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
|
static void test_instr_NEON(int printlog)
{
unsigned long long copy_size;
unsigned long flags;
long i, j, avg, pld_dst;
unsigned long long t1, t2, t_avg;
unsigned long temp;
unsigned long result[10];
unsigned long long t_result[10];
int best_pld[512];
char output_line[380];
char buf[32];
char *src, *dst;
copy_size = 256;
/* copy_size = 1024*8; */
/* copy_size = 1024*64; */
/* copy_size = 1024*1024; */
#if 1
src = (char*)((unsigned long)buffer_src+ 0x10 & ~0xf);
dst = (char*)((unsigned long)buffer_dst+ 0x10 & ~0xf);
#else
src = (char*)((unsigned long)buffer_src+ 0x8 & ~0x7);
dst = (char*)((unsigned long)buffer_dst+ 0x8 & ~0x7);
#endif
if (printlog == 1)
printk(KERN_EMERG"\n\n\r == ARM64 memcpy === \n 0x%lx-->0x%lx\n 0x%lx-->0x%lx\n\r", (unsigned long)buffer_src, (unsigned long)buffer_dst, (unsigned long)src, (unsigned long)dst);
/* for(i = 0; i< 256 + 16 + 4; i++, copy_size += 256){ */
i = 0;
while (i < 256 + 16 + 4) {
int line_offset = 0;
if (i < 256) {
copy_size = 256 + i * 256; /* inc 256 byte from 0~64 KBytes */
} else if (i < 256 + 16) {
copy_size = 1024 * 64 + (i - 256) * 1024 * 64; /* inc 64Kbyte form 64KB~1MB */
} else if (i < 256 + 16 + 4) {
copy_size = 1024 * 1024 + (i - 256 - 16) * 1024 * 1024; /* inc 1MB from 1MB~4MB */
}
i++;
//mdelay(5);
preempt_disable();
local_irq_save(flags);
/* for(i = 0; i< 8; i++, copy_size += 1024*8){ */
/* for(i = 0; i< 16; i++, copy_size += 1024*64){ */
/* for(i = 0; i< 4; i++, copy_size += 1024*1024){ */
inner_dcache_flush_all();
if(printlog == 1){
memset(buf, 0, 32);
sprintf(buf," \n%llu :",copy_size);
snprintf(output_line+line_offset, 320 - line_offset, "%s",buf);
line_offset += strlen(buf);
}
/* cached */
for(j= 0; j< copy_size; j++)
dst[j]=src[j];
avg = 0;
/* no pld */
for (j = 0; j < 8; j++) {
mdelay(3);
if(flush_cache == 1)
inner_dcache_flush_all();
t1 = sched_clock();
#if 0
asm volatile(
"mov x0, %0\n"
"mov x1, %1\n"
"mov x2, %2\n"
"mov x3, %3\n"
"subs x2, x2, #128\n"
/* There are at least 128 bytes to copy. */
"ldp x7, x8, [x1, #0]\n"
"sub x0, x0, #16 \n" /* Pre-bias. */
"ldp x9, x10, [x1, #16]\n"
"ldp x11, x12, [x1, #32]\n"
"ldp x13, x14, [x1, #48]!\n" /* src += 64 - Pre-bias. */
"1:\n"
"stp x7, x8, [x0, #16]\n"
"ldp x7, x8, [x1, #16]\n"
"stp x9, x10, [x0, #32]\n"
"ldp x9, x10, [x1, #32]\n"
"stp x11, x12, [x0, #48]\n"
"ldp x11, x12, [x1, #48]\n"
"stp x13, x14, [x0, #64]!\n"
"ldp x13, x14, [x1, #64]!\n"
"subs x2, x2, #64\n"
"b.ge 1b\n"
"stp x7, x8, [x0, #16]\n"
"stp x9, x10, [x0, #32]\n"
"stp x11, x12, [x0, #48]\n"
"stp x13, x14, [x0, #64]!\n"
: : "r" (dst), "r"(src), "r"(copy_size),"r"(pld_dst)
: "x0", "x1", "x2", "x3", "x7", "x8", "x9", "x10", "x11", "x12", "x13", "x14");
#else
asm volatile(
"mov x0, %0\n"
"mov x1, %1\n"
"mov x2, %2\n"
"mov x3, %3\n"
"subs x2, x2, #128\n"
/* There are at least 128 bytes to copy. */
"ldur q1, [x1, #0]\n"
"sub x0, x0, #16 \n" /* Pre-bias. */
"ldur q2, [x1, #16]\n"
"ldur q3, [x1, #32]\n"
"ldur q4, [x1, #48]\n"
"add x1, x1, #48\n"
"1:\n"
"stur q1, [x0, #16]\n"
"ldur q1, [x1, #16]\n"
"stur q2, [x0, #32]\n"
"ldur q2, [x1, #32]\n"
"stur q3, [x0, #48]\n"
"ldur q3, [x1, #48]\n"
"stur q4, [x0, #64]\n"
