# RUN: llc -mtriple=hexagon -run-pass pipeliner -debug-only=pipeliner %s -o /dev/null 2>&1 | FileCheck %s
# REQUIRES: asserts

# Test that there are no loop-carried dependencies between all memory instructions.

# CHECK:      SU(0):   %8:intregs = PHI %1:intregs, %bb.1, %9:intregs, %bb.2
# CHECK-NEXT:   # preds left
# CHECK-NEXT:   # succs left
# CHECK-NEXT:   # rdefs left
# CHECK-NEXT:   Latency
# CHECK-NEXT:   Depth
# CHECK-NEXT:   Height
# CHECK-NEXT:   Successors:
# CHECK-DAG:      SU(6): Data Latency=0 Reg=%8
# CHECK-DAG:      SU(5): Data Latency=0 Reg=%8
# CHECK-DAG:      SU(3): Data Latency=0 Reg=%8
# CHECK-DAG:      SU(6): Anti Latency=1
# CHECK-NEXT: SU(1):   %10:intregs = PHI %2:intregs, %bb.1, %11:intregs, %bb.2
# CHECK-NEXT:   # preds left
# CHECK-NEXT:   # succs left
# CHECK-NEXT:   # rdefs left
# CHECK-NEXT:   Latency
# CHECK-NEXT:   Depth
# CHECK-NEXT:   Height
# CHECK-NEXT:   Successors:
# CHECK-DAG:      SU(7): Data Latency=0 Reg=%10
# CHECK-DAG:      SU(4): Data Latency=0 Reg=%10
# CHECK-DAG:      SU(2): Data Latency=0 Reg=%10
# CHECK-DAG:      SU(7): Anti Latency=1
# CHECK-NEXT: SU(2):   %12:hvxvr = V6_vL32b_ai %10:intregs, 0 :: (load (s1024) from %ir.iptr.09, !tbaa !4)
# CHECK-NEXT:   # preds left
# CHECK-NEXT:   # succs left
# CHECK-NEXT:   # rdefs left
# CHECK-NEXT:   Latency
# CHECK-NEXT:   Depth
# CHECK-NEXT:   Height
# CHECK-NEXT:   Predecessors:
# CHECK-NEXT:     SU(1): Data Latency=0 Reg=%10
# CHECK-NEXT:   Successors:
# CHECK-NEXT:     SU(3): Data Latency=0 Reg=%12
# CHECK-NEXT: SU(3):   V6_vS32b_ai %8:intregs, 0, %12:hvxvr :: (store (s1024) into %ir.optr.010, !tbaa !4)
# CHECK-NEXT:   # preds left
# CHECK-NEXT:   # succs left
# CHECK-NEXT:   # rdefs left
# CHECK-NEXT:   Latency
# CHECK-NEXT:   Depth
# CHECK-NEXT:   Height
# CHECK-NEXT:   Predecessors:
# CHECK-DAG:      SU(2): Data Latency=0 Reg=%12
# CHECK-DAG:      SU(0): Data Latency=0 Reg=%8
# CHECK-NEXT: SU(4):   %13:hvxvr = V6_vL32b_ai %10:intregs, 128 :: (load (s1024) from %ir.cgep, !tbaa !4)
# CHECK-NEXT:   # preds left
# CHECK-NEXT:   # succs left
# CHECK-NEXT:   # rdefs left
# CHECK-NEXT:   Latency
# CHECK-NEXT:   Depth
# CHECK-NEXT:   Height
# CHECK-NEXT:   Predecessors:
# CHECK-NEXT:     SU(1): Data Latency=0 Reg=%10
# CHECK-NEXT:   Successors:
# CHECK-NEXT:     SU(5): Data Latency=0 Reg=%13
# CHECK-NEXT: SU(5):   V6_vS32b_ai %8:intregs, 128, %13:hvxvr :: (store (s1024) into %ir.cgep3, !tbaa !4)




