[MachineScheduler] Reduce reordering due to mem op clustering

[foad]

Summary

Mem op clustering adds a weak edge in the DAG between two loads or
stores that should be clustered, but the direction of this edge is
pretty arbitrary (it depends on the sort order of MemOpInfo, which
represents the operands of a load or store). This often means that two
loads or stores will get reordered even if they would naturally have
been scheduled together anyway, which leads to test case churn and goes
against the scheduler's "do no harm" philosophy.

The fix makes sure that the direction of the edge always matches the
original code order of the instructions.

[merge_guards_bot]

Unit tests: pass. 61818 tests passed, 0 failed and 781 were skipped.

clang-tidy: unknown.

clang-format: pass.

Build artifacts: diff.json, clang-format.patch, CMakeCache.txt, console-log.txt, test-results.xml

[rampitec]

LGTM

llvm/lib/CodeGen/MachineScheduler.cpp

void BaseMemOpClusterMutation::clusterNeighboringMemOps(
   if (MemOpRecords.size() < 2)
     return;
 
   llvm::sort(MemOpRecords);
   unsigned ClusterLength = 1;
   for (unsigned Idx = 0, End = MemOpRecords.size(); Idx < (End - 1); ++Idx) {
     SUnit *SUa = MemOpRecords[Idx].SU;
     SUnit *SUb = MemOpRecords[Idx+1].SU;
+    if (SUa->NodeNum > SUb->NodeNum)
+      std::swap(SUa, SUb);
     if (TII->shouldClusterMemOps(*MemOpRecords[Idx].BaseOp,
                                  *MemOpRecords[Idx + 1].BaseOp,
                                  ClusterLength) &&
         DAG->addEdge(SUb, SDep(SUa, SDep::Cluster))) {
       LLVM_DEBUG(dbgs() << "Cluster ld/st SU(" << SUa->NodeNum << ") - SU("
                         << SUb->NodeNum << ")\n");
       // Copy successor edges from SUa to SUb. Interleaving computation
       // dependent on SUa can prevent load combining due to register reuse.

llvm/test/CodeGen/AArch64/aarch64-stp-cluster.ll

 ; REQUIRES: asserts
 ; RUN: llc < %s -mtriple=arm64-linux-gnu -mcpu=cortex-a57 -verify-misched -debug-only=machine-scheduler -aarch64-enable-stp-suppress=false -o - 2>&1 > /dev/null | FileCheck %s
 
 ; CHECK: ********** MI Scheduling **********
 ; CHECK-LABEL: stp_i64_scale:%bb.0
-; CHECK:Cluster ld/st SU(4) - SU(3)
+; CHECK:Cluster ld/st SU(3) - SU(4)
 ; CHECK:Cluster ld/st SU(2) - SU(5)
 ; CHECK:SU(4):   STRXui %1:gpr64, %0:gpr64common, 1
 ; CHECK:SU(3):   STRXui %1:gpr64, %0:gpr64common, 2
 ; CHECK:SU(2):   STRXui %1:gpr64, %0:gpr64common, 3
 ; CHECK:SU(5):   STRXui %1:gpr64, %0:gpr64common, 4
 define i64 @stp_i64_scale(i64* nocapture %P, i64 %v) {
 entry:
   %arrayidx = getelementptr inbounds i64, i64* %P, i64 3
   store i64 %v, i64* %arrayidx
   %arrayidx1 = getelementptr inbounds i64, i64* %P, i64 2
   store i64 %v, i64* %arrayidx1
   %arrayidx2 = getelementptr inbounds i64, i64* %P, i64 1
   store i64 %v, i64* %arrayidx2
   %arrayidx3 = getelementptr inbounds i64, i64* %P, i64 4
   store i64 %v, i64* %arrayidx3
   ret i64 %v
 }
 
 ; CHECK: ********** MI Scheduling **********
 ; CHECK-LABEL: stp_i32_scale:%bb.0
-; CHECK:Cluster ld/st SU(4) - SU(3)
+; CHECK:Cluster ld/st SU(3) - SU(4)
 ; CHECK:Cluster ld/st SU(2) - SU(5)
 ; CHECK:SU(4):   STRWui %1:gpr32, %0:gpr64common, 1
 ; CHECK:SU(3):   STRWui %1:gpr32, %0:gpr64common, 2
 ; CHECK:SU(2):   STRWui %1:gpr32, %0:gpr64common, 3
 ; CHECK:SU(5):   STRWui %1:gpr32, %0:gpr64common, 4
 define i32 @stp_i32_scale(i32* nocapture %P, i32 %v) {
 entry:
   %arrayidx = getelementptr inbounds i32, i32* %P, i32 3
   store i32 %v, i32* %arrayidx
   %arrayidx1 = getelementptr inbounds i32, i32* %P, i32 2
   store i32 %v, i32* %arrayidx1
   %arrayidx2 = getelementptr inbounds i32, i32* %P, i32 1
   store i32 %v, i32* %arrayidx2
   %arrayidx3 = getelementptr inbounds i32, i32* %P, i32 4
   store i32 %v, i32* %arrayidx3
   ret i32 %v
 }
 
 ; CHECK:********** MI Scheduling **********
 ; CHECK-LABEL:stp_i64_unscale:%bb.0 entry
-; CHECK:Cluster ld/st SU(5) - SU(2)
-; CHECK:Cluster ld/st SU(4) - SU(3)
-; CHECK:SU(5):   STURXi %1:gpr64, %0:gpr64common, -32
+; CHECK:Cluster ld/st SU(2) - SU(5)
+; CHECK:Cluster ld/st SU(3) - SU(4)
 ; CHECK:SU(2):   STURXi %1:gpr64, %0:gpr64common, -24
-; CHECK:SU(4):   STURXi %1:gpr64, %0:gpr64common, -16
 ; CHECK:SU(3):   STURXi %1:gpr64, %0:gpr64common, -8
+; CHECK:SU(4):   STURXi %1:gpr64, %0:gpr64common, -16
+; CHECK:SU(5):   STURXi %1:gpr64, %0:gpr64common, -32
 define void @stp_i64_unscale(i64* nocapture %P, i64 %v) #0 {
 entry:
   %arrayidx = getelementptr inbounds i64, i64* %P, i64 -3
   store i64 %v, i64* %arrayidx
   %arrayidx1 = getelementptr inbounds i64, i64* %P, i64 -1
   store i64 %v, i64* %arrayidx1
   %arrayidx2 = getelementptr inbounds i64, i64* %P, i64 -2
   store i64 %v, i64* %arrayidx2
   %arrayidx3 = getelementptr inbounds i64, i64* %P, i64 -4
   store i64 %v, i64* %arrayidx3
   ret void
 }
 
