optimize-dmbs-v7.ll
1.6 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
; RUN: llc -O1 < %s -mtriple=armv7 -mattr=+db | FileCheck %s
@x1 = global i32 0, align 4
@x2 = global i32 0, align 4
define void @test() {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%i.013 = phi i32 [ 1, %entry ], [ %inc6, %for.body ]
store atomic i32 %i.013, i32* @x1 seq_cst, align 4
store atomic i32 %i.013, i32* @x1 seq_cst, align 4
store atomic i32 %i.013, i32* @x2 seq_cst, align 4
%inc6 = add nsw i32 %i.013, 1
%exitcond = icmp eq i32 %inc6, 2
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
; The for.body contains 3 seq_cst stores.
; Hence it should have 3 dmb;str;dmb sequences with the middle dmbs collapsed
; CHECK: %for.body
; CHECK-NOT: str
; CHECK: dmb
; CHECK-NOT: dmb
; CHECK: str
; CHECK-NOT: str
; CHECK: dmb
; CHECK-NOT: dmb
; CHECK: str
; CHECK-NOT: str
; CHECK: dmb
; CHECK-NOT: dmb
; CHECK: str
; CHECK-NOT: str
; CHECK: dmb
; CHECK-NOT: dmb
; CHECK-NOT: str
; CHECK: %for.end
}
define void @test2() {
call void @llvm.arm.dmb(i32 11)
tail call void @test()
call void @llvm.arm.dmb(i32 11)
ret void
; the call should prevent the two dmbs from collapsing
; CHECK: test2:
; CHECK: dmb
; CHECK-NEXT: bl
; CHECK-NEXT: dmb
}
define void @test3() {
call void @llvm.arm.dmb(i32 11)
call void @llvm.arm.dsb(i32 9)
call void @llvm.arm.dmb(i32 11)
ret void
; the call should prevent the two dmbs from collapsing
; CHECK: test3:
; CHECK: dmb
; CHECK-NEXT: dsb
; CHECK-NEXT: dmb
}
declare void @llvm.arm.dmb(i32)
declare void @llvm.arm.dsb(i32)