Pavel Snajdr

Results 173 comments of Pavel Snajdr

Sorry, I guess I should have said earlier I'm on something akin to a parental leave practically, I don't have the time budget to move this forward and I think...

@ahrens well the only problem here were the rollback tests, which in their entirety, when done properly, should really outweight SLOC-wise the core of the OverlayFS-related work itself. If you...

@didrocks if you guys have some time to pitch in with some help writing some ZFS tests, that's actually all that's blocking that work to be merged. One major TODO...

@didrocks https://github.com/openzfs/zfs/pull/9600#discussion_r455366047 Thank you for your offer! ;) Edit: also, the OverlayFS tests themselves in https://github.com/openzfs/zfs/pull/9414 (currently @https://github.com/openzfs/zfs/pull/9414/commits/5ce120c3f04fff4e9d425ca4c9e51edb970225da) are pretty weak. Those also need a bit of attention & love.

This bug is still relevant for me with current zfs master (as of today); lockdep reveals this, it is probably relevant, I'm investigating further: ``` [ 140.025805] ====================================================== [ 140.025806]...

Proposed fix, testing: ``` diff --git a/module/zfs/zil.c b/module/zfs/zil.c index 9dc20ba14..cf4d6d9dc 100644 --- a/module/zfs/zil.c +++ b/module/zfs/zil.c @@ -2513,8 +2513,8 @@ zil_commit_waiter_timeout(zilog_t *zilog, zil_commit_waiter_t *zcw) * elsewhere. */ mutex_exit(&zcw->zcw_lock); - mutex_enter(&zilog->zl_issuer_lock); mutex_enter(&zcw->zcw_lock);...

Oh, it's not going to be that simple :-D

After `noinline`ing all functions in `zil.c`: ``` [ 59.868232] ====================================================== [ 59.868233] WARNING: possible circular locking dependency detected [ 59.868236] 5.9.0-rc1+ #51 Tainted: G O [ 59.868237] ------------------------------------------------------ [ 59.868239]...

It took a while, but the problem has manifested again, running with ~2 months old git HEAD: ``` [[email protected]:~]# for pid in `ct ps -o pid 18205`; do pid=$(ps aux...

Well, our issue seems to be only relevant when core-dumping - https://github.com/openzfs/zfs/pull/10875#issuecomment-767880997