Re: [PATCH] mm/shmem: fix uninitialized folio in shmem_symlink
From: Baolin Wang
Date: Mon Jan 05 2026 - 22:48:08 EST
On 1/5/26 9:58 PM, Brian Foster wrote:
On Thu, Dec 25, 2025 at 06:08:08PM +0800, Baolin Wang wrote:
On 2025/12/25 12:04, Barry Song wrote:
On Thu, Dec 25, 2025 at 6:01 AM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
On Wed, Dec 24, 2025 at 10:40:27PM +1300, Barry Song wrote:
From: Barry Song <baohua@xxxxxxxxxx>
Uninitialized folio allocated in shmem_symlink() may be accessed
during swap-out, causing KMSAN BUG:
This would be an unfortunate way to fix it. The vast majority of
symlinks are short, and we'll never access past the \0 in normal
operation, so we'll be dirtying a lot of cachelines essentially to (1)
shut up an automated tool and (2) optimise a corner case.
How about this instead which delays zeroing to swapout?
Matthew, thank you very much for your review, even during Christmas.
I would like to wish you a happy holiday!
I am not quite sure, as shm symlinks do not seem very common. Since
allocating a folio requires a symname longer than 128 bytes (where
128 == SHORT_SYMLINK_LEN), such cases appear even rarer.
BTW, do we need to migrate the owner_2 flag in folio_migrate_flags()?
If so, I am not quite sure it is worth changing the hotpath to
accommodate this.
+1. At least for me, using the 'PG_owner_2' flag alone to mark this uncommon
case doesn't seem quite worthwhile.
Also JFYI the post-eof swapout zeroing work (still pending) looks to me
like it would cover the swapout time case [1]. That's just if you wanted
to go that route here; creation time zeroing for the large symlink case
seems reasonable enough to me as well.
IMHO, your post-eof zeroing work is intended to zero !uptodate or beyond EOF folios before swap-out, however, the symlink folio is uptodate and not beyond the EOF. I am not sure if it will make your code more complicated when you want to cover this symlink case.
Why I prefer Barry's fix: First, the symlink folio is marked Uptodate after copying the symlink name, but the whole folio hasn’t been initialized, which seems unreasonable to me. Second, as I said before, using the 'PG_owner_2' flag to mark this uncommon case doesn’t seem worthwhile. Currently, IIUC the 'PG_owner_2' is only used by btrfs; if we ever want to remove the 'PG_owner_2', this uncommon symlink case shouldn’t block its removal.
BTW, as I said before, I've reviewed and tested your post-eof work[1]. Maybe you could resend the patch set so that Hugh can take a look when he has time.
[1] https://lore.kernel.org/linux-mm/20251121152246.1023918-3-bfoster@xxxxxxxxxx/
diff --git a/mm/shmem.c b/mm/shmem.c
index ec6c01378e9d..f3b3be1b50fe 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1636,6 +1636,13 @@ int shmem_writeout(struct folio *folio, struct swap_iocb **plug,
folio_mark_uptodate(folio);
}
+ /* Zero out symlink tails to help with compression */
+ if (folio_test_owner_2(folio)) {
+ struct inode *inode = folio->mapping->host;
+ folio_zero_segment(folio, inode->i_size, folio_size(folio));
+ folio_clear_owner_2(folio);
+ }
+
if (!folio_alloc_swap(folio)) {
bool first_swapped = shmem_recalc_inode(inode, 0, nr_pages);
int error;
@@ -4133,6 +4140,7 @@ static int shmem_symlink(struct mnt_idmap *idmap, struct inode *dir,
memcpy(folio_address(folio), symname, len);
folio_mark_uptodate(folio);
folio_mark_dirty(folio);
+ folio_set_owner_2(folio);
folio_unlock(folio);
folio_put(folio);
}
Thanks
Barry