Apple Network Server MacOS based ROMs found

[trag]

Hmm... what do we think of this?

I equipped two different DIMMs with those back around 2008. I had zero luck getting them to work.

That was the start of the PEx ROM project. James in Britain sent me a few sets of those sockets with the agreement that I would make him a PEx ROM DIMM, but I could never get the socketed version to work. That is part of why it took so long.

If you're already ordered them, I hope you have better luck.

Given the situation with the ROMs does anyone want DIMMs made at this point? I had a couple of requests; one on this topic and one in PM.

 

[mizerable]

I equipped two different DIMMs with those back around 2008. I had zero luck getting them to work.

That was the start of the PEx ROM project. James in Britain sent me a few sets of those sockets with the agreement that I would make him a PEx ROM DIMM, but I could never get the socketed version to work. That is part of why it took so long.

If you're already ordered them, I hope you have better luck.

Given the situation with the ROMs does anyone want DIMMs made at this point? I had a couple of requests; one on this topic and one in PM.

I did already order the ones on ebay, ill lyk if i have any sucess. i also am not particularly confident, especially having messed around with these chips already, but if it does work, itll be nice for when/if we can make a modified ROM image, im just kinda tired of soldering lol. i feel like, using a binder clip to get the appropriate force would prolly work, since thats almost def the issue...hmmm...i can figure this out, looks arnt important right?

Im personally for now good on a rom, i have all the parts coming to make another, hopefully i dont have issues with the board again, i cannot for the life of me to get the one i had to work, gave up, ordered a new board.

A lot went wrong with the second rom, 4 torn pads, worked tempararily but was buggy, desoldering the roms again to reflash got me nowhere, and even using the faster chips i used oringially didnt work.

Kinda at wits end, ordered 2 more rom sticks, gonna make one with the sockets on it, regardless of it working or not, and a normal one that hopefully works, and then im not making another, this shit hurts my brain idk how you do it man. doing 2 has killed me, but then again soldering and desoldering 10 times on the same board with no success blows.

 

[trag]

Kinda at wits end, ordered 2 more rom sticks, gonna make one with the sockets on it, regardless of it working or not, and a normal one that hopefully works, and then im not making another, this shit hurts my brain idk how you do it man. doing 2 has killed me, but then again soldering and desoldering 10 times on the same board with no success blows.

It helps when you get the reward of having them work. In my experience, reworking these boards more than once is usually pretty dicey. I don't understand it either, because on visual inspection, they look fine, but then they just don't work.

I assembled almost 200 of these when I was making and selling Beige G3 Rev. C ROMs back in the early 2000s...

Remember to relax your shoulders and don't tense up when you're hunched over. Being able to not hunch over is even better.

 

[mizerable]

It helps when you get the reward of having them work. In my experience, reworking these boards more than once is usually pretty dicey. I don't understand it either, because on visual inspection, they look fine, but then they just don't work.

I assembled almost 200 of these when I was making and selling Beige G3 Rev. C ROMs back in the early 2000s...

Remember to relax your shoulders and don't tense up when you're hunched over. Being able to not hunch over is even better.

I work standing, i dont think i could sit n solder, wouldnt hit me right...

Good to know that my decision that the board may just be bad, is based in truth, i was starting to go crazy not being able to find anything wrong...

Thanks man, will update with what does and doesnt work.

 

[Mr. Macintosh]

The work being done here is incredible!!! Thank you so much @trag for making these and @ClassicHasClass for the very very detailed testing! Plus @mizerable for additional testing.

The fact that we have come this far already is amazing. Yes both current rom sims have pluses and minuses, but something tells me with the brilliant work of @joevt and others, maybe one day they might be able to fix some of these things. Think about it, even last year we never thought that it would ever be possible to run macOS on an ANS, and here we are!

 

[trag]

So far I have a two requests for ROMs. I'm going to leave the request line open until 5/22/26 (Friday) and then fire up the soldering pencil.

It looks like there are four different ANS/MacOS ROMs in the .zip file that joevt kindly created.

https://tinkerdifferent.com/threads/apple-network-server-macos-based-roms-found.4756/post-46005

I've seen discussion/requests for 2.0 and 1.1.20. Is there some reason there is no interest in 1.1.22 or 2.26B6? I can't seem to sort it out in the discussion.

 

[ClassicHasClass]

1.1.22 is the production ROM, which definitely won't work.

