Predator99 wrote on 2020-01-17, 17:49:

Yes, very easy? Its a ROM in a ROM. Look at starting adress 50000h. The SCSI-ROM on the Mainboard starts with "55 AA". 3rd byte is "40". Thats the length of the ROM *512. Thats 64d * 512 = 32kb.

The Adaptec-standalone-ROM starts with "55 AA", then "46". Therefore its bigger in size.

You have to check if there is some space left in the 512kb mainboard BIOS to fit the Adaptec ROM in it...

That looks like a nice machine, could anyone of you try Beos 5 on it?
The Personal Edition should support the dual cpu as long as you boot it with a boot floppy and not the Win9x shortcut.

Predator99 wrote on 2020-01-17, 17:50:

Predator99 wrote on 2020-01-17, 17:49:

Yes, very easy? Its a ROM in a ROM. Look at starting adress 50000h. The SCSI-ROM on the Mainboard starts with "55 AA". 3rd byte is "40". Thats the length of the ROM *512. Thats 64d * 512 = 32kb.

The Adaptec-standalone-ROM starts with "55 AA", then "46". Therefore its bigger in size.

You have to check if there is some space left in the 512kb mainboard BIOS to fit the Adaptec ROM in it...

I have taken a look at the structure of that 512kb ROM:

00000-03F00: emtpy
03F00-05F00: 55 AA 10 - 2000 bytes ELOG?
05F00-40000: empty
40000-48000: 55 AA 40 - 8000 bytes VGA
48000-4D000: 55 AA 28 - 5000 bytes ?
50000-58000: 55 AA 40 - 8000 bytes SCSI
58000-80000 Main BIOS?

I assume: The Main BIOS should stay where it is. The other parts you can shift around as you like. Try to fill the hole at 5F00 with something, maybe XT-IDE to test..
There is also enough space to put other ROMs in it. Play with it and report if it works 😉

Predator99. A few things you have not been told: There are 2 scsi controllers (AIC-7880 UW and AIC-6370) on the mainboard, each has it's own op-rom inside main rom. I have been studying the bios for about a week and can tell you that 048000 starts the AIC-7880 bios [5755AA28E9C82800 / WU.(..(. ] and ends at 04C531 and is compressed. The AIC-6370 is what starts at 050000 and is not compressed. Pulling the 7880 code out shows about 18k, pulling the 6370 code shows 32k.

1total bios file size: 
207ffff hex or 524287 bytes
3
4half that is 03ffff or 26243 and all zero's (00) from start to 04FFFF except 
5for 03F00 thru 05F00 for 8k for EISA ?
6
7040000 : CL VGA bios
855AA40E9C02D000000000000000000000000000000000
9U.@..-......................~2IBM VGA Compat
10ends at 047DB3
11
12048000: 7880 bios
135755AA28E9C828
14WU.(..(.
15ends at 04C531
16
17050000: 6370 bios
1855AA40E9CF0090E9
19U.@.....&UAdaptec.AIC6370.`...........P
20ends at 55555
21
22061ACE: main bios
23BC.......p`....\...f..J.w.v.4.04 Cop.yright 1.985-94 P.hoenix
24ends at 07BFFF
25
2607E05B: bootblock 
27H...BCPFCP.........07/02/96.@................t.. ...
28to end 

String starting at: 04BE40 (near end of 048000 - 04C531 code)
h: 41646170746563204149432D3738383020556C7472612F556C74726120572042494F532076312E3234
a: Adaptec AIC-7880 Ultra/Ultra W BIOS v1.24

OK, didnt read the whole thread 😉

What happens if you shift these segments around inside the file? Are they still executed?

Predator99, I looked through my network cards and have the following with boot ROM sockets,

EISA - 3c597 = DIP32 (currently installed in the Proserva)
ISA - IBM 100/10 ISA = DIP28
PCI - 3c905B-TX = DIP32
PCI - DFE-538TX = DIP28
PCI - DFE-530TX = PLC32

I have EEPROM's in 0.25 mbit, 0.5 mbit, 1 mbit, 2 mbit, and 4 mbit sizes. The Adaptec v1.34 BIOS is 0.5 mbit in size.

What you are saying is that I need to write the Adaptec v1.34 BIOS to the an EEPROM, install that EEPROM in a network card, ensure that the boot rom feature of the network card is enabled, disable the motherboard's SCSI BIOS, then boot the motherboard and the onboard Adaptec SCSI controller will now use the BIOS on the network card? Please correct me if I have misunderstood anything.

