From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753411Ab0EKPhq (ORCPT ); Tue, 11 May 2010 11:37:46 -0400 Received: from s15228384.onlinehome-server.info ([87.106.30.177]:54543 "EHLO mail.x86-64.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751261Ab0EKPhm (ORCPT ); Tue, 11 May 2010 11:37:42 -0400 Date: Tue, 11 May 2010 17:37:57 +0200 From: Borislav Petkov To: Peter Zijlstra Cc: Ingo Molnar , Lin Ming , Frederic Weisbecker , "eranian@gmail.com" , "Gary.Mohr@Bull.com" , Corey Ashford , "arjan@linux.intel.com" , "Zhang, Yanmin" , Paul Mackerras , "David S. Miller" , Russell King , Paul Mundt , lkml , Arnaldo Carvalho de Melo Subject: Re: [RFC][PATCH 3/9] perf: export registerred pmus via sysfs Message-ID: <20100511153757.GA14408@aftab> References: <1273483623.15998.57.camel@minggr.sh.intel.com> <1273484401.5605.3333.camel@twins> <1273486313.15998.76.camel@minggr.sh.intel.com> <1273486708.5605.3342.camel@twins> <1273487195.15998.85.camel@minggr.sh.intel.com> <1273490824.5605.3379.camel@twins> <20100510114311.GA6449@elte.hu> <20100511141511.GA14034@aftab> <1273587915.1810.1.camel@laptop> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1273587915.1810.1.camel@laptop> Organization: Advanced Micro Devices =?iso-8859-1?Q?GmbH?= =?iso-8859-1?Q?=2C_Einsteinring_24=2C_85609_Dornach_bei_M=FCnchen=2C_Gesc?= =?iso-8859-1?Q?h=E4ftsf=FChrer=3A_Thomas_M=2E_McCoy=2C_Giuliano_Meroni=2C?= =?iso-8859-1?Q?_Andrew_Bowd=2C_Sitz=3A_Dornach=2C_Gemeinde_Aschheim=2C_La?= =?iso-8859-1?Q?ndkreis_M=FCnchen=2C_Registergericht_M=FCnchen?= =?iso-8859-1?Q?=2C?= HRB Nr. 43632 User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Peter Zijlstra Date: Tue, May 11, 2010 at 10:25:15AM -0400 > On Tue, 2010-05-11 at 16:15 +0200, Borislav Petkov wrote: > > Another point I have is that MCEs don't need pmus so we should consider > > having the ability to decouple events from pmus. > > Strictly speaking tracepoints are software events, which run off of a > software 'pmu'. So no, we can't decouple, they need a 'pmu' context. We could make this configurable depending on the severity of the error. I'm guessing for further event handling through the perf infrastructure we cannot run without a sw pmu context but on critical conditions we need to run as fast and as sparingly as possible so I'm thinking maybe adding some specially tailored callbacks to the MCE tracepoint trace_mce_record, as Steven suggested. > > What you basically want to have is a tracepoint which is "persistent," > > as Ingo suggested earlier, and it buffers MCEs occurring at any time > > into a ring buffer until a userspace daemon or similar sucks that data > > out for processing (critical stuff is handled differently, of course). > > And this should work on any x86 hw supporting MCA without hw perf > > monitoring features. > > Try building x86 without perf hw support :/ I dunno, maybe decoupling wouldn't be necessary per se but I was simply pointing out that MCEs shouldn't necessarily depend on the presence on hardware performance counters. -- Regards/Gruss, Boris. -- Advanced Micro Devices, Inc. Operating Systems Research Center