diff --git a/llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp b/llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp
--- a/llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp
@@ -822,7 +822,7 @@
   Register Denom = MI.getOperand(4).getReg();
   unsigned ChooseDenom = MI.getOperand(5).getImm();
 
-  Register Src0 = ChooseDenom != 0 ? Numer : Denom;
+  Register Src0 = ChooseDenom == 0 ? Numer : Denom;
 
   auto MIB = BuildMI(*MBB, &MI, DL, TII.get(Opc), Dst0)
     .addDef(Dst1)
diff --git a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
--- a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
@@ -6007,7 +6007,7 @@
     // intrinsic has the numerator as the first operand to match a normal
     // division operation.
 
-    SDValue Src0 = Param->isAllOnesValue() ? Numerator : Denominator;
+    SDValue Src0 = !Param->isAllOnesValue() ? Numerator : Denominator;
 
     return DAG.getNode(AMDGPUISD::DIV_SCALE, DL, Op->getVTList(), Src0,
                        Denominator, Numerator);
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.div.scale.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.div.scale.ll
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.div.scale.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.div.scale.ll
@@ -18,7 +18,7 @@
 ; GFX7-NEXT:    buffer_load_dword v0, v[0:1], s[4:7], 0 addr64 offset:4
 ; GFX7-NEXT:    s_mov_b32 s6, -1
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v0, v0, v2
+; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v2, v0, v2
 ; GFX7-NEXT:    s_mov_b64 s[2:3], s[6:7]
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7-NEXT:    s_endpgm
@@ -38,7 +38,7 @@
 ; GFX8-NEXT:    flat_load_dword v0, v[0:1]
 ; GFX8-NEXT:    flat_load_dword v1, v[2:3]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v1, v1, v0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, v1, v0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -60,7 +60,7 @@
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    global_load_dword v1, v[2:3], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, v1, v1, v0
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, v0, v1, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -92,7 +92,7 @@
 ; GFX7-NEXT:    buffer_load_dword v0, v[0:1], s[4:7], 0 addr64 offset:4
 ; GFX7-NEXT:    s_mov_b32 s6, -1
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v2, v0, v2
+; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v0, v0, v2
 ; GFX7-NEXT:    s_mov_b64 s[2:3], s[6:7]
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7-NEXT:    s_endpgm
@@ -112,7 +112,7 @@
 ; GFX8-NEXT:    flat_load_dword v0, v[0:1]
 ; GFX8-NEXT:    flat_load_dword v1, v[2:3]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, v1, v0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v1, v1, v0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -134,7 +134,7 @@
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    global_load_dword v1, v[2:3], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, v0, v1, v0
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, v1, v1, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -169,7 +169,7 @@
 ; GFX7-NEXT:    flat_load_dwordx2 v[0:1], v[0:1]
 ; GFX7-NEXT:    flat_load_dwordx2 v[2:3], v[2:3]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[2:3], v[2:3], v[0:1]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[2:3], v[0:1]
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX7-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
@@ -191,7 +191,7 @@
 ; GFX8-NEXT:    flat_load_dwordx2 v[0:1], v[0:1]
 ; GFX8-NEXT:    flat_load_dwordx2 v[2:3], v[2:3]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[2:3], v[2:3], v[0:1]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[2:3], v[0:1]
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX8-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX8-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
@@ -214,7 +214,7 @@
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    global_load_dwordx2 v[2:3], v[2:3], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, v[2:3], v[2:3], v[0:1]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, v[0:1], v[2:3], v[0:1]
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX10-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
@@ -249,7 +249,7 @@
 ; GFX7-NEXT:    flat_load_dwordx2 v[0:1], v[0:1]
 ; GFX7-NEXT:    flat_load_dwordx2 v[2:3], v[2:3]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[2:3], v[0:1]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[2:3], v[2:3], v[0:1]
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX7-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
@@ -271,7 +271,7 @@
 ; GFX8-NEXT:    flat_load_dwordx2 v[0:1], v[0:1]
 ; GFX8-NEXT:    flat_load_dwordx2 v[2:3], v[2:3]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[2:3], v[0:1]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[2:3], v[2:3], v[0:1]
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX8-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX8-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
@@ -294,7 +294,7 @@
