myne wrote on 2025-05-30, 02:07:

Lpc is a bus though. […]
Show full quote

Lpc is a bus though.

You can't hardwire data selectively to one device and not the other.
Control pins like ldrq are obviously different.

Chatgpt, talking with the confidence of an expert, and the understanding of a parrot.

I'd check the soldering again.

You don't need LDRQ at all for unisound to init card - proven on Asus P5Q (attached).
ChatGPT while not being Kant's knowledge-class wisdom reveals some breadcrumbs/clues.
Like you right, imagine MultiIO was connected in some way and also requires init with proper address windows (several) set. Where, is there override method ? And it's also on LPC. Then imagine there's good old days where you have two devices listening on the same ISA port and both responding on port read. Also there's PCIe-PCI bridge.
I have another one mobo with the same issue - Asus Z87-C, with the same PCIe to PCI bridge as Z97M-Plus
Maybe there's some substractive decoding routed here ?
I also have Gryphon Z87 which maybe considered as a copy of Maximus VI Gene, will check nearby hours is UNISOUND works here.
Friend of mine will check P5Q multiplexing ability of LDRQ1/GPIO23 - will it work (by soldering directly, without port programming), because GPIO23 connected down the road to the raid logic within ICH.

vsharun wrote on 2025-05-30, 06:05:

You don't need LDRQ at all for unisound to init card - proven on Asus P5Q (attached).

Never said you did. I said it was a control line which iirc are/or optionally can be point to point vs bus.

vsharun wrote on 2025-05-30, 06:05:

Friend of mine will check P5Q multiplexing ability of LDRQ1/GPIO23 - will it work (by soldering directly, without port programming), because GPIO23 connected down the road to the raid logic within ICH.

I... don't think you can mux it 'live' like that. It's an option for MFGs. At least disable the raid.

myne wrote on 2025-05-30, 06:15:

vsharun wrote on 2025-05-30, 06:05:

You don't need LDRQ at all for unisound to init card - proven on Asus P5Q (attached).

Never said you did. I said it was a control line which iirc are/or optionally can be point to point vs bus.

Then why it didn't work on Z97M-Plus/Z87-C and B75 mentioned.

myne wrote on 2025-05-30, 06:15:

vsharun wrote on 2025-05-30, 06:05:

Friend of mine will check P5Q multiplexing ability of LDRQ1/GPIO23 - will it work (by soldering directly, without port programming), because GPIO23 connected down the road to the raid logic within ICH.

I... don't think you can mux it 'live' like that. It's an option for MFGs. At least disable the raid.

I don't care about RAID, this is clean experiment.
I need help with compiled for DOS tool where I may change LDRQ1 <-> GPIO23 pin.

No idea.
Bad soldering?
Wrong components?

OP posted his github with config for that, and someone else a page or so back posted theirs.

It's ISA PNP init blocked.
I have connected Gryphon Z87 (without LDRQ soldered) and put PicoGUS in it.
saphisa 200 fc 300 fc 1d0 0f and voila.
There's no soldering at all, mean on the motherboard or in cables. This adapter pulled out from "working everything" Asrock Z87M-pro4 setup.
UPD: MPU-401 on PicoGUS are okay.
UPD2: Tiido's setymf sees the card (YMF719 Addonics) but can't init.
UPD3: seems most port ranges are okay, and the solution is - to switch ESS card from PnP to no PnP mode. I have BTC 1857L which contains 93C66 EEPROM and according to TRW quick init quide supports programming via EEPR1868 with some images around according to the doc.

vsharun wrote on 2025-05-30, 08:10:

It's ISA PNP init blocked.

Yeah this is the problem at the moment. However, unlike Z97, B75 does have native PCI interface.
Perhaps this is something ASUS-specific, considering ISA PnP worked on ASRock and Gigabyte boards.

vsharun wrote on 2025-05-30, 08:10:

saphisa 200 fc 300 fc 1d0 0f and voila.

Be careful with 300 range as you may interfere with VGA functionality on top of COM/LPT ports. That's why by default two slots were used for the 300 range, one for MIDI and the other for FM.
To save a slot for other use cases (e.g. WSS) I once used "300 bc", though it's not a 100% good idea as it gets in the way of the I/O range used by non-VGA video (Monochrome, EGA, etc.).
Another way would be "300 8c" if you intend to put MIDI on 300h instead of 330h, provided that games do not hardcode MIDI I/O address.

LSS10999 wrote on 2025-05-30, 09:22:

vsharun wrote on 2025-05-30, 08:10:

It's ISA PNP init blocked.

Yeah this is the problem at the moment. However, unlike Z97, B75 does have native PCI interface.
Perhaps this is something ASUS-specific, considering ISA PnP worked on ASRock and Gigabyte boards.

Is there a way to avoid PnP init via system routine ? (reading/writing ports 203/279/A79). ChatGPT suggests there's some way to convert PnP cards with EEPROM to nonPNP (most of YMF719 has, at least one of my ESS1868 has it and I managed to find eeprom write utility, but without structure and sample roms ?), can't find anything about that on the internet though.

LSS10999 wrote on 2025-05-30, 09:22:

Be careful with 300 range as you may interfere with VGA functionality on top of COM/LPT ports. That's why by default two slots w […]
Show full quote

vsharun wrote on 2025-05-30, 08:10:

saphisa 200 fc 300 fc 1d0 0f and voila.

