Re: [PATCH v2 1/1] drm/amdgpu: cleanup vce,vcn,uvd ring selftests

["Christian König" <ckoenig.leichtzumerken@gmail.com>]

Am 08.01.21 um 16:42 schrieb Nirmoy:
>
> On 1/8/21 4:25 PM, Christian König wrote:
>> Am 18.12.20 um 15:40 schrieb Nirmoy Das:
>>> Use amdgpu_sa_bo instead of amdgpu_bo.
>>>
>>> v2:
>>> * do not initialize bo to get hint from compiler for -Wuninitialized
>>> * pass NULL fence to amdgpu_sa_bo_free if fence is undefined.
>>
>> Found a major bug in this which is probably the reason why that never 
>> worked before.
>>
>> See below.
>>
>>>
>>> Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
>>> ---
>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c | 56 
>>> +++++++------------------
>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c | 17 ++++----
>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 47 ++++++++++-----------
>>>   3 files changed, 45 insertions(+), 75 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c 
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
>>> index 8b989670ed66..13450a3df044 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
>>> @@ -1057,7 +1057,7 @@ int amdgpu_uvd_ring_parse_cs(struct 
>>> amdgpu_cs_parser *parser, uint32_t ib_idx)
>>>       return 0;
>>>   }
>>>
>>> -static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct 
>>> amdgpu_bo *bo,
>>> +static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct 
>>> amdgpu_sa_bo *bo,
>>>                      bool direct, struct dma_fence **fence)
>>>   {
>>>       struct amdgpu_device *adev = ring->adev;
>>> @@ -1071,19 +1071,6 @@ static int amdgpu_uvd_send_msg(struct 
>>> amdgpu_ring *ring, struct amdgpu_bo *bo,
>>>       unsigned offset_idx = 0;
>>>       unsigned offset[3] = { UVD_BASE_SI, 0, 0 };
>>>
>>> -    amdgpu_bo_kunmap(bo);
>>> -    amdgpu_bo_unpin(bo);
>>> -
>>> -    if (!ring->adev->uvd.address_64_bit) {
>>> -        struct ttm_operation_ctx ctx = { true, false };
>>> -
>>> -        amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_VRAM);
>>> -        amdgpu_uvd_force_into_uvd_segment(bo);
>>> -        r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
>>> -        if (r)
>>> -            goto err;
>>> -    }
>>> -
>>>       r = amdgpu_job_alloc_with_ib(adev, 64, direct ? 
>>> AMDGPU_IB_POOL_DIRECT :
>>>                        AMDGPU_IB_POOL_DELAYED, &job);
>>>       if (r)
>>> @@ -1101,7 +1088,7 @@ static int amdgpu_uvd_send_msg(struct 
>>> amdgpu_ring *ring, struct amdgpu_bo *bo,
>>>       data[3] = PACKET0(offset[offset_idx] + UVD_NO_OP, 0);
>>>
>>>       ib = &job->ibs[0];
>>> -    addr = amdgpu_bo_gpu_offset(bo);
>>> +    addr = amdgpu_sa_bo_gpu_addr(bo);
>>>       ib->ptr[0] = data[0];
>>>       ib->ptr[1] = addr;
>>>       ib->ptr[2] = data[1];
>>> @@ -1115,33 +1102,17 @@ static int amdgpu_uvd_send_msg(struct 
>>> amdgpu_ring *ring, struct amdgpu_bo *bo,
>>>       ib->length_dw = 16;
>>>
>>>       if (direct) {
>>> -        r = dma_resv_wait_timeout_rcu(bo->tbo.base.resv,
>>> -                            true, false,
>>> -                            msecs_to_jiffies(10));
>>> -        if (r == 0)
>>> -            r = -ETIMEDOUT;
>>> -        if (r < 0)
>>> -            goto err_free;
>>> -
>>>           r = amdgpu_job_submit_direct(job, ring, &f);
>>>           if (r)
>>>               goto err_free;
>>>       } else {
>>> -        r = amdgpu_sync_resv(adev, &job->sync, bo->tbo.base.resv,
>>> -                     AMDGPU_SYNC_ALWAYS,
>>> -                     AMDGPU_FENCE_OWNER_UNDEFINED);
>>> -        if (r)
>>> -            goto err_free;
>>> -
>>>           r = amdgpu_job_submit(job, &adev->uvd.entity,
>>>                         AMDGPU_FENCE_OWNER_UNDEFINED, &f);
>>>           if (r)
