File: callbr-asm-branch-folding.ll

package info (click to toggle)
llvm-toolchain-14 1%3A14.0.6-12
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 1,496,180 kB
  • sloc: cpp: 5,593,972; ansic: 986,872; asm: 585,869; python: 184,223; objc: 72,530; lisp: 31,119; f90: 27,793; javascript: 9,780; pascal: 9,762; sh: 9,482; perl: 7,468; ml: 5,432; awk: 3,523; makefile: 2,538; xml: 953; cs: 573; fortran: 567
file content (148 lines) | stat: -rw-r--r-- 4,884 bytes parent folder | download | duplicates (3)
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
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu | FileCheck %s

; This test hung in the BranchFolding pass during asm-goto bring up

@e = dso_local global i32 0
@j = dso_local global i32 0

define dso_local void @n(i32* %o, i32 %p, i32 %u) nounwind {
; CHECK-LABEL: n:
; CHECK:       # %bb.0: # %entry
; CHECK-NEXT:    pushq %rbp
; CHECK-NEXT:    pushq %r15
; CHECK-NEXT:    pushq %r14
; CHECK-NEXT:    pushq %r13
; CHECK-NEXT:    pushq %r12
; CHECK-NEXT:    pushq %rbx
; CHECK-NEXT:    pushq %rax
; CHECK-NEXT:    movl %edx, %ebx
; CHECK-NEXT:    movl %esi, %r12d
; CHECK-NEXT:    movq %rdi, %r15
; CHECK-NEXT:    callq c
; CHECK-NEXT:    movl %eax, %r13d
; CHECK-NEXT:    movq %r15, %rdi
; CHECK-NEXT:    callq l
; CHECK-NEXT:    testl %eax, %eax
; CHECK-NEXT:    jne .LBB0_9
; CHECK-NEXT:  # %bb.1: # %if.end
; CHECK-NEXT:    movl %ebx, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill
; CHECK-NEXT:    cmpl $0, e(%rip)
; CHECK-NEXT:    # implicit-def: $ebx
; CHECK-NEXT:    # implicit-def: $r14d
; CHECK-NEXT:    je .LBB0_4
; CHECK-NEXT:  # %bb.2: # %if.then4
; CHECK-NEXT:    movslq %r12d, %rdi
; CHECK-NEXT:    callq m
; CHECK-NEXT:    # implicit-def: $ebx
; CHECK-NEXT:    # implicit-def: $ebp
; CHECK-NEXT:  .LBB0_3: # %r
; CHECK-NEXT:    callq c
; CHECK-NEXT:    movl %ebp, %r14d
; CHECK-NEXT:  .LBB0_4: # %if.end8
; CHECK-NEXT:    movl %ebx, %edi
; CHECK-NEXT:    callq i
; CHECK-NEXT:    movl %eax, %ebp
; CHECK-NEXT:    orl %r14d, %ebp
; CHECK-NEXT:    andl $4, %ebx
; CHECK-NEXT:    testl %r13d, %r13d
; CHECK-NEXT:    jne .LBB0_3
; CHECK-NEXT:  # %bb.5: # %if.end12
; CHECK-NEXT:    testl %ebp, %ebp
; CHECK-NEXT:    je .LBB0_8
; CHECK-NEXT:  # %bb.6: # %if.then14
; CHECK-NEXT:    movl {{[-0-9]+}}(%r{{[sb]}}p), %eax # 4-byte Reload
; CHECK-NEXT:    #APP
; CHECK-NEXT:    #NO_APP
; CHECK-NEXT:    jmp .LBB0_9
; CHECK-NEXT:  .Ltmp0: # Block address taken
; CHECK-NEXT:  # %bb.7: # %if.then20.critedge
; CHECK-NEXT:    movl j(%rip), %edi
; CHECK-NEXT:    movslq %eax, %rcx
; CHECK-NEXT:    movl $1, %esi
; CHECK-NEXT:    movq %r15, %rdx
; CHECK-NEXT:    addq $8, %rsp
; CHECK-NEXT:    popq %rbx
; CHECK-NEXT:    popq %r12
; CHECK-NEXT:    popq %r13
; CHECK-NEXT:    popq %r14
; CHECK-NEXT:    popq %r15
; CHECK-NEXT:    popq %rbp
; CHECK-NEXT:    jmp k # TAILCALL
; CHECK-NEXT:  .LBB0_8: # %if.else
; CHECK-NEXT:    incq 0
; CHECK-NEXT:  .LBB0_9: # %cleanup
; CHECK-NEXT:    addq $8, %rsp
; CHECK-NEXT:    popq %rbx
; CHECK-NEXT:    popq %r12
; CHECK-NEXT:    popq %r13
; CHECK-NEXT:    popq %r14
; CHECK-NEXT:    popq %r15
; CHECK-NEXT:    popq %rbp
; CHECK-NEXT:    retq
entry:
  %call = tail call i32 @c()
  %call1 = tail call i32 @l(i32* %o)
  %tobool = icmp eq i32 %call1, 0
  br i1 %tobool, label %if.end, label %cleanup

if.end:                                           ; preds = %entry
  %0 = load i32, i32* @e
  %tobool3 = icmp eq i32 %0, 0
  br i1 %tobool3, label %if.end8, label %if.then4, !prof !0

if.then4:                                         ; preds = %if.end
  %conv5 = sext i32 %p to i64
  %call6 = tail call i32 @m(i64 %conv5)
  br label %r

r:                                                ; preds = %if.end8, %if.then4
  %flags.0 = phi i32 [ undef, %if.then4 ], [ %and, %if.end8 ]
  %major.0 = phi i32 [ undef, %if.then4 ], [ %or, %if.end8 ]
  %call7 = tail call i32 @c()
  br label %if.end8

if.end8:                                          ; preds = %r, %if.end
  %flags.1 = phi i32 [ %flags.0, %r ], [ undef, %if.end ]
  %major.1 = phi i32 [ %major.0, %r ], [ undef, %if.end ]
  %call9 = tail call i32 @i(i32 %flags.1)
  %or = or i32 %call9, %major.1
  %and = and i32 %flags.1, 4
  %tobool10 = icmp eq i32 %call, 0
  br i1 %tobool10, label %if.end12, label %r

if.end12:                                         ; preds = %if.end8
  %tobool13 = icmp eq i32 %or, 0
  br i1 %tobool13, label %if.else, label %if.then14

if.then14:                                        ; preds = %if.end12
  callbr void asm sideeffect "", "i,~{dirflag},~{fpsr},~{flags}"(i8* blockaddress(@n, %if.then20.critedge))
          to label %cleanup [label %if.then20.critedge]

if.then20.critedge:                               ; preds = %if.then14
  %1 = load i32, i32* @j
  %conv21 = sext i32 %u to i64
  %call22 = tail call i32 @k(i32 %1, i64 1, i32* %o, i64 %conv21)
  br label %cleanup

if.else:                                          ; preds = %if.end12
  %2 = load i64, i64* null
  %inc = add i64 %2, 1
  store i64 %inc, i64* null
  br label %cleanup

cleanup:                                          ; preds = %if.else, %if.then20.critedge, %if.then14, %entry
  ret void
}

declare dso_local i32 @c()

declare dso_local i32 @l(i32*)

declare dso_local i32 @m(i64)

declare dso_local i32 @i(i32)

declare dso_local i32 @k(i32, i64, i32*, i64)

!0 = !{!"branch_weights", i32 2000, i32 1}