[PATCH 1/2] drm/xe/exec: move fence reservation

[Souza, Jose]

On Wed, 2023-12-13 at 17:47 +0000, Matthew Auld wrote:
> We currently assume that we can upfront know exactly how many fence
> slots we will need at the start of the exec, however the TTM bo_validate
> can itself consume numerous fence slots, and due to how the
> dma_resv_reserve_fences() works it only ensures that at least that many
> fence slots are available. With this it is quite possible that TTM
> steals some of the fence slots and then when it comes time to do the vma
> binding and final exec stage we are lacking enough fence slots, leading
> to some nasty BUG_ON(). A simple fix is to reserve our own fences later,
> after the validate stage.
> 
> References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/698

Tested-by: José Roberto de Souza <jose.souza at intel.com>

> Signed-off-by: Matthew Auld <matthew.auld at intel.com>
> Cc: Thomas Hellström <thomas.hellstrom at linux.intel.com>
> Cc: Matthew Brost <matthew.brost at intel.com>
> ---
>  drivers/gpu/drm/xe/xe_exec.c | 40 ++++++++++++++++++++++++++++++++++--
>  1 file changed, 38 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c
> index ba92e5619da3..63e82e5285bc 100644
> --- a/drivers/gpu/drm/xe/xe_exec.c
> +++ b/drivers/gpu/drm/xe/xe_exec.c
> @@ -96,7 +96,44 @@
>  
>  static int xe_exec_fn(struct drm_gpuvm_exec *vm_exec)
>  {
> -	return drm_gpuvm_validate(vm_exec->vm, &vm_exec->exec);
> +	struct xe_vm *vm = container_of(vm_exec->vm, struct xe_vm, gpuvm);
> +	struct drm_gem_object *obj;
> +	unsigned long index;
> +	int num_fences;
> +	int ret;
> +
> +	ret = drm_gpuvm_validate(vm_exec->vm, &vm_exec->exec);
> +	if (ret)
> +		return ret;
> +
> +	/*
> +	 * 1 fence slot for the final submit, and one more for every per-tile
> +	 * bind. Note that there are potentially many vma per object/dma-resv,
> +	 * however the fence slot will just be re-used, since they are largely
> +	 * the same timeline and the seqno should be in order.
> +	 */
> +	num_fences = 1 + vm->xe->info.tile_count;
> +
> +	/*
> +	 * We don't know upfront exactly how many fence slots we will need at
> +	 * the start of the exec, since the TTM bo_validate above can consume
> +	 * numerous fence slots. Also due to how the dma_resv_reserve_fences()
> +	 * works it only ensures that at least that many fence slots are
> +	 * available i.e if there are already 10 slots available and we reserve
> +	 * two more, it can just noop without reserving anything.  With this it
> +	 * is quite possible that TTM steals some of the fence slots and then
> +	 * when it comes time to do the vma binding and final exec stage we are
> +	 * lacking enough fence slots, leading to some nasty BUG_ON() when
> +	 * adding the fences. Hence just add our own fences here, after the
> +	 * validate stage.
> +	 */
> +	drm_exec_for_each_locked_object(&vm_exec->exec, index, obj) {
> +		ret = dma_resv_reserve_fences(obj->resv, num_fences);
> +		if (ret)
> +			return ret;
> +	}
> +
> +	return 0;
>  }
>  
>  int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
> @@ -189,7 +226,6 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
>  	}
>  
>  	vm_exec.vm = &vm->gpuvm;
> -	vm_exec.num_fences = 1 + vm->xe->info.tile_count;
>  	vm_exec.flags = DRM_EXEC_INTERRUPTIBLE_WAIT;
>  	if (xe_vm_in_lr_mode(vm)) {
>  		drm_exec_init(exec, vm_exec.flags);