1; RUN: llc %s -o - -mtriple=aarch64-unknown -mattr=+fuse-aes,+crypto | FileCheck %s
2; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=generic -mattr=+crypto | FileCheck %s
3; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=cortex-a53 | FileCheck %s
4; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=cortex-a57 | FileCheck %s
5; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=cortex-a72 | FileCheck %s
6; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=cortex-a73 | FileCheck %s
7; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=exynos-m1  | FileCheck %s
8; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=exynos-m2  | FileCheck %s
9; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=exynos-m3  | FileCheck %s
10; RUN: llc %s -o - -mtriple=aarch64-unknown -mcpu=exynos-m4  | FileCheck %s
11
12declare <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %d, <16 x i8> %k)
13declare <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %d)
14declare <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %d, <16 x i8> %k)
15declare <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %d)
16
17define void @aesea(<16 x i8>* %a0, <16 x i8>* %b0, <16 x i8>* %c0, <16 x i8> %d, <16 x i8> %e) {
18  %d0 = load <16 x i8>, <16 x i8>* %a0
19  %a1 = getelementptr inbounds <16 x i8>, <16 x i8>* %a0, i64 1
20  %d1 = load <16 x i8>, <16 x i8>* %a1
21  %a2 = getelementptr inbounds <16 x i8>, <16 x i8>* %a0, i64 2
22  %d2 = load <16 x i8>, <16 x i8>* %a2
23  %a3 = getelementptr inbounds <16 x i8>, <16 x i8>* %a0, i64 3
24  %d3 = load <16 x i8>, <16 x i8>* %a3
25  %k0 = load <16 x i8>, <16 x i8>* %b0
26  %e00 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %d0, <16 x i8> %k0)
27  %f00 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e00)
28  %e01 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %d1, <16 x i8> %k0)
29  %f01 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e01)
30  %e02 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %d2, <16 x i8> %k0)
31  %f02 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e02)
32  %e03 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %d3, <16 x i8> %k0)
33  %f03 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e03)
34  %b1 = getelementptr inbounds <16 x i8>, <16 x i8>* %b0, i64 1
35  %k1 = load <16 x i8>, <16 x i8>* %b1
36  %e10 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f00, <16 x i8> %k1)
37  %f10 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e00)
38  %e11 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f01, <16 x i8> %k1)
39  %f11 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e01)
40  %e12 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f02, <16 x i8> %k1)
41  %f12 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e02)
42  %e13 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f03, <16 x i8> %k1)
43  %f13 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e03)
44  %b2 = getelementptr inbounds <16 x i8>, <16 x i8>* %b0, i64 2
45  %k2 = load <16 x i8>, <16 x i8>* %b2
46  %e20 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f10, <16 x i8> %k2)
47  %f20 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e10)
48  %e21 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f11, <16 x i8> %k2)
49  %f21 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e11)
50  %e22 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f12, <16 x i8> %k2)
51  %f22 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e12)
52  %e23 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f13, <16 x i8> %k2)
53  %f23 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e13)
54  %b3 = getelementptr inbounds <16 x i8>, <16 x i8>* %b0, i64 3
55  %k3 = load <16 x i8>, <16 x i8>* %b3
56  %e30 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f20, <16 x i8> %k3)
57  %f30 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e20)
58  %e31 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f21, <16 x i8> %k3)
59  %f31 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e21)
