[llvm] [InstCombine] Ensure Safe Handling of Flags in foldFNegIntoConstant (PR #94148)

[Jay Foad via llvm-commits]

jayfoad wrote:

> Since it was breaking the `fsub` test with the `reassoc` flag, I conditionally propagated the `FNegOp` FMF flag.

Maybe we should just update the expected output for that test? It is not at all clear to me that propagating `reassoc` from the fsub to the fmul is desirable or even correct.

https://github.com/llvm/llvm-project/pull/94148