AMDGPU/SI: Assembler: Unify parsing/printing of operands.

Summary:
The goal is for each operand type to have its own parse function and
at the same time share common code for tracking state as different
instruction types share operand types (e.g. glc/glc_flat, etc).

Introduce parseAMDGPUOperand which can parse any optional operand.
DPP and Clamp/OMod have custom handling for now. Sam also suggested
to have class hierarchy for operand types instead of table. This
can be done in separate change.

Remove parseVOP3OptionalOps, parseDS*OptionalOps, parseFlatOptionalOps,
parseMubufOptionalOps, parseDPPOptionalOps.
Reduce number of definitions of AsmOperand's and MatchClasses' by using common base class.
Rename AsmMatcher/InstPrinter methods accordingly.
Print immediate type when printing parsed immediate operand.
Use 'off' if offset/index register is unused instead of skipping it to make it more readable (also agreed with SP3).
Update tests.

Reviewers: tstellarAMD, SamWot, artem.tamazov

Subscribers: qcolombet, arsenm, llvm-commits

Differential Revision: http://reviews.llvm.org/D19584

llvm-svn: 268015

31 files changed, 352 insertions, 353 deletions

diff --git a/llvm/test/CodeGen/AMDGPU/captured-frame-index.ll b/llvm/test/CodeGen/AMDGPU/captured-frame-index.ll
index 6cf19677812..3a0e8f31dc3 100644
--- a/llvm/test/CodeGen/AMDGPU/captured-frame-index.ll
+++ b/llvm/test/CodeGen/AMDGPU/captured-frame-index.ll
@@ -123,10 +123,10 @@ define void @stored_fi_to_global(float* addrspace(1)* %ptr) #0 {
 ; GCN: buffer_store_dword v{{[0-9]+}}, v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offen
 
 ; GCN: v_mov_b32_e32 [[FI1:v[0-9]+]], 4{{$}}
-; GCN: buffer_store_dword [[FI1]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN: buffer_store_dword [[FI1]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 
 ; GCN-DAG: v_mov_b32_e32 [[FI2:v[0-9]+]], 8{{$}}
-; GCN: buffer_store_dword [[FI2]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN: buffer_store_dword [[FI2]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 define void @stored_fi_to_global_2_small_objects(float* addrspace(1)* %ptr) #0 {
   %tmp0 = alloca float
   %tmp1 = alloca float
@@ -150,7 +150,7 @@ define void @stored_fi_to_global_2_small_objects(float* addrspace(1)* %ptr) #0 {
 
 ; GCN-DAG: s_add_i32 [[BASE_1_OFF_1:s[0-9]+]], 0, 56
 ; GCN-DAG: v_mov_b32_e32 [[V_BASE_1_OFF_1:v[0-9]+]], [[BASE_1_OFF_1]]
-; GCN: buffer_store_dword [[V_BASE_1_OFF_1]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN: buffer_store_dword [[V_BASE_1_OFF_1]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 define void @stored_fi_to_global_huge_frame_offset(i32* addrspace(1)* %ptr) #0 {
   %tmp0 = alloca [4096 x i32]
   %tmp1 = alloca [4096 x i32]
diff --git a/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes-flat.ll b/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes-flat.ll
index 1c5bed3b905..82f88a07930 100644
--- a/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes-flat.ll
+++ b/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes-flat.ll
@@ -42,7 +42,7 @@ done:
 ; OPT: br label
 
 ; GCN-LABEL: {{^}}test_sink_noop_addrspacecast_flat_to_global_i32:
-; CI: buffer_load_dword {{v[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:28
+; CI: buffer_load_dword {{v[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:28
 define void @test_sink_noop_addrspacecast_flat_to_global_i32(i32 addrspace(4)* %out, i32 addrspace(4)* %in, i32 %cond) {
 entry:
   %out.gep = getelementptr i32, i32 addrspace(4)* %out, i64 999999
diff --git a/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes.ll b/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes.ll
index f2b83fbe632..ae419a6a353 100644
--- a/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes.ll
+++ b/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes.ll
@@ -40,7 +40,7 @@ done:
 
 ; GCN-LABEL: {{^}}test_sink_global_small_max_i32_ds_offset:
 ; GCN: s_and_saveexec_b64
-; GCN: buffer_load_sbyte {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, s{{[0-9]+$}}
+; GCN: buffer_load_sbyte {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, s{{[0-9]+$}}
 ; GCN: {{^}}BB1_2:
 ; GCN: s_or_b64 exec
 define void @test_sink_global_small_max_i32_ds_offset(i32 addrspace(1)* %out, i8 addrspace(1)* %in) {
@@ -67,7 +67,7 @@ done:
 
 ; GCN-LABEL: {{^}}test_sink_global_small_max_mubuf_offset:
 ; GCN: s_and_saveexec_b64
-; GCN: buffer_load_sbyte {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:4095{{$}}
+; GCN: buffer_load_sbyte {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:4095{{$}}
 ; GCN: {{^}}BB2_2:
 ; GCN: s_or_b64 exec
 define void @test_sink_global_small_max_mubuf_offset(i32 addrspace(1)* %out, i8 addrspace(1)* %in) {
@@ -94,7 +94,7 @@ done:
 
 ; GCN-LABEL: {{^}}test_sink_global_small_max_plus_1_mubuf_offset:
 ; GCN: s_and_saveexec_b64
-; GCN: buffer_load_sbyte {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, s{{[0-9]+$}}
+; GCN: buffer_load_sbyte {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, s{{[0-9]+$}}
 ; GCN: {{^}}BB3_2:
 ; GCN: s_or_b64 exec
 define void @test_sink_global_small_max_plus_1_mubuf_offset(i32 addrspace(1)* %out, i8 addrspace(1)* %in) {
diff --git a/llvm/test/CodeGen/AMDGPU/ctpop.ll b/llvm/test/CodeGen/AMDGPU/ctpop.ll
index 839b893035a..32b071d18ab 100644
--- a/llvm/test/CodeGen/AMDGPU/ctpop.ll
+++ b/llvm/test/CodeGen/AMDGPU/ctpop.ll
@@ -250,8 +250,8 @@ define void @v_ctpop_i32_add_var_inv(i32 addrspace(1)* noalias %out, i32 addrspa
 }
 
 ; FUNC-LABEL: {{^}}v_ctpop_i32_add_vvar_inv:
-; GCN-DAG: buffer_load_dword [[VAL:v[0-9]+]], s[{{[0-9]+:[0-9]+}}], {{0$}}
-; GCN-DAG: buffer_load_dword [[VAR:v[0-9]+]], s[{{[0-9]+:[0-9]+}}], 0 offset:16
+; GCN-DAG: buffer_load_dword [[VAL:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], {{0$}}
+; GCN-DAG: buffer_load_dword [[VAR:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], 0 offset:16
 ; SI: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
 ; VI: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
 ; GCN: buffer_store_dword [[RESULT]],
diff --git a/llvm/test/CodeGen/AMDGPU/fdiv.f64.ll b/llvm/test/CodeGen/AMDGPU/fdiv.f64.ll
index 9fe3755a32b..3343b681b9f 100644
--- a/llvm/test/CodeGen/AMDGPU/fdiv.f64.ll
+++ b/llvm/test/CodeGen/AMDGPU/fdiv.f64.ll
@@ -4,8 +4,8 @@
 
 
 ; COMMON-LABEL: {{^}}fdiv_f64:
-; COMMON-DAG: buffer_load_dwordx2 [[NUM:v\[[0-9]+:[0-9]+\]]], {{s\[[0-9]+:[0-9]+\]}}, 0
-; COMMON-DAG: buffer_load_dwordx2 [[DEN:v\[[0-9]+:[0-9]+\]]], {{s\[[0-9]+:[0-9]+\]}}, 0 offset:8
+; COMMON-DAG: buffer_load_dwordx2 [[NUM:v\[[0-9]+:[0-9]+\]]], off, {{s\[[0-9]+:[0-9]+\]}}, 0
+; COMMON-DAG: buffer_load_dwordx2 [[DEN:v\[[0-9]+:[0-9]+\]]], off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:8
 ; CI-DAG: v_div_scale_f64 [[SCALE0:v\[[0-9]+:[0-9]+\]]], {{s\[[0-9]+:[0-9]+\]}}, [[DEN]], [[DEN]], [[NUM]]
 ; CI-DAG: v_div_scale_f64 [[SCALE1:v\[[0-9]+:[0-9]+\]]], vcc, [[NUM]], [[DEN]], [[NUM]]
 
diff --git a/llvm/test/CodeGen/AMDGPU/fmax3.f64.ll b/llvm/test/CodeGen/AMDGPU/fmax3.f64.ll
index 6d2e273f11d..9bbfe1e95c5 100644
--- a/llvm/test/CodeGen/AMDGPU/fmax3.f64.ll
+++ b/llvm/test/CodeGen/AMDGPU/fmax3.f64.ll
@@ -4,9 +4,9 @@
 declare double @llvm.maxnum.f64(double, double) nounwind readnone
 
 ; SI-LABEL: {{^}}test_fmax3_f64:
-; SI-DAG: buffer_load_dwordx2 [[REGA:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+:[0-9]+}}], 0{{$}}
-; SI-DAG: buffer_load_dwordx2 [[REGB:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+:[0-9]+}}], 0 offset:8
-; SI-DAG: buffer_load_dwordx2 [[REGC:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+:[0-9]+}}], 0 offset:16
+; SI-DAG: buffer_load_dwordx2 [[REGA:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+:[0-9]+}}], 0{{$}}
+; SI-DAG: buffer_load_dwordx2 [[REGB:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+:[0-9]+}}], 0 offset:8
+; SI-DAG: buffer_load_dwordx2 [[REGC:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+:[0-9]+}}], 0 offset:16
 ; SI: v_max_f64 [[REGA]], [[REGA]], [[REGB]]
 ; SI: v_max_f64 [[RESULT:v\[[0-9]+:[0-9]+\]]], [[REGA]], [[REGC]]
 ; SI: buffer_store_dwordx2 [[RESULT]],
diff --git a/llvm/test/CodeGen/AMDGPU/global_atomics.ll b/llvm/test/CodeGen/AMDGPU/global_atomics.ll
index 752ded85351..30b197a3604 100644
--- a/llvm/test/CodeGen/AMDGPU/global_atomics.ll
+++ b/llvm/test/CodeGen/AMDGPU/global_atomics.ll
@@ -3,7 +3,7 @@
 
