Date: Thu, 14 Sep 2006 03:19:47 +0200 From: Max Laier <max@love2party.net> To: Martin Blapp <mb@imp.ch> Cc: src-committers@freebsd.org, jhb@freebsd.org, cvs-src@freebsd.org, "Christian S.J. Peron" <csjp@freebsd.org>, cvs-all@freebsd.org, bde@freebsd.org, phk@freebsd.org Subject: Re: cvs commit: src/sys/kern kern_exit.c (DEVFS bug) ? Message-ID: <200609140319.57943.max@love2party.net> In-Reply-To: <20060913234613.S1494@godot.imp.ch> References: <200609131547.k8DFlrmv012940@repoman.freebsd.org> <20060913234613.S1494@godot.imp.ch>
next in thread | previous in thread | raw e-mail | index | archive | help
--nextPart15493826.0phNsZjQF4
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
On Wednesday 13 September 2006 23:54, Martin Blapp wrote:
> Hi all,
>
> The kernel I've made with 'mtx_assert(&Giant, MA_OWNED);' in all tty
> functions works fine. I'm not able to crash it yet, even under 24 hour
> load.
Strange. Probably the asserts screw up the timing just enough to "fix"=20
the problem. Could you try to narrow down the asserts to the places=20
where t_session (we still agree that's the problem?) is accessed?
> > But I have also seen what appears to be strange interactions or races
> > between devfs and the TTY code before the Giant push down here, which
> > was causing me some problems. After some discussions with John and
> > Bruce, it looks like the manipulation of t_session should be
> > protected by Giant.
>
> I Agree. This backout doesn't solve this panic here. I got this panic
> on FreeBSD 5.3 RELEASE and 5.4 RELEASE too. Exactly the same trace.
> And revision 1.272 never made it into RELENG_5.
>
> #1 0xc066355e in boot (howto=3D260) at
> /usr/src/sys/kern/kern_shutdown.c:409 #2 0xc06638b5 in panic
> (fmt=3D0xc0891732 "%s") at /usr/src/sys/kern/kern_shutdown.c:565 #3=20
> 0xc085c6b6 in trap_fatal (frame=3D0xed6e4ab8, eva=3D4) at
> /usr/src/sys/i386/i386/trap.c:836 #4 0xc085c3bf in trap_pfault
> (frame=3D0xed6e4ab8, usermode=3D0, eva=3D4) at
> /usr/src/sys/i386/i386/trap.c:744 #5 0xc085bfb5 in trap (frame=3D
> {tf_fs =3D 8, tf_es =3D 40, tf_ds =3D -1063714776, tf_edi =3D
> -1064042304, tf_esi =3D 0, tf_ebp =3D -311538944, tf_isp =3D -311538972,
> tf_ebx
> =3D -967615488, tf_edx =3D -1063651212, tf_ecx =3D -941099136, tf_eax =3D=
0,
> tf_trapno =3D 12, tf_err =3D 0, tf_eip =3D -1066845359, tf_cs =3D 32,
> tf_eflags =3D 66194, tf_esp =3D -967615488, tf_ss =3D 0})
> at /usr/src/sys/i386/i386/trap.c:434
> #6 0xc0848bea in calltrap () at /usr/src/sys/i386/i386/exception.s:139
> #7 0xc0693b51 in ttymodem (tp=3D0xc6535c00, flag=3D-1063651212) at
> /usr/src/sys/kern/tty.c:1659 #8 0xc0698362 in ptcclose (dev=3D0x0,
> flags=3D3, fmt=3D8192, td=3D0xc7e7f780) at linedisc.h:136 #9 0xc0638a6f =
in
> giant_close (dev=3D0xcb3c1100, fflag=3D3, devtype=3D8192, td=3D0xc7e7f780=
) at
> /usr/src/sys/kern/kern_conf.c:266 #10 0xc06162bf in devfs_close
> (ap=3D0xed6e4b7c) at /usr/src/sys/fs/devfs/devfs_vnops.c:287 #11
> 0xc086dc1c in VOP_CLOSE_APV (vop=3D0x0, a=3D0xc099f874) at vnode_if.c:426
> #12 0xc06c87e2 in vn_close (vp=3D0xc9cdf660, flags=3D3, file_cred=3D0x0,
> td=3D0xc7e7f780) at vnode_if.h:227 #13 0xc06c974a in vn_closefile
> (fp=3D0xc6fc5438, td=3D0xc7e7f780) at /usr/src/sys/kern/vfs_vnops.c:865 #=
14
> 0xc06162e7 in devfs_close_f (fp=3D0xc6fc5438, td=3D0xc7e7f780) at
> /usr/src/sys/fs/devfs/devfs_vnops.c:297 #15 0xc0642cdc in fdrop_locked
> (fp=3D0xc6fc5438, td=3D0xc7e7f780) at file.h:295 #16 0xc0642c29 in fdrop
> (fp=3D0xc6fc5438, td=3D0xc7e7f780) at /usr/src/sys/kern/kern_descrip.c:21=
22
> #17 0xc06411c7 in closef (fp=3D0xc6fc5438, td=3D0xc7e7f780) at
> /usr/src/sys/kern/kern_descrip.c:1942 #18 0xc063e329 in close
> (td=3D0xc7e7f780, uap=3D0x0) at /usr/src/sys/kern/kern_descrip.c:1007
>
> > Back out one of the Giant removals from revision 1.272. Giant was
> > not here to protect the vnode, it was present to synchronize access
> > to TTY session information between exit(2) and the TTY code. While we
> > are here, note that Giant is required for TTY protection.
> >
> > Clue from: bde
> > Discussed with: jhb
> > MFC after: 1 week
> >
> > Revision Changes Path
> > 1.290 +2 -2 src/sys/kern/kern_exit.c
=2D-=20
/"\ Best regards, | mlaier@freebsd.org
\ / Max Laier | ICQ #67774661
X http://pf4freebsd.love2party.net/ | mlaier@EFnet
/ \ ASCII Ribbon Campaign | Against HTML Mail and News
--nextPart15493826.0phNsZjQF4
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (FreeBSD)
iD8DBQBFCK49XyyEoT62BG0RAvskAJ967IWmbciDMAFm98kiRy1fT2mTYQCfWLUE
i0dPqU6k6BUhCaat11JzzeA=
=2OtX
-----END PGP SIGNATURE-----
--nextPart15493826.0phNsZjQF4--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200609140319.57943.max>