Bug#741412: exim4: process crashed with signal 11 while delivering

[Christian Hilgers]

Hi Marc,

this is to sumup the past days of testing and collecting data.

>From comparing with the other two Bug reporters, we found that we 
all use a 32bit pae Kernel.  Two of us run VIA CPUs:

from System 1 /proc/cpuinfo 
processor       : 0
vendor_id       : CentaurHauls
cpu family      : 6
model           : 13
model name      : VIA Eden Processor 1200MHz

System 2
vendor_id       : CentaurHauls
cpu family      : 6
model           : 13
model name      : VIA Eden Processor 1000MHz


I was able to create a core and gdb showed me this one line not beeing 
like these: #1  0xada689b9 in ?? ()


[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/i386-linux-gnu/i686/cmov/libthread_db.so.1".
Core was generated by `/usr/sbin/exim4 -v -Mc 1Z2OhU-0001Ke-5q'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0xb69ec614 in padlock_cbc_encrypt () at elf/e_padlock-x86.s:437
437     elf/e_padlock-x86.s: No such file or directory.
(gdb) 

memtest did not indicate any HW Error, but I agree this might be HW releated.
But not related to faulty HW but to the way VIA CPUs differ from Intel ones.

exim does SEGV while sending against one TLS enabled mailhub.
When I move over to eg. strato as mailhub the SEGV is gone.

I tried to reproduce the behaviour on two freshly installed VMs (32bit and 64bit)
and on a pae running Laptop. I failed allways. I cannot provide and reproducer.


Finally I rebuild exim with OpenSSL instaed of GnuTLS and 
after installing the deb exim did never SEGV again.

>From above reasons (memtest OK, Problem gone with OpenSSL,
Problem gone with different mailhub) I ask to forward 
this Bug Report to GnuTLS.


Kind regards

Christian
-- 
Christian Hilgers                       chris at familie-hilgers.com