2.2b6 could be interesting, but doesn't have many differences from the 2.2NT ROM which does not seem to boot MacOS either, at least not in my testing.

 

[trag]

1.1.22 is the production ROM, which definitely won't work.

2.2b6 could be interesting, but doesn't have many differences from the 2.2NT ROM which does not seem to boot MacOS either, at least not in my testing.

By production ROM, do you mean shipping ROM for machines intended to run AIX? Mr. Macintosh needs a plain AIX ROM for his ANS. He lacks even that.

Is 2.2b6 a MacOS ROM at all, then?

 

[ClassicHasClass]

No, I don't think any of the 2.2 ROMs will work for MacOS, only 2.0 and 1.1.20.1. Love to be proven wrong, but I couldn't get it to boot on 2.2 and the others are not significantly different.

Yes, 1.1.22 is the shipping ROM for booting AIX, which is what's in my production 500 (which boots AIX).

 

[eric]

Finally have time to look into this more and getting a ROM in the mail soon wanted to make a summary/validate my understanding of the current landscape of ANS roms - please let me know if anything in the tables below is incorrect or missing a detail and I'll update (and update the first post as well):

ANS ROM Versions Summary

ROM	OF Version	Boots Mac OS?	Boots AIX?	Internal Video	Internal SCSI (53C825)	MACE Ethernet	Notes
1.1.20.1	1.1.20.1	Yes (with caveats)	Yes	No	No	No	Pre-production / EVT. Best CPU perf — properly handles L2 cache & parity RAM. Requires PCI video card + external SCSI boot. Warns "MacOS is currently unsupported, use at your own risk."
1.1.22	1.1.22	No	Yes	n/a (AIX)	n/a (AIX)	n/a (AIX)	Production shipping ROM. <bye> removed from AAPL,ROM load word; open always returns false → blocks Mac OS boot. First 3 MiB identical to PM9500 v2.
2.0	2.0 (077d.7dd0.1)	Yes	No	Yes (.Display_Video_Cirrus_54M30 ndrv)	Yes (apple53C8xx ndrv, model NCR,825A)	Yes	The "MacOS ROM." Unique 3 MiB (not shared with TNT). Does NOT properly handle L2 cache (~75% slower CPU than 1.1.20.1), doesn't set RAM parity, bogus devaliases — appears unfinished. Adds MSR bit 16 clear in <bye>. Rhapsody starts but doesn't complete.
2.26B6	2.26	No	likely yes	No	No	No	Version-2 beta. First 3 MiB identical to PM9500 v2 / 1.0.5. Adds mouse device alias, pe-loader package, many minor additions. Broken open from 1.1.22 + normal load from 1.1.20.1. First copy received had errors in byte lane 2 in first 2 MiB.
2.26NT	2.26	No	?	No	No	No	Windows NT ROM. NT-specific OF words: init-nt-vars, maybe-read-nt-part (creates NVRAM section for NT). Adds not-pci-pci-alloc? flag and size word on fat-files/iso-9660-files. SETUPLDR loads but fails with CLAIM failed — missing HAL per Rairii. First 3 MiB matches 1.1.20.1/1.1.22; last 1 MiB very similar to 2.26B6.

Known ROMs

ROM	Version word	32-bit CSum	MD5
1.1.20.1	077d.28f2.1	962f6c13	23a7bbda5681be978972517d4758bbdd
1.1.22	077d.28f2.1	962f6c13	676809c236138574282fa8416c6c5a6d
2.0	077d.7dd0.1	49b2be8f	83be626679b5c74a606cd31443063a53
2.26B6	077d.28f2.1	9630c68b	d1d3720c38143eb2bac86b8393318529
2.26NT	077d.28f2.1	962f6c13	ad405e01c663340c479668c70f741f1b

 

[cc333]

This huge development almost makes me want to get an ANS so I can try it for myself.

Almost....

c

 

[joevt]

Finally have time to look into this more and getting a ROM in the mail soon wanted to make a summary/validate my understanding of the current landscape of ANS roms - please let me know if anything in the tables below is incorrect or missing a detail and I'll update (and update the first post as well):

ANS ROM Versions Summary

I think only 2.26B6 has first 3 MiB of 9500 v2 as indicated by the 32-bit checksum which covers only the first 3 MiB of the ROM. The ROMs include a 64-bit checksum that covers the entire 4 MiB.

