[AMDGPU] Optimze old value of v_mov_b32_dpp

We can eliminate old value if bound_ctrl = 1 and row_mask = bank_mask = 0xf.
This is alternative implementation working with the intrinsic in InstCombine.
Original review for past-ISel optimization: D46570.

Differential Revision: https://reviews.llvm.org/D46596

llvm-svn: 332956
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
index 2516179..8b7c282 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
@@ -3407,6 +3407,23 @@
     // amdgcn.kill(i1 1) is a no-op
     return eraseInstFromFunction(CI);
   }
+  case Intrinsic::amdgcn_update_dpp: {
+    Value *Old = II->getArgOperand(0);
+
+    auto BC = dyn_cast<ConstantInt>(II->getArgOperand(5));
+    auto RM = dyn_cast<ConstantInt>(II->getArgOperand(3));
+    auto BM = dyn_cast<ConstantInt>(II->getArgOperand(4));
+    if (!BC || !RM || !BM ||
+        BC->isZeroValue() ||
+        RM->getZExtValue() != 0xF ||
+        BM->getZExtValue() != 0xF ||
+        isa<UndefValue>(Old))
+      break;
+
+    // If bound_ctrl = 1, row mask = bank mask = 0xf we can omit old value.
+    II->setOperand(0, UndefValue::get(Old->getType()));
+    return II;
+  }
   case Intrinsic::stackrestore: {
     // If the save is right next to the restore, remove the restore.  This can
     // happen when variable allocas are DCE'd.
diff --git a/llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll b/llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll
index c8a0520..e66e054 100644
--- a/llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll
+++ b/llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll
@@ -1692,5 +1692,34 @@
   ret void
 }
 
+; --------------------------------------------------------------------
+; llvm.amdgcn.update.dpp.i32
+; --------------------------------------------------------------------
+
+declare i32 @llvm.amdgcn.update.dpp.i32(i32, i32, i32, i32, i32, i1)
+
+; CHECK-LABEL: {{^}}define amdgpu_kernel void @update_dpp_no_combine(
+; CHECK: @llvm.amdgcn.update.dpp.i32(i32 %in1, i32 %in2, i32 1, i32 1, i32 1, i1 false)
+define amdgpu_kernel void @update_dpp_no_combine(i32 addrspace(1)* %out, i32 %in1, i32 %in2) {
+  %tmp0 = call i32 @llvm.amdgcn.update.dpp.i32(i32 %in1, i32 %in2, i32 1, i32 1, i32 1, i1 0)
+  store i32 %tmp0, i32 addrspace(1)* %out
+  ret void
+}
+
+; CHECK-LABEL: {{^}}define amdgpu_kernel void @update_dpp_drop_old(
+; CHECK: @llvm.amdgcn.update.dpp.i32(i32 undef, i32 %in2, i32 3, i32 15, i32 15, i1 true)
+define amdgpu_kernel void @update_dpp_drop_old(i32 addrspace(1)* %out, i32 %in1, i32 %in2) {
+  %tmp0 = call i32 @llvm.amdgcn.update.dpp.i32(i32 %in1, i32 %in2, i32 3, i32 15, i32 15, i1 1)
+  store i32 %tmp0, i32 addrspace(1)* %out
+  ret void
+}
+
+; CHECK-LABEL: {{^}}define amdgpu_kernel void @update_dpp_undef_old(
+; CHECK: @llvm.amdgcn.update.dpp.i32(i32 undef, i32 %in1, i32 4, i32 15, i32 15, i1 true)
+define amdgpu_kernel void @update_dpp_undef_old(i32 addrspace(1)* %out, i32 %in1) {
+  %tmp0 = call i32 @llvm.amdgcn.update.dpp.i32(i32 undef, i32 %in1, i32 4, i32 15, i32 15, i1 1)
+  store i32 %tmp0, i32 addrspace(1)* %out
+  ret void
+}
 
 ; CHECK: attributes #5 = { convergent }