From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751310Ab3JaIPO (ORCPT ); Thu, 31 Oct 2013 04:15:14 -0400 Received: from mail-ee0-f51.google.com ([74.125.83.51]:45286 "EHLO mail-ee0-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753150Ab3JaIPL (ORCPT ); Thu, 31 Oct 2013 04:15:11 -0400 Date: Thu, 31 Oct 2013 09:15:07 +0100 From: Ingo Molnar To: "Michael S. Tsirkin" Cc: Peter Zijlstra , Paul Mackerras , Ingo Molnar , Arnaldo Carvalho de Melo , Thomas Gleixner , "H. Peter Anvin" , x86@kernel.org, Stephane Eranian , Andi Kleen , linux-kernel@vger.kernel.org Subject: Re: [PATCH] x86: print microcode revision on PEBs errors Message-ID: <20131031081507.GA8922@gmail.com> References: <20131030193150.GA7765@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20131030193150.GA7765@redhat.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org * Michael S. Tsirkin wrote: > On Sandy bridge CPUs with old microcode, PEBs > fails and suggests a microcode update. > Print out the required and the actual revision to > make it easier to figure out what's wrong. > > Signed-off-by: Michael S. Tsirkin > --- > arch/x86/kernel/cpu/perf_event_intel.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c > index f31a165..81f2789 100644 > --- a/arch/x86/kernel/cpu/perf_event_intel.c > +++ b/arch/x86/kernel/cpu/perf_event_intel.c > @@ -2105,6 +2105,8 @@ static int intel_snb_pebs_broken(int cpu) > } > } > > + pr_info("PEBS checking: microcode 0x%x min legal 0x%x\n", > + cpu_data(cpu).microcode, rev); That's not a very informative message though. Something like: "x86/perf/intel: PEBS turned off due to too old microcode version 0x%x\n" "x86/perf/intel: Minimum required microcode version for PEBS: 0x%x\n" would work better I think. Also, only output this once, and only output it if the check _fails_. > return (cpu_data(cpu).microcode < rev); Btw., please also fix the return statement, it doesn't need parentheses. Thanks, Ingo