68kMLA Classic Interface

dev /packages/bootpath-search
: mypatch { ; loc0 loc1 loc2 loc3 loc4 } loc4 6 = if loc0 findandadd-block-usb then ;
' findandadd-block 3ac + ' mypatch c + brpatch

For the boot option, I would like to get linux working if that's possible. About the firmware, I probably just read it wrong

I think I found a command that should work for any operating system:

probe-usb boot usb<port ID>/disk:<partition number>,\\:tbxi

I found it here: https://forums.macrumors.com/threads/guide-new-method-booting-from-usb-on-powerpc-macs.2403368/

4.2.8f1, 4.5.7f1, and 4.6.0f1 should all be able to boot USB.

Of the firmwares I have, the ones that support USB mass storage but don't scan for USB are built on or after 09/23/04.
09/23/04 4.8.7f1 iBook G4
03/23/05 4.8.9f4 Mac mini G4
07/12/05 4.9.4f1 Mac mini G4
09/22/05 4.9.5f3 PowerBook G4 DLSD
09/30/05 5.2.7f1 G5 Quad

Huh....so 4.2.8f1, 4.5.7f1, and 4.6.0f1 do scan for USB at boot and present it in the picker? I swear I just read another thread where people were saying they saw USB in the picker on pre Mac mini, but maybe I'm imagining things.

• 08/19/99 1.2f2 : The latest ROM that I have that can't boot USB is from a Yikes G4. The bootpath-search package is similar to later versions but doesn't include code for USB and FireWire.
• 06/13/00 3.0f3 : The first ROM that I don't have that is known to boot USB is from a iMac DV+.
• 03/20/01 4.1.7f4 : The earliest ROM I have that can boot USB is from an iBook G3.
  The bootpath-search package hasn't changed much since this version.
• 09/23/04 4.8.7f1 : The first ROM I have that can boot USB but doesn't include USB in the boot menu because a couple lines of code were removed which my patch restores.

06/13/00 3.0f3 : The first ROM that I don't have that is known to boot USB is from a iMac DV+.

I think I found a command that should work for any operating system:

probe-usb boot usb<port ID>/disk:<partition number>,\\:tbxi

I found it here: https://forums.macrumors.com/threads/guide-new-method-booting-from-usb-on-powerpc-macs.2403368/

Not all linux in stalls use a :tbxi as the boot loader.

Most of the later ones use an EFL binary( grub.elf ).

I wish I could make heads or tails of Forth yet...but...

Trying this out for a mini, PowerMac10,1. Doesn't crash but doesn't seem to probe USB at all with the G5 patch. How did you discover the patch values, basically detoking the firmware?

I imagine the procedure is to dump the firmware, detok, and from that determine how to inject findandadd-block-usb again?

Along the way, learning how to understand Forth I suppose...

What firmware version does your Mac Mini have? You can dump the ROM using the info at #2

My DumpMacRom.sh script disassembles the Open Firmware assembly part of ROM, and it detokenizes and converts to Forth the fcode parts of ROM.

For making patches, you usually patch the assembled Forth words, rather than the originating fcode. So for this I dump the entire Open Firmware dictionary where the assembled forth words exist in RAM and use DumpMacRom.sh to disassemble that.

There's notes at
https://forums.macrumors.com/thread...l-work-in-a-beige-power-macintosh-g3.2303689/

Open Firmware documentation exists at
https://openfirmware.info/Welcome_to_OpenBIOS
https://openfirmware.info/IEEE_1275-1994 <-- get this one
https://openfirmware.info/Bindings <-- get these if you're working with a particular device type
https://openfirmware.info/On_the_Net

0 > : mypatch { ; loc0 loc1 loc2 loc3 loc4 } loc4 6 = if loc0 findandadd-block-usb then ;  ok
0 > xsee mypatch 
ffa3b9e8: 967efffc 
ffa3b9ec: 7e6802a6 
ffa3b9f0: 4be0b859 {0}                   \ {0} means don't pop any parameters from the stack - doesn't matter because the patch skips this
ffa3b9f4: 4be0b8f5 {4@}                  \ branch from findandadd-block goes to here; we want to examine the fifth local variable
ffa3b9f8: 4be11cb9 6
ffa3b9fc: 4be0e1f5 = 
ffa3ba00: 4be0b629 (b?branch) 
ffa3ba04: 4800000c 
ffa3ba08: 4be0b8a1 {0@} 
ffa3ba0c: 4bf0b1bd findandadd-block-usb 
ffa3ba10: 4be0aa90                       \ this exit takes the place of the exit that we patched in findandadd-block

findandadd-block appears to be the same in 4.8.9f4 (Mac mini G4) and 5.2.7f1 (Quad G5).

I suppose we might need to look at the assembly code to make sure they compile the same.