diff --git a/llvm/lib/Target/AMDGPU/BUFInstructions.td b/llvm/lib/Target/AMDGPU/BUFInstructions.td
--- a/llvm/lib/Target/AMDGPU/BUFInstructions.td
+++ b/llvm/lib/Target/AMDGPU/BUFInstructions.td
@@ -1737,7 +1737,7 @@
 defm : MUBUFScratchStorePat <BUFFER_STORE_DWORDX4_OFFEN, BUFFER_STORE_DWORDX4_OFFSET, v4i32, store_private, VReg_128>;
 
 
-let OtherPredicates = [D16PreservesUnusedBits, DisableFlatScratch] in {
+let OtherPredicates = [HasD16LoadStore, DisableFlatScratch] in {
  // Hiding the extract high pattern in the PatFrag seems to not
  // automatically increase the complexity.
 let AddedComplexity = 1 in {
diff --git a/llvm/lib/Target/AMDGPU/DSInstructions.td b/llvm/lib/Target/AMDGPU/DSInstructions.td
--- a/llvm/lib/Target/AMDGPU/DSInstructions.td
+++ b/llvm/lib/Target/AMDGPU/DSInstructions.td
@@ -778,7 +778,7 @@
 defm : DSAtomicWritePat_mc <DS_WRITE_B32, i32, "atomic_store_local_32">;
 defm : DSAtomicWritePat_mc <DS_WRITE_B64, i64, "atomic_store_local_64">;
 
-let OtherPredicates = [D16PreservesUnusedBits] in {
+let OtherPredicates = [HasD16LoadStore] in {
 def : DSWritePat <DS_WRITE_B16_D16_HI, i32, store_hi16_local>;
 def : DSWritePat <DS_WRITE_B8_D16_HI, i32, truncstorei8_hi16_local>;
 }
diff --git a/llvm/lib/Target/AMDGPU/FLATInstructions.td b/llvm/lib/Target/AMDGPU/FLATInstructions.td
--- a/llvm/lib/Target/AMDGPU/FLATInstructions.td
+++ b/llvm/lib/Target/AMDGPU/FLATInstructions.td
@@ -1027,10 +1027,12 @@
 def : FlatStorePat <FLAT_STORE_BYTE, truncstorei8_flat, i16>;
 def : FlatStorePat <FLAT_STORE_SHORT, store_flat, i16>;
 
-let OtherPredicates = [D16PreservesUnusedBits] in {
+let OtherPredicates = [HasD16LoadStore] in {
 def : FlatStorePat <FLAT_STORE_SHORT_D16_HI, truncstorei16_hi16_flat, i32>;
 def : FlatStorePat <FLAT_STORE_BYTE_D16_HI, truncstorei8_hi16_flat, i32>;
+}
 
+let OtherPredicates = [D16PreservesUnusedBits] in {
 def : FlatLoadPat_D16 <FLAT_LOAD_UBYTE_D16_HI, az_extloadi8_d16_hi_flat, v2i16>;
 def : FlatLoadPat_D16 <FLAT_LOAD_UBYTE_D16_HI, az_extloadi8_d16_hi_flat, v2f16>;
 def : FlatLoadPat_D16 <FLAT_LOAD_SBYTE_D16_HI, sextloadi8_d16_hi_flat, v2i16>;
@@ -1186,10 +1188,12 @@
 defm : GlobalFLATStorePats <GLOBAL_STORE_SHORT, store_global, i16>;
 defm : GlobalFLATStorePats <GLOBAL_STORE_DWORDX3, store_global, v3i32>;
 
-let OtherPredicates = [D16PreservesUnusedBits] in {
+let OtherPredicates = [HasD16LoadStore] in {
 defm : GlobalFLATStorePats <GLOBAL_STORE_SHORT_D16_HI, truncstorei16_hi16_global, i32>;
 defm : GlobalFLATStorePats <GLOBAL_STORE_BYTE_D16_HI, truncstorei8_hi16_global, i32>;
+}
 
