Since gfx allocations tend to be large, unmovable and disposable, report
the allocation failure back to userspace as an ENOMEM rather than incur
the oomkiller. We have already tried to make room by purging our own
cached gfx objects, and the oomkiller doesn't attribute ownership of gfx
objects so will likely pick the wrong candidate. Instead, let userspace
see the ENOMEM.
Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
---
drivers/gpu/drm/i915/i915_gem.c | 15 ++++++++++++---
1 file changed, 12 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index bb65072f688d..9d710bd8c6c9 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2321,7 +2321,7 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
st->nents = 0;
for (i = 0; i < page_count; i++) {
page = shmem_read_mapping_page_gfp(mapping, i, gfp);
- if (IS_ERR(page)) {
+ if (unlikely(IS_ERR(page))) {
i915_gem_shrink(dev_priv,
page_count,
I915_SHRINK_BOUND |
@@ -2329,12 +2329,21 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
I915_SHRINK_PURGEABLE);
page = shmem_read_mapping_page_gfp(mapping, i, gfp);
}
- if (IS_ERR(page)) {
+ if (unlikely(IS_ERR(page))) {
+ gfp_t reclaim;
+
/* We've tried hard to allocate the memory by reaping
* our own buffer, now let the real VM do its job and
* go down in flames if truly OOM.
+ *
+ * However, since graphics tend to be disposable,
+ * defer the oom here by reporting the ENOMEM back
+ * to userspace.
*/
- page = shmem_read_mapping_page(mapping, i);
+ reclaim = mapping_gfp_constraint(mapping, 0);
+ reclaim |= __GFP_NORETRY; /* reclaim, but no oom */
+
+ page = shmem_read_mapping_page_gfp(mapping, i, gfp);
if (IS_ERR(page)) {
ret = PTR_ERR(page);
goto err_sg;
--
2.11.0
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
