Re: Alpha issues (SRM installation)
On Sat, Feb 12, 2000 at 12:19:44AM +0100, Christian Meder wrote:
> On Wed, Feb 09, 2000 at 07:38:01PM -0500, David Huggins-Daines wrote:
> > Hi,
> >
> > I've been doing a few things to make installing Debian on Alphas that boot
> > from SRM firmware a bit more pleasant, and I'm wondering about how to
> > proceed with a few things.
>
> Hey, that's great. I checked in our alpha subarch stuff (it's just really
> the old stuff a bit cleaned up) in CVS this afternoon. The next TODO items
> for Alpha are:
>
> * fixing the alphaspecific bootfloppies bugs
Which ones are these again (aside from usability issues like selecting the
correct fdisk and boot method)
> * make harddisk bootable and make rescue floppy
I did these for SRM on Thursday (and actually tested/fixed them on Friday :)
Should we try to do something for MILO installs as well? The problem is
that people might be loading MILO in any number of different ways (it could
be in firmware, on a DOS partition, or from SRM even) so there doesn't seem
to be any general method for making a machine bootable with MILO. Also
there isn't much point in making a MILO bootfloppy (or an SRM one even,
actually) I don't think, since the kernel ("rescue") disk is basically the
same thing...
That said, we could check for a DOS partition and offer to copy MILO and
linload.exe to it.
> * boot from CDROM
Do we have potato CD scripts (for any architecture) yet? Who is working on
this?
> * update the Alpha documentation
Yes - I think I can do this. This is probably the most important thing, as
slink's documentation was nearly useless...
> If you're tackling the second issue that would be great. Could you discuss
> your ideas with Flynn Marquardt <flynn@isr.uni-stuttgart.de> ? He's done
> a lot of work and testing with our Alpha boot-floppies setup and has got a
> couple of ideas for the boot issues too.
I've Cc:'ed him - the code that I've got so far is in
utilities/dbootstrap/bootconfig.c, utilities/dbootstrap/partition_config.c,
and utilities/dbootstrap/main.c (where the SRM/MILO detection should go -
it's not there yet, and just defaults to SRM :)
Reply to: