stable pc164 milo hangs on 'show'? (was: Re: initrd/cramfs/etc (was Re: srm nightmare, milo & kernel future questions)
More updates on my continuing adventures to run 2.4.21 kernels robustly on
pc164...
I think I'm going to go back to MILO unless somebody can come up with
another idea here for making SRM act right on reboot/halt.
So, I couldn't get SRM to save the boot variables, so I decided to go back
to ARC+MILO. I rebuilt 2.4.21 with the legacy start address and my scsi
driver compiled in so that I didn't need an initrd to load the module. I
reflashed the nvram with ARC (after repartitioning my main drive under the
running OS, which is always fun, back to PC from BSD) and used the latest
pc164 milo from stable. The new milo now hangs on scanning the scsi
bus...it would load the aic7xxx driver, find the host adapter, and just sit
there.
I went back to the milo from potato, which is built on some ancient 2.0
kernel, and it worked fine. It obviously doesn't understand initrd, so I'm
booting without it right now.
Any ideas why the newer milo won't show the devices correctly (or do
anything, since it has to scan the devices (and therefor hangs) before
booting, etc.)? It finds the floppy and the ide drive fine, it just hangs
on the scsi bus (after the host adapter printout).
However, now with arc+milo I have a system that will reboot and halt
consistently on command, which is kinda important for remote administration.
I could have saved about 70+ hours of messing around if I'd just built
2.4.21 with my scsi driver (to avoid initrd) at the legacy start address
(to work with milo) from the start (after installing gcc-3.2, of course)
and kept using milo. It would have "just worked". Hindsight and all that.
Thanks for everybody's help, and I'm still interested in running a newer
milo and initrd (although I'm not sure why) if I can figure out the scsi
scan problem.
Chris
Reply to: