Index: llvm/trunk/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp
===================================================================
--- llvm/trunk/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp
+++ llvm/trunk/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp
@@ -311,7 +311,7 @@
     }
 
     LaneOffset = B.CreateIntrinsic(Intrinsic::amdgcn_wwm, Ty, NewV);
-    NewV = B.CreateBinOp(Op, NewV, SetInactive);
+    NewV = B.CreateBinOp(Op, SetInactive, NewV);
 
     // Read the value from the last lane, which has accumlated the values of
     // each active lane in the wavefront. This will be our new value with which
Index: llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_buffer.ll
===================================================================
--- llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_buffer.ll
+++ llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_buffer.ll
@@ -112,7 +112,7 @@
 ; GFX7LESS-NOT: v_mbcnt_hi_u32_b32
 ; GFX7LESS-NOT: s_bcnt1_i32_b64
 ; GFX7LESS: buffer_atomic_sub v{{[0-9]+}}
-; GFX8MORE: v_mov_b32_dpp v[[wave_shr1:[0-9]+]], v{{[0-9]+}} wave_shr:1 row_mask:0xf bank_mask:0xf
+; GFX8MORE: v_mov_b32_dpp v[[wave_shr1:[0-9]+]], v[[sub_value:[0-9]+]] wave_shr:1 row_mask:0xf bank_mask:0xf
 ; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v[[wave_shr1]] row_shr:1 row_mask:0xf bank_mask:0xf
 ; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v[[wave_shr1]] row_shr:2 row_mask:0xf bank_mask:0xf
 ; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v[[wave_shr1]] row_shr:3 row_mask:0xf bank_mask:0xf
@@ -120,7 +120,8 @@
 ; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v{{[0-9]+}} row_shr:8 row_mask:0xf bank_mask:0xc
 ; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v{{[0-9]+}} row_bcast:15 row_mask:0xa bank_mask:0xf
 ; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v{{[0-9]+}} row_bcast:31 row_mask:0xc bank_mask:0xf
-; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v{{[0-9]+}}, 63
+; GFX8MORE: v_sub_u32_e32 v[[sub_value]],{{( vcc,)?}} v[[sub_value]], v{{[0-9]+}}
+; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v[[sub_value]], 63
 ; GFX8MORE: v_mov_b32{{(_e[0-9]+)?}} v[[value:[0-9]+]], s[[scalar_value]]
 ; GFX8MORE: buffer_atomic_sub v[[value]]
 define amdgpu_kernel void @sub_i32_varying_vdata(i32 addrspace(1)* %out, <4 x i32> %inout) {
Index: llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
===================================================================
--- llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+++ llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
@@ -133,7 +133,9 @@
 ; GFX7LESS-NOT: v_mbcnt_hi_u32_b32
 ; GFX7LESS-NOT: s_bcnt1_i32_b64
 ; GFX7LESS: buffer_atomic_sub v{{[0-9]+}}
-; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v{{[0-9]+}}, 63
+; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v[[sub_value:[0-9]+]] wave_shr:1 row_mask:0xf bank_mask:0xf
+; GFX8MORE: v_sub_u32_e32 v[[sub_value]],{{( vcc,)?}} v[[sub_value]], v{{[0-9]+}}
+; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v[[sub_value]], 63
 ; GFX8MORE: v_mov_b32{{(_e[0-9]+)?}} v[[value:[0-9]+]], s[[scalar_value]]
 ; GFX8MORE: buffer_atomic_sub v[[value]]
 define amdgpu_kernel void @sub_i32_varying(i32 addrspace(1)* %out, i32 addrspace(1)* %inout) {
Index: llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_local_pointer.ll
===================================================================
--- llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_local_pointer.ll
+++ llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_local_pointer.ll
@@ -136,7 +136,9 @@
 ; GFX7LESS-NOT: v_mbcnt_hi_u32_b32
 ; GFX7LESS-NOT: s_bcnt1_i32_b64
 ; GFX7LESS: ds_sub_rtn_u32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}
-; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v{{[0-9]+}}, 63
+; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v[[sub_value:[0-9]+]] wave_shr:1 row_mask:0xf bank_mask:0xf
+; GFX8MORE: v_sub_u32_e32 v[[sub_value]],{{( vcc,)?}} v[[sub_value]], v{{[0-9]+}}
+; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v[[sub_value]], 63
 ; GFX8MORE: v_mov_b32{{(_e[0-9]+)?}} v[[value:[0-9]+]], s[[scalar_value]]
 ; GFX8MORE: ds_sub_rtn_u32 v{{[0-9]+}}, v{{[0-9]+}}, v[[value]]
 define amdgpu_kernel void @sub_i32_varying(i32 addrspace(1)* %out) {
Index: llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_raw_buffer.ll
===================================================================
--- llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_raw_buffer.ll
+++ llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_raw_buffer.ll
@@ -104,7 +104,9 @@
 ; GFX7LESS-NOT: v_mbcnt_hi_u32_b32
 ; GFX7LESS-NOT: s_bcnt1_i32_b64
 ; GFX7LESS: buffer_atomic_sub v{{[0-9]+}}
-; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v{{[0-9]+}}, 63
+; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v[[sub_value:[0-9]+]] wave_shr:1 row_mask:0xf bank_mask:0xf
+; GFX8MORE: v_sub_u32_e32 v[[sub_value]],{{( vcc,)?}} v[[sub_value]], v{{[0-9]+}}
+; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v[[sub_value]], 63
 ; GFX8MORE: v_mov_b32{{(_e[0-9]+)?}} v[[value:[0-9]+]], s[[scalar_value]]
 ; GFX8MORE: buffer_atomic_sub v[[value]]
 define amdgpu_kernel void @sub_i32_varying_vdata(i32 addrspace(1)* %out, <4 x i32> %inout) {
Index: llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_struct_buffer.ll
===================================================================
--- llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_struct_buffer.ll
+++ llvm/trunk/test/CodeGen/AMDGPU/atomic_optimizations_struct_buffer.ll
@@ -117,7 +117,9 @@
 ; GFX7LESS-NOT: v_mbcnt_hi_u32_b32
 ; GFX7LESS-NOT: s_bcnt1_i32_b64
 ; GFX7LESS: buffer_atomic_sub v{{[0-9]+}}
-; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v{{[0-9]+}}, 63
+; GFX8MORE: v_mov_b32_dpp v{{[0-9]+}}, v[[sub_value:[0-9]+]] wave_shr:1 row_mask:0xf bank_mask:0xf
+; GFX8MORE: v_sub_u32_e32 v[[sub_value]],{{( vcc,)?}} v[[sub_value]], v{{[0-9]+}}
+; GFX8MORE: v_readlane_b32 s[[scalar_value:[0-9]+]], v[[sub_value]], 63
 ; GFX8MORE: v_mov_b32{{(_e[0-9]+)?}} v[[value:[0-9]+]], s[[scalar_value]]
 ; GFX8MORE: buffer_atomic_sub v[[value]]
 define amdgpu_kernel void @sub_i32_varying_vdata(i32 addrspace(1)* %out, <4 x i32> %inout) {