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    global_load_dwordx2 v[2:3], v[2:3], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, v[0:1], v[2:3], v[0:1]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, v[2:3], v[2:3], v[0:1]
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX10-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
@@ -327,7 +327,7 @@
 ; GFX7-NEXT:    s_mov_b32 s2, -1
 ; GFX7-NEXT:    s_mov_b64 s[6:7], s[2:3]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], v0, v0, s8
+; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], s8, v0, s8
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[4:7], 0
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -344,7 +344,7 @@
 ; GFX8-NEXT:    v_addc_u32_e32 v1, vcc, v3, v1, vcc
 ; GFX8-NEXT:    flat_load_dword v0, v[0:1]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[0:1], v0, v0, s0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[0:1], s0, v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s5
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -364,7 +364,7 @@
 ; GFX10-NEXT:    v_add_co_ci_u32_e32 v1, vcc_lo, v3, v1, vcc_lo
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s0, v0, v0, s0
+; GFX10-NEXT:    v_div_scale_f32 v2, s0, s0, v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s5
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -395,7 +395,7 @@
 ; GFX7-NEXT:    s_mov_b32 s2, -1
 ; GFX7-NEXT:    s_mov_b64 s[6:7], s[2:3]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], s8, v0, s8
+; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], v0, v0, s8
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[4:7], 0
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -412,7 +412,7 @@
 ; GFX8-NEXT:    v_addc_u32_e32 v1, vcc, v3, v1, vcc
 ; GFX8-NEXT:    flat_load_dword v0, v[0:1]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[0:1], s0, v0, s0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[0:1], v0, v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s5
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -432,7 +432,7 @@
 ; GFX10-NEXT:    v_add_co_ci_u32_e32 v1, vcc_lo, v3, v1, vcc_lo
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s0, s0, v0, s0
+; GFX10-NEXT:    v_div_scale_f32 v2, s0, v0, v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s5
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -463,7 +463,7 @@
 ; GFX7-NEXT:    s_mov_b32 s2, -1
 ; GFX7-NEXT:    s_mov_b64 s[6:7], s[2:3]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], s8, s8, v0
+; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], v0, s8, v0
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[4:7], 0
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -480,7 +480,7 @@
 ; GFX8-NEXT:    v_addc_u32_e32 v1, vcc, v3, v1, vcc
 ; GFX8-NEXT:    flat_load_dword v0, v[0:1]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[0:1], s0, s0, v0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[0:1], v0, s0, v0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s5
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -500,7 +500,7 @@
 ; GFX10-NEXT:    v_add_co_ci_u32_e32 v1, vcc_lo, v3, v1, vcc_lo
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s0, s0, s0, v0
+; GFX10-NEXT:    v_div_scale_f32 v2, s0, v0, s0, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s5
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -531,7 +531,7 @@
 ; GFX7-NEXT:    s_mov_b32 s2, -1
 ; GFX7-NEXT:    s_mov_b64 s[6:7], s[2:3]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], v0, s8, v0
+; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], s8, s8, v0
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[4:7], 0
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -548,7 +548,7 @@
 ; GFX8-NEXT:    v_addc_u32_e32 v1, vcc, v3, v1, vcc
 ; GFX8-NEXT:    flat_load_dword v0, v[0:1]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[0:1], v0, s0, v0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[0:1], s0, s0, v0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s5
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -568,7 +568,7 @@
 ; GFX10-NEXT:    v_add_co_ci_u32_e32 v1, vcc_lo, v3, v1, vcc_lo
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s0, v0, s0, v0
+; GFX10-NEXT:    v_div_scale_f32 v2, s0, s0, s0, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s5
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -600,7 +600,7 @@
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s4
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[0:1], v[0:1], v[0:1], s[0:1]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[0:1], s[0:1], v[0:1], s[0:1]
 ; GFX7-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -619,7 +619,7 @@
 ; GFX8-NEXT:    v_mov_b32_e32 v2, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[0:1], v[0:1], v[0:1], s[0:1]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[0:1], s[0:1], v[0:1], s[0:1]
 ; GFX8-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
 ; GFX8-NEXT:    s_endpgm
 ;
@@ -639,7 +639,7 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s0, v[0:1], v[0:1], s[0:1]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s0, s[0:1], v[0:1], s[0:1]
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
 ; GFX10-NEXT:    s_endpgm
   %tid = call i32 @llvm.amdgcn.workitem.id.x()
@@ -669,7 +669,7 @@
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s4