 ; CHECK:********** MI Scheduling **********
 ; CHECK-LABEL:stp_i32_unscale:%bb.0 entry
-; CHECK:Cluster ld/st SU(5) - SU(2)
-; CHECK:Cluster ld/st SU(4) - SU(3)
-; CHECK:SU(5):   STURWi %1:gpr32, %0:gpr64common, -16
+; CHECK:Cluster ld/st SU(2) - SU(5)
+; CHECK:Cluster ld/st SU(3) - SU(4)
 ; CHECK:SU(2):   STURWi %1:gpr32, %0:gpr64common, -12
-; CHECK:SU(4):   STURWi %1:gpr32, %0:gpr64common, -8
 ; CHECK:SU(3):   STURWi %1:gpr32, %0:gpr64common, -4
+; CHECK:SU(4):   STURWi %1:gpr32, %0:gpr64common, -8
+; CHECK:SU(5):   STURWi %1:gpr32, %0:gpr64common, -16
 define void @stp_i32_unscale(i32* nocapture %P, i32 %v) #0 {
 entry:
   %arrayidx = getelementptr inbounds i32, i32* %P, i32 -3
   store i32 %v, i32* %arrayidx
   %arrayidx1 = getelementptr inbounds i32, i32* %P, i32 -1
   store i32 %v, i32* %arrayidx1
   %arrayidx2 = getelementptr inbounds i32, i32* %P, i32 -2
   store i32 %v, i32* %arrayidx2

llvm/test/CodeGen/AArch64/arm64-ldp-cluster.ll

define i64 @ldp_sext_int(i32* %p) nounwind {
   %sexttmp1 = sext i32 %tmp1 to i64
   %add = add nsw i64 %sexttmp1, %sexttmp
   ret i64 %add
 }
 
 ; Test ldur clustering.
 ; CHECK: ********** MI Scheduling **********
 ; CHECK-LABEL: ldur_int:%bb.0
-; CHECK: Cluster ld/st SU(2) - SU(1)
+; CHECK: Cluster ld/st SU(1) - SU(2)
 ; CHECK: SU(1):   %{{[0-9]+}}:gpr32 = LDURWi
 ; CHECK: SU(2):   %{{[0-9]+}}:gpr32 = LDURWi
 define i32 @ldur_int(i32* %a) nounwind {
   %p1 = getelementptr inbounds i32, i32* %a, i32 -1
   %tmp1 = load i32, i32* %p1, align 2
   %p2 = getelementptr inbounds i32, i32* %a, i32 -2
   %tmp2 = load i32, i32* %p2, align 2
   %tmp3 = add i32 %tmp1, %tmp2

llvm/test/CodeGen/AArch64/arm64-memset-inline.ll

; CHECK-NEXT:  bl something
   call void @llvm.memset.p0i8.i32(i8* %cast, i8 -86, i32 12, i1 false)
   call void @something(i8* %cast)
   ret void
 }
 
 define void @memset_16_stack() {
 ; CHECK-LABEL: memset_16_stack:
 ; CHECK:       mov x8, #-6148914691236517206
-; CHECK-NEXT:  str x8, [sp, #-32]!
 ; CHECK-NEXT:  mov x0, sp
 ; CHECK-NEXT:  stp x8, x30, [sp, #8]
+; CHECK-NEXT:  str x8, [sp]
 ; CHECK-NEXT:  bl something
   %buf = alloca [16 x i8], align 1
   %cast = bitcast [16 x i8]* %buf to i8*
   call void @llvm.memset.p0i8.i32(i8* %cast, i8 -86, i32 16, i1 false)
   call void @something(i8* %cast)
   ret void
 }
 

llvm/test/CodeGen/AArch64/expand-select.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=aarch64-unknown-linux-gnu -O3 %s -o - | FileCheck %s
 
 define void @foo(i32 %In1, <2 x i128> %In2, <2 x i128> %In3, <2 x i128> *%Out) {
 ; CHECK-LABEL: foo:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    and w8, w0, #0x1
 ; CHECK-NEXT:    fmov s0, wzr
-; CHECK-NEXT:    ldp x10, x9, [sp, #8]
 ; CHECK-NEXT:    fmov s1, w8
-; CHECK-NEXT:    ldr x8, [sp]
+; CHECK-NEXT:    ldp x8, x9, [sp, #8]
+; CHECK-NEXT:    ldr x10, [sp]
 ; CHECK-NEXT:    cmeq v0.4s, v1.4s, v0.4s
 ; CHECK-NEXT:    fmov w11, s0
 ; CHECK-NEXT:    tst w11, #0x1
 ; CHECK-NEXT:    csel x11, x2, x6, ne
 ; CHECK-NEXT:    csel x12, x3, x7, ne
-; CHECK-NEXT:    csel x8, x4, x8, ne
-; CHECK-NEXT:    csel x10, x5, x10, ne
-; CHECK-NEXT:    stp x8, x10, [x9, #16]
+; CHECK-NEXT:    csel x10, x4, x10, ne
+; CHECK-NEXT:    csel x8, x5, x8, ne
+; CHECK-NEXT:    stp x10, x8, [x9, #16]
 ; CHECK-NEXT:    stp x11, x12, [x9]
 ; CHECK-NEXT:    ret
   %cond = and i32 %In1, 1
   %cbool = icmp eq i32 %cond, 0
   %res = select i1 %cbool, <2 x i128> %In2, <2 x i128> %In3
   store <2 x i128> %res, <2 x i128> *%Out
 
   ret void

llvm/test/CodeGen/AArch64/global-merge-group-by-use.ll

 
 @m4 = internal global i32 0, align 4
 @n4 = internal global i32 0, align 4
 
 ; CHECK-LABEL: f4:
 define void @f4(i32 %a1, i32 %a2, i32 %a3) #0 {
 ; CHECK-NEXT: adrp x8, [[SET3]]@PAGE
 ; CHECK-NEXT: add x8, x8, [[SET3]]@PAGEOFF
-; CHECK-NEXT: stp w2, w0, [x8]
-; CHECK-NEXT: str w1, [x8, #8]
+; CHECK-NEXT: stp w0, w1, [x8, #4]
+; CHECK-NEXT: str w2, [x8]
 ; CHECK-NEXT: ret
   store i32 %a1, i32* @m4, align 4
   store i32 %a2, i32* @n4, align 4
   store i32 %a3, i32* @n3, align 4
   ret void
 }
 
 ; Finally, check that we don't do anything with one-element global sets.

llvm/test/CodeGen/AArch64/machine-scheduler.mir

entry:
     %and = and i64 %2, 4294967295
     %add = add nuw nsw i64 %and, %conv
     ret i64 %add
   }
 ...
 ---
 # CHECK-LABEL: name: load_imp-def
 # CHECK: bb.0.entry:
-# CHECK: LDRWui $x0, 0
 # CHECK: LDRWui $x0, 1
+# CHECK: LDRWui $x0, 0
 # CHECK: STRWui $w1, $x0, 2
 name: load_imp-def
 tracksRegLiveness: true
 body: |
   bb.0.entry:
     liveins: $w1, $x0
     $w8 = LDRWui $x0, 1, implicit-def $x8  :: (load 4 from %ir.0)
     STRWui killed $w1, $x0, 2 :: (store 4 into %ir.arrayidx1)
     $w9 = LDRWui killed $x0, 0, implicit-def $x9  :: (load 4 from %ir.arrayidx19, align 8)
     $x0 = ADDXrr killed $x9, killed $x8
     RET_ReallyLR implicit $x0
 ...
 

