Don't compute FnAbi for LLVM intrinsics in backends

[bjorn3]

This removes support for extern "unadjusted" for anything other than LLVM intrinsics. It only makes sense in the context of calling LLVM intrinsics anyway as it exposes the way the LLVM backend internally represents types. Perhaps it should be renamed to extern "llvm-intrinsic"?

Follow up to #148533

[bjorn3]

cc @sayantn this would make #140763 a tiny bit simpler I think. You don't have to modify compiler/rustc_codegen_llvm/src/abi.rs anymore.

[rustbot]

These commits modify compiler targets.
(See the Target Tier Policy.)

Some changes occurred in compiler/rustc_codegen_gcc

cc @antoyo, @GuillaumeGomez

[rustbot]

r? @wesleywiser

rustbot has assigned @wesleywiser.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[bjorn3]

I've removed the removal of computing an FnAbi for extern "unadjusted" as miri still depends on it. Will remove it in a follow up PR.

[sayantn]

Thanks a lot @bjorn3 ❤️

View changes since this review

[sayantn]

why this assertion? I mean I get the point that there are very few LLVM intrinsics that use vararg arguments, but I don't get what exactly fails for vararg functions

[bjorn3]

Because varargs are not handled. For varargs the LLVM function signature must only contain the non-varargs + an indication that the function is a vararg. Given that afaik we don't currently use any LLVM intrinsics that use varargs, I didn't implement that.