6 ms·

I worked on the original port (did about half the kernel stuff) - it was our standard SystemV with berkeley sockets, and wrote some of Apple specific stuff (ker

I worked on the original port (did about half the kernel stuff) - it was our standard SystemV with berkeley sockets, and wrote some of Apple specific stuff (kernel event manager, appletalk etc) - one thing I was particularly proud of was on the fly loadable unix device drivers, somewhat ahead of their time - feel free to ask me questions The Mac compatibility stuff (done at Apple) essentially ran in a single unix thread - really a VM for the mac OS7 world - it ran in user mode (mac OS apps usually ran in kernel mode) and emulated exceptions

jamesfmilne 8y ago
Hi Taniwha I’m working on a new Ethernet card for the Mac SE/30, and I’d love to be able to get if working with A/UX 3. https://www.mactothefuture.org/ https://www.mactothefuture.org/ If you had any pointers on how one could write an Ethernet driver for A/UX I’d be very appreciative! :-)
- Taniwha 8y ago
  oooh .... sadly I haven't had access to kernel source for maybe 30 years, it's been a long time essentially you need to write a BSD style networking driver (of the era, so likely 4.1) along with interrupts/etc I think without source it's going to be really difficult - on the other hand you can use autoconfig to load it into an old kernel, hook up the interrupts and set it running
  - jamesfmilne 8y ago
    Cheers. I'll see what I can dig out of the fossil record about autoconfig ;-)
    - Taniwha 8y ago
      autoconfig is essentially a front end like the linux module loading tools, it makes some fake COFF files containing the kernel symbol table and some glue code and links them against the driver(s) you want to load, then it loads the code into kernel space, patches the block/char major tables and calls a driver's init routine Sadly it's not as functional as we would like, I actually wrote it for UniSoft as a way we could sell drivers without building kernels every time, someone at Apple saw it and not only demanded we include it in A/UX but also demanded that they own it .... so I wrote another one for Apple, it worked differently and was barely functional - Apple could have had the original better one for free if they'd dropped the demand that they owned it
- npunt 8y ago
  Been watching this project for a while (via 68kmla), haven't seen any updates recently but excited to about any progress you make!
  - jamesfmilne 8y ago
    Yeah, sorry about that! Real life got in the way a bit this year. I've got the card sending & receiving packets properly now, but having a few issues with the CPLD, making the the machine crash sometimes accessing the card. Hopefully once that's fixed I can make a rev2 of the board and release some schematics & drivers.
    - npunt 8y ago
      Nice to hear you're making progress. Given these are 30 years old, what's a few months here or there, right? :) Look forward to seeing v2.
bluejekyll 8y ago
I guess the only question I have, is do you have an opinion on why this OS wasn’t chosen as the means of modernizing MacOS, rather than the Copeland effort? It sounds like you had a huge headstart on what turned out to be a similar strategy with virtualized Mac OS 9 on Mac OS X. I could see it being a case of wanting something shiny and new. Thanks for your response!
- Taniwha 8y ago
  I didn't work for Apple, I worked for UniSoft who did the Unix port to the Mac2 (we got half of the original batch of Mac 2 proto boards, I got to debug [and fix] the hardware). After we handed it over the group who did the UI work were pretty small within Apple - I think it was mostly politics, my view of Apple in those days (and the few years after) was that everything was politics, I remember the firewire guys coming around and shilling for supportive developer comments to try and keep their project afloat at one point. I'd guess that 2/3 of every cool thing designed at Apple got shelved, people who had poured several years of their lives would walk. A/UX died more slowly, switching to the PPC killed it, Apple decided not to do a Unix port (we probably could have done one faster than getting the MacOS working)
  - bluejekyll 8y ago
    excellent, thanks for the insight.

jamesfmilne 8y ago

Hi Taniwha I’m working on a new Ethernet card for the Mac SE/30, and I’d love to be able to get if working with A/UX 3. https://www.mactothefuture.org/ https://www.mactothefuture.org/ If you had any pointers on how one could write an Ethernet driver for A/UX I’d be very appreciative! :-)

Taniwha 8y ago
oooh .... sadly I haven't had access to kernel source for maybe 30 years, it's been a long time essentially you need to write a BSD style networking driver (of the era, so likely 4.1) along with interrupts/etc I think without source it's going to be really difficult - on the other hand you can use autoconfig to load it into an old kernel, hook up the interrupts and set it running
- jamesfmilne 8y ago
  Cheers. I'll see what I can dig out of the fossil record about autoconfig ;-)
  - Taniwha 8y ago
    autoconfig is essentially a front end like the linux module loading tools, it makes some fake COFF files containing the kernel symbol table and some glue code and links them against the driver(s) you want to load, then it loads the code into kernel space, patches the block/char major tables and calls a driver's init routine Sadly it's not as functional as we would like, I actually wrote it for UniSoft as a way we could sell drivers without building kernels every time, someone at Apple saw it and not only demanded we include it in A/UX but also demanded that they own it .... so I wrote another one for Apple, it worked differently and was barely functional - Apple could have had the original better one for free if they'd dropped the demand that they owned it
npunt 8y ago
Been watching this project for a while (via 68kmla), haven't seen any updates recently but excited to about any progress you make!
- jamesfmilne 8y ago
  Yeah, sorry about that! Real life got in the way a bit this year. I've got the card sending & receiving packets properly now, but having a few issues with the CPLD, making the the machine crash sometimes accessing the card. Hopefully once that's fixed I can make a rev2 of the board and release some schematics & drivers.
  - npunt 8y ago
    Nice to hear you're making progress. Given these are 30 years old, what's a few months here or there, right? :) Look forward to seeing v2.

bluejekyll 8y ago

I guess the only question I have, is do you have an opinion on why this OS wasn’t chosen as the means of modernizing MacOS, rather than the Copeland effort? It sounds like you had a huge headstart on what turned out to be a similar strategy with virtualized Mac OS 9 on Mac OS X. I could see it being a case of wanting something shiny and new. Thanks for your response!

Taniwha 8y ago
I didn't work for Apple, I worked for UniSoft who did the Unix port to the Mac2 (we got half of the original batch of Mac 2 proto boards, I got to debug [and fix] the hardware). After we handed it over the group who did the UI work were pretty small within Apple - I think it was mostly politics, my view of Apple in those days (and the few years after) was that everything was politics, I remember the firewire guys coming around and shilling for supportive developer comments to try and keep their project afloat at one point. I'd guess that 2/3 of every cool thing designed at Apple got shelved, people who had poured several years of their lives would walk. A/UX died more slowly, switching to the PPC killed it, Apple decided not to do a Unix port (we probably could have done one faster than getting the MacOS working)
- bluejekyll 8y ago
  excellent, thanks for the insight.