Subject: Re: Please Revert newlock2
To: Zafer Aydogan <>
From: Bill Studenmund <>
List: tech-kern
Date: 02/21/2007 09:07:38
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Feb 21, 2007 at 09:57:05AM +0100, Zafer Aydogan wrote:
> I'm agreeing with Mouse.
> It should be possible to leave  "M:N" as an compile option, for those,
> like Bucky, who work with it.

Do you have a patch that does this? If it's good and clean, it will most=20
likely be included.

As I noted to Bucky, the problem is how well the SA code integrates with=20
the rest of the kernel.

The SA code was not removed from the rest of the kernel because someone=20
didn't like it or someone decided that 1:1 threading was 3133t. It was=20
removed because we didn't have the resources to make it work w/o biglock.=
Think about it, it was easier to add a whole new pthread methodology than=
to fix SA.

Further, the whole rest of the kenrel changed with newlock2. So how the SA=
code works needs to change, even on a UP system. No one has stepped up to=
do this.

As above, if you want to get SA working on a UP system, we'd probably take=
the patch for a while. Long-term, though, we need everything to work well=

Take care,


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

Version: GnuPG v1.4.3 (NetBSD)