Bug#1082727: llvm-18: LLVM miscompiles certain floating point library calls

[Timothy Pearson]

Upstream bug report: [ https://github.com/llvm/llvm-project/issues/108904 | 
https://github.com/llvm/llvm-project/issues/108904 ] 
Upstream pull request: https://github.com/llvm/llvm-project/pull/109134