|
From: <gup...@ne...> - 2010-06-14 15:19:02
|
I have a supervisory process that runs a bunch of python executables, boost and some shared objects that I built in C. Normally this supervisory process kicks up 6 child processes. I am trying to run valgrind with -tool-callgrind and --trace-children=yes option. I have only 256MB RAM on board and I am running a highly stripped linux kernel. When I ran valgrind natively on the supervisory process with --trace-children=yes option, it started some of the python child processes but it crashed soon after. Julian/ Bart suggested I bump up the Active Stack size to 2MB instead of the 64K by default since the symbols names in boost libraries are very long... When I did that there were no more early crashes but 5 of the 6 child processes that are critical to my functionality do not come up. Perhaps the small amount of RAM is an issue? Is there a way in which I can run valgrind so that it instruments specific shared objects in my supervisory process and ignores the other shared objects it finds in the process? Thanks SVG ____________________________________________________________ Refinance Now 4.0% FIXED! $160,000 Mortgage for $633/mo. Free. No Obligation. Get 4 Quotes! http://thirdpartyoffers.netzero.net/TGL3241/4c16485d1d40735350est06vuc |