> On Tue, 10 Jul 2001, Hayden, Bill wrote:
> 
> : > (Of course, the long term solution is to switch to a booter
> : > like BootX that
> : > can boot on everything and bypass the need to go to OFW 
> : > and type in commands.)
> :
> : Just for the record, which BootX are you talking about?  
> There has been
> : preliminary work to get Apple's BootX booting NetBSD,
> 
> That's the one.
> 
> : which has currently reached a sticking point due to the 
> : fact that NetBSD continues to use the OF console after boot.
> 
> If, by "continues to use the OF console," you mean "uses the 
> OF console for
> printf() and keyboard interaction", that is relatively simple 
> to fix (just
> code tedium, mainly).  The Shark port, for instance, starts 
> from OFW on ARM,
> but buffers kernel messages until the "pc0" console starts up 
> -- at which
> point, all the messages get dumped to that console.
> 
> However, if you mean "uses the OF callbacks to get system 
> hardware info,"
> and BootX doesn't allow that, I guess we're screwed.

You should probably check out the Darwin-Developers thread which discussed
this.  It's only about 5 messages long, but very informative.  You can see
it by going to <http://www.darwinfo.org/devlist.php3> and doing a search for
"quietsce".  Unfortunately, the consensus was that the OF interface needs to
go away permanently once the kernel is called, which is apparently not at
all how NetBSD currently does things.  There was a hack solution presented
which may at least get people out of OF hell, though.  (Basically, it's a
"is this NetBSD?" test which would then not blow away OF)

Bill Hayden