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[0:1], s[0:1], v[0:1], s[0:1]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[0:1], v[0:1], v[0:1], s[0:1]
 ; GFX7-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -688,7 +688,7 @@
 ; GFX8-NEXT:    v_mov_b32_e32 v2, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[0:1], s[0:1], v[0:1], s[0:1]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[0:1], v[0:1], v[0:1], s[0:1]
 ; GFX8-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
 ; GFX8-NEXT:    s_endpgm
 ;
@@ -708,7 +708,7 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s0, s[0:1], v[0:1], s[0:1]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s0, v[0:1], v[0:1], s[0:1]
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
 ; GFX10-NEXT:    s_endpgm
   %tid = call i32 @llvm.amdgcn.workitem.id.x()
@@ -738,7 +738,7 @@
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s4
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[0:1], s[0:1], s[0:1], v[0:1]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[0:1], v[0:1], s[0:1], v[0:1]
 ; GFX7-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -757,7 +757,7 @@
 ; GFX8-NEXT:    v_mov_b32_e32 v2, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[0:1], s[0:1], s[0:1], v[0:1]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[0:1], v[0:1], s[0:1], v[0:1]
 ; GFX8-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
 ; GFX8-NEXT:    s_endpgm
 ;
@@ -777,7 +777,7 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s0, s[0:1], s[0:1], v[0:1]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s0, v[0:1], s[0:1], v[0:1]
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
 ; GFX10-NEXT:    s_endpgm
   %tid = call i32 @llvm.amdgcn.workitem.id.x()
@@ -807,7 +807,7 @@
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s4
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[0:1], v[0:1], s[0:1], v[0:1]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[0:1], s[0:1], s[0:1], v[0:1]
 ; GFX7-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -826,7 +826,7 @@
 ; GFX8-NEXT:    v_mov_b32_e32 v2, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[0:1], v[0:1], s[0:1], v[0:1]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[0:1], s[0:1], s[0:1], v[0:1]
 ; GFX8-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
 ; GFX8-NEXT:    s_endpgm
 ;
@@ -846,7 +846,7 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s5
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s0, v[0:1], s[0:1], v[0:1]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s0, s[0:1], s[0:1], v[0:1]
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
 ; GFX10-NEXT:    s_endpgm
   %tid = call i32 @llvm.amdgcn.workitem.id.x()
@@ -870,7 +870,7 @@
 ; GFX7-NEXT:    s_mov_b32 s7, 0xf000
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX7-NEXT:    v_mov_b32_e32 v0, s0
-; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], v0, v0, s2
+; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], s2, v0, s2
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[4:7], 0
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -881,7 +881,7 @@
 ; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s3
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, v0, s2
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], s2, v0, s2
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -894,7 +894,7 @@
 ; GFX10-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX10-NEXT:    ; implicit-def: $vcc_hi
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, s3, s3, s2
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, s2, s3, s2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -915,7 +915,7 @@
 ; GFX7-NEXT:    s_mov_b32 s7, 0xf000
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX7-NEXT:    v_mov_b32_e32 v0, s0
-; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], s2, v0, s2
+; GFX7-NEXT:    v_div_scale_f32 v0, s[0:1], v0, v0, s2
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[4:7], 0
 ; GFX7-NEXT:    s_endpgm
 ;
@@ -926,7 +926,7 @@
 ; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s3
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], s2, v0, s2
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, v0, s2
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -939,7 +939,7 @@
 ; GFX10-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX10-NEXT:    ; implicit-def: $vcc_hi
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, s2, s3, s2
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, s3, s3, s2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -959,7 +959,7 @@
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX7-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX7-NEXT:    v_mov_b32_e32 v1, s5
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[0:1], s[2:3]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], s[2:3], v[0:1], s[2:3]
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX7-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
@@ -973,7 +973,7 @@
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s5
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[0:1], s[2:3]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], s[2:3], v[0:1], s[2:3]
