Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 25 Sep 2009 23:03:06 +0200
From:      Pawel Jakub Dawidek <pjd@FreeBSD.org>
To:        Adam McDougall <mcdouga9@egr.msu.edu>
Cc:        svn-src-all@freebsd.org
Subject:   Re: svn commit: r196300 - in stable/8/sys: . amd64/include/xen	cddl/contrib/opensolaris cddl/contrib/opensolaris/uts/common/fs/zfs contrib/dev/acpica	contrib/pf dev/ata dev/cxgb dev/mfi dev/xen/netfron...
Message-ID:  <20090925210305.GB1597@garage.freebsd.pl>
In-Reply-To: <4ABD247D.40201@egr.msu.edu>
References:  <200908170923.n7H9NR7t056904@svn.freebsd.org> <4ABD247D.40201@egr.msu.edu>

next in thread | previous in thread | raw e-mail | index | archive | help

--2B/JsCI69OhZNC5r
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Sep 25, 2009 at 04:13:49PM -0400, Adam McDougall wrote:
> Pawel Jakub Dawidek wrote:
> >Author: pjd
> >Date: Mon Aug 17 09:23:27 2009
> >New Revision: 196300
> >URL: http://svn.freebsd.org/changeset/base/196300
> >
> >Log:
> >  MFC r196299:
> > =20
> >  - We need to recycle vnode instead of freeing znode.
> > =20
> >  Submitted by:	avg
> > =20
> >  - Add missing vnode interlock unlock.
> >  - Remove redundant znode locking.
> > =20
> >  Approved by:	re (kib)
> >
>=20
> Two quick questions:
>=20
> - Do you think this fix would improve or fix the panic below?
>   (transcribed by hand since it did not dump and I only have a
>    screenshot)
> - Are the many misc ZFS fixes that went into 9.0 and 8.0 going to
>   be MFC'ed to 7-stable sometime, including the one above? I have
>   lucked out pretty well so far since May.

At this point I'm not going to MFC them into stable/7, although marck@
seems to be interested in doing it.

> Sleeping thread (tid 100161), pid 919) owns a non-sleepable lock
> sched_switch() at sched_switch+0x1fe
> mi_switch() at mi_switch+0x18e
> sleepq_wait() at sleepq_wait+0x3b
> _sx_xlock_hard() at _sx_xlock_hard+0x1ee
> _sx_xlock() at _sx_xlock+0x4e
> rrw_enter() at rrw_enter+0xa4
> zfs_freebsd_getattr() at zfs_freebsd_getattr+0x9d
> filt_vfsread() at filt_vfsread+_0x5e
> knote() at knote+0xc2

It looks like kqueue is calling into ZFS with mutex held, which is
bad... No, this bug is not fixed.

--=20
Pawel Jakub Dawidek                       http://www.wheel.pl
pjd@FreeBSD.org                           http://www.FreeBSD.org
FreeBSD committer                         Am I Evil? Yes, I Am!

--2B/JsCI69OhZNC5r
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.4 (FreeBSD)

iD8DBQFKvTAJForvXbEpPzQRAn6sAKD0O14wAuQRFaIsIByTJ323Y9dLhACgtAlU
20C/i8lor84jqMQ3uSWQDUE=
=wTwg
-----END PGP SIGNATURE-----

--2B/JsCI69OhZNC5r--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090925210305.GB1597>