Horun, I was able to install NT4 w/SP6 and it shows 2 CPUs in Task Manager. Will attempt to re-install W2K to ensure 2 CPUs are identified. EDIT: Installed W2K full version (not upgrade) and it still only shows 1 CPU in Task Manager. Present on the HDD in other partitions were NT4 and Win98SE. Could installing W2K past the 8 GB barrier have something to do with the missing CPU? Or Win98SE being present? Both seem unlikely. You are using BIOS 1.00.14.WL014 on your system right? Really puzzled as to what I am doing wrong.

EDIT: What I have not had success with is getting the AIC-6370 controller to properly boot any bootable CD disc with any SCSI CD-ROM drive. Have you had success with this? I've been having to use the NT4 and W2K installation floppy diskettes.

Predator99 wrote on 2020-01-18, 08:41:

OK, didnt read the whole thread 😉

What happens if you shift these segments around inside the file? Are they still executed?

Have not tried shifting any thing inside the rom bin. Only concern is that the main BIOS would be decompressing the 7880 op-rom and may have a pointer to its location, where the VGA and 6370 op-roms are already non-compressed so would not need any extra handling. Am going to re-read the "Unlock Hidden menu in Phoenix BIOS Setup Menu Tutorial.pdf" and a couple other BIOS manuals before doing any editing. Just wish the eeprom was not a flat soldered SMC chip.

feipoa wrote on 2020-01-18, 11:57:

Horun, I was able to install NT4 w/SP6 and it shows 2 CPUs in Task Manager. Will attempt to re-install W2K to ensure 2 CPUs are identified. EDIT: Installed W2K full version (not upgrade) and it still only shows 1 CPU in Task Manager. Present on the HDD in other partitions were NT4 and Win98SE. Could installing W2K past the 8 GB barrier have something to do with the missing CPU? Or Win98SE being present? Both seem unlikely. You are using BIOS 1.00.14.WL014 on your system right? Really puzzled as to what I am doing wrong.

EDIT: What I have not had success with is getting the AIC-6370 controller to properly boot any bootable CD disc with any SCSI CD-ROM drive. Have you had success with this? I've been having to use the NT4 and W2K installation floppy diskettes.

Also have issues booting off any cdrom and cd disk. Tried both Win98se, Win2000, NT4 and none booted properly. Win2k booted but got an error after loading NTLDR. the other gave non-bootable errors.
I ended up installing Win2k from Win98se to a partition below the 8Gb barrier, that way I have a boot menu on C: to choose either Win2k or Win98.
Actually this board is running bios x08, the flakey board was upgraded to x014....

All my Adaptec PCI SCSI adapters' BIOSes are packed as well. I don't believe they are unpacked by any external hardware though, just self unpacking at runtime.

maxtherabbit wrote on 2020-01-18, 16:26:

All my Adaptec PCI SCSI adapters' BIOSes are packed as well. I don't believe they are unpacked by any external hardware though, just self unpacking at runtime.

Agree. As they start with the "55AA" they are standalone. You can replace them as you like, no need to care about the compression.

If the Eeprom is soldered I would be very careful indeed. If you flash something wrong you wont be able to start the PC anymore.

feipoa wrote on 2020-01-18, 11:57:

What you are saying is that I need to write the Adaptec v1.34 BIOS to the an EEPROM, install that EEPROM in a network card, ensure that the boot rom feature of the network card is enabled, disable the motherboard's SCSI BIOS, then boot the motherboard and the onboard Adaptec SCSI controller will now use the BIOS on the network card? Please correct me if I have misunderstood anything.

Thats exactly what I propose and there is no risk by doing so. If this doesnt work the 1.34 will not be suitable anyway...

Predator99 wrote on 2020-01-18, 16:53:

Agree. As they start with the "55AA" they are standalone. You can replace them as you like, no need to care about the compressio […]
Show full quote

maxtherabbit wrote on 2020-01-18, 16:26:

All my Adaptec PCI SCSI adapters' BIOSes are packed as well. I don't believe they are unpacked by any external hardware though, just self unpacking at runtime.

Agree. As they start with the "55AA" they are standalone. You can replace them as you like, no need to care about the compression.

