[PATCH v3 06/10] mm/kasan: Support non-blocking GFP in kasan_populate_vmalloc()

From: Uladzislau Rezki (Sony)
Date: Wed Oct 01 2025 - 15:27:45 EST


A "gfp_mask" is already passed to kasan_populate_vmalloc() as
an argument to respect GFPs from callers and KASAN uses it for
its internal allocations.

But apply_to_page_range() function ignores GFP flags due to a
hard-coded mask.

Wrap the call with memalloc_apply_gfp_scope()/memalloc_restore_scope()
so that non-blocking GFP flags(GFP_ATOMIC, GFP_NOWAIT) are respected.

Cc: Andrey Ryabinin <ryabinin.a.a@xxxxxxxxx>
Cc: Alexander Potapenko <glider@xxxxxxxxxx>
Reviewed-by: Baoquan He <bhe@xxxxxxxxxx>
Reviewed-by: Andrey Ryabinin <ryabinin.a.a@xxxxxxxxx>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>
---
mm/kasan/shadow.c | 12 ++----------
1 file changed, 2 insertions(+), 10 deletions(-)

diff --git a/mm/kasan/shadow.c b/mm/kasan/shadow.c
index 5d2a876035d6..a30d84bfdd52 100644
--- a/mm/kasan/shadow.c
+++ b/mm/kasan/shadow.c
@@ -377,18 +377,10 @@ static int __kasan_populate_vmalloc(unsigned long start, unsigned long end, gfp_
* page tables allocations ignore external gfp mask, enforce it
* by the scope API
*/
- if ((gfp_mask & (__GFP_FS | __GFP_IO)) == __GFP_IO)
- flags = memalloc_nofs_save();
- else if ((gfp_mask & (__GFP_FS | __GFP_IO)) == 0)
- flags = memalloc_noio_save();
-
+ flags = memalloc_apply_gfp_scope(gfp_mask);
ret = apply_to_page_range(&init_mm, start, nr_pages * PAGE_SIZE,
kasan_populate_vmalloc_pte, &data);
-
- if ((gfp_mask & (__GFP_FS | __GFP_IO)) == __GFP_IO)
- memalloc_nofs_restore(flags);
- else if ((gfp_mask & (__GFP_FS | __GFP_IO)) == 0)
- memalloc_noio_restore(flags);
+ memalloc_restore_scope(flags);

___free_pages_bulk(data.pages, nr_pages);
if (ret)
--
2.47.3