Be careful with 300 range as you may interfere with VGA functionality on top of COM/LPT ports. That's why by default two slots were used for the 300 range, one for MIDI and the other for FM.
To save a slot for other use cases (e.g. WSS) I once used "300 bc", though it's not a 100% good idea as it gets in the way of the I/O range used by non-VGA video (Monochrome, EGA, etc.).
Another way would be "300 8c" if you intend to put MIDI on 300h instead of 330h, provided that games do not hardcode MIDI I/O address.

That was PoC only.
I suppose help from UNISOUND developer might be useful here. Is it possible at all to init card directly avoiding system PnP init procedure.

upd: There manual for ESS1869 (the same for 1868) "bypass key" method of configuration where no PnP available:
https://www.philscomputerlab.com/uploads/3/7/ … 9techmanual.pdf
Page 28/29
Content of EEPROM's 512Bytes pages 89 and more.
No signs for "Legacy mode startup" though.

https://www.intel.com/content/dam/www/public/ … h-datasheet.pdf

That's where you'll find answers about isa registers

This might help too
https://download.microsoft.com/download/1/6/1 … 456c/pnpisa.rtf
I feel like pg 7 is a good place to start

myne wrote on 2025-05-30, 11:53:

https://www.intel.com/content/dam/www/public/ … h-datasheet.pdf […]
Show full quote

https://www.intel.com/content/dam/www/public/ … h-datasheet.pdf

That's where you'll find answers about isa registers

This might help too
https://download.microsoft.com/download/1/6/1 … 456c/pnpisa.rtf
I feel like pg 7 is a good place to start

pnpdump from isapnptools returns "no boards found" across the whole range 0x273-0x3FB (I have tried from 0x203 - the same)
quite interesting how debug looks like:
for ports range forwarded by sapphisa I get FF's on all reads from all ports during pnpdump, for range not forwarded - 00's

upd: obviously the same sequence done in DEBUG gaves nothing. Blind configuration also gaves nothing.
Seems debug board required for 279 and A79 ports writes detection.

Just curious, are those decode ranges ( 1 and 2), while not being used (mobo has no com/lpt ports) can be reprogrammed ? Are they readable ?
Maybe 279/A79 are trapped here ?

Iirc the OP used a z97.
Perhaps, a different brand, but it can't hurt to ask for a dump of their chipset registers for comparison.

vsharun wrote on 2025-05-31, 12:39:

Just curious, are those decode ranges ( 1 and 2), while not being used (mobo has no com/lpt ports) can be reprogrammed ? Are they readable ?
Maybe 279/A79 are trapped here ?

I think you're on to something with the generic decode ranges

myne wrote on 2025-05-31, 14:00:

vsharun wrote on 2025-05-31, 12:39:

Just curious, are those decode ranges ( 1 and 2), while not being used (mobo has no com/lpt ports) can be reprogrammed ? Are they readable ?
Maybe 279/A79 are trapped here ?

I think you're on to something with the generic decode ranges

Thank man. This points me there, the situation is very-very close to our case: when port 279 set to read only, then no PnP ISA devices can be configured, as they require port 279 to be writable.

101 = 278h–27Fh (port 279h is read only) and 678h–67Fh

The only issue IDK how to read or write those registers, as I have only DOS (because only from DOS or win98 this should work)

wpcredit might work in 98

not sure why you're looking at the printer port

myne wrote on 2025-05-31, 19:01:

wpcredit might work in 98

not sure why you're looking at the printer port

Because 279 is the main configuration port for ISA PnP process as stated in the document pnpisa.rtf
Unbelievable: I managed to read bus 0/dev 31/function 240 on Asus and Asrock motherboards. And they're full of differences on "Reserved" bits.
In both cases bits 8-9 LPT decode range stays 00, whatever type/port I set in Asrock BIOS.
Asus offset 80 (4bytes) - 3C 0F 00 10

1offset 80 81 82 83
20011 1100 0000 1111 0000 0000 0001 0000
30001 0100 0000 0000 0000 0000 0000 0000
40001 0100 0000 0101 0000 0001 0000 0000

^^^ Asrock 1400 0000, then asrock LPT and com enabled 1405 0100
... on clean start. The differences only in "reserved" bits.

vsharun wrote on 2025-05-31, 18:09:

The only issue IDK how to read or write those registers, as I have only DOS (because only from DOS or win98 this should work)

You need setpci for DOS.
The syntax is a bit weird but it works.
https://linux.die.net/man/8/setpci

vsharun wrote on 2025-05-31, 18:09:

Thank man. This points me there, the situation is very-very close to our case: when port 279 set to read only, then no PnP ISA devices can be configured, as they require port 279 to be writable.

I don't think this is the problem. Port 279 being read-only is intended for parallel port use cases, if no generic decode range is set to this region. While ISA PnP writes to port 279, it has nothing to do with LPT port.

Maybe leaving the LPT port setting in BIOS to "Auto" (which is the default) is not a good idea?

LSS10999 wrote on Yesterday, 02:50:

vsharun wrote on 2025-05-31, 18:09:

Thank man. This points me there, the situation is very-very close to our case: when port 279 set to read only, then no PnP ISA devices can be configured, as they require port 279 to be writable.

I don't think this is the problem. Port 279 being read-only is intended for parallel port use cases, if no generic decode range is set to this region. While ISA PnP writes to port 279, it has nothing to do with LPT port.

Maybe leaving the LPT port setting in BIOS to "Auto" (which is the default) is not a good idea?

Asus Gryphon Z87 has no BIOS settings for LPT/COM because there's no COM/LPT ports. It's Asrock Z87M-Pro4 settings for parallel/serial, which affects LPC configuration registers 80-83h.
Next week or two I expect to get back my oscilloscope and will check is writing to port 279/A79 (debug script with lot of O) reflects somehow within ISA slot or LADx LPC pins.