From owner-freebsd-current  Thu Oct 24  4:59:20 2002
Delivered-To: freebsd-current@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id 0142237B401
	for <current@freebsd.org>; Thu, 24 Oct 2002 04:59:19 -0700 (PDT)
Received: from mail03.svc.cra.dublin.eircom.net (mail03.svc.cra.dublin.eircom.net [159.134.118.19])
	by mx1.FreeBSD.org (Postfix) with SMTP id 057A443E65
	for <current@freebsd.org>; Thu, 24 Oct 2002 04:59:18 -0700 (PDT)
	(envelope-from pmedwards@eircom.net)
Received: (qmail 81566 messnum 527871 invoked from network[159.134.237.90/chester.eircom.net]); 24 Oct 2002 11:59:16 -0000
Received: from chester.eircom.net (HELO webmail.eircom.net) (159.134.237.90)
  by mail03.svc.cra.dublin.eircom.net (qp 81566) with SMTP; 24 Oct 2002 11:59:16 -0000
From: "Peter Edwards" <pmedwards@eircom.net>
To: current@freebsd.org
Subject: Floating point problems
Date: Thu, 24 Oct 2002 12:54:47 +0100
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 8bit
X-Originating-IP: 62.17.151.61
X-Mailer: Eircom Net CRC Webmail (http://www.eircom.net/)
Organization: Eircom Net (http://www.eircom.net/)
Message-Id: <20021024115918.057A443E65@mx1.FreeBSD.org>
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
List-ID: <freebsd-current.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo@FreeBSD.ORG?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo@FreeBSD.ORG?subject=subscribe%20freebsd-current>
List-Unsubscribe: <mailto:majordomo@FreeBSD.ORG?subject=unsubscribe%20freebsd-current>
X-Loop: FreeBSD.ORG

Hi,
There was some discussion about issues with interactions between the floating 
point context and signal handling in a thread a week or so ago, and a suggestion 
that someone try and get a simple test that would fail. I was surprised how
easy it was: The following program just spins calculating the value of 6.0 / 
3.0, and traps SIGINT.

If you run it on -current (as of a few hours ago), 99% of the time, hitting 
ctl-C will cause the program to exit with an error. A 4.5 kernel never causes 
any problems.

I'm pretty sure this is what's causing the stalls and crashes in X. I've taken
stack traces of crashes, and from "spinning" processes, and I can spot NaNs on
the stack that shouldn't be there, etc.

I just thought it might be useful to send this, to firmly take the blame
away from the X server, and give someone a way of reproducing the bug.

fptest.c:

#include <math.h>
#include <stdio.h>
#include <signal.h>
#include <assert.h>


int count = 0;
void sigint(int sig)
{
    write(2, "INT\n", 4);
    count++;
}

int
main(int argc, char **argv)
{
    double x = 6.0;
    double y = 3.0;
    double d;
    double err;
    struct sigaction sa;

    sa.sa_handler = sigint;
    sa.sa_flags = 0;
    sigemptyset(&sa.sa_mask);
    sigaction(SIGINT, &sa, 0);

    while (count < 30) {
        d = x / y;
        err = 2.0 - d;
        if (err != 0.0) {
            fprintf(stderr, "err %f!\n", err);
            exit(-1);
        }
    }
}


-- 
Peter Edwards.


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