http://defect.opensolaris.org/bz/show_bug.cgi?id=1986
Darren J Moffat <darrenm at opensolaris.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |FIXINPROGRESS
--- Comment #2 from Darren J Moffat <darrenm at opensolaris.org> 2008-05-21
07:32:46 ---
Lets try again with another fix!
This time I'm going to keep the spa_keystore_remove() call in the same place
but I'll save the spa_t and objid at the start of the function. I think what
was happening that causes the panic is that they are made invalid by some of
the earlier calls in dsl_dataset_destroy().
--- a/usr/src/uts/common/fs/zfs/dsl_dataset.c Wed May 21 10:50:34 2008 +0100
+++ b/usr/src/uts/common/fs/zfs/dsl_dataset.c Wed May 21 13:49:41 2008 +0100
@@ -871,7 +871,8 @@ dsl_dataset_destroy(dsl_dataset_t *ds, v
dsl_sync_task_group_t *dstg;
objset_t *os;
dsl_dir_t *dd;
- uint64_t obj;
+ uint64_t obj, cryptkeyobj = ds->ds_object;
+ spa_t *spa = dsl_dataset_get_spa(ds);
if (ds->ds_open_refcount != DS_REF_MAX) {
if (dsl_dataset_tryupgrade(ds, DS_MODE_PRIMARY,
@@ -959,8 +960,7 @@ dsl_dataset_destroy(dsl_dataset_t *ds, v
/*
* Remove the key from the keystore for encrypted datasets.
*/
- if ((err = spa_keystore_remove(dsl_dataset_get_spa(ds),
- ds->ds_object)) != 0) {
+ if ((err = spa_keystore_remove(spa, cryptkeyobj)) != 0) {
return (err);
}
I can reproduce the panic with the same stack trace without the above fix but
I've succesfully created and destroyed tank/fs 400+ times without seeing this
panic. So hopefully the fix is correct this time.
--
Configure bugmail: http://defect.opensolaris.org/bz/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
