[llvm] [LangRef] Clarify specification for float min/max operations (PR #172012)

[Phoebe Wang via llvm-commits]

phoebewang wrote:

Ping @nikic . It's one week before LLVM 22.x branch now. I don't see my concerns got addressed, so I'm opposed to the current statement regarding signed zero. Is it able to make consensus on back to the last revision in a short time? Otherwise, reverting #112852 seems the straightforward solution.

https://github.com/llvm/llvm-project/pull/172012