CVE-2024-26629

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

nfsd: fix RELEASE_LOCKOWNER

The test on so_count in nfsd4_release_lockowner() is nonsense and harmful. Revert to using check_for_locks(), changing that to not sleep.

First: harmful. As is documented in the kdoc comment for nfsd4_release_lockowner(), the test on so_count can transiently return a false positive resulting in a return of NFS4ERR_LOCKS_HELD when in fact no locks are held. This is clearly a protocol violation and with the Linux NFS client it can cause incorrect behaviour.

If RELEASE_LOCKOWNER is sent while some other thread is still processing a LOCK request which failed because, at the time that request was received, the given owner held a conflicting lock, then the nfsd thread processing that LOCK request can hold a reference (conflock) to the lock owner that causes nfsd4_release_lockowner() to return an incorrect error.

The Linux NFS client ignores that NFS4ERR_LOCKS_HELD error because it never sends NFS4_RELEASE_LOCKOWNER without first releasing any locks, so it knows that the error is impossible. It assumes the lock owner was in fact released so it feels free to use the same lock owner identifier in some later locking request.

When it does reuse a lock owner identifier for which a previous RELEASE failed, it will naturally use a lock_seqid of zero. However the server, which didnt release the lock owner, will expect a larger lock_seqid and so will respond with NFS4ERR_BAD_SEQID.

So clearly it is harmful to allow a false positive, which testing so_count allows.

The test is nonsense because … well… it doesnt mean anything.

so_count is the sum of three different counts. 1/ the set of states listed on so_stateids 2/ the set of active vfs locks owned by any of those states 3/ various transient counts such as for conflicting locks.

When it is tested against 2 it is clear that one of these is the transient reference obtained by find_lockowner_str_locked(). It is not clear what the other one is expected to be.

In practice, the count is often 2 because there is precisely one state on so_stateids. If there were more, this would fail.

In my testing I see two circumstances when RELEASE_LOCKOWNER is called. In one case, CLOSE is called before RELEASE_LOCKOWNER. That results in all the lock states being removed, and so the lockowner being discarded (it is removed when there are no more references which usually happens when the lock state is discarded). When nfsd4_release_lockowner() finds that the lock owner doesnt exist, it returns success.

The other case shows an so_count of 2 and precisely one state listed in so_stateid. It appears that the Linux client uses a separate lock owner for each file resulting in one lock state per lock owner, so this test on 2 is safe. For another client it might not be safe.

So this patch changes check_for_locks() to use the (newish) find_any_file_locked() so that it doesnt take a reference on the nfs4_file and so never calls nfsd_file_put(), and so never sleeps. With this check is it safe to restore the use of check_for_locks() rather than testing so_count against the mysterious 2.

Affected Software

References

• https://git.kernel.org/stable/c/8f5b860de87039b007e84a28a5eefc888154e098
• https://git.kernel.org/stable/c/99fb654d01dc3f08b5905c663ad6c89a9d83302f
• https://git.kernel.org/stable/c/b7d2eee1f53899b53f069bba3a59a419fc3d331b
• https://git.kernel.org/stable/c/c6f8b3fcc62725e4129f2c0fd550d022d4a7685a
• https://git.kernel.org/stable/c/e4cf8941664cae2f89f0189c29fe2ce8c6be0d03
• https://git.kernel.org/stable/c/edcf9725150e42beeca42d085149f4c88fa97afd