llvm/test/CodeGen/AMDGPU/bitcast-constant-to-vector.ll

 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
 
 ; GCN-LABEL: {{^}}cast_constant_i64_to_build_vector_v4i16:
-; GCN: global_store_dwordx2
-; GCN: global_store_dword v
 ; GCN: global_store_short
+; GCN: global_store_dword v
+; GCN: global_store_dwordx2
 define amdgpu_kernel void @cast_constant_i64_to_build_vector_v4i16(i8 addrspace(1)* nocapture %data) {
 entry:
   store i8 72, i8 addrspace(1)* %data, align 1
   %arrayidx1 = getelementptr inbounds i8, i8 addrspace(1)* %data, i64 1
   store i8 101, i8 addrspace(1)* %arrayidx1, align 1
   %arrayidx2 = getelementptr inbounds i8, i8 addrspace(1)* %data, i64 2
   store i8 108, i8 addrspace(1)* %arrayidx2, align 1
   %arrayidx3 = getelementptr inbounds i8, i8 addrspace(1)* %data, i64 3

llvm/test/CodeGen/AMDGPU/byval-frame-setup.ll

 ; GCN-NOT: s_add_u32 s32, s32, 0x800
 
 
 ; GCN-DAG: buffer_store_dword [[LOAD0]], off, s[0:3], s32{{$}}
 ; GCN-DAG: buffer_store_dword [[LOAD1]], off, s[0:3], s32 offset:4
 ; GCN-DAG: buffer_store_dword [[LOAD2]], off, s[0:3], s32 offset:8
 ; GCN-DAG: buffer_store_dword [[LOAD3]], off, s[0:3], s32 offset:12
 
+; GCN: buffer_load_dword [[LOAD4:v[0-9]+]], off, s[0:3], s34 offset:16
 ; GCN: buffer_load_dword [[LOAD5:v[0-9]+]], off, s[0:3], s34 offset:20
 ; GCN: buffer_load_dword [[LOAD6:v[0-9]+]], off, s[0:3], s34 offset:24
 ; GCN: buffer_load_dword [[LOAD7:v[0-9]+]], off, s[0:3], s34 offset:28
-; GCN: buffer_load_dword [[LOAD4:v[0-9]+]], off, s[0:3], s34 offset:16
 
 ; GCN-DAG: buffer_store_dword [[LOAD4]], off, s[0:3], s32 offset:16
 ; GCN-DAG: buffer_store_dword [[LOAD5]], off, s[0:3], s32 offset:20
 ; GCN-DAG: buffer_store_dword [[LOAD6]], off, s[0:3], s32 offset:24
 ; GCN-DAG: buffer_store_dword [[LOAD7]], off, s[0:3], s32 offset:28
 
 ; GCN: s_swappc_b64
 ; GCN-NOT: v_readlane_b32 s32

 
 ; GCN-NOT: s_add_u32 s32, s32, 0x800
 
 ; GCN-DAG: buffer_store_dword [[LOAD0]], off, s[0:3], s32{{$}}
 ; GCN-DAG: buffer_store_dword [[LOAD1]], off, s[0:3], s32 offset:4
 ; GCN-DAG: buffer_store_dword [[LOAD2]], off, s[0:3], s32 offset:8
 ; GCN-DAG: buffer_store_dword [[LOAD3]], off, s[0:3], s32 offset:12
 
+; GCN: buffer_load_dword [[LOAD4:v[0-9]+]], off, s[0:3], s34 offset:16
 ; GCN: buffer_load_dword [[LOAD5:v[0-9]+]], off, s[0:3], s34 offset:20
 ; GCN: buffer_load_dword [[LOAD6:v[0-9]+]], off, s[0:3], s34 offset:24
 ; GCN: buffer_load_dword [[LOAD7:v[0-9]+]], off, s[0:3], s34 offset:28
-; GCN: buffer_load_dword [[LOAD4:v[0-9]+]], off, s[0:3], s34 offset:16
 
 ; GCN: s_waitcnt vmcnt(0)
 ; GCN-DAG: buffer_store_dword [[LOAD4]], off, s[0:3], s32 offset:16
 ; GCN-DAG: buffer_store_dword [[LOAD5]], off, s[0:3], s32 offset:20
 ; GCN-DAG: buffer_store_dword [[LOAD6]], off, s[0:3], s32 offset:24
 ; GCN-DAG: buffer_store_dword [[LOAD7]], off, s[0:3], s32 offset:28
 
 ; GCN: s_swappc_b64

llvm/test/CodeGen/AMDGPU/call-argument-types.ll

 define amdgpu_kernel void @stack_passed_arg_alignment_v32i32_f64(<32 x i32> %val, double %tmp) #0 {
 entry:
   call void @stack_passed_f64_arg(<32 x i32> %val, double %tmp)
   ret void
 }
 
 ; GCN-LABEL: {{^}}tail_call_byval_align16:
 ; GCN-NOT: s32
-; GCN-NOT: buffer_store_dword v33
-; GCN: buffer_store_dword v32, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
-; GCN-NOT: buffer_store_dword v33
+; GCN: buffer_store_dword v32, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
+; GCN: buffer_store_dword v33, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GCN: buffer_load_dword v32, off, s[0:3], s32 offset:20
+; GCN: buffer_load_dword v33, off, s[0:3], s32 offset:16
 
 ; GCN: s_getpc_b64
 
 ; GCN: buffer_store_dword v32, off, s[0:3], s32 offset:4
-; GCN: buffer_load_dword v32, off, s[0:3], s32 offset:16
-; GCN: buffer_load_dword v32, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GCN: buffer_store_dword v33, off, s[0:3], s32{{$}}
+; GCN: buffer_load_dword v33, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GCN: buffer_load_dword v32, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GCN-NOT: s32
 ; GCN: s_setpc_b64
 define void @tail_call_byval_align16(<32 x i32> %val, double %tmp) #0 {
 entry:
   %alloca = alloca double, align 8, addrspace(5)
   tail call void @byval_align16_f64_arg(<32 x i32> %val, double addrspace(5)* byval align 16 %alloca)
   ret void
 }

