From owner-freebsd-stable  Fri Mar 24 14:18:29 2000
Delivered-To: freebsd-stable@freebsd.org
Received: from mail.hcisp.net (Stargate.hcisp.net [208.60.89.18])
	by hub.freebsd.org (Postfix) with SMTP id 8679637B57A
	for <freebsd-stable@freebsd.org>; Fri, 24 Mar 2000 14:18:20 -0800 (PST)
	(envelope-from tim@mysql.com)
Received: (qmail 12569 invoked from network); 24 Mar 2000 22:24:27 -0000
Received: from modem9.hcisp.net (HELO threads.polyesthetic.msg) (208.60.89.75)
  by stargate.hcisp.net with SMTP; 24 Mar 2000 22:24:27 -0000
Received: (qmail 37534 invoked by uid 1001); 24 Mar 2000 22:17:16 -0000
From: "Thimble Smith" <tim@mysql.com>
Date: Fri, 24 Mar 2000 17:17:16 -0500
To: freebsd-stable@freebsd.org
Subject: pthread_kill() not killing thread
Message-ID: <20000324171716.E37030@threads.polyesthetic.msg>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 1.0.1i
Sender: owner-freebsd-stable@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

Hi.  I'm running 4.0-STABLE cvsup'd on Tuesday (21 Mar).  It
looks like pthread_kill() isn't killing the thread; I could be
doing something wrong, of course.  Here's the output of my test
program:

tim:/tmp$ cc -o simple simple.c -pthread
tim:/tmp$ ./simple
main-thrd: letting test thread start up...
test-body: unblocking signal and looping...
test-body: sleeping...
main-thrd: sleeping...
test-body: sleeping...
main-thrd: killing test thread...
main-thrd: waiting a sec...
test-body: sleeping...
main-thrd: unblocking TERM signal here...
main-thrd: waiting a sec...
test-body: sleeping...
main-thrd: exiting
test-body: sleeping...
test-body: sleeping...
test-body: sleeping...
test-body: sleeping...
test-body: returning
tim:/tmp$ 


The TERM signal never gets delivered to the test thread.  Here
is the code:

#include <assert.h>
#include <pthread.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>

pthread_t test_thread;
sigset_t set;

void *test_body(void *arg)
{
    int i;

    fprintf(stderr, "test-body: looping...\n");
    /* this SIG_UNBLOCK doesn't have any effect */
    /*assert( pthread_sigmask(SIG_UNBLOCK, &set, NULL) == 0);*/
    for (i = 0; i < 8; ++i) {
	fprintf(stderr, "test-body: sleeping...\n");
	sleep(1);
    }

    fprintf(stderr, "test-body: returning\n");
    return NULL;
}

int main (void)
{
    pthread_attr_t attrs;

    sigemptyset(&set);
    sigaddset(&set, SIGTERM);
    assert( pthread_sigmask(SIG_UNBLOCK, &set, NULL) == 0);

    assert( pthread_attr_init(&attrs) == 0);
    assert( pthread_attr_setdetachstate(&attrs, PTHREAD_CREATE_DETACHED) == 0);
    assert( pthread_create(&test_thread, &attrs, test_body, NULL) == 0 );
    assert( pthread_attr_destroy(&attrs) == 0);

    fprintf(stderr, "main-thrd: letting test thread start up...\n");
    sleep(1);

    /* I should have this uncommented for a real program, but it
       doesn't make any difference here - the signal isn't delivered. */
    /*assert( pthread_sigmask(SIG_BLOCK, &set, NULL) == 0);*/

    fprintf(stderr, "main-thrd: sleeping...\n");
    sleep(1);

    fprintf(stderr, "main-thrd: killing test thread...\n");
    assert( pthread_kill(test_thread, SIGTERM) == 0);

    fprintf(stderr, "main-thrd: waiting a sec...\n");
    sleep(1);

    /* With LinuxThreads, this would cause the program to terminate.
       But with -stable, this has no effect. */
    fprintf(stderr, "main-thrd: unblocking TERM signal here...\n");
    assert( pthread_sigmask(SIG_UNBLOCK, &set, NULL) == 0);

    fprintf(stderr, "main-thrd: waiting a sec...\n");
    sleep(1);

    fprintf(stderr, "main-thrd: exiting\n");

    pthread_exit(NULL);
    /*exit(EXIT_SUCCESS);*/
}


The extra pthread_sigmask(SIG_UNBLOCK, &set) that is in test_body()
doesn't make any difference.

Am I doing something wrong?  Or should I file a problem report?

Thanks,

Tim


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message