 
 ; FUNC-LABEL: {{^}}atomic_add_i32_offset:
-; GCN: buffer_atomic_add v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_add v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_add_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -12,7 +12,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_add_i32_ret_offset:
-; GCN: buffer_atomic_add [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_add [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_add_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -48,7 +48,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_add_i32:
-; GCN: buffer_atomic_add v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_add v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_add_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile add i32 addrspace(1)* %out, i32 %in seq_cst
@@ -56,7 +56,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_add_i32_ret:
-; GCN: buffer_atomic_add [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_add [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_add_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -88,7 +88,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_and_i32_offset:
-; GCN: buffer_atomic_and v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_and v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_and_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -97,7 +97,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_and_i32_ret_offset:
-; GCN: buffer_atomic_and [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_and [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_and_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -132,7 +132,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_and_i32:
-; GCN: buffer_atomic_and v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_and v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_and_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile and i32 addrspace(1)* %out, i32 %in seq_cst
@@ -140,7 +140,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_and_i32_ret:
-; GCN: buffer_atomic_and [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_and [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_and_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -172,7 +172,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_sub_i32_offset:
-; GCN: buffer_atomic_sub v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_sub v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_sub_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -181,7 +181,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_sub_i32_ret_offset:
-; GCN: buffer_atomic_sub [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_sub [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_sub_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -216,7 +216,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_sub_i32:
-; GCN: buffer_atomic_sub v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_sub v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_sub_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile sub i32 addrspace(1)* %out, i32 %in seq_cst
@@ -224,7 +224,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_sub_i32_ret:
-; GCN: buffer_atomic_sub [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_sub [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_sub_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -256,7 +256,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_max_i32_offset:
-; GCN: buffer_atomic_smax v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_smax v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_max_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -265,7 +265,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_max_i32_ret_offset:
-; GCN: buffer_atomic_smax [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_smax [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_max_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -300,7 +300,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_max_i32:
-; GCN: buffer_atomic_smax v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_smax v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_max_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile max i32 addrspace(1)* %out, i32 %in seq_cst
@@ -308,7 +308,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_max_i32_ret:
-; GCN: buffer_atomic_smax [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_smax [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_max_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -340,7 +340,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_umax_i32_offset:
-; GCN: buffer_atomic_umax v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_umax v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_umax_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -349,7 +349,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_umax_i32_ret_offset:
-; GCN: buffer_atomic_umax [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_umax [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_umax_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -384,7 +384,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_umax_i32:
-; GCN: buffer_atomic_umax v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_umax v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_umax_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile umax i32 addrspace(1)* %out, i32 %in seq_cst
@@ -392,7 +392,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_umax_i32_ret:
-; GCN: buffer_atomic_umax [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_umax [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_umax_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -424,7 +424,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_min_i32_offset:
-; GCN: buffer_atomic_smin v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_smin v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_min_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -433,7 +433,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_min_i32_ret_offset:
-; GCN: buffer_atomic_smin [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_smin [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_min_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -468,7 +468,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_min_i32:
-; GCN: buffer_atomic_smin v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_smin v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_min_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile min i32 addrspace(1)* %out, i32 %in seq_cst
@@ -476,7 +476,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_min_i32_ret:
-; GCN: buffer_atomic_smin [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_smin [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_min_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -508,7 +508,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_umin_i32_offset:
-; GCN: buffer_atomic_umin v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_umin v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_umin_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -517,7 +517,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_umin_i32_ret_offset:
-; GCN: buffer_atomic_umin [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_umin [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_umin_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -552,7 +552,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_umin_i32:
-; GCN: buffer_atomic_umin v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_umin v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_umin_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile umin i32 addrspace(1)* %out, i32 %in seq_cst
@@ -560,7 +560,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_umin_i32_ret:
-; SI: buffer_atomic_umin [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; SI: buffer_atomic_umin [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_umin_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -592,7 +592,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_or_i32_offset:
-; GCN: buffer_atomic_or v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_or v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_or_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -601,7 +601,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_or_i32_ret_offset:
-; GCN: buffer_atomic_or [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_or [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_or_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -636,7 +636,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_or_i32:
-; GCN: buffer_atomic_or v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_or v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_or_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile or i32 addrspace(1)* %out, i32 %in seq_cst
@@ -644,7 +644,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_or_i32_ret:
-; GCN: buffer_atomic_or [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_or [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_or_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -676,7 +676,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_xchg_i32_offset:
-; GCN: buffer_atomic_swap v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_swap v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_xchg_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -685,7 +685,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_xchg_i32_ret_offset:
-; GCN: buffer_atomic_swap [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_swap [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_xchg_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -719,7 +719,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_xchg_i32:
-; GCN: buffer_atomic_swap v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_swap v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_xchg_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile xchg i32 addrspace(1)* %out, i32 %in seq_cst
@@ -727,7 +727,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_xchg_i32_ret:
-; GCN: buffer_atomic_swap [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_swap [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_xchg_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -761,7 +761,7 @@ entry:
 ; CMP_SWAP
 
 ; FUNC-LABEL: {{^}}atomic_cmpxchg_i32_offset:
-; GCN: buffer_atomic_cmpswap v[{{[0-9]+}}:{{[0-9]+}}], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_cmpswap v[{{[0-9]+}}:{{[0-9]+}}], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_cmpxchg_i32_offset(i32 addrspace(1)* %out, i32 %in, i32 %old) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -770,7 +770,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_cmpxchg_i32_ret_offset:
-; GCN: buffer_atomic_cmpswap v{{\[}}[[RET:[0-9]+]]{{:[0-9]+}}], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_cmpswap v{{\[}}[[RET:[0-9]+]]{{:[0-9]+}}], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword v[[RET]]
 define void @atomic_cmpxchg_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in, i32 %old) {
 entry:
@@ -806,7 +806,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_cmpxchg_i32:
-; GCN: buffer_atomic_cmpswap v[{{[0-9]+:[0-9]+}}], s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_cmpswap v[{{[0-9]+:[0-9]+}}], off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_cmpxchg_i32(i32 addrspace(1)* %out, i32 %in, i32 %old) {
 entry:
   %0  = cmpxchg volatile i32 addrspace(1)* %out, i32 %old, i32 %in seq_cst seq_cst
@@ -814,7 +814,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_cmpxchg_i32_ret:
-; GCN: buffer_atomic_cmpswap v{{\[}}[[RET:[0-9]+]]:{{[0-9]+}}], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_cmpswap v{{\[}}[[RET:[0-9]+]]:{{[0-9]+}}], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword v[[RET]]
 define void @atomic_cmpxchg_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in, i32 %old) {
 entry:
@@ -848,7 +848,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_xor_i32_offset:
-; GCN: buffer_atomic_xor v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
+; GCN: buffer_atomic_xor v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16{{$}}
 define void @atomic_xor_i32_offset(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 4
@@ -857,7 +857,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_xor_i32_ret_offset:
-; GCN: buffer_atomic_xor [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_xor [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_xor_i32_ret_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -892,7 +892,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_xor_i32:
-; GCN: buffer_atomic_xor v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_xor v{{[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_xor_i32(i32 addrspace(1)* %out, i32 %in) {
 entry:
   %0  = atomicrmw volatile xor i32 addrspace(1)* %out, i32 %in seq_cst
@@ -900,7 +900,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_xor_i32_ret:
-; GCN: buffer_atomic_xor [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_xor [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_xor_i32_ret(i32 addrspace(1)* %out, i32 addrspace(1)* %out2, i32 %in) {
 entry:
@@ -933,7 +933,7 @@ entry:
 
 ; ATOMIC_LOAD
 ; FUNC-LABEL: {{^}}atomic_load_i32_offset:
-; SI: buffer_load_dword [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; SI: buffer_load_dword [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; VI: flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_load_i32_offset(i32 addrspace(1)* %in, i32 addrspace(1)* %out) {
@@ -945,7 +945,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_load_i32:
-; SI: buffer_load_dword [[RET:v[0-9]+]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; SI: buffer_load_dword [[RET:v[0-9]+]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; VI: flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc
 ; GCN: buffer_store_dword [[RET]]
 define void @atomic_load_i32(i32 addrspace(1)* %in, i32 addrspace(1)* %out) {
@@ -981,7 +981,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_load_i64_offset:
-; SI: buffer_load_dwordx2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; SI: buffer_load_dwordx2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; VI: flat_load_dwordx2 [[RET:v\[[0-9]+:[0-9]\]]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_load_i64_offset(i64 addrspace(1)* %in, i64 addrspace(1)* %out) {
@@ -993,7 +993,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_load_i64:
-; SI: buffer_load_dwordx2 [[RET:v\[[0-9]+:[0-9]\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; SI: buffer_load_dwordx2 [[RET:v\[[0-9]+:[0-9]\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; VI: flat_load_dwordx2 [[RET:v\[[0-9]+:[0-9]\]]], v[{{[0-9]+}}:{{[0-9]+}}] glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_load_i64(i64 addrspace(1)* %in, i64 addrspace(1)* %out) {
@@ -1030,7 +1030,7 @@ entry:
 
 ; ATOMIC_STORE
 ; FUNC-LABEL: {{^}}atomic_store_i32_offset:
-; SI: buffer_store_dword {{v[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
+; SI: buffer_store_dword {{v[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:16 glc{{$}}
 ; VI: flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}} glc{{$}}
 define void @atomic_store_i32_offset(i32 %in, i32 addrspace(1)* %out) {
 entry:
@@ -1040,7 +1040,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_store_i32:
-; SI: buffer_store_dword {{v[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc{{$}}
+; SI: buffer_store_dword {{v[0-9]+}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc{{$}}
 ; VI: flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}} glc{{$}}
 define void @atomic_store_i32(i32 %in, i32 addrspace(1)* %out) {
 entry:
@@ -1070,7 +1070,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_store_i64_offset:
-; SI: buffer_store_dwordx2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; SI: buffer_store_dwordx2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; VI: flat_store_dwordx2 [[RET:v\[[0-9]+:[0-9]\]]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
 define void @atomic_store_i64_offset(i64 %in, i64 addrspace(1)* %out) {
 entry:
@@ -1080,7 +1080,7 @@ entry:
 }
 
 ; FUNC-LABEL: {{^}}atomic_store_i64:
-; SI: buffer_store_dwordx2 {{v\[[0-9]+:[0-9]\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; SI: buffer_store_dwordx2 {{v\[[0-9]+:[0-9]\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; VI: flat_store_dwordx2 {{v\[[0-9]+:[0-9]\]}}, v[{{[0-9]+}}:{{[0-9]+}}] glc
 define void @atomic_store_i64(i64 %in, i64 addrspace(1)* %out) {
 entry:
diff --git a/llvm/test/CodeGen/AMDGPU/global_atomics_i64.ll b/llvm/test/CodeGen/AMDGPU/global_atomics_i64.ll
index 3eabec45145..bc37825e59f 100644
--- a/llvm/test/CodeGen/AMDGPU/global_atomics_i64.ll
+++ b/llvm/test/CodeGen/AMDGPU/global_atomics_i64.ll
@@ -3,7 +3,7 @@
 