>>>               goto err_free;
>>>       }
>>>
>>> -    amdgpu_bo_fence(bo, f, false);
>>> -    amdgpu_bo_unreserve(bo);
>>> -    amdgpu_bo_unref(&bo);
>>> +    amdgpu_sa_bo_free(adev, &bo, f);
>>>
>>>       if (fence)
>>>           *fence = dma_fence_get(f);
>>> @@ -1153,8 +1124,7 @@ static int amdgpu_uvd_send_msg(struct 
>>> amdgpu_ring *ring, struct amdgpu_bo *bo,
>>>       amdgpu_job_free(job);
>>>
>>>   err:
>>> -    amdgpu_bo_unreserve(bo);
>>> -    amdgpu_bo_unref(&bo);
>>> +    amdgpu_sa_bo_free(adev, &bo, NULL);
>>>       return r;
>>>   }
>>>
>>> @@ -1165,16 +1135,17 @@ int amdgpu_uvd_get_create_msg(struct 
>>> amdgpu_ring *ring, uint32_t handle,
>>>                     struct dma_fence **fence)
>>>   {
>>>       struct amdgpu_device *adev = ring->adev;
>>> -    struct amdgpu_bo *bo = NULL;
>>> +    struct amdgpu_sa_bo *bo;
>>>       uint32_t *msg;
>>>       int r, i;
>>>
>>> -    r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE,
>>> -                      AMDGPU_GEM_DOMAIN_VRAM,
>>> -                      &bo, NULL, (void **)&msg);
>>> +    r = amdgpu_sa_bo_new(&adev->ib_pools[AMDGPU_IB_POOL_DIRECT],
>>> +                 &bo, 1024, PAGE_SIZE);
>>
>> This 1024 here is wrong. The size is in bytes and not double words. 
>> So this needs to be 4096.
>>
>> This only worked because amdgpu_bo_create_reserved always allocated a 
>> full page instead of 1024 bytes.
>>
>> Same problem on all other occasions.
>
>
> I wonder why it worked on newer cards.

Maybe just coincident, but there must be more problems. Even with the 
size fixed I get a hang of the IB tests during boot.

Christian.

>
>
>>
>> Additional to that the direct pool should only be used if we submit 
>> directly, e.g. during a GPU reset.
>>
>> Otherwise we need to use the normal pool.
>
>
> I will resend with above changes.
>
>
> Thanks,
>
> Nirmoy
>
>
>>
>> Regards,
>> Christian.
>>
>>
>>> +
>>>       if (r)
>>>           return r;
>>>
>>> +    msg = amdgpu_sa_bo_cpu_addr(bo);
>>>       /* stitch together an UVD create msg */
>>>       msg[0] = cpu_to_le32(0x00000de4);
>>>       msg[1] = cpu_to_le32(0x00000000);
>>> @@ -1197,16 +1168,17 @@ int amdgpu_uvd_get_destroy_msg(struct 
>>> amdgpu_ring *ring, uint32_t handle,
>>>                      bool direct, struct dma_fence **fence)
>>>   {
>>>       struct amdgpu_device *adev = ring->adev;
>>> -    struct amdgpu_bo *bo = NULL;
>>> +    struct amdgpu_sa_bo *bo;
>>>       uint32_t *msg;
>>>       int r, i;
>>>
>>> -    r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE,
>>> -                      AMDGPU_GEM_DOMAIN_VRAM,
>>> -                      &bo, NULL, (void **)&msg);
>>> +    r = amdgpu_sa_bo_new(&adev->ib_pools[AMDGPU_IB_POOL_DIRECT],
>>> +                 &bo, 1024, PAGE_SIZE);
>>> +
>>>       if (r)
>>>           return r;
>>>
>>> +    msg = amdgpu_sa_bo_cpu_addr(bo);
>>>       /* stitch together an UVD destroy msg */
>>>       msg[0] = cpu_to_le32(0x00000de4);
>>>       msg[1] = cpu_to_le32(0x00000002);
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c 
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
>>> index 0d5284b936e4..bce29d6975d3 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
>>> @@ -81,7 +81,7 @@ MODULE_FIRMWARE(FIRMWARE_VEGA20);
>>>
>>>   static void amdgpu_vce_idle_work_handler(struct work_struct *work);
>>>   static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, 
>>> uint32_t handle,
>>> -                     struct amdgpu_bo *bo,
>>> +                     struct amdgpu_sa_bo *bo,
>>>                        struct dma_fence **fence);
>>>   static int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, 
>>> uint32_t handle,
>>>                         bool direct, struct dma_fence **fence);
