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
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=riscv32 -mattr=+d -target-abi=ilp32 -verify-machineinstrs < %s \
; RUN: | FileCheck -check-prefix=RV32IFD %s
; RUN: llc -mtriple=riscv32 -mattr=+zdinx -target-abi=ilp32 -verify-machineinstrs < %s \
; RUN: | FileCheck -check-prefix=RV32IZFINXZDINX %s
; Basic correctness checks for calling convention lowering for RV32D. This can
; be somewhat error-prone for soft-float RV32D due to the fact that f64 is legal
; but i64 is not, and there is no instruction to move values directly between
; the GPRs and 64-bit FPRs.
define double @callee_double_inreg(double %a, double %b) nounwind {
; RV32IFD-LABEL: callee_double_inreg:
; RV32IFD: # %bb.0:
; RV32IFD-NEXT: addi sp, sp, -16
; RV32IFD-NEXT: sw a2, 8(sp)
; RV32IFD-NEXT: sw a3, 12(sp)
; RV32IFD-NEXT: fld fa5, 8(sp)
; RV32IFD-NEXT: sw a0, 8(sp)
; RV32IFD-NEXT: sw a1, 12(sp)
; RV32IFD-NEXT: fld fa4, 8(sp)
; RV32IFD-NEXT: fadd.d fa5, fa4, fa5
; RV32IFD-NEXT: fsd fa5, 8(sp)
; RV32IFD-NEXT: lw a0, 8(sp)
; RV32IFD-NEXT: lw a1, 12(sp)
; RV32IFD-NEXT: addi sp, sp, 16
; RV32IFD-NEXT: ret
;
; RV32IZFINXZDINX-LABEL: callee_double_inreg:
; RV32IZFINXZDINX: # %bb.0:
; RV32IZFINXZDINX-NEXT: fadd.d a0, a0, a2
; RV32IZFINXZDINX-NEXT: ret
%1 = fadd double %a, %b
ret double %1
}
; TODO: code quality for loading and then passing f64 constants is poor.
define double @caller_double_inreg() nounwind {
; RV32IFD-LABEL: caller_double_inreg:
; RV32IFD: # %bb.0:
; RV32IFD-NEXT: addi sp, sp, -16
; RV32IFD-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
; RV32IFD-NEXT: lui a0, 262236
; RV32IFD-NEXT: lui a2, 377487
; RV32IFD-NEXT: lui a3, 262364
; RV32IFD-NEXT: addi a1, a0, 655
; RV32IFD-NEXT: addi a0, a2, 1475
; RV32IFD-NEXT: addi a3, a3, 655
; RV32IFD-NEXT: mv a2, a0
; RV32IFD-NEXT: call callee_double_inreg
; RV32IFD-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
; RV32IFD-NEXT: addi sp, sp, 16
; RV32IFD-NEXT: ret
;
; RV32IZFINXZDINX-LABEL: caller_double_inreg:
; RV32IZFINXZDINX: # %bb.0:
; RV32IZFINXZDINX-NEXT: addi sp, sp, -16
; RV32IZFINXZDINX-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
; RV32IZFINXZDINX-NEXT: lui a0, 262236
; RV32IZFINXZDINX-NEXT: lui a2, 377487
; RV32IZFINXZDINX-NEXT: lui a3, 262364
; RV32IZFINXZDINX-NEXT: addi a1, a0, 655
; RV32IZFINXZDINX-NEXT: addi a0, a2, 1475
; RV32IZFINXZDINX-NEXT: addi a3, a3, 655
; RV32IZFINXZDINX-NEXT: mv a2, a0
; RV32IZFINXZDINX-NEXT: call callee_double_inreg
; RV32IZFINXZDINX-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
; RV32IZFINXZDINX-NEXT: addi sp, sp, 16
; RV32IZFINXZDINX-NEXT: ret
%1 = call double @callee_double_inreg(double 2.720000e+00, double 3.720000e+00)
ret double %1
}
define double @callee_double_split_reg_stack(i32 %a, i64 %b, i64 %c, double %d, double %e) nounwind {
