Add missing gfx6 run line

Author: arsenm

llvm/test/CodeGen/AMDGPU/GlobalISel/atomic_load_global.ll

@@ -1,9 +1,20 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -mcpu=tahiti < %s | FileCheck -check-prefixes=GCN,GFX6 %s
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -mcpu=kaveri < %s | FileCheck -check-prefixes=GCN,GFX7 %s
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -mcpu=fiji < %s | FileCheck -check-prefixes=GCN,GFX8 %s
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 < %s | FileCheck -check-prefixes=GCN,GFX9 %s
 
 define i8 @atomic_load_global_monotonic_i8(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i8:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i8:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -29,6 +40,16 @@ define i8 @atomic_load_global_monotonic_i8(ptr addrspace(1) %ptr) {
 }
 
 define i32 @atomic_load_global_monotonic_i8_zext_to_i32(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i8_zext_to_i32:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i8_zext_to_i32:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -55,6 +76,18 @@ define i32 @atomic_load_global_monotonic_i8_zext_to_i32(ptr addrspace(1) %ptr) {
 }
 
 define i32 @atomic_load_global_monotonic_i8_sext_to_i32(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i8_sext_to_i32:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_sbyte v2, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    v_mov_b32_e32 v0, v2
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i8_sext_to_i32:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -88,6 +121,16 @@ define i32 @atomic_load_global_monotonic_i8_sext_to_i32(ptr addrspace(1) %ptr) {
 }
 
 define i16 @atomic_load_global_monotonic_i8_zext_to_i16(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i8_zext_to_i16:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i8_zext_to_i16:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -114,6 +157,18 @@ define i16 @atomic_load_global_monotonic_i8_zext_to_i16(ptr addrspace(1) %ptr) {
 }
 
 define i16 @atomic_load_global_monotonic_i8_sext_to_i16(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i8_sext_to_i16:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_sbyte v2, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    v_mov_b32_e32 v0, v2
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i8_sext_to_i16:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -147,6 +202,16 @@ define i16 @atomic_load_global_monotonic_i8_sext_to_i16(ptr addrspace(1) %ptr) {
 }
 
 define i16 @atomic_load_global_monotonic_i16(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i16:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ushort v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i16:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -172,6 +237,16 @@ define i16 @atomic_load_global_monotonic_i16(ptr addrspace(1) %ptr) {
 }
 
 define i32 @atomic_load_global_monotonic_i16_zext_to_i32(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i16_zext_to_i32:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i16_zext_to_i32:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -198,6 +273,18 @@ define i32 @atomic_load_global_monotonic_i16_zext_to_i32(ptr addrspace(1) %ptr)
 }
 
 define i32 @atomic_load_global_monotonic_i16_sext_to_i32(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i16_sext_to_i32:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_sbyte v2, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    buffer_load_ushort v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    v_mov_b32_e32 v0, v2
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i16_sext_to_i32:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -231,6 +318,16 @@ define i32 @atomic_load_global_monotonic_i16_sext_to_i32(ptr addrspace(1) %ptr)
 }
 
 define half @atomic_load_global_monotonic_f16(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_f16:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ushort v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_f16:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -256,6 +353,16 @@ define half @atomic_load_global_monotonic_f16(ptr addrspace(1) %ptr) {
 }
 
 define bfloat @atomic_load_global_monotonic_bf16(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_bf16:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ushort v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_bf16:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -281,6 +388,16 @@ define bfloat @atomic_load_global_monotonic_bf16(ptr addrspace(1) %ptr) {
 }
 
 define i32 @atomic_load_global_monotonic_f16_zext_to_i32(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_f16_zext_to_i32:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_f16_zext_to_i32:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -308,6 +425,16 @@ define i32 @atomic_load_global_monotonic_f16_zext_to_i32(ptr addrspace(1) %ptr)
 }
 
 define i32 @atomic_load_global_monotonic_bf16_zext_to_i32(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_bf16_zext_to_i32:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_bf16_zext_to_i32:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -335,6 +462,17 @@ define i32 @atomic_load_global_monotonic_bf16_zext_to_i32(ptr addrspace(1) %ptr)
 }
 
 define i32 @atomic_load_global_monotonic_i16_d16_hi_shift(ptr addrspace(1) %ptr) {
+; GFX6-LABEL: atomic_load_global_monotonic_i16_d16_hi_shift:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    v_lshlrev_b32_e32 v0, 16, v0
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i16_d16_hi_shift:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -365,6 +503,23 @@ define i32 @atomic_load_global_monotonic_i16_d16_hi_shift(ptr addrspace(1) %ptr)
 }
 
 define <2 x i16> @atomic_load_global_monotonic_i16_d16_hi_vector_insert(ptr addrspace(1) %ptr, <2 x i16> %vec) {
+; GFX6-LABEL: atomic_load_global_monotonic_i16_d16_hi_vector_insert:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    v_lshlrev_b32_e32 v1, 16, v3
+; GFX6-NEXT:    v_and_b32_e32 v2, 0xffff, v2
+; GFX6-NEXT:    v_or_b32_e32 v1, v1, v2
+; GFX6-NEXT:    v_and_b32_e32 v1, 0xffff, v1
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    v_lshlrev_b32_e32 v0, 16, v0
+; GFX6-NEXT:    v_or_b32_e32 v0, v1, v0
+; GFX6-NEXT:    v_lshrrev_b32_e32 v1, 16, v0
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i16_d16_hi_vector_insert:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -403,6 +558,19 @@ define <2 x i16> @atomic_load_global_monotonic_i16_d16_hi_vector_insert(ptr addr
 }
 
 define i32 @atomic_load_global_monotonic_i16_d16_lo_or(ptr addrspace(1) %ptr, i16 %high) {
+; GFX6-LABEL: atomic_load_global_monotonic_i16_d16_lo_or:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    v_and_b32_e32 v1, 0xffff, v2
+; GFX6-NEXT:    v_lshlrev_b32_e32 v1, 16, v1
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    v_or_b32_e32 v0, v1, v0
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i16_d16_lo_or:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@@ -440,6 +608,22 @@ define i32 @atomic_load_global_monotonic_i16_d16_lo_or(ptr addrspace(1) %ptr, i1
 }
 
 define <2 x i16> @atomic_load_global_monotonic_i16_d16_lo_vector_insert(ptr addrspace(1) %ptr, <2 x i16> %vec) {
+; GFX6-LABEL: atomic_load_global_monotonic_i16_d16_lo_vector_insert:
+; GFX6:       ; %bb.0:
+; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0x100f000
+; GFX6-NEXT:    s_mov_b64 s[4:5], 0
+; GFX6-NEXT:    buffer_load_ubyte v0, v[0:1], s[4:7], 0 addr64 glc
+; GFX6-NEXT:    v_lshlrev_b32_e32 v1, 16, v3
+; GFX6-NEXT:    v_and_b32_e32 v2, 0xffff, v2
+; GFX6-NEXT:    v_or_b32_e32 v1, v1, v2
+; GFX6-NEXT:    v_and_b32_e32 v1, 0xffff0000, v1
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    v_or_b32_e32 v0, v1, v0
+; GFX6-NEXT:    v_lshrrev_b32_e32 v1, 16, v0
+; GFX6-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX7-LABEL: atomic_load_global_monotonic_i16_d16_lo_vector_insert:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)