 
 ; GCN-LABEL: {{^}}atomic_add_i64_offset:
-; GCN: buffer_atomic_add_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_add_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_add_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -12,7 +12,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_add_i64_ret_offset:
-; GCN: buffer_atomic_add_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_add_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_add_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -47,7 +47,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_add_i64:
-; GCN: buffer_atomic_add_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_add_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_add_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile add i64 addrspace(1)* %out, i64 %in seq_cst
@@ -55,7 +55,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_add_i64_ret:
-; GCN: buffer_atomic_add_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_add_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_add_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -87,7 +87,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_and_i64_offset:
-; GCN: buffer_atomic_and_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_and_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_and_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -96,7 +96,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_and_i64_ret_offset:
-; GCN: buffer_atomic_and_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_and_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_and_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -131,7 +131,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_and_i64:
-; GCN: buffer_atomic_and_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_and_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_and_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile and i64 addrspace(1)* %out, i64 %in seq_cst
@@ -139,7 +139,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_and_i64_ret:
-; GCN: buffer_atomic_and_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_and_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_and_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -171,7 +171,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_sub_i64_offset:
-; GCN: buffer_atomic_sub_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_sub_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_sub_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -180,7 +180,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_sub_i64_ret_offset:
-; GCN: buffer_atomic_sub_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_sub_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_sub_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -215,7 +215,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_sub_i64:
-; GCN: buffer_atomic_sub_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_sub_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_sub_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile sub i64 addrspace(1)* %out, i64 %in seq_cst
@@ -223,7 +223,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_sub_i64_ret:
-; GCN: buffer_atomic_sub_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_sub_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_sub_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -255,7 +255,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_max_i64_offset:
-; GCN: buffer_atomic_smax_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_smax_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_max_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -264,7 +264,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_max_i64_ret_offset:
-; GCN: buffer_atomic_smax_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_smax_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_max_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -299,7 +299,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_max_i64:
-; GCN: buffer_atomic_smax_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_smax_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_max_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile max i64 addrspace(1)* %out, i64 %in seq_cst
@@ -307,7 +307,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_max_i64_ret:
-; GCN: buffer_atomic_smax_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_smax_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_max_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -339,7 +339,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_umax_i64_offset:
-; GCN: buffer_atomic_umax_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_umax_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_umax_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -348,7 +348,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_umax_i64_ret_offset:
-; GCN: buffer_atomic_umax_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_umax_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_umax_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -383,7 +383,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_umax_i64:
-; GCN: buffer_atomic_umax_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_umax_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_umax_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile umax i64 addrspace(1)* %out, i64 %in seq_cst
@@ -391,7 +391,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_umax_i64_ret:
-; GCN: buffer_atomic_umax_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_umax_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_umax_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -423,7 +423,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_min_i64_offset:
-; GCN: buffer_atomic_smin_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_smin_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_min_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -432,7 +432,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_min_i64_ret_offset:
-; GCN: buffer_atomic_smin_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_smin_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_min_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -467,7 +467,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_min_i64:
-; GCN: buffer_atomic_smin_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_smin_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_min_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile min i64 addrspace(1)* %out, i64 %in seq_cst
@@ -475,7 +475,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_min_i64_ret:
-; GCN: buffer_atomic_smin_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_smin_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_min_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -507,7 +507,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_umin_i64_offset:
-; GCN: buffer_atomic_umin_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_umin_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_umin_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -516,7 +516,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_umin_i64_ret_offset:
-; GCN: buffer_atomic_umin_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_umin_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_umin_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -551,7 +551,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_umin_i64:
-; GCN: buffer_atomic_umin_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_umin_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_umin_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile umin i64 addrspace(1)* %out, i64 %in seq_cst
@@ -559,7 +559,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_umin_i64_ret:
-; CI: buffer_atomic_umin_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; CI: buffer_atomic_umin_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_umin_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -591,7 +591,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_or_i64_offset:
-; GCN: buffer_atomic_or_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_or_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_or_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -600,7 +600,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_or_i64_ret_offset:
-; GCN: buffer_atomic_or_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_or_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_or_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -635,7 +635,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_or_i64:
-; GCN: buffer_atomic_or_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_or_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_or_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile or i64 addrspace(1)* %out, i64 %in seq_cst
@@ -643,7 +643,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_or_i64_ret:
-; GCN: buffer_atomic_or_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_or_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_or_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -675,7 +675,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_xchg_i64_offset:
-; GCN: buffer_atomic_swap_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_swap_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_xchg_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -684,7 +684,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_xchg_i64_ret_offset:
-; GCN: buffer_atomic_swap_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_swap_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_xchg_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -718,7 +718,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_xchg_i64:
-; GCN: buffer_atomic_swap_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_swap_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_xchg_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile xchg i64 addrspace(1)* %out, i64 %in seq_cst
@@ -726,7 +726,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_xchg_i64_ret:
-; GCN: buffer_atomic_swap_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_swap_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_xchg_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -758,7 +758,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_xor_i64_offset:
-; GCN: buffer_atomic_xor_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
+; GCN: buffer_atomic_xor_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32{{$}}
 define void @atomic_xor_i64_offset(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %gep = getelementptr i64, i64 addrspace(1)* %out, i64 4
@@ -767,7 +767,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_xor_i64_ret_offset:
-; GCN: buffer_atomic_xor_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_xor_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 offset:32 glc{{$}}
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_xor_i64_ret_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
@@ -802,7 +802,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_xor_i64:
-; GCN: buffer_atomic_xor_x2 v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
+; GCN: buffer_atomic_xor_x2 v{{\[[0-9]+:[0-9]+\]}}, off, s[{{[0-9]+}}:{{[0-9]+}}], 0{{$}}
 define void @atomic_xor_i64(i64 addrspace(1)* %out, i64 %in) {
 entry:
   %tmp0  = atomicrmw volatile xor i64 addrspace(1)* %out, i64 %in seq_cst
@@ -810,7 +810,7 @@ entry:
 }
 
 ; GCN-LABEL: {{^}}atomic_xor_i64_ret:
-; GCN: buffer_atomic_xor_x2 [[RET:v\[[0-9]+:[0-9]+\]]], s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
+; GCN: buffer_atomic_xor_x2 [[RET:v\[[0-9]+:[0-9]+\]]], off, s[{{[0-9]+}}:{{[0-9]+}}], 0 glc
 ; GCN: buffer_store_dwordx2 [[RET]]
 define void @atomic_xor_i64_ret(i64 addrspace(1)* %out, i64 addrspace(1)* %out2, i64 %in) {
 entry:
diff --git a/llvm/test/CodeGen/AMDGPU/half.ll b/llvm/test/CodeGen/AMDGPU/half.ll
index 409b2ec8a29..af409afc496 100644
--- a/llvm/test/CodeGen/AMDGPU/half.ll
+++ b/llvm/test/CodeGen/AMDGPU/half.ll
@@ -13,10 +13,10 @@ define void @load_f16_arg(half addrspace(1)* %out, half %arg) #0 {
 }
 
 ; GCN-LABEL: {{^}}load_v2f16_arg:
-; GCN-DAG: buffer_load_ushort [[V0:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:44
-; GCN-DAG: buffer_load_ushort [[V1:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:46
-; GCN-DAG: buffer_store_short [[V0]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_store_short [[V1]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:2{{$}}
+; GCN-DAG: buffer_load_ushort [[V0:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:44
+; GCN-DAG: buffer_load_ushort [[V1:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:46
+; GCN-DAG: buffer_store_short [[V0]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_store_short [[V1]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:2{{$}}
 ; GCN: s_endpgm
 define void @load_v2f16_arg(<2 x half> addrspace(1)* %out, <2 x half> %arg) #0 {
   store <2 x half> %arg, <2 x half> addrspace(1)* %out
@@ -280,8 +280,8 @@ define void @global_extload_f16_to_f32(float addrspace(1)* %out, half addrspace(
 }
 
 ; GCN-LABEL: {{^}}global_extload_v2f16_to_v2f32:
-; GCN-DAG: buffer_load_ushort [[LOAD0:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_load_ushort [[LOAD1:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:2{{$}}
+; GCN-DAG: buffer_load_ushort [[LOAD0:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_load_ushort [[LOAD1:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:2{{$}}
 ; GCN-DAG: v_cvt_f32_f16_e32 v[[CVT0:[0-9]+]], [[LOAD0]]
 ; GCN-DAG: v_cvt_f32_f16_e32 v[[CVT1:[0-9]+]], [[LOAD1]]
 ; GCN-DAG: buffer_store_dwordx2 v{{\[}}[[CVT0]]:[[CVT1]]{{\]}}
@@ -378,8 +378,8 @@ define void @global_extload_f16_to_f64(double addrspace(1)* %out, half addrspace
 }
 
 ; GCN-LABEL: {{^}}global_extload_v2f16_to_v2f64:
-; GCN-DAG: buffer_load_ushort [[LOAD0:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_load_ushort [[LOAD1:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:2{{$}}
+; GCN-DAG: buffer_load_ushort [[LOAD0:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_load_ushort [[LOAD1:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:2{{$}}
 ; GCN-DAG: v_cvt_f32_f16_e32 v[[CVT0:[0-9]+]], [[LOAD0]]
 ; GCN-DAG: v_cvt_f32_f16_e32 v[[CVT1:[0-9]+]], [[LOAD1]]
 ; GCN-DAG: v_cvt_f64_f32_e32 v{{\[}}[[CVT2_LO:[0-9]+]]:[[CVT2_HI:[0-9]+]]{{\]}}, v[[CVT0]]
@@ -407,8 +407,8 @@ define void @global_extload_v2f16_to_v2f64(<2 x double> addrspace(1)* %out, <2 x
 ; GCN: v_cvt_f64_f32_e32
 ; GCN-NOT: v_cvt_f64_f32_e32
 
-; GCN-DAG: buffer_store_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16
+; GCN-DAG: buffer_store_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16
 ; GCN: s_endpgm
 define void @global_extload_v3f16_to_v3f64(<3 x double> addrspace(1)* %out, <3 x half> addrspace(1)* %in) #0 {
   %val = load <3 x half>, <3 x half> addrspace(1)* %in
diff --git a/llvm/test/CodeGen/AMDGPU/invariant-load-no-alias-store.ll b/llvm/test/CodeGen/AMDGPU/invariant-load-no-alias-store.ll
index 2a01a621fc4..a70da5be5f5 100644
--- a/llvm/test/CodeGen/AMDGPU/invariant-load-no-alias-store.ll
+++ b/llvm/test/CodeGen/AMDGPU/invariant-load-no-alias-store.ll
@@ -21,7 +21,7 @@ define void @test_merge_store_constant_i16_invariant_global_pointer_load(i16 add
 ; GCN-LABEL: {{^}}test_merge_store_constant_i16_invariant_constant_pointer_load:
 ; GCN: s_load_dwordx2 s{{\[}}[[SPTR_LO:[0-9]+]]:[[SPTR_HI:[0-9]+]]{{\]}}
 ; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 0x1c8007b
-; GCN: buffer_store_dword [[K]], s{{\[}}[[SPTR_LO]]:
+; GCN: buffer_store_dword [[K]], off, s{{\[}}[[SPTR_LO]]:
 define void @test_merge_store_constant_i16_invariant_constant_pointer_load(i16 addrspace(1)* addrspace(2)* dereferenceable(4096) nonnull %in) #0 {
   %ptr = load i16 addrspace(1)*, i16 addrspace(1)* addrspace(2)* %in, !invariant.load !0
   %ptr.1 = getelementptr i16, i16 addrspace(1)* %ptr, i64 1
@@ -32,4 +32,4 @@ define void @test_merge_store_constant_i16_invariant_constant_pointer_load(i16 a
 
 !0 = !{}
 
-attributes #0 = { nounwind }
\ No newline at end of file
+attributes #0 = { nounwind }
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.SI.load.dword.ll b/llvm/test/CodeGen/AMDGPU/llvm.SI.load.dword.ll
index 0d9ec45a894..d0cc00d81b4 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.SI.load.dword.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.SI.load.dword.ll
@@ -7,7 +7,7 @@
 ; FIXME: Out of bounds immediate offset crashes
 
 ; CHECK-LABEL: {{^}}main:
-; CHECK: buffer_load_dword {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 glc slc
+; CHECK: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 glc slc
 ; CHECK: buffer_load_dword {{v[0-9]+}}, {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 offen glc slc
 ; CHECK: buffer_load_dword {{v[0-9]+}}, {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 idxen glc slc
 ; CHECK: buffer_load_dword {{v[0-9]+}}, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, 0 idxen offen glc slc
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.dec.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.dec.ll
index 48f3fbad042..d17179e96bc 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.dec.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.dec.ll
@@ -49,7 +49,7 @@ define void @lds_atomic_dec_noret_i32_offset(i32 addrspace(3)* %ptr) nounwind {
 
 ; GCN-LABEL: {{^}}global_atomic_dec_ret_i32:
 ; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
-; GCN: buffer_atomic_dec [[K]], s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
+; GCN: buffer_atomic_dec [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
 define void @global_atomic_dec_ret_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %ptr) #0 {
   %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 42)
   store i32 %result, i32 addrspace(1)* %out
@@ -58,7 +58,7 @@ define void @global_atomic_dec_ret_i32(i32 addrspace(1)* %out, i32 addrspace(1)*
 
 ; GCN-LABEL: {{^}}global_atomic_dec_ret_i32_offset:
 ; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
-; GCN: buffer_atomic_dec [[K]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_dec [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16 glc{{$}}
 define void @global_atomic_dec_ret_i32_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %ptr) #0 {
   %gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
   %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
@@ -67,7 +67,7 @@ define void @global_atomic_dec_ret_i32_offset(i32 addrspace(1)* %out, i32 addrsp
 }
 
 ; FUNC-LABEL: {{^}}global_atomic_dec_noret_i32:
-; GCN: buffer_atomic_dec [[K]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN: buffer_atomic_dec [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 define void @global_atomic_dec_noret_i32(i32 addrspace(1)* %ptr) nounwind {
   %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 42)
   ret void
@@ -75,7 +75,7 @@ define void @global_atomic_dec_noret_i32(i32 addrspace(1)* %ptr) nounwind {
 
