Return-path: Received: from xc.sipsolutions.net ([83.246.72.84]:59926 "EHLO sipsolutions.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753120AbYLIPDl (ORCPT ); Tue, 9 Dec 2008 10:03:41 -0500 Subject: Re: [PATCH v2] p54usb: rewriting rx/tx routines to make use of usb_anchor's facilities From: Johannes Berg To: "John W. Linville" Cc: Christian Lamparter , linux-wireless@vger.kernel.org, Larry Finger In-Reply-To: <20081209145439.GB3653@tuxdriver.com> References: <200812091514.37634.chunkeey@web.de> <20081209143530.GA3653@tuxdriver.com> <1228833999.1029.0.camel@johannes.berg> <20081209145439.GB3653@tuxdriver.com> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-Tr7sJdldxvcZCTbQ7ZdS" Date: Tue, 09 Dec 2008 16:03:05 +0100 Message-Id: <1228834985.1029.2.camel@johannes.berg> (sfid-20081209_160344_533591_2BEC647B) Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: --=-Tr7sJdldxvcZCTbQ7ZdS Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Tue, 2008-12-09 at 09:54 -0500, John W. Linville wrote: > On Tue, Dec 09, 2008 at 03:46:39PM +0100, Johannes Berg wrote: > > On Tue, 2008-12-09 at 09:35 -0500, John W. Linville wrote: > > > On Tue, Dec 09, 2008 at 03:14:37PM +0100, Christian Lamparter wrote: > > > > Alan Stern found several flaws in p54usb's implementation and annot= ated:=20 > > > > "usb_kill_urb() and similar routines do not expect an URB's complet= ion > > > > routine to deallocate it. This is almost obvious -- if the URB is = deallocated > > > > before the completion routine returns then there's no way for usb_k= ill_urb > > > > to detect when the URB actually is complete." > > > >=20 > > > > This patch addresses all known limitations in the old implementatio= n and fixes > > > > khub's "use-after-freed" hang, when SLUB debug's poisoning option i= s enabled. > > > >=20 > > > > Signed-off-by: Christian Lamparter > > > > Cc: stable@kernel.org > > >=20 > > > That's a big patch for so late in the cycle. Is there a bugzilla ope= n for it? > >=20 > > No, but we can open one ;) It's strange, but nobody but me apparently > > found the bug so far, even though it has been in there forever. Might b= e > > due to slub debugging being disabled by default even if you enable > > SLUB_DEBUG in Kconfig (you also need to enable SLUB_DEBUG_ON) >=20 > It might be helpful to have open a bug that includes a backtrace or somet= hing. There's no backtrace in that sense, it just hangs. I dumped it with sysrq-w but that wasn't too helpful... johannes --=-Tr7sJdldxvcZCTbQ7ZdS Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Comment: Johannes Berg (powerbook) iQIcBAABAgAGBQJJPoilAAoJEKVg1VMiehFYZ84QAKxNuepF7NXj2t+YHD6WauWI gdPXJUZKK9/DptegCqT4pbKPWxOGnWSaPyBiV7y8qa383zIpkSLFE7lie8v4oYc1 wRFZ/h0RQobiCXczHaeDYqTqF6BfMwAhQz5mU8bPrJKXGJu8fj/IxWZ2+dOveWc8 KPadKSyDI9fh/4P1tZOy+HVW21NotPJWFTrIJyYUPoyVVJm8cgkp0CVWy54ZeVde ir1K1HyklKWxAfCTxMxb40ZuLW80aEj3JtpHfZCUWNzV1ev1Lo73hgbTS+jGhqsD lzYwdK1Lxgp/AOqGrPxFCxZusy06ArS+fyklItfeBqv2rA59QlYdD5XD+8fZ6gzk 7PxWMhbOCg4p8PY8VICQWxQctLNnZLAI1yqmNDQ83WZzzFdcs9e+nfumlHLp+bXJ Mf84+vDoGub/mN1IlIlKMR6VoghMhNKS018OIZXi8OPNaUBR7Bot9tCDu3lc+KfZ 6oEssXwzNdXbI4D2Pf29yX00UVh4NpwG27PW+SnhjUvZwPd5FtaFMRp8x6/PNhgc PJ+k5pd4hLGJNAL6Owcx+A74CPU97nipASYznGS5/4joOmZoXMdYeRO70jiKeyuz lE8mQb6l9Pzk3hNKVoe7yjZS5CC0TUJ9och/5g/wmO8BKimrPGPC0Yew4/61YRkA At2YUZ1LXre5wwM4zMhn =U0MA -----END PGP SIGNATURE----- --=-Tr7sJdldxvcZCTbQ7ZdS--