-
Notifications
You must be signed in to change notification settings - Fork 3
/
monitoring.c
483 lines (397 loc) · 12.9 KB
/
monitoring.c
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
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
/* ----------------------------------------------------------------------- *
*
* Copyright 2016 Clémentine Maurice
* Copyright 2021 Guillaume Didier
*
* 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, Inc., 51 Franklin St, Fifth Floor,
* Boston MA 02110-1301, USA; either version 2 of the License, or
* (at your option) any later version. *
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
* ----------------------------------------------------------------------- */
#define _GNU_SOURCE
#include <errno.h>
#include <sched.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include "arch.h"
#include "global_variables.h"
#include "monitoring.h"
#include "poke.h"
#include "rdmsr.h"
#include "util.h"
#include "wrmsr.h"
#define SIZE_HIST (600)
int monitor_single_address_clflush(uintptr_t addr, int print) {
uintptr_t paddr = read_pagemap("/proc/self/pagemap", addr);
unsigned long cores_package = cores_per_package();
int *map_coreid = mapping_coreid();
int *map_apicid = mapping_apicid();
int i, j, thread;
size_t hit_histogram[SIZE_HIST];
int nb_tries = 50 * 1024;
/*
* Execute some code on every core (!= every thread)
*/
int *core_used;
core_used = (int *)calloc(cores_package, sizeof(int));
unsigned long current_apicid = -1;
int current_core = -1;
int mask;
int n = threads_per_package();
int slice = 0;
for (thread = 0; thread < n; thread++) {
mask = thread;
cpu_set_t my_set; // Define your cpu_set bit mask.
CPU_ZERO(&my_set); // Initialize it all to 0, i.e. no CPUs selected.
CPU_SET(mask, &my_set); // set the bit that represents core
// Set affinity of this process to mask
if (sched_setaffinity(0, sizeof(cpu_set_t), &my_set) == -1) {
printf("Error with sched_setaffinity\n");
exit(-1);
}
current_apicid = current_apic();
current_core = apicid2coreid(current_apicid, map_apicid, map_coreid);
memset(hit_histogram, 0, SIZE_HIST * sizeof(*hit_histogram));
if (core_used[current_core] == 0)
// Code to execute in every core
{
// Construct clflush hit histogram
for (i = 0; i < nb_tries; ++i) {
size_t d = flush_hit((char *)addr);
hit_histogram[MIN(599, d)]++;
for (j = 0; j < 1; ++j)
sched_yield();
}
// Print histogram for each core if not sure of what the threshold values should
// be
//#define DEBUG
#ifdef DEBUG
for (i = 145; i < 180; ++i) {
printf("%3d: %15zu\n", i, hit_histogram[i]);
}
#endif
// Based on the historgram, is the address on the same slice as the
// executing core?
if (same_slice(hit_histogram)) {
slice = current_core;
}
// Current core has been used (do not remove)
core_used[current_core] = 1;
}
}
// Pretty print
if (print) {
print_bin(paddr);
printf(" %d\n", slice);
}
return slice;
}
int monitor_single_address_core(uintptr_t addr, int print) {
int i;
// Disable counters
uint64_t val[] = {val_disable_ctrs};
wrmsr_on_cpu_0(msr_unc_perf_global_ctr, 1, val);
// Reset counters
val[0] = val_reset_ctrs;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_unc_cbo_per_ctr0[i], 1, val);
}
// Select event to monitor
val[0] = val_select_evt_core;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_unc_cbo_perfevtsel0[i], 1, val);
}
// Enable counting
val[0] = val_enable_ctrs;
// val[0] = val_enable_ctrs;
wrmsr_on_cpu_0(msr_unc_perf_global_ctr, 1, val);
// Launch program to monitor
uintptr_t paddr;
paddr = poke(addr);
/*
// Disable counters
val[0] = val_disable_ctrs;
wrmsr_on_cpu_0(msr_unc_perf_global_ctr,1,val);
*/
// Read counter
int *cboxes = calloc(max_slices, sizeof(int));
int *cboxes_tri = calloc(max_slices, sizeof(int));
int res_temp;
for (i = 0; i < nb_cores; i++) {
res_temp = rdmsr_on_cpu_0(msr_unc_cbo_per_ctr0[i]);
cboxes[i] = ((res_temp - nb_pokes) < 0) ? 0 : res_temp - nb_pokes;
cboxes_tri[i] = ((res_temp - nb_pokes) < 0) ? 0 : res_temp - nb_pokes;
}
int slice = 0;
int first = 0;
int second = 0;
float percent;
// Interpreting the results
//
// Finding the slice in which the address is
for (i = 0; i < max_slices; i++) {
if (cboxes[i] > cboxes[slice]) {
slice = i;
}
}
// Calculate the ratio between the first and the second result
// to estimate the error
quicksort(cboxes_tri, 0, max_slices - 1);
first = cboxes_tri[max_slices - 1];
second = cboxes_tri[max_slices - 2];
percent = ((float)second) / ((float)first) * 100;
// Pretty print
if (print) {
print_bin(paddr);
printf(" %d %6.2f", slice, percent);
for (i = 0; i < nb_cores; i++) {
printf(" % 6d", cboxes[i]);
}
printf("\n");
}
free(cboxes);
free(cboxes_tri);
return slice;
}
int monitor_single_address_fast(uintptr_t addr) {
// Session monitoring
//
// The whole setup is explained in the section 2.1.2 of the manual (p15)
// Beware: it is written to reset all counters after enabling monitoring and
// selecting event to monitor, while the reset should be done before
int i;
// Freeze box counters
uint64_t val[] = {val_box_freeze};
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctr0[i], 1, val);
}
// Reset counters
val[0] = val_box_reset;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Enable counting
val[0] = val_enable_counting;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctl0[i], 1, val);
