Date: Mon, 30 May 2005 22:15:02 +0300 From: Ruslan Ermilov <ru@FreeBSD.org> To: Ed Maste <emaste@phaedrus.sandvine.ca> Cc: freebsd-current@FreeBSD.org, Gleb Smirnoff <glebius@FreeBSD.org> Subject: Re: SMP, DEVICE_POLLING and fxp LOR Message-ID: <20050530191502.GC57333@ip.net.ua> In-Reply-To: <20050530134443.GA9132@sandvine.com> References: <20050530134443.GA9132@sandvine.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--+nBD6E3TurpgldQp Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Mon, May 30, 2005 at 09:44:43AM -0400, Ed Maste wrote: > A few days ago I posted to -stable about a LOR observed on our 5.4-based > tree with SMP, DEVICE_POLLING and fxp: > http://lists.freebsd.org/pipermail/freebsd-stable/2005-May/015376.html > Contrary to my earlier assertion, em does not have a problem. >=20 > We do have a fair amount of local diffs, so I've now confirmed this is ea= sily > reproduced on unmodified -CURRENT as of 05/28. >=20 > My config file has > options SMP > options DEVICE_POLLING > device fxp > (full config file at http://www.maste.org/~ed/sv/STOCK.1 ) >=20 > lock order reversal > 1st 0xa0783860 Giant (Giant) @ /d2/emaste/cvs_head/src/sys/kern/kern_pol= l.c:332 > 2nd 0xa23e22a0 fxp0 (network driver) @ /d2/emaste/cvs_head/src/sys/dev/f= xp/if_fxp.c:1473 > KDB: stack backtrace: > kdb_backtrace(0,ffffffff,a07a6168,a07a4750,a0759160) at 0xa0585789 =3D kd= b_backtrace+0x29 > witness_checkorder(a23e22a0,9,a070a173,5c1) at 0xa058f74c =3D witness_che= ckorder+0x55c > _mtx_lock_flags(a23e22a0,0,a070a173,5c1) at 0xa05652f8 =3D _mtx_lock_flag= s+0x60 > fxp_poll(a23e2000,1,5) at 0xa04fa001 =3D fxp_poll+0x29 > netisr_poll(0) at 0xa0567070 =3D netisr_poll+0x188 > swi_net(0) at 0xa05d9041 =3D swi_net+0x81 > ithread_loop(a2309000,c5d8fd38,a2309000,a055aba8,0) at 0xa055acc8 =3D ith= read_loop+0x120 > fork_exit(a055aba8,a2309000,c5d8fd38) at 0xa055a0f0 =3D fork_exit+0xa0 > fork_trampoline() at 0xa06ad31c =3D fork_trampoline+0x8 >=20 > The ether_poll_deregister in fxp_stop() looks to be the culprit. >=20 This LOR is well-known and will go away as soon as we commit our WIP changes for polling(4). I have no firm plans when it happens though. Cheers, --=20 Ruslan Ermilov ru@FreeBSD.org FreeBSD committer --+nBD6E3TurpgldQp Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (FreeBSD) iD8DBQFCm2Y1qRfpzJluFF4RAuk1AJwLTp+tQ01RY0qcFmfsiFESQ9sMLQCfSslr sRLImZP1PwMQVQO2lVOyFow= =Xb4s -----END PGP SIGNATURE----- --+nBD6E3TurpgldQp--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050530191502.GC57333>