Here is HOT-433 BIOS with a new IRQ table (F000:0) taken from Octek HIPPO's BIOS (which has IRQ table at (F000:19F4).

Wow thank you very much!
I will test it during the weekend.

kmeaw wrote on 2024-10-09, 02:35:

Here is HOT-433 BIOS with a new IRQ table (F000:0) taken from Octek HIPPO's BIOS (which has IRQ table at (F000:19F4).

Hello
so I have flashed the bios but the result is the same as my bios edit attemps. Board is not booting and makes beep error NO Graphics card and than non stop short beep.

Any sugestions?
Thank you

Could you tell the exact version of the original BIOS that works on your board and has the correct IRQ table?
Do you have a POST code card to check which POST function has failed?

Why not use a PCI IDE controller? I've had great luck with Promise ultra 100's (one of that last I can confirm to work on most 486s), or obviously you could use a slower (66) model. These can really wake a 486 up since the onboard ide is always painfully slow. An ultra txt2 will even work on wonky pci frequencies like 50 or 60 mhz.

bertrammatrix wrote on 2024-10-14, 21:14:

Why not use a PCI IDE controller? I've had great luck with Promise ultra 100's (one of that last I can confirm to work on most 486s), or obviously you could use a slower (66) model. These can really wake a 486 up since the onboard ide is always painfully slow. An ultra txt2 will even work on wonky pci frequencies like 50 or 60 mhz.

Yes in time I will try it, but now I dont have one and the board has only 3 PCI slots and I am using them all.

kmeaw wrote on 2024-10-14, 20:16:

Could you tell the exact version of the original BIOS that works on your board and has the correct IRQ table?
Do you have a POST code card to check which POST function has failed?

Hey sorry for long repply. I was bussy lately.
I am using this bios https://theretroweb.com/motherboard/bios/hipp … 98139842119.zip. Only with EDO options enabled via MODBIN.
I was not able to open the bios with AWBEDIT, but if you load and save it with MODBIN than its possible. If it helps.

Check the link with video of the beep. I dont have yet the POST code card.
Please keep in mind that the Hippo has 3 PCI slots and the Shuttle 4 PCI slots. I dont know maybee this cound be the case.

https://1drv.ms/v/s!Alo_fdsfM0Nsh_4Aikqg7vIyP6tETw?e=cLE4Ld

Thank you for your help.

Thank you for the video.
I have located the code which produces exactly the sequence of the beeps you have - it's the checksum test at F000:EB33.
Here is another version with an updated checksum (1DFFE: 82 EF -> 5D CA).

kmeaw wrote on 2024-10-21, 22:36:

Thank you for the video.
I have located the code which produces exactly the sequence of the beeps you have - it's the checksum test at F000:EB33.
Here is another version with an updated checksum (1DFFE: 82 EF -> 5D CA).

Hi
so the bios is working but still no IRQs for the PCI cards. ISA soundcard is fine.

In the attachement is report from Speedsys and here are some pictures https://1drv.ms/f/s!Alo_fdsfM0Nsh_4Ko6JqMnkRlrIrGQ?e=pQc0qc

thx

kmeaw wrote on 2024-10-21, 22:36:

Thank you for the video.
I have located the code which produces exactly the sequence of the beeps you have - it's the checksum test at F000:EB33.
Here is another version with an updated checksum (1DFFE: 82 EF -> 5D CA).

I have added more pictures to https://1drv.ms/f/s!Alo_fdsfM0Nsh_4Ko6JqMnkRlrIrGQ?e=pQc0qc

In Modbin there is no change in the PCI configuration table SHMODIRQ and the PCI configuration in the original bios is completly different.

Also I have noticed in the chipset registry defaults that the HIPPO bios has two times 805E index and the 9046 index is different from the Shuttle bios.
Dont know if its relevant.

Can you tell me what tools are you using or give me some hints what are you exatly doing? 😀
I have read these articles https://sites.google.com/site/pinczakko/home

Thank you for your help

PCI slot configuration table is different between 2A4X5H21 (F000:4A90) and 2A4X5O0HC (F000:7A9F). I have copied the table to 2A4X5H21 - fortunatelly it has less slots, if it had more, then I would need to relocate it.
Here is another patched version. Unfortunatelly 86Box does not assign the interrupts even with your original BIOS so I cannot properly test it.

I am using Modbin4 to unpack and repack, Hiew 6.11 for interactive editing, AwdHack as a Modbin wrapper to run Hiew just before repacking (for this purpose I use Hiew 2.51 as it requires much less low memory), ndisasm for listings, less and grep for paging and filtering.

The workflow for a pair of compressed 128KB BIOS is:
1) open a bios with Modbin4, it unpacks the image and creates a temporary file called "original.tmp" that I would use later;
2) run "ndisasm -b 16 -o 0xe0000 original.tmp" and keep the listing;
3) repeat steps 1-2 for another BIOS, compare the listings by finding rare instructions - to find PCI slot configuration table, I used this regular expression: "mov e..,0x40......$" - the following mov si,imm32 points to the table;
4) open both files in Hiew 6.11, go to code mode (enter, enter), set offset (ctrl-f5, e0000), go (f5) to the code that was just found, set a bookmark (+), walk around (pgup/pgdn/0-9) to check the surrounding code and then return back (-) if needed;
5) make a plan of the changes - make instruction patches in a temporary copy using Hiew to get it assembled if needed to get the hex codes;
6) run AwdHack - it would run Modbin4, then just use the top option to update the file - AwdHack would intercept that and drop you into Hiew 2.51;
7) apply the planned changes in hex (remember there is no offset option, so E000:0 is just 0 but F000:0 is 10000), save and exit Hiew;
😎 AwdHack would pass the changed original.tmp back to Modbin4, which would repack it.