1.0.5     077d.28f1.1  96cd923d  c241cd82bf90797a  dfebb8fdad4124e02608429d98bf349b   PM 7200 & 7500 & 8500 & 9500 (v1)

1.0.5     077d.28f2.1  9630c68b  4db4a42fea3b53b3  2623a0c438045ea04d2cc67310c97743   PM 7200 & 7500 & 8500 & 9500 (v2), SuperMac S900 (v2)
2.26B6    077d.28f2.1  9630c68b  a71fb907dd180b8a  d1d3720c38143eb2bac86b8393318529         

1.1.20.1  077d.28f2.1  962f6c13  c60da96de537f08a  23a7bbda5681be978972517d4758bbdd     
1.1.22    077d.28f2.1  962f6c13  d540b3dd5bcf9caa  676809c236138574282fa8416c6c5a6d      
2.26NT    077d.28f2.1  962f6c13  50348b3d0126096b  ad405e01c663340c479668c70f741f1b      

2.0       077d.7dd0.1  49b2be8f  5f2aeeb25507b2cb  83be626679b5c74a606cd31443063a53

 

[ClassicHasClass]

Okay, I have internal video working for Mac OS on 1.1.20.1 with the 2.0 ndrv. I created an extension shell for it in ResEdit and grafted the code fragment into the data fork with a hex editor, and it worked on the first try.

I'm now hacking the Startup Disk cdev to make the restarts better.

@joevt , are you able to find the MACE driver in the 2.0 ROM? I can find it in the DeclROM but I can't find a DRVR or ndrv for it, unless I've been skipping over it inadvertently.

 

[joevt]

Okay, I have internal video working for Mac OS on 1.1.20.1 with the 2.0 ndrv. I created an extension shell for it in ResEdit and grafted the code fragment into the data fork with a hex editor, and it worked on the first try.

Video ndrv is from the ndrv_506_53m30.pef file produced by tbxi dump.
Are you going to try the same for ndrv_-20424_SimNCR53c8xx.pef to support internal SCSI?

I'm now hacking the Startup Disk cdev to make the restarts better.

Create a OFpt resource with Open Firmware patches for 1.1.20.1. The first unused resource ID is 137.

Create a new entry in the OFtc resource for the ANS. For 1.1.20.1, the identifier for the ANS is AAPL,ShinerESB ? The entry should include the new OFpt resource ID.

If you're not making a new OFpt, then you don't need to add AAPL,ShinerESB to OFtc.

If you have an ANS that uses one of the existing identifiers (AAPL,9500 for example), then you have to change the corresponding OFpt to include code that checks for the TNT version of Open Firmware before doing any of the patches which are probably not applicable to the ANS.

I wonder what the OFtb resource is for?

derez -only 'OFtb' "/Volumes/Classic/System Folder/Control Panels/Startup Disk"
data 'OFtb' (128) {
	$"006E 0080 0196 FFFF"                                /* .n.Ä.ñˇˇ */
};

derez -only 'OFtc' "/Volumes/Classic/System Folder/Control Panels/Startup Disk" | sed -nE 's/\$"([0-9A-F ]+).*/\1/p' | xxd -p -r | xxd -c 38
00000000: 4141 504c 2c37 3330 3000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0080  AAPL,7300.............................
00000026: 4141 504c 2c37 3530 3000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0080  AAPL,7500.............................
0000004c: 4141 504c 2c38 3530 3000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0080  AAPL,8500.............................
00000072: 4141 504c 2c39 3530 3000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0080  AAPL,9500.............................
00000098: 4141 504c 2c39 3730 3000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0081  AAPL,9700.............................
000000be: 4141 504c 2c47 6f73 7361 6d65 722c 3531 3000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0082  AAPL,Gossamer,510.....................
000000e4: 4141 504c 2c50 6f77 6572 4d61 6320 4733 2c35 3130 0000 0000 0000 0000 0000 0000 0000 0000 0083  AAPL,PowerMac G3,510..................
0000010a: 4141 504c 2c50 6f77 6572 426f 6f6b 3139 3938 2c33 3132 0000 0000 0000 0000 0000 0000 0000 0084  AAPL,PowerBook1998,312................
00000130: 4141 504c 2c50 6f77 6572 426f 6f6b 3139 3938 2c33 3134 0000 0000 0000 0000 0000 0000 0000 0085  AAPL,PowerBook1998,314................
00000156: 4141 504c 2c33 3430 302f 3234 3030 2c33 3036 0000 0000 0000 0000 0000 0000 0000 0000 0000 0086  AAPL,3400/2400,306....................
0000017c: 4141 504c 2c33 3530 302c 3331 3300 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0087  AAPL,3500,313.........................
000001a2: 4141 504c 2c65 3430 3700 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0088  AAPL,e407.............................

