Matt Sealey wrote:
As far as I recall, it booted into single user mode, then they couldn't get the openfirmware ethernet to work - plus a few firmware bugs may have stalled progress (some of these are fixed now). We have suggested eschewing the OF port for now and using x86 drivers (NetBSD 2.0 supports everything except the Firewire and GigE on other hardware) instead.
Neither ethernet nor hard disk worked. I didn't debug it fully, but I'm pretty sure it was related to the SmartFirmware bug that caused double-length numbers to be stacked in reverse order. That would cause things like OF_seek() to fail.
This could be handled with #ifdefs in sys/arch/powerpc/powerpc/openfirm.c, but this solution would be The Bad Thing.
The clean solution is fixing SmartFirmware, of course. I don't know the status of SF, but I can try again in case that bug has been fixed.
On January 9th, that's tomorrow, the hardware specs and schematics become available from Power.org to registered members (free signup). This may spur someone to look back at it, with renewed information and so on.
Well, this is good news. It probably isn't necessary at this stage (as you said, other platforms already drive almost all the hardware), but will certainly be helpful for all OSs considering Pegasos.
If anyone is considering reviving NetBSD/Pegasos contact me. The old repository is still available and we could easily bring those changes to a more recent 3.x tree.