Debian GNU/Hurd Adds Experimental 32-bit SMP Kernel & Rust Compiler

[phoronix]

Phoronix: Debian GNU/Hurd Adds Experimental 32-bit SMP Kernel & Rust Compiler

The GNU Hurd team has put out their Q2'2024 status update to outline recent activity around this micro-kernel platform...

https://www.phoronix.com/news/Hurd-Progress-Q2-2024

[waxhead]

Personally I am a fan of the ***theoretical*** benefits of a microkernel. So glad to hear some news about gnu/hurd for once.

Not sure how Hurd compares to Minix3 these days, but it would be great to see some of the Minix3 concepts in a Debian OS at some point. Maybe Hurd will be the key to getting that done sooner or ...well... Later

[Eirikr1848]

Personally I am a fan of the ***theoretical*** benefits of a microkernel. So glad to hear some news about gnu/hurd for once.

Not sure how Hurd compares to Minix3 these days, but it would be great to see some of the Minix3 concepts in a Debian OS at some point. Maybe Hurd will be the key to getting that done sooner or ...well... Later

Hey if Hurd is moving forward with finally having multi core support?

Then Debian/Minix or even Debian/Illumos are not far fetched.

[Noitatsidem]

Still sad that they killed debian GNU/kfreebsd - that perhaps was the more interesting project. Imagine a kernel with ZFS support on a distro with the enormous prebuilt archive that is the debian repos.

Either way, SMP support on hurd is an exciting development!

[kpedersen]

Still sad that they killed debian GNU/kfreebsd - that perhaps was the more interesting project. Imagine a kernel with ZFS support on a distro with the enormous prebuilt archive that is the debian repos.

I was always interested in the other way round. An extremely compatible kernel with fantastic hardware support. However, with a userland that doesn't smell of wee.

[Ironmask]

I know they didn't say they're adding Rust *into* the project itself, but that honestly might revive some interest and get new developers on the project. I mentioned in another HURD thread that HURD is unpopular because there's nothing new, exciting or experimental about it, it's only purpose is to be a replacement for a proprietary kernel, and Linux did that long ago. Adding Rust into existing projects is a relatively new field of exploration (language-oriented programming has always been a fascination to people but there has never been a good excuse to put it into practice until now). Exploring ways to augment HURD with the language and accelerate it's development would probably draw some interest and new blood.

[skeevy420]

Still sad that they killed debian GNU/kfreebsd - that perhaps was the more interesting project. Imagine a kernel with ZFS support on a distro with the enormous prebuilt archive that is the debian repos.

Either way, SMP support on hurd is an exciting development!

That's called CachyOS. Ubuntu also has ZFS support, but it's nowhere near as good as what CachyOS offers. There's also NixOS, but that's not as easy or user friendly as Ubuntu or CachyOS.

[aviallon]

That's called CachyOS. Ubuntu also has ZFS support, but it's nowhere near as good as what CachyOS offers. There's also NixOS, but that's not as easy or user friendly as Ubuntu or CachyOS.

Depends on who sets up your NixOS install 😁
NixOS is probably the most innovative distribution of all time.

[ayumu]

Reminder that the hurd has not yet moved away from Mach, thus it has not progressed nor will it progress.

I would look at healthier systems like Genode or even Managarm instead.

[kpedersen]

Reminder that the hurd has not yet moved away from Mach, thus it has not progressed nor will it progress.

Apple's XNU kernel underpinning macOS is a derivative of Mach. That seems to have progressed pretty well.

Mach is actually an increadibly interesting design that allows compromise and pragmatism with the whole microkernel stuff. I think if the GNU/FOSS ecosystem truely wanted to (and didn't have the decent Linux kernel already), it would have worked well.