; RV32IFD-LABEL: callee_double_split_reg_stack:
; RV32IFD: # %bb.0:
; RV32IFD-NEXT: addi sp, sp, -16
; RV32IFD-NEXT: lw a0, 16(sp)
; RV32IFD-NEXT: sw a7, 8(sp)
; RV32IFD-NEXT: sw a0, 12(sp)
; RV32IFD-NEXT: fld fa5, 8(sp)
; RV32IFD-NEXT: sw a5, 8(sp)
; RV32IFD-NEXT: sw a6, 12(sp)
; RV32IFD-NEXT: fld fa4, 8(sp)
; RV32IFD-NEXT: fadd.d fa5, fa4, fa5
; RV32IFD-NEXT: fsd fa5, 8(sp)
; RV32IFD-NEXT: lw a0, 8(sp)
; RV32IFD-NEXT: lw a1, 12(sp)
; RV32IFD-NEXT: addi sp, sp, 16
; RV32IFD-NEXT: ret
;
; RV32IZFINXZDINX-LABEL: callee_double_split_reg_stack:
; RV32IZFINXZDINX: # %bb.0:
; RV32IZFINXZDINX-NEXT: mv a0, a7
; RV32IZFINXZDINX-NEXT: lw a1, 0(sp)
; RV32IZFINXZDINX-NEXT: mv a3, a6
; RV32IZFINXZDINX-NEXT: mv a2, a5
; RV32IZFINXZDINX-NEXT: fadd.d a0, a2, a0
; RV32IZFINXZDINX-NEXT: ret
%1 = fadd double %d, %e
ret double %1
}
define double @caller_double_split_reg_stack() nounwind {
; RV32IFD-LABEL: caller_double_split_reg_stack:
; RV32IFD: # %bb.0:
; RV32IFD-NEXT: addi sp, sp, -16
; RV32IFD-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
; RV32IFD-NEXT: lui a2, 262510
; RV32IFD-NEXT: lui a3, 262446
; RV32IFD-NEXT: lui a4, 713032
; RV32IFD-NEXT: li a0, 1
; RV32IFD-NEXT: li a1, 2
; RV32IFD-NEXT: addi a2, a2, 327
; RV32IFD-NEXT: addi a6, a3, 327
; RV32IFD-NEXT: addi a5, a4, -1311
; RV32IFD-NEXT: li a3, 3
; RV32IFD-NEXT: sw a2, 0(sp)
; RV32IFD-NEXT: li a2, 0
; RV32IFD-NEXT: li a4, 0
; RV32IFD-NEXT: mv a7, a5
; RV32IFD-NEXT: call callee_double_split_reg_stack
; RV32IFD-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
; RV32IFD-NEXT: addi sp, sp, 16
; RV32IFD-NEXT: ret
;
; RV32IZFINXZDINX-LABEL: caller_double_split_reg_stack:
; RV32IZFINXZDINX: # %bb.0:
; RV32IZFINXZDINX-NEXT: addi sp, sp, -16
; RV32IZFINXZDINX-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
; RV32IZFINXZDINX-NEXT: lui a2, 262510
; RV32IZFINXZDINX-NEXT: lui a3, 262446
; RV32IZFINXZDINX-NEXT: lui a4, 713032
; RV32IZFINXZDINX-NEXT: li a0, 1
; RV32IZFINXZDINX-NEXT: li a1, 2
; RV32IZFINXZDINX-NEXT: addi a2, a2, 327
; RV32IZFINXZDINX-NEXT: addi a6, a3, 327
; RV32IZFINXZDINX-NEXT: addi a5, a4, -1311
; RV32IZFINXZDINX-NEXT: li a3, 3
; RV32IZFINXZDINX-NEXT: sw a2, 0(sp)
; RV32IZFINXZDINX-NEXT: li a2, 0
; RV32IZFINXZDINX-NEXT: li a4, 0
; RV32IZFINXZDINX-NEXT: mv a7, a5
; RV32IZFINXZDINX-NEXT: call callee_double_split_reg_stack
; RV32IZFINXZDINX-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
; RV32IZFINXZDINX-NEXT: addi sp, sp, 16
; RV32IZFINXZDINX-NEXT: ret
%1 = call double @callee_double_split_reg_stack(i32 1, i64 2, i64 3, double 4.72, double 5.72)
ret double %1
}
define double @callee_double_stack(i64 %a, i64 %b, i64 %c, i64 %d, double %e, double %f) nounwind {
; RV32IFD-LABEL: callee_double_stack:
; RV32IFD: # %bb.0:
; RV32IFD-NEXT: addi sp, sp, -16
; RV32IFD-NEXT: fld fa5, 24(sp)
