I hope in the attached archive you find the information you need in this case.
Here is, what i typed to get this dumped:
Code:
0 > : dumpbytes ( start size ) ok
0 ] cr bounds ?do i 3f and 0= if cr then i c@ 2 u.r loop cr ; ok
0 > @startvec dup here - negate dumpbytes
I assume @startvec is ff808000. I'll see if I can disassemble it properly.
It's somehow "a little" bigger than expected. Hope, it is the right thing anyway.
The dictionary is where all the words and variables and constants and buffers and device nodes and properties are stored. All the fcode that's in the rom is compiled into those bits in the dictionary. What the dictionary doesn't have is the fcode that created those bits.
Example fcode:
Code:
12 value myvalue
: mycolon ." hello" ;
myword
The dictionary includes the compiled myvalue and mycolon but it does not include the fcode above that created them or the call to myword.
If this is correct, then i can try for the G3 later too.
I don't think we need the dictionary for G3. The G3 rom is easy to extract fcode from. The dictionary would include device node stuff but that's mostly available from dump-device-tree... The dictionary would help to make patches (alternative to the dumping bytes for individual words like we did for patching
initscc) but we don't need to make patches.
the card did not give any signal to the display and was not "seen" by System Profiler present as pci-card.
Is this different than previous behaviour? The point of this rom was to log stuff, not change behaviour.
Did this and it seems to have worked so far. Here is, what "probe-all" comes up with:
I think I can use this info to make my emulation get further.
Code:
w_0_16 @00000004.b = 1000000
w_0_16 @00001850.b = 1000000
Why would a value ≥ 0x100 be written to a byte? I need to check my logging to make sure I put the correct letter "b,w,l" for the size...
Code:
r_0_16 @00600804.l = 0
w_0_16 @00600804.b = 80
w_0_16 @00600860.b = 0
ok
I wonder why it stopped without an abort message...
In the future you'll want to put log output into a file since it will get very large.
In fact, when installed, it startet to log read/write transactions to OF just like in the new World Mac.
We definitely need info from working New World Mac case for compare. So we need to work on a method.
Yes, the Sawtooth has a Stealth Serial Port. This is how i connect to the Beige.
Is it usable by Open Firmware? It might be easier to use than telnet. We wouldn't need to do
pci-probe-mask stuff.
Try this:
scca io
I think the default is 57600 baud.
If it works, then maybe we can do these:
setenv input-device scca
setenv output-device scca
But it depends on if New World Mac has the
(install-console) command to enable logging before probe-all in nvramrc (before nvramrc finishes).
Before you put
(install-console) in
nvramrc of sawtooth to enable logging to serial device before nvramrc finishes then we need to make sure it exists as a word:
see (install-console)
see install-console
If you see something like this:
Code:
0 > see (install-console) (install-console), huh??
ok
0 > see install-console defer install-console
: ^ff822bc0
" open KEYBOARD" time-stamp ^ff811d40 0= if
^ff811030 open-dev to ^ff811d40
then you need to use
ff822bc0 execute (or whatever address/execution token it says
install-console is deferring to) instead of
(install-console) in the nvramrc.
