Re: [PATCH] drm/amdgpu: handle pipeline sync without a VM fence

On 8/14/26 19:29, Alex Deucher wrote:
> If we end up emitting a VM fence keep pipeline sync
> associated with that fence.  If not, emit them as
> part of the IB fence.
>
> v2: fix need_pipe_sync handling
>
> Cc: David Rosca <[email protected]>
> Fixes: cb1e657ccac8 ("drm/amdgpu: handle GDS and SPM without a VM fence")
> Signed-off-by: Alex Deucher <[email protected]>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c | 6 +++++-
>   drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 8 +++++---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 2 +-
>   3 files changed, 11 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> index da4dc489e80bd..360e6f00cb7c0 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> @@ -222,7 +222,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs,
>   		vm_af = job->hw_vm_fence;
>   		/* VM sequence */
>   		vm_af->ib_wptr = ring->wptr;
> -		amdgpu_vm_flush(ring, job, need_pipe_sync, &emit_spm_needed,
> +		amdgpu_vm_flush(ring, job, &need_pipe_sync, &emit_spm_needed,
>   				&emit_gds_needed);
>   		vm_af->ib_dw_size =
>   			amdgpu_ring_get_dw_distance(ring, vm_af->ib_wptr, ring->wptr);
> @@ -235,6 +235,10 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs,
>   	if (ring->funcs->insert_start)
>   		ring->funcs->insert_start(ring);
>   
> +	/* this may have been handled by amdgpu_vm_flush */
> +	if (need_pipe_sync)
> +		amdgpu_ring_emit_pipeline_sync(ring);
> +
>   	if (emit_spm_needed)
>   		adev->gfx.rlc.funcs->update_spm_vmid(adev, ring->xcc_id, ring, job->vmid);
>   
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
> index 71050a86bcc3a..b7d0461184d62 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
> @@ -772,7 +772,7 @@ bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring,
>    * Emit a VM flush when it is necessary.
>    */
>   void amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
> -		     bool need_pipe_sync, bool *emit_spm_needed,
> +		     bool *need_pipe_sync, bool *emit_spm_needed,
>   		     bool *emit_gds_needed)
>   {
>   	struct amdgpu_device *adev = ring->adev;
> @@ -827,7 +827,7 @@ void amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
>   	if (gds_switch_needed && emit_fence)
>   		*emit_gds_needed = false;
>   
> -	if (!vm_flush_needed && !gds_switch_needed && !need_pipe_sync &&
> +	if (!vm_flush_needed && !gds_switch_needed && !(*need_pipe_sync) &&

The only time need_pipe_sync in this check makes a difference is when 
only pasid_mapping_needed (which is missing from this condition, is that 
intended?) is true and the rest *_needed are false. Then emit_fence is 
true and pipeline_sync is emitted in this function which looks fine.

If pasid_mapping_needed and all other *_needed are false, then 
emit_fence is false and the rest of the function effectively does 
nothing. emit_pipeline_sync will be called from amdgpu_ib_schedule. 
While this works, I think it would be better to return early here?

David

>   	    !cleaner_shader_needed && !spm_update_needed)
>   		return;
>   
> @@ -847,8 +847,10 @@ void amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
>   		patch = amdgpu_ring_init_cond_exec(ring,
>   						   ring->cond_exe_gpu_addr);
>   
> -	if (need_pipe_sync)
> +	if (emit_fence && *need_pipe_sync) {
>   		amdgpu_ring_emit_pipeline_sync(ring);
> +		*need_pipe_sync = false;
> +	}
>   
>   	if (cleaner_shader_needed)
>   		ring->funcs->emit_cleaner_shader(ring);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
> index 7f2ba728e3ed3..d32183cd9e0fc 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
> @@ -512,7 +512,7 @@ int amdgpu_vm_validate(struct amdgpu_device *adev, struct amdgpu_vm *vm,
>   		       int (*callback)(void *p, struct amdgpu_bo *bo),
>   		       void *param);
>   void amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
> -		     bool need_pipe_sync, bool *emit_spm_needed,
> +		     bool *need_pipe_sync, bool *emit_spm_needed,
>   		     bool *emit_gds_needed);
>   int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
>   			  struct amdgpu_vm *vm, bool immediate);