[clang] [clang] Fix invalid CodeGen for atomic fetch min/max on floating point values (PR #162011)

[Eli Friedman via cfe-commits]

https://github.com/efriedma-quic requested changes to this pull request.

This is already under review in #160330.  (And you can't use fcmp.)

https://github.com/llvm/llvm-project/pull/162011