llvm/test/CodeGen/AMDGPU/chain-hi-to-lo.ll

 ; GCN-LABEL: vload2_private:
 ; GCN:       ; %bb.0: ; %entry
 ; GCN-NEXT:    s_add_u32 flat_scratch_lo, s6, s9
 ; GCN-NEXT:    s_addc_u32 flat_scratch_hi, s7, 0
 ; GCN-NEXT:    s_load_dwordx4 s[4:7], s[4:5], 0x0
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    v_mov_b32_e32 v2, s4
 ; GCN-NEXT:    v_mov_b32_e32 v3, s5
-; GCN-NEXT:    global_load_dword v4, v[2:3], off
-; GCN-NEXT:    global_load_ushort v2, v[2:3], off offset:4
+; GCN-NEXT:    global_load_ushort v4, v[2:3], off offset:4
+; GCN-NEXT:    global_load_dword v2, v[2:3], off
 ; GCN-NEXT:    v_mov_b32_e32 v0, s6
 ; GCN-NEXT:    v_mov_b32_e32 v1, s7
-; GCN-NEXT:    s_waitcnt vmcnt(1)
-; GCN-NEXT:    buffer_store_short v4, off, s[0:3], s9 offset:4
-; GCN-NEXT:    buffer_store_short_d16_hi v4, off, s[0:3], s9 offset:6
-; GCN-NEXT:    s_waitcnt vmcnt(2)
-; GCN-NEXT:    buffer_store_short v2, off, s[0:3], s9 offset:8
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    buffer_store_short v2, off, s[0:3], s9 offset:4
+; GCN-NEXT:    buffer_store_short_d16_hi v2, off, s[0:3], s9 offset:6
+; GCN-NEXT:    buffer_store_short v4, off, s[0:3], s9 offset:8
 ; GCN-NEXT:    buffer_load_ushort v2, off, s[0:3], s9 offset:4
 ; GCN-NEXT:    buffer_load_ushort v4, off, s[0:3], s9 offset:6
 ; GCN-NEXT:    s_waitcnt vmcnt(1)
 ; GCN-NEXT:    v_and_b32_e32 v2, 0xffff, v2
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_mov_b32_e32 v3, v4
 ; GCN-NEXT:    buffer_load_short_d16_hi v3, off, s[0:3], s9 offset:8
 ; GCN-NEXT:    v_lshl_or_b32 v2, v4, 16, v2

llvm/test/CodeGen/AMDGPU/cvt_f32_ubyte.ll

 ; VI-NEXT:    v_add_u32_e32 v0, vcc, s0, v0
 ; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v2, vcc, 3, v0
 ; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v4, vcc, 2, v0
 ; VI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v6, vcc, 1, v0
 ; VI-NEXT:    v_addc_u32_e32 v7, vcc, 0, v1, vcc
-; VI-NEXT:    flat_load_ubyte v0, v[0:1]
-; VI-NEXT:    flat_load_ubyte v1, v[6:7]
-; VI-NEXT:    flat_load_ubyte v4, v[4:5]
 ; VI-NEXT:    flat_load_ubyte v2, v[2:3]
+; VI-NEXT:    flat_load_ubyte v3, v[4:5]
+; VI-NEXT:    flat_load_ubyte v4, v[6:7]
+; VI-NEXT:    flat_load_ubyte v0, v[0:1]
+; VI-NEXT:    s_waitcnt vmcnt(3) lgkmcnt(3)
+; VI-NEXT:    v_lshlrev_b32_e32 v1, 8, v2
 ; VI-NEXT:    s_waitcnt vmcnt(2) lgkmcnt(2)
-; VI-NEXT:    v_lshlrev_b32_e32 v1, 8, v1
-; VI-NEXT:    v_or_b32_e32 v0, v1, v0
+; VI-NEXT:    v_or_b32_e32 v1, v1, v3
+; VI-NEXT:    s_waitcnt vmcnt(1) lgkmcnt(1)
+; VI-NEXT:    v_lshlrev_b32_e32 v2, 8, v4
+; VI-NEXT:    v_lshlrev_b32_e32 v3, 16, v1
 ; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; VI-NEXT:    v_lshlrev_b32_e32 v2, 8, v2
-; VI-NEXT:    v_or_b32_e32 v2, v2, v4
-; VI-NEXT:    v_lshlrev_b32_e32 v3, 16, v2
+; VI-NEXT:    v_or_b32_e32 v0, v2, v0
+; VI-NEXT:    v_cvt_f32_ubyte0_e32 v2, v1
 ; VI-NEXT:    v_cvt_f32_ubyte1_e32 v1, v0
-; VI-NEXT:    v_cvt_f32_ubyte0_e32 v2, v2
 ; VI-NEXT:    v_cvt_f32_ubyte0_e32 v0, v0
 ; VI-NEXT:    v_cvt_f32_ubyte3_e32 v3, v3
 ; VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[4:7], 0
 ; VI-NEXT:    s_endpgm
   %tid = call i32 @llvm.amdgcn.workitem.id.x()
   %gep = getelementptr <4 x i8>, <4 x i8> addrspace(1)* %in, i32 %tid
   %load = load <4 x i8>, <4 x i8> addrspace(1)* %gep, align 1
   %cvt = uitofp <4 x i8> %load to <4 x float>

 ; VI-NEXT:    s_waitcnt lgkmcnt(0)
 ; VI-NEXT:    v_mov_b32_e32 v1, s1
 ; VI-NEXT:    v_add_u32_e32 v0, vcc, s0, v0
 ; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v2, vcc, 1, v0
 ; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v4, vcc, 3, v0
 ; VI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v1, vcc
+; VI-NEXT:    flat_load_ubyte v9, v[2:3]
 ; VI-NEXT:    flat_load_ubyte v10, v[4:5]
-; VI-NEXT:    flat_load_ubyte v11, v[2:3]
 ; VI-NEXT:    v_add_u32_e32 v2, vcc, 2, v0
 ; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v4, vcc, 5, v0
 ; VI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v6, vcc, 4, v0
 ; VI-NEXT:    v_addc_u32_e32 v7, vcc, 0, v1, vcc
-; VI-NEXT:    v_add_u32_e32 v8, vcc, 6, v0
-; VI-NEXT:    v_addc_u32_e32 v9, vcc, 0, v1, vcc
-; VI-NEXT:    flat_load_ubyte v0, v[0:1]
-; VI-NEXT:    flat_load_ubyte v1, v[8:9]
-; VI-NEXT:    flat_load_ubyte v7, v[6:7]
-; VI-NEXT:    flat_load_ubyte v4, v[4:5]
+; VI-NEXT:    flat_load_ubyte v8, v[0:1]
+; VI-NEXT:    v_add_u32_e32 v0, vcc, 6, v0
+; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
 ; VI-NEXT:    flat_load_ubyte v2, v[2:3]
+; VI-NEXT:    flat_load_ubyte v3, v[4:5]
+; VI-NEXT:    flat_load_ubyte v4, v[6:7]
+; VI-NEXT:    flat_load_ubyte v0, v[0:1]
 ; VI-NEXT:    s_waitcnt vmcnt(6) lgkmcnt(6)
-; VI-NEXT:    v_lshlrev_b32_e32 v5, 8, v10
+; VI-NEXT:    v_lshlrev_b32_e32 v1, 8, v9
 ; VI-NEXT:    s_waitcnt vmcnt(5) lgkmcnt(5)
-; VI-NEXT:    v_lshlrev_b32_e32 v3, 8, v11
-; VI-NEXT:    s_waitcnt vmcnt(4) lgkmcnt(4)
-; VI-NEXT:    v_or_b32_e32 v0, v3, v0
-; VI-NEXT:    s_waitcnt vmcnt(3) lgkmcnt(3)
-; VI-NEXT:    v_cvt_f32_ubyte0_e32 v6, v1
+; VI-NEXT:    v_lshlrev_b32_e32 v5, 8, v10
+; VI-NEXT:    s_waitcnt vmcnt(2) lgkmcnt(2)
+; VI-NEXT:    v_lshlrev_b32_e32 v3, 8, v3
 ; VI-NEXT:    s_waitcnt vmcnt(1) lgkmcnt(1)
-; VI-NEXT:    v_lshlrev_b32_e32 v4, 8, v4
+; VI-NEXT:    v_or_b32_e32 v4, v3, v4
 ; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; VI-NEXT:    v_cvt_f32_ubyte0_e32 v6, v0
+; VI-NEXT:    v_or_b32_e32 v0, v1, v8
 ; VI-NEXT:    v_or_b32_sdwa v1, v5, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
 ; VI-NEXT:    v_or_b32_e32 v0, v1, v0
-; VI-NEXT:    v_or_b32_e32 v4, v4, v7
 ; VI-NEXT:    v_and_b32_e32 v5, 0xffff0000, v0
 ; VI-NEXT:    v_or_b32_e32 v4, v4, v5
 ; VI-NEXT:    v_cvt_f32_ubyte1_e32 v5, v4
 ; VI-NEXT:    v_cvt_f32_ubyte3_e32 v3, v0
 ; VI-NEXT:    v_cvt_f32_ubyte2_e32 v2, v0
 ; VI-NEXT:    v_cvt_f32_ubyte1_e32 v1, v0
 ; VI-NEXT:    v_cvt_f32_ubyte0_e32 v0, v0
 ; VI-NEXT:    v_cvt_f32_ubyte0_e32 v4, v4

