Index: lib/CodeGen/SelectionDAG/DAGCombiner.cpp
===================================================================
--- lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -4539,15 +4539,17 @@
   }
 
   // fold (shl (add x, c1), c2) -> (add (shl x, c2), c1 << c2)
+  // fold (shl (or x, c1), c2) -> (or (shl x, c2), c1 << c2)
   // Variant of version done on multiply, except mul by a power of 2 is turned
   // into a shift.
   APInt Val;
-  if (N1C && N0.getOpcode() == ISD::ADD && N0.getNode()->hasOneUse() &&
+  if (N1C && (N0.getOpcode() == ISD::ADD || N0.getOpcode() == ISD::OR) &&
+      N0.getNode()->hasOneUse() &&
       (isa<ConstantSDNode>(N0.getOperand(1)) ||
        isConstantSplatVector(N0.getOperand(1).getNode(), Val))) {
     SDValue Shl0 = DAG.getNode(ISD::SHL, SDLoc(N0), VT, N0.getOperand(0), N1);
     SDValue Shl1 = DAG.getNode(ISD::SHL, SDLoc(N1), VT, N0.getOperand(1), N1);
-    return DAG.getNode(ISD::ADD, SDLoc(N), VT, Shl0, Shl1);
+    return DAG.getNode(N0.getOpcode(), SDLoc(N), VT, Shl0, Shl1);
   }
 
   // fold (shl (mul x, c1), c2) -> (mul x, c1 << c2)
Index: test/CodeGen/AMDGPU/llvm.amdgcn.ds.bpermute.ll
===================================================================
--- test/CodeGen/AMDGPU/llvm.amdgcn.ds.bpermute.ll
+++ test/CodeGen/AMDGPU/llvm.amdgcn.ds.bpermute.ll
@@ -21,4 +21,27 @@
   ret void
 }
 
+; CHECK-LABEL: {{^}}ds_bpermute_add_shl:
+; CHECK: ds_bpermute_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset:4
+; CHECK: s_waitcnt lgkmcnt
+define void @ds_bpermute_add_shl(i32 addrspace(1)* %out, i32 %base_index, i32 %src) nounwind {
+  %index = add i32 %base_index, 1
+  %byte_index = shl i32 %index, 2
+  %bpermute = call i32 @llvm.amdgcn.ds.bpermute(i32 %byte_index, i32 %src) #0
+  store i32 %bpermute, i32 addrspace(1)* %out, align 4
+  ret void
+}
+
+; CHECK-LABEL: {{^}}ds_bpermute_or_shl:
+; CHECK: ds_bpermute_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset:4
+; CHECK: s_waitcnt lgkmcnt
+define void @ds_bpermute_or_shl(i32 addrspace(1)* %out, i32 %base_index, i32 %src) nounwind {
+  %masked = and i32 %base_index, 62
+  %index = or i32 %masked, 1
+  %byte_index = shl i32 %index, 2
+  %bpermute = call i32 @llvm.amdgcn.ds.bpermute(i32 %byte_index, i32 %src) #0
+  store i32 %bpermute, i32 addrspace(1)* %out, align 4
+  ret void
+}
+
 attributes #0 = { nounwind readnone convergent }
Index: test/CodeGen/AMDGPU/shl.ll
===================================================================
--- test/CodeGen/AMDGPU/shl.ll
+++ test/CodeGen/AMDGPU/shl.ll
@@ -377,4 +377,31 @@
   ret void
 }
 
+; FUNC-LABEL: {{^}}shl_or_k:
+; SI: s_lshl_b32 [[SHL:s[0-9]+]], s{{[0-9]+}}, 2
+; SI: s_or_b32 [[OR:s[0-9]+]], [[SHL]], 4
+; SI: v_mov_b32_e32 [[VOR:v[0-9]+]], [[OR]]
+; SI: buffer_store_dword [[VOR]]
+define void @shl_or_k(i32 addrspace(1)* %out, i32 %in) {
+  %tmp0 = or i32 %in, 1
+  %tmp2 = shl i32 %tmp0, 2
+  store i32 %tmp2, i32 addrspace(1)* %out
+  ret void
+}
+
+; FUNC-LABEL: {{^}}shl_or_k_two_uses:
+; SI: s_or_b32 [[OR:s[0-9]+]], s{{[0-9]+}}, 1
+; SI: s_lshl_b32 [[SHL:s[0-9]+]], [[OR]], 2
+; SI-DAG: v_mov_b32_e32 [[VOR:v[0-9]+]], [[OR]]
+; SI-DAG: v_mov_b32_e32 [[VSHL:v[0-9]+]], [[SHL]]
+; SI-DAG: buffer_store_dword [[VOR]]
+; SI-DAG: buffer_store_dword [[VSHL]]
+define void @shl_or_k_two_uses(i32 addrspace(1)* %out0, i32 addrspace(1)* %out1, i32 %in) {
+  %tmp0 = or i32 %in, 1
+  %tmp2 = shl i32 %tmp0, 2
+  store i32 %tmp2, i32 addrspace(1)* %out0
+  store i32 %tmp0, i32 addrspace(1)* %out1
+  ret void
+}
+
 attributes #0 = { nounwind readnone }