Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754665Ab3H1Q5E (ORCPT ); Wed, 28 Aug 2013 12:57:04 -0400 Received: from hrndva-omtalb.mail.rr.com ([71.74.56.122]:18989 "EHLO hrndva-omtalb.mail.rr.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752677Ab3H1Q5C (ORCPT ); Wed, 28 Aug 2013 12:57:02 -0400 X-Authority-Analysis: v=2.0 cv=DqnUCRD+ c=1 sm=0 a=Sro2XwOs0tJUSHxCKfOySw==:17 a=Drc5e87SC40A:10 a=jxQzypR4wcoA:10 a=5SG0PmZfjMsA:10 a=kj9zAlcOel0A:10 a=meVymXHHAAAA:8 a=KGjhK52YXX0A:10 a=qWrlR9Lw5PkA:10 a=TiaNW0bfAAAA:8 a=NEAV23lmAAAA:8 a=dtPOJR9oX_xSaj3Fd34A:9 a=CjuIK1q_8ugA:10 a=B0KwX_7kBuMA:10 a=Sro2XwOs0tJUSHxCKfOySw==:117 X-Cloudmark-Score: 0 X-Authenticated-User: X-Originating-IP: 67.255.60.225 Date: Wed, 28 Aug 2013 12:57:01 -0400 From: Steven Rostedt To: Vince Weaver Cc: Dave Jones , Linux Kernel Subject: Re: FTRACE_WARN_ON((rec->flags & ~FTRACE_FL_MASK) == 0)) Message-ID: <20130828125701.1ca6ccd0@gandalf.local.home> In-Reply-To: References: <20130828034627.GA30596@redhat.com> <20130828103101.3f5679bc@gandalf.local.home> <20130828162321.GA14689@redhat.com> X-Mailer: Claws Mail 3.9.2 (GTK+ 2.24.20; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1706 Lines: 37 On Wed, 28 Aug 2013 12:50:36 -0400 (EDT) Vince Weaver wrote: > On Wed, 28 Aug 2013, Dave Jones wrote: > > Quite often just rerunning that last syscall that caused the oops/warn > > isn't sufficient to trigger an issue. (Though it may be for this specific > > bug that may not be the case..) > > > > Vince has a variant of trinity focussed just on perf which also has some > > neat replay/bisecting capabilities to narrow down test cases. > > I think I might need to add something like that at some point. > > Yes, you can get the perf_fuzzer code here: > git clone https://github.com/deater/perf_event_tests.git > in the fuzzer directory > > Bisecting down which perf_event_open() call causes problems is still very > labor intensive even with the other tools I've added to help. Even if the > bug is triggered right away (within the first 100,000 calls or so) it can > take hours to narrow things down to the two or three syscalls needed to > reproduce the problem. Especially if you need to reboot after triggering > the issue. Note, I have an idea that it's not even that many perf calls. I have a good idea what perf syscalls with what options are the issue. I think the issue is the calling order that is the problem. Dave, I'm assuming that trinidy does things as threads, such that it may be two threads calling perf with the same descriptor, and if we don't have the proper locks, things can get bad, right? -- Steve -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/