Bug 7496

Summary:	valgrind has to be patched
Product:	Mageia	Reporter:	François PELLEGRINI <pelegrin>
Component:	RPM Packages	Assignee:	Shlomi Fish <shlomif>
Status:	RESOLVED INVALID	QA Contact:
Severity:	normal
Priority:	Normal	Keywords:	PATCH
Version:	2
Target Milestone:	---
Hardware:	i586
OS:	Linux
Whiteboard:
Source RPM:	valgrind	CVE:
Status comment:

Description François PELLEGRINI 2012-09-16 19:48:00 CEST

Hello all,
Valgrind is no longer functioning in Mageia 2.
The error I get when running it is :

valgrind: m_debuginfo/readdwarf.c:2152 (copy_convert_CfiExpr_tree): Assertion 'srcix >= 0 && srcix < VG_(sizeXA)(srcxa)' failed.

This has already be diagnosed in other distros, e.g. :
https://lists.ubuntu.com/archives/foundations-bugs/2011-August/016879.html

Revelant patches have to be applied so that it can function again, before Mageia 3 is available with a new version of Valgrind.

Thx !

f.p.

Manuel Hiebel 2012-09-23 21:11:30 CEST

Keywords: (none) => PATCH
Assignee: bugsquad => shlomif
Source RPM: (none) => valgrind

Comment 1 Shlomi Fish 2012-09-23 21:43:30 CEST

(In reply to comment #0)
> Hello all,
> Valgrind is no longer functioning in Mageia 2.
> The error I get when running it is :
> 
> valgrind: m_debuginfo/readdwarf.c:2152 (copy_convert_CfiExpr_tree): Assertion
> 'srcix >= 0 && srcix < VG_(sizeXA)(srcxa)' failed.
> 
> This has already be diagnosed in other distros, e.g. :
> https://lists.ubuntu.com/archives/foundations-bugs/2011-August/016879.html
> 
> Revelant patches have to be applied so that it can function again, before
> Mageia 3 is available with a new version of Valgrind.
> 
> Thx !
> 
> f.p.

Hi, does it happen only on i586? Because valgrind seems to work fine here on x86-64 Mageia 2.

Regards,

-- Shlomi Fish

Comment 2 François PELLEGRINI 2012-09-23 22:34:18 CEST

I'm sorry I can't tell : I only have a 586 architectures around.  :-(

Thx,

f.p.

Comment 3 Shlomi Fish 2012-09-23 23:25:24 CEST

Hi,

(In reply to comment #2)
> I'm sorry I can't tell : I only have a 586 architectures around.  :-(
> 

OK, I'll try in a VM.

Regards,

-- Shlomi Fish

> Thx,
> 
> f.p.

Comment 4 Shlomi Fish 2012-09-24 00:31:53 CEST

(In reply to comment #3)
> Hi,
> 
> (In reply to comment #2)
> > I'm sorry I can't tell : I only have a 586 architectures around.  :-(
> > 
> 
> OK, I'll try in a VM.
> 
> Regards,
> 
> -- Shlomi Fish
> 
> > Thx,
> > 
> > f.p.

OK, tested in a VirtualBox i586 Mageia 2 VM (with all updated from updates_testing) and valgrind appears to be working fine. So I cannot reproduce this problem.

Regards,

-- Shlomi Fish

Comment 5 François PELLEGRINI 2012-09-24 09:39:43 CEST

Ok. Thx for trying. Let's check.
My versions are :

valgrind-3.7.0-1.mga2
gcc-4.6.3-2.mga2
gcc-cpp-4.6.3-2.mga2
libgcc1-4.6.3-2.mga2

What are yours ?

Regards,


f.p.

Comment 6 Shlomi Fish 2012-09-24 12:26:51 CEST

(In reply to comment #5)
> Ok. Thx for trying. Let's check.
> My versions are :
> 
> valgrind-3.7.0-1.mga2
> gcc-4.6.3-2.mga2
> gcc-cpp-4.6.3-2.mga2
> libgcc1-4.6.3-2.mga2
> 
> What are yours ?
> 

I have exactly the same versions in my VM. :-( .

BTW, does the bug happen when valgrind detects a memory error?

Regards,

-- Shlomi Fish

> Regards,
> 
> 
> f.p.

Comment 7 François PELLEGRINI 2012-09-24 13:23:02 CEST

Argh ! I should get back to basics before bothering people first. :-(
A "which" made me realize that I was pointing to an older revision of valgrind in /usr/local/bin .
*Really* sorry for taking your time with such a silly thing.
Thanks again.
f.p.

Status: NEW => RESOLVED
Resolution: (none) => INVALID