This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Initialize the entire obstack struct [BZ #17919]
- From: Mark Wielaard <mjw at redhat dot com>
- To: "Carlos O'Donell" <carlos at redhat dot com>
- Cc: Siddhesh Poyarekar <siddhesh at redhat dot com>, bug-gnulib at gnu dot org, libc-alpha at sourceware dot org
- Date: Wed, 04 Feb 2015 00:25:18 +0100
- Subject: Re: [PATCH] Initialize the entire obstack struct [BZ #17919]
- Authentication-results: sourceware.org; auth=none
- References: <20150203145649 dot GG1528 at spoyarek dot pnq dot redhat dot com> <54D0F799 dot 2030204 at redhat dot com> <20150203163831 dot GL1528 at spoyarek dot pnq dot redhat dot com> <54D0FC16 dot 7090009 at redhat dot com> <20150203170511 dot GM1528 at spoyarek dot pnq dot redhat dot com> <54D10FEE dot 4000307 at redhat dot com> <54D1101C dot 80902 at redhat dot com>
On Tue, 2015-02-03 at 13:14 -0500, Carlos O'Donell wrote:
> On 02/03/2015 01:14 PM, Carlos O'Donell wrote:
> > However, I see that GCC thinks this is a valgrind bug.
> >
> > If valgrind is simply looking at the comparison to make
> > the warning then it falls into the 'false positive' category.
> > In which case I think Valgrind should set up an exception for
> > this warning on s390.
>
> To be clear, I think nothing needs to be done now except file
> an upstream valgrind PR.
Yes please:
https://bugs.kde.org/enter_bug.cgi?product=valgrind&format=guided
This sounds a bit like https://bugs.kde.org/show_bug.cgi?id=308427 "s390
memcheck reports tsearch conditional jump or move depends on
uninitialized value". But that bug got fixed some time ago and should
not be in the latest version of valgrind. It does contain a nice
description of how best to report such an issue.
Just run under valgrind with --vgdb-error=0 and then attach with gdb,
through target remote | vgdb. Please include the disassembly output from
gdb around the code that produces the warning. And if possible use the
valgrind gdb monitor commands to figure out which bits valgrind thinks
are (un)defined.
http://valgrind.org/docs/manual/mc-manual.html#mc-manual.monitor-commands
Thanks,
Mark