xref: /llvm-project/llvm/test/CodeGen/RISCV/umulo-128-legalisation-lowering.ll (revision 9122c5235ec85ce0c0ad337e862b006e7b349d84)
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc < %s -mtriple=riscv32 -mattr=+m | FileCheck %s --check-prefixes=RISCV32
3
4define { i128, i8 } @muloti_test(i128 %l, i128 %r) #0 {
5; RISCV32-LABEL: muloti_test:
6; RISCV32:       # %bb.0: # %start
7; RISCV32-NEXT:    addi sp, sp, -32
8; RISCV32-NEXT:    sw s0, 28(sp) # 4-byte Folded Spill
9; RISCV32-NEXT:    sw s1, 24(sp) # 4-byte Folded Spill
10; RISCV32-NEXT:    sw s2, 20(sp) # 4-byte Folded Spill
11; RISCV32-NEXT:    sw s3, 16(sp) # 4-byte Folded Spill
12; RISCV32-NEXT:    sw s4, 12(sp) # 4-byte Folded Spill
13; RISCV32-NEXT:    sw s5, 8(sp) # 4-byte Folded Spill
14; RISCV32-NEXT:    lw a4, 0(a1)
15; RISCV32-NEXT:    lw t0, 4(a1)
16; RISCV32-NEXT:    lw a3, 8(a1)
17; RISCV32-NEXT:    lw a1, 12(a1)
18; RISCV32-NEXT:    lw a6, 0(a2)
19; RISCV32-NEXT:    lw a5, 4(a2)
20; RISCV32-NEXT:    lw a7, 8(a2)
21; RISCV32-NEXT:    lw a2, 12(a2)
22; RISCV32-NEXT:    mulhu t1, a4, a6
23; RISCV32-NEXT:    mul t2, t0, a6
24; RISCV32-NEXT:    mulhu t3, t0, a6
25; RISCV32-NEXT:    mul t4, a4, a5
26; RISCV32-NEXT:    mulhu t5, a4, a5
27; RISCV32-NEXT:    mul s2, t0, a5
28; RISCV32-NEXT:    mul t6, a7, a4
29; RISCV32-NEXT:    mul s3, a3, a6
30; RISCV32-NEXT:    mul s0, t0, a7
31; RISCV32-NEXT:    mul s1, a2, a4
32; RISCV32-NEXT:    mul s4, a5, a3
33; RISCV32-NEXT:    add s1, s1, s0
34; RISCV32-NEXT:    mul s0, a1, a6
35; RISCV32-NEXT:    add s4, s0, s4
36; RISCV32-NEXT:    mulhu s5, t0, a5
37; RISCV32-NEXT:    add t1, t2, t1
38; RISCV32-NEXT:    sltu t2, t1, t2
39; RISCV32-NEXT:    add t2, t3, t2
40; RISCV32-NEXT:    mulhu s0, a7, a4
41; RISCV32-NEXT:    add t1, t4, t1
42; RISCV32-NEXT:    sltu t3, t1, t4
43; RISCV32-NEXT:    add t3, t5, t3
44; RISCV32-NEXT:    mulhu t5, a3, a6
45; RISCV32-NEXT:    add t4, s3, t6
46; RISCV32-NEXT:    add s1, s0, s1
47; RISCV32-NEXT:    add t6, t5, s4
48; RISCV32-NEXT:    sltu s3, t4, s3
49; RISCV32-NEXT:    add t3, t2, t3
50; RISCV32-NEXT:    sltu t2, t3, t2
51; RISCV32-NEXT:    add s5, s5, t2
52; RISCV32-NEXT:    add s4, t6, s1
53; RISCV32-NEXT:    add t3, s2, t3
54; RISCV32-NEXT:    add t2, t3, t4
55; RISCV32-NEXT:    sltu s2, t3, s2
56; RISCV32-NEXT:    sltu t4, t2, t3
57; RISCV32-NEXT:    add s2, s5, s2
58; RISCV32-NEXT:    add s3, s4, s3
59; RISCV32-NEXT:    add t3, s2, s3
60; RISCV32-NEXT:    add t3, t3, t4