derez -p -only xxx "/Volumes/Classic/System Folder/Control Panels/Startup Disk" 2>&1 | grep OFpt | sed -E 's/.*Skipping //'
'OFpt' (128, "PowerSurge").
'OFpt' (129, "PowerExpress").
'OFpt' (130, "Gossamer").
'OFpt' (131, "PowerMac G3").
'OFpt' (132, "Wallstreet").
'OFpt' (133, "Mainstreet").
'OFpt' (134, "Hooper").
'OFpt' (135, "Kanga").
'OFpt' (136, "Alchemy").

@joevt , are you able to find the MACE driver in the 2.0 ROM? I can find it in the DeclROM but I can't find a DRVR or ndrv for it, unless I've been skipping over it inadvertently.

MACE is sound compression.
mace is Ethernet controller.
I think you're talking about Ethernet controller.

DeclData of Power Mac ROMs is not very large (0x130 bytes).

FHeaderRec:
FEFFFFEC
       fhDirOffset: 00 FFFFF4 = FEFFFFE0
          fhLength: 304 ; Start of ROM: FEFFFED0
             fhCRC: 341BACA4 = ok
          fhROMRev: 1 = romRevision
          fhFormat: 1 = appleFormat
          fhTstPat: 5A932BC7 = testPattern
        fhReserved: 00
       fhByteLanes: 0F = byte lanes: 0,1,2,3

Slot Resource Directory from Header:
FEFFFFE0
  0    3E    FFFFEC FEFFFFCC                   
	  0    01    FFFFF8 FEFFFFC4 ; sRsrcType       0001=CatBoard 0000=TypBoard 0000=DrSWBoard 0000=DrHWBoard
	  1    02    FFFFE6 FEFFFFB6 ; sRsrcName       "Macintosh 4A"
	  2    20    000670          ; boardId         0670 = TNTBoardID
	  3    24    FFFFCA FEFFFFA2 ; vendorInfo      
		  0    01    FFFFBC FEFFFF5E ; vendorID        "Copyright © 1986-1995 by Apple Computer, Inc.  All Rights Reserved."
		  1    03    FFFF9E FEFFFF44 ; revLevel        "Macintosh CPU Family 6.0"
		  2    04    FFFF84 FEFFFF2E ; partNum         "DeclROM for OpenTxpt"
		  3    05    FFFF64 FEFFFF12 ; date            "Wednesday, January 14, 1998"
		  4    FF    000000 FEFFFFB2 ; endOfList       
	  4    FF    000000 FEFFFFDC ; endOfList       
  1    FD    FFFF1A FEFFFEFE                   
	  0    01    FFFFF8 FEFFFEF6 ; sRsrcType       0004=CatNetwork 0001=TypEthernet 0001=DrSWApple 0004=DrHWproductC
	  1    02    FFFFCE FEFFFED0 ; sRsrcName       "Network_Ethernet_Apple_OpenTransport"
	  2    07    000000          ; sRsrcFlags      0000 2:f32BitMode=0  1:fOpenAtStart=0
	  3    08    000002          ; sRsrcHWDevId    02
	  4    FF    000000 FEFFFF0E ; endOfList       
  2    FF    000000 FEFFFFE8 ; endOfList       

minAddr = FEFFFED0   maxAddr = FEFFFFE8

I'm not sure what this Network_Ethernet_Apple_OpenTransport slot resource is trying to convey. It's just a slot resource device id, same as in in the 9500 v2 ROM.

