Exim kernel segfaults - Error 4

[Marc Perkel]

Does anyone know what would cause Exim to segfault with Error 4? What is
error 4?

Mar  6 14:02:41 pascal kernel: exim[26817]: segfault at 0000000000000000
rip 0000003254278eb0 rsp 00007fff0dd7f178 error 4
Mar  6 14:02:53 pascal kernel: exim[18416]: segfault at 0000000000000000
rip 0000003254278eb0 rsp 00007fff0dd7f178 error 4
Mar  6 14:02:56 pascal kernel: exim[23424]: segfault at 0000000000000000
rip 0000003254278eb0 rsp 00007fff0dd7f178 error 4

I'm getting a lot of these.

I'm running 64 bit fedora 8 with latest stock Fedora kernel and the 4.68
version of Exim.

Thanks in advance.



--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

On Thu, 2008-03-06 at 14:06 -0800, Marc Perkel wrote:
> Does anyone know what would cause Exim to segfault with Error 4? What is
> error 4?

/usr/include/asm-generic/errno-base.h

Graeme


--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[Marc Perkel]

Let me see if I can ask the question more precisely. Is there something
wrong with Exim or is it something else. No other applications on my
servers are generating segfaults. So I'm trying to figure out what's
broken. If it's Exim then I'll fill out a bug report. If it's something
with my servers then I'll fix that.

Since it's onlt Exim that's segfaulting it makes me think there's an
issue with Exim.

Is anyone else seeing these problems?

Marc Perkel wrote:

> Does anyone know what would cause Exim to segfault with Error 4? What is
> error 4?
>
> Mar  6 14:02:41 pascal kernel: exim[26817]: segfault at 0000000000000000
> rip 0000003254278eb0 rsp 00007fff0dd7f178 error 4
> Mar  6 14:02:53 pascal kernel: exim[18416]: segfault at 0000000000000000
> rip 0000003254278eb0 rsp 00007fff0dd7f178 error 4
> Mar  6 14:02:56 pascal kernel: exim[23424]: segfault at 0000000000000000
> rip 0000003254278eb0 rsp 00007fff0dd7f178 error 4
>
> I'm getting a lot of these.
>
> I'm running 64 bit fedora 8 with latest stock Fedora kernel and the 4.68
> version of Exim.
>
> Thanks in advance.
>
>
>
>  

--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[Renaud Allard]

Hello,

Marc Perkel wrote:
> Let me see if I can ask the question more precisely. Is there something
> wrong with Exim or is it something else. No other applications on my
> servers are generating segfaults. So I'm trying to figure out what's
> broken. If it's Exim then I'll fill out a bug report. If it's something
> with my servers then I'll fix that.
>
> Since it's onlt Exim that's segfaulting it makes me think there's an
> issue with Exim.
>

Try disabling SELINUX to find out if it is not something related to
permissions in default SELINUX configuration.


--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[Martin A. Brooks-3]

Marc Perkel wrote:
> I'm running 64 bit fedora 8 with latest stock Fedora kernel and the 4.68
> version of Exim.
>  

You don't mention if this is exim from a package or handrolled.  
Whenever I find a redhat-like box doing weird things, I first try
disabling selinux.


--
 Martin A. Brooks |  http://www.antibodymx.net/ | Anti-spam & anti-virus
    Consultant    |  martin@...      | filtering. Inoculate
  antibodymx.net  |  m: [phone]           | your mail system.


--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[Marc Perkel]

Renaud Allard wrote:

> Hello,
>
> Marc Perkel wrote:
>  
>> Let me see if I can ask the question more precisely. Is there something
>> wrong with Exim or is it something else. No other applications on my
>> servers are generating segfaults. So I'm trying to figure out what's
>> broken. If it's Exim then I'll fill out a bug report. If it's something
>> with my servers then I'll fix that.
>>
>> Since it's onlt Exim that's segfaulting it makes me think there's an
>> issue with Exim.
>>
>>    
>
> Try disabling SELINUX to find out if it is not something related to
> permissions in default SELINUX configuration.
>  

Thanks but it's already disabled.

Interestingly enough my servers running under OpenZV don't have segfaults.

--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[Marc Perkel]

Martin A. Brooks wrote:

> Marc Perkel wrote:
>> I'm running 64 bit fedora 8 with latest stock Fedora kernel and the
>> 4.68 version of Exim.
>>  
>
> You don't mention if this is exim from a package or handrolled.  
> Whenever I find a redhat-like box doing weird things, I first try
> disabling selinux.
>
>

Actually - it is hand rolled. The only significant change I can think of
is adding the domain keys stuff. Could that be it? And - thanks for your
help.



--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[Martin A. Brooks-3]

Marc Perkel wrote:
> Actually - it is hand rolled.

Alas you just reduced your debug pool to one machine: yours.



--

 Martin A. Brooks |  http://www.antibodymx.net/ | Anti-spam & anti-virus
    Consultant    |  martin@...      | filtering. Inoculate
  antibodymx.net  |  m: [phone]           | your mail system.



--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[Marc Perkel]

Martin A. Brooks wrote:
> Marc Perkel wrote:
>> Actually - it is hand rolled.
>
> Alas you just reduced your debug pool to one machine: yours.
>

It's a problem on several machines. Not just one. And it's an Exim
problem that should be addressed.


--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[Martin A. Brooks-3]

Marc Perkel wrote:
>> Alas you just reduced your debug pool to one machine: yours.
>>
>>    
>
> It's a problem on several machines. Not just one.

Do you see how this is the first time you've mentioned that this happens
on more than one machine?


>  And it's an Exim problem that should be addressed.
>  

That has yet to be proven.  It may just be you've misconfigured all the
affected machines in that same way.

Did turning off selinux have any effect?


--

 Martin A. Brooks |  http://www.antibodymx.net/ | Anti-spam & anti-virus
    Consultant    |  martin@...      | filtering. Inoculate
  antibodymx.net  |  m: [phone]           | your mail system.



--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Marc Perkel wrote:
>
> Martin A. Brooks wrote:
>> Marc Perkel wrote:
>>> Actually - it is hand rolled.
>> Alas you just reduced your debug pool to one machine: yours.
>>
>
> It's a problem on several machines. Not just one. And it's an Exim
> problem that should be addressed.

I don't suppose you have a core dump to inspect which function is
causing the segfault? Does it happen during a particular time of the
receipt/delivery? If you can narrow down when it's happening by adding a
few log_write calls to the code that would help.

  You should see what I did to the libPNG/FreeImage to track down a
segfault when they get cross linked :P Loading one 9k image causes
almost 2Mb of debug output all to locate an assigned long int value
being compared to the middle of a char string due to a struct being a
different size at compile time to runtime.

--
The Exim Manual
http://www.exim.org/docs.html
http://www.exim.org/exim-html-current/doc/html/spec_html/index.html

--
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

[svwh-support]

We are having a similar failure of exim.

cPanel 11.24.5-C38506 - WHM 11.24.2 - X 3.9
CENTOS 5.3 x86_64 standard

Although we have several other cPanel server with a similar configuration this is the only one running 64bit.

Aug 21 06:20:53 sh05 kernel: pure-quotacheck[26471]: segfault at 00007fff0342fff0 rip 0000003bf6c42755 rsp 00007fff0342fff0 error 6
Aug 21 08:20:06 sh05 kernel: exim[24275]: segfault at 000000000000004a rip 000000000000004a rsp 00007fff3da412f8 error 4
Aug 21 08:20:09 sh05 kernel: exim[24872]: segfault at 0000000000000008 rip 0000003bf6c6a13d rsp 00007fffc6f74348 error 4