Since Linux is moving forward into the 6.1+ versions and Rust seems to be accepted as additional kernel code language we might need to move away from a "simple" GCC based toolchain to something like a LLVM toolchain as soon as any Rust code actually becomes part of the kernel and we move our targets to lets say 6.1 since that's likely the next LTS (or maybe we skip that version and wait for 6.5 or whatever) but that makes it even more inevitable that we need a "new" toolchain.
I know we have something LLVM-bpf but can that be used (or extended) to build a full working (flashable) image and if so: how?
Honestly, it doesn't seem viable for LLVM to replace GCC anytime soon as it doesn't support all of the architectures nor do I honestly see a point in doing so.
Rust code that is in 6.1 is negligible, nothing depends on it so whats the point?
Rust uses rustc anyway, though GCC merged a frontend for Rust in GCC13.
Unless somebody puts in the effort to have everything properly compiling with LLVM+Clang its not gonna happen soon nor is there need for a "new" toolchain.
Ah, I missed that GCC 13 merged Rust support. Moving to GCC 13 down the line should be a much easier transition compared to going the LLVM route.
I do understand that no Rust code yet but that is bound the change. I can even see parts of the code being rewritten in Rust to be "memory save" which is one of the biggest selling point to prefer Rust over C. Other points like more efficient because you could do the same in less lines of code seems more about how fast you can type and how familiar one is with the language.
I will mark this as "Solved" or better yet "Not needed"
BTW I was googling some more about RUST since I don't know that language (yet) and was thinking: this looks a lot like Modula-2. And to my surprise GCC 13 also has that included now