 ; FUNC-LABEL: {{^}}global_atomic_dec_noret_i32_offset:
 ; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
-; GCN: buffer_atomic_dec [[K]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
+; GCN: buffer_atomic_dec [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
 define void @global_atomic_dec_noret_i32_offset(i32 addrspace(1)* %ptr) nounwind {
   %gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
   %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
@@ -166,7 +166,7 @@ define void @lds_atomic_dec_noret_i64_offset(i64 addrspace(3)* %ptr) nounwind {
 ; GCN-LABEL: {{^}}global_atomic_dec_ret_i64:
 ; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
 ; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
-; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
+; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
 define void @global_atomic_dec_ret_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
   %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 42)
   store i64 %result, i64 addrspace(1)* %out
@@ -176,7 +176,7 @@ define void @global_atomic_dec_ret_i64(i64 addrspace(1)* %out, i64 addrspace(1)*
 ; GCN-LABEL: {{^}}global_atomic_dec_ret_i64_offset:
 ; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
 ; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
-; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32 glc{{$}}
 define void @global_atomic_dec_ret_i64_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
   %gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
   %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
@@ -187,7 +187,7 @@ define void @global_atomic_dec_ret_i64_offset(i64 addrspace(1)* %out, i64 addrsp
 ; FUNC-LABEL: {{^}}global_atomic_dec_noret_i64:
 ; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
 ; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
-; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 define void @global_atomic_dec_noret_i64(i64 addrspace(1)* %ptr) nounwind {
   %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 42)
   ret void
@@ -196,7 +196,7 @@ define void @global_atomic_dec_noret_i64(i64 addrspace(1)* %ptr) nounwind {
 ; FUNC-LABEL: {{^}}global_atomic_dec_noret_i64_offset:
 ; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
 ; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
-; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32{{$}}
+; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32{{$}}
 define void @global_atomic_dec_noret_i64_offset(i64 addrspace(1)* %ptr) nounwind {
   %gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
   %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.inc.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.inc.ll
index d79c37cbfba..93aa1faba2a 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.inc.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.inc.ll
@@ -49,7 +49,7 @@ define void @lds_atomic_inc_noret_i32_offset(i32 addrspace(3)* %ptr) nounwind {
 
 ; GCN-LABEL: {{^}}global_atomic_inc_ret_i32:
 ; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
-; GCN: buffer_atomic_inc [[K]], s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
+; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
 define void @global_atomic_inc_ret_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %ptr) #0 {
   %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 42)
   store i32 %result, i32 addrspace(1)* %out
@@ -58,7 +58,7 @@ define void @global_atomic_inc_ret_i32(i32 addrspace(1)* %out, i32 addrspace(1)*
 
 ; GCN-LABEL: {{^}}global_atomic_inc_ret_i32_offset:
 ; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
-; GCN: buffer_atomic_inc [[K]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16 glc{{$}}
+; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16 glc{{$}}
 define void @global_atomic_inc_ret_i32_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %ptr) #0 {
   %gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
   %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
@@ -67,7 +67,7 @@ define void @global_atomic_inc_ret_i32_offset(i32 addrspace(1)* %out, i32 addrsp
 }
 
 ; FUNC-LABEL: {{^}}global_atomic_inc_noret_i32:
-; GCN: buffer_atomic_inc [[K]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 define void @global_atomic_inc_noret_i32(i32 addrspace(1)* %ptr) nounwind {
   %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 42)
   ret void
@@ -75,7 +75,7 @@ define void @global_atomic_inc_noret_i32(i32 addrspace(1)* %ptr) nounwind {
 
 ; FUNC-LABEL: {{^}}global_atomic_inc_noret_i32_offset:
 ; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
-; GCN: buffer_atomic_inc [[K]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
+; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
 define void @global_atomic_inc_noret_i32_offset(i32 addrspace(1)* %ptr) nounwind {
   %gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
   %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
@@ -166,7 +166,7 @@ define void @lds_atomic_inc_noret_i64_offset(i64 addrspace(3)* %ptr) nounwind {
 ; GCN-LABEL: {{^}}global_atomic_inc_ret_i64:
 ; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
 ; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
-; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
+; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
 define void @global_atomic_inc_ret_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
   %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 42)
   store i64 %result, i64 addrspace(1)* %out
@@ -176,7 +176,7 @@ define void @global_atomic_inc_ret_i64(i64 addrspace(1)* %out, i64 addrspace(1)*
 ; GCN-LABEL: {{^}}global_atomic_inc_ret_i64_offset:
 ; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
 ; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
-; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32 glc{{$}}
+; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32 glc{{$}}
 define void @global_atomic_inc_ret_i64_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
   %gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
   %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
@@ -187,7 +187,7 @@ define void @global_atomic_inc_ret_i64_offset(i64 addrspace(1)* %out, i64 addrsp
 ; FUNC-LABEL: {{^}}global_atomic_inc_noret_i64:
 ; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
 ; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
-; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 define void @global_atomic_inc_noret_i64(i64 addrspace(1)* %ptr) nounwind {
   %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 42)
   ret void
@@ -196,7 +196,7 @@ define void @global_atomic_inc_noret_i64(i64 addrspace(1)* %ptr) nounwind {
 ; FUNC-LABEL: {{^}}global_atomic_inc_noret_i64_offset:
 ; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
 ; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
-; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32{{$}}
+; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32{{$}}
 define void @global_atomic_inc_noret_i64_offset(i64 addrspace(1)* %ptr) nounwind {
   %gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
   %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.atomic.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.atomic.ll
index 1d848726533..f513095cd03 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.atomic.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.atomic.ll
@@ -2,7 +2,7 @@
 ;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck %s
 
 ;CHECK-LABEL: {{^}}test1:
-;CHECK: buffer_atomic_swap v0, s[0:3], 0 glc
+;CHECK: buffer_atomic_swap v0, off, s[0:3], 0 glc
 ;CHECK: s_waitcnt vmcnt(0)
 ;CHECK: buffer_atomic_swap v0, v1, s[0:3], 0 idxen glc
 ;CHECK: s_waitcnt vmcnt(0)
@@ -13,9 +13,9 @@
 ;CHECK: buffer_atomic_swap v0, v2, s[0:3], 0 offen offset:42 glc
 ;CHECK-DAG: s_waitcnt vmcnt(0)
 ;CHECK-DAG: s_movk_i32 [[SOFS:s[0-9]+]], 0x1fff
-;CHECK: buffer_atomic_swap v0, s[0:3], [[SOFS]] offset:1 glc
+;CHECK: buffer_atomic_swap v0, off, s[0:3], [[SOFS]] offset:1 glc
 ;CHECK: s_waitcnt vmcnt(0)
-;CHECK: buffer_atomic_swap v0, s[0:3], 0{{$}}
+;CHECK: buffer_atomic_swap v0, off, s[0:3], 0{{$}}
 define amdgpu_ps float @test1(<4 x i32> inreg %rsrc, i32 %data, i32 %vindex, i32 %voffset) {
 main_body:
   %o1 = call i32 @llvm.amdgcn.buffer.atomic.swap(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i1 0)
@@ -68,7 +68,7 @@ main_body:
 ; create copies which we don't bother to track here.
 ;
 ;CHECK-LABEL: {{^}}test3:
-;CHECK: buffer_atomic_cmpswap {{v\[[0-9]+:[0-9]+\]}}, s[0:3], 0 glc
+;CHECK: buffer_atomic_cmpswap {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], 0 glc
 ;CHECK: s_waitcnt vmcnt(0)
 ;CHECK: buffer_atomic_cmpswap {{v\[[0-9]+:[0-9]+\]}}, v2, s[0:3], 0 idxen glc
 ;CHECK: s_waitcnt vmcnt(0)
@@ -79,7 +79,7 @@ main_body:
 ;CHECK: buffer_atomic_cmpswap {{v\[[0-9]+:[0-9]+\]}}, v3, s[0:3], 0 offen offset:42 glc
 ;CHECK-DAG: s_waitcnt vmcnt(0)
 ;CHECK-DAG: s_movk_i32 [[SOFS:s[0-9]+]], 0x1fff
-;CHECK: buffer_atomic_cmpswap {{v\[[0-9]+:[0-9]+\]}}, s[0:3], [[SOFS]] offset:1 glc
+;CHECK: buffer_atomic_cmpswap {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], [[SOFS]] offset:1 glc
 define amdgpu_ps float @test3(<4 x i32> inreg %rsrc, i32 %data, i32 %cmp, i32 %vindex, i32 %voffset) {
 main_body:
   %o1 = call i32 @llvm.amdgcn.buffer.atomic.cmpswap(i32 %data, i32 %cmp, <4 x i32> %rsrc, i32 0, i32 0, i1 0)
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.load.format.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.load.format.ll
index d8ec770f9cd..9fe7161394a 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.load.format.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.load.format.ll
@@ -2,9 +2,9 @@
 ;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck %s
 
 ;CHECK-LABEL: {{^}}buffer_load:
-;CHECK: buffer_load_format_xyzw v[0:3], s[0:3], 0
-;CHECK: buffer_load_format_xyzw v[4:7], s[0:3], 0 glc
-;CHECK: buffer_load_format_xyzw v[8:11], s[0:3], 0 slc
+;CHECK: buffer_load_format_xyzw v[0:3], off, s[0:3], 0
+;CHECK: buffer_load_format_xyzw v[4:7], off, s[0:3], 0 glc
+;CHECK: buffer_load_format_xyzw v[8:11], off, s[0:3], 0 slc
 ;CHECK: s_waitcnt
 define amdgpu_ps {<4 x float>, <4 x float>, <4 x float>} @buffer_load(<4 x i32> inreg) {
 main_body:
@@ -18,7 +18,7 @@ main_body:
 }
 
 ;CHECK-LABEL: {{^}}buffer_load_immoffs:
-;CHECK: buffer_load_format_xyzw v[0:3], s[0:3], 0 offset:42
+;CHECK: buffer_load_format_xyzw v[0:3], off, s[0:3], 0 offset:42
 ;CHECK: s_waitcnt
 define amdgpu_ps <4 x float> @buffer_load_immoffs(<4 x i32> inreg) {
 main_body:
@@ -27,11 +27,11 @@ main_body:
 }
 
 ;CHECK-LABEL: {{^}}buffer_load_immoffs_large:
-;CHECK-DAG: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, s[0:3], 61 offset:4095
+;CHECK-DAG: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], 61 offset:4095
 ;CHECK-DAG: s_movk_i32 [[OFS1:s[0-9]+]], 0x7fff
-;CHECK: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, s[0:3], [[OFS1]] offset:4093
+;CHECK: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], [[OFS1]] offset:4093
 ;CHECK: s_mov_b32 [[OFS2:s[0-9]+]], 0x8fff
-;CHECK: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, s[0:3], [[OFS2]] offset:1
+;CHECK: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], [[OFS2]] offset:1
 ;CHECK: s_waitcnt
 define amdgpu_ps <4 x float> @buffer_load_immoffs_large(<4 x i32> inreg) {
 main_body:
@@ -45,9 +45,9 @@ main_body:
 
 ;CHECK-LABEL: {{^}}buffer_load_immoffs_reuse:
 ;CHECK: s_movk_i32 [[OFS:s[0-9]+]], 0xfff
-;CHECK: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, s[0:3], [[OFS]] offset:65
+;CHECK: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], [[OFS]] offset:65
 ;CHECK-NOT: s_mov
-;CHECK: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, s[0:3], [[OFS]] offset:81
+;CHECK: buffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], [[OFS]] offset:81
 ;CHECK: s_waitcnt
 define amdgpu_ps <4 x float> @buffer_load_immoffs_reuse(<4 x i32> inreg) {
 main_body:
@@ -105,7 +105,7 @@ main_body:
 }
 
 ;CHECK-LABEL: {{^}}buffer_load_x:
-;CHECK: buffer_load_format_x v0, s[0:3], 0
+;CHECK: buffer_load_format_x v0, off, s[0:3], 0
 ;CHECK: s_waitcnt
 define amdgpu_ps float @buffer_load_x(<4 x i32> inreg %rsrc) {
 main_body:
@@ -114,7 +114,7 @@ main_body:
 }
 
