I am working on the first experiments for a BIOS driver for the RS232 version of FujiNet.

https://gist.github.com/tschak909/8a60d1654a6 … 3c23ad81aada8e6

First BIOS experiment fails because serial interrupt does not drain the FIFO inside the INT F5 call. What stupid thing(s) am I doing here?

-Thom

oop, I had forgotten that the turbo c runtime disables interrupts. 😀 enabling them allowed the serial traffic to flow. good. whew.

-Thom

This is a video showing the current status (as of 2024-12-26) of the RS-232-C version of FujiNet, and the corresponding PC host software (COM, FUJICOM, and the client programs, ISS, NGET, NPUT, and NC).

https://youtu.be/ekWULObGB1o

The last part of the video talks about the X1 experiment to implement a BIOS interface for FujiNet using software INT 0xF5.

* COM implements the low level RS232 routines.
* FUJICOM implements the FujiNet command to device layer.
* ISS is the International Space Station Tracker
* NGET retrieves to a file from any network endpoint.
* NPUT puts a file to any network endpoint.
* NC is a dumb terminal to any network endpoint.

NC, NGET, and NPUT use the network device, and you can pass any available protocol adapter to it, e.g. TCP, UDP, FTP, HTTP/S, SSH, TELNET, SMB, NFS, and more, with the ability to add more protocol adapters. The FujiNet does all the hard work, and presents a nice clean interface to the target PC.

The above programs use FUJICOM directly, but this is not the intent for any final user program, at all. What is intended is for a BIOS interface to be used, abstracting for any physical interface. This will be important as the ISA version of FujiNet comes into existence. The X1 code is not working, and I need eyes to help fix it.

Eventually, it is intended that the work for X1 will be used to implement a DOS device driver, to directly deal with, not only disk images, provide the virtual printer, but also to provide file level access to MS-DOS directly for any network endpoint.

GitHub Repo:
HW: https://github.com/FujiNetWIFI/fujinet-hardware
SW: https://github.com/FujiNetWIFI/fujinet-rs232

FUJIBIOS Experiment 2 (X2) has been posted here:
https://github.com/FujiNetWIFI/fujinet-rs232/tree/main/X2

it can be built with Turbo C 2.0, with

```
tcc -v fujibios fujicom com
```

What works:

* AH = 00 - Send command AL with parameters in CX, to device DL, no payload
* AH = 80 - Send command AL with len in CX, to device DL, Payload from ES:BX

What doesn't:

* AH = 40 - Read command AL with len in CX, to device DL, Payload to ES:BX

Basically, _fujicom_send_command(), fujicom_command(), and fujicom_command_write() work.

fujicom_command_read() does not. using com_getc_sync not working.

Basically, any receive isn't hanging, but am not getting any result.

-Thom

Sorry for being ignorant. I get that it is a... Kind of software defined pi-based wireless adapter over serial, yes?

Is the aim for it to behave like any old Ethernet card once configured? Or a Laplink?

If Ethernet, there's already a pi project that afik emulates an ne2000.

Not trying to rain on your parade, but maybe you can borrow that code?
Or am I missing something?

No. That's not what this is, at all, look closer.

There's a great deal of protocol offloading and abstraction, and sub-devices.

If you want a good idea of the core functionality, you can see the video we did for the Atari version, many moons ago:
https://www.youtube.com/watch?v=izdCLDQ1ZxU

Welp, following up, we actually got the first pass implemented:
https://fujinet.online/2025/02/01/programming … ibm-pc-fujinet/

Looks promising! 😃