On Tuesday 12 May 2009 11:03:49 ipif wrote: > I'm still coping with my embedded sparc system. > <snip> > Running strace the initial parts are identical: > # strace ntpdate -u <server> > <snip> > On the original system the following happens: > poll([{fd=3, events=POLLIN}], 1, 60000) = -1 EINTR (Interrupted system > call) --- SIGALRM (Alarm clock) @ 0 (0) --- > <snip>
looks like signal handling is busted. iirc, ntp acts by setting up an alarm() and then polling with short timeouts indefinitely until the SIGALRM fires. your strace output indicates this: the correct one had SIGALRM delivered while the incorrect one did not. check /proc/<pid>/status to compare signal masks between the working processes. -mike _______________________________________________ uClibc mailing list uClibc@uclibc.org http://lists.busybox.net/mailman/listinfo/uclibc