Looking at the DumpNameRegistry output for my Power Mac 8600, it shows there is a ndrv:
https://68kmla.org/bb/threads/skipping-the-startup-memory-test.50699/post-570896

/Devices/device-tree/bandit/gc/mace
    name                            mace                                                                       6D616365 00
    device_type                     network                                                                    6E657477 6F726B00
    AAPL,connector                  ethernet                                                                   65746865 726E6574 00
    reg                             00011000 00001000 00008200 00000100 00008300 00000100                      "..........Ç.......É....."
    AAPL,interrupts                 0000000E 00000002 00000003
    local-mac-address               000502A8 066B                                                              "...®.k"
    address-bits                    00000030                                                                   "...0"
    max-frame-size                  00000800
    AAPL,address                    F3011000 F3008200 F3008300                                                 "Û...Û.Ç.Û.É."
    driver-ist                      0008ECC0 00000001 0008EE60 00000003 0008EE60 00000004                      "..Ï¿......Ó`......Ó`...."
    driver,AAPL,MacOS,PowerPC       4A6F7921 70656666 70777063 00000001 B624A1BB 00000000 00000000 00000000    "Joy!peffpwpc....∂$°ª............"
                                    00030002 00000000 FFFFFFFF 00000000 00014FD4 00014FD4 00014FD4 00000080    "........ˇˇˇˇ......O‘..O‘..O‘...Ä"
                                    00010200 FFFFFFFF 00000000 00004298 000040F0 0000392E 00015060 02010300    "....ˇˇˇˇ......Bò..@..9...P`...."
                                    FFFFFFFF 00000000 00000000 00000000 000007CA 00018990 04010000 00000000    "ˇˇˇˇ............... ..âê........"
                                    7C0802A6 90010008 9421FFC0 39823F10 7CE83B78 7CC73378 7CA62B78 7C852378    "|..¶ê...î!ˇ¿9Ç?.|Ë;x|«3x|¶+x|Ö#x"
                                    7C641B78 806C0000 48013B05 60000000 38210040 80010008 7C0803A6 4E800020    "|d.xÄl..H.;.`...8!.@Ä...|..¶NÄ. "
                                    8062FF2C 4E800020 7C0802A6 93E1FFFC 90010008 9421FFC0 3BE3FFFF 8122FF28    "Äbˇ,NÄ. |..¶ì·ˇ¸ê...î!ˇ¿;„ˇˇÅ"ˇ("
                                    281F0008 38820300 41810080 57E5103A 7CA4282E 7CA903A6 4E800420 80690000    "(...8Ç..AÅ.ÄWÂ.:|§(.|©.¶NÄ. Äi.."
                                    2C030000 41820064 38800000 48014021 60000000 48000054 80690000 2C030000    ",...AÇ.d8Ä..H.@!`...H..TÄi..,..."
                                    41820048 38800000 48014005 60000000 48000038 80690000 2C030000 4182002C    "AÇ.H8Ä..H.@.`...H..8Äi..,...AÇ.,"
                                    38800001 48013FE9 60000000 4800001C 80690000 2C030000 41820010 38800001    "8Ä..H.?È`...H...Äi..,...AÇ..8Ä.."
                                    48013FCD 60000000 38600000 38210040 80010008 7C0803A6 83E1FFFC 4E800020    "H.?Õ`...8`..8!.@Ä...|..¶É·ˇ¸NÄ. "
                                    38823F10 80640000 2C030000 4D820020 7C0802A6 93A1FFF4 93C1FFF8 93E1FFFC    "8Ç?.Äd..,...MÇ. |..¶ì°ˇÙì¡ˇ¯ì·ˇ¸"
                                    90010008 9421FFB0 3BC23F28 7C9F2378 801E0006 2C000000 4182001C 3BA00000    "ê...î!ˇ∞;¬?(|ü#xÄ...,...AÇ..;†.."
                                    7FC3F378 480146FD 80410014 93BE0006 807F0000 48014011 60000000 907F0000    ".√ÛxH.F˝ÄA..ìæ..Ä...H.@.`...ê..."
                                    2C030000 4182002C 38820348 38600000 480063DD 60000000 38820348 38600000    ",...AÇ.,8Ç.H8`..H.c›`...8Ç.H8`.." … total size = 102746
    driver-descriptor               6D74656A 00000000 046D6163 65000000 00000000 00000000 00000000 00000000    "mtej.....mace..................."
                                    00000000 00000000 02306002 00000004 04656E65 74000000 00000000 00000000    ".........0`......enet..........."
                                    00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
                                    00000000 00000000 00000000 00000000 00000001 6F74616E 000A0B01 02306002    "....................otan.....0`."
    driver-ptr                      00B0B880                                                                   ".∞∏Ä"

However, this is not a driver that exists in the ROM. Therefore it is in the System file or in an extension.
B624A1BB is a time stamp.