If the Eeprom is soldered I would be very careful indeed. If you flash something wrong you wont be able to start the PC anymore.

feipoa wrote on 2020-01-18, 11:57:

What you are saying is that I need to write the Adaptec v1.34 BIOS to the an EEPROM, install that EEPROM in a network card, ensure that the boot rom feature of the network card is enabled, disable the motherboard's SCSI BIOS, then boot the motherboard and the onboard Adaptec SCSI controller will now use the BIOS on the network card? Please correct me if I have misunderstood anything.

Thats exactly what I propose and there is no risk by doing so. If this doesnt work the 1.34 will not be suitable anyway...

The only way to disable the onboard scsi bios is thru the EISA config tool, there are no jumpers to disable anything. I can confirm that a standard 2940uw with BIOS 1.34 works just fine and the onboard does not load it's rom as long as no HD's are connected to it. I will try to load a good eeprom with v1.34 and put it on a PCI NIC and see what happens, have a bunch with 28pin and 32pin empty sockets. May have to order some sst27SF512 if I can't find a good one in the pile.

Replying to feipoa's PM.

Predator99's idea of using a separate card to load a SCSI option ROM seems possible, if you'll manage to properly execute it from a NIC, then it'd be the safest method of BIOS "patch".

Horun is right about the location of the AIC-7880's option ROM image, except it starts precisely at 48000H and ends at 4FFFFH (32KB).

The two latest firmware updates for the standalone 2940UW card - ver. 1.34.1 and 2.20.0 (36KB and 41KB respectively) are already larger than the reserved space for AIC-7880's option ROM in the NEC's BIOS, thus it wouldn't be such a simple task to patch BIOS with larger images, using only a HEX editor w/o messing it up completely (no Phoenix BIOS editor that I found, could recognise NEC's BIOS). It's possible to relocate other option ROMs to different addresses and make enough room for larger AIC-7880's images, but I wouldn't reccomend experimenting with a motherboard in a way which could easily brick it.

I took a slightly different approach of patching system BIOS, instead of risking to corrupt original BIOS image with non-standard sized ROMs, I had extracted AIC-7880's option ROM from Asus P2L97-DS mainboard's BIOS, ver. 1001. It's exactly 32KB in size and seems to be of 1.32 version (according to the option ROM's name), though I'm not entirely sure if that's the case, there's no mention of its version anywhere in the image. New option ROM was inserted into 48000-50000H 48000-4FFFFH range, 8 bit checksum is 0.

Attaching modded NEC Proserva V BIOS (ver. 1.00.14) with some extracted option ROMs for AIC-7880.

EXPERIMENT AT YOUR OWN RISK.

*Edit: next option ROM starts at 50000H, it's just a typo, BIOS is patched correctly 😉

Thanks ! if I can get the spare board to boot proper will flash it with your bios. It hangs in POST since just after I got it and can't figure out why yet...

SSTV2, thanks a lot for your clever solution. As Horun has two of these boards, he will probably be the first one to test out your modified BIOS. While the 4 mbit BIOS on the Proserva board is in the same package as the EEPROM we did for the Pentium Pro board last year, the location is a lot more inconvenient for soldering and would require removing other components. So let's see how Horun makes out :) How did you get the idea to check the BIOS of an Asus board?

I'll probably try the NIC ROM trick, however I am currently quite disturbed why my W2K installation shows only one processor (yet MPS Multiprocessor kernel) and NT4 shows two processors, yet Horun's system with W2K shows two CPUs. Some ideas I had to circumvent this were, a) install NT4 then install W2K over NT4, b) install W2K on a fresh drive within int 13 translation limitation of 8 GB, c) install w2K on a dual PIII system on a fresh harddrive, then move the drive into the Proserva box to see if one of the CPUs disappear, d) try the original 1.00.10 WLB08 BIOS rather than the v14 BIOS (Horun had used the v8 BIOS). Unfortunately I no longer have the v8 BIOS and would need Horun to share his. EDIT1: Horun, could you let me know if you had APM enabled or disabled when installing W2K? Mine was enabled and am thinking maybe that messed up 2nd CPU detection? EDIT2: Solved: I disabled APM in the BIOS and now Task Manager shows 2 CPUs. I guess your APM was disabled.

feipoa wrote on 2020-01-19, 05:14:

SSTV2, thanks a lot for your clever solution. As Horun has two of these boards, he will probably be the first one to test out your modified BIOS. While the 4 mbit BIOS on the Proserva board is in the same package as the EEPROM we did for the Pentium Pro board last year, the location is a lot more inconvenient for soldering and would require removing other components.

