Re: [PATCH 1/1] nfsd: fix possible badness in FREE_STATEID

[Chuck Lever <chuck.lever@oracle.com>]

On Fri, Oct 04, 2024 at 06:04:03PM -0400, Olga Kornievskaia wrote:
> When multiple FREE_STATEIDs are sent for the same delegation stateid,
> it can lead to a possible either use-after-tree or counter refcount
> underflow errors.
> 
> In nfsd4_free_stateid() under the client lock we find a delegation
> stateid, however the code drops the lock before calling nfs4_put_stid(),
> that allows another FREE_STATE to find the stateid again. The first one
> will proceed to then free the stateid which leads to either
> use-after-free or decrementing already zerod counter.
> 
> CC: stable@vger.kernel.org

I assume that the broken commit is pretty old, but this fix does not
apply before v6.9 (where sc_status is introduced). I can add
"# v6.9+" to the Cc: stable tag.

But what do folks think about a Fixes: tag?

Could be e1ca12dfb1be ("NFSD: added FREE_STATEID operation"), but
that doesn't have the switch statement, which was added by
2da1cec713bc ("nfsd4: simplify free_stateid").


> Signed-off-by: Olga Kornievskaia <okorniev@redhat.com>
> ---
>  fs/nfsd/nfs4state.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> index ac1859c7cc9d..56b261608af4 100644
> --- a/fs/nfsd/nfs4state.c
> +++ b/fs/nfsd/nfs4state.c
> @@ -7154,6 +7154,7 @@ nfsd4_free_stateid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
>  	switch (s->sc_type) {
>  	case SC_TYPE_DELEG:
>  		if (s->sc_status & SC_STATUS_REVOKED) {
> +			s->sc_status |= SC_STATUS_CLOSED;
>  			spin_unlock(&s->sc_lock);
>  			dp = delegstateid(s);
>  			list_del_init(&dp->dl_recall_lru);
> -- 
> 2.43.5
> 

-- 
Chuck Lever