[llvm] LangRef: Clarify llvm.minnum and llvm.maxnum about sNaN and signed zero (PR #112852)

[Phoebe Wang via llvm-commits]

phoebewang wrote:

@wzssyqa Just noticed this change due to #166912. I don't agree with the sentences:

> Although C23 clarified the behavior about sNaN and +0.0/-0.0: (NUM or NaN) vs sNaN -> qNaN +0.0 vs -0.0 -> either one of +0.0/-0.0 It is the same the IEEE754-2008's maxNUM and minNUM. Not all implementation work as expected.
> 
> Since some architectures such as aarch64/MIPSr6/LoongArch, have instructions that implements +0.0>-0.0. So Let's define llvm.minnum and llvm.maxnum to IEEE754-2008 with +0.0>-0.0.
> 
> The architectures without such instructions can implements `NSZ` flavor to speed up, and the frontend, such as clang, can call them with `nsz` attribute.

Because:

1. `nsz` attribute is not emitted by the frontend so far;
2. So it is a behavioural change to targets that doesn't implement +0.0>-0.0 like X86 instead of speed up problem;

We should not change LangRef if not all targets are ready for it.

@nikic, I noticed you have a revert commit, which never landed. I agree with your points in the commit. Any reason you decided not to revert it?

https://github.com/llvm/llvm-project/pull/112852