nigel@ind.tansu.com.au wrote:
> 
> 	For the curious (or just in case I get hit by a bus or something)
> the problem seems to be in a function TurnCachesOff() in ufs_test.c.
> 
> 	Now, Booter 1.9.3b3 had an assembler version of TurnCachesOff().
> Booter 1.9.4 had both the assembler and a new C version of this function,
> selected by a #if block. Booter 1.9.6 and later had just the assembler
> one again (this was about the time Brian Gaeke stopped maintaining the
> Booter source code).
> 
> 	The really strange part:
> 
> 1) A few people reported that Booters _earlier_ than 1.9.4 also worked.
> 
> 2) Why would the assembler code not work [stop working?] _just_ on the
>    IIvx machines? They are not the only machines with an '030 and FPU.
> 
>   	Anyway, for your edification, here is the code in question -
> from Booter 1.9.4's ufs_test.c:

Not that I doubt your assessment of the problem in any way, but why does
turning the caches off via any method hang the machine in the booter?  For
that matter, I thought that NetBSD turned the caches off very early in
locore.s....is it necessary to turn them off in the booter as well?  Just
trying to figure this out... ;-)

Thanks for getting this working!

Later.

-- 
Colin Wood                                 cwood@ichips.intel.com
Component Design Engineer - MD6                 Intel Corporation
-----------------------------------------------------------------
I speak only on my own behalf, not for my employer.