PT-2026-152612
CVE-2026-99732
·
Published
2026-11-20
·
Updated
2026-11-20
Description by
Score
N/A
Not availableNo severity ratings or metrics are available. When they are, we'll update the corresponding info on the page.
In the Linux kernel, the following vulnerability has been resolved:
mm/filemap: read mapping->a ops->free folio under i lock
memory failure() -> ... -> filemap remove folio() can read mapping->a ops
after the inode has been freed. Once the folio is deleted from the
pagecache it no longer pins the mapping, so after i lock is dropped
e.g., a concurrent final iput() can evict and free the inode before
filemap free folio() looks up mapping->a ops->free folio.
Most callers hold a reference on the inode, but not all.
Documentation/filesystems/vfs.rst explicitly states that ->free folio()
must not assume the original address space mapping still exists.
Fix it by loading ->free folio while i lock is still held, similar to
what folio unmap invalidate() does.
Inline filemap free folio() into its remaining caller such that
delete from page cache batch() also just loads the callback before
dropping the i lock.
The bug was found by code inspection and has not been reproduced.
Fix generated by an LLM and sent by Josef Bacik in private.
[ david: get rid of filemap free folio() entirely; improve patch description ]
Found an issue in the description? Have something to add? Feel free to write us 👾