diff options
| author | Gunnar Beutner <[email protected]> | 2011-10-11 09:59:02 -0700 |
|---|---|---|
| committer | Brian Behlendorf <[email protected]> | 2011-10-11 09:59:19 -0700 |
| commit | 64c075c3f4381df02d206008b94498248cc49324 (patch) | |
| tree | 3da8bdf5d574f1086273da179b4d40810f9005bd | |
| parent | 763b2f3b57af1febb0fd72c4976992e672a350b4 (diff) | |
Properly destroy work items in spl_kmem_cache_destroy()
In a non-debug build the ASSERT() would be optimized away
which could cause pending work items to not be cancelled.
We must also use cancel_delayed_work_sync() rather than just
cancel_delayed_work() to actually wait until work items have
completed. Otherwise they might accidentally access free'd
memory.
Signed-off-by: Brian Behlendorf <[email protected]>
Closes ZFS bugs #279, #62, #363, #418
| -rw-r--r-- | module/spl/spl-kmem.c | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/module/spl/spl-kmem.c b/module/spl/spl-kmem.c index d71ab11bb..a916f7ef0 100644 --- a/module/spl/spl-kmem.c +++ b/module/spl/spl-kmem.c @@ -1491,10 +1491,10 @@ spl_kmem_cache_destroy(spl_kmem_cache_t *skc) up_write(&spl_kmem_cache_sem); /* Cancel any and wait for any pending delayed work */ - ASSERT(!test_and_set_bit(KMC_BIT_DESTROY, &skc->skc_flags)); - cancel_delayed_work(&skc->skc_work); + VERIFY(!test_and_set_bit(KMC_BIT_DESTROY, &skc->skc_flags)); + cancel_delayed_work_sync(&skc->skc_work); for_each_online_cpu(i) - cancel_delayed_work(&skc->skc_mag[i]->skm_work); + cancel_delayed_work_sync(&skc->skc_mag[i]->skm_work); flush_scheduled_work(); |