; RUN: llc -march=hexagon < %s -pipeliner-experimental-cg=true | FileCheck %s
; Test that when we order instructions in a packet we check for
; order dependences so that the source of an order dependence
; appears before the destination.
; CHECK: loop0(.LBB0_[[LOOP:.]],
; CHECK: .LBB0_[[LOOP]]:
; CHECK: = memw
; CHECK: = memw
; CHECK: memw({{.*}}) =
; CHECK: = memw
; CHECK: = memw
; CHECK: endloop0
@g0 = external hidden unnamed_addr constant [19 x i8], align 1
; Function Attrs: nounwind optsize
declare i32 @f0(ptr nocapture readonly, ...) #0
; Function Attrs: nounwind optsize
declare void @f1(ptr, ptr, ptr nocapture readnone) #0
; Function Attrs: argmemonly nounwind
declare ptr @llvm.hexagon.circ.stw(ptr, i32, i32, i32) #1
; Function Attrs: nounwind optsize
define void @f2(ptr %a0, ptr %a1, ptr %a2) #0 {
b0:
%v0 = alloca i32, align 4
call void @f1(ptr %a2, ptr %a0, ptr %v0) #2
br label %b1
b1: ; preds = %b1, %b0
%v2 = phi i32 [ 0, %b0 ], [ %v13, %b1 ]
%v3 = phi ptr [ %a2, %b0 ], [ %v16, %b1 ]
%v4 = phi i32 [ 0, %b0 ], [ %v14, %b1 ]
%v5 = load i32, ptr %a1, align 4, !tbaa !0
%v6 = add nsw i32 %v2, %v5
%v7 = load i32, ptr %v3, align 4, !tbaa !0
%v8 = tail call ptr @llvm.hexagon.circ.stw(ptr %a1, i32 %v7, i32 150995968, i32 4) #3
%v10 = load i32, ptr %v3, align 4, !tbaa !0
%v11 = add nsw i32 %v6, %v10
%v12 = load i32, ptr %v8, align 4, !tbaa !0
%v13 = add nsw i32 %v11, %v12
%v14 = add nsw i32 %v4, 1
%v15 = icmp eq i32 %v14, 2
%v16 = getelementptr i32, ptr %v3, i32 1
br i1 %v15, label %b2, label %b1
b2: ; preds = %b1
%v17 = tail call i32 (ptr, ...) @f0(ptr @g0, i32 %v13) #4
ret void
}
attributes #0 = { nounwind optsize "target-cpu"="hexagonv55" }
attributes #1 = { argmemonly nounwind }
attributes #2 = { optsize }
attributes #3 = { nounwind }
attributes #4 = { nounwind optsize }
!0 = !{!1, !1, i64 0}
!1 = !{!"int", !2, i64 0}
!2 = !{!"omnipotent char", !3, i64 0}
!3 = !{!"Simple C/C++ TBAA"}