From owner-cvs-lib  Sat May 17 23:11:27 1997
Return-Path: <owner-cvs-lib>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id XAA12648
          for cvs-lib-outgoing; Sat, 17 May 1997 23:11:27 -0700 (PDT)
Received: from spinner.DIALix.COM (spinner.dialix.com [192.203.228.67])
          by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id XAA12637;
          Sat, 17 May 1997 23:11:11 -0700 (PDT)
Received: from spinner.DIALix.COM (localhost.dialix.com.au [127.0.0.1])
          by spinner.DIALix.COM with ESMTP id OAA04447;
          Sun, 18 May 1997 14:10:47 +0800 (WST)
Message-Id: <199705180610.OAA04447@spinner.DIALix.COM>
X-Mailer: exmh version 2.0gamma 1/27/96
To: Bruce Evans <bde@zeta.org.au>
cc: cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-lib@FreeBSD.ORG,
        peter@FreeBSD.ORG
Subject: Re: cvs commit: src/lib/libc/gen usleep.c 
In-reply-to: Your message of "Sun, 18 May 1997 13:01:28 +1000."
             <199705180301.NAA20270@godzilla.zeta.org.au> 
Date: Sun, 18 May 1997 14:10:45 +0800
From: Peter Wemm <peter@spinner.DIALix.COM>
Sender: owner-cvs-lib@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

Bruce Evans wrote:
>  Modified files:
>    lib/libc/gen         usleep.c 
>  Log:
>  if nanosleep returns too early, loop.  usleep() does not have a return
>  value, it appears as though the semantics of usleep are that it doesn't
>  return early.  (only in the nanosleep code - the setitimer code does this
>  already)
>
> Where do those semantics appear?  The old behaviour was to return
> early.  The man page said otherwise, but so does the BSD4.4-Lite man
  ^^^^^
Are you sure of that?  Both sleep() and usleep() do a 'while
(!ringring) sigpause(...)'..  To my reading, it won't return early ever...

The sleep(3) implementation was the same, it didn't ever return early 
either as far as I can see.  There's no way out except for a longjump from 
within some other signal handler.

My observations of the svr4 libc suggest that it uses alarm() and a
sigpause() once and doesn't sit in a while loop..  So, sleep(3) will return
early.  The svr4 version of usleep() (presumably based on sunos 4.x code)
appears to use setitimer and do the same while (!finished) sigpause(..).

> page for sleep(3).  usleep.3 is a clone of sleep.3 in in BSD4.4-Lite.
> Someone fixed the worst bugs in sleep.3 but didn't change usleep.3.
> The actual behaviour hasn't changed since 386BSD.

It seems to me that somebody cloned the usleep(3) code to replace 
sleep(3), presumably this was csrg, but I couldn't guess how far back.

> Bruce

Cheers,
-Peter