From owner-freebsd-alpha  Thu Jun  7 15:18:41 2001
Delivered-To: freebsd-alpha@freebsd.org
Received: from duke.cs.duke.edu (duke.cs.duke.edu [152.3.140.1])
	by hub.freebsd.org (Postfix) with ESMTP
	id 2F8CE37B401; Thu,  7 Jun 2001 15:18:38 -0700 (PDT)
	(envelope-from gallatin@cs.duke.edu)
Received: from grasshopper.cs.duke.edu (grasshopper.cs.duke.edu [152.3.145.30])
	by duke.cs.duke.edu (8.9.3/8.9.3) with ESMTP id SAA27119;
	Thu, 7 Jun 2001 18:18:37 -0400 (EDT)
Received: (from gallatin@localhost)
	by grasshopper.cs.duke.edu (8.11.3/8.9.1) id f57MI7355072;
	Thu, 7 Jun 2001 18:18:07 -0400 (EDT)
	(envelope-from gallatin@cs.duke.edu)
From: Andrew Gallatin <gallatin@cs.duke.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <15135.64927.633480.754465@grasshopper.cs.duke.edu>
Date: Thu, 7 Jun 2001 18:18:07 -0400 (EDT)
To: Alfred Perlstein <alfred@freebsd.org>
Cc: jhb@freebsd.org, freebsd-alpha@freebsd.org
Subject: Re: Wow, have I got a stack for you!
In-Reply-To: <20010607164933.Y1832@superconductor.rush.net>
References: <15135.58016.753596.764419@grasshopper.cs.duke.edu>
	<20010607164933.Y1832@superconductor.rush.net>
X-Mailer: VM 6.75 under 21.1 (patch 12) "Channel Islands" XEmacs Lucid
Sender: owner-freebsd-alpha@FreeBSD.ORG
Precedence: bulk
List-ID: <freebsd-alpha.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo?subject=subscribe%20freebsd-alpha>
List-Unsubscribe: <mailto:majordomo?subject=unsubscribe%20freebsd-alpha>
X-Loop: FreeBSD.org


Alfred Perlstein writes:
 > * Andrew Gallatin <gallatin@cs.duke.edu> [010607 16:25] wrote:
 > > 
 > > Take a look at the following trap, courtesy the all-singing /
 > > all-dancing ddb traceback from NetBSD.
 > > 
 > > We're trapping in pmap_insert_entry(), then trapping in vm_fault1().
 > > We start to print out the trap info, but then we get a clock
 > > interrupt.  The clock interrupt then traps.  Witness does its thing
 > > and brings this mess to a screeching halt. (it doesn't hurt that
 > > we're finally at ipl 7).
 > > 
 > > It sure looks like somebody is blowing away some low-level vm state.
 > > 
 > > A patch to improve alpha ddb is up for review at
 > > http://people.freebsd.org/~gallatin/ddb.diff
 > 
 > The way I almost stabilized i386 was by placing an assertion in
 > almost every pmap function to assert that the vm mutex was owned.

The key here is "almost" -- is there any documentation on which
functions should assert that vm_mtx is owned?

 > I was then able to track down direct callers of pmap as well as
 > oevrsights in the vm system that I had missed.

The only i386 pmap locking I've seen is is mapdev, which isn't
required on alpha (it maps the phys into K0SEG, which is basically an
| ).

Drew

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