This is the mail archive of the
gdb@sourceware.org
mailing list for the GDB project.
Shared memory and coredumps
- From: Lukas Sinker <lusnkr at gmail dot com>
- To: gdb at sourceware dot org
- Date: Wed, 7 Apr 2010 11:08:55 +0200
- Subject: Shared memory and coredumps
Hello,
how to analyze a coredump that has a shared memory included? I'm
actually having problems with getting such core created. I've tried
the following:
ps -eLf | grep myprocessname | grep -v grep | while read artm; do
??????? pid=`echo "$artm" | gawk '{ print $4 }'`
??????? echo 0xffff > /proc/$pid/coredump_filter
??????? echo -n "coredump_filter for $pid set.. "
??????? cat /proc/$pid/coredump_filter
done
The result, e.g.:
coredump_filter for 16434 set.. 0000001f
coredump_filter for 16446 set.. 0000001f
coredump_filter for 16466 set.. 0000001f
coredump_filter for 16467 set.. 0000001f
and then I trigger the core dumping process. The resulting core
doesn't include my 2GB shared memory address space. The address space
is included in e.g. top's (or htop's) view. It's mmap()-ped by using a
/dev/somecharacterdevice. Why it isn't included in the core dump?
Lukas