On Fri, 5 Jan 2007, Michael Schmitz wrote:
The Q650 and Q700 are usually OK, but the 650 sometimes fails with the
same SCSI failure that kills the 630 on almost every boot (but even
2.6.7 had that bug, just harder to trigger. It only fails during
device probing. If you get past that, it is solid.).
Some stray interrupt coming in while SCSI is probing, and messing up
things in the interrupt handler? Does it hang with interrupts disabled?
Could be. It is a very racey bug, in that it is intermittent and is worse
on some models than others. And with 2.6.18-4 it is harder to trigger than
2.6.19 on the LC630 (but it still happens sometimes). I swapped hard disk,
cable and terminator but no change.
No idea what changed since 2.6.10 ... the IIci perhaps fails due to
screen memory getting in the way in the low memory bank. While I think
about it - one major change was the discontiguous memory support. But
that usually improved things.
You are right. Discontiguous memory fixed the early boot hang on the IIsi
and IIci. Thanks.
By disabling CONFIG_MAC_SCSI, I can NFS boot to a prompt. Now I can test
my patches on the non-quadra macs. I'd really like to finish with ADB and
move on to the real time clock, so I'm not about to tackle the SCSI bugs.
Around 2.6.10 I could also boot the 630 from IDE, but I noticed from
around 2.6.16 and later the disk never spins up again after penguin
spins it down, and can't be used ("no such device" or something like
that). Pity about that, it was faster than the SCSI driver.
I found the IDE problem, I did not have CONFIG_IDE_GENERIC enabled :-/
CONFIG_BLK_DEV_MAC_IDE probably needs to depend on CONFIG_IDE_GENERIC.
However, there is still a very old bug there, where heavy IDE and SONIC
traffic together cause all the NuBus interrupts (which are only SONIC &
IDE on this machine) to cease altogether. I'll probably do some more work
on this, but I'm not optimistic; others have tried and failed --
http://marc.theaimsgroup.com/?l=netbsd-port-mac68k&m=96498911504667&w=2
And it would appear that IDE used to be polled from the VIA1 IRQ handler
(I guess the F108 chip is another of Apple's mysteries...)
-f
-
To unsubscribe from this list: send the line "unsubscribe linux-m68k" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
