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
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
|
; RUN: llc < %s -asm-verbose=false | FileCheck %s
; RUN: llc < %s -asm-verbose=false -fast-isel | FileCheck %s
target datalayout = "e-m:e-p:32:32-i64:64-n32:64-S128"
target triple = "wasm32-unknown-unknown"
; CHECK-LABEL: alloca32:
; Check that there is an extra local for the stack pointer.
; CHECK: .local i32, i32, i32, i32{{$}}
define void @alloca32() {
; CHECK: i32.const [[L1:.+]]=, __stack_pointer
; CHECK-NEXT: i32.load [[L1]]=, 0([[L1]])
; CHECK-NEXT: i32.const [[L2:.+]]=, 16
; CHECK-NEXT: i32.sub [[SP:.+]]=, [[L1]], [[L2]]
%retval = alloca i32
; CHECK: i32.const $push[[L3:.+]]=, 0
; CHECK: i32.store {{.*}}=, 12([[SP]]), $pop[[L3]]
store i32 0, i32* %retval
; CHECK: i32.const [[L4:.+]]=, 16
; CHECK-NEXT: i32.add [[SP]]=, [[SP]], [[L4]]
; CHECK-NEXT: i32.const [[L5:.+]]=, __stack_pointer
; CHECK-NEXT: i32.store [[SP]]=, 0([[L5]]), [[SP]]
ret void
}
; CHECK-LABEL: alloca3264:
; CHECK: .local i32, i32, i32, i32{{$}}
define void @alloca3264() {
; CHECK: i32.const [[L1:.+]]=, __stack_pointer
; CHECK-NEXT: i32.load [[L1]]=, 0([[L1]])
; CHECK-NEXT: i32.const [[L2:.+]]=, 16
; CHECK-NEXT: i32.sub [[SP:.+]]=, [[L1]], [[L2]]
%r1 = alloca i32
%r2 = alloca double
; CHECK: i32.const $push[[L3:.+]]=, 0
; CHECK: i32.store {{.*}}=, 12([[SP]]), $pop[[L3]]
store i32 0, i32* %r1
; CHECK: i64.const $push[[L4:.+]]=, 0
; CHECK: i64.store {{.*}}=, 0([[SP]]), $pop[[L4]]
store double 0.0, double* %r2
; CHECK: i32.const [[L4:.+]]=, 16
; CHECK-NEXT: i32.add [[SP]]=, [[SP]], [[L4]]
; CHECK-NEXT: i32.const [[L5:.+]]=, __stack_pointer
; CHECK-NEXT: i32.store [[SP]]=, 0([[L5]]), [[SP]]
ret void
}
; CHECK-LABEL: allocarray:
; CHECK: .local i32, i32, i32, i32, i32{{$}}
define void @allocarray() {
; CHECK-NEXT: i32.const [[L1:.+]]=, __stack_pointer
; CHECK-NEXT: i32.load [[L1]]=, 0([[L1]])
; CHECK-NEXT: i32.const [[L2:.+]]=, 32{{$}}
; CHECK-NEXT: i32.sub [[SP:.+]]=, [[L1]], [[L2]]
; CHECK-NEXT: i32.const [[L2]]=, __stack_pointer{{$}}
; CHECK-NEXT: i32.store [[SP]]=, 0([[L2]]), [[SP]]
%r = alloca [5 x i32]
; CHECK-NEXT: i32.const $push[[L4:.+]]=, 12
; CHECK-NEXT: i32.const [[L5:.+]]=, 12
; CHECK-NEXT: i32.add [[L5]]=, [[SP]], [[L5]]
; CHECK-NEXT: i32.add $push[[L6:.+]]=, [[L5]], $pop[[L4]]
; CHECK-NEXT: i32.const $push[[L9:.+]]=, 1{{$}}
; CHECK-NEXT: i32.store $push[[L10:.+]]=, 12([[SP]]), $pop[[L9]]{{$}}
; CHECK-NEXT: i32.store $discard=, 0($pop3), $pop[[L10]]{{$}}
%p = getelementptr [5 x i32], [5 x i32]* %r, i32 0, i32 0
store i32 1, i32* %p
%p2 = getelementptr [5 x i32], [5 x i32]* %r, i32 0, i32 3
store i32 1, i32* %p2
; CHECK-NEXT: i32.const [[L7:.+]]=, 32
; CHECK-NEXT: i32.add [[SP]]=, [[SP]], [[L7]]
; CHECK-NEXT: i32.const [[L8:.+]]=, __stack_pointer
; CHECK-NEXT: i32.store [[SP]]=, 0([[L8]]), [[SP]]
ret void
}
declare void @ext_func(i64* %ptr)
; CHECK-LABEL: non_mem_use
define void @non_mem_use(i8** %addr) {
; CHECK: i32.const [[L2:.+]]=, 48
; CHECK-NEXT: i32.sub [[SP:.+]]=, {{.+}}, [[L2]]
%buf = alloca [27 x i8], align 16
%r = alloca i64
%r2 = alloca i64
; %r is at SP+8
; CHECK: i32.const [[OFF:.+]]=, 8
; CHECK-NEXT: i32.add [[ARG1:.+]]=, [[SP]], [[OFF]]
; CHECK-NEXT: call ext_func@FUNCTION, [[ARG1]]
call void @ext_func(i64* %r)
; %r2 is at SP+0, no add needed
; CHECK-NEXT: call ext_func@FUNCTION, [[SP]]
call void @ext_func(i64* %r2)
; Use as a value, but in a store
; %buf is at SP+16
; CHECK: i32.const [[OFF:.+]]=, 16
; CHECK-NEXT: i32.add [[VAL:.+]]=, [[SP]], [[OFF]]
; CHECK-NEXT: i32.store {{.*}}=, 0($0), [[VAL]]
%gep = getelementptr inbounds [27 x i8], [27 x i8]* %buf, i32 0, i32 0
store i8* %gep, i8** %addr
ret void
}
; CHECK-LABEL: allocarray_inbounds:
; CHECK: .local i32, i32, i32, i32{{$}}
define void @allocarray_inbounds() {
; CHECK: i32.const [[L1:.+]]=, __stack_pointer
; CHECK-NEXT: i32.load [[L1]]=, 0([[L1]])
; CHECK-NEXT: i32.const [[L2:.+]]=, 32
; CHECK-NEXT: i32.sub [[SP:.+]]=, [[L1]], [[L2]]
%r = alloca [5 x i32]
; CHECK: i32.const $push[[L3:.+]]=, 1
; CHECK: i32.store {{.*}}=, 12([[SP]]), $pop[[L3]]
%p = getelementptr inbounds [5 x i32], [5 x i32]* %r, i32 0, i32 0
store i32 1, i32* %p
; This store should have both the GEP and the FI folded into it.
