Commit 9e304a18 authored by Brian Geffon's avatar Brian Geffon Committed by Ville Syrjälä
Browse files

drm/i915: Fix page cleanup on DMA remap failure 

When converting to folios the cleanup path of shmem_get_pages() was
missed. When a DMA remap fails and the max segment size is greater than
PAGE_SIZE it will attempt to retry the remap with a PAGE_SIZEd segment
size. The cleanup code isn't properly using the folio apis and as a
result isn't handling compound pages correctly.

v2 -> v3:
(Ville) Just use shmem_sg_free_table() as-is in the failure path of
shmem_get_pages(). shmem_sg_free_table() will clear mapping unevictable
but it will be reset when it retries in shmem_sg_alloc_table().

v1 -> v2:
(Ville) Fixed locations where we were not clearing mapping unevictable.

Cc: stable@vger.kernel.org
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
Cc: Vidya Srinivas <vidya.srinivas@intel.com>
Link: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/13487
Link: https://lore.kernel.org/lkml/20250116135636.410164-1-bgeffon@google.com/


Fixes: 0b62af28 ("i915: convert shmem_sg_free_table() to use a folio_batch")
Signed-off-by: default avatarBrian Geffon <bgeffon@google.com>
Suggested-by: default avatarTomasz Figa <tfiga@google.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250127204332.336665-1-bgeffon@google.com


Reviewed-by: default avatarJonathan Cavitt <jonathan.cavitt@intel.com>
Tested-by: default avatarVidya Srinivas <vidya.srinivas@intel.com>
Signed-off-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
parent 431b742e

drivers/gpu/drm/i915/gem/i915_gem_shmem.c

+1 −5
Original line number Diff line number Diff line
@@ -209,8 +209,6 @@ static int shmem_get_pages(struct drm_i915_gem_object *obj) 
	struct address_space *mapping = obj->base.filp->f_mapping;

	unsigned int max_segment = i915_sg_segment_size(i915->drm.dev);

	struct sg_table *st;

	struct sgt_iter sgt_iter;

	struct page *page;

	int ret;



	/*
@@ -239,9 +237,7 @@ static int shmem_get_pages(struct drm_i915_gem_object *obj) 
		 * for PAGE_SIZE chunks instead may be helpful.

		 */

		if (max_segment > PAGE_SIZE) {

			for_each_sgt_page(page, sgt_iter, st)

				put_page(page);

			sg_free_table(st);

			shmem_sg_free_table(st, mapping, false, false);

			kfree(st);



			max_segment = PAGE_SIZE;