From owner-freebsd-current  Mon Sep  7 13:11:49 1998
Return-Path: <owner-freebsd-current@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id NAA27936
          for freebsd-current-outgoing; Mon, 7 Sep 1998 13:11:49 -0700 (PDT)
          (envelope-from owner-freebsd-current@FreeBSD.ORG)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.15.68.22])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id NAA27928
          for <current@FreeBSD.ORG>; Mon, 7 Sep 1998 13:11:47 -0700 (PDT)
          (envelope-from bde@godzilla.zeta.org.au)
Received: (from bde@localhost)
	by godzilla.zeta.org.au (8.8.7/8.8.7) id GAA16751;
	Tue, 8 Sep 1998 06:11:31 +1000
Date: Tue, 8 Sep 1998 06:11:31 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199809072011.GAA16751@godzilla.zeta.org.au>
To: ache@nagual.pp.ru, current@FreeBSD.ORG
Subject: Re: new sleep.c & SIGALRM problem
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

>When SIGALRM comes at INT_MAX seconds, new sleep code kill the program
>instead of stopping sleep.

This doesn't matter if ints are 32 bits, because the error can't be
detected without sleeping for a INT_MAX seconds (about 68 years).
The previous version was correct enough for the same reason.

>Previous code works, so LONG_MAX comparison
>must be restored, it is pretty legal in C.

The LONG_MAX was broken on alphas because time_t is int32_t on alphas,
so UINT_MAX < LONG_MAX but tv_sec = UINT_MAX overflows to -1 and
sleep(UINT_MAX) returns immediately.  This should be handled by
defining TIME_T_MAX in a header file and comparing with it instead
of LONG_MAX or INT_MAX.  There is no need to loop, because time_t
must have a range large enough to cover all interesting times.

Bruce

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