 ;CHECK-LABEL: {{^}}buffer_load_xy:
-;CHECK: buffer_load_format_xy v[0:1], s[0:3], 0
+;CHECK: buffer_load_format_xy v[0:1], off, s[0:3], 0
 ;CHECK: s_waitcnt
 define amdgpu_ps <2 x float> @buffer_load_xy(<4 x i32> inreg %rsrc) {
 main_body:
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.load.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.load.ll
index 3d5f69f0745..a547f007ce2 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.load.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.load.ll
@@ -2,9 +2,9 @@
 ;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck %s
 
 ;CHECK-LABEL: {{^}}buffer_load:
-;CHECK: buffer_load_dwordx4 v[0:3], s[0:3], 0
-;CHECK: buffer_load_dwordx4 v[4:7], s[0:3], 0 glc
-;CHECK: buffer_load_dwordx4 v[8:11], s[0:3], 0 slc
+;CHECK: buffer_load_dwordx4 v[0:3], off, s[0:3], 0
+;CHECK: buffer_load_dwordx4 v[4:7], off, s[0:3], 0 glc
+;CHECK: buffer_load_dwordx4 v[8:11], off, s[0:3], 0 slc
 ;CHECK: s_waitcnt
 define amdgpu_ps {<4 x float>, <4 x float>, <4 x float>} @buffer_load(<4 x i32> inreg) {
 main_body:
@@ -18,7 +18,7 @@ main_body:
 }
 
 ;CHECK-LABEL: {{^}}buffer_load_immoffs:
-;CHECK: buffer_load_dwordx4 v[0:3], s[0:3], 0 offset:42
+;CHECK: buffer_load_dwordx4 v[0:3], off, s[0:3], 0 offset:42
 ;CHECK: s_waitcnt
 define amdgpu_ps <4 x float> @buffer_load_immoffs(<4 x i32> inreg) {
 main_body:
@@ -28,7 +28,7 @@ main_body:
 
 ;CHECK-LABEL: {{^}}buffer_load_immoffs_large:
 ;CHECK: s_movk_i32 [[OFFSET:s[0-9]+]], 0x1fff
-;CHECK: buffer_load_dwordx4 v[0:3], s[0:3], [[OFFSET]] offset:1
+;CHECK: buffer_load_dwordx4 v[0:3], off, s[0:3], [[OFFSET]] offset:1
 ;CHECK: s_waitcnt
 define amdgpu_ps <4 x float> @buffer_load_immoffs_large(<4 x i32> inreg) {
 main_body:
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.store.format.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.store.format.ll
index 63dcd841383..555a1d23ebe 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.store.format.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.store.format.ll
@@ -2,9 +2,9 @@
 ;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck %s
 
 ;CHECK-LABEL: {{^}}buffer_store:
-;CHECK: buffer_store_format_xyzw v[0:3], s[0:3], 0
-;CHECK: buffer_store_format_xyzw v[4:7], s[0:3], 0 glc
-;CHECK: buffer_store_format_xyzw v[8:11], s[0:3], 0 slc
+;CHECK: buffer_store_format_xyzw v[0:3], off, s[0:3], 0
+;CHECK: buffer_store_format_xyzw v[4:7], off, s[0:3], 0 glc
+;CHECK: buffer_store_format_xyzw v[8:11], off, s[0:3], 0 slc
 define amdgpu_ps void @buffer_store(<4 x i32> inreg, <4 x float>, <4 x float>, <4 x float>) {
 main_body:
   call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 0, i32 0, i1 0, i1 0)
@@ -14,7 +14,7 @@ main_body:
 }
 
 ;CHECK-LABEL: {{^}}buffer_store_immoffs:
-;CHECK: buffer_store_format_xyzw v[0:3], s[0:3], 0 offset:42
+;CHECK: buffer_store_format_xyzw v[0:3], off, s[0:3], 0 offset:42
 define amdgpu_ps void @buffer_store_immoffs(<4 x i32> inreg, <4 x float>) {
 main_body:
   call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 0, i32 42, i1 0, i1 0)
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.store.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.store.ll
index c935f9b520a..5ae255c7a26 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.store.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.buffer.store.ll
@@ -2,9 +2,9 @@
 ;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck %s
 
 ;CHECK-LABEL: {{^}}buffer_store:
-;CHECK: buffer_store_dwordx4 v[0:3], s[0:3], 0
-;CHECK: buffer_store_dwordx4 v[4:7], s[0:3], 0 glc
-;CHECK: buffer_store_dwordx4 v[8:11], s[0:3], 0 slc
+;CHECK: buffer_store_dwordx4 v[0:3], off, s[0:3], 0
+;CHECK: buffer_store_dwordx4 v[4:7], off, s[0:3], 0 glc
+;CHECK: buffer_store_dwordx4 v[8:11], off, s[0:3], 0 slc
 define amdgpu_ps void @buffer_store(<4 x i32> inreg, <4 x float>, <4 x float>, <4 x float>) {
 main_body:
   call void @llvm.amdgcn.buffer.store.v4f32(<4 x float> %1, <4 x i32> %0, i32 0, i32 0, i1 0, i1 0)
@@ -14,7 +14,7 @@ main_body:
 }
 
 ;CHECK-LABEL: {{^}}buffer_store_immoffs:
-;CHECK: buffer_store_dwordx4 v[0:3], s[0:3], 0 offset:42
+;CHECK: buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:42
 define amdgpu_ps void @buffer_store_immoffs(<4 x i32> inreg, <4 x float>) {
 main_body:
   call void @llvm.amdgcn.buffer.store.v4f32(<4 x float> %1, <4 x i32> %0, i32 0, i32 42, i1 0, i1 0)
diff --git a/llvm/test/CodeGen/AMDGPU/merge-stores.ll b/llvm/test/CodeGen/AMDGPU/merge-stores.ll
index a8948374b0c..17b4af818f8 100644
--- a/llvm/test/CodeGen/AMDGPU/merge-stores.ll
+++ b/llvm/test/CodeGen/AMDGPU/merge-stores.ll
@@ -231,8 +231,8 @@ define void @merge_global_store_2_adjacent_loads_i32(i32 addrspace(1)* %out, i32
 }
 
 ; GCN-LABEL: {{^}}merge_global_store_2_adjacent_loads_i32_nonzero_base:
-; GCN: buffer_load_dwordx2 [[LOAD:v\[[0-9]+:[0-9]+\]]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
-; GCN: buffer_store_dwordx2 [[LOAD]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
+; GCN: buffer_load_dwordx2 [[LOAD:v\[[0-9]+:[0-9]+\]]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
+; GCN: buffer_store_dwordx2 [[LOAD]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
 define void @merge_global_store_2_adjacent_loads_i32_nonzero_base(i32 addrspace(1)* %out, i32 addrspace(1)* %in) #0 {
   %in.gep.0 = getelementptr i32, i32 addrspace(1)* %in, i32 2
   %in.gep.1 = getelementptr i32, i32 addrspace(1)* %in, i32 3
@@ -334,8 +334,8 @@ define void @merge_global_store_4_adjacent_loads_f32(float addrspace(1)* %out, f
 }
 
 ; GCN-LABEL: {{^}}merge_global_store_4_adjacent_loads_i32_nonzero_base:
-; GCN: buffer_load_dwordx4 [[LOAD:v\[[0-9]+:[0-9]+\]]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:44
-; GCN: buffer_store_dwordx4 [[LOAD]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:28
+; GCN: buffer_load_dwordx4 [[LOAD:v\[[0-9]+:[0-9]+\]]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:44
+; GCN: buffer_store_dwordx4 [[LOAD]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:28
 define void @merge_global_store_4_adjacent_loads_i32_nonzero_base(i32 addrspace(1)* %out, i32 addrspace(1)* %in) #0 {
   %in.gep.0 = getelementptr i32, i32 addrspace(1)* %in, i32 11
   %in.gep.1 = getelementptr i32, i32 addrspace(1)* %in, i32 12
@@ -640,13 +640,13 @@ define void @merge_global_store_8_constants_i32(i32 addrspace(1)* %out) {
 
 ; GCN-LABEL: {{^}}copy_v3i32_align4:
 ; GCN-NOT: SCRATCH_RSRC_DWORD
-; GCN-DAG: buffer_load_dword v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
-; GCN-DAG: buffer_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_load_dword v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
+; GCN-DAG: buffer_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 ; GCN-NOT: offen
 ; GCN: s_waitcnt vmcnt
 ; GCN-NOT: offen
-; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_store_dword v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
+; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_store_dword v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
 
 ; GCN: ScratchSize: 0{{$}}
 define void @copy_v3i32_align4(<3 x i32> addrspace(1)* noalias %out, <3 x i32> addrspace(1)* noalias %in) #0 {
@@ -657,13 +657,13 @@ define void @copy_v3i32_align4(<3 x i32> addrspace(1)* noalias %out, <3 x i32> a
 
 ; GCN-LABEL: {{^}}copy_v3i64_align4:
 ; GCN-NOT: SCRATCH_RSRC_DWORD
-; GCN-DAG: buffer_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
+; GCN-DAG: buffer_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
 ; GCN-NOT: offen
 ; GCN: s_waitcnt vmcnt
 ; GCN-NOT: offen
-; GCN-DAG: buffer_store_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
+; GCN-DAG: buffer_store_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
 ; GCN: ScratchSize: 0{{$}}
 define void @copy_v3i64_align4(<3 x i64> addrspace(1)* noalias %out, <3 x i64> addrspace(1)* noalias %in) #0 {
   %vec = load <3 x i64>, <3 x i64> addrspace(1)* %in, align 4
@@ -673,13 +673,13 @@ define void @copy_v3i64_align4(<3 x i64> addrspace(1)* noalias %out, <3 x i64> a
 
 ; GCN-LABEL: {{^}}copy_v3f32_align4:
 ; GCN-NOT: SCRATCH_RSRC_DWORD
-; GCN-DAG: buffer_load_dword v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
-; GCN-DAG: buffer_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_load_dword v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
+; GCN-DAG: buffer_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
 ; GCN-NOT: offen
 ; GCN: s_waitcnt vmcnt
 ; GCN-NOT: offen
-; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_store_dword v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
+; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_store_dword v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:8
 ; GCN: ScratchSize: 0{{$}}
 define void @copy_v3f32_align4(<3 x float> addrspace(1)* noalias %out, <3 x float> addrspace(1)* noalias %in) #0 {
   %vec = load <3 x float>, <3 x float> addrspace(1)* %in, align 4
@@ -690,13 +690,13 @@ define void @copy_v3f32_align4(<3 x float> addrspace(1)* noalias %out, <3 x floa
 
 ; GCN-LABEL: {{^}}copy_v3f64_align4:
 ; GCN-NOT: SCRATCH_RSRC_DWORD
-; GCN-DAG: buffer_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
+; GCN-DAG: buffer_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
 ; GCN-NOT: offen
 ; GCN: s_waitcnt vmcnt
 ; GCN-NOT: offen
-; GCN-DAG: buffer_store_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
+; GCN-DAG: buffer_store_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; GCN-DAG: buffer_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
 ; GCN: ScratchSize: 0{{$}}
 define void @copy_v3f64_align4(<3 x double> addrspace(1)* noalias %out, <3 x double> addrspace(1)* noalias %in) #0 {
   %vec = load <3 x double>, <3 x double> addrspace(1)* %in, align 4
diff --git a/llvm/test/CodeGen/AMDGPU/mubuf.ll b/llvm/test/CodeGen/AMDGPU/mubuf.ll
index 4208c53363d..277b5882eb2 100644
--- a/llvm/test/CodeGen/AMDGPU/mubuf.ll
+++ b/llvm/test/CodeGen/AMDGPU/mubuf.ll
@@ -8,7 +8,7 @@ declare i32 @llvm.amdgcn.workitem.id.x() readnone
 
 ; MUBUF load with an immediate byte offset that fits into 12-bits
 ; CHECK-LABEL: {{^}}mubuf_load0:
-; CHECK: buffer_load_dword v{{[0-9]}}, s[{{[0-9]+:[0-9]+}}], 0 offset:4 ; encoding: [0x04,0x00,0x30,0xe0
+; CHECK: buffer_load_dword v{{[0-9]}}, off, s[{{[0-9]+:[0-9]+}}], 0 offset:4 ; encoding: [0x04,0x00,0x30,0xe0
 define void @mubuf_load0(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
 entry:
   %0 = getelementptr i32, i32 addrspace(1)* %in, i64 1
@@ -19,7 +19,7 @@ entry:
 
