Hi !

On Tue, 22 Feb 2000, Todd Whitesel wrote:
> >don't get this cache running. It's different from KA660 cache enabling. But the
> >box runs fine, just a bit slow, and SCSI is missing on all 4000s (NCR5394).
> 
> When you attempt to enable the cache, does it simply not do anything, or does
> it do something bad (like crashing the machine) ?

Last tries were according to the KA660 TM, i enabled the primary cache the same
way as it's done there. The CPU is the same, most stuff is IPR-controlled, and
the only mapped I/O registers seem to show very useful values. The result:
If i enable the cache (also if i enable just one bank), the kernel crashes
very soon after mounting the root FS. Without cache, it runs stable. The same
init routine works very fine on the KA660, cache is definitely active there,
"make depend" runs faster than on my KA670, which should be faster than the
KA660. So the VLC cache is controlled a bit different, but i don't know this
difference. Nobody seems to have a TM for this box, and i don't want to pay
about 150-200$, this is the price, i saw some time ago for the TM.

> Is it at all possible for us to boot-trace VMS and figure out what the heck
> it does to turn the cache on?

On a newer PROM, this could be done fine with XDELTA, but the VLC doesn't have
such a superb command, and there are some CASE-instructions leading to machine-
dependent initialization. (Yes, i did XDELTA on VMS for the KA53 support).

> Are there any avenues for real hardware docs that we haven't tried yet?

Did someone try hardware docs ? Does someone _HAVE_ hardware docs ?

> Argh. There must be something we can do.

You're right. I did the main support about 1 year ago, but up to now, i didn't
find a way to get cache running.

...Michael

-- 
 Michael@camaronet.de, Rottweil/Germany, http://www.camaronet.de
 Visit the german VAXpage: http://www.vaxpower.de