diff options
| author | Brian Behlendorf <behlendorf1@llnl.gov> | 2012-12-19 21:51:44 +0000 |
|---|---|---|
| committer | Brian Behlendorf <behlendorf1@llnl.gov> | 2012-12-20 17:57:39 +0000 |
| commit | 65d56083b4617a4cade0cff68cbbaf68114169d6 (patch) | |
| tree | 7707c7061ed4be1f0100649d0f7ff6fe6c3e756a | |
| parent | d5446cfc528262ae3a41da31a8524c9d2e793f45 (diff) | |
Fix zpool on zvol lock inversion deadlock
In all but one case the spa_namespace_lock is taken before the
bdev->bd_mutex lock. But Linux __blkdev_get() function calls
fops->open() with the bdev->bd_mutex lock held and we must
somehow still safely acquire the spa_namespace_lock.
To avoid a potential lock inversion deadlock we preemptively
try to take the spa_namespace_lock(). Normally it will not
be contended and this is safe because spa_open_common() handles
the case where the caller already holds the spa_namespace_lock.
When it is contended we risk a lock inversion if we were to
block waiting for the lock. Luckily, the __blkdev_get()
function allows us to return -ERESTARTSYS which will result in
bdev->bd_mutex being dropped, reacquired, and fops->open() being
called again. This process can be repeated safely until both
locks are acquired.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Jorgen Lundman <lundman@lundman.net>
Closes #612
| -rw-r--r-- | module/zfs/zvol.c | 28 |
1 files changed, 28 insertions, 0 deletions
diff --git a/module/zfs/zvol.c b/module/zfs/zvol.c index 7a448f1948fa..d4d533f0231d 100644 --- a/module/zfs/zvol.c +++ b/module/zfs/zvol.c @@ -891,11 +891,39 @@ zvol_first_open(zvol_state_t *zv) { objset_t *os; uint64_t volsize; + int locked = 0; int error; uint64_t ro; + /* + * In all other cases the spa_namespace_lock is taken before the + * bdev->bd_mutex lock. But in this case the Linux __blkdev_get() + * function calls fops->open() with the bdev->bd_mutex lock held. + * + * To avoid a potential lock inversion deadlock we preemptively + * try to take the spa_namespace_lock(). Normally it will not + * be contended and this is safe because spa_open_common() handles + * the case where the caller already holds the spa_namespace_lock. + * + * When it is contended we risk a lock inversion if we were to + * block waiting for the lock. Luckily, the __blkdev_get() + * function allows us to return -ERESTARTSYS which will result in + * bdev->bd_mutex being dropped, reacquired, and fops->open() being + * called again. This process can be repeated safely until both + * locks are acquired. + */ + if (!mutex_owned(&spa_namespace_lock)) { + locked = mutex_tryenter(&spa_namespace_lock); + if (!locked) + return (-ERESTARTSYS); + } + /* lie and say we're read-only */ error = dmu_objset_own(zv->zv_name, DMU_OST_ZVOL, 1, zvol_tag, &os); + + if (locked) + mutex_exit(&spa_namespace_lock); + if (error) return (-error); |