[llvm] [AMDGPU][True16][CodeGen] Implement sgpr folding in true16 (PR #128929)

[Christudasan Devadasan via llvm-commits]

================
@@ -1073,9 +1095,43 @@ void SIFoldOperandsImpl::foldOperand(
         UseMI->getOperand(0).getReg().isVirtual() &&
         !UseMI->getOperand(1).getSubReg()) {
       LLVM_DEBUG(dbgs() << "Folding " << OpToFold << "\n into " << *UseMI);
+      unsigned Size = TII->getOpSize(*UseMI, 1);
       Register UseReg = OpToFold.getReg();
       UseMI->getOperand(1).setReg(UseReg);
-      UseMI->getOperand(1).setSubReg(OpToFold.getSubReg());
+      unsigned SubRegIdx = OpToFold.getSubReg();
+      // Hack to allow 32-bit SGPRs to be folded into True16 instructions
+      // Remove this if 16-bit SGPRs (i.e. SGPR_LO16) are added to the
+      // VS_16RegClass
+      //
+      // Excerpt from AMDGPUGenRegisterInfo.inc
+      // NoSubRegister, //0
+      // hi16, // 1
+      // lo16, // 2
+      // sub0, // 3
+      // ...
+      // sub1, // 11
+      // sub1_hi16, // 12
+      // sub1_lo16, // 13
+      static_assert(AMDGPU::sub1_hi16 == 12, "Subregister layout has changed");
----------------
cdevadas wrote:

I see your point, and I don't have a better solution to recommend here. But it looks like any sub-reg layout change in the future needed a fixup in this hardcoded value.

https://github.com/llvm/llvm-project/pull/128929