Re: [PATCH 11/18] drm/amdgpu: independence for the amdgpu_eviction_fence!

[Philipp Stanner <phasta@mailbox.org>]

On Thu, 2025-11-13 at 15:51 +0100, Christian König wrote:
> This should allow amdgpu_fences to outlive the amdgpu module.

amdgpu_eviction_fences you mean.

And I would drop the "should" from this and the previous patch. We're
confident it works, aren't we? ^^'

P.

> 
> Signed-off-by: Christian König <christian.koenig@amd.com>
> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c | 3 +--
>  drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h | 1 -
>  2 files changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c
> index 23d7d0b0d625..95ee22c43ceb 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c
> @@ -167,9 +167,8 @@ amdgpu_eviction_fence_create(struct amdgpu_eviction_fence_mgr *evf_mgr)
>  
>  	ev_fence->evf_mgr = evf_mgr;
>  	get_task_comm(ev_fence->timeline_name, current);
> -	spin_lock_init(&ev_fence->lock);
>  	dma_fence_init64(&ev_fence->base, &amdgpu_eviction_fence_ops,
> -			 &ev_fence->lock, evf_mgr->ev_fence_ctx,
> +			 NULL, evf_mgr->ev_fence_ctx,
>  			 atomic_inc_return(&evf_mgr->ev_fence_seq));
>  	return ev_fence;
>  }
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h
> index fcd867b7147d..fb70efb54338 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h
> @@ -27,7 +27,6 @@
>  
>  struct amdgpu_eviction_fence {
>  	struct dma_fence base;
> -	spinlock_t	 lock;
>  	char		 timeline_name[TASK_COMM_LEN];
>  	struct amdgpu_eviction_fence_mgr *evf_mgr;
>  };