Partial cores using Linux "pipe" core_pattern

[Paul Smith]

On Mon, 2009-05-18 at 09:25 +0200, Andi Kleen wrote:
> Paul Smith <psmith@gnu.org> writes:
> >
> > Does anyone have any thoughts about where I can look next to try to
> > figure out what's going on?  Ideas or knowledge about limitations of the
> > kernel's core_pattern pipe capability, such as timing issues etc., that
> > might be leaving me with short cores?
> 
> Sounds more like a kernel problem to me. It would be good if you
> could find out in what way the cores are truncated, then do a proper
> kernel bug report.
> 
> Or add printks to the kernel coredumper and try to figure out this
> way where/why it stops.

Heh, I was afraid you'd say that :-).  OK, I'll dig into the kernel
coredumper and add some printks etc.

I do have problems getting a readily reproducible test case.  It seems
timing related; I tried replacing my gzopen()/etc. calls with
uncompressed open()/etc. and that seemed to help; no more short cores.

And yet at other times I can dump the core through gzopen() no problem.

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <psmith@gnu.org>          Find some GNU make tips at:
 http://www.gnu.org                      http://make.mad-scientist.us
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist