Woohoo!

putting together a mapper array to go from usb hid codes to ps2 codes is quite a slow and boring process.

I know there are some shortcuts I can take but doing it this way gives me a good idea of the issues I have to solve with this.
It was also interesting to see how different the XT and AT scancodes are. The XT release code is the same as the press code with 128(DEC) added thus just flipping a bit in binary, meanwhile AT sends a code to say "depressed" followed by the press code.

so yeah. wooooo

BTW it's worth mentioning that I'm writing the keyboard side of things with XT in mind but I don't have immediate plans to support XT largely before I don't have an XT class machine or interface to test with.

It really will be universal then!

I put in the tables I made in excel and after some testing I have discovered that my spare USB keyboard is broken, yay!

Alrighty,
Nothing too fun or exciting done. Just some scaffolding stuff but hopefully it means that I'll get the USB keyboard buttons sent to over ps/2 which would be another fun thing achieved.
Had to figure out bitwise operations again, boy those things make no sense to me no matter how many times I use them.

You can do et!

So what is the full supported configuration list now? And future supported configurations?

Sphere478 wrote on 2022-10-20, 01:42:

You can do et!

So what is the full supported configuration list now? And future supported configurations?

I mean the hope is USB mouse to serial mouse, usb keyboard to ps/2 keyboard with a potential option for xt support. Honestly, I want to keep the keyboard side of things transparent so I'm not going to have anywhere near the same amount of configuration options for the keyboard as I have for the mouse. Though I may have to come up with something for keyboard buttons that have been lost over the years. Press F to mourn the calculator button.

but if you meant the scan codes, attached is a print out of the excel table I have with all the tinyUSB -> ps2 scan codes I could get.

Well it is now writing to the computer from the ps/2 port.
It's actually working surprisingly well for the moment.

The translation table seems to work, I was very surprised when I pressed the "right mouse click" button on my keyboard that it actually worked. I was very confused for a while when only one device from my usb hubs was picked up but that was just a TinyUSB setting because of course my most confusing and vague issues stem from TinyUSB being weird.

There's some polish to do and I have to get it working in a way that continues to work when you open the terminal but overall the bulk is coming together quicker than I thought it would.

CalamityLime wrote on 2022-10-19, 15:58:

BTW it's worth mentioning that I'm writing the keyboard side of things with XT in mind but I don't have immediate plans to support XT largely before I don't have an XT class machine or interface to test with.

I'm kinda tempted to send over an XT m/b with an 8-bit VGA... A USB->XT KB adapter would be a huge thing to me. I don't even know if it already exists, though... I'm not a big ebay jockey. Thanks to the lack of funds, mostly...

There is PS/2 to XT Adapter 😀

alvaro84 wrote on 2022-10-23, 04:47:

I'm kinda tempted to send over an XT m/b with an 8-bit VGA... A USB->XT KB adapter would be a huge thing to me. I don't even know if it already exists, though... I'm not a big ebay jockey. Thanks to the lack of funds, mostly...

I appreciate the sentiment!
On my end I'm kinda tempted to send out a unit to someone who would be willing to test the xt support.

Like matze said there various PS2/at -> xt adapters out there, some are a little over priced tbh. I'm probably going to look to open sourced ones to figure out how the initialization works, i think the transmission of data is almost the same just with a little slower transmission rate, maybe. I'd have to look into it a bit more.

But I did read that IBM xt keyboards and clone xt keyboards have a different number of start bits to show that they are different. I wonder why? Are there xt's by IBM that reject clone XT keyboards? Or did big blue not want their XT keyboards to work on clones?

This may be a bit of a scattered update.

Didn't get much of a chance to work on this project today. Some home improvement didn't quite work out, yay.

Still, I've ordered some more ps/2 cables and some bits for a potential tester unit and I've decided to learn something about buck-boost regulators by eventually putting together a voltage blaster that makes a -5v from +12v. So thats something to look forward to, hopefully. Who knows might even try to track down a scope somewhere to measure ripple, it's been on my todo list for while now.

I assembled some units today. I'm tempted to offer one as a testing unit but I'm not sure, would people even be interested in that?

CalamityLime wrote on 2022-10-24, 22:05:

I assembled some units today. I'm tempted to offer one as a testing unit but I'm not sure, would people even be interested in that?

I could maybe do some testing. But I don’t have a xt

hey there.

Currently work is a bit all over the place with 2 events back to back so I have no idea how time will be to work on things.

I have been researching the XT protocol and I'm happy to see that it is well documented especially between these two links.
https://github.com/tmk/tmk_keyboard/wiki/IBM- … yboard-Protocol
http://www.ccgcpu.com/2019/02/14/the-xt-part- … rnish-keyboard/

So today's coding time was spent working out timings and writing up an XT send keycode function that should work plus cleaning up the ps/2 code a bit. with it imported from ps2dev, it's a little all over the place while I get it organised.

I'm still waiting on bits to arrive from amazon to put together testing units. This is already a huge loss for me to make so might as well make it decent where possible.

On that note, I was thinking today about the time I tried to decompile and edit serial mouse drivers so they would work with higher baud rates. Essentially it *should* be a case of finding serial latch registers and editing the values but for complicated drivers there might be more to it.
Does anyone here think that they might be able to do that? I tried and couldn't figure out how to use ghidra. Even if you know if a different decompiler that a noob could use to edit and re-package a driver.

I did manage to make modified CTmouse drivers for DOS but It would be fantastic if we could use higher baud rates under windows 95 or something.

I took a look at modding windows 95 mouse drivers. I think whatever needs to be edited is somewhere in mouse.drv but I really have no idea what I'm looking at in there.

welp, I tried again with modding win 9x drivers and got nowhere.

I then moved onto ReactOS drivers thinking that they could be use on win nt. Modding the baud rate for the reactos serial mouse drive looks simple enough, and compilation takes a while but it works away on it's own just fine. I couldn't figure out how to make the modded sermouse file from reactOS install on winxp. If I replaced the file live, the driver information updates but it stops working after a reboot (it would be the using the ms one from memory)

Then I thought that I'm being too hopeful and to just get the modded drivers working in reactOS first. Then reactOs live cd doesn't pick up the serial mouse with stock drivers, that seems to be due to the fact that the liveCD holds the com line for debugging and to turn that off I would have to modify an ini file. Expect it's a liveCD. I tried unpacking and repacking the livecd iso but that resulted in a bootable iso that errored out on boot. I know how to repack an iso and keep it bootable but it just didn't want to work with reactos. I tried installing from the liveCD and it errored out. I tried using the bootCD and it doesn't boot.

I tried this for the release and nightly builds. Not the first time reactos didn't work I guess I have to wait for a working release?

I've also found an older MS keyboard in the bin today. It has some media keys (like the glorious dedicated calculator button!) that I was hoping to map to the ps/2 scan codes but it looks like tinyUSB doesn't support media keys easily. It's well able to send them when it's mimicking a keyboard but not receive them easily. this should be interesting.

So not a very cheerful update but eh

Update within update,
While typing this out I remembered that I had reactos working on the same hardware in question in the past. so I tried an older release and it worked! it was missing USB drivers so I had to use my adapter to translate usb to ps2, that half worked. I had to debug a bit to get it working right with ps2 so I had to patch that.
While reactOS was installing, the power led of the adapter started flashing meaning it is in terminal mode. I guess reactos prodded the com port looking for something. To my surprise, they keyboard still worked. Something I thought I'd have to do some big reshuffling to do. I investigated that just to find that I done that work already and forgot. So that's handy

The older reactos installed (after one fail), it didn't have a working com port driver though and trying to install one makes the system hang.

so it's a series of ups and downs. and I'm very tired now so goodnight.

I have a question

how many people here have playstation controllers? If so, which ones?

CalamityLime wrote on 2022-10-31, 17:19:

I have a question

how many people here have playstation controllers? If so, which ones?

I have usb xbox 360 style ones

Sphere478 wrote on 2022-10-31, 18:21:

CalamityLime wrote on 2022-10-31, 17:19:

I have a question

how many people here have playstation controllers? If so, which ones?

I have usb xbox 360 style ones

I'd like to get them working but they use xinput which doesn't work natively with tinyusb, so that's more work unfortunately.
Personally I really only use xbox or ps2 controllers when I use controllers.