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX8-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX8-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
@@ -986,7 +986,7 @@
 ; GFX10-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX10-NEXT:    ; implicit-def: $vcc_hi
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, s[4:5], s[4:5], s[2:3]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, s[2:3], s[4:5], s[2:3]
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX10-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
@@ -1006,7 +1006,7 @@
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX7-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX7-NEXT:    v_mov_b32_e32 v1, s5
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], s[2:3], v[0:1], s[2:3]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[0:1], s[2:3]
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX7-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
@@ -1020,7 +1020,7 @@
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s4
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s5
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], s[2:3], v[0:1], s[2:3]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[0:1], s[2:3]
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX8-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX8-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
@@ -1033,7 +1033,7 @@
 ; GFX10-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX10-NEXT:    ; implicit-def: $vcc_hi
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, s[2:3], s[4:5], s[2:3]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, s[4:5], s[4:5], s[2:3]
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX10-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
@@ -1057,7 +1057,7 @@
 ; GFX7-NEXT:    buffer_load_dword v0, v[0:1], s[4:7], 0 addr64
 ; GFX7-NEXT:    s_mov_b32 s6, -1
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v0, v0, 1.0
+; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], 1.0, v0, 1.0
 ; GFX7-NEXT:    s_mov_b64 s[2:3], s[6:7]
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7-NEXT:    s_endpgm
@@ -1074,7 +1074,7 @@
 ; GFX8-NEXT:    v_addc_u32_e32 v1, vcc, v3, v1, vcc
 ; GFX8-NEXT:    flat_load_dword v0, v[0:1]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, v0, 1.0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], 1.0, v0, 1.0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -1093,7 +1093,7 @@
 ; GFX10-NEXT:    v_add_co_ci_u32_e32 v1, vcc_lo, v3, v1, vcc_lo
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, v0, v0, 1.0
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, 1.0, v0, 1.0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -1121,7 +1121,7 @@
 ; GFX7-NEXT:    buffer_load_dword v0, v[0:1], s[4:7], 0 addr64
 ; GFX7-NEXT:    s_mov_b32 s6, -1
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], 2.0, 2.0, v0
+; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v0, 2.0, v0
 ; GFX7-NEXT:    s_mov_b64 s[2:3], s[6:7]
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7-NEXT:    s_endpgm
@@ -1138,7 +1138,7 @@
 ; GFX8-NEXT:    v_addc_u32_e32 v1, vcc, v3, v1, vcc
 ; GFX8-NEXT:    flat_load_dword v0, v[0:1]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], 2.0, 2.0, v0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, 2.0, v0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -1157,7 +1157,7 @@
 ; GFX10-NEXT:    v_add_co_ci_u32_e32 v1, vcc_lo, v3, v1, vcc_lo
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, 2.0, 2.0, v0
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, v0, 2.0, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -1188,7 +1188,7 @@
 ; GFX7-NEXT:    s_waitcnt vmcnt(1)
 ; GFX7-NEXT:    v_and_b32_e32 v1, 0x7fffffff, v2
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v0, v0, v1
+; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v1, v0, v1
 ; GFX7-NEXT:    s_mov_b64 s[2:3], s[6:7]
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7-NEXT:    s_endpgm
@@ -1210,7 +1210,7 @@
 ; GFX8-NEXT:    s_waitcnt vmcnt(1) lgkmcnt(1)
 ; GFX8-NEXT:    v_and_b32_e32 v0, 0x7fffffff, v0
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v1, v1, v0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, v1, v0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -1234,7 +1234,7 @@
 ; GFX10-NEXT:    s_waitcnt vmcnt(1)
 ; GFX10-NEXT:    v_and_b32_e32 v0, 0x7fffffff, v0
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, v1, v1, v0
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, v0, v1, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -1269,7 +1269,7 @@
 ; GFX7-NEXT:    s_mov_b32 s6, -1
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    v_and_b32_e32 v0, 0x7fffffff, v0
-; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v0, v0, v2
+; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v2, v0, v2
 ; GFX7-NEXT:    s_mov_b64 s[2:3], s[6:7]
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7-NEXT:    s_endpgm
@@ -1290,7 +1290,7 @@
 ; GFX8-NEXT:    flat_load_dword v1, v[2:3]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
 ; GFX8-NEXT:    v_and_b32_e32 v1, 0x7fffffff, v1
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v1, v1, v0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, v1, v0
