From owner-cvs-all  Thu May  6 12: 6:42 1999
Delivered-To: cvs-all@freebsd.org
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.26.10.9])
	by hub.freebsd.org (Postfix) with ESMTP
	id 2C6E01506D; Thu,  6 May 1999 12:06:37 -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 FAA06299;
	Fri, 7 May 1999 05:06:35 +1000
Date: Fri, 7 May 1999 05:06:35 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199905061906.FAA06299@godzilla.zeta.org.au>
To: hibma@skylink.it, peter@netplex.com.au
Subject: Re: cvs commit: src/sys/sys kernel.h
Cc: cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org
Sender: owner-cvs-all@FreeBSD.ORG
Precedence: bulk

>Anyway, tsleep() used to work because curproc was set to &proc0 very early,
>and the idle context was viable then, which meant that it did a spl0() for
>the halt and woke up on clock or hardware interrupts... (!)
>
>Hmm.. if I'm interpreting that right, we were breaking the splhigh() around
>the configure process and have been for a while...  I'll revert here and
>see exactly where it's happening..

tsleep() begins with the evil hack

	(if cold || panicstr) { splx(safepri); splx(s); return}

This hides problems with uninitialised and NULL curproc's.  We probably
uncovered some bugs by moving the setting of curproc after the unsetting
of cold.  Note that the hack breaks splhigh().  I think early calls to
tsleep() should be fatal.

Bruce


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