[llvm-branch-commits] [llvm] AMDGPU/GlobalISel: RBLegalize rules for load (PR #112882)

[Matt Arsenault via llvm-branch-commits]

================
@@ -119,6 +210,53 @@ void RegBankLegalizeHelper::lower(MachineInstr &MI,
     MI.eraseFromParent();
     break;
   }
+  case SplitLoad: {
+    LLT DstTy = MRI.getType(MI.getOperand(0).getReg());
+    LLT V8S16 = LLT::fixed_vector(8, S16);
+    LLT V4S32 = LLT::fixed_vector(4, S32);
+    LLT V2S64 = LLT::fixed_vector(2, S64);
+
+    if (DstTy == LLT::fixed_vector(8, S32))
----------------
arsenm wrote:

Can you rework this into be a function that returns the type to use for the load? 

https://github.com/llvm/llvm-project/pull/112882