60  %e32 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f22, <16 x i8> %k3)
61  %f32 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e22)
62  %e33 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f23, <16 x i8> %k3)
63  %f33 = call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %e23)
64  %g0 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f30, <16 x i8> %d)
65  %h0 = xor <16 x i8> %g0, %e
66  %g1 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f31, <16 x i8> %d)
67  %h1 = xor <16 x i8> %g1, %e
68  %g2 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f32, <16 x i8> %d)
69  %h2 = xor <16 x i8> %g2, %e
70  %g3 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %f33, <16 x i8> %d)
71  %h3 = xor <16 x i8> %g3, %e
72  store <16 x i8> %h0, <16 x i8>* %c0
73  %c1 = getelementptr inbounds <16 x i8>, <16 x i8>* %c0, i64 1
74  store <16 x i8> %h1, <16 x i8>* %c1
75  %c2 = getelementptr inbounds <16 x i8>, <16 x i8>* %c0, i64 2
76  store <16 x i8> %h2, <16 x i8>* %c2
77  %c3 = getelementptr inbounds <16 x i8>, <16 x i8>* %c0, i64 3
78  store <16 x i8> %h3, <16 x i8>* %c3
79  ret void
80
81; CHECK-LABEL: aesea:
82; CHECK: aese [[VA:v[0-7].16b]], {{v[0-7].16b}}
83; CHECK-NEXT: aesmc [[VA]], [[VA]]
84; CHECK: aese [[VB:v[0-7].16b]], {{v[0-7].16b}}
85; CHECK-NEXT: aesmc [[VB]], [[VB]]
86; CHECK: aese [[VC:v[0-7].16b]], {{v[0-7].16b}}
87; CHECK-NEXT: aesmc [[VC]], [[VC]]
88; CHECK: aese [[VD:v[0-7].16b]], {{v[0-7].16b}}
89; CHECK-NEXT: aesmc [[VD]], [[VD]]
90; CHECK: aese [[VE:v[0-7].16b]], {{v[0-7].16b}}
91; CHECK-NEXT: aesmc [[VE]], [[VE]]
92; CHECK: aese [[VF:v[0-7].16b]], {{v[0-7].16b}}
93; CHECK-NEXT: aesmc [[VF]], [[VF]]
94; CHECK: aese [[VG:v[0-7].16b]], {{v[0-7].16b}}
95; CHECK-NEXT: aesmc [[VG]], [[VG]]
96; CHECK: aese [[VH:v[0-7].16b]], {{v[0-7].16b}}
97; CHECK-NEXT: aesmc [[VH]], [[VH]]
98; CHECK-NOT: aesmc
99}
100
101define void @aesda(<16 x i8>* %a0, <16 x i8>* %b0, <16 x i8>* %c0, <16 x i8> %d, <16 x i8> %e) {
102  %d0 = load <16 x i8>, <16 x i8>* %a0
103  %a1 = getelementptr inbounds <16 x i8>, <16 x i8>* %a0, i64 1
104  %d1 = load <16 x i8>, <16 x i8>* %a1
105  %a2 = getelementptr inbounds <16 x i8>, <16 x i8>* %a0, i64 2
106  %d2 = load <16 x i8>, <16 x i8>* %a2
107  %a3 = getelementptr inbounds <16 x i8>, <16 x i8>* %a0, i64 3
108  %d3 = load <16 x i8>, <16 x i8>* %a3
109  %k0 = load <16 x i8>, <16 x i8>* %b0
110  %e00 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %d0, <16 x i8> %k0)
111  %f00 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e00)
112  %e01 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %d1, <16 x i8> %k0)
113  %f01 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e01)
114  %e02 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %d2, <16 x i8> %k0)
115  %f02 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e02)
116  %e03 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %d3, <16 x i8> %k0)
117  %f03 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e03)
118  %b1 = getelementptr inbounds <16 x i8>, <16 x i8>* %b0, i64 1
119  %k1 = load <16 x i8>, <16 x i8>* %b1
120  %e10 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f00, <16 x i8> %k1)
121  %f10 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e00)
122  %e11 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f01, <16 x i8> %k1)
123  %f11 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e01)
124  %e12 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f02, <16 x i8> %k1)
125  %f12 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e02)
126  %e13 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f03, <16 x i8> %k1)
127  %f13 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e03)
128  %b2 = getelementptr inbounds <16 x i8>, <16 x i8>* %b0, i64 2
129  %k2 = load <16 x i8>, <16 x i8>* %b2
130  %e20 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f10, <16 x i8> %k2)
131  %f20 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e10)
132  %e21 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f11, <16 x i8> %k2)
