From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759300AbZANCCe (ORCPT ); Tue, 13 Jan 2009 21:02:34 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754682AbZANCCZ (ORCPT ); Tue, 13 Jan 2009 21:02:25 -0500 Received: from mga02.intel.com ([134.134.136.20]:6010 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753726AbZANCCZ (ORCPT ); Tue, 13 Jan 2009 21:02:25 -0500 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.37,261,1231142400"; d="asc'?scan'208";a="378238209" Subject: Re: x86/mce merge, integration hickup + crash, design thoughts From: Huang Ying To: Ingo Molnar Cc: Andi Kleen , Thomas Gleixner , "linux-kernel@vger.kernel.org" , "H. Peter Anvin" In-Reply-To: <20090113174522.GA26965@elte.hu> References: <20081227155019.GA15493@elte.hu> <20081227225102.GA17822@elte.hu> <49594411.60000@linux.intel.com> <20090113174522.GA26965@elte.hu> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-rKYHO/CVuNE869qZ0+Y5" Date: Wed, 14 Jan 2009 10:02:21 +0800 Message-Id: <1231898541.5937.143.camel@yhuang-dev.sh.intel.com> Mime-Version: 1.0 X-Mailer: Evolution 2.22.3.1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org --=-rKYHO/CVuNE869qZ0+Y5 Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Wed, 2009-01-14 at 01:45 +0800, Ingo Molnar wrote: > * Andi Kleen wrote: >=20 > > Ingo Molnar wrote: > > > >>>> A far more useful design for handling MCE events would be to feed=20 > >>>> them into printk logging. > >>> If there's ASCII logging it should be separate from normal printk. > >> > >> Well, why?=20 > > > > Mostly because the problem is not a kernel issue. Especially large=20 > > systems with a lot of memory can generate a lot of corrected events (on= e=20 > > bit flips in DIMMs are not that uncommon) and it's not good to mix that= =20 > > all up into other kernel messages. It also makes it more clear that it'= s=20 > > not a kernel problem, but a hardware problem. I've got feedback over th= e=20 > > years that confirm this sight. >=20 > Is your argument that syslog is not suitable for the logging of hw events= ? >=20 > If that is your argument then the answer is to extend syslog with those=20 > aspects, instead of widening the quirky /dev based mce ABIs to achieve=20 > something similar. In current /dev based mce ABI implementation, syslog is used for logging hw events, not though printk, but through /dev/mcelog and /sbin/mcelog. For uncorrected MCE, they should be logged via printk. But for corrected MCE, there could be thousands/millions ones (imagining you have a DIMM with one data pin corrupted). I don't think it's a good idea to blend these hardware events with other kernel software events in printk. Best Regards, Huang Ying --=-rKYHO/CVuNE869qZ0+Y5 Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEABECAAYFAkltR6oACgkQKhFGF+eHlphXDgCfWxRpfYcAiGiZi4f24hWRpU4Q eqMAnRVeQch8CyHAgNoTVm2LnkSXwws2 =AlIH -----END PGP SIGNATURE----- --=-rKYHO/CVuNE869qZ0+Y5--