G5 fan problems return moving to 2.6.15 with dual processor 2.7GHz machine

[Brian D. Carlstrom]

Benjamin Herrenschmidt writes:
 > Might be something in that prom_init.c fix that broke... it would be
 > really nice if you could give a try with the console and find out what
 > it is ... Unfortunately, I don't have access to one of these machines
 > with the "problem" at the moment...

Well, I added several prom_printf calls to prom_init.c's
fixup_device_tree routine. I assumed I would spot these scrolling by
during boot before what appears to be the video mode switch. However, I
didn't see anything, but I wasn't sure if it wasn't just going by too
fast.

I tried using PROM_BUG to halt the output, but that just resulted in
returning to an OpenFirmware prompt, although with a white background
instead of the usual black background when I go their from yaboot with
'o'.

I also tried putting a "while (1) ;" after one of my prom_printf, in
case the illegal instruction used by PROM_BUG was causing the output to
be lost, since it was clearing the screen to display the OpenFirmware
prompt. However then I just got a pure white screen. So clearly in both
cases it was running my changed code, but I see no output.

I tried reviewing some OpenFirmware doc, looking at their talk of
debugging via serial and telnet, but that all seemed to be a dead end,
although I learned much more about the device tree. :)

Clearly I could theoretically debug by moving the while(1); around to
see what branches are being taken, but since I'm away from the machines
today, I figured I'd ask how I'm expected to use prom_printf, before
returning to debugging tomorrow. Sorry my lack of ppc experience is
showing.

-bri