[llvm] AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (PR #104486)

[Matt Arsenault via llvm-commits]

================
@@ -2180,7 +2180,7 @@ bool AMDGPULegalizerInfo::legalizeCustom(
   case TargetOpcode::G_CTLZ_ZERO_UNDEF:
     return legalizeCTLZ_ZERO_UNDEF(MI, MRI, B);
   case TargetOpcode::G_INTRINSIC_FPTRUNC_ROUND:
-    return legalizeFPTruncRound(MI, B);
+    return legalizeFPTruncRound(MI, MRI, B);
----------------
arsenm wrote:

Don't need to add this parameter, you can just get MRI from the MachineIRBuilder 

https://github.com/llvm/llvm-project/pull/104486