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX8-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX8-NEXT:    flat_store_dword v[0:1], v2
@@ -1313,7 +1313,7 @@
 ; GFX10-NEXT:    global_load_dword v1, v[2:3], off
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_and_b32_e32 v1, 0x7fffffff, v1
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, v1, v1, v0
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, v0, v1, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
@@ -1339,7 +1339,7 @@
 ; GFX7-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x9
 ; GFX7-NEXT:    v_mov_b32_e32 v0, 0x41000000
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], s0, s0, v0
+; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v0, s0, v0
 ; GFX7-NEXT:    s_mov_b32 s2, -1
 ; GFX7-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[0:3], 0
@@ -1348,7 +1348,7 @@
 ; GFX8-LABEL: test_div_scale_f32_val_undef_val:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    v_mov_b32_e32 v0, 0x41000000
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], s0, s0, v0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, s0, v0
 ; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
@@ -1359,10 +1359,11 @@
 ; GFX10-LABEL: test_div_scale_f32_val_undef_val:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
+; GFX10-NEXT:    v_mov_b32_e32 v0, 0x41000000
 ; GFX10-NEXT:    ; implicit-def: $vcc_hi
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, v0, s0, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, s0, s0, 0x41000000
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
 ; GFX10-NEXT:    s_endpgm
@@ -1378,7 +1379,7 @@
 ; GFX7-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x9
 ; GFX7-NEXT:    v_mov_b32_e32 v0, 0x41000000
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], v0, v0, s0
+; GFX7-NEXT:    v_div_scale_f32 v0, s[2:3], s0, v0, s0
 ; GFX7-NEXT:    s_mov_b32 s2, -1
 ; GFX7-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7-NEXT:    buffer_store_dword v0, off, s[0:3], 0
@@ -1387,7 +1388,7 @@
 ; GFX8-LABEL: test_div_scale_f32_undef_val_val:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    v_mov_b32_e32 v0, 0x41000000
-; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], v0, v0, s0
+; GFX8-NEXT:    v_div_scale_f32 v2, s[2:3], s0, v0, s0
 ; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    v_mov_b32_e32 v0, s0
@@ -1398,11 +1399,10 @@
 ; GFX10-LABEL: test_div_scale_f32_undef_val_val:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
-; GFX10-NEXT:    v_mov_b32_e32 v0, 0x41000000
 ; GFX10-NEXT:    ; implicit-def: $vcc_hi
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    v_div_scale_f32 v2, s2, v0, v0, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s0
+; GFX10-NEXT:    v_div_scale_f32 v2, s2, s0, 0x41000000, s0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s1
 ; GFX10-NEXT:    global_store_dword v[0:1], v2, off
 ; GFX10-NEXT:    s_endpgm
@@ -1455,7 +1455,7 @@
 ; GFX7-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x9
 ; GFX7-NEXT:    s_mov_b32 s2, 0
 ; GFX7-NEXT:    s_mov_b32 s3, 0x40200000
-; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[0:1], s[2:3]
+; GFX7-NEXT:    v_div_scale_f64 v[0:1], s[2:3], s[2:3], v[0:1], s[2:3]
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX7-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX7-NEXT:    v_mov_b32_e32 v2, s0
@@ -1466,7 +1466,7 @@
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    s_mov_b32 s2, 0
 ; GFX8-NEXT:    s_mov_b32 s3, 0x40200000
-; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], v[0:1], v[0:1], s[2:3]
+; GFX8-NEXT:    v_div_scale_f64 v[0:1], s[2:3], s[2:3], v[0:1], s[2:3]
 ; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    v_mov_b32_e32 v3, s1
@@ -1481,7 +1481,7 @@
 ; GFX10-NEXT:    s_mov_b32 s3, 0x40200000
 ; GFX10-NEXT:    ; implicit-def: $vcc_hi
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, s[0:1], s[0:1], s[2:3]
+; GFX10-NEXT:    v_div_scale_f64 v[0:1], s2, s[2:3], s[0:1], s[2:3]
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s1
 ; GFX10-NEXT:    v_mov_b32_e32 v2, s0
 ; GFX10-NEXT:    global_store_dwordx2 v[2:3], v[0:1], off
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.div.scale.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.div.scale.ll
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.div.scale.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.div.scale.ll
@@ -19,7 +19,7 @@
   %a = load volatile float, float addrspace(1)* %gep.0, align 4
   %b = load volatile float, float addrspace(1)* %gep.1, align 4
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 false) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 true) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -39,7 +39,7 @@
   %a = load volatile float, float addrspace(1)* %gep.0, align 4
   %b = load volatile float, float addrspace(1)* %gep.1, align 4
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 true) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 false) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -59,7 +59,7 @@
   %a = load volatile double, double addrspace(1)* %gep.0, align 8
   %b = load volatile double, double addrspace(1)* %gep.1, align 8
 
