CVE-2026-89708 PUBLISHED

nfsd: RCU-protect cl_cb_session to fix use-after-free on session teardown

Assigner: Linux
Reserved: 11.09.2026 Published: 11.09.2026 Updated: 13.09.2026

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

nfsd: RCU-protect cl_cb_session to fix use-after-free on session teardown

After a DESTROY_SESSION the per-session teardown path can free a session while rpciod still holds an inflight callback rpc_task that dereferences clp->cl_cb_session. nfsd4_probe_callback_sync() flushes cl_callback_wq, but once nfsd4_run_cb_work() has called rpc_call_async() the rpc_task lives on rpciod; flushing the workqueue does not wait for it. rpc_shutdown_client() does drain rpciod tasks, but uses a 1-second wait_event_timeout — tasks stuck in rpc_delay() (e.g. 2-second NFS4ERR_DELAY retries) can outlive the drain.

<pre>destroy path rpciod ------------ ------ unhash_session(ses) nfsd4_probe_callback_sync(clp) flush_workqueue(cl_callback_wq) /* returns; rpc_task still live */ nfsd4_put_session_locked(ses) free_session(ses) -> kfree(ses) nfsd4_cb_sequence_done() reads cb_clp->cl_cb_session /* freed slab */ </pre>

A second window exists in nfsd4_process_cb_update(). When __nfsd4_find_backchannel() returns NULL because unhash_session() has already removed the destroyed session from cl_sessions, setup_callback_client() takes the v4.1 early return so clp->cl_cb_session = ses never fires and the field retains a pointer to the about-to-be-freed session.

Fix both by converting cl_cb_session to an RCU-protected pointer:

  • Move the cl_cb_session = ses assignment in setup_callback_client() to after rpc_create() succeeds, so it is only published when a working backchannel exists. Clear cl_cb_session on the error return in nfsd4_process_cb_update(). Both stores use rcu_assign_pointer().

  • Annotate cl_cb_session with __rcu. All rpciod-side readers use rcu_read_lock()/rcu_dereference() and check for NULL, bailing to the appropriate error or requeue path: encode_cb_sequence4args(), decode_cb_sequence4resok(), nfsd41_cb_get_slot(), nfsd41_cb_release_slot(), nfsd4_cb_prepare(), and nfsd4_cb_sequence_done().

  • Switch __free_session() from kfree() to kfree_rcu() so the session slab is not reclaimed until after an RCU grace period, guaranteeing that rpciod readers inside rcu_read_lock() never dereference freed memory.

  • Pass the session pointer to the nfsd_cb_seq_status and nfsd_cb_free_slot tracepoints instead of having them re-read cl_cb_session.

  • nfsd4_cb_prepare() calls rpc_exit() when the session is NULL, routing through the done/release path to requeue the callback.

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 UAF is in nfsd NFSv4.1 session teardown (nfsd4_destroy_session) and the rpciod callback path (nfsd4_cb_sequence_done). DESTROY_SESSION and the inflight CB_COMPOUND run over TCP/2049, so a remote NFS client reaches the bug with no local access. AC:L - The attacker controls both sides: they keep a callback rpc_task on rpciod (e.g. NFS4ERR_DELAY so rpc_delay is 2s) and send DESTROY_SESSION. flush_workqueue does not wait for rpciod, and rpc_shutdown_client's 1s drain can expire first, so the window is attacker-driven and retryable. PR:N - Typical nfsd exports use AUTH_SYS/AUTH_NULL, so RPC identities are self-asserted. EXCHANGE_ID, CREATE_SESSION with a backchannel, OPEN, and DESTROY_SESSION succeed for any host allowed by the export; nfsd4_mach_creds_match is a no-op under default SP4_NONE. No local account is required. UI:N - The attacking NFS client issues CREATE_SESSION, OPEN/CB_RECALL traffic, NFS4ERR_DELAY replies, and DESTROY_SESSION itself. No victim on the server must mount a filesystem, open a file, or take any other action. S:U - The use-after-free corrupts a kmalloc'd nfsd4_session in the host nfsd kernel. Impact stays in that kernel security authority and does not cross a VM, IOMMU, or sandbox boundary. C:H - After free_session kfree's the session, rpciod still reads se_sessionid and se_cb_seq_nr in encode_cb_sequence4args and decode_cb_sequence4resok. Reclaim of that slab lets the attacker disclose kernel memory via the callback they receive. Per kernel guidance a UAF is C:H. I:H - The same dangling cl_cb_session is written by update_cb_slot_table (attacker-controlled target_highest_slotid), grab_slot/nfsd41_cb_release_slot (se_cb_slot_avail and se_lock), and nfsd4_cb_sequence_done (se_cb_seq_nr++). That is a UAF write primitive for heap corruption and control-flow hijack. A:H - Use-after-free of struct nfsd4_session from an rpciod callback oopses or panics the NFS server even without full exploitation, and the attacker can repeat DESTROY_SESSION against inflight callbacks to deny service.

Product Status

Vendor Linux
Product Linux
Versions Default: unaffected
  • affected from dcbeaa68dbbdacbbb330a86c7fc95a28473fc209 to f164eb52b6f3cbf40f07fe379f9f421f88e02f76 (excl.)
  • affected from dcbeaa68dbbdacbbb330a86c7fc95a28473fc209 to 13bdd486c3aad4fc19e6d8b9c3556a4b4c190b25 (excl.)
  • affected from dcbeaa68dbbdacbbb330a86c7fc95a28473fc209 to 01c5d5f58a5db9b0ee5afba2e49d3157788687b2 (excl.)
Vendor Linux
Product Linux
Versions Default: affected
  • Version 2.6.38 is affected
  • unaffected from 0 to 2.6.38 (excl.)
  • unaffected from 6.18.50 to 6.18.* (incl.)
  • unaffected from 7.2.4 to 7.2.* (incl.)
  • unaffected from 7.3-rc1 to * (incl.)

References