* Eric W Biederman <[EMAIL PROTECTED]> [020715 13:38]:
> For the most part I don't think moving to a newer codebase will be all
> that hard, but will take several days work. Especially if you want to
> support etherboot this is something worth doing.
I bet my patch back from some weeks ago is quite far from complete, but
it was half a day of work - any idea how far it is from the right path?
I will try to work on this as soon as I have a working image.
> So (a) we need to move the palcode into LinuxBIOS. This should fix
> the Linux booting Linux issues.
This is a good idea - all other firmware implementations (namely srm,
alphabios, arc and arcsbios) do the same here. being the only ones that
do it different will not make things easier. and you do not want to
bloat the kernel with PALcode stuff.
> (b) Either setup the whacky virtual addressing expected by the Alpha
> architeture, or have a small kernel patch that can live without it.
> Probably the latter is easier.
Milo sets up some linear 1:1 addressing which should be fairly easy to
realize. This works even when flashing milo to ROM as an AlphaBIOS
supplement.
> Why do you wish to port etherboot to the Alpha. I'm just trying to
> get a feel for what you are after.
Etherboot is perfect for easy development. Instead of flashing a new
ELF image to flash every time you try, you can just update the image
on the server and reboot the machine. Most of this testing can be done
on intel though once LinuxBIOS itself is working again I guess
Stefan
--
The x86 isn't all that complex - it just doesn't make a lot of
sense. -- Mike Johnson, Leader of 80x86 Design at AMD
Microprocessor Report (1994)
