diff options
| author | ab-oe <[email protected]> | 2016-02-26 08:33:44 +0100 |
|---|---|---|
| committer | Brian Behlendorf <[email protected]> | 2016-02-29 09:07:27 -0800 |
| commit | 513168abd267f84ebcb688b4e1d42a2abd4fedce (patch) | |
| tree | e06741f3535e627968c944fe2c84dc7295e68608 | |
| parent | 9f500936c82137ef3a57c53013894f622dcec14e (diff) | |
Make zvol update volsize operation synchronous.
There is a race condition when new transaction group is added
to dp->dp_dirty_datasets list by the zap_update in the zvol_update_volsize.
Meanwhile, before these dirty data are synchronized, the receive process
can cause that dmu_recv_end_sync is executed. Then finally dirty data
are going to be synchronized but the synchronization ends with the NULL
pointer dereference error.
Signed-off-by: ab-oe <[email protected]>
Signed-off-by: Brian Behlendorf <[email protected]>
Closes #4116
| -rw-r--r-- | module/zfs/zvol.c | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/module/zfs/zvol.c b/module/zfs/zvol.c index c90e4ec6b..034cf6a6a 100644 --- a/module/zfs/zvol.c +++ b/module/zfs/zvol.c @@ -276,6 +276,7 @@ zvol_update_volsize(uint64_t volsize, objset_t *os) { dmu_tx_t *tx; int error; + uint64_t txg; ASSERT(MUTEX_HELD(&zvol_state_lock)); @@ -287,11 +288,14 @@ zvol_update_volsize(uint64_t volsize, objset_t *os) dmu_tx_abort(tx); return (SET_ERROR(error)); } + txg = dmu_tx_get_txg(tx); error = zap_update(os, ZVOL_ZAP_OBJ, "size", 8, 1, &volsize, tx); dmu_tx_commit(tx); + txg_wait_synced(dmu_objset_pool(os), txg); + if (error == 0) error = dmu_free_long_range(os, ZVOL_OBJ, volsize, DMU_OBJECT_END); |