 ; MUBUF load with the largest possible immediate offset
 ; CHECK-LABEL: {{^}}mubuf_load1:
-; CHECK: buffer_load_ubyte v{{[0-9]}}, s[{{[0-9]+:[0-9]+}}], 0 offset:4095 ; encoding: [0xff,0x0f,0x20,0xe0
+; CHECK: buffer_load_ubyte v{{[0-9]}}, off, s[{{[0-9]+:[0-9]+}}], 0 offset:4095 ; encoding: [0xff,0x0f,0x20,0xe0
 define void @mubuf_load1(i8 addrspace(1)* %out, i8 addrspace(1)* %in) {
 entry:
   %0 = getelementptr i8, i8 addrspace(1)* %in, i64 4095
@@ -31,7 +31,7 @@ entry:
 ; MUBUF load with an immediate byte offset that doesn't fit into 12-bits
 ; CHECK-LABEL: {{^}}mubuf_load2:
 ; CHECK: s_movk_i32 [[SOFFSET:s[0-9]+]], 0x1000
-; CHECK: buffer_load_dword v{{[0-9]}}, s[{{[0-9]+:[0-9]+}}], [[SOFFSET]] ; encoding: [0x00,0x00,0x30,0xe0
+; CHECK: buffer_load_dword v{{[0-9]}}, off, s[{{[0-9]+:[0-9]+}}], [[SOFFSET]] ; encoding: [0x00,0x00,0x30,0xe0
 define void @mubuf_load2(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
 entry:
   %0 = getelementptr i32, i32 addrspace(1)* %in, i64 1024
@@ -92,7 +92,7 @@ main_body:
 
 ; MUBUF store with an immediate byte offset that fits into 12-bits
 ; CHECK-LABEL: {{^}}mubuf_store0:
-; CHECK: buffer_store_dword v{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0 offset:4 ; encoding: [0x04,0x00,0x70,0xe0
+; CHECK: buffer_store_dword v{{[0-9]}}, off, s[{{[0-9]:[0-9]}}], 0 offset:4 ; encoding: [0x04,0x00,0x70,0xe0
 define void @mubuf_store0(i32 addrspace(1)* %out) {
 entry:
   %0 = getelementptr i32, i32 addrspace(1)* %out, i64 1
@@ -102,7 +102,7 @@ entry:
 
 ; MUBUF store with the largest possible immediate offset
 ; CHECK-LABEL: {{^}}mubuf_store1:
-; CHECK: buffer_store_byte v{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0 offset:4095 ; encoding: [0xff,0x0f,0x60,0xe0
+; CHECK: buffer_store_byte v{{[0-9]}}, off, s[{{[0-9]:[0-9]}}], 0 offset:4095 ; encoding: [0xff,0x0f,0x60,0xe0
 
 define void @mubuf_store1(i8 addrspace(1)* %out) {
 entry:
@@ -114,7 +114,7 @@ entry:
 ; MUBUF store with an immediate byte offset that doesn't fit into 12-bits
 ; CHECK-LABEL: {{^}}mubuf_store2:
 ; CHECK: s_movk_i32 [[SOFFSET:s[0-9]+]], 0x1000
-; CHECK: buffer_store_dword v{{[0-9]}}, s[{{[0-9]:[0-9]}}], [[SOFFSET]] ; encoding: [0x00,0x00,0x70,0xe0
+; CHECK: buffer_store_dword v{{[0-9]}}, off, s[{{[0-9]:[0-9]}}], [[SOFFSET]] ; encoding: [0x00,0x00,0x70,0xe0
 define void @mubuf_store2(i32 addrspace(1)* %out) {
 entry:
   %0 = getelementptr i32, i32 addrspace(1)* %out, i64 1024
@@ -135,14 +135,14 @@ entry:
 }
 
 ; CHECK-LABEL: {{^}}store_sgpr_ptr:
-; CHECK: buffer_store_dword v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, 0
+; CHECK: buffer_store_dword v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0
 define void @store_sgpr_ptr(i32 addrspace(1)* %out) #0 {
   store i32 99, i32 addrspace(1)* %out, align 4
   ret void
 }
 
 ; CHECK-LABEL: {{^}}store_sgpr_ptr_offset:
-; CHECK: buffer_store_dword v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:40
+; CHECK: buffer_store_dword v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:40
 define void @store_sgpr_ptr_offset(i32 addrspace(1)* %out) #0 {
   %out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 10
   store i32 99, i32 addrspace(1)* %out.gep, align 4
@@ -151,7 +151,7 @@ define void @store_sgpr_ptr_offset(i32 addrspace(1)* %out) #0 {
 
 ; CHECK-LABEL: {{^}}store_sgpr_ptr_large_offset:
 ; CHECK: s_mov_b32 [[SOFFSET:s[0-9]+]], 0x20000
-; CHECK: buffer_store_dword v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, [[SOFFSET]]
+; CHECK: buffer_store_dword v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, [[SOFFSET]]
 define void @store_sgpr_ptr_large_offset(i32 addrspace(1)* %out) #0 {
   %out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 32768
   store i32 99, i32 addrspace(1)* %out.gep, align 4
@@ -160,7 +160,7 @@ define void @store_sgpr_ptr_large_offset(i32 addrspace(1)* %out) #0 {
 
 ; CHECK-LABEL: {{^}}store_sgpr_ptr_large_offset_atomic:
 ; CHECK: s_mov_b32 [[SOFFSET:s[0-9]+]], 0x20000
-; CHECK: buffer_atomic_add v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, [[SOFFSET]]
+; CHECK: buffer_atomic_add v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, [[SOFFSET]]
 define void @store_sgpr_ptr_large_offset_atomic(i32 addrspace(1)* %out) #0 {
   %gep = getelementptr i32, i32 addrspace(1)* %out, i32 32768
   %val = atomicrmw volatile add i32 addrspace(1)* %gep, i32 5 seq_cst
diff --git a/llvm/test/CodeGen/AMDGPU/reduce-load-width-alignment.ll b/llvm/test/CodeGen/AMDGPU/reduce-load-width-alignment.ll
index c255b05324b..c902cb9e1df 100644
--- a/llvm/test/CodeGen/AMDGPU/reduce-load-width-alignment.ll
+++ b/llvm/test/CodeGen/AMDGPU/reduce-load-width-alignment.ll
@@ -25,7 +25,7 @@ define void @reduce_i64_align_4_bitcast_v2i32_elt0(i32 addrspace(1)* %out, i64 a
 }
 
 ; GCN-LABEL: {{^}}reduce_i64_align_4_bitcast_v2i32_elt1:
-; GCN: buffer_load_dword [[VAL:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:4
+; GCN: buffer_load_dword [[VAL:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:4
 ; GCN: buffer_store_dword [[VAL]]
 define void @reduce_i64_align_4_bitcast_v2i32_elt1(i32 addrspace(1)* %out, i64 addrspace(1)* %in) #0 {
   %a = load i64, i64 addrspace(1)* %in, align 4
diff --git a/llvm/test/CodeGen/AMDGPU/schedule-global-loads.ll b/llvm/test/CodeGen/AMDGPU/schedule-global-loads.ll
index 41ff202245c..faf03c3c3b8 100644
--- a/llvm/test/CodeGen/AMDGPU/schedule-global-loads.ll
+++ b/llvm/test/CodeGen/AMDGPU/schedule-global-loads.ll
@@ -7,8 +7,8 @@
 ; ordering the loads so that the lower address loads come first.
 
 ; FUNC-LABEL: {{^}}cluster_global_arg_loads:
-; SI-DAG: buffer_load_dword [[REG0:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; SI-DAG: buffer_load_dword [[REG1:v[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0 offset:4
+; SI-DAG: buffer_load_dword [[REG0:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
+; SI-DAG: buffer_load_dword [[REG1:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:4
 ; SI: buffer_store_dword [[REG0]]
 ; SI: buffer_store_dword [[REG1]]
 define void @cluster_global_arg_loads(i32 addrspace(1)* %out0, i32 addrspace(1)* %out1, i32 addrspace(1)* %ptr) #0 {
diff --git a/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll b/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
index 34b85cdf632..14fe99c4190 100644
--- a/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
+++ b/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
@@ -184,12 +184,12 @@ define void @reorder_local_offsets(i32 addrspace(1)* nocapture %out, i32 addrspa
 }
 
 ; FUNC-LABEL: @reorder_global_offsets
-; CI: buffer_load_dword {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:400
-; CI: buffer_load_dword {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:404
-; CI: buffer_store_dword {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:12
-; CI: buffer_load_dword {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:12
-; CI: buffer_store_dword {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:400
-; CI: buffer_store_dword {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:404
+; CI: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:400
+; CI: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:404
+; CI: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:12
+; CI: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:12
+; CI: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:400
+; CI: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:404
 ; CI: buffer_store_dword
 ; CI: s_endpgm
 define void @reorder_global_offsets(i32 addrspace(1)* nocapture %out, i32 addrspace(1)* noalias nocapture readnone %gptr, i32 addrspace(1)* noalias nocapture %ptr0) #0 {
diff --git a/llvm/test/CodeGen/AMDGPU/subreg-coalescer-undef-use.ll b/llvm/test/CodeGen/AMDGPU/subreg-coalescer-undef-use.ll
index 1e7d37aa384..289d6f4eee8 100644
--- a/llvm/test/CodeGen/AMDGPU/subreg-coalescer-undef-use.ll
+++ b/llvm/test/CodeGen/AMDGPU/subreg-coalescer-undef-use.ll
@@ -18,7 +18,7 @@ target triple="amdgcn--"
 ; CHECK: s_or_b64 exec, exec, s[2:3]
 ; CHECK-NEXT: s_mov_b32 s7, 0xf000
 ; CHECK-NEXT: s_mov_b32 s6, -1
-; CHECK-NEXT: buffer_store_dword v1, s[4:7], 0
+; CHECK-NEXT: buffer_store_dword v1, off, s[4:7], 0
 ; CHECK-NEXT: s_endpgm
 define void @foobar(float %a0, float %a1, float addrspace(1)* %out) nounwind {
 entry:
diff --git a/llvm/test/CodeGen/AMDGPU/v_mac.ll b/llvm/test/CodeGen/AMDGPU/v_mac.ll
index ff05027ff0d..4a209ca8392 100644
--- a/llvm/test/CodeGen/AMDGPU/v_mac.ll
+++ b/llvm/test/CodeGen/AMDGPU/v_mac.ll
@@ -2,9 +2,9 @@
 ; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=VI -check-prefix=GCN %s
 
 ; GCN-LABEL: {{^}}mac_vvv:
-; GCN: buffer_load_dword [[A:v[0-9]+]], s[{{[0-9]+:[0-9]+}}], 0{{$}}
-; GCN: buffer_load_dword [[B:v[0-9]+]], s[{{[0-9]+:[0-9]+}}], 0 offset:4
-; GCN: buffer_load_dword [[C:v[0-9]+]], s[{{[0-9]+:[0-9]+}}], 0 offset:8
+; GCN: buffer_load_dword [[A:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], 0{{$}}
+; GCN: buffer_load_dword [[B:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], 0 offset:4
+; GCN: buffer_load_dword [[C:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], 0 offset:8
 ; GCN: v_mac_f32_e32 [[C]], [[B]], [[A]]
 ; GCN: buffer_store_dword [[C]]
 define void @mac_vvv(float addrspace(1)* %out, float addrspace(1)* %in) {
diff --git a/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot-compute.ll b/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot-compute.ll
index b3031669e50..12afaa01004 100644
--- a/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot-compute.ll
+++ b/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot-compute.ll
@@ -27,7 +27,7 @@
 ; VIMESA-NEXT: s_mov_b32 s15, 0x980000
 
