On 21 Apr 2009 at 6:46, Peter Hjalmarsson wrote: > Thanks for the information. > This time I ran > strace -ff -o ~/fah6-strace ./fah6 > > The part I ran previously was just the core which is run by fah6 after > ensuring there is a workunit and all that. fah6-strace.795 seems to be > the log for the core, and 796 its failing child.
it seems that this binary is linked against the old linuxthreads glibc and fails to allocate the thread stacks due to randomization, so just disable that one PaX feature on it and it should work (as it apparently does) and also feel free to tell the folding guys that it's time to move into the 21st century and switch to nptl. at least that one doesn't want to allocate executable stacks.