>>> @@ -437,7 +437,7 @@ void amdgpu_vce_free_handles(struct 
>>> amdgpu_device *adev, struct drm_file *filp)
>>>    * Open up a stream for HW test
>>>    */
>>>   static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, 
>>> uint32_t handle,
>>> -                     struct amdgpu_bo *bo,
>>> +                     struct amdgpu_sa_bo *bo,
>>>                        struct dma_fence **fence)
>>>   {
>>>       const unsigned ib_size_dw = 1024;
>>> @@ -454,7 +454,7 @@ static int amdgpu_vce_get_create_msg(struct 
>>> amdgpu_ring *ring, uint32_t handle,
>>>
>>>       ib = &job->ibs[0];
>>>
>>> -    addr = amdgpu_bo_gpu_offset(bo);
>>> +    addr = amdgpu_sa_bo_gpu_addr(bo);
>>>
>>>       /* stitch together an VCE create msg */
>>>       ib->length_dw = 0;
>>> @@ -1130,16 +1130,16 @@ int amdgpu_vce_ring_test_ring(struct 
>>> amdgpu_ring *ring)
>>>   int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring, long timeout)
>>>   {
>>>       struct dma_fence *fence = NULL;
>>> -    struct amdgpu_bo *bo = NULL;
>>> +    struct amdgpu_sa_bo *bo = NULL;
>>> +    struct amdgpu_device *adev = ring->adev;
>>>       long r;
>>>
>>>       /* skip vce ring1/2 ib test for now, since it's not reliable */
>>>       if (ring != &ring->adev->vce.ring[0])
>>>           return 0;
>>>
>>> -    r = amdgpu_bo_create_reserved(ring->adev, 512, PAGE_SIZE,
>>> -                      AMDGPU_GEM_DOMAIN_VRAM,
>>> -                      &bo, NULL, NULL);
>>> +    r = amdgpu_sa_bo_new(&adev->ib_pools[AMDGPU_IB_POOL_DIRECT],
>>> +                 &bo, 512, PAGE_SIZE);
>>>       if (r)
>>>           return r;
>>>
>>> @@ -1158,8 +1158,7 @@ int amdgpu_vce_ring_test_ib(struct amdgpu_ring 
>>> *ring, long timeout)
>>>           r = 0;
>>>
>>>   error:
>>> +    amdgpu_sa_bo_free(adev, &bo, fence);
>>>       dma_fence_put(fence);
>>> -    amdgpu_bo_unreserve(bo);
>>> -    amdgpu_bo_unref(&bo);
>>>       return r;
>>>   }
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c 
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
>>> index 4a77c7424dfc..1e46b2f895ba 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
>>> @@ -488,7 +488,7 @@ int amdgpu_vcn_dec_sw_ring_test_ring(struct 
>>> amdgpu_ring *ring)
>>>   }
>>>
>>>   static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring,
>>> -                   struct amdgpu_bo *bo,
>>> +                   struct amdgpu_sa_bo *bo,
>>>                      struct dma_fence **fence)
>>>   {
>>>       struct amdgpu_device *adev = ring->adev;
>>> @@ -504,7 +504,8 @@ static int amdgpu_vcn_dec_send_msg(struct 
>>> amdgpu_ring *ring,
>>>           goto err;
>>>
>>>       ib = &job->ibs[0];
>>> -    addr = amdgpu_bo_gpu_offset(bo);
>>> +    addr = amdgpu_sa_bo_gpu_addr(bo);
>>> +
>>>       ib->ptr[0] = PACKET0(adev->vcn.internal.data0, 0);
>>>       ib->ptr[1] = addr;
>>>       ib->ptr[2] = PACKET0(adev->vcn.internal.data1, 0);
>>> @@ -521,9 +522,7 @@ static int amdgpu_vcn_dec_send_msg(struct 
>>> amdgpu_ring *ring,
>>>       if (r)
>>>           goto err_free;
>>>
>>> -    amdgpu_bo_fence(bo, f, false);
>>> -    amdgpu_bo_unreserve(bo);
>>> -    amdgpu_bo_unref(&bo);
>>> +    amdgpu_sa_bo_free(adev, &bo, f);
>>>
>>>       if (fence)
>>>           *fence = dma_fence_get(f);
>>> @@ -535,25 +534,27 @@ static int amdgpu_vcn_dec_send_msg(struct 
>>> amdgpu_ring *ring,
>>>       amdgpu_job_free(job);
>>>
>>>   err:
>>> -    amdgpu_bo_unreserve(bo);
>>> -    amdgpu_bo_unref(&bo);
>>> +    amdgpu_sa_bo_free(adev, &bo, NULL);
>>>       return r;
>>>   }
>>>
>>>   static int amdgpu_vcn_dec_get_create_msg(struct amdgpu_ring *ring, 
>>> uint32_t handle,
>>> -                     struct amdgpu_bo **bo)
>>> +                     struct amdgpu_sa_bo **bo)
>>>   {
