Date: Tue, 9 Jun 2009 19:30:05 +0300 From: Kostik Belousov <kostikbel@gmail.com> To: John Baldwin <jhb@freebsd.org> Cc: Yuri Pankov <yuri.pankov@gmail.com>, freebsd-current@freebsd.org, Paul Saab <ps@freebsd.org> Subject: Re: panic: knlist not locked, but should be Message-ID: <20090609163005.GD75569@deviant.kiev.zoral.com.ua> In-Reply-To: <200906091059.15278.jhb@freebsd.org> References: <20090609110540.GA1290@darklight.homeunix.org> <200906090937.37562.jhb@freebsd.org> <20090609135145.GB1290@darklight.homeunix.org> <200906091059.15278.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--pQhZXvAqiZgbeUkD Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Tue, Jun 09, 2009 at 10:59:14AM -0400, John Baldwin wrote: > On Tuesday 09 June 2009 9:51:45 am Yuri Pankov wrote: > > On Tue, Jun 09, 2009 at 09:37:37AM -0400, John Baldwin wrote: > > > On Tuesday 09 June 2009 7:05:40 am Yuri Pankov wrote: > > > > Hi, > > > >=20 > > > > I'm getting the following panic when running Firefox+nspluginwrapper > > > > (which crashes a lot, but paniced the box only two times, if it's e= ven > > > > related). -CURRENT/amd64 20080607. > > > >=20 > > > > Unread portion of the kernel message buffer: > > > > panic: knlist not locked, but should be > > > > cpuid =3D 0 > > > > KDB: stack backtrace: > > > > db_trace_self_wrapper() at db_trace_self_wrapper+0x2a > > > > panic() at panic+0x182 > > > > knote() at knote+0xea > > > > VOP_WRITE_APV() at VOP_WRITE_APV+0x152 > > > > vn_rdwr() at vn_rdwr+0x21d > > > > vn_rdwr_inchunks() at vn_rdwr_inchunks+0xc2 > > > > elf32_coredump() at elf32_coredump+0x109 > > > > sigexit() at sigexit+0x810 > > > > postsig() at postsig+0x32f > > > > ast() at ast+0x3ac > > > > doreti_ast() at doreti_ast+0x1f > > >=20 > > > Are you using ZFS? > > >=20 > > > --=20 > > > John Baldwin > >=20 > > Yes, sorry for not mentioning it. >=20 > This appears to be an interaction with the recent changes to use shared v= node=20 > locks for writes on ZFS. Hmm, I think it may be ok to use a shared vnode= =20 > lock for kevents on vnodes though. The vnode interlock should be suffici= ent=20 > locking for what little work the kevent filters do. As a quick hack for = now=20 > the MNT_SHARED_WRITES() stuff could avoid using shared locks 'if=20 > (!VN_KNLIST_EMPTY(vp))', but I think the longer term fix is to not use th= e=20 > vnode locks for vnode kevents, but use the interlock instead. filt_vfsread() needs at least shared lock on the vnode since it uses VOP_GETATTR. --pQhZXvAqiZgbeUkD Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (FreeBSD) iEYEARECAAYFAkoujg0ACgkQC3+MBN1Mb4igZQCfW4LuuyTrvH2V4r4rT8vI8qry k4QAn3iCa+IltJ0u/KMwRzhU5kEXMWee =Fq1D -----END PGP SIGNATURE----- --pQhZXvAqiZgbeUkD--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090609163005.GD75569>