 ; VI-NEXT:    s_waitcnt lgkmcnt(0)
 ; VI-NEXT:    v_mov_b32_e32 v1, s1
 ; VI-NEXT:    v_add_u32_e32 v0, vcc, s0, v0
 ; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v2, vcc, 1, v0
 ; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v1, vcc
 ; VI-NEXT:    v_add_u32_e32 v4, vcc, 3, v0
 ; VI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v1, vcc
-; VI-NEXT:    v_add_u32_e32 v6, vcc, 2, v0
-; VI-NEXT:    v_addc_u32_e32 v7, vcc, 0, v1, vcc
-; VI-NEXT:    flat_load_ubyte v0, v[0:1]
-; VI-NEXT:    flat_load_ubyte v1, v[6:7]
-; VI-NEXT:    flat_load_ubyte v4, v[4:5]
+; VI-NEXT:    flat_load_ubyte v6, v[0:1]
+; VI-NEXT:    v_add_u32_e32 v0, vcc, 2, v0
+; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
 ; VI-NEXT:    flat_load_ubyte v2, v[2:3]
+; VI-NEXT:    flat_load_ubyte v3, v[4:5]
+; VI-NEXT:    flat_load_ubyte v0, v[0:1]
+; VI-NEXT:    s_waitcnt vmcnt(2) lgkmcnt(2)
+; VI-NEXT:    v_lshlrev_b32_e32 v1, 8, v2
 ; VI-NEXT:    s_waitcnt vmcnt(1) lgkmcnt(1)
-; VI-NEXT:    v_lshlrev_b32_e32 v3, 8, v4
+; VI-NEXT:    v_lshlrev_b32_e32 v2, 8, v3
+; VI-NEXT:    v_or_b32_e32 v1, v1, v6
 ; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; VI-NEXT:    v_lshlrev_b32_e32 v2, 8, v2
-; VI-NEXT:    v_or_b32_e32 v4, v2, v0
-; VI-NEXT:    v_or_b32_sdwa v0, v3, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; VI-NEXT:    v_or_b32_e32 v0, v0, v4
+; VI-NEXT:    v_or_b32_sdwa v0, v2, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
+; VI-NEXT:    v_or_b32_e32 v0, v0, v1
 ; VI-NEXT:    v_cvt_f32_ubyte3_e32 v3, v0
 ; VI-NEXT:    v_cvt_f32_ubyte2_e32 v2, v0
 ; VI-NEXT:    v_cvt_f32_ubyte0_e32 v0, v0
-; VI-NEXT:    v_cvt_f32_ubyte1_e32 v1, v4
+; VI-NEXT:    v_cvt_f32_ubyte1_e32 v1, v1
 ; VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[4:7], 0
 ; VI-NEXT:    s_endpgm
   %tid = call i32 @llvm.amdgcn.workitem.id.x()
   %gep = getelementptr <4 x i8>, <4 x i8> addrspace(1)* %in, i32 %tid
   %load = load <4 x i8>, <4 x i8> addrspace(1)* %gep, align 1
   %ext = zext <4 x i8> %load to <4 x i32>
   %cvt = uitofp <4 x i32> %ext to <4 x float>
   store <4 x float> %cvt, <4 x float> addrspace(1)* %out, align 16

llvm/test/CodeGen/AMDGPU/ds_read2.ll

 
 @bar = addrspace(3) global [4 x i64] undef, align 4
 
 ; GCN-LABEL: @load_misaligned64_constant_offsets
 ; CI-DAG: s_mov_b32 m0
 ; GFX9-NOT: m0
 
 ; GCN-DAG: v_mov_b32_e32 [[PTR:v[0-9]+]], bar@abs32@lo{{$}}
-; GCN: ds_read2_b32 v{{\[[0-9]+:[0-9]+\]}}, [[PTR]] offset1:1
 ; GCN: ds_read2_b32 v{{\[[0-9]+:[0-9]+\]}}, [[PTR]] offset0:2 offset1:3
+; GCN: ds_read2_b32 v{{\[[0-9]+:[0-9]+\]}}, [[PTR]] offset1:1
 define amdgpu_kernel void @load_misaligned64_constant_offsets(i64 addrspace(1)* %out) {
   %val0 = load i64, i64 addrspace(3)* getelementptr inbounds ([4 x i64], [4 x i64] addrspace(3)* @bar, i32 0, i32 0), align 4
   %val1 = load i64, i64 addrspace(3)* getelementptr inbounds ([4 x i64], [4 x i64] addrspace(3)* @bar, i32 0, i32 1), align 4
   %sum = add i64 %val0, %val1
   store i64 %sum, i64 addrspace(1)* %out, align 8
   ret void
 }
 

llvm/test/CodeGen/AMDGPU/frame-index-elimination.ll

 
 declare void @func(<4 x float> addrspace(5)* nocapture) #0
 
 ; undef flag not preserved in eliminateFrameIndex when handling the
 ; stores in the middle block.
 
 ; GCN-LABEL: {{^}}undefined_stack_store_reg:
 ; GCN: s_and_saveexec_b64
-; GCN: buffer_store_dword v{{[0-9]+}}, off, s[0:3], s34 offset:
 ; GCN: buffer_store_dword v0, off, s[0:3], s34 offset:
 ; GCN: buffer_store_dword v0, off, s[0:3], s34 offset:
 ; GCN: buffer_store_dword v0, off, s[0:3], s34 offset:
+; GCN: buffer_store_dword v{{[0-9]+}}, off, s[0:3], s34 offset:
 define void @undefined_stack_store_reg(float %arg, i32 %arg1) #0 {
 bb:
   %tmp = alloca <4 x float>, align 16, addrspace(5)
   %tmp2 = insertelement <4 x float> undef, float %arg, i32 0
   store <4 x float> %tmp2, <4 x float> addrspace(5)* undef
   %tmp3 = icmp eq i32 %arg1, 0
   br i1 %tmp3, label %bb4, label %bb5
 

llvm/test/CodeGen/AMDGPU/insert-subvector-unused-scratch.ll

 ; RUN: llc -mtriple amdgcn-amd-- -mcpu=bonaire -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
 
 ; Before the fix that this test was committed with, this code would leave
 ; an unused stack slot, causing ScratchSize to be non-zero.
 