133  %f21 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e11)
134  %e22 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f12, <16 x i8> %k2)
135  %f22 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e12)
136  %e23 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f13, <16 x i8> %k2)
137  %f23 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e13)
138  %b3 = getelementptr inbounds <16 x i8>, <16 x i8>* %b0, i64 3
139  %k3 = load <16 x i8>, <16 x i8>* %b3
140  %e30 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f20, <16 x i8> %k3)
141  %f30 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e20)
142  %e31 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f21, <16 x i8> %k3)
143  %f31 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e21)
144  %e32 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f22, <16 x i8> %k3)
145  %f32 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e22)
146  %e33 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f23, <16 x i8> %k3)
147  %f33 = call <16 x i8> @llvm.aarch64.crypto.aesimc(<16 x i8> %e23)
148  %g0 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f30, <16 x i8> %d)
149  %h0 = xor <16 x i8> %g0, %e
150  %g1 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f31, <16 x i8> %d)
151  %h1 = xor <16 x i8> %g1, %e
152  %g2 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f32, <16 x i8> %d)
153  %h2 = xor <16 x i8> %g2, %e
154  %g3 = call <16 x i8> @llvm.aarch64.crypto.aesd(<16 x i8> %f33, <16 x i8> %d)
155  %h3 = xor <16 x i8> %g3, %e
156  store <16 x i8> %h0, <16 x i8>* %c0
157  %c1 = getelementptr inbounds <16 x i8>, <16 x i8>* %c0, i64 1
158  store <16 x i8> %h1, <16 x i8>* %c1
159  %c2 = getelementptr inbounds <16 x i8>, <16 x i8>* %c0, i64 2
160  store <16 x i8> %h2, <16 x i8>* %c2
161  %c3 = getelementptr inbounds <16 x i8>, <16 x i8>* %c0, i64 3
162  store <16 x i8> %h3, <16 x i8>* %c3
163  ret void
164
165; CHECK-LABEL: aesda:
166; CHECK: aesd [[VA:v[0-7].16b]], {{v[0-7].16b}}
167; CHECK-NEXT: aesimc [[VA]], [[VA]]
168; CHECK: aesd [[VB:v[0-7].16b]], {{v[0-7].16b}}
169; CHECK-NEXT: aesimc [[VB]], [[VB]]
170; CHECK: aesd [[VC:v[0-7].16b]], {{v[0-7].16b}}
171; CHECK-NEXT: aesimc [[VC]], [[VC]]
172; CHECK: aesd [[VD:v[0-7].16b]], {{v[0-7].16b}}
173; CHECK-NEXT: aesimc [[VD]], [[VD]]
174; CHECK: aesd [[VE:v[0-7].16b]], {{v[0-7].16b}}
175; CHECK-NEXT: aesimc [[VE]], [[VE]]
176; CHECK: aesd [[VF:v[0-7].16b]], {{v[0-7].16b}}
177; CHECK-NEXT: aesimc [[VF]], [[VF]]
178; CHECK: aesd [[VG:v[0-7].16b]], {{v[0-7].16b}}
179; CHECK-NEXT: aesimc [[VG]], [[VG]]
180; CHECK: aesd [[VH:v[0-7].16b]], {{v[0-7].16b}}
181; CHECK-NEXT: aesimc [[VH]], [[VH]]
182; CHECK-NOT: aesimc
183}
184
185define void @aes_load_store(<16 x i8> *%p1, <16 x i8> *%p2 , <16 x i8> *%p3) {
186entry:
187  %x1 = alloca <16 x i8>, align 16
188  %x2 = alloca <16 x i8>, align 16
189  %x3 = alloca <16 x i8>, align 16
190  %x4 = alloca <16 x i8>, align 16
191  %x5 = alloca <16 x i8>, align 16
192  %in1 = load <16 x i8>, <16 x i8>* %p1, align 16
193  store <16 x i8> %in1, <16 x i8>* %x1, align 16
194  %aese1 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %in1, <16 x i8> %in1) #2
195  %in2 = load <16 x i8>, <16 x i8>* %p2, align 16
196  %aesmc1= call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %aese1) #2
197  %aese2 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %in1, <16 x i8> %in2) #2
198  store <16 x i8> %aesmc1, <16 x i8>* %x3, align 16
199  %in3 = load <16 x i8>, <16 x i8>* %p3, align 16
200  %aesmc2= call <16 x i8> @llvm.aarch64.crypto.aesmc(<16 x i8> %aese2) #2
201  %aese3 = call <16 x i8> @llvm.aarch64.crypto.aese(<16 x i8> %aesmc2, <16 x i8> %in3) #2
202  store <16 x i8> %aese3, <16 x i8>* %x5, align 16
203  ret void
204
205; CHECK-LABEL: aes_load_store:
206; CHECK: aese [[VA:v[0-7].16b]], {{v[0-7].16b}}
207; CHECK-NEXT: aesmc [[VA]], [[VA]]
208; CHECK: aese [[VB:v[0-7].16b]], {{v[0-7].16b}}
209; CHECK-NEXT: aesmc [[VB]], [[VB]]
210; CHECK-NOT: aesmc
211}
212