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 false) nounwind readnone
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 true) nounwind readnone
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void
@@ -79,7 +79,7 @@
   %a = load volatile double, double addrspace(1)* %gep.0, align 8
   %b = load volatile double, double addrspace(1)* %gep.1, align 8
 
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 true) nounwind readnone
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 false) nounwind readnone
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void
@@ -97,7 +97,7 @@
 
   %b = load float, float addrspace(1)* %gep, align 4
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 false) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 true) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -115,7 +115,7 @@
 
   %b = load float, float addrspace(1)* %gep, align 4
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 true) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 false) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -133,7 +133,7 @@
 
   %a = load float, float addrspace(1)* %gep, align 4
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 false) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 true) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -151,7 +151,7 @@
 
   %a = load float, float addrspace(1)* %gep, align 4
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 true) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 false) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -169,7 +169,7 @@
 
   %b = load double, double addrspace(1)* %gep, align 8
 
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 false) nounwind readnone
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 true) nounwind readnone
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void
@@ -187,7 +187,7 @@
 
   %b = load double, double addrspace(1)* %gep, align 8
 
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 true) nounwind readnone
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 false) nounwind readnone
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void
@@ -205,7 +205,7 @@
 
   %a = load double, double addrspace(1)* %gep, align 8
 
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 false) nounwind readnone
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 true) nounwind readnone
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void
@@ -223,7 +223,7 @@
 
