Sven Luther wrote:
On Wed, Mar 31, 2004 at 03:16:31PM +1000, Benjamin Herrenschmidt wrote:Do you have the System.map that corresponds to this kernel ? It's needed to decode the numbers above.It's the standard woody powerpc kernel, from the current CD I think (downloaded it a couple of days ago). There's a copy of the sysmap on the CD; I'll email it direct to you (rather than clutter up the list with attachments) just in case you can't find it.Well, looking at where it fails, it seems it's a HW error triggered by the NCR driver in some way. Difficult to tell precisely what's up though, definitely strange. Woody is a fairly old kernel though, what more recent did you try ?In particular, could you try the debian-installer daily builds : http://people.debian.org/~luther/debian-installer/daily-powerpc-built/current/powerpc-small/floppy/
OK, neither of those boot images work at all. Both of them give me a grey screen with Tux in the centre, and I can hear the floppy drive stepping, but eventually that stops and nothing happens after that. I assume it should ask me for the next disk somewhere along the line?
I have made a little progress using BootX, though. I tried removing the ethernet board just in case it may have been the problem, and now I don't get the kernel crash at that point any more. I'm afraid I've led you all on a wild goose chase, and I apologise for that. I would have removed that board earlier, but the fact that the machine crashed just after initialising the SCSI board and drives, and the fact that I had been getting crashes at that point before (now, when I think about it, when I was trying to boot from the woody install floppy, before I tried BootX), and the fact that I couldn't see how that would affect things, all made me feel that it wasn't the cause of the crashes. Jeez, is the PCI bus always *that* sensitive?
Anyway, I'm back to having it crash during install. This time I got as far as the dhcp initialisation (which went ok, according to my dhcp server log). I then got the "dhcp went ok" dialog with the continue button, but it had locked up at that point and hitting keys on the keyboard had no effect. That's the weird part: it locks up at different spots, which again fit well with the "funny SCSI board" theory. I'd wonder if I had a hardware failure on my hands, but the thing works well under MacOS.
Sigh. Well, at least that's _some_ progress. Again, apologies. .....Ron -- Ron Murray (rjmx@rjmx.net) http://www.rjmx.net/~ronGPG Public Key Fingerprint: F2C1 FC47 5EF7 0317 133C D66B 8ADA A3C4 D86C 74DE