 ; GCN-LABEL: store_v3i32:
-; GCN:        ds_read_b64
 ; GCN:        ds_read_b32
+; GCN:        ds_read_b64
 ; GCN:        ds_write_b32
 ; GCN:        ds_write_b64
 ; GCN: ScratchSize: 0
 define amdgpu_kernel void @store_v3i32(<3 x i32> addrspace(3)* %out, <3 x i32> %a) nounwind {
   %val = load <3 x i32>, <3 x i32> addrspace(3)* %out
   %val.1 = add <3 x i32> %a, %val
   store <3 x i32> %val.1, <3 x i32> addrspace(3)* %out, align 16
   ret void
 }
 
 ; GCN-LABEL: store_v5i32:
-; GCN:        ds_read2_b64
 ; GCN:        ds_read_b32
+; GCN:        ds_read2_b64
 ; GCN:        ds_write_b32
 ; GCN:        ds_write2_b64
 ; GCN: ScratchSize: 0
 define amdgpu_kernel void @store_v5i32(<5 x i32> addrspace(3)* %out, <5 x i32> %a) nounwind {
   %val = load <5 x i32>, <5 x i32> addrspace(3)* %out
   %val.1 = add <5 x i32> %a, %val
   store <5 x i32> %val.1, <5 x i32> addrspace(3)* %out, align 16
   ret void
 }
 
 

llvm/test/CodeGen/AMDGPU/insert_vector_elt.ll

 ; SI-NEXT:    v_mov_b32_e32 v8, s16
 ; SI-NEXT:    v_mov_b32_e32 v9, s17
 ; SI-NEXT:    v_mov_b32_e32 v10, s18
 ; SI-NEXT:    v_mov_b32_e32 v11, s19
 ; SI-NEXT:    v_mov_b32_e32 v12, s20
 ; SI-NEXT:    v_mov_b32_e32 v13, s21
 ; SI-NEXT:    v_mov_b32_e32 v14, s22
 ; SI-NEXT:    v_mov_b32_e32 v15, s23
-; SI-NEXT:    buffer_store_dwordx4 v[8:11], off, s[0:3], s7 offset:96
 ; SI-NEXT:    buffer_store_dwordx4 v[12:15], off, s[0:3], s7 offset:112
-; SI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], s7 offset:64
+; SI-NEXT:    buffer_store_dwordx4 v[8:11], off, s[0:3], s7 offset:96
 ; SI-NEXT:    buffer_store_dwordx4 v[4:7], off, s[0:3], s7 offset:80
+; SI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], s7 offset:64
 ; SI-NEXT:    v_or_b32_e32 v16, s4, v16
 ; SI-NEXT:    v_mov_b32_e32 v0, 0
 ; SI-NEXT:    v_mov_b32_e32 v1, 0x40200000
 ; SI-NEXT:    buffer_store_dwordx2 v[0:1], v16, s[0:3], s7 offen
 ; SI-NEXT:    buffer_load_dwordx4 v[0:3], off, s[0:3], s7 offset:64
 ; SI-NEXT:    buffer_load_dwordx4 v[4:7], off, s[0:3], s7 offset:80
 ; SI-NEXT:    buffer_load_dwordx4 v[8:11], off, s[0:3], s7 offset:96
 ; SI-NEXT:    buffer_load_dwordx4 v[12:15], off, s[0:3], s7 offset:112

 ; VI-NEXT:    v_mov_b32_e32 v8, s16
 ; VI-NEXT:    v_mov_b32_e32 v9, s17
 ; VI-NEXT:    v_mov_b32_e32 v10, s18
 ; VI-NEXT:    v_mov_b32_e32 v11, s19
 ; VI-NEXT:    v_mov_b32_e32 v12, s20
 ; VI-NEXT:    v_mov_b32_e32 v13, s21
 ; VI-NEXT:    v_mov_b32_e32 v14, s22
 ; VI-NEXT:    v_mov_b32_e32 v15, s23
-; VI-NEXT:    buffer_store_dwordx4 v[8:11], off, s[0:3], s7 offset:96
 ; VI-NEXT:    buffer_store_dwordx4 v[12:15], off, s[0:3], s7 offset:112
-; VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], s7 offset:64
+; VI-NEXT:    buffer_store_dwordx4 v[8:11], off, s[0:3], s7 offset:96
 ; VI-NEXT:    buffer_store_dwordx4 v[4:7], off, s[0:3], s7 offset:80
+; VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], s7 offset:64
 ; VI-NEXT:    v_or_b32_e32 v16, s4, v16
 ; VI-NEXT:    v_mov_b32_e32 v0, 0
 ; VI-NEXT:    v_mov_b32_e32 v1, 0x40200000
 ; VI-NEXT:    buffer_store_dwordx2 v[0:1], v16, s[0:3], s7 offen
 ; VI-NEXT:    buffer_load_dwordx4 v[0:3], off, s[0:3], s7 offset:64
 ; VI-NEXT:    buffer_load_dwordx4 v[4:7], off, s[0:3], s7 offset:80
 ; VI-NEXT:    buffer_load_dwordx4 v[8:11], off, s[0:3], s7 offset:96
 ; VI-NEXT:    buffer_load_dwordx4 v[12:15], off, s[0:3], s7 offset:112

llvm/test/CodeGen/AMDGPU/kernel-args.ll

define amdgpu_kernel void @struct_argument_alignment({i32, i64} %arg0, i8, {i32, i64} %arg1) {
   store volatile i64 %val3, i64 addrspace(1)* null
   ret void
 }
 
 ; No padding between i8 and next struct, but round up at end to 4 byte
 ; multiple.
 ; FUNC-LABEL: {{^}}packed_struct_argument_alignment:
 ; HSA-GFX9: kernarg_segment_byte_size = 28
-; HSA-GFX9: global_load_dword v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, off offset:13
 ; HSA-GFX9: global_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, off offset:17
+; HSA-GFX9: global_load_dword v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, off offset:13
 ; HSA-GFX9: s_load_dword s{{[0-9]+}}, s[4:5], 0x0
 ; HSA-GFX9: s_load_dwordx2 s{{\[[0-9]+:[0-9]+\]}}, s[4:5], 0x4
 define amdgpu_kernel void @packed_struct_argument_alignment(<{i32, i64}> %arg0, i8, <{i32, i64}> %arg1) {
   %val0 = extractvalue <{i32, i64}> %arg0, 0
   %val1 = extractvalue <{i32, i64}> %arg0, 1
   %val2 = extractvalue <{i32, i64}> %arg1, 0
   %val3 = extractvalue <{i32, i64}> %arg1, 1
   store volatile i32 %val0, i32 addrspace(1)* null

llvm/test/CodeGen/AMDGPU/kernel-argument-dag-lowering.ll

define amdgpu_kernel void @struct_argument_alignment({i32, i64} %arg0, i8, {i32, i64} %arg1) {
   store volatile i64 %val3, i64 addrspace(1)* null
   ret void
 }
 