>>>       struct amdgpu_device *adev = ring->adev;
>>>       uint32_t *msg;
>>>       int r, i;
>>>
>>>       *bo = NULL;
>>> -    r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE,
>>> -                      AMDGPU_GEM_DOMAIN_VRAM,
>>> -                      bo, NULL, (void **)&msg);
>>> +
>>> +    r = amdgpu_sa_bo_new(&adev->ib_pools[AMDGPU_IB_POOL_DIRECT],
>>> +                 bo, 1024, PAGE_SIZE);
>>> +
>>>       if (r)
>>>           return r;
>>>
>>> +    msg = amdgpu_sa_bo_cpu_addr(*bo);
>>> +
>>>       msg[0] = cpu_to_le32(0x00000028);
>>>       msg[1] = cpu_to_le32(0x00000038);
>>>       msg[2] = cpu_to_le32(0x00000001);
>>> @@ -575,18 +576,19 @@ static int 
>>> amdgpu_vcn_dec_get_create_msg(struct amdgpu_ring *ring, uint32_t hand
>>>   }
>>>
>>>   static int amdgpu_vcn_dec_get_destroy_msg(struct amdgpu_ring 
>>> *ring, uint32_t handle,
>>> -                      struct amdgpu_bo **bo)
>>> +                      struct amdgpu_sa_bo **bo)
>>>   {
>>>       struct amdgpu_device *adev = ring->adev;
>>>       uint32_t *msg;
>>>       int r, i;
>>>
>>>       *bo = NULL;
>>> -    r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE,
>>> -                      AMDGPU_GEM_DOMAIN_VRAM,
>>> -                      bo, NULL, (void **)&msg);
>>> +    r = amdgpu_sa_bo_new(&adev->ib_pools[AMDGPU_IB_POOL_DIRECT],
>>> +                 bo, 1024, PAGE_SIZE);
>>> +
>>>       if (r)
>>>           return r;
>>> +    msg = amdgpu_sa_bo_cpu_addr(*bo);
>>>
>>>       msg[0] = cpu_to_le32(0x00000028);
>>>       msg[1] = cpu_to_le32(0x00000018);
>>> @@ -603,7 +605,7 @@ static int amdgpu_vcn_dec_get_destroy_msg(struct 
>>> amdgpu_ring *ring, uint32_t han
>>>   int amdgpu_vcn_dec_ring_test_ib(struct amdgpu_ring *ring, long 
>>> timeout)
>>>   {
>>>       struct dma_fence *fence = NULL;
>>> -    struct amdgpu_bo *bo;
>>> +    struct amdgpu_sa_bo *bo;
>>>       long r;
>>>
>>>       r = amdgpu_vcn_dec_get_create_msg(ring, 1, &bo);
>>> @@ -633,7 +635,7 @@ int amdgpu_vcn_dec_ring_test_ib(struct 
>>> amdgpu_ring *ring, long timeout)
>>>   }
>>>
>>>   static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring,
>>> -                   struct amdgpu_bo *bo,
>>> +                   struct amdgpu_sa_bo *bo,
>>>                      struct dma_fence **fence)
>>>   {
>>>       struct amdgpu_vcn_decode_buffer *decode_buffer = NULL;
>>> @@ -651,7 +653,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct 
>>> amdgpu_ring *ring,
>>>           goto err;
>>>
>>>       ib = &job->ibs[0];
>>> -    addr = amdgpu_bo_gpu_offset(bo);
>>> +    addr = amdgpu_sa_bo_gpu_addr(bo);
>>>       ib->length_dw = 0;
>>>
>>>       ib->ptr[ib->length_dw++] = sizeof(struct 
>>> amdgpu_vcn_decode_buffer) + 8;
>>> @@ -671,9 +673,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct 
>>> amdgpu_ring *ring,
>>>       if (r)
>>>           goto err_free;
>>>
>>> -    amdgpu_bo_fence(bo, f, false);
>>> -    amdgpu_bo_unreserve(bo);
>>> -    amdgpu_bo_unref(&bo);
>>> +    amdgpu_sa_bo_free(adev, &bo, f);
>>>
>>>       if (fence)
>>>           *fence = dma_fence_get(f);
>>> @@ -685,15 +685,14 @@ static int amdgpu_vcn_dec_sw_send_msg(struct 
>>> amdgpu_ring *ring,
>>>       amdgpu_job_free(job);
>>>
>>>   err:
>>> -    amdgpu_bo_unreserve(bo);
>>> -    amdgpu_bo_unref(&bo);
>>> +    amdgpu_sa_bo_free(adev, &bo, NULL);
>>>       return r;
>>>   }
>>>
>>>   int amdgpu_vcn_dec_sw_ring_test_ib(struct amdgpu_ring *ring, long 
>>> timeout)
>>>   {
>>>       struct dma_fence *fence = NULL;
>>> -    struct amdgpu_bo *bo;
>>> +    struct amdgpu_sa_bo *bo;
>>>       long r;
>>>
>>>       r = amdgpu_vcn_dec_get_create_msg(ring, 1, &bo);
>>> -- 
>>> 2.29.2
>>>
>>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx