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
|
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -passes='sroa<preserve-cfg>' -S | FileCheck %s --check-prefixes=CHECK,CHECK-PRESERVE-CFG
; RUN: opt < %s -passes='sroa<modify-cfg>' -S | FileCheck %s --check-prefixes=CHECK,CHECK-MODIFY-CFG
; This test checks that SROA runs mem2reg on scalable vectors.
define <vscale x 16 x i1> @alloca_nxv16i1(<vscale x 16 x i1> %pg) {
; CHECK-LABEL: @alloca_nxv16i1(
; CHECK-NEXT: ret <vscale x 16 x i1> [[PG:%.*]]
;
%pg.addr = alloca <vscale x 16 x i1>
store <vscale x 16 x i1> %pg, ptr %pg.addr
%1 = load <vscale x 16 x i1>, ptr %pg.addr
ret <vscale x 16 x i1> %1
}
define <vscale x 16 x i8> @alloca_nxv16i8(<vscale x 16 x i8> %vec) {
; CHECK-LABEL: @alloca_nxv16i8(
; CHECK-NEXT: ret <vscale x 16 x i8> [[VEC:%.*]]
;
%vec.addr = alloca <vscale x 16 x i8>
store <vscale x 16 x i8> %vec, ptr %vec.addr
%1 = load <vscale x 16 x i8>, ptr %vec.addr
ret <vscale x 16 x i8> %1
}
; Test scalable alloca that can't be promoted. Mem2Reg only considers
; non-volatile loads and stores for promotion.
define <vscale x 16 x i8> @unpromotable_alloca(<vscale x 16 x i8> %vec) {
; CHECK-LABEL: @unpromotable_alloca(
; CHECK-NEXT: [[VEC_ADDR:%.*]] = alloca <vscale x 16 x i8>, align 16
; CHECK-NEXT: store volatile <vscale x 16 x i8> [[VEC:%.*]], ptr [[VEC_ADDR]], align 16
; CHECK-NEXT: [[TMP1:%.*]] = load volatile <vscale x 16 x i8>, ptr [[VEC_ADDR]], align 16
; CHECK-NEXT: ret <vscale x 16 x i8> [[TMP1]]
;
%vec.addr = alloca <vscale x 16 x i8>
store volatile <vscale x 16 x i8> %vec, ptr %vec.addr
%1 = load volatile <vscale x 16 x i8>, ptr %vec.addr
ret <vscale x 16 x i8> %1
}
; Test we bail out when using an alloca of a fixed-length vector (VLS) that was
; bitcasted to a scalable vector.
define <vscale x 4 x i32> @cast_alloca_to_svint32_t(<vscale x 4 x i32> %type.coerce) {
; CHECK-LABEL: @cast_alloca_to_svint32_t(
; CHECK-NEXT: [[TYPE:%.*]] = alloca <16 x i32>, align 64
; CHECK-NEXT: [[TYPE_ADDR:%.*]] = alloca <16 x i32>, align 64
; CHECK-NEXT: store <vscale x 4 x i32> [[TYPE_COERCE:%.*]], ptr [[TYPE]], align 16
; CHECK-NEXT: [[TYPE1:%.*]] = load <16 x i32>, ptr [[TYPE]], align 64
; CHECK-NEXT: store <16 x i32> [[TYPE1]], ptr [[TYPE_ADDR]], align 64
; CHECK-NEXT: [[TMP1:%.*]] = load <16 x i32>, ptr [[TYPE_ADDR]], align 64
; CHECK-NEXT: [[TMP2:%.*]] = load <vscale x 4 x i32>, ptr [[TYPE_ADDR]], align 16
; CHECK-NEXT: ret <vscale x 4 x i32> [[TMP2]]
;
%type = alloca <16 x i32>
%type.addr = alloca <16 x i32>
store <vscale x 4 x i32> %type.coerce, ptr %type
%type1 = load <16 x i32>, ptr %type
store <16 x i32> %type1, ptr %type.addr
%1 = load <16 x i32>, ptr %type.addr
%2 = load <vscale x 4 x i32>, ptr %type.addr
ret <vscale x 4 x i32> %2
}
; When casting from VLA to VLS via memory check we bail out when producing a
; GEP where the element type is a scalable vector.
define <vscale x 4 x i32> @cast_alloca_from_svint32_t() {
; CHECK-LABEL: @cast_alloca_from_svint32_t(
; CHECK-NEXT: [[RETVAL_COERCE:%.*]] = alloca <vscale x 4 x i32>, align 16
; CHECK-NEXT: store <16 x i32> undef, ptr [[RETVAL_COERCE]], align 16
; CHECK-NEXT: [[TMP1:%.*]] = load <vscale x 4 x i32>, ptr [[RETVAL_COERCE]], align 16
; CHECK-NEXT: ret <vscale x 4 x i32> [[TMP1]]
;
%retval = alloca <16 x i32>
%retval.coerce = alloca <vscale x 4 x i32>
call void @llvm.memcpy.p0.p0.i64(ptr align 16 %retval.coerce, ptr align 16 %retval, i64 64, i1 false)
%1 = load <vscale x 4 x i32>, ptr %retval.coerce
ret <vscale x 4 x i32> %1
}
; Test we bail out when using an alloca of a fixed-length vector (VLS) that was
; bitcasted to a scalable vector.
define void @select_load_alloca_to_svdouble_t() {
; CHECK-LABEL: @select_load_alloca_to_svdouble_t(
; CHECK-NEXT: [[Z:%.*]] = alloca <16 x half>, align 32
; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 0, 0
; CHECK-NEXT: [[COND:%.*]] = select i1 [[CMP]], ptr [[Z]], ptr null
; CHECK-NEXT: [[VAL:%.*]] = load <vscale x 2 x double>, ptr [[COND]], align 16
; CHECK-NEXT: ret void
;
%z = alloca <16 x half>
%cmp = icmp eq i32 0, 0
%cond = select i1 %cmp, ptr %z, ptr null
%val = load <vscale x 2 x double>, ptr %cond, align 16
ret void
}
define void @select_store_alloca_to_svdouble_t(<vscale x 2 x double> %val) {
; CHECK-LABEL: @select_store_alloca_to_svdouble_t(
; CHECK-NEXT: [[Z:%.*]] = alloca <16 x half>, align 32
; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 0, 0
; CHECK-NEXT: [[COND:%.*]] = select i1 [[CMP]], ptr [[Z]], ptr null
; CHECK-NEXT: store <vscale x 2 x double> [[VAL:%.*]], ptr [[COND]], align 16
; CHECK-NEXT: ret void
;
%z = alloca <16 x half>
%cmp = icmp eq i32 0, 0
%cond = select i1 %cmp, ptr %z, ptr null
store <vscale x 2 x double> %val, ptr %cond, align 16
ret void
}
declare void @llvm.memcpy.p0.p0.i64(ptr nocapture, ptr nocapture, i64, i1) nounwind
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; CHECK-MODIFY-CFG: {{.*}}
; CHECK-PRESERVE-CFG: {{.*}}
|
