Subject: Re: Context switch revamp
To: None <tech-kern@netbsd.org>
From: Jason R Thorpe <thorpej@wasabisystems.com>
List: tech-kern
Date: 09/23/2002 10:08:41
On Fri, Sep 13, 2002 at 08:33:39AM +1200, Gregory McGarry wrote:

First of all, I would like to apologize for how long it took me to repond,
here...

...and now, on to business...

 > I'm keen to separate the scheduler and the context switching
 > functionality.  This is useful for the following reasons:
 > 
 > 	- simplifies the possibility of replacing the scheduler;
 > 	- aligns better with the scheduler activations changes;
 > 	- allows microtime optimisations if switching to self; and
 > 	- cleans up the horrendous cpu_switch() code.

I think this is all great ... in fact, I started thinking about this
for the "newlock" branch (so that turnstiles didn't have to interlock
with the normal scheduler lock)...

However ... I think this is the wrong time to make this change (and, sigh,
I see this email is a little late -- you've already checked it in).  There
are a couple of reasons for this:

	1. This is going to really make life difficult for the nathanw_sa
	   branch, and is thus going to make it more difficult to get it
	   in final shape to merge down to the mainline.

	2. Some of the changes you propose overlap, to some extent,
	   with changes on the nathanw_sa branch.  I would like to
	   see some coordination between the two.

	3. There have been reports of some problems with the code as
	   currently in the tree.

I'd rather see this developed on a branch of its own, and eliminate all
the conditional __HAVE_CHOOSEPROC stuff.

As such, I think it would be best if these changes were backed out from the
trunk .. but I do think developing this on a branch would be a good idea,
because I definitly think it's the direction we want to go.

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>