 ; No padding between i8 and next struct, but round up at end to 4 byte
 ; multiple.
 ; FUNC-LABEL: {{^}}packed_struct_argument_alignment:
 ; HSA-VI: kernarg_segment_byte_size = 28
-; HSA-VI: global_load_dword v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, off offset:13
 ; HSA-VI: global_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, off offset:17
+; HSA-VI: global_load_dword v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, off offset:13
 ; HSA-VI: s_load_dword s{{[0-9]+}}, s[4:5], 0x0
 ; HSA-VI: s_load_dwordx2 s{{\[[0-9]+:[0-9]+\]}}, s[4:5], 0x4
 define amdgpu_kernel void @packed_struct_argument_alignment(<{i32, i64}> %arg0, i8, <{i32, i64}> %arg1) {
   %val0 = extractvalue <{i32, i64}> %arg0, 0
   %val1 = extractvalue <{i32, i64}> %arg0, 1
   %val2 = extractvalue <{i32, i64}> %arg1, 0
   %val3 = extractvalue <{i32, i64}> %arg1, 1
   store volatile i32 %val0, i32 addrspace(1)* null

llvm/test/CodeGen/AMDGPU/max.i16.ll

 ; VI-NEXT:    v_mov_b32_e32 v3, s1
 ; VI-NEXT:    v_add_u32_e32 v2, vcc, s0, v6
 ; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
 ; VI-NEXT:    v_add_u32_e32 v4, vcc, 4, v0
 ; VI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v1, vcc
 ; VI-NEXT:    v_mov_b32_e32 v7, s5
 ; VI-NEXT:    v_add_u32_e32 v6, vcc, s4, v6
 ; VI-NEXT:    v_addc_u32_e32 v7, vcc, 0, v7, vcc
-; VI-NEXT:    flat_load_dword v8, v[0:1]
-; VI-NEXT:    flat_load_ushort v9, v[4:5]
+; VI-NEXT:    flat_load_ushort v8, v[4:5]
+; VI-NEXT:    flat_load_dword v9, v[0:1]
 ; VI-NEXT:    v_add_u32_e32 v0, vcc, 4, v2
 ; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v3, vcc
-; VI-NEXT:    flat_load_dword v2, v[2:3]
 ; VI-NEXT:    flat_load_ushort v0, v[0:1]
+; VI-NEXT:    flat_load_dword v1, v[2:3]
 ; VI-NEXT:    v_add_u32_e32 v4, vcc, 4, v6
 ; VI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v7, vcc
 ; VI-NEXT:    s_waitcnt vmcnt(1) lgkmcnt(1)
-; VI-NEXT:    v_max_i16_e32 v1, v8, v2
-; VI-NEXT:    v_max_i16_sdwa v2, v8, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
-; VI-NEXT:    v_or_b32_e32 v1, v1, v2
+; VI-NEXT:    v_max_i16_e32 v0, v8, v0
 ; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; VI-NEXT:    v_max_i16_e32 v0, v9, v0
-; VI-NEXT:    flat_store_dword v[6:7], v1
+; VI-NEXT:    v_max_i16_e32 v2, v9, v1
+; VI-NEXT:    v_max_i16_sdwa v1, v9, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
+; VI-NEXT:    v_or_b32_e32 v1, v2, v1
 ; VI-NEXT:    flat_store_short v[4:5], v0
+; VI-NEXT:    flat_store_dword v[6:7], v1
 ; VI-NEXT:    s_endpgm
 ;
 ; GFX9-LABEL: v_test_imax_sge_v3i16:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[0:1], 0x24
 ; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x34
 ; GFX9-NEXT:    v_lshlrev_b32_e32 v4, 3, v0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)

llvm/test/CodeGen/AMDGPU/promote-constOffset-to-imm.ll

 ; GFX8:    flat_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}]
 ;
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
-; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
-; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:2048
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:3072
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
+; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
+; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:2048
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:3072
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
 ;
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
-; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:-2048
-; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
-; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:-2048
+; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
+; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:-2048
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
+; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
 entry:
    %call = tail call i64 @_Z13get_global_idj(i32 0)
    %conv = and i64 %call, 255
    %id = shl i64 %call, 7
    %idx.ext11 = and i64 %id, 4294934528
    %add.ptr12 = getelementptr inbounds i8, i8 addrspace(1)* %buffer, i64 %idx.ext11
    %addr = bitcast i8 addrspace(1)* %add.ptr12 to i32 addrspace(1)*
 

 
 define amdgpu_kernel void @Offset64(i8 addrspace(1)*  %buffer) {
 ; GCN-LABEL: Offset64:
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ;
-; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:2048
-; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:-4096
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
+; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:2048
+; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ;
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:-2048
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 entry:
   %call = tail call i64 @_Z13get_global_idj(i32 0)
   %conv = and i64 %call, 255

 ;
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:2048
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:3072
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX9:    global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
 ;
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:-2048
-; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
 ; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off{{$}}
+; GFX10:   global_load_dword {{v[0-9]+}}, v[{{[0-9]+:[0-9]+}}], off offset:1024
 entry:
   %call = tail call i64 @_Z13get_global_idj(i32 0)
   %conv = and i64 %call, 255
   %a0 = shl i64 %call, 7
   %idx.ext11 = and i64 %a0, 4294934528
   %add.ptr12 = getelementptr inbounds i8, i8 addrspace(1)* %buffer, i64 %idx.ext11
   %saddr = bitcast i8 addrspace(1)* %add.ptr12 to i32 addrspace(1)*
 

 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ;
-; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:2048
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:2048
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
-; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:2048
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
+; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
+; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:2048
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:2048
 ;
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off{{$}}
entry:
   ret void
 }
 
 define hidden amdgpu_kernel void @negativeoffset(i8 addrspace(1)* nocapture %buffer) {
 ; GCN-LABEL: negativeoffset:
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ; GFX8:    flat_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}]
 ;
-; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off
 ; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:2048
+; GFX9:    global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off
 ;
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off offset:-2048
 ; GFX10:   global_load_dwordx2 v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], off
 entry:
   %call = tail call i64 @_Z13get_global_idj(i32 0) #2
   %conv = and i64 %call, 255
   %0 = shl i64 %call, 7
   %idx.ext11 = and i64 %0, 4294934528

llvm/test/CodeGen/AMDGPU/shift-i128.ll

 ; GCN-NEXT:    v_mov_b32_e32 v6, s8
 ; GCN-NEXT:    v_mov_b32_e32 v10, s2
 ; GCN-NEXT:    v_cndmask_b32_e32 v6, v6, v10, vcc
 ; GCN-NEXT:    v_mov_b32_e32 v10, s14
 ; GCN-NEXT:    v_cndmask_b32_e64 v6, v6, v10, s[0:1]
 ; GCN-NEXT:    v_mov_b32_e32 v10, 16
 ; GCN-NEXT:    v_cndmask_b32_e32 v4, 0, v4, vcc
 ; GCN-NEXT:    v_mov_b32_e32 v11, 0
-; GCN-NEXT:    flat_store_dwordx4 v[8:9], v[0:3]
 ; GCN-NEXT:    flat_store_dwordx4 v[10:11], v[4:7]
+; GCN-NEXT:    flat_store_dwordx4 v[8:9], v[0:3]
 ; GCN-NEXT:    s_endpgm
   %shift = shl <2 x i128> %lhs, %rhs
   store <2 x i128> %shift, <2 x i128> addrspace(1)* null
   ret void
 }
 
 define amdgpu_kernel void @s_lshr_v2i128_ss(<2 x i128> %lhs, <2 x i128> %rhs) {
 ; GCN-LABEL: s_lshr_v2i128_ss:

