Tracking Issue for asm_experimental_arch

[Amanieu]

The feature gate for the issue is #![feature(asm_experimental_arch)].

Summary

This feature tracks asm! and global_asm! support for the following architectures:

• Arm64ec: Stabilize Arm64EC inline assembly #131781
• NVPTX
• PowerPC: Stabilize ppc inline assembly #147996
• Hexagon
• MIPS32r2 and MIPS64r2
• wasm32, more discussion at Moving WebAssembly inline assembly forward #136382
• BPF
• SPIR-V
• AVR
• MSP430
• s390x: Stabilize s390x inline assembly #131258

Status

Each architecture needs to be reviewed before stabilization:

• It must have clobber_abi.
• It must be possible to clobber every register that is normally clobbered by a function call.
• Generally review that the exposed register classes make sense.

See #93335 (comment) for a detailed todo list.

[solomatov]

Is there any chance that wasm32 assembly support will be stabilized any time soon? Are there any tasks where contributions might speed this up?

[matveitolochko]

Is there any news about AVR ? If not, where and how can I contribute to this?

[tgross35]

The top post mentions the following:

Each architecture needs to be reviewed before stabilization:
* It must have clobber_abi.
* It must be possible to clobber every register that is normally clobbered by a function call.
* Generally review that the exposed register classes make sense.

Anyone can post a stabilization (or status) report answering the above, probably with examples/tests, and that would be a good place for any platform to start. It would also be good to make sure there aren't any open LLVM issues, and mention it there.

The bar for enabling AVR should be pretty low since it is tier 3. I am not a decision maker here but I think any of the tier 3 targets could probably be stabilized with just the report and some real-world usage backing it up.

I'm not sure what exactly would be needed for the tier 2 and above targets, e.g. WASM as requested by @solomatov. Having a better picture of LLVM support is more important. @alexcrichton maybe you have an idea about the WASM picture here, or know who would?

[tgross35]

Cc @ecnelises with the above in case you are interested in moving ppc assembly forward, based on linked issues it seems like there may be some ecosystem use cases.