xref: /llvm-project/llvm/test/CodeGen/RISCV/rvv/rv64-spill-vector-csr.ll (revision b6c0f1bfa79a3a32d841ac5ab1f94c3aee3b5d90) 
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc -mtriple=riscv64 -mattr=+v,+d -mattr=+d -O0 < %s \
3; RUN:    | FileCheck --check-prefix=SPILL-O0 %s
4; RUN: llc -mtriple=riscv64 -mattr=+v,+d -mattr=+d -O2 < %s \
5; RUN:    | FileCheck --check-prefix=SPILL-O2 %s
6; RUN: llc -mtriple=riscv64 -mattr=+v,+d -mattr=+d -riscv-v-vector-bits-max=128 -O2 < %s \
7; RUN:    | FileCheck --check-prefix=SPILL-O2-VLEN128 %s
8; RUN: llc -mtriple=riscv64 -mattr=+v,+d,+zcmp -O2 < %s \
9; RUN:    | FileCheck --check-prefix=SPILL-O2-ZCMP %s
10
11
12@.str = private unnamed_addr constant [6 x i8] c"hello\00", align 1
13
14define <vscale x 1 x double> @foo(<vscale x 1 x double> %a, <vscale x 1 x double> %b, <vscale x 1 x double> %c, i64 %gvl) nounwind
15; SPILL-O0-LABEL: foo:
16; SPILL-O0:       # %bb.0:
17; SPILL-O0-NEXT:    addi sp, sp, -48
18; SPILL-O0-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
19; SPILL-O0-NEXT:    csrr a1, vlenb
20; SPILL-O0-NEXT:    slli a1, a1, 1
21; SPILL-O0-NEXT:    sub sp, sp, a1
22; SPILL-O0-NEXT:    sd a0, 16(sp) # 8-byte Folded Spill
23; SPILL-O0-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
24; SPILL-O0-NEXT:    vmv1r.v v10, v9
25; SPILL-O0-NEXT:    vmv1r.v v9, v8
26; SPILL-O0-NEXT:    csrr a1, vlenb
27; SPILL-O0-NEXT:    add a1, sp, a1
28; SPILL-O0-NEXT:    addi a1, a1, 32
29; SPILL-O0-NEXT:    vs1r.v v9, (a1) # Unknown-size Folded Spill
30; SPILL-O0-NEXT:    # implicit-def: $v8
31; SPILL-O0-NEXT:    vsetvli zero, a0, e64, m1, tu, ma
32; SPILL-O0-NEXT:    vfadd.vv v8, v9, v10
33; SPILL-O0-NEXT:    addi a0, sp, 32
34; SPILL-O0-NEXT:    vs1r.v v8, (a0) # Unknown-size Folded Spill
35; SPILL-O0-NEXT:    lui a0, %hi(.L.str)
36; SPILL-O0-NEXT:    addi a0, a0, %lo(.L.str)
37; SPILL-O0-NEXT:    call puts
38; SPILL-O0-NEXT:    addi a1, sp, 32
39; SPILL-O0-NEXT:    vl1r.v v10, (a1) # Unknown-size Folded Reload
40; SPILL-O0-NEXT:    csrr a1, vlenb
41; SPILL-O0-NEXT:    add a1, sp, a1
42; SPILL-O0-NEXT:    addi a1, a1, 32
43; SPILL-O0-NEXT:    vl1r.v v9, (a1) # Unknown-size Folded Reload
44; SPILL-O0-NEXT:    # kill: def $x11 killed $x10
45; SPILL-O0-NEXT:    ld a0, 16(sp) # 8-byte Folded Reload
46; SPILL-O0-NEXT:    # implicit-def: $v8
47; SPILL-O0-NEXT:    vsetvli zero, a0, e64, m1, tu, ma
48; SPILL-O0-NEXT:    vfadd.vv v8, v9, v10
49; SPILL-O0-NEXT:    csrr a0, vlenb
50; SPILL-O0-NEXT:    slli a0, a0, 1
51; SPILL-O0-NEXT:    add sp, sp, a0
52; SPILL-O0-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
53; SPILL-O0-NEXT:    addi sp, sp, 48
54; SPILL-O0-NEXT:    ret
55;
56; SPILL-O2-LABEL: foo:
57; SPILL-O2:       # %bb.0:
58; SPILL-O2-NEXT:    addi sp, sp, -32
59; SPILL-O2-NEXT:    sd ra, 24(sp) # 8-byte Folded Spill
60; SPILL-O2-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
61; SPILL-O2-NEXT:    csrr a1, vlenb
62; SPILL-O2-NEXT:    slli a1, a1, 1
63; SPILL-O2-NEXT:    sub sp, sp, a1
64; SPILL-O2-NEXT:    mv s0, a0
65; SPILL-O2-NEXT:    addi a1, sp, 16
66; SPILL-O2-NEXT:    vs1r.v v8, (a1) # Unknown-size Folded Spill
67; SPILL-O2-NEXT:    vsetvli zero, a0, e64, m1, ta, ma
68; SPILL-O2-NEXT:    vfadd.vv v9, v8, v9
69; SPILL-O2-NEXT:    csrr a0, vlenb
70; SPILL-O2-NEXT:    add a0, sp, a0
71; SPILL-O2-NEXT:    addi a0, a0, 16
72; SPILL-O2-NEXT:    vs1r.v v9, (a0) # Unknown-size Folded Spill
73; SPILL-O2-NEXT:    lui a0, %hi(.L.str)
74; SPILL-O2-NEXT:    addi a0, a0, %lo(.L.str)
75; SPILL-O2-NEXT:    call puts
76; SPILL-O2-NEXT:    csrr a0, vlenb
77; SPILL-O2-NEXT:    add a0, sp, a0
78; SPILL-O2-NEXT:    addi a0, a0, 16
79; SPILL-O2-NEXT:    vl1r.v v8, (a0) # Unknown-size Folded Reload
80; SPILL-O2-NEXT:    addi a0, sp, 16
81; SPILL-O2-NEXT:    vl1r.v v9, (a0) # Unknown-size Folded Reload
82; SPILL-O2-NEXT:    vsetvli zero, s0, e64, m1, ta, ma
83; SPILL-O2-NEXT:    vfadd.vv v8, v9, v8
84; SPILL-O2-NEXT:    csrr a0, vlenb
85; SPILL-O2-NEXT:    slli a0, a0, 1
86; SPILL-O2-NEXT:    add sp, sp, a0
87; SPILL-O2-NEXT:    ld ra, 24(sp) # 8-byte Folded Reload
88; SPILL-O2-NEXT:    ld s0, 16(sp) # 8-byte Folded Reload
89; SPILL-O2-NEXT:    addi sp, sp, 32
90; SPILL-O2-NEXT:    ret
91;
92; SPILL-O2-VLEN128-LABEL: foo:
93; SPILL-O2-VLEN128:       # %bb.0:
94; SPILL-O2-VLEN128-NEXT:    addi sp, sp, -32
95; SPILL-O2-VLEN128-NEXT:    sd ra, 24(sp) # 8-byte Folded Spill
96; SPILL-O2-VLEN128-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
97; SPILL-O2-VLEN128-NEXT:    addi sp, sp, -32
98; SPILL-O2-VLEN128-NEXT:    mv s0, a0
99; SPILL-O2-VLEN128-NEXT:    addi a1, sp, 16
100; SPILL-O2-VLEN128-NEXT:    vs1r.v v8, (a1) # Unknown-size Folded Spill
101; SPILL-O2-VLEN128-NEXT:    vsetvli zero, a0, e64, m1, ta, ma
102; SPILL-O2-VLEN128-NEXT:    vfadd.vv v9, v8, v9
103; SPILL-O2-VLEN128-NEXT:    addi a0, sp, 32
104; SPILL-O2-VLEN128-NEXT:    vs1r.v v9, (a0) # Unknown-size Folded Spill
105; SPILL-O2-VLEN128-NEXT:    lui a0, %hi(.L.str)
106; SPILL-O2-VLEN128-NEXT:    addi a0, a0, %lo(.L.str)
107; SPILL-O2-VLEN128-NEXT:    call puts
108; SPILL-O2-VLEN128-NEXT:    addi a0, sp, 32
109; SPILL-O2-VLEN128-NEXT:    vl1r.v v8, (a0) # Unknown-size Folded Reload
110; SPILL-O2-VLEN128-NEXT:    addi a0, sp, 16
