Subject: Re: MP/Threading...
To: Greywolf <greywolf@starwolf.com>
From: Alistair Crooks <agc@pkgsrc.org>
List: tech-kern
Date: 03/21/2001 09:31:06
On Tue, Mar 20, 2001 at 10:36:02AM -0800, Greywolf wrote:
> On Tue, 20 Mar 2001, Hubert Feyrer wrote:
> 
> # Date: Tue, 20 Mar 2001 19:34:52 +0100 (MET)
> # From: Hubert Feyrer <hubert.feyrer@informatik.fh-regensburg.de>
> # To: Greywolf <greywolf@starwolf.com>
> # Cc: BSD Kernel <tech-kern@netbsd.org>
> # Subject: Re: MP/Threading...
> #
> #
> # I hear this rumour that you will work on fine grained locking?
> # That's excellent news, when will you be done?
> 
> Well, that answers my question.
> 
> Apologies for having my .doofus flag set.  I'll shut up now.
> 
> [and IFF I can find time and understand how locking works, I'd love
>  to help; I'm just not confident of either my temporal availability
>  nor of my skills.]

I disagree with Hubert here, but that's what makes life fun.

As Bill Sommerfeld explained in a different piece of mail, there's
considerable work involved in doing any form of SMP.  Personally, I've
always been a proponent of the "small steps" approach, rather than the
"big bang" method, and I think this is yet another example where we
should walk before we run. Sure, the aim is fine-grained locking, but
I'd like to get there, and, as a means of achieving the aim, biglock
would seem to be a good way of getting there.

Regards,
Alistair