[PATCH] drm/amdgpu: add dep_sync for amdgpu job

[Christian König]

Am 18.05.2017 um 10:18 schrieb Chunming Zhou:
> The fence in dep_sync cannot be optimized.
>
> Change-Id: Ica1924ad4fe991c0f13438ab521036f60544afcc
> Signed-off-by: Chunming Zhou <David1.Zhou at amd.com>

A bit more commit message wouldn't hurt, but either way the patch is 
Reviewed-by: Christian König <christian.koenig at amd.com>

Regards,
Christian.

> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  2 +-
>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 17 +++++++++++------
>   3 files changed, 13 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index 8351dd2..49bcafd 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -1162,6 +1162,7 @@ struct amdgpu_job {
>   	struct amdgpu_vm	*vm;
>   	struct amdgpu_ring	*ring;
>   	struct amdgpu_sync	sync;
> +	struct amdgpu_sync	dep_sync;
>   	struct amdgpu_sync	sched_sync;
>   	struct amdgpu_ib	*ibs;
>   	struct fence		*fence; /* the hw fence */
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index 2f5ef94..dce3ed6 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -1048,7 +1048,7 @@ static int amdgpu_cs_dependencies(struct amdgpu_device *adev,
>   		}
>   	}
>   
> -	return amdgpu_sem_add_cs(p->ctx, p->job->ring, &p->job->sync);
> +	return amdgpu_sem_add_cs(p->ctx, p->job->ring, &p->job->dep_sync);
>   }
>   
>   static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> index 0c177bb..12f3207 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> @@ -63,6 +63,7 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
>   	(*job)->num_ibs = num_ibs;
>   
>   	amdgpu_sync_create(&(*job)->sync);
> +	amdgpu_sync_create(&(*job)->dep_sync);
>   	amdgpu_sync_create(&(*job)->sched_sync);
>   
>   	return 0;
> @@ -102,6 +103,7 @@ static void amdgpu_job_free_cb(struct amd_sched_job *s_job)
>   
>   	fence_put(job->fence);
>   	amdgpu_sync_free(&job->sync);
> +	amdgpu_sync_free(&job->dep_sync);
>   	amdgpu_sync_free(&job->sched_sync);
>   	kfree(job);
>   }
> @@ -112,6 +114,7 @@ void amdgpu_job_free(struct amdgpu_job *job)
>   
>   	fence_put(job->fence);
>   	amdgpu_sync_free(&job->sync);
> +	amdgpu_sync_free(&job->dep_sync);
>   	amdgpu_sync_free(&job->sched_sync);
>   	kfree(job);
>   }
> @@ -144,9 +147,16 @@ static struct fence *amdgpu_job_dependency(struct amd_sched_job *sched_job)
>   	struct amdgpu_job *job = to_amdgpu_job(sched_job);
>   	struct amdgpu_vm *vm = job->vm;
>   
> -	struct fence *fence = amdgpu_sync_get_fence(&job->sync);
> +	struct fence *fence = amdgpu_sync_get_fence(&job->dep_sync);
>   	int r;
>   
> +	if (amd_sched_dependency_optimized(fence, sched_job->s_entity)) {
> +		r = amdgpu_sync_fence(job->adev, &job->sched_sync, fence);
> +		if (r)
> +			DRM_ERROR("Error adding fence to sync (%d)\n", r);
> +	}
> +	if (!fence)
> +		fence = amdgpu_sync_get_fence(&job->sync);
>   	while (fence == NULL && vm && !job->vm_id) {
>   		struct amdgpu_ring *ring = job->ring;
>   
> @@ -159,11 +169,6 @@ static struct fence *amdgpu_job_dependency(struct amd_sched_job *sched_job)
>   		fence = amdgpu_sync_get_fence(&job->sync);
>   	}
>   
> -	if (amd_sched_dependency_optimized(fence, sched_job->s_entity)) {
> -		r = amdgpu_sync_fence(job->adev, &job->sched_sync, fence);
> -		if (r)
> -			DRM_ERROR("Error adding fence to sync (%d)\n", r);
> -	}
>   	return fence;
>   }
>