On 3/4/25 7:08 PM, Jonathan Cavitt wrote:
The page fault handler should reject write/atomic access to read only
VMAs. Add code to handle this in handle_pagefault after the VMA lookup.
Fixes: 3d420e9fa848 ("drm/xe: Rework GPU page fault handling")
Signed-off-by: Jonathan Cavitt <jonathan.cavitt@xxxxxxxxx>
Suggested-by: Matthew Brost <matthew.brost@xxxxxxxxx>
---
drivers/gpu/drm/xe/xe_gt_pagefault.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/drivers/gpu/drm/xe/xe_gt_pagefault.c b/drivers/gpu/drm/xe/xe_gt_pagefault.c
index 17d69039b866..f608a765fa7c 100644
--- a/drivers/gpu/drm/xe/xe_gt_pagefault.c
+++ b/drivers/gpu/drm/xe/xe_gt_pagefault.c
@@ -235,6 +235,11 @@ static int handle_pagefault(struct xe_gt *gt, struct pagefault *pf)
goto unlock_vm;
}
+ if (xe_vma_read_only(vma) && pf->access_type != ACCESS_TYPE_READ) {
one question, if the PTE Present bit is disabled and the page fault is
caused by atomic load/store eu instruction, will the GuC deliver
pagefault request to KMD with ACCESS_TYPE_READ/ACCESS_TYPE_WRITE Fault
type instead of ACCESS_TYPE_ATOMIC?
G.G.
+ err = -EPERM;
+ goto unlock_vm;
+ }
+
err = handle_vma_pagefault(gt, pf, vma);
unlock_vm:
