Index: lib/Target/AMDGPU/AMDGPUISelLowering.h
===================================================================
--- lib/Target/AMDGPU/AMDGPUISelLowering.h
+++ lib/Target/AMDGPU/AMDGPUISelLowering.h
@@ -302,6 +302,7 @@
   /// Pointer to the start of the shader's constant data.
   CONST_DATA_PTR,
   SENDMSG,
+  SENDMSGHALT,
   INTERP_MOV,
   INTERP_P1,
   INTERP_P2,
Index: lib/Target/AMDGPU/AMDGPUISelLowering.cpp
===================================================================
--- lib/Target/AMDGPU/AMDGPUISelLowering.cpp
+++ lib/Target/AMDGPU/AMDGPUISelLowering.cpp
@@ -2717,6 +2717,7 @@
   NODE_NAME_CASE(KILL)
   case AMDGPUISD::FIRST_MEM_OPCODE_NUMBER: break;
   NODE_NAME_CASE(SENDMSG)
+  NODE_NAME_CASE(SENDMSGHALT)
   NODE_NAME_CASE(INTERP_MOV)
   NODE_NAME_CASE(INTERP_P1)
   NODE_NAME_CASE(INTERP_P2)
Index: lib/Target/AMDGPU/AMDGPUInstrInfo.td
===================================================================
--- lib/Target/AMDGPU/AMDGPUInstrInfo.td
+++ lib/Target/AMDGPU/AMDGPUInstrInfo.td
@@ -246,6 +246,10 @@
                     SDTypeProfile<0, 1, [SDTCisInt<0>]>,
                     [SDNPHasChain, SDNPInGlue]>;
 
+def AMDGPUsendmsghalt : SDNode<"AMDGPUISD::SENDMSGHALT",
+                    SDTypeProfile<0, 1, [SDTCisInt<0>]>,
+                    [SDNPHasChain, SDNPInGlue]>;
+
 def AMDGPUinterp_mov : SDNode<"AMDGPUISD::INTERP_MOV",
                         SDTypeProfile<1, 3, [SDTCisFP<0>]>,
                         [SDNPInGlue]>;
Index: lib/Target/AMDGPU/SIISelLowering.cpp
===================================================================
--- lib/Target/AMDGPU/SIISelLowering.cpp
+++ lib/Target/AMDGPU/SIISelLowering.cpp
@@ -2295,6 +2295,12 @@
     return DAG.getNode(AMDGPUISD::SENDMSG, DL, MVT::Other, Chain,
                        Op.getOperand(2), Glue);
   }
+  case AMDGPUIntrinsic::SI_sendmsghalt: {
+    Chain = copyToM0(DAG, Chain, DL, Op.getOperand(3));
+    SDValue Glue = Chain.getValue(1);
+    return DAG.getNode(AMDGPUISD::SENDMSGHALT, DL, MVT::Other, Chain,
+                       Op.getOperand(2), Glue);
+  }
   case AMDGPUIntrinsic::SI_tbuffer_store: {
     SDValue Ops[] = {
       Chain,
Index: lib/Target/AMDGPU/SIInsertWaits.cpp
===================================================================
--- lib/Target/AMDGPU/SIInsertWaits.cpp
+++ lib/Target/AMDGPU/SIInsertWaits.cpp
@@ -491,7 +491,7 @@
     return;
 
   // There must be "S_NOP 0" between an instruction writing M0 and S_SENDMSG.
-  if (LastInstWritesM0 && I->getOpcode() == AMDGPU::S_SENDMSG) {
+  if (LastInstWritesM0 && (I->getOpcode() == AMDGPU::S_SENDMSG || I->getOpcode() == AMDGPU::S_SENDMSGHALT)) {
     BuildMI(MBB, I, DebugLoc(), TII->get(AMDGPU::S_NOP)).addImm(0);
     LastInstWritesM0 = false;
     return;
@@ -591,7 +591,8 @@
       // but we also want to wait for any other outstanding transfers before
       // signalling other hardware blocks
       if (I->getOpcode() == AMDGPU::S_BARRIER ||
-          I->getOpcode() == AMDGPU::S_SENDMSG)
+          I->getOpcode() == AMDGPU::S_SENDMSG ||
+          I->getOpcode() == AMDGPU::S_SENDMSGHALT)
         Required = LastIssued;
       else
         Required = handleOperands(*I);
Index: lib/Target/AMDGPU/SIInstructions.td
===================================================================
--- lib/Target/AMDGPU/SIInstructions.td
+++ lib/Target/AMDGPU/SIInstructions.td
@@ -508,9 +508,11 @@
   def S_SENDMSG : SOPP <0x00000010, (ins SendMsgImm:$simm16), "s_sendmsg $simm16",
       [(AMDGPUsendmsg (i32 imm:$simm16))]
   >;
-} // End Uses = [EXEC, M0]
 
