Skip to contentSkip to conversation

NFS: fix mount/umount race in nlmclnt.

PATCH In review NeilBrown 2 participants 2 messages Mar 19, 2019

Subsystems (2)
lore
Share & tools
lore

NeilBrown

author
#0 lore raw

PATCH 1/1 NFS: fix mount/umount race in nlmclnt.

If the last NFSv3 unmount from a given host races with a mount from the same host, we can destroy an nlm_host that is still in use.

Specifically nlmclnt_lookup_host() can increment h_count on an nlm_host that nlmclnt_release_host() has just successfully called refcount_dec_and_test() on.
Once nlmclnt_lookup_host() drops the mutex, nlm_destroy_host_lock() will be called to destroy the nlmclnt which is now in use again.

The cause of the problem is that the dec_and_test happens outside the locked region. This is easily fixed by using refcount_dec_and_mutex_lock().

Fixes 8ea6ecc8b075 ("lockd: Create client-side nlm_host cache")
Cc [email protected] (v2.6.38+)
---
 fs/lockd/host.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
fs/lockd/host.c +1 -2
index 93fb7cf0b92b..f0b5c987d6ae 100644
@@ -290,12 +290,11 @@ void nlmclnt_release_host(struct nlm_host *host)
 
 	WARN_ON_ONCE(host->h_server);
 
-	if (refcount_dec_and_test(&host->h_count)) {
+	if (refcount_dec_and_mutex_lock(&host->h_count, &nlm_host_mutex)) {
 		WARN_ON_ONCE(!list_empty(&host->h_lockowners));
 		WARN_ON_ONCE(!list_empty(&host->h_granted));
 		WARN_ON_ONCE(!list_empty(&host->h_reclaim));
 
-		mutex_lock(&nlm_host_mutex);
 		nlm_destroy_host_locked(host);
 		mutex_unlock(&nlm_host_mutex);
 	}
Signature · 1 line
2.14.0.rc0.dirty
#1 lore raw
Quoted 39 lines · NeilBrown

If the last NFSv3 unmount from a given host races with a mount from the
same host, we can destroy an nlm_host that is still in use.

Specifically nlmclnt_lookup_host() can increment h_count on an nlm_host that nlmclnt_release_host() has just successfully called refcount_dec_and_test() on.
Once nlmclnt_lookup_host() drops the mutex, nlm_destroy_host_lock() will be called to destroy the nlmclnt which is now in use again.

The cause of the problem is that the dec_and_test happens outside the locked region. This is easily fixed by using refcount_dec_and_mutex_lock().

Fixes 8ea6ecc8b075 ("lockd: Create client-side nlm_host cache")
Cc [email protected] (v2.6.38+)
---
 fs/lockd/host.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
fs/lockd/host.c +1 -2
index 93fb7cf0b92b..f0b5c987d6ae 100644
@@ -290,12 +290,11 @@ void nlmclnt_release_host(struct nlm_host
 *host)
 
 	WARN_ON_ONCE(host->h_server);
 
-	if (refcount_dec_and_test(&host->h_count)) {
+	if (refcount_dec_and_mutex_lock(&host->h_count,
 &nlm_host_mutex)) {
 		WARN_ON_ONCE(!list_empty(&host->h_lockowners));
 		WARN_ON_ONCE(!list_empty(&host->h_granted));
 		WARN_ON_ONCE(!list_empty(&host->h_reclaim));
 
-		mutex_lock(&nlm_host_mutex);
 		nlm_destroy_host_locked(host);

fs/lockd/host.c void nlmclnt_release_host(struct nlm_host *host) · line 300

Quoted 2 lines
fs/lockd/host.c +0 -0
 		mutex_unlock(&nlm_host_mutex);
 	}

Thanks Neil! Applying to linux-next.

Signature · 3 lines
Trond Myklebust
Linux NFS client maintainer, Hammerspace
[email protected]