Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive

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

segmantation fault error

Kristian Hermansen wrote:
> On 7/9/07, Eugene Gorelik <eugene.gorelik-Re5JQEeQqe8AvxtiuMwx3w at> wrote:
>> All memcheck test are passed.
> Right, so it probably isn't a memory error.
>> Yes it is not repeatable, when segfault happen it always points to
>> 0000000000000003 rip 0000003a9a970810.
>> Ran gdb and analyzed core files and again all segfaults point to the same
>> memory address, seems like 100% physical memory issue.
> So it is a bug in this once program.  But, I you should check out the
> stack trace in gdb with the 'bt' command.  It could very well be a zip
> bug, as I have reported many in zip, some on amd64.  

I find it hard to believe that a bad zip binary (even a malicious one) could
cause a kernel crash, unless it happens to be triggering a kernel bug.  But
then it would affect all his machines, and he said he had 29 good ones...

> If you are getting lots of segfaults in other programs, it could be
> possible that your libc has been corrupted. I would try something like
> this. Reinstall all the installed applications on the system. In 
> Debian/Ubuntu's APT, I would do this...
> $ sudo aptitude reinstall ~i~Afeisty
> There must be a similar command for yum, right?  Where are the yum
> experts?  That will ensure that you have replaced all the binaries on
> your disk with presumed good builds.  The above command says
> "reinstall all packages currently marked as installed in the feisty
> archive"...

Before you re-install anything, you can see if your binaries are modified by
verifying the md5sum of your binaries against the version from the installed
RPM with 'rpm -V zip' (if you see a '5' in the summary for a file, then it's
md5sum is different; there is no output at all if everything matches the RPM
db).  (note that if you do this for glibc, you'll need to be root to check one
of the files, you'll know you have permission issues if you see a '?')

Does it only happen with disk access?  If memory checks out, it might be a
faulty ide controller, or something else motherboard-related.


This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!

Boston Linux & Unix /