-def S_SENDMSGHALT : SOPP <0x00000011, (ins SendMsgImm:$simm16), "s_sendmsghalt $simm16">;
+  def S_SENDMSGHALT : SOPP <0x00000011, (ins SendMsgImm:$simm16), "s_sendmsghalt $simm16",
+      [(AMDGPUsendmsghalt (i32 imm:$simm16))]
+  >;
+} // End Uses = [EXEC, M0]
 def S_TRAP : SOPP <0x00000012, (ins i16imm:$simm16), "s_trap $simm16">;
 def S_ICACHE_INV : SOPP <0x00000013, (ins), "s_icache_inv"> {
 	let simm16 = 0;
Index: lib/Target/AMDGPU/SIIntrinsics.td
===================================================================
--- lib/Target/AMDGPU/SIIntrinsics.td
+++ lib/Target/AMDGPU/SIIntrinsics.td
@@ -52,6 +52,7 @@
     [IntrReadMem, IntrArgMemOnly]>;
 
   def int_SI_sendmsg : Intrinsic <[], [llvm_i32_ty, llvm_i32_ty], []>;
+  def int_SI_sendmsghalt : Intrinsic <[], [llvm_i32_ty, llvm_i32_ty], []>;
 
   // Fully-flexible SAMPLE instruction.
   class SampleRaw : Intrinsic <
Index: test/CodeGen/AMDGPU/llvm.SI.sendmsg-m0.ll
===================================================================
--- test/CodeGen/AMDGPU/llvm.SI.sendmsg-m0.ll
+++ test/CodeGen/AMDGPU/llvm.SI.sendmsg-m0.ll
@@ -12,6 +12,18 @@
   ret void
 }
 
+; GCN-LABEL: {{^}}main_halt:
+; GCN: s_mov_b32 m0, s0
+; VI-NEXT: s_nop 0
+; GCN-NEXT: s_sendmsghalt sendmsg(MSG_INTERRUPT)
+; GCN-NEXT: s_endpgm
+
+define  void @main_halt(i32 inreg %a) #0 {
+  call void @llvm.SI.sendmsghalt(i32 1, i32 %a)
+  ret void
+}
+
 declare void @llvm.SI.sendmsg(i32, i32) #0
+declare void @llvm.SI.sendmsghalt(i32, i32) #0
 
 attributes #0 = { nounwind }
Index: test/CodeGen/AMDGPU/llvm.SI.sendmsg.ll
===================================================================
--- test/CodeGen/AMDGPU/llvm.SI.sendmsg.ll
+++ test/CodeGen/AMDGPU/llvm.SI.sendmsg.ll
@@ -4,21 +4,35 @@
 ; CHECK-LABEL: {{^}}main:
 ; CHECK: s_mov_b32 m0, 0
 ; CHECK-NOT: s_mov_b32 m0
+; CHECK: s_sendmsg sendmsg(MSG_INTERRUPT)
 ; CHECK: s_sendmsg sendmsg(MSG_GS, GS_OP_EMIT, 0)
 ; CHECK: s_sendmsg sendmsg(MSG_GS, GS_OP_CUT, 1)
 ; CHECK: s_sendmsg sendmsg(MSG_GS, GS_OP_EMIT_CUT, 2)
 ; CHECK: s_sendmsg sendmsg(MSG_GS_DONE, GS_OP_NOP)
+; CHECK: s_sendmsghalt sendmsg(MSG_INTERRUPT)
+; CHECK: s_sendmsghalt sendmsg(MSG_GS, GS_OP_EMIT, 0)
+; CHECK: s_sendmsghalt sendmsg(MSG_GS, GS_OP_CUT, 1)
+; CHECK: s_sendmsghalt sendmsg(MSG_GS, GS_OP_EMIT_CUT, 2)
+; CHECK: s_sendmsghalt sendmsg(MSG_GS_DONE, GS_OP_NOP)
 
 define void @main() {
 main_body:
+  call void @llvm.SI.sendmsg(i32 1, i32 0);
   call void @llvm.SI.sendmsg(i32 34, i32 0);
   call void @llvm.SI.sendmsg(i32 274, i32 0);
   call void @llvm.SI.sendmsg(i32 562, i32 0);
   call void @llvm.SI.sendmsg(i32 3, i32 0);
+
+  call void @llvm.SI.sendmsghalt(i32 1, i32 0);
+  call void @llvm.SI.sendmsghalt(i32 34, i32 0);
+  call void @llvm.SI.sendmsghalt(i32 274, i32 0);
+  call void @llvm.SI.sendmsghalt(i32 562, i32 0);
+  call void @llvm.SI.sendmsghalt(i32 3, i32 0);
   ret void
 }
 
 ; Function Attrs: nounwind
 declare void @llvm.SI.sendmsg(i32, i32) #0
+declare void @llvm.SI.sendmsghalt(i32, i32) #0
 
 attributes #0 = { nounwind }