+let OtherPredicates = [D16PreservesUnusedBits] in {
 defm : GlobalFLATLoadPats_D16 <GLOBAL_LOAD_UBYTE_D16_HI, az_extloadi8_d16_hi_global, v2i16>;
 defm : GlobalFLATLoadPats_D16 <GLOBAL_LOAD_UBYTE_D16_HI, az_extloadi8_d16_hi_global, v2f16>;
 defm : GlobalFLATLoadPats_D16 <GLOBAL_LOAD_SBYTE_D16_HI, sextloadi8_d16_hi_global, v2i16>;
@@ -1291,10 +1295,12 @@
 defm : ScratchFLATStorePats <SCRATCH_STORE_SHORT, store_private, i16>;
 defm : ScratchFLATStorePats <SCRATCH_STORE_DWORDX3, store_private, v3i32>;
 
-let OtherPredicates = [D16PreservesUnusedBits, HasFlatScratchInsts, EnableFlatScratch] in {
+let OtherPredicates = [HasD16LoadStore, HasFlatScratchInsts, EnableFlatScratch] in {
 defm : ScratchFLATStorePats <SCRATCH_STORE_SHORT_D16_HI, truncstorei16_hi16_private, i32>;
 defm : ScratchFLATStorePats <SCRATCH_STORE_BYTE_D16_HI, truncstorei8_hi16_private, i32>;
+}
 
+let OtherPredicates = [D16PreservesUnusedBits, HasFlatScratchInsts, EnableFlatScratch] in {
 defm : ScratchFLATLoadPats_D16 <SCRATCH_LOAD_UBYTE_D16_HI, az_extloadi8_d16_hi_private, v2i16>;
 defm : ScratchFLATLoadPats_D16 <SCRATCH_LOAD_UBYTE_D16_HI, az_extloadi8_d16_hi_private, v2f16>;
 defm : ScratchFLATLoadPats_D16 <SCRATCH_LOAD_SBYTE_D16_HI, sextloadi8_d16_hi_private, v2i16>;
diff --git a/llvm/test/CodeGen/AMDGPU/store-hi16.ll b/llvm/test/CodeGen/AMDGPU/store-hi16.ll
--- a/llvm/test/CodeGen/AMDGPU/store-hi16.ll
+++ b/llvm/test/CodeGen/AMDGPU/store-hi16.ll
@@ -1,5 +1,5 @@
 ; RUN: llc -march=amdgcn -mcpu=gfx900 -amdgpu-sroa=0 -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,GFX900,GFX9,GFX900-MUBUF %s
-; RUN: llc -march=amdgcn -mcpu=gfx906 -amdgpu-sroa=0 -mattr=-promote-alloca,+sram-ecc -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,GFX906,GFX9,NO-D16-HI %s
+; RxN: llc -march=amdgcn -mcpu=gfx906 -amdgpu-sroa=0 -mattr=-promote-alloca,+sram-ecc -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,GFX900,GFX9 %s
 ; RUN: llc -march=amdgcn -mcpu=fiji -amdgpu-sroa=0 -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,GFX803,NO-D16-HI %s
 ; RUN: llc -march=amdgcn -mcpu=gfx900 -amdgpu-sroa=0 -mattr=-promote-alloca -amdgpu-enable-flat-scratch -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,GFX900,GFX9,GFX900-FLATSCR %s
 
@@ -10,7 +10,6 @@
 
 ; NO-D16-HI-NEXT: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803-NEXT: flat_store_short v[0:1], v2
-; GFX906-NEXT: global_store_short v[0:1], v2, off
 
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
@@ -30,7 +29,6 @@
 
 ; NO-D16-HI-NEXT: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803-NEXT: flat_store_short v[0:1], v2
-; GFX906-NEXT: global_store_short v[0:1], v2, off
 
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
@@ -50,7 +48,6 @@
 
 ; NO-D16-HI-NEXT: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803-NEXT: flat_store_short v[0:1], v2
-; GFX906-NEXT: global_store_short v[0:1], v2, off
 
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
@@ -69,7 +66,6 @@
 
 ; NO-D16-HI-NEXT: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803-NEXT: flat_store_byte v[0:1], v2
-; GFX906-NEXT: global_store_byte v[0:1], v2, off
 
