Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 22 Dec 1997 15:21:27 +1030
From:      Mike Smith <mike@smith.net.au>
To:        John-Mark Gurney <gurney_j@resnet.uoregon.edu>
Cc:        Mike Smith <mike@smith.net.au>, FreeBSD Hackers <freebsd-hackers@freebsd.org>
Subject:   Re: converting drivers to dynamic memory... 
Message-ID:  <199712220451.PAA00692@word.smith.net.au>
In-Reply-To: Your message of "Sat, 20 Dec 1997 23:33:51 -0800." <19971220233351.10621@hydrogen.nike.efn.org> 

next in thread | previous in thread | raw e-mail | index | archive | help
> Mike Smith scribbled this message on Dec 21:
> 
> my point was that this "framework" that you said was debunking my changes
> ISN'T complete and working, even though you said it was... I will be
> needing it in a couple months...

To be honest, I made no claim as to its functionality, merely its 
inevitability.  It's hardly reasonable to be making loud accusations of 
debunkment when you're talking about a hypothetical scheme yourself.

> > This is entirely contingent on the acceptance of devfs as "the way to 
> > go", and the encompassing of the various prerequisite tasks (such as 
> > persistence and fixing buggy devfs support in drivers).
> 
> so, until that happens, can I procede with phase one (that Darren Reed
> was so nice to name :) )...  almost ALL of that work will be able to
> easily translate once devfs is complete..

What was "phase one"?  I'm hardly going to stop you doing something 
yourself, but you should consider whether the work is worth the effort.

> > I think that devfs will be a goer for 3.0.  I don't know what sort of 
> > timetable you're on for your bus restructuring, but I suspect you may 
> > be looking at 3.1 for that.
> 
> I'm hoping for having the bus/device code completely written and
> functional with in the next two months...  of course, after the bus/device
> code is done, it will require massive changes to the device drivers...

I would recommend working on a minimal functional subset of drivers for 
the new model, such that people can test the framework before you spam 
it into -current.  You should think long and hard about how you can 
support old-style drivers in the new environment as you suggested,
as you'll make some *very* unhappy vendors otherwise.

mike




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199712220451.PAA00692>