Date: Tue, 10 Mar 1998 13:42:32 +1100 From: Bruce Evans <bde@zeta.org.au> To: freebsd-stable@FreeBSD.ORG, serge@yahoo-inc.com Cc: dk@genesyslab.com Subject: Re: portmap failures Message-ID: <199803100242.NAA28188@godzilla.zeta.org.au>
index | next in thread | raw e-mail
>does anybody know smth about 'portmap' or related kernel bugs ? >I have 2.2.x(some SNAP before 2.2.5R) and 2.2.5 as well as some 2.1 systems >on which portmap dies periodically. I can not catch the rules of those >failures, i.e. what causes them to fail, but here is the message from >/var/log/messages: > >Mar 5 07:47:38 serge portmap[87]: svc_run: - select failed: No child processes >Mar 5 07:47:38 serge /kernel: pid 87 (portmap), uid 1: exited on signal 6 >Mar 5 07:47:38 serge portmap[87]: svc_run returned unexpectedly >Mar 7 15:26:50 serge portmap[5014]: svc_run: - select failed: No child processes >Mar 7 15:26:50 serge /kernel: pid 5014 (portmap), uid 1: exited on signal 6 >Mar 7 15:26:50 serge portmap[5014]: svc_run returned unexpectedly > >the deal is select() should *not* return "no child processes" according to >man pages... This seems to be caused by portmap's SIGCHLD handler not preserving errno. This may be a library bug. POSIX.1 says that it is safe to call wait() from a signal handler. However, portmap() actually calls wait3(), and all syscalls traditionally clobber errno, so robust signal handlers should preserve errno explicitly even if they call a "safe" function. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the messagehome | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199803100242.NAA28188>