"add x0, x0, #64\n"
"ldur q4, [x1, #64]\n"
"add x1, x1, #64\n"
"subs x2, x2, #64\n"
"b.ge 1b\n"
"stur q1, [x0, #16]\n"
"stur q2, [x0, #32]\n"
"stur q3, [x0, #48]\n"
"stur q4, [x0, #64]\n"
: : "r" (dst), "r"(src), "r"(copy_size),"r"(pld_dst)
: "x0", "x1", "x2", "x3", "x7", "x8", "x9", "x10", "x11", "x12", "x13", "x14");
#endif
t2 = sched_clock();
t_result[j] = t2 - t1;
} /* 10 times loop */
t_avg = 0;
for (j = 0; j < 8; j++) {
t_avg += t_result[j];
}
t_avg = t_avg >> 3;
if(printlog == 1){
//printk(KERN_CONT" %d ", avg );
memset(buf, 0, 32);
snprintf(buf,32, " %llu ",t_avg);
snprintf(output_line+line_offset, 320 - line_offset, "%s %d",buf, flush_cache);
line_offset += strlen(buf);
}
// for (pld_dst = 0; pld_dst < 64 * 16 * 2; pld_dst += 64*2) {
for (pld_dst = 64 * 16 * 2 -128 ; pld_dst >=0 ; pld_dst -= 64*2) {
avg = 0;
for (j = 0; j < 8; j++) {
mdelay(3);
if (flush_cache == 1)
inner_dcache_flush_all();
t1 = sched_clock();
#if 0
asm volatile(
"mov x0, %0\n"
"mov x1, %1\n"
"mov x2, %2\n"
"mov x3, %3\n"
"subs x2, x2, #128\n"
/* There are at least 128 bytes to copy. */
"ldp x7, x8, [x1, #0]\n"
"sub x0, x0, #16 \n" /* Pre-bias. */
"ldp x9, x10, [x1, #16]\n"
"ldp x11, x12, [x1, #32]\n"
"ldp x13, x14, [x1, #48]!\n" /* src += 64 - Pre-bias. */
"1:\n"
"prfm pldl2keep, [x1,x3]\n"
"stp x7, x8, [x0, #16]\n"
"ldp x7, x8, [x1, #16]\n"
"stp x9, x10, [x0, #32]\n"
"ldp x9, x10, [x1, #32]\n"
"stp x11, x12, [x0, #48]\n"
"ldp x11, x12, [x1, #48]\n"
"stp x13, x14, [x0, #64]!\n"
"ldp x13, x14, [x1, #64]!\n"
"subs x2, x2, #64\n"
"b.ge 1b\n"
"stp x7, x8, [x0, #16]\n"
"stp x9, x10, [x0, #32]\n"
"stp x11, x12, [x0, #48]\n"
"stp x13, x14, [x0, #64]!\n"
: : "r" (dst), "r"(src), "r"(copy_size),"r"(pld_dst)
: "x0", "x1", "x2", "x3", "x7", "x8", "x9", "x10", "x11", "x12", "x13", "x14");
#else
asm volatile(
"mov x0, %0\n"
"mov x1, %1\n"
"mov x2, %2\n"
"mov x3, %3\n"
"subs x2, x2, #128\n"
/* There are at least 128 bytes to copy. */
"ldur q1, [x1, #0]\n"
"sub x0, x0, #16 \n" /* Pre-bias. */
"ldur q2, [x1, #16]\n"
"ldur q3, [x1, #32]\n"
"ldur q4, [x1, #48]\n"
"add x1, x1, #48\n"
"1:\n"
"prfm pldl1keep, [x1,x3]\n"
"stur q1, [x0, #16]\n"
"ldur q1, [x1, #16]\n"
"stur q2, [x0, #32]\n"
"ldur q2, [x1, #32]\n"
"stur q3, [x0, #48]\n"
"ldur q3, [x1, #48]\n"
"stur q4, [x0, #64]\n"
"add x0, x0, #64\n"
"ldur q4, [x1, #64]\n"
"add x1, x1, #64\n"
"subs x2, x2, #64\n"
"b.ge 1b\n"
"stur q1, [x0, #16]\n"
"stur q2, [x0, #32]\n"
"stur q3, [x0, #48]\n"
"stur q4, [x0, #64]\n"
: : "r" (dst), "r"(src), "r"(copy_size),"r"(pld_dst)
: "x0", "x1", "x2", "x3", "x7", "x8", "x9", "x10", "x11", "x12", "x13", "x14");
#endif
t2 = sched_clock();
t_result[j] = t2 - t1;
} /* 10 times loop */
t_avg = 0;
for (j = 0; j < 8; j++) {
t_avg += t_result[j];
}
t_avg = t_avg >> 3;
if(printlog == 1){
//printk(KERN_CONT" %d ", avg );
memset(buf, 0, 32);
snprintf(buf,32, " %llu ",t_avg);
snprintf(output_line+line_offset, 320 - line_offset, "%s",buf);
line_offset += strlen(buf);
}
} /* pld dist loop */
local_irq_restore(flags);
preempt_enable();
printk(KERN_EMERG "%s %d", output_line, flush_cache);
}
if (printlog == 1)
pr_err("\n\r ====NEON instruction test done ==== flush_cache:%d\n", flush_cache);
}
|
