I consider SRM to be the way to go, and i think instead the convoluted
hacks that things like MILO had to go through are the rsult of using the
ARC/ALphaBIOS _instead_ of using SRM. Now that Compaq has released SRMs
for so many more models and eliminated the need to buy the developer kit,
i see no reason to consider ARC anymore _at all_. It was a cheap second
best rigged up so that NT could work on Alphas, and I dont see why we need
to try to cater on NT's inadequacies and start using ARC/AlphaBIOS or the
workarounds to it like MILO. SRM is the 'high class' firmware, whatever,
it is available and is superior to ARC/AB. BSD has used SRM since the
beginning- why should we dilute that now, especially when some of the last
arguments for it have lost their bite? 

just my $.02, 
isildur

On Fri, 3 Dec 1999, Matthew Jacob wrote:

> 
> 
> ---------- Forwarded message ----------
> Date: Fri, 3 Dec 1999 11:47:27 -0800
> From: Richard Henderson <rth@twiddle.net>
> Reply-To: axp-list@redhat.com
> To: "warp@xs4all.nl" <warp@xs4all.nl>
> Cc: axp-list@redhat.com
> Subject: Re: Q: Compaq, *BSD and 'Linux-only' AlphaBIOS
> Resent-Date: 3 Dec 1999 21:15:44 -0000
> Resent-From: axp-list@redhat.com
> Resent-cc: recipient list not shown: ;
> 
> On Thu, Dec 02, 1999 at 02:02:29PM +0100, warp@xs4all.nl wrote:
> > 1) Is anybody planning to make MILO work with NetBSD*;
> 
> >From what I could tell from browsing the OpenBSD source, the only
> thing standing in the way of the BSDs booting off MILO is their
> reliance on SRM callbacks for printing to the console during the
> early boot process.  And some bits that read the SRM environment
> variables to snag the kernel command line and such.
> 
> It shouldn't be terribly hard to either implement the SRM callbacks
> in MILO or (conditionally) avoid them in the BSD kernel.  The former
> would of course be more helpful.

certainly, if anyone _needs_ to use BSD with MILO for some strange reason,
it'd be better to modify MILO than BSD.