Re: Nasty non-recursive lockmgr panic on softdep only enabled UFS partition when filesystem full

2011/5/4 Kostik Belousov <kostikbel_at_gmail.com>:
> On Tue, May 03, 2011 at 11:58:49PM -0700, Garrett Cooper wrote:
>> On Tue, May 3, 2011 at 11:42 PM, Garrett Cooper <yanegomi_at_gmail.com> wrote:
>> > On Tue, May 3, 2011 at 10:59 PM, Kirk McKusick <mckusick_at_mckusick.com> wrote:
>> >>> Date: Tue, 3 May 2011 22:40:26 -0700
>> >>> Subject: Nasty non-recursive lockmgr panic on softdep only enabled UFS
>> >>>  partition when filesystem full
>> >>> From: Garrett Cooper <yanegomi_at_gmail.com>
>> >>> To: Jeff Roberson <jeff_at_freebsd.org>,
>> >>>         Marshall Kirk McKusick <mckusick_at_mckusick.com>
>> >>> Cc: FreeBSD Current <freebsd-current_at_freebsd.org>
>> >>>
>> >>> Hi Jeff and Dr. McKusick,
>> >>>     Ran into this panic when /usr ran out of space doing a make
>> >>> universe on amd64/r221219 (it took ~15 minutes for the panic to occur
>> >>> after the filesystem ran out of space -- wasn't quite sure what it was
>> >>> doing at the time):
>> >>>
>> >>> ...
>> >>>
>> >>>     Let me know what other commands you would like for me to run in kgdb.
>> >>> Thanks,
>> >>> -Garrett
>> >>
>> >> You did not indicate whether you are running an 8.X system or a 9-current
>> >> system. It would be helpful to know that.
>> >
>> > I've actually been running CURRENT for a few years now, but you're right --
>> > I didn't mention that part.
>> >
>> >> Jeff thinks that there may be a potential race in the locking code for
>> >> softdep_request_cleanup. If so, this patch for 9-current should fix it:
>> >>
>> >> Index: ffs_softdep.c
>> >> ===================================================================
>> >> --- ffs_softdep.c       (revision 221385)
>> >> +++ ffs_softdep.c       (working copy)
>> >> _at__at_ -11380,7 +11380,8 _at__at_
>> >>                                continue;
>> >>                        }
>> >>                        MNT_IUNLOCK(mp);
>> >> -                       if (vget(lvp, LK_EXCLUSIVE | LK_INTERLOCK, curthread)) {
>> >> +                       if (vget(lvp, LK_EXCLUSIVE | LK_NOWAIT | LK_INTERLOCK,
>> >> +                           curthread)) {
>> >>                                MNT_ILOCK(mp);
>> >>                                continue;
>> >>                        }
>> >>
>> >> If you are running an 8.X system, hopefully you will be able to apply it.
>> >
>> >    I've applied it, rebuilt and installed the kernel, and trying to
>> > repro the case again. Will let you know how things go!
>>
>>     Happened again with the change. It's really easy to repro:
>>
>> 1. Get a filesystem with UFS+SU
>> 2. Execute something that does a large number of small writes to a partition.
>> 3. 'dd if=/dev/zero of=FOO bs=10m' on the same partition
>>
>>     The kernel will panic with the issue I discussed above.
>> Thanks!
>
> Jeff' change is required to avoid LORs, but it is not sufficient to
> prevent recursion. We must skip the vnode supplied as a parameter to
> softdep_request_cleanup(). Theoretically, other vnodes might be also
> locked by curthread, thus I think the change below is needed. Try this.
>
> diff --git a/sys/ufs/ffs/ffs_softdep.c b/sys/ufs/ffs/ffs_softdep.c
> index a6d4441..25fa5d6 100644
> --- a/sys/ufs/ffs/ffs_softdep.c
> +++ b/sys/ufs/ffs/ffs_softdep.c
> _at__at_ -11380,7 +11380,9 _at__at_ retry:
>                                continue;
>                        }
>                        MNT_IUNLOCK(mp);
> -                       if (vget(lvp, LK_EXCLUSIVE | LK_INTERLOCK, curthread)) {
> +                       if (VOP_ISLOCKED(lvp) ||
> +                           vget(lvp, LK_EXCLUSIVE | LK_INTERLOCK | LK_NOWAIT,
> +                           curthread)) {
>                                MNT_ILOCK(mp);
>                                continue;
>                        }

    Ok. I'll let the make universe I have going run to completion, and
once I get back home later on, I'll take a look at repro'ing this
again with the above patch applied.
Thanks!
-Garrett