Tom Riddle wrote:
> An update after discussions with the newboot team. SPARC platforms and
> PPC/ODW are fairly similar with the OBP/OF piece in place. The newboot move
> to SPARC is on the roadmap, but they have no immediate plans to implement.
> This leaves with 2 choices.
>
> The quickest path with the least amount of code change would be leveraging
> ufsboot as a binary to copy in krtld/unix/genunix and go. So use OF directly
> or GRUB2 to load a PPC based ufsboot. A more elegant way would be to
> integrate ufsboot/netboot into GRUB2. Some of the 2.5.1 code may be useful
> here but we have to verify. But this is not where newboot is going longterm.
>
> The other way would be to write a multiboot type of loader (ie; AMD64) more
> inline with the newboot arch, becoming less of in fact independent on OBP for
> callbacks. This would encompass a port of the ramdisk driver with multiboot
> loading krtld, etc and go. Benefits of portability are obvious and 2.5.1
> levergae is less relevant.
>
> thoughts?
Tom,
First of all thanks for update. Fast communication is critical for
the community project success.
As for the choices you've presented - that is exactly the situation
as we discussed it on this list around a year ago - when the PowerPC
community was created. Taking the ufsboot route may seem to be the
quick way to bootstrap the activity, however it has significant
drawback of requiring working disk driver to be able to move beyond
the point where the root disk driver chain is being assembled.
Bearing in mind that PegasosII has ATA disk and the Solaris ATA driver
isn't opensourced it may be a major showstopper for us. Netboot
is a obvious alternative as all the components are in current source
tree and significant similarity between SPARC OBP and Peg' SmartFirmware
(both are implementations of OpenFirmware) simplifies port to Polaris.
Taking GRUB2 route is very attractive too. Mainly because we can even
boot from disk in that case - GRUB2 will bring the boot archive from
the disk and (while not being able to mount the disk) we can boot,
practically, to the fully operational state of the OS. And network
root is still an [easy] option. Downside - GRUB2 in its current state
needs some work to be able to boot Solaris. And Solaris itself needs
some work too. And then we can add the ATA driver based on the state
of the whole [S]ATA framework at the moment.
I have strong feeling that GRUB2 way worth being chosen.
As a side note I want to add that Sun's road map (both short- and
long-term) WRT GRUB in Solaris is totally unclear. I never saw
that UFS support was submitted to GRUB. And the GRUB itself in Solaris
stopped at 0.95, while the GRUB itself went ahead to 0.97.
It would be very beneficial for [our] community if engineers from
NewBoot Team shared their road map and plans with the rest of us.
I am taking the freedom of CC'ing Jan Setje-Eilers. Jan, if think
you can participate in the discussion, please do so.
Regards,
Cyril
