Re: low-level Rust, I've been really enjoying following this series:
It shows a variety of ways to safely wrap unsafe code and then provide strongly-typed safe interfaces to the hardware, no pointer twiddling required. A particularly good example is the section "Some Clever Solution" in the "Page Tables" post.
Ooh nice guide. Thanks for the link!
I was looking at the printing to screen section:
http://os.phil-opp.com/printing-to-screen.html
It is so helpful to have the explanation of what is going and why the code looks the way it is, what are some alternatives and why those wouldn't work.
Sorry, I think I downvoted you by accident. HN's mobile interface makes scrolling risky... :)
Re:Robigalia: seL4 and Rust
by dikaiosune on Wednesday January 06, 2016 @06:21PM
(#10854718)
We've all been there :).
Multicore support + verification is in the works, see https://ssrg.nicta.com.au/publications/nictaabstracts/Peters...
More info about L4 micro-kernel:
https://sel4.systems/Info/FAQ/
Now user-level here means drivers as well. So that would be interesting how it would work with Rust. Will it all default to unsafe code and just end up with pointer twiddling and memory overwrites and all the same errors. I hope not.
Currently it seems that the verification mode had DMA turned off and for SMP they run a kernel instance on each core and then send messages to each other via shared memory. That probably means not load balancing or not scheduling?