bug-gnulib
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: helgrind reported data race for test-lock is it a false positive ?


From: Eric Blake
Subject: Re: helgrind reported data race for test-lock is it a false positive ?
Date: Thu, 16 Jul 2015 06:56:56 -0600
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.0.1

On 07/16/2015 05:56 AM, Normand wrote:
> Hi there,
> I am looking at the test-lock failure reported by (1) while findutils
> building in opensuse for ppc64le architecture.
> Helgrind is reporting data race for the three *_checker_done volatiles
> as summarized in (2)
> 
> Is is a false positive or a real problem ?

Probably a real problem.  test-lock is notorious for causing slow
execution on multi-processor setups, to the point that several projects
explicitly disable it from running (at least libvirt does [1]); it may
be that the helgrind analysis here explains why the test is so painfully
slow on multiprocessor systems.

[1] libvirt.org/git/?p=libvirt.git;a=commitdiff;h=25fb4c6

But I'm not an expert enough in what the test was trying to do to easily
respond to this report, so maybe someone else can help chime in.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

Attachment: signature.asc
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]