; CHECK-NEXT: i32.store {{.*}}=, 24([[SP]]), $pop
%p2 = getelementptr inbounds [5 x i32], [5 x i32]* %r, i32 0, i32 3
store i32 1, i32* %p2
; CHECK: i32.const [[L7:.+]]=, 32
; CHECK-NEXT: i32.add [[SP]]=, [[SP]], [[L7]]
; CHECK-NEXT: i32.const [[L8:.+]]=, __stack_pointer
; CHECK-NEXT: i32.store [[SP]]=, 0([[L7]]), [[SP]]
ret void
}
; CHECK-LABEL: dynamic_alloca:
define void @dynamic_alloca(i32 %alloc) {
; CHECK: i32.const [[L0:.+]]=, __stack_pointer
; CHECK-NEXT: i32.load [[SP:.+]]=, 0([[L0]])
; CHECK-NEXT: copy_local [[FP:.+]]=, [[SP]]
; Target independent codegen bumps the stack pointer
; FIXME: we need to write the value back to memory
%r = alloca i32, i32 %alloc
; Target-independent codegen also calculates the store addr
store i32 0, i32* %r
; CHECK: i32.const [[L3:.+]]=, __stack_pointer
; CHECK-NEXT: i32.store [[SP]]=, 0([[L3]]), [[FP]]
ret void
}
; CHECK-LABEL: dynamic_static_alloca:
define void @dynamic_static_alloca(i32 %alloc) {
; CHECK: i32.const [[L0:.+]]=, __stack_pointer
; CHECK-NEXT: i32.load [[L0]]=, 0([[L0]])
; CHECK-NEXT: i32.const [[L2:.+]]=, 16
; CHECK-NEXT: i32.sub [[SP:.+]]=, [[L0]], [[L2]]
; CHECK-NEXT: copy_local [[FP:.+]]=, [[SP]]
; CHECK-NEXT: i32.const [[L3:.+]]=, __stack_pointer
; CHECK-NEXT: i32.store {{.*}}=, 0([[L3]]), [[SP]]
%r1 = alloca i32
%r = alloca i32, i32 %alloc
store i32 0, i32* %r
; CHECK: i32.const [[L3:.+]]=, 16
; CHECK: i32.add [[SP]]=, [[FP]], [[L3]]
; CHECK: i32.const [[L4:.+]]=, __stack_pointer
; CHECK-NEXT: i32.store [[SP]]=, 0([[L4]]), [[SP]]
ret void
}
; The use of the alloca in a phi causes a CopyToReg DAG node to be generated,
; which has to have special handling because CopyToReg can't have a FI operand
; CHECK-LABEL: copytoreg_fi:
define void @copytoreg_fi(i1 %cond, i32* %b) {
entry:
; CHECK: i32.const [[L2:.+]]=, 16
; CHECK-NEXT: i32.sub [[SP:.+]]=, {{.+}}, [[L2]]
%addr = alloca i32
; CHECK: i32.const [[OFF:.+]]=, 12
; CHECK-NEXT: i32.add [[ADDR:.+]]=, [[SP]], [[OFF]]
; CHECK-NEXT: copy_local [[COPY:.+]]=, [[ADDR]]
br label %body
body:
%a = phi i32* [%addr, %entry], [%b, %body]
store i32 1, i32* %a
; CHECK: i32.store {{.*}}, 0([[COPY]]),
br i1 %cond, label %body, label %exit
exit:
ret void
}
declare void @use_i8_star(i8*)
declare i8* @llvm.frameaddress(i32)
; Test __builtin_frame_address(0).
; TODO: When the prolog/epilog sequences are optimized, refine these checks to
; be more specific.
; CHECK-LABEL: frameaddress_0:
; CHECK: __stack_pointer
; CHECK: load
; CHECK: call use_i8_star
; CHECK: __stack_pointer
; CHECK: store
define void @frameaddress_0() {
%t = call i8* @llvm.frameaddress(i32 0)
call void @use_i8_star(i8* %t)
ret void
}
; Test __builtin_frame_address(1).
; CHECK-LABEL: frameaddress_1:
; CHECK-NEXT: i32.const $push0=, 0{{$}}
; CHECK-NEXT: call use_i8_star@FUNCTION, $pop0{{$}}
; CHECK-NEXT: return{{$}}
define void @frameaddress_1() {
%t = call i8* @llvm.frameaddress(i32 1)
call void @use_i8_star(i8* %t)
ret void
}
; TODO: test over-aligned alloca
|