 
-; GCN: buffer_store_dword {{v[0-9]+}}, s[12:15], s16 offset:{{[0-9]+}} ; 4-byte Folded Spill
+; GCN: buffer_store_dword {{v[0-9]+}}, off, s[12:15], s16 offset:{{[0-9]+}} ; 4-byte Folded Spill
 
 ; GCN: buffer_store_dwordx4 {{v\[[0-9]+:[0-9]+\]}}, {{v[0-9]+}}, s[12:15], s16 offen offset:{{[0-9]+}}
 ; GCN: buffer_load_dwordx4 {{v\[[0-9]+:[0-9]+\]}}, {{v[0-9]+}}, s[12:15], s16 offen offset:{{[0-9]+}}
diff --git a/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll b/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll
index b755b786501..851a9f85ed6 100644
--- a/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll
+++ b/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll
@@ -19,8 +19,8 @@
 ; VI-NEXT: s_mov_b32 s15, 0x980000
 
 ; s12 is offset user SGPR
-; GCN: buffer_store_dword {{v[0-9]+}}, s[12:15], s11 offset:{{[0-9]+}} ; 16-byte Folded Spill
-; GCN: buffer_load_dword v{{[0-9]+}}, s[12:15], s11 offset:{{[0-9]+}} ; 16-byte Folded Reload
+; GCN: buffer_store_dword {{v[0-9]+}}, off, s[12:15], s11 offset:{{[0-9]+}} ; 16-byte Folded Spill
+; GCN: buffer_load_dword v{{[0-9]+}}, off, s[12:15], s11 offset:{{[0-9]+}} ; 16-byte Folded Reload
 
 ; GCN: NumVgprs: 256
 ; GCN: ScratchSize: 1024
diff --git a/llvm/test/MC/AMDGPU/ds-err.s b/llvm/test/MC/AMDGPU/ds-err.s
index 52c2740bec2..3951efbb60f 100644
--- a/llvm/test/MC/AMDGPU/ds-err.s
+++ b/llvm/test/MC/AMDGPU/ds-err.s
@@ -2,15 +2,15 @@
 // RUN: not llvm-mc -arch=amdgcn -mcpu=SI %s 2>&1 | FileCheck %s
 
 // offset too big
-// CHECK: invalid operand for instruction
+// CHECK: error: invalid operand for instruction
 ds_add_u32 v2, v4 offset:1000000000
 
 // offset0 twice
-// CHECK:  error: not a valid operand.
+// CHECK:  error: invalid operand for instruction
 ds_write2_b32 v2, v4, v6 offset0:4 offset0:8
 
 // offset1 twice
-// CHECK:  error: not a valid operand.
+// CHECK:  error: invalid operand for instruction
 ds_write2_b32 v2, v4, v6 offset1:4 offset1:8
 
 // offset0 too big
diff --git a/llvm/test/MC/AMDGPU/mubuf.s b/llvm/test/MC/AMDGPU/mubuf.s
index 84953319c53..9c5d4923308 100644
--- a/llvm/test/MC/AMDGPU/mubuf.s
+++ b/llvm/test/MC/AMDGPU/mubuf.s
@@ -14,34 +14,33 @@
 // load - immediate offset only
 //===----------------------------------------------------------------------===//
 
-buffer_load_dword v1, s[4:7], s1
-// SICI: buffer_load_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x30,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_dword v1, off, s[4:7], s1
+// SICI: buffer_load_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x30,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_dword v1, s[4:7], s1 offset:4
-// SICI: buffer_load_dword v1, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x30,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_dword v1, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_dword v1, off, s[4:7], s1 offset:4
+// SICI: buffer_load_dword v1, off, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x30,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_dword v1, s[4:7], s1 offset:4 glc
-// SICI: buffer_load_dword v1, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x30,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_dword v1, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x50,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_dword v1, off, s[4:7], s1 offset:4 glc
+// SICI: buffer_load_dword v1, off, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x30,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x50,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_dword v1, s[4:7], s1 offset:4 slc
-// SICI: buffer_load_dword v1, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x30,0xe0,0x00,0x01,0x41,0x01]
-// VI:   buffer_load_dword v1, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x52,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_dword v1, off, s[4:7], s1 offset:4 slc
+// SICI: buffer_load_dword v1, off, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x30,0xe0,0x00,0x01,0x41,0x01]
+// VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x52,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_dword v1, s[4:7], s1 offset:4 tfe
-// SICI: buffer_load_dword v1, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x30,0xe0,0x00,0x01,0x81,0x01]
-// VI:   buffer_load_dword v1, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x50,0xe0,0x00,0x01,0x81,0x01]
+buffer_load_dword v1, off, s[4:7], s1 offset:4 tfe
+// SICI: buffer_load_dword v1, off, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x30,0xe0,0x00,0x01,0x81,0x01]
+// VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x50,0xe0,0x00,0x01,0x81,0x01]
 
-buffer_load_dword v1, s[4:7], s1 glc tfe
-// SICI: buffer_load_dword v1, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x30,0xe0,0x00,0x01,0x81,0x01]
-// VI:   buffer_load_dword v1, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x50,0xe0,0x00,0x01,0x81,0x01]
-
-buffer_load_dword v1, s[4:7], s1 offset:4 glc slc tfe
-// SICI: buffer_load_dword v1, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x30,0xe0,0x00,0x01,0xc1,0x01]
-// VI:   buffer_load_dword v1, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x52,0xe0,0x00,0x01,0x81,0x01]
+buffer_load_dword v1, off, s[4:7], s1 glc tfe
+// SICI: buffer_load_dword v1, off, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x30,0xe0,0x00,0x01,0x81,0x01]
+// VI:   buffer_load_dword v1, off, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x50,0xe0,0x00,0x01,0x81,0x01]
 
+buffer_load_dword v1, off, s[4:7], s1 offset:4 glc slc tfe
+// SICI: buffer_load_dword v1, off, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x30,0xe0,0x00,0x01,0xc1,0x01]
+// VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x52,0xe0,0x00,0x01,0x81,0x01]
 
 //===----------------------------------------------------------------------===//
 // load - vgpr offset
@@ -175,33 +174,33 @@ buffer_load_dword v1, v[2:3], s[4:7], s1 addr64 offset:4 glc slc tfe
 // store - immediate offset only
 //===----------------------------------------------------------------------===//
 
-buffer_store_dword v1, s[4:7], s1
-// SICI: buffer_store_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_dword v1, off, s[4:7], s1
+// SICI: buffer_store_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_dword v1, s[4:7], s1 offset:4
-// SICI: buffer_store_dword v1, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_dword v1, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_dword v1, off, s[4:7], s1 offset:4
+// SICI: buffer_store_dword v1, off, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_dword v1, s[4:7], s1 offset:4 glc
-// SICI: buffer_store_dword v1, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x70,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_dword v1, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x70,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_dword v1, off, s[4:7], s1 offset:4 glc
+// SICI: buffer_store_dword v1, off, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x70,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x70,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_dword v1, s[4:7], s1 offset:4 slc
-// SICI: buffer_store_dword v1, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x41,0x01]
-// VI:   buffer_store_dword v1, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x72,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_dword v1, off, s[4:7], s1 offset:4 slc
+// SICI: buffer_store_dword v1, off, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x41,0x01]
+// VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x72,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_dword v1, s[4:7], s1 offset:4 tfe
-// SICI: buffer_store_dword v1, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x81,0x01]
-// VI:   buffer_store_dword v1, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x81,0x01]
+buffer_store_dword v1, off, s[4:7], s1 offset:4 tfe
+// SICI: buffer_store_dword v1, off, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x81,0x01]
+// VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x81,0x01]
 
-buffer_store_dword v1, s[4:7], s1 glc tfe
-// SICI: buffer_store_dword v1, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x70,0xe0,0x00,0x01,0x81,0x01]
-// VI:   buffer_store_dword v1, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x70,0xe0,0x00,0x01,0x81,0x01]
+buffer_store_dword v1, off, s[4:7], s1 glc tfe
+// SICI: buffer_store_dword v1, off, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x70,0xe0,0x00,0x01,0x81,0x01]
+// VI:   buffer_store_dword v1, off, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x70,0xe0,0x00,0x01,0x81,0x01]
 
-buffer_store_dword v1, s[4:7], s1 offset:4 glc slc tfe
-// SICI: buffer_store_dword v1, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x70,0xe0,0x00,0x01,0xc1,0x01]
-// VI:   buffer_store_dword v1, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x72,0xe0,0x00,0x01,0x81,0x01]
+buffer_store_dword v1, off, s[4:7], s1 offset:4 glc slc tfe
+// SICI: buffer_store_dword v1, off, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x70,0xe0,0x00,0x01,0xc1,0x01]
+// VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x72,0xe0,0x00,0x01,0x81,0x01]
 
 //===----------------------------------------------------------------------===//
 // store - vgpr offset
@@ -335,85 +334,85 @@ buffer_store_dword v1, v[2:3], s[4:7], s1 addr64 offset:4 glc slc tfe
 // Instructions
 //===----------------------------------------------------------------------===//
 
-buffer_load_format_x v1, s[4:7], s1
-// SICI: buffer_load_format_x v1, s[4:7], s1 ; encoding: [0x00,0x00,0x00,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_format_x v1, s[4:7], s1 ; encoding: [0x00,0x00,0x00,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_format_x v1, off, s[4:7], s1
+// SICI: buffer_load_format_x v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x00,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_format_x v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x00,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_format_xy v[1:2], s[4:7], s1
-// SICI: buffer_load_format_xy v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x04,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_format_xy v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x04,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_format_xy v[1:2], off, s[4:7], s1
+// SICI: buffer_load_format_xy v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x04,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_format_xy v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x04,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_format_xyz v[1:3], s[4:7], s1
-// SICI: buffer_load_format_xyz v[1:3], s[4:7], s1 ; encoding: [0x00,0x00,0x08,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_format_xyz v[1:3], s[4:7], s1 ; encoding: [0x00,0x00,0x08,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_format_xyz v[1:3], off, s[4:7], s1
+// SICI: buffer_load_format_xyz v[1:3], off, s[4:7], s1 ; encoding: [0x00,0x00,0x08,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_format_xyz v[1:3], off, s[4:7], s1 ; encoding: [0x00,0x00,0x08,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_format_xyzw v[1:4], s[4:7], s1
-// SICI: buffer_load_format_xyzw v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x0c,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_format_xyzw v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x0c,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_format_xyzw v[1:4], off, s[4:7], s1
+// SICI: buffer_load_format_xyzw v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x0c,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_format_xyzw v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x0c,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_format_x v1, s[4:7], s1
-// SICI: buffer_store_format_x v1, s[4:7], s1 ; encoding: [0x00,0x00,0x10,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_format_x v1, s[4:7], s1 ; encoding: [0x00,0x00,0x10,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_format_x v1, off, s[4:7], s1
+// SICI: buffer_store_format_x v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x10,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_format_x v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x10,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_format_xy v[1:2], s[4:7], s1
-// SICI: buffer_store_format_xy v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x14,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_format_xy v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x14,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_format_xy v[1:2], off, s[4:7], s1
+// SICI: buffer_store_format_xy v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x14,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_format_xy v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x14,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_format_xyz v[1:3], s[4:7], s1
-// SICI: buffer_store_format_xyz v[1:3], s[4:7], s1 ; encoding: [0x00,0x00,0x18,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_format_xyz v[1:3], s[4:7], s1 ; encoding: [0x00,0x00,0x18,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_format_xyz v[1:3], off, s[4:7], s1
+// SICI: buffer_store_format_xyz v[1:3], off, s[4:7], s1 ; encoding: [0x00,0x00,0x18,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_format_xyz v[1:3], off, s[4:7], s1 ; encoding: [0x00,0x00,0x18,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_format_xyzw v[1:4], s[4:7], s1
-// SICI: buffer_store_format_xyzw v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x1c,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_format_xyzw v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x1c,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_format_xyzw v[1:4], off, s[4:7], s1
+// SICI: buffer_store_format_xyzw v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x1c,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_format_xyzw v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x1c,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_ubyte v1, s[4:7], s1
-// SICI: buffer_load_ubyte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x20,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_ubyte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x40,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_ubyte v1, off, s[4:7], s1
+// SICI: buffer_load_ubyte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x20,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_ubyte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x40,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_sbyte v1, s[4:7], s1
-// SICI: buffer_load_sbyte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x24,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_sbyte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x44,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_sbyte v1, off, s[4:7], s1
+// SICI: buffer_load_sbyte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x24,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_sbyte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x44,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_ushort v1, s[4:7], s1
-// SICI: buffer_load_ushort v1, s[4:7], s1 ; encoding: [0x00,0x00,0x28,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_ushort v1, s[4:7], s1 ; encoding: [0x00,0x00,0x48,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_ushort v1, off, s[4:7], s1
+// SICI: buffer_load_ushort v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x28,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_ushort v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x48,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_sshort v1, s[4:7], s1
-// SICI: buffer_load_sshort v1, s[4:7], s1 ; encoding: [0x00,0x00,0x2c,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_sshort v1, s[4:7], s1 ; encoding: [0x00,0x00,0x4c,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_sshort v1, off, s[4:7], s1
+// SICI: buffer_load_sshort v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x2c,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_sshort v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x4c,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_dword v1, s[4:7], s1
-// SICI: buffer_load_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x30,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_dword v1, off, s[4:7], s1
+// SICI: buffer_load_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x30,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_dwordx2 v[1:2], s[4:7], s1
-// SICI: buffer_load_dwordx2 v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x34,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_dwordx2 v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x54,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_dwordx2 v[1:2], off, s[4:7], s1
+// SICI: buffer_load_dwordx2 v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x34,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_dwordx2 v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x54,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_load_dwordx4 v[1:4], s[4:7], s1
-// SICI: buffer_load_dwordx4 v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x38,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_dwordx4 v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x5c,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_dwordx4 v[1:4], off, s[4:7], s1
+// SICI: buffer_load_dwordx4 v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x38,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_dwordx4 v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x5c,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_byte v1, s[4:7], s1
-// SICI: buffer_store_byte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x60,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_byte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x60,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_byte v1, off, s[4:7], s1
+// SICI: buffer_store_byte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x60,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_byte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x60,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_short v1, s[4:7], s1
-// SICI: buffer_store_short v1, s[4:7], s1 ; encoding: [0x00,0x00,0x68,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_short v1, s[4:7], s1 ; encoding: [0x00,0x00,0x68,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_short v1, off, s[4:7], s1
+// SICI: buffer_store_short v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x68,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_short v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x68,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_dword v1 s[4:7], s1
-// SICI: buffer_store_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_dword v1, off, s[4:7], s1
+// SICI: buffer_store_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_dwordx2 v[1:2], s[4:7], s1
-// SICI: buffer_store_dwordx2 v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x74,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_dwordx2 v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x74,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_dwordx2 v[1:2], off, s[4:7], s1
+// SICI: buffer_store_dwordx2 v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x74,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_dwordx2 v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x74,0xe0,0x00,0x01,0x01,0x01]
 
