linux-kernel - Re: perf: bisected sampling bug in Linux 4.11-rc1

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <alpine.DEB.2.20.1707152259180.30679@macbook-air>
Date:   Sat, 15 Jul 2017 23:03:06 -0400 (EDT)
From:   Vince Weaver <vincent.weaver@...ne.edu>
To:     Ingo Molnar <mingo@...nel.org>
cc:     linux-kernel@...r.kernel.org,
        Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Stephane Eranian <eranian@...il.com>,
        Jiri Olsa <jolsa@...hat.com>,
        Arnaldo Carvalho de Melo <acme@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: perf: bisected sampling bug in Linux 4.11-rc1

On Sat, 15 Jul 2017, Vince Weaver wrote:

> Although there is a separate issue also introduced in 4.11-rc1 that still 
> fails a different testcase.  I'm in the middle of bisecting that one and 
> probably won't have the result of the bisect until Monday.

I went and bisected the other issue anyway.  It wasn't in 4.11-rc1, but in 
4.12-rc4.  Yes, I have a test that triggered the
	c1582c231ea041 perf/core: Drop kernel samples even though :u is specified
issue.

The test was actually losing 7% of its samples, so not just a single 
sample here or there.  The test had two events sampling, one at 100k and 
one at 200k so I guess every other sample would have two samples 
immediately back-to-back which must make it more likely to stray into the 
kernel.

Vince