To verify the changes, I use vbindiff (for visual inspection) and 86Box (to actually run the modded image).

If I need some more complex rework, I would use Ghidra and spend a few hours mapping the structures and subroutines.

kmeaw wrote on 2024-10-22, 17:33:

PCI slot configuration table is different between 2A4X5H21 (F000:4A90) and 2A4X5O0HC (F000:7A9F). I have copied the table to 2A4 […]
Show full quote

PCI slot configuration table is different between 2A4X5H21 (F000:4A90) and 2A4X5O0HC (F000:7A9F). I have copied the table to 2A4X5H21 - fortunatelly it has less slots, if it had more, then I would need to relocate it.
Here is another patched version. Unfortunatelly 86Box does not assign the interrupts even with your original BIOS so I cannot properly test it.

I am using Modbin4 to unpack and repack, Hiew 6.11 for interactive editing, AwdHack as a Modbin wrapper to run Hiew just before repacking (for this purpose I use Hiew 2.51 as it requires much less low memory), ndisasm for listings, less and grep for paging and filtering.

The workflow for a pair of compressed 128KB BIOS is:
1) open a bios with Modbin4, it unpacks the image and creates a temporary file called "original.tmp" that I would use later;
2) run "ndisasm -b 16 -o 0xe0000 original.tmp" and keep the listing;
3) repeat steps 1-2 for another BIOS, compare the listings by finding rare instructions - to find PCI slot configuration table, I used this regular expression: "mov e..,0x40......$" - the following mov si,imm32 points to the table;
4) open both files in Hiew 6.11, go to code mode (enter, enter), set offset (ctrl-f5, e0000), go (f5) to the code that was just found, set a bookmark (+), walk around (pgup/pgdn/0-9) to check the surrounding code and then return back (-) if needed;
5) make a plan of the changes - make instruction patches in a temporary copy using Hiew to get it assembled if needed to get the hex codes;
6) run AwdHack - it would run Modbin4, then just use the top option to update the file - AwdHack would intercept that and drop you into Hiew 2.51;
7) apply the planned changes in hex (remember there is no offset option, so E000:0 is just 0 but F000:0 is 10000), save and exit Hiew;
😎 AwdHack would pass the changed original.tmp back to Modbin4, which would repack it.

To verify the changes, I use vbindiff (for visual inspection) and 86Box (to actually run the modded image).

If I need some more complex rework, I would use Ghidra and spend a few hours mapping the structures and subroutines.

again, no IRQs.
pictures and log in the share folder. SHMOD2.TXT

Is there anything I can help you?

Does the PCI device listing look the same as in your shmod2.jpg file?
Do you perform an ESCD reset after flashing a new version of BIOS?

kmeaw wrote on 2024-10-22, 19:58:

Does the PCI device listing look the same as in your shmod2.jpg file?
Do you perform an ESCD reset after flashing a new version of BIOS?

yes I clear bios and reset configuration to ENABLE
also tried to assign IRQs manualy.

with the original bios there is no listing like in the Shuttle bios. You can see the IRQs in the HIPPO.JPG.

kmeaw wrote on 2024-10-22, 19:58:

Does the PCI device listing look the same as in your shmod2.jpg file?
Do you perform an ESCD reset after flashing a new version of BIOS?

Maybee this could be the problem.
On the original bios after clear cmos there is always Updating ESCD.

On the mod bios even when I select reset configuration or change everything manualy, it never shows.

Here is another version to test with changed $PIR table.

kmeaw wrote on 2024-10-22, 22:50:

Here is another version to test with changed $PIR table.

You are amazing! Its working! Thank you very much!

Can you please check the share folder there is a video and the clock is acting weird. 😀

Also the PNP/PCI configuration page is missing 2 lines. This can be fixed with Modbin right?

Can I or can you upload the bios to Retroweb? Maybee put some signature inside.

Once again thank you. I owe you! Do you like Czech beer or candy? I can send you some.

freezecze wrote on 2024-10-23, 06:58:

You are amazing! Its working! Thank you very much! […]
Show full quote

kmeaw wrote on 2024-10-22, 22:50:

Here is another version to test with changed $PIR table.

You are amazing! Its working! Thank you very much!

Can you please check the share folder there is a video and the clock is acting weird. 😀

Also the PNP/PCI configuration page is missing 2 lines. This can be fixed with Modbin right?

Can I or can you upload the bios to Retroweb? Maybee put some signature inside.

Once again thank you. I owe you! Do you like Czech beer or candy? I can send you some.

Clearing CMOS once more solved the problem with clock.

kmeaw wrote on 2024-10-22, 22:50:

Here is another version to test with changed $PIR table.

I've done a successful PCI slots - IRQ table modification once, but for Award BIOS version 4.51PG. Is it valid for Award BIOS ver. 4.50P(G)?

Never heard of this AwdHack tool, but unfortunately I can't find it for download.

freezecze wrote on 2024-10-23, 06:58:

Also the PNP/PCI configuration page is missing 2 lines. This can be fixed with Modbin right?

It can be fixed by MODBIN, but only if these two lines are hidden in the BIOS settings menu.