CVE-2026-99366 PUBLISHED

nfsd: don't modify a session slot when replaying its cached reply

Assigner: Linux
Reserved: 25.09.2026 Published: 19.11.2026 Updated: 22.11.2026

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

nfsd: don't modify a session slot when replaying its cached reply

nfsd4_sequence() claims a session slot by setting NFSD4_SLOT_INUSE under nn->client_lock. A reply served from the slot's reply cache does not claim it, and that distinction lived only in cstate->status, which nfsd4_sequence() set to nfserr_replay_cache. Commit cc028a10a48c ("NFSD: Hoist status code encoding into XDR encoder functions") moved nfsd4_proc_compound()'s cstate->status assignment below the out: label, and the replay path's goto out was the one path that relied on skipping it. The test in nfsd4_sequence_done() therefore no longer identifies a replay, and every replay now stores its reply and clears NFSD4_SLOT_INUSE as though it owned the slot.

NFSD4_SLOT_INUSE is the interlock: check_slot_seqid() rejects every sequence id for a slot that is in use, before replay_matches_cache() runs. A replay never sets it, so two replays can be in flight on the same slot at once. One can read slot->sl_cred under nn->client_lock while the other's completion frees and rebuilds it from the XDR encoder without that lock. Two completions can also collide with each other and release the same group_info twice. free_svc_cred() leaves cr_uid, cr_gid and cr_flavor intact, so the reader passes every earlier test in same_creds() and dereferences a NULL cr_group_info. From a 6.12.91 production server:

BUG: kernel NULL pointer dereference, address: 0000000000000004 CPU: 63 UID: 0 PID: 39015 Comm: nfsd RIP: 0010:same_creds+0x38/0xa0 [nfsd] RDX: 0000000000000000 Call Trace: nfsd4_sequence+0x6a8/0x910 [nfsd] nfsd4_proc_compound+0x345/0x670 [nfsd] nfsd_dispatch+0x100/0x220 [nfsd] svc_process_common+0x311/0x700 [sunrpc] svc_process+0x131/0x1c0 [sunrpc] svc_recv+0x7ef/0x9c0 [sunrpc] nfsd+0xa3/0x100 [nfsd] Kernel panic - not syncing: Fatal exception

Since v6.14 a live session's slot table can shrink, and the unowned store becomes a use-after-free write. nfsd4_sequence() defers the shrink while a slot is in use, but it tests NFSD4_SLOT_INUSE, which a replay does not set, so free_session_slots() can kfree() the slot the replay still holds in cstate->slot.

Record the claim in cstate->slot_owned when nfsd4_sequence() accepts a request and test that in nfsd4_sequence_done(), so only the request that claimed the slot updates its cached reply and clears NFSD4_SLOT_INUSE. svc_generic_init_request() zeroes the compound response before each request, so the flag starts clear.

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 attacker sends an NFSv4.1 COMPOUND whose first operation is OP_SEQUENCE, carrying sessionid, slotid, and seqid, over TCP. nfsd_dispatch() calls nfsd4_proc_compound(), which calls nfsd4_sequence(); when seqid equals slot->sl_seqid the replay path runs, and nfs4svc_encode_compoundres() calls nfsd4_sequence_done() on that slot. AC:L - The replay arm of nfsd4_sequence() never sets NFSD4_SLOT_INUSE, and check_slot_seqid() rejects another SEQUENCE for the slot only while that flag is set. The attacker sends two overlapping COMPOUND replays of one slot seqid, so both calls reach nfsd4_store_cache_entry() and free_svc_cred() on the same slot->sl_cred. PR:N - svcauth_unix_accept() copies the RPC_AUTH_SYS uid, gid, and supplementary groups from the call and checks only an AUTH_NULL verifier. The attacker uses that asserted identity for EXCHANGE_ID, CREATE_SESSION, and the replayed OP_SEQUENCE; nfsd4_sequence() accepts it with no capability or GSS check. UI:N - The attacker completes the attack by sending EXCHANGE_ID, CREATE_SESSION, and the overlapping SEQUENCE replays on its own connections. S:U - Impact is confined to the kernel that handled the COMPOUND, the same security authority as the corrupted nfsd4_slot and its embedded svc_cred. C:H - free_svc_cred() releases slot->sl_cred.cr_group_info without nn->client_lock while another replay, in replay_matches_cache(), calls same_creds() and groups_equal() on that pointer. groups_equal() reads ngroups and gid[] from a group_info that put_group_info() may already have passed to groups_free(). I:H - Both replay completions call free_svc_cred() on one slot->sl_cred, so both call put_group_info() on its cr_group_info; groups_free() kvfrees it while the other completion still uses it. A replay also leaves NFSD4_SLOT_INUSE unset, so after nfsd_slot_shrinker_scan() lowers se_target_maxslots, free_session_slots() can kfree the slot and nfsd4_store_cache_entry() writes sl_cred and sl_data through the dangling cstate->slot. A:H - init_svc_cred() in free_svc_cred() clears cr_group_info but leaves cr_uid, cr_gid, and cr_flavor, so same_creds() passes those checks and groups_equal() dereferences NULL. That oops is the fatal nfsd exception reported from nfsd4_sequence().

Product Status

Vendor Linux
Product Linux
Versions Default: unaffected
  • affected from cc028a10a48c3c555d7772d02f56eea9f86fdf79 to 065764e256ebff20201dca5bc6ff075c6cb62e17 (excl.)
  • affected from cc028a10a48c3c555d7772d02f56eea9f86fdf79 to 0b5149b7ad8ccd64c61e887e2cf1af068b6210c9 (excl.)
  • affected from cc028a10a48c3c555d7772d02f56eea9f86fdf79 to 4e3c12151f13c4e51eb381f9377835fb94b9f56a (excl.)
  • affected from cc028a10a48c3c555d7772d02f56eea9f86fdf79 to 2987838317501ff67b672ac2a441fe34b7729f3d (excl.)
Vendor Linux
Product Linux
Versions Default: affected
  • Version 5.10 is affected
  • unaffected from 0 to 5.10 (excl.)
  • unaffected from 6.12.117 to 6.12.* (incl.)
  • unaffected from 6.18.61 to 6.18.* (incl.)
  • unaffected from 7.3.2 to 7.3.* (incl.)
  • unaffected from 7.4-rc1 to * (incl.)

References