Ok I found the problem ...
There was a configuration missing inside the kernel:
CONFIG_ELF_CORE
General setup --->
-*- Configure standard kernel features (expert users) --->
[*] Enable ELF core dumps
Now it's working as expected, sorry for all the noise!
Kind regards,
Guy Morand
-----Ursprüngliche Nachricht-----
Von: Morand, Guy
Gesendet: Dienstag, 17. März 2015 12:55
An: '[email protected]'
Betreff: core dump
Hello ptxdist developers,
My application is crashing and I cannot get the core dump. To narrow down the
problem, I wrote a simple program:
int main(int argc, char* argv[])
{
int *a = 0;
*a = 0;
return 0;
}
Then I make sure that the core dump is generated correctly:
# ulimit -c unlimited
# echo /tmp/%e.%p.core > /proc/sys/kernel/core_pattern # ulimit -c unlimited #
cat /proc/sys/kernel/core_pattern /tmp/%e.%p.core
Run my test buggy software:
# ./test
Segmentation fault
And no core dump is generated:
# ls /tmp/
I was expecting to have an indication about the core dump being generated on
segmentation fault, as it happens on my host system, I have this error message:
Segmentation fault (core dumped)
Core dump is activated in the kernel:
$ cat .config | grep COREDUMP
CONFIG_COREDUMP=y
I have no setui, setgid bit set for my application:
# ls -l
-rwxr-xr-x 1 xuser 1000 10481 Mar 17 11:25 test
Moreover, I'm pretty sure there is enough space in temp to generate the core
dump:
# df -h
Filesystem Size Used Available Use% Mounted on
192.168.3.10:/home/admin/bsp/platform-ifc/root
25.3G 21.1G 2.9G 88% /
devtmpfs 61.1M 0 61.1M 0% /dev
tmpfs 61.1M 0 61.1M 0% /dev/shm
tmpfs 61.1M 124.0K 61.0M 0% /run
tmpfs 61.1M 0 61.1M 0% /sys/fs/cgroup
none 61.1M 0 61.1M 0% /var/tmp
none 61.1M 0 61.1M 0% /var/lock
tmpfs 61.1M 124.0K 61.0M 0% /var/run
none 61.1M 632.0K 60.5M 1% /var/log
none 61.1M 4.0K 61.1M 0% /tmp
/dev/sda1 7.5G 1.3M 7.5G 0% /mnt/usbstick
What am I missing?
Kind regards,
Guy
--
ptxdist mailing list
[email protected]
