Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

ELKS: Linux for 16-bit Intel Processors

github.com

Post reply on HN
Print view View original

91–100 of 119 posts

Re: ELKS: Linux for 16-bit Intel Processors

#91

Post by ghaerr »

ikari_pl wrote: post #59
ghaerr wrote: post #47

Thanks for the comments! ELKS will run on an 8088, but also runs on any x86 PC using the legacy "real mode" which runs in 16-bit segmented architecture without an MMU or any hardware protection. It can be fun to boot a PC and see a close resemblance to Linux, but run the way things used to be, using only 16-bits. It'll also run in ROM, e.g. using a 8018x CPU w/onboard PIT and PIC. The point of ELKS today is about "sm…

This is what I've been looking for! Can you fit a workable version on a 1.44 floppy?

Yes, one can actually boot and run a minimal system on 360k floppy, but if you want networking you need 720k. The native compiler might fit on 1440k but needs a lot more for any development if the C library header files etc are wanted.

Re: ELKS: Linux for 16-bit Intel Processors

#92

Post by ghaerr »

cricalix wrote: post #66
ikari_pl wrote: post #59

Earlier quoted context omitted.

This is what I've been looking for! Can you fit a workable version on a 1.44 floppy?

Around 1996/97 timeframe, you could fit a kernel and userspace. I remember building a 1.44 setup that booted a compressed kernel and had enough user space tooling to bring up telnet, ftp, and the radio stack to drive the long haul radio cards (we were replacing JNOS [0] IIRC at an ISP ). Even had writable space at the end of the floppy (the kernel etc were readonly) to write overrides of the config; poor man's overla…

That's pretty cool about the radio stack!

We don't support a compressed kernel anymore but have a way to compress user executables which saves about 30%. Sadly, even with straightforward decompression, that process on ancient 8088's sometimes takes longer than reading an uncompressed executable. But we're finally at the point of having "too much stuff" to fit on a floppy. And of course everyone wants games. (Don't ask about Doom: yes we have it, no it doesn't fit on a floppy :)

Re: ELKS: Linux for 16-bit Intel Processors

#93

Post by ghaerr »

ikari_pl wrote: post #68
cricalix wrote: post #66

Earlier quoted context omitted.

Around 1996/97 timeframe, you could fit a kernel and userspace. I remember building a 1.44 setup that booted a compressed kernel and had enough user space tooling to bring up telnet, ftp, and the radio stack to drive the long haul radio cards (we were replacing JNOS [0] IIRC at an ISP ). Even had writable space at the end of the floppy (the kernel etc were readonly) to write overrides of the config; poor man's overla…

I'm totally going to run it on an Amstrad PPC640 soon :)))

Yes it runs on that Amstrad. A funny story about the Amstrad, at one point we added divide-by-zero trap handler in the kernel for user space apps. When the Amstrad reboots via our 'shutdown', its gets a div zero exception in its own BIOS (which at the time prohibited the reboot, lol).

Re: ELKS: Linux for 16-bit Intel Processors

#94

Post by nolist_policy »

anthk wrote: post #78
LeFantome wrote: post #39

We are removing 32 bit support from the Linux kernel because nobody uses that hardware anymore. Yet, this 16 bit version has commits from today.

Hyperbola GNU/Linux, Parabola and Guix still have 32 bit support. And I still use a 32 bit computer. And for sure the 2nd and 3rd world use 32 bit devices.

I bet more than 50% of computer e waste is 64 bit machines nowadays.

Re: ELKS: Linux for 16-bit Intel Processors

#95

Post by ghaerr »

dcminter wrote: post #88
cricalix wrote: post #66

Earlier quoted context omitted.

Around 1996/97 timeframe, you could fit a kernel and userspace. I remember building a 1.44 setup that booted a compressed kernel and had enough user space tooling to bring up telnet, ftp, and the radio stack to drive the long haul radio cards (we were replacing JNOS [0] IIRC at an ISP ). Even had writable space at the end of the floppy (the kernel etc were readonly) to write overrides of the config; poor man's overla…

I'm 99% sure the original kernel and thus the boot&root disk Linux (the original "distribution" I guess) only ran on the 386 & up as it required & used the memory management capabilities. There were some forks that could run without the mmu (micro-Linux I think?) but as I recall it they came quite a bit later. Edit: Ah, close, this: https://en.m.wikipedia.org/wiki/%CE%9CClinux

We dropped all support for 286 or 386+ protected mode/paging etc, as well as produce only 8088/8086 instructions so it'll run on any x86 (including the PCjr with its peculiarities, remember that?) running in real mode only without MMU. Of course, that means any program can write anywhere, so more care is taken towards program correctness, which is kind of fun.

