[llvm] [AMDGPU][True16][CodeGen] S_PACK_XX_B32_B16 lowering for true16 mode (PR #162389)

[Brox Chen via llvm-commits]

================
@@ -9115,6 +9115,63 @@ void SIInstrInfo::movePackToVALU(SIInstrWorklist &Worklist,
   MachineOperand &Src1 = Inst.getOperand(2);
   const DebugLoc &DL = Inst.getDebugLoc();
 
+  if (ST.useRealTrue16Insts()) {
+    Register SrcReg0 = Src0.getReg();
+    Register SrcReg1 = Src1.getReg();
+
+    if (!RI.isVGPR(MRI, SrcReg0)) {
----------------
broxigarchen wrote:

This is called in the si-fix-sgpr-copies pass. It's in the moveToVALU sequence that the compiler decide the move the whole def-use chain to VALU

https://github.com/llvm/llvm-project/pull/162389