[llvm-branch-commits] [llvm] [AMDGPU][True16] Add legalization/selection handling for G_MERGE_VALUES of 2 s16 -> s32 (PR #200082)

[Petar Avramovic via llvm-branch-commits]

================
@@ -461,6 +461,8 @@ class AMDGPUInstructionSelector final : public InstructionSelector {
   /// Match an any extend from a 32-bit value to 64-bit.
   Register matchAnyExtendFromS32(Register Reg) const;
 
+  bool selectS16MergeToS32(MachineInstr &MI) const;
----------------
petar-avramovic wrote:

move before selectG_MERGE_VALUES

https://github.com/llvm/llvm-project/pull/200082