; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -S -passes=loop-rotate -o - -verify-loop-info -verify-dom-info | FileCheck %s
@d = external global i64, align 8
@f = external global i32, align 4
@g = external global i32, align 4
@i = external global i32, align 4
@h = external global i32, align 4
define i32 @o() #0 {
; CHECK-LABEL: @o(
; CHECK-NEXT: bb:
; CHECK-NEXT: [[I:%.*]] = alloca [1 x i32], align 4
; CHECK-NEXT: [[I1:%.*]] = load ptr, ptr @d, align 8
; CHECK-NEXT: [[I33:%.*]] = load i32, ptr @f, align 4
; CHECK-NEXT: [[I44:%.*]] = icmp eq i32 [[I33]], 0
; CHECK-NEXT: br i1 [[I44]], label [[BB15:%.*]], label [[BB5_LR_PH:%.*]]
; CHECK: bb5.lr.ph:
; CHECK-NEXT: br label [[BB5:%.*]]
; CHECK: bb5:
; CHECK-NEXT: [[I35:%.*]] = phi i32 [ [[I33]], [[BB5_LR_PH]] ], [ [[I3:%.*]], [[M_EXIT:%.*]] ]
; CHECK-NEXT: [[I6:%.*]] = icmp ult i32 [[I35]], 4
; CHECK-NEXT: [[I7:%.*]] = zext i1 [[I6]] to i32
; CHECK-NEXT: store i32 [[I7]], ptr @g, align 4
; CHECK-NEXT: [[I9:%.*]] = call i32 @n(ptr nonnull [[I]], ptr [[I1]])
; CHECK-NEXT: [[I10:%.*]] = icmp eq i32 [[I9]], 0
; CHECK-NEXT: br i1 [[I10]], label [[THREAD_PRE_SPLIT:%.*]], label [[BB5_BB15_CRIT_EDGE:%.*]]
; CHECK: thread-pre-split:
; CHECK-NEXT: [[DOTPR:%.*]] = load i32, ptr @i, align 4
; CHECK-NEXT: [[I12:%.*]] = icmp eq i32 [[DOTPR]], 0
; CHECK-NEXT: br i1 [[I12]], label [[M_EXIT]], label [[BB13_LR_PH:%.*]]
; CHECK: bb13.lr.ph:
; CHECK-NEXT: br label [[BB13:%.*]]
; CHECK: bb13:
; CHECK-NEXT: [[DOT11:%.*]] = phi i32 [ undef, [[BB13_LR_PH]] ], [ [[I14:%.*]], [[J_EXIT_I:%.*]] ]
; CHECK-NEXT: callbr void asm sideeffect "", "!i,~{dirflag},~{fpsr},~{flags}"() #[[ATTR1:[0-9]+]]
; CHECK-NEXT: to label [[J_EXIT_I]] [label %bb13.m.exit_crit_edge]
; CHECK: j.exit.i:
; CHECK-NEXT: [[I14]] = tail call i32 asm "", "={ax},~{dirflag},~{fpsr},~{flags}"() #[[ATTR2:[0-9]+]]
; CHECK-NEXT: br i1 [[I12]], label [[BB11_M_EXIT_CRIT_EDGE:%.*]], label [[BB13]]
; CHECK: bb13.m.exit_crit_edge:
; CHECK-NEXT: [[SPLIT:%.*]] = phi i32 [ [[DOT11]], [[BB13]] ]
; CHECK-NEXT: br label [[M_EXIT]]
; CHECK: bb11.m.exit_crit_edge:
; CHECK-NEXT: [[SPLIT2:%.*]] = phi i32 [ [[I14]], [[J_EXIT_I]] ]
; CHECK-NEXT: br label [[M_EXIT]]
; CHECK: m.exit:
; CHECK-NEXT: [[DOT1_LCSSA:%.*]] = phi i32 [ [[SPLIT]], [[BB13_M_EXIT_CRIT_EDGE:%.*]] ], [ [[SPLIT2]], [[BB11_M_EXIT_CRIT_EDGE]] ], [ undef, [[THREAD_PRE_SPLIT]] ]
; CHECK-NEXT: store i32 [[DOT1_LCSSA]], ptr @h, align 4
; CHECK-NEXT: [[I3]] = load i32, ptr @f, align 4
; CHECK-NEXT: [[I4:%.*]] = icmp eq i32 [[I3]], 0
; CHECK-NEXT: br i1 [[I4]], label [[BB2_BB15_CRIT_EDGE:%.*]], label [[BB5]]
; CHECK: bb5.bb15_crit_edge:
; CHECK-NEXT: br label [[BB15]]
; CHECK: bb2.bb15_crit_edge:
; CHECK-NEXT: br label [[BB15]]
; CHECK: bb15:
; CHECK-NEXT: ret i32 undef
;
bb:
%i = alloca [1 x i32], align 4
%i1 = load ptr, ptr @d, align 8
br label %bb2
bb2: ; preds = %m.exit, %bb
%i3 = load i32, ptr @f, align 4
%i4 = icmp eq i32 %i3, 0
br i1 %i4, label %bb15, label %bb5
bb5: ; preds = %bb2
%i6 = icmp ult i32 %i3, 4
%i7 = zext i1 %i6 to i32
store i32 %i7, ptr @g, align 4
%i9 = call i32 @n(ptr nonnull %i, ptr %i1)
%i10 = icmp eq i32 %i9, 0
br i1 %i10, label %thread-pre-split, label %bb15
thread-pre-split: ; preds = %bb5
%.pr = load i32, ptr @i, align 4
br label %bb11
bb11: ; preds = %j.exit.i, %thread-pre-split
%.1 = phi i32 [ %i14, %j.exit.i ], [ undef, %thread-pre-split ]
%i12 = icmp eq i32 %.pr, 0
br i1 %i12, label %m.exit, label %bb13
bb13: ; preds = %bb11
callbr void asm sideeffect "", "!i,~{dirflag},~{fpsr},~{flags}"() #1
to label %j.exit.i [label %m.exit]
j.exit.i: ; preds = %bb13
%i14 = tail call i32 asm "", "={ax},~{dirflag},~{fpsr},~{flags}"() #2
br label %bb11
m.exit: ; preds = %bb13, %bb11
%.1.lcssa = phi i32 [ %.1, %bb13 ], [ %.1, %bb11 ]
store i32 %.1.lcssa, ptr @h, align 4
br label %bb2
bb15: ; preds = %bb5, %bb2
ret i32 undef
}
declare i32 @n(ptr, ptr) #0
attributes #0 = { "use-soft-float"="false" }
attributes #1 = { nounwind }
attributes #2 = { nounwind readnone }