[Libvir] Core dump while executing virsh in RHEL5 .

[Daniel Veillard]

On Mon, Oct 15, 2007 at 01:10:45PM +0100, Richard W.M. Jones wrote:
> Did this problem get fixed while I was away?
> 
> What I'm seeing in Veerendra's valgrind log are the following suspicious 
> messages, although the line numbers don't correspond to the earlier line 
> numbers from gdb:

  yeah, I looked at it but could not find the right mapping

> ==17508== Invalid read of size 8
> ==17508==    at 0x4C5D0BB: doRemoteOpen (remote_internal.c:323)
> ==17508==    by 0x4C5EABB: remoteNetworkOpen (remote_internal.c:2392)
> ==17508==    by 0x4C395DA: do_open (libvirt.c:447)
> ==17508==    by 0x40A80D: main (virsh.c:4507)
> 
>     if (uri->user) {
>         username = strdup (uri->user);		<--- line 323
>         if (!username) goto out_of_memory;
>     }
> 
> 
> ==17508== Invalid write of size 8
> ==17508==    at 0x4C5D455: doRemoteOpen (remote_internal.c:761)
> ==17508==    by 0x4C5EABB: remoteNetworkOpen (remote_internal.c:2392)
> ==17508==    by 0x4C395DA: do_open (libvirt.c:447)
> ==17508==    by 0x40A80D: main (virsh.c:4507)
> 
>     if (query_out) *query_out = NULL;		<-- line 761
> 
> As I understand the valgrind message, these indicate that the memory 
> being read/written is not valid (ie. outside any allocated malloc block 
> or static memory), although I don't understand how those lines could 
> generate that error.

  yesh I could not understand either, I was afraid the compilation with
optimization would lead to the skew of lines, and only a valgrind of 
a debug version possibly recompiled locally would be really okay to
check this out. And then the issue got off my radar :-\

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
veillard at redhat.com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/