I have completely forgotten that you've mentioned the type of the ROM chip NEC uses in the PM. In that case MB is protected from BIOS corruption, you can flash modded BIOS w/o fear of bricking it.

feipoa wrote on 2020-01-19, 05:14:

How did you get the idea to check the BIOS of an Asus board?

Due to newer firmare being larger than there's space for AIC-7880's module in NEC's BIOS, I've looked for OEM alternatives and after some googling, this Asus MB appeared in the search results, it also has an integrated AIC-7880 IC in it, so I gave its BIOS a go.

Also, Adaptec's firmare update "readme" file states this:

1CAUTION:
2 
3This BIOS update is NOT to be used for updating motherboard embedded
4SCSI controllers or OEM adapters. Some OEM adapters may have the letter
5'S' (designating a 'Special' BIOS version ) when the BIOS is displayed 
6on the screen, during the systems booting process.
7
8
9WARNING:
10
11Using this utility for OEM adapters, designed-in controller chips, 
12or other Adaptec adapters, may render those Adapters UNUSABLE.

feipoa wrote on 2020-01-19, 05:14:

try the original 1.00.10 WLB08 BIOS rather than the v14 BIOS (Horun had used the v8 BIOS). Unfortunately I no longer have the v8 BIOS and would need Horun to share his.

In case anyone need s the original BIOS that came on the NEC it is in the attached. You have to rename it (BIOS_08.ROM) to just BIOS.ROM or Phlash will not work.
Included an updated version of PHLASH (4.0 Release 1.57, the one from ZD website BIOS files is 4.0 Release 1.17).
Tested everything ! Flashed up and down w/o issue.

Hopefully I'll have time late this evening to try out the NIC boot rom, then try SSTV2's BIOS. I assume that the NIC needs to have an "enable boot from boot rom" option for this to work?

For the NIC boot ROM, should I use AIC-7880 ROM from Asus v1001 BIOS.bin (32 kbyte) or the Adaptec supplied afuv1343-29401343.bin (64 kbyte)? Based on the Adaptec firmware warning, I'm guessing I should use the Asus supplied BIOS?

feipoa wrote on 2020-01-19, 22:00:

For the NIC boot ROM, should I use AIC-7880 ROM from Asus v1001 BIOS.bin (32 kbyte) or the Adaptec supplied afuv1343-29401343.bin (64 kbyte)? Based on the Adaptec firmware warning, I'm guessing I should use the Asus supplied BIOS?

Yes the Asus bin file, the Adaptec file is same as in my PCI 2940UW and cannot be used in ours boards ROM so no need trying it.

I decided to use a D-Link DFE-538TX PCI network card with a DIP-28 socket.

First, without an EEPROM installed in the network card, I ran the Proserva's EISA configuration utility to ensure the 538TX card was setup to an available IRQ, IRQ 15 in this case. I also disabled the onboard AIC-7880 SCSI controller from the EISA configuration utility. Then I ran D-Link's 'diag.exe' utility from a DOS boot diskette. Here I enabled the boot ROM feature and set it to 32 Kbyte.

Next I shutdown the Proserva system and programmed the 32 Kbyte AIC-7880 Asus BIOS to a Winbond W27E257 using an external programmer. Installed the W27E257 into the 538TX network card, put the 538TX back into the Proserva's motherboard and powered up. I did not see any evidence that the Asus SCSI BIOS loaded - there was no on-screen display of any sort and the system did not attempt to boot from the SCSI HDD. Did I do anything wrong up to this point?

I re-ran D-Link's 'diag.exe' utility, but received an error message saying that there was no D-Link NIC in the system.

I powered down and removed the W27E257 EEPROM. Powered back up, ran 'diag.exe', and disabled the boot ROM. NIC still appears to work if the boot ROM BIOS is not installed.

I also have an EISA NIC in the system, a 3C597. It has a boot ROM socket, but I only have one of these cards so don't want to goof it up. It's boot rom is configured from the EISA configuration utility.

What is the size of your eprom? Also 32Kb?
Maybe the NIC performs a check if the ROM is valid?
Can you test NIC with ROM in another (standard) Mainboard?