Re: [PATCH v2 3/4] drm/xe: Bump prefetch BO LRU via GPUSVM devmem callback

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] xe: build failure in xe_svm_devmem_lru_bump when PAGEMAP is disabled
--

commit 940efd09f1bf622d948db6a47af2bd9f2667fa6e
Author: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>

drm/xe: Bump prefetch BO LRU via GPUSVM devmem callback

This patch adds xe_svm_devmem_lru_bump() and uses it as a devmem callback
in the prefetch worker. This allows each backing BO to be moved to the
LRU tail during get_pages.

> diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
> index f39e647512ad9..56c82ff9b39ef 100644
> --- a/drivers/gpu/drm/xe/xe_svm.c
> +++ b/drivers/gpu/drm/xe/xe_svm.c
[ ... ]
> @@ -1625,6 +1626,33 @@ int xe_svm_range_get_pages(struct xe_vm *vm, struct xe_svm_range *range,
>  	return err;
>  }
>  
> +/**
> + * xe_svm_devmem_lru_bump() - Move a range's backing BO to the TTM LRU tail
> + * @devmem_allocation: The device-memory allocation backing the range's pages
> + *
> + * Intended as a &drm_gpusvm_ctx.devmem_fn. Runs under the GPUSVM notifier lock;
> + * the dma-resv trylock avoids inverting against eviction/shrinker, which take
> + * dma-resv before the notifier lock. A contended BO is simply skipped.
> + *
> + * A shared device_private_page_owner can hand us allocations backed by a
> + * foreign driver; skip anything that is not an xe BO before casting.
> + */
> +void xe_svm_devmem_lru_bump(struct drm_pagemap_devmem *devmem_allocation)
> +{
> +	struct xe_bo *bo;
> +
> +	if (devmem_allocation->ops != &dpagemap_devmem_ops)

[Severity: Low]
Does this cause a static build failure when CONFIG_DRM_XE_PAGEMAP is
disabled? 

The function xe_svm_devmem_lru_bump() is compiled unconditionally here,
but it references dpagemap_devmem_ops and calls to_xe_bo(). Since both of
these are omitted when CONFIG_DRM_XE_PAGEMAP is disabled, this will halt
the compiler with undeclared identifier errors in configurations where
CONFIG_DRM_XE_GPUSVM=y but CONFIG_DRM_XE_PAGEMAP=n.

> +		return;
> +
> +	bo = to_xe_bo(devmem_allocation);
> +
> +	if (!dma_resv_trylock(bo->ttm.base.resv))
> +		return;
> +
> +	ttm_bo_move_to_lru_tail_unlocked(&bo->ttm);
> +	dma_resv_unlock(bo->ttm.base.resv);
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260921093651.2995928-6-himal.prasad.ghimiray@intel.com?part=3