[llvm] Update foldFMulReassoc to respect absent fast-math flags (PR #88589)

[Joshua Cranmer via llvm-commits]

https://github.com/jcranmer-intel commented:

It kind of would be nice if we could get a matcher that checked the FMF flags of everything in the chain of instructions, but I don't think the matcher infrastructure is set up in a way that makes such a thing possible.

But overall, this is something desperately needed in more of our fast-math optimizatoins.

https://github.com/llvm/llvm-project/pull/88589