Re: ELKS: Linux for 16-bit Intel Processors

#96

Post by ghaerr »

tialaramex wrote: post #18

Cool. A quarter of a century ago a friend of mine worked on this and you can still see his name in some copyright notices (Alistair Riddoch) Most recently Al was learning Rust because he needs that for his current role, it might be fun to see whether you can write an ELKS target for Rust.

Is Alistair Riddoch still around? I've been wondering about him, his name is all over the early dev86 and ELKS C library code :)

Re: ELKS: Linux for 16-bit Intel Processors

#97

Post by yjftsjthsd-h »

marcodiego wrote: post #2

What would be really interesting: porting DOSEMU 1.4.0 to it. This would give us a maintained unix-like OS combined with a huge abandonware dos library turning those old machines into something fun and maybe useful.

I don't know that it's really Unix like, but I feel like freedos covers that use?

Re: ELKS: Linux for 16-bit Intel Processors

#98

Post by tialaramex »

ghaerr wrote: post #96
tialaramex wrote: post #18

Cool. A quarter of a century ago a friend of mine worked on this and you can still see his name in some copyright notices (Alistair Riddoch) Most recently Al was learning Rust because he needs that for his current role, it might be fun to see whether you can write an ELKS target for Rust.

Is Alistair Riddoch still around? I've been wondering about him, his name is all over the early dev86 and ELKS C library code :)

Well I mean, he's a friend and in fact hosted the NYE party I was at so, in that sense he's still around. I might see him later this week and if so I'll try to remember to mention that ELKS is still going.

Re: ELKS: Linux for 16-bit Intel Processors

#99

Post by LeFantome »

happycube wrote: post #89
ghaerr wrote: post #47

Thanks for the comments! ELKS will run on an 8088, but also runs on any x86 PC using the legacy "real mode" which runs in 16-bit segmented architecture without an MMU or any hardware protection. It can be fun to boot a PC and see a close resemblance to Linux, but run the way things used to be, using only 16-bits. It'll also run in ROM, e.g. using a 8018x CPU w/onboard PIT and PIC. The point of ELKS today is about "sm…

Very awesome work indeed... I haven't looked at ELKS in a long time and seeing everything it can do now is heartwarming. The 8018x/ROM mode sounds like it could make for a fun breadboarding project, not that I'll likely wind up doing it myself. Is there any way to run it in emulation? Running native compilers on XT/286 hardware does indeed sound sloooow. At least it can be tested, probably automatically, in emulators…

If you go back 15 years, to January 2010, you find x86-64 hardware that you could install modern Linux on as easily as running the installer off a USB stick. The unibody 2009 MacBook Pro was out by then, so your computer could even look totally normal at a Starbuck’s today (and connect to the WiFi just fine). That machine will take 8 GB of RAM and run the COSMIC desktop written in Rust for Wayland (run it well-I have done it).

Go back another 15 years to 1995 and you have Pentium desktops. Linux was created just a few years earlier for similar but even more modest hardware (as was Windows NT). Not every modern Linux distro will still run on Pentiums but many will. You may have to burn a CD but up-to-date Antix, Q4OS, or Adelie releases will still work out of the box. Certainly NetBSD will. Not everything will still run on 32 bit machines and RAM is going to be a problem. The “modern web” would bring those machines to their knees. Still, surprisingly modern.

Go back another 15 to January 1980 and you have to wait almost two years for hardware that will even run ELKS. Crazy.

Re: ELKS: Linux for 16-bit Intel Processors

#100

Post by tengwar2 »

usr1106 wrote: post #46
tengwar2 wrote: post #29

Earlier quoted context omitted.

The Psions could do execute-in-place, including for third party software on ROM SSD - i.e. they did not have any fundamental distinction between working memory and disk storage.

Yes, NOR flash was common in these kind of devices and phones until maybe 2005 or so. So you could run code without loading it to RAM first. However, I believe RAM was still needed for stacks and heaps. Writing them to NOR flash sounds too slow.

No, I really do mean that they used the disks as ROM. No writing to it. This worked because you can just put the stack and heap at a difference position in the memory map from the executable code.
Post reply on HN