Subject: Re: 3.99.7 system crashed while shutting down
To: Matt Thomas <>
From: Bill Studenmund <>
List: tech-kern
Date: 08/24/2005 13:01:35
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sun, Aug 21, 2005 at 09:28:21AM -0700, Matt Thomas wrote:
> Chuck Silvers wrote:
> >I looked into this, the problem is a race between VOP_REVOKE() and=20

> I've heard that in OSX 10.4, the VOP locking protocol has been redone so =
> actually works properly.  Maybe we should look at using their protocol.

The problem isn't the locking protocol, it's that we don't follow the
protocol we now have. That fix removed an invariant (vp->v_specinfo is
always good) w/o changing all the other uses. And the invariant really=20
doesn't involve locking, so changing the locking won't really help. :-|

In a lot of ways, it doesn't matter what protocol we follow, as long as we=
actually follow it.

I don't object to us considering other VOP layer formalisms. I'm actually=
interested in moving more towards the SVR4 formalism, but Tiger's could be=
interesting too. However I think we should not change formalisms just=20
because we don't code to the formalism we have. :-)

If we run into a problem that is intractable with our formalism, that's=20
different. :-) But I don't think we've hit such a thing yet.

Take care,


Content-Type: application/pgp-signature
Content-Disposition: inline

Version: GnuPG v1.2.3 (NetBSD)