[llvm] [NFC][AMDGPU] Let IR level callers query the FMA/FMAD predicates (PR #213310)

[Matt Arsenault via llvm-commits]

================
@@ -7421,7 +7433,7 @@ LLT SITargetLowering::getPreferredShiftAmountTy(LLT Ty) const {
 // a fast fma device and require denormals.
 //
 bool SITargetLowering::isFMAFasterThanFMulAndFAdd(
-    EVT VT, bool FlushF32Denormals, bool FlushF64F16Denormals) const {
+    EVT VT, const SIModeRegisterDefaults &Mode) const {
----------------
arsenm wrote:

I'd rather keep this in terms of DenormalFPEnv or the machine function, not spread references to SIModeRegisterDefaults

https://github.com/llvm/llvm-project/pull/213310