Re: Immedate panic when booting on a 14" iBook G4 1.33GHz

[cjb%brushtail.apana.org.au@localhost (Chris Baird)]

 > This looks like we crashed somewhere in
 > arch/powerpc/oea/oea_machdep.c::oea_batinit(), in or shortly after the
 > call to mem_regions() ( which lives on powerpc/ofw_machdep.c ). I'd
 > sprinkle printf("%s %d\n", __func__, __LINE__); to narrow down where
 > exactly the crash happens. Also set options DEBUG for extra console
 > spam.

Tracing it that way finds it breaking in an openfirmware() call during
set_timebase() ..

mem_regions 202
ofwoea_initppc 296
Loading initial symtab....
set_timebase 325
OF_peer 75

/sys/arch/powerpc/powerpc/openfirm.c:OF_peer()
	[...]
	ofw_stack();
	args.phandle = phandle;
 printf("%s %d\n", __func__, __LINE__);  // line 75
	if (openfirmware(&args) == -1)
	   return 0;
 printf("%s %d\n", __func__, __LINE__);
	return args.sibling;

(where &args would be {const char*="peer", int nargs=1, int
nreturns=1, int phandle=0, int sibling=xx} )

So... it's bought the farm during a real-mode OFW call to find the
root of the system's .properties, yet I would assume there's been a
few hundred of those already by that stage of the boot (writing to
console, loading from HDD, etc.)

--
Chris Baird,, <cjb%brushtail.apana.org.au@localhost>