   %a = load double, double addrspace(1)* %gep, align 8
 
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 true) nounwind readnone
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 false) nounwind readnone
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void
@@ -237,7 +237,7 @@
 ; SI: buffer_store_dword [[RESULT0]]
 ; SI: s_endpgm
 define amdgpu_kernel void @test_div_scale_f32_all_scalar_1(float addrspace(1)* %out, [8 x i32], float %a, [8 x i32], float %b) nounwind {
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 false) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 true) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -251,7 +251,7 @@
 ; SI: buffer_store_dword [[RESULT0]]
 ; SI: s_endpgm
 define amdgpu_kernel void @test_div_scale_f32_all_scalar_2(float addrspace(1)* %out, [8 x i32], float %a, [8 x i32], float %b) nounwind {
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 true) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b, i1 false) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -266,7 +266,7 @@
 ; SI: buffer_store_dwordx2 [[RESULT0]]
 ; SI: s_endpgm
 define amdgpu_kernel void @test_div_scale_f64_all_scalar_1(double addrspace(1)* %out, [8 x i32], double %a, [8 x i32], double %b) nounwind {
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 false) nounwind readnone
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 true) nounwind readnone
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void
@@ -281,7 +281,7 @@
 ; SI: buffer_store_dwordx2 [[RESULT0]]
 ; SI: s_endpgm
 define amdgpu_kernel void @test_div_scale_f64_all_scalar_2(double addrspace(1)* %out, [8 x i32], double %a, [8 x i32], double %b) nounwind {
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 true) nounwind readnone
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double %a, double %b, i1 false) nounwind readnone
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void
@@ -297,7 +297,7 @@
   %gep.0 = getelementptr float, float addrspace(1)* %in, i32 %tid
   %a = load float, float addrspace(1)* %gep.0, align 4
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float 1.0, float %a, i1 false) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float 1.0, float %a, i1 true) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -313,7 +313,7 @@
   %gep.0 = getelementptr float, float addrspace(1)* %in, i32 %tid
   %a = load float, float addrspace(1)* %gep.0, align 4
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float 2.0, i1 false) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float 2.0, i1 true) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -336,7 +336,7 @@
 
   %a.fabs = call float @llvm.fabs.f32(float %a) nounwind readnone
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a.fabs, float %b, i1 false) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a.fabs, float %b, i1 true) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -359,7 +359,7 @@
 
   %b.fabs = call float @llvm.fabs.f32(float %b) nounwind readnone
 
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b.fabs, i1 false) nounwind readnone
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float %a, float %b.fabs, i1 true) nounwind readnone
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -369,7 +369,7 @@
 ; SI: s_mov_b32 [[K:s[0-9]+]], 0x41000000
 ; SI: v_div_scale_f32 v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, [[K]], v{{[0-9]+}}, [[K]]
 define amdgpu_kernel void @test_div_scale_f32_val_undef_val(float addrspace(1)* %out) #0 {
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float 8.0, float undef, i1 false)
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float 8.0, float undef, i1 true)
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -379,7 +379,7 @@
 ; SI: s_mov_b32 [[K:s[0-9]+]], 0x41000000
 ; SI: v_div_scale_f32 v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, [[K]], [[K]], v{{[0-9]+}}
 define amdgpu_kernel void @test_div_scale_f32_undef_val_val(float addrspace(1)* %out) #0 {
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float undef, float 8.0, i1 false)
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float undef, float 8.0, i1 true)
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -389,7 +389,7 @@
 ; SI-NOT: v0
 ; SI: v_div_scale_f32 v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, s0, s0, v0
 define amdgpu_kernel void @test_div_scale_f32_undef_undef_val(float addrspace(1)* %out) #0 {
-  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float undef, float undef, i1 false)
+  %result = call { float, i1 } @llvm.amdgcn.div.scale.f32(float undef, float undef, i1 true)
   %result0 = extractvalue { float, i1 } %result, 0
   store float %result0, float addrspace(1)* %out, align 4
   ret void
@@ -400,7 +400,7 @@
 ; SI-DAG: s_mov_b32 s[[K_HI:[0-9]+]], 0x40200000
 ; SI: v_div_scale_f64 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, s{{\[}}[[K_LO]]:[[K_HI]]{{\]}}, v[0:1], s{{\[}}[[K_LO]]:[[K_HI]]{{\]}}
 define amdgpu_kernel void @test_div_scale_f64_val_undef_val(double addrspace(1)* %out) #0 {
-  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double 8.0, double undef, i1 false)
+  %result = call { double, i1 } @llvm.amdgcn.div.scale.f64(double 8.0, double undef, i1 true)
   %result0 = extractvalue { double, i1 } %result, 0
   store double %result0, double addrspace(1)* %out, align 8
   ret void