CVE-2026-72366 PUBLISHED

netfs: Fix netfs_create_write_req() to handle async cache object creation

Assigner: Linux
Reserved: 09.08.2026 Published: 15.08.2026 Updated: 17.08.2026

In the Linux kernel, the following vulnerability has been resolved:

netfs: Fix netfs_create_write_req() to handle async cache object creation

netfs_create_write_req() will skip caching if the fscache cookie is disabled, but this is a problem because async cache object creation might not have got far enough yet that has been enabled - thereby causing the call to fscache_begin_write_operation() to be skipped.

Fix this by removing the checks on the cookie and delegating this to fscache_begin_write_operation().

Metrics

CVSS Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
CVSS Score: 9.8

AV:N - The bug is in netfs client write/cache coordination for CIFS, NFS+fscache, Ceph, AFS, and 9p. A malicious remote filesystem server can deliver file data and control protocol timing to drive async fscache object creation and subsequent kernel writeback over the network. AC:L - The attacker controls both sides of the race by interleaving reads and writes or triggering writeback while the fscache cookie is still in CREATING/LOOKING_UP; the async-creation window can be hit reliably with retries, not conditions outside attacker influence. PR:N - A malicious NFS/SMB/Ceph server needs no privileges on the victim client; once a fscache-enabled mount to that server exists, exploitation requires only ordinary file access or automated writeback, not real root or init-namespace capabilities. UI:N - After a triggering read from the remote server, kernel writeback can invoke netfs_writepages without further victim action; an attacker with file access can also drive the race directly through concurrent read/write syscalls. S:U - Impact is kernel heap/folio corruption during local pagecache I/O within the same kernel security authority; it does not inherently cross VM, container, IOMMU, or sandbox boundaries. C:H - Skipping fscache_begin_write_operation desynchronizes cache and writeback streams; the same netfs/fscache subsystem previously produced UAFs when PG_private_2 coordination failed, enabling disclosure via freed/reused folio memory. I:H - Inconsistent dual-stream write completion can free or reuse folios while still referenced, yielding heap corruption primitives that related netfs fscache bugs have leveraged for arbitrary kernel writes and privilege escalation. A:H - The related netfs fscache fix documented kernel oops, ceph caps warnings, RCU expedited stalls, and UAF crashes from this coordination failure class; skipping required cache setup can oops, panic, or hang the system during writeback.

Product Status

Vendor Linux
Product Linux
Versions Default: unaffected
  • affected from 7b589a9b45ae32aa9d7bece597490e141198d7a6 to 8ab75e445c161c4cb504aa98e20ce1dd1ecd8e9a (excl.)
  • affected from 7b589a9b45ae32aa9d7bece597490e141198d7a6 to 1188a9846fadeacf9d1430b528e5217f749d665b (excl.)
  • affected from 7b589a9b45ae32aa9d7bece597490e141198d7a6 to dbd6f56d975b23241b7bbb11bb8f562af548a0aa (excl.)
Vendor Linux
Product Linux
Versions Default: affected
  • Version 6.11 is affected
  • unaffected from 0 to 6.11 (excl.)
  • unaffected from 6.18.40 to 6.18.* (incl.)
  • unaffected from 7.1.5 to 7.1.* (incl.)
  • unaffected from 7.2 to * (incl.)

References