AMDGPU: Remove pointless wrapper nodes for init.exec intrinsics

llvm-svn: 371364

diff --git a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
index 29a3d24..ed2a98c 100644 (file)
--- a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
+++ b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
@@ -175,6 +175,7 @@ def int_amdgcn_implicit_buffer_ptr :
 
 // Set EXEC to the 64-bit value given.
 // This is always moved to the beginning of the basic block.
+// FIXME: Should be mangled for wave size.
 def int_amdgcn_init_exec : Intrinsic<[],
   [llvm_i64_ty],      // 64-bit literal constant
   [IntrConvergent, ImmArg<0>]>;

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp b/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
index 04f1da2..83086db 100644 (file)
--- a/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
@@ -4351,8 +4351,6 @@ const char* AMDGPUTargetLowering::getTargetNodeName(unsigned Opcode) const {
   NODE_NAME_CASE(KILL)
   NODE_NAME_CASE(DUMMY_CHAIN)
   case AMDGPUISD::FIRST_MEM_OPCODE_NUMBER: break;
-  NODE_NAME_CASE(INIT_EXEC)
-  NODE_NAME_CASE(INIT_EXEC_FROM_INPUT)
   NODE_NAME_CASE(INTERP_MOV)
   NODE_NAME_CASE(INTERP_P1)
   NODE_NAME_CASE(INTERP_P2)

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h b/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h
index b0ca554..54ef263 100644 (file)
--- a/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h
+++ b/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h
@@ -479,8 +479,6 @@ enum NodeType : unsigned {
   BUILD_VERTICAL_VECTOR,
   /// Pointer to the start of the shader's constant data.
   CONST_DATA_PTR,
-  INIT_EXEC,
-  INIT_EXEC_FROM_INPUT,
   INTERP_MOV,
   INTERP_P1,
   INTERP_P2,

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td b/llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td
index 7da37f9..8f6cafa 100644 (file)
--- a/llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td
+++ b/llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td
@@ -330,15 +330,6 @@ def AMDGPUfdot2 : SDNode<"AMDGPUISD::FDOT2",
 
 def AMDGPUperm : SDNode<"AMDGPUISD::PERM", AMDGPUDTIntTernaryOp, []>;
 
-def AMDGPUinit_exec : SDNode<"AMDGPUISD::INIT_EXEC",
-                      SDTypeProfile<0, 1, [SDTCisInt<0>]>,
-                      [SDNPHasChain, SDNPInGlue]>;
-
-def AMDGPUinit_exec_from_input : SDNode<"AMDGPUISD::INIT_EXEC_FROM_INPUT",
-                                 SDTypeProfile<0, 2,
-                                 [SDTCisInt<0>, SDTCisInt<1>]>,
-                                 [SDNPHasChain, SDNPInGlue]>;
-
 def AMDGPUinterp_mov : SDNode<"AMDGPUISD::INTERP_MOV",
                         SDTypeProfile<1, 3, [SDTCisFP<0>]>,
                         [SDNPInGlue]>;

diff --git a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
index 7430e87..ba60311 100644 (file)
--- a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
@@ -6765,14 +6765,6 @@ SDValue SITargetLowering::LowerINTRINSIC_VOID(SDValue Op,
       AMDGPUISD::EXPORT : AMDGPUISD::EXPORT_DONE;
     return DAG.getNode(Opc, DL, Op->getVTList(), Ops);
   }
-  case Intrinsic::amdgcn_init_exec: {
-    return DAG.getNode(AMDGPUISD::INIT_EXEC, DL, MVT::Other, Chain,
-                       Op.getOperand(2));
-  }
-  case Intrinsic::amdgcn_init_exec_from_input: {
-    return DAG.getNode(AMDGPUISD::INIT_EXEC_FROM_INPUT, DL, MVT::Other, Chain,
-                       Op.getOperand(2), Op.getOperand(3));
-  }
   case Intrinsic::amdgcn_s_barrier: {
     if (getTargetMachine().getOptLevel() > CodeGenOpt::None) {
       const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();

diff --git a/llvm/lib/Target/AMDGPU/SIInstructions.td b/llvm/lib/Target/AMDGPU/SIInstructions.td
index 934b50b..90fbb17 100644 (file)
--- a/llvm/lib/Target/AMDGPU/SIInstructions.td
+++ b/llvm/lib/Target/AMDGPU/SIInstructions.td
@@ -356,6 +356,7 @@ def SI_INIT_EXEC : SPseudoInstSI <
   let WaveSizePredicate = isWave64;
 }
 
+// FIXME: Intrinsic should be mangled for wave size.
 def SI_INIT_EXEC_LO : SPseudoInstSI <
   (outs), (ins i32imm:$src), []> {
   let Defs = [EXEC_LO];
@@ -609,21 +610,19 @@ def : GCNPat <
 >;
 
 def : GCNPat <
-  (AMDGPUinit_exec i64:$src),
-  (SI_INIT_EXEC (as_i64imm $src))
-> {
+  (int_amdgcn_init_exec i64:$src),
+  (SI_INIT_EXEC (as_i64imm $src))> {
   let WaveSizePredicate = isWave64;
 }
 
 def : GCNPat <
-  (AMDGPUinit_exec i64:$src),
-  (SI_INIT_EXEC_LO (as_i32imm $src))
-> {
+  (int_amdgcn_init_exec i64:$src),
+  (SI_INIT_EXEC_LO (as_i32imm $src))> {
   let WaveSizePredicate = isWave32;
 }
 
 def : GCNPat <
-  (AMDGPUinit_exec_from_input i32:$input, i32:$shift),
+  (int_amdgcn_init_exec_from_input i32:$input, i32:$shift),
   (SI_INIT_EXEC_FROM_INPUT (i32 $input), (as_i32imm $shift))
 >;