I was 'stepping' through arm/arm/cpufunc.c last week
and noticed that it might panic even when it says
it can't; see below (guess the real question is
"what to do?").

So, how about fixing the code, or fixing the comment.
The latter might be easier.

The difference is quite old and is here:
http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/arch/arm/arm/cpufunc.c.d
iff?r1=1.36&r2=1.37&f=h

Luckily, my poor kernel doesn't get stuck here..
(It bombs elsewhere ;-).


/*
 * Cannot panic here as we may not have a console yet ...
 */

int
set_cpufuncs()
{
	cputype = cpufunc_id();
	cputype &= CPU_ID_CPU_MASK;

...

	/*
	 * Bzzzz. And the answer was ...
	 */
	panic("No support for this CPU type (%08x) in kernel", cputype);
	return(ARCHITECTURE_NOT_PRESENT);