61; RISCV32-NEXT:    beq t3, s2, .LBB0_2
62; RISCV32-NEXT:  # %bb.1: # %start
63; RISCV32-NEXT:    sltu t4, t3, s2
64; RISCV32-NEXT:  .LBB0_2: # %start
65; RISCV32-NEXT:    sltu s0, s1, s0
66; RISCV32-NEXT:    snez s1, t0
67; RISCV32-NEXT:    snez s2, a2
68; RISCV32-NEXT:    sltu t5, t6, t5
69; RISCV32-NEXT:    mulhu t6, a2, a4
70; RISCV32-NEXT:    mulhu t0, t0, a7
71; RISCV32-NEXT:    or a2, a7, a2
72; RISCV32-NEXT:    snez a7, a5
73; RISCV32-NEXT:    mul a4, a4, a6
74; RISCV32-NEXT:    mulhu a6, a1, a6
75; RISCV32-NEXT:    mulhu a5, a5, a3
76; RISCV32-NEXT:    or a3, a3, a1
77; RISCV32-NEXT:    snez a1, a1
78; RISCV32-NEXT:    and s1, s2, s1
79; RISCV32-NEXT:    snez t6, t6
80; RISCV32-NEXT:    snez t0, t0
81; RISCV32-NEXT:    and a1, a1, a7
82; RISCV32-NEXT:    snez a6, a6
83; RISCV32-NEXT:    snez a5, a5
84; RISCV32-NEXT:    snez a2, a2
85; RISCV32-NEXT:    snez a3, a3
86; RISCV32-NEXT:    or a7, s1, t6
87; RISCV32-NEXT:    or a1, a1, a6
88; RISCV32-NEXT:    and a2, a3, a2
89; RISCV32-NEXT:    or a3, a7, t0
90; RISCV32-NEXT:    or a1, a1, a5
91; RISCV32-NEXT:    or a3, a3, s0
92; RISCV32-NEXT:    or a1, a1, t5
93; RISCV32-NEXT:    or a1, a2, a1
94; RISCV32-NEXT:    or a1, a1, a3
95; RISCV32-NEXT:    or a1, a1, t4
96; RISCV32-NEXT:    andi a1, a1, 1
97; RISCV32-NEXT:    sw a4, 0(a0)
98; RISCV32-NEXT:    sw t1, 4(a0)
99; RISCV32-NEXT:    sw t2, 8(a0)
100; RISCV32-NEXT:    sw t3, 12(a0)
101; RISCV32-NEXT:    sb a1, 16(a0)
102; RISCV32-NEXT:    lw s0, 28(sp) # 4-byte Folded Reload
103; RISCV32-NEXT:    lw s1, 24(sp) # 4-byte Folded Reload
104; RISCV32-NEXT:    lw s2, 20(sp) # 4-byte Folded Reload
105; RISCV32-NEXT:    lw s3, 16(sp) # 4-byte Folded Reload
106; RISCV32-NEXT:    lw s4, 12(sp) # 4-byte Folded Reload
107; RISCV32-NEXT:    lw s5, 8(sp) # 4-byte Folded Reload
108; RISCV32-NEXT:    addi sp, sp, 32
109; RISCV32-NEXT:    ret
110start:
111  %0 = tail call { i128, i1 } @llvm.umul.with.overflow.i128(i128 %l, i128 %r) #2
112  %1 = extractvalue { i128, i1 } %0, 0
113  %2 = extractvalue { i128, i1 } %0, 1
114  %3 = zext i1 %2 to i8
115  %4 = insertvalue { i128, i8 } undef, i128 %1, 0
116  %5 = insertvalue { i128, i8 } %4, i8 %3, 1
117  ret { i128, i8 } %5
118}
119
120; Function Attrs: nounwind readnone speculatable
121declare { i128, i1 } @llvm.umul.with.overflow.i128(i128, i128) #1
122
123attributes #0 = { nounwind readnone }
124attributes #1 = { nounwind readnone speculatable }
125attributes #2 = { nounwind }
126