Subject: Re: Floating point in the kernel
To: Ken Nakata <email@example.com>
From: David Brownlee <firstname.lastname@example.org>
Date: 09/18/1998 17:52:14
On Sat, 19 Sep 1998, Ken Nakata wrote:
> On Fri, 18 Sep 1998 15:42:04 -0700 (PDT), David Brownlee wrote:
> > I could be missing something here, but since nothing currently in
> > the kernel uses FPU, and all we have to preserve is user data
> > in FPU registers then if you are not using them in an interrupt
> > context, you can just save and restore around each use of the
> > registers?
> Even if so, it does not scale *at all*. So, in Linux or even FreeBSD,
> maybe, but I wouldn't want an ugly hack like this in NetBSD!
If its in an LKM for a DSP chip then I dont see a real problem
a) It doesn't need to scale - just work for that DSP chip.
b) It doesn't impact any other part of the kernel source.
> Isn't it exactly the kind of things NetBSD has been avoiding? "Well,
> it works _for me_" kind of hacks?
Welcome agaon to conflicting goals :) - its a solution that
should work while not impacting anyone else, and keeping the
main tree 'clean'. A clean method of better integrating it would
obviously be preferrable, but this is one 'solution' (or so).
> Oh, I wouldn't mind if someone made such a thing publically available
> as an unofficial patch, provided that its users knew what they were
> doing. But, no, please not in the CVS tree.
-=- "Promise carved in stone, deeper than the sea" -=-