Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 8 Mar 2006 10:21:09 +0800
From:      David Xu <davidxu@freebsd.org>
To:        freebsd-threads@freebsd.org
Cc:        john gladkih <john@kak-sam.to>, freebsd-gnats-submit@freebsd.org
Subject:   Re: threads/94176: KSE: sigwait doesn't recieve SIGWINCH sent by pthread_kill() or kill -WINCH
Message-ID:  <200603081021.09351.davidxu@freebsd.org>
In-Reply-To: <200603071617.k27GHSM5000487@www.freebsd.org>
References:  <200603071617.k27GHSM5000487@www.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday 08 March 2006 00:17, john gladkih wrote:
> 
> >Number:         94176
> >Category:       threads
> >Synopsis:       KSE: sigwait doesn't recieve SIGWINCH sent by 
pthread_kill() or kill -WINCH
> >Confidential:   no
> >Severity:       serious
> >Priority:       medium
> >Responsible:    freebsd-threads
> >State:          open
> >Quarter:        
> >Keywords:       
> >Date-Required:
> >Class:          sw-bug
> >Submitter-Id:   current-users
> >Arrival-Date:   Tue Mar 07 16:20:04 GMT 2006
> >Closed-Date:
> >Last-Modified:
> >Originator:     john gladkih
> >Release:        5.4-RELEASE, 5.3-RELEASE
> >Organization:
> >Environment:
> FreeBSD freebsd54.i.drweb.ru 5.4-RELEASE FreeBSD 5.4-RELEASE #0: Sun May  8 
10:21:06 UTC 2005     
root@harlow.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC  i386
> 
> >Description:
> sigwait() doesn't recieve SIGWINCH (SIGIO and maybe some others signal) sent 
by pthread_kill() from another thread or even from kill -WINCH. there is no 
such issue on 5.2.1 and older with libc_r.

Good luck! I have been waiting for this PR for a few years.
The problem you've hit is that default signal action of  these signals 
is ignore, which causes kernel immediately discard them at the time
the signals are generated, I believe libc_r sets handlers for
these signals, it causes kernel to always deliver these signals to
userland. POSIX does not specify if signal should be discarded immediately
or remain pending if its action is ignore. this means sigwaiting a signal
which action is ignore is not reliable or portable.
http://www.opengroup.org/onlinepubs/000095399/functions/xsh_chap02_04.html

Traditionally, FreeBSD immediately discards a signal if its action is
ignore except that process is being debugged. I am not sure I should
change this history behavior. To work around the problem you've seen,
you can set a dummy action for those signals you want to wait for.

David Xu



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200603081021.09351.davidxu>