On Thu, 28 Jan 1999, Eduardo E. Horvath wrote:

> On Wed, 27 Jan 1999, Alfred Perlstein wrote:
> 
> > On Wed, 27 Jan 1999, Eduardo E. Horvath wrote:
> > 
> > > This creates the mapping and leaves the virtual address on the stack.
> > > Then `cd' back to the zs node and type:
> > > 
> > > encode-in " address" property
> > 
> > ok cd /zs
> > ok encode-in " address" property
> > encode-in ?
> > ok encode " address" property
> > Fast Data Access MMU miss
> 
> Sorry, typo, that should read:
> 
> encode-int " address" property

hrm:

Initializing Memory |
ok showstack
ok cd /zs@f,1000000
ok 1fef1000000 sbus 4 memmap
fffae000 ok encode-int " address" property
ok cd /zs@f,1100000
ok 1fef1100000 sbus 4 memmap
fffac000 ok encode-int " address" property
ok .properties
port-a-rts-dtr-off
reg                      0000000f 01100000 00000004
interrupts               00000028
device_type              serial
name                     zs
ok             


it seems to allow it, but doesn't store the properties in the zs nodes...
and the kernel still panics

anything else to try?

well i guess i _could_ hardcode it into the kernel, but i was hoping to
avoid that... time to unpack some linux code and take a peek as well.

I think i may call the prom people tomorrow to ask them what the deal is.

The little i have learned so far (forth is nifty) puts PC bios's to shame
though, very impressive.

thank you,
-Alfred