec_send_msg(struct 
>> amdgpu_ring *ring,
>>       struct amdgpu_job *job;
>>       struct amdgpu_ib *ib;
>>       uint64_t addr;
>> +    void *msg = NULL;
>>       int i, r;
>>
>>       r = amdgpu_job_alloc_with_ib(adev, 64,
>> @@ -474,6 +475,7 @@ static int amdgpu_vcn_dec_send_msg(struct 
>> amdgpu_ring *ring,
>>
>>       ib = &job->ibs[0];
>>       addr = amdgpu_bo_gpu_offset(bo);
>> +    msg = amdgpu_bo_kptr(bo);
>>       ib->ptr[0] = PACKET0(adev->vcn.internal.data0, 0);
>>       ib->ptr[1] = addr;
>>       ib->ptr[2] = PACKET0(adev->vcn.internal.data1, 0);
>> @@ -492,7 +494,7 @@ static int amdgpu_vcn_dec_send_msg(struct 
>> amdgpu_ring *ring,
>>
>>       amdgpu_bo_fence(bo, f, false);
>>       amdgpu_bo_unreserve(bo);
>> -    amdgpu_bo_unref(&bo);
>> +    amdgpu_bo_free_kernel(&bo, NULL, (void **)&msg);
>>
>>       if (fence)
>>           *fence = dma_fence_get(f);
>> @@ -505,7 +507,7 @@ static int amdgpu_vcn_dec_send_msg(struct 
>> amdgpu_ring *ring,
>>
>>   err:
>>       amdgpu_bo_unreserve(bo);
>> -    amdgpu_bo_unref(&bo);
>> +    amdgpu_bo_free_kernel(&bo, NULL, (void **)&msg);
>>       return r;
>>   }
>>
>> @@ -761,6 +763,7 @@ int amdgpu_vcn_enc_ring_test_ib(struct 
>> amdgpu_ring *ring, long timeout)
>>   error:
>>       dma_fence_put(fence);
>>       amdgpu_bo_unreserve(bo);
>> -    amdgpu_bo_unref(&bo);
>> +    amdgpu_bo_free_kernel(&bo, NULL, NULL);
>> +
>>       return r;
>>   }
>> -- 
>> 2.29.2
>>



More information about the dri-devel mailing list