 ; GCN-NEXT:    v_cndmask_b32_e64 v5, v4, v5, s[0:1]
 ; GCN-NEXT:    v_mov_b32_e32 v4, s8
 ; GCN-NEXT:    v_mov_b32_e32 v10, s2
 ; GCN-NEXT:    v_cndmask_b32_e32 v4, v4, v10, vcc
 ; GCN-NEXT:    v_mov_b32_e32 v10, s12
 ; GCN-NEXT:    v_cndmask_b32_e64 v4, v4, v10, s[0:1]
 ; GCN-NEXT:    v_mov_b32_e32 v10, 16
 ; GCN-NEXT:    v_mov_b32_e32 v11, 0
-; GCN-NEXT:    flat_store_dwordx4 v[8:9], v[0:3]
 ; GCN-NEXT:    flat_store_dwordx4 v[10:11], v[4:7]
+; GCN-NEXT:    flat_store_dwordx4 v[8:9], v[0:3]
 ; GCN-NEXT:    s_endpgm
   %shift = lshr <2 x i128> %lhs, %rhs
   store <2 x i128> %shift, <2 x i128> addrspace(1)* null
   ret void
 }
 
 define amdgpu_kernel void @s_ashr_v2i128_ss(<2 x i128> %lhs, <2 x i128> %rhs) {
 ; GCN-LABEL: s_ashr_v2i128_ss:

 ; GCN-NEXT:    v_cndmask_b32_e64 v5, v4, v5, s[0:1]
 ; GCN-NEXT:    v_mov_b32_e32 v4, s8
 ; GCN-NEXT:    v_mov_b32_e32 v10, s2
 ; GCN-NEXT:    v_cndmask_b32_e32 v4, v4, v10, vcc
 ; GCN-NEXT:    v_mov_b32_e32 v10, s12
 ; GCN-NEXT:    v_cndmask_b32_e64 v4, v4, v10, s[0:1]
 ; GCN-NEXT:    v_mov_b32_e32 v10, 16
 ; GCN-NEXT:    v_mov_b32_e32 v11, 0
-; GCN-NEXT:    flat_store_dwordx4 v[8:9], v[0:3]
 ; GCN-NEXT:    flat_store_dwordx4 v[10:11], v[4:7]
+; GCN-NEXT:    flat_store_dwordx4 v[8:9], v[0:3]
 ; GCN-NEXT:    s_endpgm
   %shift = ashr <2 x i128> %lhs, %rhs
   store <2 x i128> %shift, <2 x i128> addrspace(1)* null
   ret void
 }
 

llvm/test/CodeGen/AMDGPU/vgpr-descriptor-waterfall-loop-idom-update.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -march=amdgcn -mcpu=gfx1010 | FileCheck %s --check-prefix=GCN
 
 define void @vgpr_descriptor_waterfall_loop_idom_update(<4 x i32>* %arg) {
 ; GCN-LABEL: vgpr_descriptor_waterfall_loop_idom_update:
 ; GCN:       ; %bb.0: ; %entry
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-NEXT:    ; implicit-def: $vcc_hi
 ; GCN-NEXT:  BB0_1: ; %bb0
 ; GCN-NEXT:    ; =>This Loop Header: Depth=1
 ; GCN-NEXT:    ; Child Loop BB0_2 Depth 2
-; GCN-NEXT:    v_add_co_u32_e64 v4, vcc_lo, v0, 8
+; GCN-NEXT:    v_add_co_u32_e64 v2, vcc_lo, v0, 8
 ; GCN-NEXT:    s_mov_b32 s5, exec_lo
-; GCN-NEXT:    v_add_co_ci_u32_e32 v5, vcc_lo, 0, v1, vcc_lo
-; GCN-NEXT:    flat_load_dwordx2 v[2:3], v[0:1]
-; GCN-NEXT:    flat_load_dwordx2 v[4:5], v[4:5]
+; GCN-NEXT:    v_add_co_ci_u32_e32 v3, vcc_lo, 0, v1, vcc_lo
+; GCN-NEXT:    flat_load_dwordx2 v[2:3], v[2:3]
+; GCN-NEXT:    flat_load_dwordx2 v[4:5], v[0:1]
 ; GCN-NEXT:  BB0_2: ; Parent Loop BB0_1 Depth=1
 ; GCN-NEXT:    ; => This Inner Loop Header: Depth=2
-; GCN-NEXT:    s_waitcnt vmcnt(1) lgkmcnt(1)
-; GCN-NEXT:    v_readfirstlane_b32 s8, v2
-; GCN-NEXT:    v_readfirstlane_b32 s9, v3
 ; GCN-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GCN-NEXT:    v_readfirstlane_b32 s10, v4
-; GCN-NEXT:    v_readfirstlane_b32 s11, v5
-; GCN-NEXT:    v_cmp_eq_u64_e32 vcc_lo, s[8:9], v[2:3]
-; GCN-NEXT:    v_cmp_eq_u64_e64 s4, s[10:11], v[4:5]
+; GCN-NEXT:    v_readfirstlane_b32 s8, v4
+; GCN-NEXT:    v_readfirstlane_b32 s9, v5
+; GCN-NEXT:    v_readfirstlane_b32 s10, v2
+; GCN-NEXT:    v_readfirstlane_b32 s11, v3
+; GCN-NEXT:    v_cmp_eq_u64_e32 vcc_lo, s[8:9], v[4:5]
+; GCN-NEXT:    v_cmp_eq_u64_e64 s4, s[10:11], v[2:3]
 ; GCN-NEXT:    s_and_b32 s4, vcc_lo, s4
 ; GCN-NEXT:    s_and_saveexec_b32 s4, s4
 ; GCN-NEXT:    s_nop 0
 ; GCN-NEXT:    buffer_store_dword v0, v0, s[8:11], 0 offen
 ; GCN-NEXT:    v_nop
 ; GCN-NEXT:    s_xor_b32 exec_lo, exec_lo, s4
 ; GCN-NEXT:    s_cbranch_execnz BB0_2
 ; GCN-NEXT:  ; %bb.3: ; in Loop: Header=BB0_1 Depth=1