Searching my Classic System Folder, I find a mace driver with a later timestamp B8C77E0E. I guess this means my DumpNameRegistry example is from an earlier OS.

grep -r mace "/Volumes/Classic/System Folder"
Binary file /Volumes/Classic/System Folder/Extensions/Apple Enet matches
Binary file /Volumes/Classic/System Folder/Extensions/Microsoft Internet Library matches

mpw DumpPEF -a -do All -pi u -format on -ntb -dialect PPC32 "/Volumes/Classic/System Folder/Extensions/Apple Enet" > "/tmp/Apple Enet.dumppef.txt"

grep "member" "/tmp/Apple Enet.dumppef.txt"
  { /* array memberArray: 7 elements */
		"mace",                                                 /* member name               */
		"bmac",                                                 /* member name               */
		"uprGMACLib",                                           /* member name               */
		"uprdecLib",                                            /* member name               */
		"pci1011,14",                                           /* member name               */
		"pci1011,9",                                            /* member name               */
		"uprBMAC+Lib"                                           /* member name               */

grep "dateTimeStamp" "/tmp/Apple Enet.dumppef.txt"
00010 dateTimeStamp    = 0xB8C77E0E (Wed Mar 27 05:47:58 2002) date/time stamp
00010 dateTimeStamp    = 0xB1F5C96A (Tue Aug 11 03:11:54 1998) date/time stamp
00010 dateTimeStamp    = 0xB8C77E16 (Wed Mar 27 05:48:06 2002) date/time stamp
00010 dateTimeStamp    = 0xB8C77E1C (Wed Mar 27 05:48:12 2002) date/time stamp
00010 dateTimeStamp    = 0xB8C77E2F (Wed Mar 27 05:48:31 2002) date/time stamp
00010 dateTimeStamp    = 0xB8C77E42 (Wed Mar 27 05:48:50 2002) date/time stamp
00010 dateTimeStamp    = 0xB8C77E49 (Wed Mar 27 05:48:57 2002) date/time stamp

Anyway, do a DumpNameRegistry for your ANS to see what's different. Does it show a gc/mace device?

One difference between TNT and Shiner in Open Firmware is that init-mace for ANS stores a 1 in the PLSCC (PLS Config Control) register instead of a 7. I don't know what that means. DingusPPC has this warning message:

void MaceController::write(uint8_t reg_offset, uint8_t value)
...
    case MaceReg::PLS_Config_Ctrl:
        if (value != 7)
            LOG_F(WARNING, "%s: unsupported transceiver interface 0x%X in PLSCC",
                  this->name.c_str(), value);
        break;

init-mace only happens when you try to use mace in Open Firmware (for tftp which is for net booting?). Therefore, patching Open Firmware to change init-mace is not going to help anything.

I wonder if the mace ndrv is storing 1 or 7 to PLSCC ?

The DingusPPC source code says mace is related to Am79C940. The Am79C940 Datasheet says bit 2 and 1 of the PLSCC register are PORTSEL. The ANS is selecting the default value 0b00 for AUI. 0b01 is for 10Base-T. 0b10 is for DAI Port. TNT machines use 0b11 for GPSI.

 

[ClassicHasClass]

Yes, I'll do the SCSI driver as well, though that should be easy using the same process as the video ndrv and it doesn't remove the external SCSI boot requirement, so I'm going to do that last. For me getting the Ethernet working is more important.

The video extension is based on the 53m30 ndrv, but I note for the record that the name appears to be a typo (it's a 54M30 and is detected as such).

There is a /bandit/gc/mace visible in Open Firmware. Where did you find the offset for that code fragment?

Why would your DumpNameRegistry tool do anything with the ROM? Isn't that for NVRAM?

As far as Startup Disk, I'm not convinced that OFtc is doing any matching, or at least in the manner hypothesised here. Since the ANS with the preproduction ROMs should come up as a 9500, I monkeypatched the AAPL,9500 string to AAPL,XX00 and I still got the PowerSurge script. I could patch them all but I'm thinking I'll just rewrite the PowerSurge script. All the scripts are apparently doing is setting Open Firmware variables.

Still, the string `ttya` appears nowhere in Startup Disk, so something else must be doing that. The way around it might be an nvramrc that resets input-device and output-device but that seems inelegant.