-buffer_store_dwordx4 v[1:4], s[4:7], s1
-// SICI: buffer_store_dwordx4 v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x78,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_store_dwordx4 v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x7c,0xe0,0x00,0x01,0x01,0x01]
+buffer_store_dwordx4 v[1:4], off, s[4:7], s1
+// SICI: buffer_store_dwordx4 v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x78,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_store_dwordx4 v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x7c,0xe0,0x00,0x01,0x01,0x01]
 
 //===----------------------------------------------------------------------===//
 // Cache invalidation
diff --git a/llvm/test/MC/AMDGPU/reg-syntax-extra.s b/llvm/test/MC/AMDGPU/reg-syntax-extra.s
index fb7778f0585..996ced66b35 100644
--- a/llvm/test/MC/AMDGPU/reg-syntax-extra.s
+++ b/llvm/test/MC/AMDGPU/reg-syntax-extra.s
@@ -50,6 +50,6 @@ v_rcp_f64 [v1,v2], [v2,v3]
 // SICI: v_rcp_f64_e32 v[1:2], v[2:3] ; encoding: [0x02,0x5f,0x02,0x7e]
 // VI:   v_rcp_f64_e32 v[1:2], v[2:3] ; encoding: [0x02,0x4b,0x02,0x7e]
 
-buffer_load_dwordx4 [v1,v2,v3,v4], [s4,s5,s6,s7], s1
-// SICI: buffer_load_dwordx4 v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x38,0xe0,0x00,0x01,0x01,0x01]
-// VI:   buffer_load_dwordx4 v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x5c,0xe0,0x00,0x01,0x01,0x01]
+buffer_load_dwordx4 [v1,v2,v3,v4], off, [s4,s5,s6,s7], s1
+// SICI: buffer_load_dwordx4 v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x38,0xe0,0x00,0x01,0x01,0x01]
+// VI:   buffer_load_dwordx4 v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x5c,0xe0,0x00,0x01,0x01,0x01]
diff --git a/llvm/test/MC/Disassembler/AMDGPU/mubuf_vi.txt b/llvm/test/MC/Disassembler/AMDGPU/mubuf_vi.txt
index a54e62c3028..830fa5371b7 100644
--- a/llvm/test/MC/Disassembler/AMDGPU/mubuf_vi.txt
+++ b/llvm/test/MC/Disassembler/AMDGPU/mubuf_vi.txt
@@ -1,24 +1,24 @@
 # RUN: llvm-mc -arch=amdgcn -mcpu=tonga -disassemble -show-encoding < %s | FileCheck %s -check-prefix=VI
 
-# VI:   buffer_load_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x50 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_dword v1, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
 0x04 0x00 0x50 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_dword v1, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x50,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x50,0xe0,0x00,0x01,0x01,0x01]
 0x04 0x40 0x50 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_dword v1, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x52,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x52,0xe0,0x00,0x01,0x01,0x01]
 0x04 0x00 0x52 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_dword v1, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x50,0xe0,0x00,0x01,0x81,0x01]
+# VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x50,0xe0,0x00,0x01,0x81,0x01]
 0x04 0x00 0x50 0xe0 0x00 0x01 0x81 0x01
 
-# VI:   buffer_load_dword v1, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x50,0xe0,0x00,0x01,0x81,0x01]
+# VI:   buffer_load_dword v1, off, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x50,0xe0,0x00,0x01,0x81,0x01]
 0x00 0x40 0x50 0xe0 0x00 0x01 0x81 0x01
 
-# VI:   buffer_load_dword v1, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x52,0xe0,0x00,0x01,0x81,0x01]
+# VI:   buffer_load_dword v1, off, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x52,0xe0,0x00,0x01,0x81,0x01]
 0x04 0x40 0x52 0xe0 0x00 0x01 0x81 0x01
 
 # VI:   buffer_load_dword v1, v2, s[4:7], s1 offen ; encoding: [0x00,0x10,0x50,0xe0,0x02,0x01,0x01,0x01]
@@ -84,25 +84,25 @@
 # VI:   buffer_load_dword v1, v[2:3], s[4:7], s1 idxen offen offset:4 glc slc tfe ; encoding: [0x04,0x70,0x52,0xe0,0x02,0x01,0x81,0x01]
 0x04 0x70 0x52 0xe0 0x02 0x01 0x81 0x01
 
-# VI:   buffer_store_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x70 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_dword v1, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
 0x04 0x00 0x70 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_dword v1, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x70,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 glc ; encoding: [0x04,0x40,0x70,0xe0,0x00,0x01,0x01,0x01]
 0x04 0x40 0x70 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_dword v1, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x72,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 slc ; encoding: [0x04,0x00,0x72,0xe0,0x00,0x01,0x01,0x01]
 0x04 0x00 0x72 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_dword v1, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x81,0x01]
+# VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 tfe ; encoding: [0x04,0x00,0x70,0xe0,0x00,0x01,0x81,0x01]
 0x04 0x00 0x70 0xe0 0x00 0x01 0x81 0x01
 
-# VI:   buffer_store_dword v1, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x70,0xe0,0x00,0x01,0x81,0x01]
+# VI:   buffer_store_dword v1, off, s[4:7], s1 glc tfe ; encoding: [0x00,0x40,0x70,0xe0,0x00,0x01,0x81,0x01]
 0x00 0x40 0x70 0xe0 0x00 0x01 0x81 0x01
 
-# VI:   buffer_store_dword v1, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x72,0xe0,0x00,0x01,0x81,0x01]
+# VI:   buffer_store_dword v1, off, s[4:7], s1 offset:4 glc slc tfe ; encoding: [0x04,0x40,0x72,0xe0,0x00,0x01,0x81,0x01]
 0x04 0x40 0x72 0xe0 0x00 0x01 0x81 0x01
 
 # VI:   buffer_store_dword v1, v2, s[4:7], s1 offen ; encoding: [0x00,0x10,0x70,0xe0,0x02,0x01,0x01,0x01]
@@ -168,64 +168,64 @@
 # VI:   buffer_store_dword v1, v[2:3], s[4:7], s1 idxen offen offset:4 glc slc tfe ; encoding: [0x04,0x70,0x72,0xe0,0x02,0x01,0x81,0x01]
 0x04 0x70 0x72 0xe0 0x02 0x01 0x81 0x01
 
-# VI:   buffer_load_format_x v1, s[4:7], s1 ; encoding: [0x00,0x00,0x00,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_format_x v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x00,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x00 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_format_xy v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x04,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_format_xy v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x04,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x04 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_format_xyz v[1:3], s[4:7], s1 ; encoding: [0x00,0x00,0x08,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_format_xyz v[1:3], off, s[4:7], s1 ; encoding: [0x00,0x00,0x08,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x08 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_format_xyzw v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x0c,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_format_xyzw v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x0c,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x0c 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_format_x v1, s[4:7], s1 ; encoding: [0x00,0x00,0x10,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_format_x v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x10,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x10 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_format_xy v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x14,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_format_xy v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x14,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x14 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_format_xyz v[1:3], s[4:7], s1 ; encoding: [0x00,0x00,0x18,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_format_xyz v[1:3], off, s[4:7], s1 ; encoding: [0x00,0x00,0x18,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x18 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_format_xyzw v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x1c,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_format_xyzw v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x1c,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x1c 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_ubyte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x40,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_ubyte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x40,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x40 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_sbyte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x44,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_sbyte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x44,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x44 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_ushort v1, s[4:7], s1 ; encoding: [0x00,0x00,0x48,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_ushort v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x48,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x48 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_sshort v1, s[4:7], s1 ; encoding: [0x00,0x00,0x4c,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_sshort v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x4c,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x4c 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x50,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x50 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_dwordx2 v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x54,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_dwordx2 v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x54,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x54 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_load_dwordx4 v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x5c,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_load_dwordx4 v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x5c,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x5c 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_byte v1, s[4:7], s1 ; encoding: [0x00,0x00,0x60,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_byte v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x60,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x60 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_short v1, s[4:7], s1 ; encoding: [0x00,0x00,0x68,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_short v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x68,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x68 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_dword v1, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_dword v1, off, s[4:7], s1 ; encoding: [0x00,0x00,0x70,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x70 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_dwordx2 v[1:2], s[4:7], s1 ; encoding: [0x00,0x00,0x74,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_dwordx2 v[1:2], off, s[4:7], s1 ; encoding: [0x00,0x00,0x74,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x74 0xe0 0x00 0x01 0x01 0x01
 
-# VI:   buffer_store_dwordx4 v[1:4], s[4:7], s1 ; encoding: [0x00,0x00,0x7c,0xe0,0x00,0x01,0x01,0x01]
+# VI:   buffer_store_dwordx4 v[1:4], off, s[4:7], s1 ; encoding: [0x00,0x00,0x7c,0xe0,0x00,0x01,0x01,0x01]
 0x00 0x00 0x7c 0xe0 0x00 0x01 0x01 0x01
 
 # VI:   buffer_wbinvl1   ; encoding: [0x00,0x00,0xf8,0xe0,0x00,0x00,0x00,0x00]