4 ms·

If we're going to write RustBSD, why not RustMinix -- and get the benefits of microkernel design? Isn't it about time we give up a few percent performance for

If we're going to write RustBSD, why not RustMinix -- and get the benefits of microkernel design? Isn't it about time we give up a few percent performance for robustness?

Pet_Ant 2y ago
Minix is sadly dead. Minix on RISC-V seems like it would have been a pedagogical gold mine and yet it hasn’t happened last I checked.
- boricj 2y ago
  xv6 is now targeting RISC-V and has displaced MINIX as a better pedagogical tool in my opinion. MINIX is dead for multiple reasons, including technical ones [1]. Simply put, its design and implementation are no longer fit for purpose on modern systems. At this point, starting over from scratch would probably be more effective than trying to address all of its issues. [1] https://news.ycombinator.com/item?id=34916261 https://news.ycombinator.com/item?id=34916261
ggm 2y ago
If it has an ABI which means POSIX tools just work, I'm fine if its a microkernel. MACH was at the heart of lots of things too. We could do capabilities based OS design you name it. 9P, I don't care. What I care about is getting an interactive shell prompt and a filesystem with . and .. and / and BOURNE shell compatible semantics, and an ability to run vi and emacs. (I am less sure plan9 people care about shell btw they went rc and then into some place I don't understand up in the 5th dimension)
- bakul 2y ago
  Did anybody mention plan9? At any rate it is not a microkernel. There is at least one effort implementing v6 unix in rust. That would probably meet your minimum requirements!
yjftsjthsd-h 2y ago
> why not RustMinix -- and get the benefits of microkernel design? A person could reasonably argue that this is redox. It's different in that it's a new OS rather than a rust-ified version of an existing thing, but still, it's a microkernel unix-ish OS in rust.

4 ms·

If we're going to write RustBSD, why not RustMinix -- and get the benefits of microkernel design? Isn't it about time we give up a few percent performance for

by GeorgeTirebiter 2y ago

If we're going to write RustBSD, why not RustMinix -- and get the benefits of microkernel design? Isn't it about time we give up a few percent performance for robustness?

Pet_Ant 2y ago
Minix is sadly dead. Minix on RISC-V seems like it would have been a pedagogical gold mine and yet it hasn’t happened last I checked.
- boricj 2y ago
  xv6 is now targeting RISC-V and has displaced MINIX as a better pedagogical tool in my opinion. MINIX is dead for multiple reasons, including technical ones [1]. Simply put, its design and implementation are no longer fit for purpose on modern systems. At this point, starting over from scratch would probably be more effective than trying to address all of its issues. [1] https://news.ycombinator.com/item?id=34916261 https://news.ycombinator.com/item?id=34916261
ggm 2y ago
If it has an ABI which means POSIX tools just work, I'm fine if its a microkernel. MACH was at the heart of lots of things too. We could do capabilities based OS design you name it. 9P, I don't care. What I care about is getting an interactive shell prompt and a filesystem with . and .. and / and BOURNE shell compatible semantics, and an ability to run vi and emacs. (I am less sure plan9 people care about shell btw they went rc and then into some place I don't understand up in the 5th dimension)
- bakul 2y ago
  Did anybody mention plan9? At any rate it is not a microkernel. There is at least one effort implementing v6 unix in rust. That would probably meet your minimum requirements!
yjftsjthsd-h 2y ago
> why not RustMinix -- and get the benefits of microkernel design? A person could reasonably argue that this is redox. It's different in that it's a new OS rather than a rust-ified version of an existing thing, but still, it's a microkernel unix-ish OS in rust.