111; SPILL-O2-VLEN128-NEXT:    vl1r.v v9, (a0) # Unknown-size Folded Reload
112; SPILL-O2-VLEN128-NEXT:    vsetvli zero, s0, e64, m1, ta, ma
113; SPILL-O2-VLEN128-NEXT:    vfadd.vv v8, v9, v8
114; SPILL-O2-VLEN128-NEXT:    addi sp, sp, 32
115; SPILL-O2-VLEN128-NEXT:    ld ra, 24(sp) # 8-byte Folded Reload
116; SPILL-O2-VLEN128-NEXT:    ld s0, 16(sp) # 8-byte Folded Reload
117; SPILL-O2-VLEN128-NEXT:    addi sp, sp, 32
118; SPILL-O2-VLEN128-NEXT:    ret
119;
120; SPILL-O2-ZCMP-LABEL: foo:
121; SPILL-O2-ZCMP:       # %bb.0:
122; SPILL-O2-ZCMP-NEXT:    cm.push {ra, s0}, -32
123; SPILL-O2-ZCMP-NEXT:    csrr a1, vlenb
124; SPILL-O2-ZCMP-NEXT:    slli a1, a1, 1
125; SPILL-O2-ZCMP-NEXT:    sub sp, sp, a1
126; SPILL-O2-ZCMP-NEXT:    mv s0, a0
127; SPILL-O2-ZCMP-NEXT:    addi a1, sp, 16
128; SPILL-O2-ZCMP-NEXT:    vs1r.v v8, (a1) # Unknown-size Folded Spill
129; SPILL-O2-ZCMP-NEXT:    vsetvli zero, a0, e64, m1, ta, ma
130; SPILL-O2-ZCMP-NEXT:    vfadd.vv v9, v8, v9
131; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
132; SPILL-O2-ZCMP-NEXT:    add a0, a0, sp
133; SPILL-O2-ZCMP-NEXT:    addi a0, a0, 16
134; SPILL-O2-ZCMP-NEXT:    vs1r.v v9, (a0) # Unknown-size Folded Spill
135; SPILL-O2-ZCMP-NEXT:    lui a0, %hi(.L.str)
136; SPILL-O2-ZCMP-NEXT:    addi a0, a0, %lo(.L.str)
137; SPILL-O2-ZCMP-NEXT:    call puts
138; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
139; SPILL-O2-ZCMP-NEXT:    add a0, a0, sp
140; SPILL-O2-ZCMP-NEXT:    addi a0, a0, 16
141; SPILL-O2-ZCMP-NEXT:    vl1r.v v8, (a0) # Unknown-size Folded Reload
142; SPILL-O2-ZCMP-NEXT:    addi a0, sp, 16
143; SPILL-O2-ZCMP-NEXT:    vl1r.v v9, (a0) # Unknown-size Folded Reload
144; SPILL-O2-ZCMP-NEXT:    vsetvli zero, s0, e64, m1, ta, ma
145; SPILL-O2-ZCMP-NEXT:    vfadd.vv v8, v9, v8
146; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
147; SPILL-O2-ZCMP-NEXT:    slli a0, a0, 1
148; SPILL-O2-ZCMP-NEXT:    add sp, sp, a0
149; SPILL-O2-ZCMP-NEXT:    cm.popret {ra, s0}, 32
150{
151   %x = call <vscale x 1 x double> @llvm.riscv.vfadd.nxv1f64.nxv1f64(<vscale x 1 x double> undef, <vscale x 1 x double> %a, <vscale x 1 x double> %b, i64 7, i64 %gvl)
152   %call = call signext i32 @puts(ptr @.str)
153   %z = call <vscale x 1 x double> @llvm.riscv.vfadd.nxv1f64.nxv1f64(<vscale x 1 x double> undef, <vscale x 1 x double> %a, <vscale x 1 x double> %x, i64 7, i64 %gvl)
154   ret <vscale x 1 x double> %z
155}
156
157declare <vscale x 1 x double> @llvm.riscv.vfadd.nxv1f64.nxv1f64(<vscale x 1 x double> %passthru, <vscale x 1 x double> %a, <vscale x 1 x double> %b, i64, i64 %gvl)
158declare i32 @puts(ptr);
159