[clang] [llvm] [AMDGPU][clang][CodeGen][opt] Add late-resolved feature identifying predicates (PR #134016)

[Alex Voicu via cfe-commits]

AlexVlx wrote:

> 
> You could default the warning to an error to make it more visible to the user. Not certain if that's a bad idea or not though.
> 

Possibly. I had to drop this for a bit so circling around again, so apologies for the late reply. In the meanwhile, @epilk pointed out something I spaced out on, which is that we have some related / somewhat similar feature in Clang already: https://clang.llvm.org/docs/LanguageExtensions.html#objective-c-available, which is permissive along the lines of what you and @erichkeane are suggesting (a bit too permissive for my colleagues, I suspect, but that can be addressed by erroring). The Sema checking around that is rather extensive and I'd like to avoid duplicating / adding that amount of code.

> > If I may be so bold as to inquire: would you and @AaronBallman be slightly less horrified if the return type variance would be replaced with returning an odd type that only knows how to `bool`ify itself in conditions? More explicitly, if instead of `void __builtin_amdgcn_processor_is(const char*)` what we see is `__amdgpu_predicate_t __builtin_amdgcn_processor_is(const char*)`, would that be somewhat less bad? There is precedent for special-ish builtins returning special-ish target types (please consider `__amdgpu_buffer_rsrc_t` for `__builtin_amdgcn_make_buffer_rsrc` or `svcount_t`)
> 
> That would be just as weird, IMO. Having something that's contextually converted to bool but only in some contexts, is going to be confusing in practice.

Well, there's an interim solution here where it'd actually look essentially like so:

```cpp
struct __amdgpu_predicate_t { explicit operator bool() const noexcept; }
```

Which forces the user to consciously cast, unless in a valid (e.g. `if` context), and allows for the Sema checking to be funneled into a single point (the cast). Obligatory handwavium godbolt for illustration: <https://gcc.godbolt.org/z/cGT7TPrjE>


https://github.com/llvm/llvm-project/pull/134016