Subj : Re: NPTL and sigwait() problem under linux
To   : comp.programming.threads
From : steve
Date : Thu Aug 18 2005 11:54 pm

In article <fg3Ne.25245$hV3.9265@nntpserver.swip.net>,
Daniele  <arkaam@everyday.com> wrote:
>
>P.S. Now I install handlers for SIGSEGV and SIGFPE after signal_init(), in the 
>main thread, and they are working fine.

Just be aware that SIGSEGV and SIGFPE, when generated by the
implementation (as opposed to being sent with kill/pthread_kill) are
generated synchronously, and thus *will* *only* get delivered to the
thread that caused them.  You can't create a signal-handling thread
to cope with those particular signals.

Also, on many implementations, masking SIGSEGV and then generating one
will result in the signal being unmasked, whether you want it or not.
If there's no handler, expect a core dump.  If there is a handler, I
would still strongly advise that the process not last much longer after
a SEGV.  There are all sorts of bad corners that you can crawl into with
SEGV handlers.
-- 
Steve Watt KD6GGD  PP-ASEL-IA          ICBM: 121W 56' 57.8" / 37N 20' 14.9"
 Internet: steve @ Watt.COM                         Whois: SW32
   Free time?  There's no such thing.  It just comes in varying prices...

.