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
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
|
/*
* include/asm-ia64/xor.h
*
* Optimized RAID-5 checksumming functions for IA-64.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2, or (at your option)
* any later version.
*
* You should have received a copy of the GNU General Public License
* (for example /usr/src/linux/COPYING); if not, write to the Free
* Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
extern void xor_ia64_2(unsigned long, unsigned long *, unsigned long *);
extern void xor_ia64_3(unsigned long, unsigned long *, unsigned long *,
unsigned long *);
extern void xor_ia64_4(unsigned long, unsigned long *, unsigned long *,
unsigned long *, unsigned long *);
extern void xor_ia64_5(unsigned long, unsigned long *, unsigned long *,
unsigned long *, unsigned long *, unsigned long *);
asm ("
.text
// Assume L2 memory latency of 6 cycles.
.proc xor_ia64_2
xor_ia64_2:
.prologue
.fframe 0
{ .mii
.save ar.pfs, r31
alloc r31 = ar.pfs, 3, 0, 13, 16
.save ar.lc, r30
mov r30 = ar.lc
.save pr, r29
mov r29 = pr
;;
}
.body
{ .mii
mov r8 = in1
mov ar.ec = 6 + 2
shr in0 = in0, 3
;;
}
{ .mmi
adds in0 = -1, in0
mov r16 = in1
mov r17 = in2
;;
}
{ .mii
mov ar.lc = in0
mov pr.rot = 1 << 16
;;
}
.rotr s1[6+1], s2[6+1], d[2]
.rotp p[6+2]
0: { .mmi
(p[0]) ld8.nta s1[0] = [r16], 8
(p[0]) ld8.nta s2[0] = [r17], 8
(p[6]) xor d[0] = s1[6], s2[6]
}
{ .mfb
(p[6+1]) st8.nta [r8] = d[1], 8
nop.f 0
br.ctop.dptk.few 0b
;;
}
{ .mii
mov ar.lc = r30
mov pr = r29, -1
}
{ .bbb
br.ret.sptk.few rp
}
.endp xor_ia64_2
.proc xor_ia64_3
xor_ia64_3:
.prologue
.fframe 0
{ .mii
.save ar.pfs, r31
alloc r31 = ar.pfs, 4, 0, 20, 24
.save ar.lc, r30
mov r30 = ar.lc
.save pr, r29
mov r29 = pr
;;
}
.body
{ .mii
mov r8 = in1
mov ar.ec = 6 + 2
shr in0 = in0, 3
;;
}
{ .mmi
adds in0 = -1, in0
mov r16 = in1
mov r17 = in2
;;
}
{ .mii
mov r18 = in3
mov ar.lc = in0
mov pr.rot = 1 << 16
;;
}
.rotr s1[6+1], s2[6+1], s3[6+1], d[2]
.rotp p[6+2]
0: { .mmi
(p[0]) ld8.nta s1[0] = [r16], 8
(p[0]) ld8.nta s2[0] = [r17], 8
(p[6]) xor d[0] = s1[6], s2[6]
;;
}
{ .mmi
(p[0]) ld8.nta s3[0] = [r18], 8
(p[6+1]) st8.nta [r8] = d[1], 8
(p[6]) xor d[0] = d[0], s3[6]
}
{ .bbb
br.ctop.dptk.few 0b
;;
}
{ .mii
mov ar.lc = r30
mov pr = r29, -1
}
{ .bbb
br.ret.sptk.few rp
}
.endp xor_ia64_3
.proc xor_ia64_4
xor_ia64_4:
.prologue
.fframe 0
{ .mii
.save ar.pfs, r31
alloc r31 = ar.pfs, 5, 0, 27, 32
.save ar.lc, r30
mov r30 = ar.lc
.save pr, r29
mov r29 = pr
;;
}
.body
{ .mii
mov r8 = in1
mov ar.ec = 6 + 2
shr in0 = in0, 3
;;
}
{ .mmi
adds in0 = -1, in0
mov r16 = in1
mov r17 = in2
;;
}
{ .mii
mov r18 = in3
mov ar.lc = in0
mov pr.rot = 1 << 16
}
{ .mfb
mov r19 = in4
;;
}
.rotr s1[6+1], s2[6+1], s3[6+1], s4[6+1], d[2]
.rotp p[6+2]
0: { .mmi
(p[0]) ld8.nta s1[0] = [r16], 8
(p[0]) ld8.nta s2[0] = [r17], 8
(p[6]) xor d[0] = s1[6], s2[6]
}
{ .mmi
(p[0]) ld8.nta s3[0] = [r18], 8
(p[0]) ld8.nta s4[0] = [r19], 8
(p[6]) xor r20 = s3[6], s4[6]
;;
}
{ .mib
(p[6+1]) st8.nta [r8] = d[1], 8
(p[6]) xor d[0] = d[0], r20
br.ctop.dptk.few 0b
;;
}
{ .mii
mov ar.lc = r30
mov pr = r29, -1
}
{ .bbb
br.ret.sptk.few rp
}
.endp xor_ia64_4
.proc xor_ia64_5
xor_ia64_5:
.prologue
.fframe 0
{ .mii
.save ar.pfs, r31
alloc r31 = ar.pfs, 6, 0, 34, 40
.save ar.lc, r30
mov r30 = ar.lc
.save pr, r29
mov r29 = pr
;;
}
.body
{ .mii
mov r8 = in1
mov ar.ec = 6 + 2
shr in0 = in0, 3
;;
}
{ .mmi
adds in0 = -1, in0
mov r16 = in1
mov r17 = in2
;;
}
{ .mii
mov r18 = in3
mov ar.lc = in0
mov pr.rot = 1 << 16
}
{ .mib
mov r19 = in4
mov r20 = in5
;;
}
.rotr s1[6+1], s2[6+1], s3[6+1], s4[6+1], s5[6+1], d[2]
.rotp p[6+2]
0: { .mmi
(p[0]) ld8.nta s1[0] = [r16], 8
(p[0]) ld8.nta s2[0] = [r17], 8
(p[6]) xor d[0] = s1[6], s2[6]
}
{ .mmi
(p[0]) ld8.nta s3[0] = [r18], 8
(p[0]) ld8.nta s4[0] = [r19], 8
(p[6]) xor r21 = s3[6], s4[6]
;;
}
{ .mmi
(p[0]) ld8.nta s5[0] = [r20], 8
(p[6+1]) st8.nta [r8] = d[1], 8
(p[6]) xor d[0] = d[0], r21
;;
}
{ .mfb
(p[6]) xor d[0] = d[0], s5[6]
nop.f 0
br.ctop.dptk.few 0b
;;
}
{ .mii
mov ar.lc = r30
mov pr = r29, -1
}
{ .bbb
br.ret.sptk.few rp
}
.endp xor_ia64_5
");
static struct xor_block_template xor_block_ia64 = {
name: "ia64",
do_2: xor_ia64_2,
do_3: xor_ia64_3,
do_4: xor_ia64_4,
do_5: xor_ia64_5,
};
#define XOR_TRY_TEMPLATES xor_speed(&xor_block_ia64)
|