--- |
  define dso_local void @foo(ptr noundef readonly captures(none) %in, ptr noalias noundef writeonly captures(none) %out, i32 noundef %width) local_unnamed_addr #0 {
  entry:
    %cmp7 = icmp sgt i32 %width, 0
    br i1 %cmp7, label %for.body.preheader, label %for.end
  
  for.body.preheader:                               ; preds = %entry
    %0 = add i32 %width, 128
    br label %for.body
  
  for.body:                                         ; preds = %for.body.preheader, %for.body
    %lsr.iv = phi i32 [ %0, %for.body.preheader ], [ %lsr.iv.next, %for.body ]
    %optr.010 = phi ptr [ %cgep4, %for.body ], [ %out, %for.body.preheader ]
    %iptr.09 = phi ptr [ %cgep5, %for.body ], [ %in, %for.body.preheader ]
    %ald = load <128 x i8>, ptr %iptr.09, align 128, !tbaa !4
    %cst = bitcast <128 x i8> %ald to <32 x i32>
    store <32 x i32> %cst, ptr %optr.010, align 128, !tbaa !4
    %cgep = getelementptr i8, ptr %iptr.09, i32 128
    %ald1 = load <128 x i8>, ptr %cgep, align 128, !tbaa !4
    %cst2 = bitcast <128 x i8> %ald1 to <32 x i32>
    %cgep3 = getelementptr i8, ptr %optr.010, i32 128
    store <32 x i32> %cst2, ptr %cgep3, align 128, !tbaa !4
    %lsr.iv.next = add i32 %lsr.iv, -128
    %cmp = icmp samesign ugt i32 %lsr.iv.next, 128
    %cgep4 = getelementptr i8, ptr %optr.010, i32 256
    %cgep5 = getelementptr i8, ptr %iptr.09, i32 256
    br i1 %cmp, label %for.body, label %for.end
  
  for.end:                                          ; preds = %for.body, %entry
    ret void
  }
  
  attributes #0 = { "target-cpu"="hexagonv60" "target-features"="+hvx-length128b,+hvxv69,+v66,-long-calls" }
  
  !llvm.module.flags = !{!0, !1, !2, !3}
  
  !0 = !{i32 1, !"wchar_size", i32 4}
  !1 = !{i32 8, !"PIC Level", i32 2}
  !2 = !{i32 7, !"PIE Level", i32 2}
  !3 = !{i32 7, !"frame-pointer", i32 2}
  !4 = !{!5, !5, i64 0}
  !5 = !{!"omnipotent char", !6, i64 0}
  !6 = !{!"Simple C/C++ TBAA"}
...
---
name:            foo
tracksRegLiveness: true
body:             |
  bb.0.entry:
    successors: %bb.1(0x50000000), %bb.3(0x30000000)
    liveins: $r0, $r1, $r2
  
    %9:intregs = COPY $r2
    %8:intregs = COPY $r1
    %7:intregs = COPY $r0
    %10:predregs = C2_cmpgti %9, 0
    J2_jumpf %10, %bb.3, implicit-def dead $pc
    J2_jump %bb.1, implicit-def dead $pc
  
  bb.1.for.body.preheader:
    successors: %bb.2(0x80000000)
  
    %0:intregs = A2_addi %9, 128
    %15:intregs = A2_addi %0, -1
    %16:intregs = S2_lsr_i_r %15, 7
    %17:intregs = COPY %16
    J2_loop0r %bb.2, %17, implicit-def $lc0, implicit-def $sa0, implicit-def $usr
  
  bb.2.for.body (machine-block-address-taken):
    successors: %bb.2(0x7c000000), %bb.3(0x04000000)
  
    %2:intregs = PHI %8, %bb.1, %5, %bb.2
    %3:intregs = PHI %7, %bb.1, %6, %bb.2
    %12:hvxvr = V6_vL32b_ai %3, 0 :: (load (s1024) from %ir.iptr.09, !tbaa !4)
    V6_vS32b_ai %2, 0, killed %12 :: (store (s1024) into %ir.optr.010, !tbaa !4)
    %13:hvxvr = V6_vL32b_ai %3, 128 :: (load (s1024) from %ir.cgep, !tbaa !4)
    V6_vS32b_ai %2, 128, killed %13 :: (store (s1024) into %ir.cgep3, !tbaa !4)
    %5:intregs = A2_addi %2, 256
    %6:intregs = A2_addi %3, 256
    ENDLOOP0 %bb.2, implicit-def $pc, implicit-def $lc0, implicit $sa0, implicit $lc0
    J2_jump %bb.3, implicit-def dead $pc
  
  bb.3.for.end:
    PS_jmpret $r31, implicit-def dead $pc
...