]> git.ipfire.org Git - thirdparty/kernel/linux.git/commit
mm/memfd: use folio_nr_pages() for shmem inode accounting
authorChenghao Duan <duanchenghao@kylinos.cn>
Thu, 26 Mar 2026 08:47:21 +0000 (16:47 +0800)
committerAndrew Morton <akpm@linux-foundation.org>
Sat, 18 Apr 2026 07:10:52 +0000 (00:10 -0700)
commited2a29dc6dcf4630ef19d588704c2ca7b46607bb
tree36d9e1c76fcd9d1ca8aaa600ce68dfc912285f36
parent7cf6d940f4032d87d9cfe6b27c0e49e309818e5d
mm/memfd: use folio_nr_pages() for shmem inode accounting

I found several modifiable points while reading the code.

This patch (of 6):

Patch series "Modify memfd_luo code", v3.

memfd_luo_retrieve_folios() called shmem_inode_acct_blocks() and
shmem_recalc_inode() with hardcoded 1 instead of the actual folio page
count.  memfd may use large folios (THP/hugepages), causing quota/limit
under-accounting and incorrect stat output.

Fix by using folio_nr_pages(folio) for both functions.

Issue found by AI review and suggested by Pratyush Yadav <pratyush@kernel.org>.
https://sashiko.dev/#/patchset/20260319012845.29570-1-duanchenghao%40kylinos.cn

Link: https://lore.kernel.org/20260326084727.118437-1-duanchenghao@kylinos.cn
Link: https://lore.kernel.org/20260326084727.118437-2-duanchenghao@kylinos.cn
Signed-off-by: Chenghao Duan <duanchenghao@kylinos.cn>
Suggested-by: Pratyush Yadav <pratyush@kernel.org>
Reviewed-by: Pasha Tatashin <pasha.tatashin@soleen.com>
Reviewed-by: Pratyush Yadav <pratyush@kernel.org>
Cc: Haoran Jiang <jianghaoran@kylinos.cn>
Cc: Mike Rapoport <rppt@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
mm/memfd_luo.c