; RV32IFD-NEXT: fld fa4, 16(sp)
; RV32IFD-NEXT: fadd.d fa5, fa4, fa5
; RV32IFD-NEXT: fsd fa5, 8(sp)
; RV32IFD-NEXT: lw a0, 8(sp)
; RV32IFD-NEXT: lw a1, 12(sp)
; RV32IFD-NEXT: addi sp, sp, 16
; RV32IFD-NEXT: ret
;
; RV32IZFINXZDINX-LABEL: callee_double_stack:
; RV32IZFINXZDINX: # %bb.0:
; RV32IZFINXZDINX-NEXT: lw a1, 12(sp)
; RV32IZFINXZDINX-NEXT: lw a3, 4(sp)
; RV32IZFINXZDINX-NEXT: lw a0, 8(sp)
; RV32IZFINXZDINX-NEXT: lw a2, 0(sp)
; RV32IZFINXZDINX-NEXT: fadd.d a0, a2, a0
; RV32IZFINXZDINX-NEXT: ret
%1 = fadd double %e, %f
ret double %1
}
define double @caller_double_stack() nounwind {
; RV32IFD-LABEL: caller_double_stack:
; RV32IFD: # %bb.0:
; RV32IFD-NEXT: addi sp, sp, -32
; RV32IFD-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
; RV32IFD-NEXT: lui a1, 262510
; RV32IFD-NEXT: lui a3, 713032
; RV32IFD-NEXT: lui a5, 262574
; RV32IFD-NEXT: li a0, 1
; RV32IFD-NEXT: li a2, 2
; RV32IFD-NEXT: li a4, 3
; RV32IFD-NEXT: li a6, 4
; RV32IFD-NEXT: addi a1, a1, 327
; RV32IFD-NEXT: addi a3, a3, -1311
; RV32IFD-NEXT: addi a5, a5, 327
; RV32IFD-NEXT: sw a3, 0(sp)
; RV32IFD-NEXT: sw a1, 4(sp)
; RV32IFD-NEXT: sw a3, 8(sp)
; RV32IFD-NEXT: sw a5, 12(sp)
; RV32IFD-NEXT: li a1, 0
; RV32IFD-NEXT: li a3, 0
; RV32IFD-NEXT: li a5, 0
; RV32IFD-NEXT: li a7, 0
; RV32IFD-NEXT: call callee_double_stack
; RV32IFD-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
; RV32IFD-NEXT: addi sp, sp, 32
; RV32IFD-NEXT: ret
;
; RV32IZFINXZDINX-LABEL: caller_double_stack:
; RV32IZFINXZDINX: # %bb.0:
; RV32IZFINXZDINX-NEXT: addi sp, sp, -32
; RV32IZFINXZDINX-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
; RV32IZFINXZDINX-NEXT: lui a1, 262510
; RV32IZFINXZDINX-NEXT: lui a3, 713032
; RV32IZFINXZDINX-NEXT: lui a5, 262574
; RV32IZFINXZDINX-NEXT: li a0, 1
; RV32IZFINXZDINX-NEXT: li a2, 2
; RV32IZFINXZDINX-NEXT: li a4, 3
; RV32IZFINXZDINX-NEXT: li a6, 4
; RV32IZFINXZDINX-NEXT: addi a1, a1, 327
; RV32IZFINXZDINX-NEXT: addi a3, a3, -1311
; RV32IZFINXZDINX-NEXT: addi a5, a5, 327
; RV32IZFINXZDINX-NEXT: sw a3, 0(sp)
; RV32IZFINXZDINX-NEXT: sw a1, 4(sp)
; RV32IZFINXZDINX-NEXT: sw a3, 8(sp)
; RV32IZFINXZDINX-NEXT: sw a5, 12(sp)
; RV32IZFINXZDINX-NEXT: li a1, 0
; RV32IZFINXZDINX-NEXT: li a3, 0
; RV32IZFINXZDINX-NEXT: li a5, 0
; RV32IZFINXZDINX-NEXT: li a7, 0
; RV32IZFINXZDINX-NEXT: call callee_double_stack
; RV32IZFINXZDINX-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
; RV32IZFINXZDINX-NEXT: addi sp, sp, 32
; RV32IZFINXZDINX-NEXT: ret
%1 = call double @callee_double_stack(i64 1, i64 2, i64 3, i64 4, double 5.72, double 6.72)
ret double %1
}
define double @func_return_double_undef() nounwind {
; RV32IFD-LABEL: func_return_double_undef:
; RV32IFD: # %bb.0:
; RV32IFD-NEXT: ret
;
; RV32IZFINXZDINX-LABEL: func_return_double_undef:
; RV32IZFINXZDINX: # %bb.0:
; RV32IZFINXZDINX-NEXT: ret
ret double undef
}
|