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
@@ -89,7 +85,6 @@
 
 ; NO-D16-HI-NEXT: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803-NEXT: flat_store_byte v[0:1], v2
-; GFX906-NEXT: global_store_byte v[0:1], v2, off
 
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
@@ -110,9 +105,6 @@
 ; GFX803-DAG: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803: flat_store_short v[0:1], v2{{$}}
 
-; GFX906-NEXT: v_lshrrev_b32_e32 v2, 16, v2
-; GFX906-NEXT: global_store_short v[0:1], v2, off
-
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
 define void @store_global_hi_v2i16_max_offset(i16 addrspace(1)* %out, i32 %arg) #0 {
@@ -134,9 +126,6 @@
 ; GFX803-DAG: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803: flat_store_short v[0:1], v{{[0-9]$}}
 
-; GFX906-NEXT: v_lshrrev_b32_e32 v2, 16, v2
-; GFX906-NEXT: global_store_short v[0:1], v2, off
-
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
 define void @store_global_hi_v2i16_min_offset(i16 addrspace(1)* %out, i32 %arg) #0 {
@@ -157,9 +146,6 @@
 ; GFX803-DAG: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803: flat_store_byte v[0:1], v{{[0-9]$}}
 
-; GFX906-NEXT: v_lshrrev_b32_e32 v2, 16, v2
-; GFX906-NEXT: global_store_byte v[0:1], v2, off
-
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
 define void @store_global_hi_v2i16_i8_max_offset(i8 addrspace(1)* %out, i32 %arg) #0 {
@@ -181,9 +167,6 @@
 ; GFX803-DAG: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803: flat_store_byte v[0:1], v{{[0-9]$}}
 
-; GFX906-NEXT: v_lshrrev_b32_e32 v2, 16, v2
-; GFX906-NEXT: global_store_byte v[0:1], v2, off
-
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
 define void @store_global_hi_v2i16_i8_min_offset(i8 addrspace(1)* %out, i32 %arg) #0 {
@@ -291,9 +274,6 @@
 ; GCN: s_waitcnt
 ; GFX900-NEXT: flat_store_short_d16_hi v[0:1], v2 offset:4094{{$}}
 
-; GFX906-NEXT: v_lshrrev_b32_e32 v2, 16, v2
-; GFX906-NEXT: flat_store_short v[0:1], v2 offset:4094
-
 ; GFX803-DAG: v_add_u32_e32
 ; GFX803-DAG: v_addc_u32_e32
 ; GFX803-DAG: v_lshrrev_b32_e32 v2, 16, v2
@@ -318,9 +298,6 @@
 ; GFX9-DAG: v_add_co_u32_e32 v{{[0-9]+}}, vcc, 0xfffff802, v
 ; GFX9-DAG: v_addc_co_u32_e32 v{{[0-9]+}}, vcc, -1, v
 
-; GFX906-DAG: v_lshrrev_b32_e32
-; GFX906: flat_store_short v[0:1], v2{{$}}
-
 ; GFX900-NEXT: flat_store_short_d16_hi v[0:1], v2{{$}}
 ; GFX803: flat_store_short v[0:1], v2{{$}}
 ; GCN-NEXT: s_waitcnt
@@ -343,9 +320,6 @@
 ; GFX803-DAG: v_addc_u32_e32
 ; GFX803: flat_store_byte v[0:1], v2{{$}}
 
-; GFX906-NEXT: v_lshrrev_b32_e32 v2, 16, v2
-; GFX906-NEXT: flat_store_byte v[0:1], v2 offset:4095{{$}}
-
 ; GCN-NEXT: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
 define void @store_flat_hi_v2i16_i8_max_offset(i8* %out, i32 %arg) #0 {
@@ -369,9 +343,6 @@
 
 ; GFX900-NEXT: flat_store_byte_d16_hi v[0:1], v2{{$}}
 
-; GFX906-DAG: v_lshrrev_b32_e32 v2, 16, v2
-; GFX906: flat_store_byte v[0:1], v2{{$}}
-
 ; GFX803-DAG: v_lshrrev_b32_e32 v2, 16, v2
 ; GFX803: flat_store_byte v[0:1], v2{{$}}