}
// Select event to monitor: umask and filter
val[0] = val_select_event;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctl0[i], 1, val);
}
val[0] = val_filter;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_filter[i], 1, val);
}
// Unfreezing box counters
val[0] = val_box_unfreeze;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Launch program to monitor
uintptr_t paddr;
paddr = poke(addr);
// Freeze box counters
val[0] = val_box_freeze;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Read counters
//
int *cboxes = calloc(max_slices, sizeof(int));
int *cboxes_tri = calloc(max_slices, sizeof(int));
int res_temp;
for (i = 0; i < nb_cores; i++) {
res_temp = rdmsr_on_cpu_0(msr_pmon_ctr0[i]);
cboxes[i] = ((res_temp - nb_pokes) < 0) ? 0 : res_temp - nb_pokes;
cboxes_tri[i] = ((res_temp - nb_pokes) < 0) ? 0 : res_temp - nb_pokes;
}
int slice = 0;
int first = 0;
int second = 0;
float percent;
// Interpreting the results
//
// Finding the slice in which the address is
for (i = 0; i < max_slices; i++) {
if (cboxes[i] > cboxes[slice]) {
slice = i;
}
}
// Calculate the ratio between the first and the second result
// to estimate the error
quicksort(cboxes_tri, 0, max_slices - 1);
first = cboxes_tri[max_slices - 1];
second = cboxes_tri[max_slices - 2];
percent = ((float)second) / ((float)first) * 100;
// Pretty print
print_bin(paddr);
printf(" %d %6.2f", slice, percent);
for (i = 0; i < nb_cores; i++) {
printf(" % 6d", cboxes[i]);
}
printf("\n");
free(cboxes);
free(cboxes_tri);
return slice;
}
int monitor_single_address_print(uintptr_t addr) {
// Session monitoring
//
// The whole setup is explained in the section 2.1.2 of the manual (p15)
// Beware: it is written to reset all counters after enabling monitoring and
// selecting event to monitor, while the reset should be done before
int i;
// Freeze box counters
uint64_t val[] = {val_box_freeze};
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctr0[i], 1, val);
}
// Reset counters
val[0] = val_box_reset;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Enable counting
val[0] = val_enable_counting;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctl0[i], 1, val);
}
// Select event to monitor: umask and filter
val[0] = val_select_event;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctl0[i], 1, val);
}
val[0] = val_filter;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_filter[i], 1, val);
}
// Unfreezing box counters
val[0] = val_box_unfreeze;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Launch program to monitor
uintptr_t paddr;
paddr = poke(addr);
// Freeze box counters
val[0] = val_box_freeze;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Read counters
//
int *cboxes = calloc(max_slices, sizeof(int));
int *cboxes_tri = calloc(max_slices, sizeof(int));
for (i = 0; i < nb_cores; i++) {
cboxes[i] = rdmsr_on_cpu_0(msr_pmon_ctr0[i]);
cboxes_tri[i] = rdmsr_on_cpu_0(msr_pmon_ctr0[i]);
}
int slice = 0;
int first = 0;
int second = 0;
float percent;
// Interpreting the results
//
// Finding the slice in which the address is
for (i = 0; i < max_slices; i++) {
if (cboxes[i] > cboxes[slice]) {
slice = i;
}
}
// Calculate the ratio between the first and the second result
// to estimate the error
quicksort(cboxes_tri, 0, max_slices - 1);
first = cboxes_tri[max_slices - 1];
second = cboxes_tri[max_slices - 2];
percent = ((float)second) / ((float)first) * 100;
// Pretty print
print_bin(paddr);
printf(" %d %6.2f", slice, percent);
for (i = 0; i < nb_cores; i++) {
printf(" % 6d", cboxes[i]);
}
printf("\n");
free(cboxes);
free(cboxes_tri);
return slice;
}
int monitor_single_address(uintptr_t addr) {
// Session monitoring
//
// The whole setup is explained in the section 2.1.2 of the manual (p15)
// Beware: it is written to reset all counters after enabling monitoring and
// selecting event to monitor, while the reset should be done before
int i;
// Freeze box counters
uint64_t val[] = {val_box_freeze};
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctr0[i], 1, val);
}
// Reset counters
val[0] = val_box_reset;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Enable counting
val[0] = val_enable_counting;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctl0[i], 1, val);
}
// Select event to monitor: umask and filter
val[0] = val_select_event;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_ctl0[i], 1, val);
}
val[0] = val_filter;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_filter[i], 1, val);
}
// Unfreezing box counters
val[0] = val_box_unfreeze;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Launch program to monitor
poke(addr);
// Freeze box counters
val[0] = val_box_freeze;
for (i = 0; i < nb_cores; i++) {
wrmsr_on_cpu_0(msr_pmon_box_ctl[i], 1, val);
}
// Read counters
//
int *cboxes = calloc(max_slices, sizeof(int));
int *cboxes_tri = calloc(max_slices, sizeof(int));
for (i = 0; i < nb_cores; i++) {
cboxes[i] = rdmsr_on_cpu_0(msr_pmon_ctr0[i]);
cboxes_tri[i] = rdmsr_on_cpu_0(msr_pmon_ctr0[i]);
}
// Interpreting the results
//
// Finding the slice in which the address is
int slice = 0;
for (i = 0; i < max_slices; i++) {
if (cboxes[i] > cboxes[slice]) {
slice = i;
}
}
free(cboxes);
free(cboxes_tri);
return slice;
}