From owner-freebsd-current@FreeBSD.ORG  Wed Sep  8 18:32:46 2004
Return-Path: <owner-freebsd-current@FreeBSD.ORG>
Delivered-To: freebsd-current@freebsd.org
Received: by hub.freebsd.org (Postfix, from userid 758)
	id 4FB7716A4CF; Wed,  8 Sep 2004 18:32:46 +0000 (GMT)
Date: Wed, 8 Sep 2004 18:32:46 +0000
From: Kris Kennaway <kris@FreeBSD.org>
To: Brian Fundakowski Feldman <green@FreeBSD.org>
Message-ID: <20040908183246.GB66994@hub.freebsd.org>
References: <413E4A71.9080400@samsco.org>
	<20040908003141.GB928@green.homeunix.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20040908003141.GB928@green.homeunix.org>
User-Agent: Mutt/1.4.1i
cc: freebsd-current@FreeBSD.org
Subject: Re: HEADS UP! Kernel debugging disabled in RELENG_5 [Fwd: cvs
	commit: src/sys/i386/conf GENERIC src/sys/pc98/conf GENERIC
	src/sys/sparc64/conf GENERIC]
X-BeenThere: freebsd-current@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
List-Id: Discussions about the use of FreeBSD-current
	<freebsd-current.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-current>,
	<mailto:freebsd-current-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-current>
List-Post: <mailto:freebsd-current@freebsd.org>
List-Help: <mailto:freebsd-current-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-current>,
	<mailto:freebsd-current-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Wed, 08 Sep 2004 18:32:46 -0000

On Tue, Sep 07, 2004 at 08:31:41PM -0400, Brian Fundakowski Feldman wrote:
> On Tue, Sep 07, 2004 at 05:55:29PM -0600, Scott Long wrote:
> > All,
> > 
> > Since we are closing in on the 5.3 release, I've removed the kernel 
> > debugging options from GENERIC in RELENG_5.  They can of course be 
> > re-enabled by adding the appropriate lines back in locally.  However,
> > from now on when filing a kernel-related bug report, make sure to
> > enable WITNESS, INVARIANTS, and DDB/KDB to help determine the problem.
> 
> It's not necessarily a good thing to disable INVARIANTS without verifying
> it has a huge cost in performance.  It would be invaluable finding more
> bugs in production, and will likely be sorely missed.  IT's also worth
> noting DDB/KDB is broken anyway in fairly large ways (on i386, at least,
> breakpoints don't work and even with the "fix" you get panics soon if
> you really do try to use them).

INVARIANTS does have quite a significant impact on performance. 

5.3 + 4BSD + INVARIANTS
     2205.39 real      3840.87 user      1583.64 sys
5.3 + 4BSD - INVARIANTS
     1862.61 real      3752.97 user      1141.81 sys

That's the 'make world' time for the same source tree built under
different kernels.

Kris