forked from pagekite/libpagekite
-
Notifications
You must be signed in to change notification settings - Fork 0
/
pkblocker.c
572 lines (502 loc) · 17.2 KB
/
pkblocker.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
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
/******************************************************************************
pkblocker.c - Blocking tasks handled outside the main event loop.
This file is Copyright 2011-2013, The Beanstalks Project ehf.
This program is free software: you can redistribute it and/or modify it under
the terms of the Apache License 2.0 as published by the Apache Software
Foundation.
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 Apache License for more details.
You should have received a copy of the Apache License along with this program.
If not, see: <http://www.apache.org/licenses/>
Note: For alternate license terms, see the file COPYING.md.
******************************************************************************/
#include "common.h"
#include "utils.h"
#include "pkerror.h"
#include "pkstate.h"
#include "pkconn.h"
#include "pkproto.h"
#include "pkblocker.h"
#include "pkmanager.h"
#include "pklogging.h"
int pkb_add_job(struct pk_job_pile* pkj, pk_job_t job, void* data)
{
int i;
PK_TRACE_FUNCTION;
pthread_mutex_lock(&(pkj->mutex));
for (i = 0; i < pkj->max; i++) {
if ((pkj->pile+i)->job == PK_NO_JOB) {
(pkj->pile+i)->job = job;
(pkj->pile+i)->data = data;
pkj->count += 1;
pthread_cond_signal(&(pkj->cond));
pthread_mutex_unlock(&(pkj->mutex));
return 1;
}
}
pthread_mutex_unlock(&(pkj->mutex));
return -1;
}
int pkb_get_job(struct pk_job_pile* pkj, struct pk_job* dest)
{
int i;
PK_TRACE_FUNCTION;
pthread_mutex_lock(&(pkj->mutex));
while (pkj->count == 0)
pthread_cond_wait(&(pkj->cond), &(pkj->mutex));
for (i = 0; i < pkj->max; i++) {
if ((pkj->pile+i)->job != PK_NO_JOB) {
dest->job = (pkj->pile+i)->job;
dest->data = (pkj->pile+i)->data;
(pkj->pile+i)->job = PK_NO_JOB;
(pkj->pile+i)->data = NULL;
pkj->count -= 1;
pthread_mutex_unlock(&(pkj->mutex));
return 1;
}
}
dest->job = PK_NO_JOB;
dest->data = NULL;
pthread_mutex_unlock(&(pkj->mutex));
return -1;
}
void pkb_clear_transient_flags(struct pk_manager* pkm)
{
int i;
struct pk_frontend* fe;
PK_TRACE_FUNCTION;
for (i = 0, fe = pkm->frontends; i < pkm->frontend_max; i++, fe++) {
fe->conn.status &= ~FE_STATUS_REJECTED;
fe->conn.status &= ~FE_STATUS_LAME;
fe->conn.status &= ~FE_STATUS_IS_FAST;
fe->conn.status &= ~FE_STATUS_IN_DNS;
}
}
void pkb_choose_frontends(struct pk_manager* pkm)
{
int i, wanted, wantn, highpri, prio;
struct pk_frontend* fe;
struct pk_frontend* highpri_fe;
PK_TRACE_FUNCTION;
/* Clear WANTED flag... */
for (i = 0, fe = pkm->frontends; i < pkm->frontend_max; i++, fe++) {
if (fe->ai) fe->conn.status &= ~(FE_STATUS_WANTED|FE_STATUS_IS_FAST);
}
/* Choose N fastest: this is inefficient, but trivially correct. */
for (wantn = 0; wantn < pkm->want_spare_frontends+1; wantn++) {
highpri_fe = NULL;
highpri = 1024000;
for (i = 0, fe = pkm->frontends; i < pkm->frontend_max; i++, fe++) {
prio = fe->priority + (25 * fe->error_count);
if ((fe->ai) &&
(fe->priority) &&
((highpri_fe == NULL) || (highpri > prio)) &&
(!(fe->conn.status & (FE_STATUS_IS_FAST
|FE_STATUS_REJECTED
|FE_STATUS_LAME)))) {
highpri_fe = fe;
highpri = prio;
}
}
if (highpri_fe != NULL)
highpri_fe->conn.status |= FE_STATUS_IS_FAST;
}
wanted = 0;
for (i = 0, fe = pkm->frontends; i < pkm->frontend_max; i++, fe++) {
/* If it's nailed up or fast: we want it. */
if ((fe->conn.status & FE_STATUS_NAILED_UP) ||
(fe->conn.status & FE_STATUS_IS_FAST)) {
fe->conn.status |= FE_STATUS_WANTED;
pk_log(PK_LOG_MANAGER_DEBUG,
"Fast or nailed up, should use %s (status=%x)",
fe->fe_hostname, fe->conn.status);
}
/* Otherwise, we don't! */
else {
fe->conn.status &= ~FE_STATUS_WANTED;
if (fe->conn.status & FE_STATUS_IN_DNS) {
pk_log(PK_LOG_MANAGER_DEBUG,
"Not wanted, but in DNS (fallback): %s (status=%x)",
fe->fe_hostname, fe->conn.status);
}
}
/* Rejecting us or going lame overrides other concerns. */
if ((fe->conn.status & FE_STATUS_REJECTED) ||
(fe->conn.status & FE_STATUS_LAME)) {
fe->conn.status &= ~FE_STATUS_WANTED;
pk_log(PK_LOG_MANAGER_DEBUG,
"Lame or rejecting, avoiding %s (status=%x)",
fe->fe_hostname, fe->conn.status);
}
/* Count how many we're aiming for. */
if (fe->conn.status & (FE_STATUS_WANTED|FE_STATUS_IN_DNS)) wanted++;
}
if (wanted) return;
/* None wanted? Uh oh, best accept anything non-broken at this point... */
for (i = 0, fe = pkm->frontends; i < pkm->frontend_max; i++, fe++) {
if ((fe->ai != NULL) &&
!(fe->conn.status & (FE_STATUS_REJECTED|FE_STATUS_LAME))) {
fe->conn.status |= FE_STATUS_WANTED;
wanted++;
pk_log(PK_LOG_MANAGER_INFO,
"No front-end wanted, randomly using %s (status=%x)",
fe->fe_hostname, fe->conn.status);
break;
}
}
if (wanted) return;
/* Still none? Crazy town. Maybe a good front-end has been marked as
* being lame because of duplicates and we've somehow forgotten it is
* in DNS? Let's at least not disconnect. */
for (i = 0, fe = pkm->frontends; i < pkm->frontend_max; i++, fe++) {
if ((fe->ai != NULL) &&
(fe->conn.sockfd > 0)) {
fe->conn.status |= FE_STATUS_WANTED;
wanted++;
pk_log(PK_LOG_MANAGER_INFO,
"No front-end wanted, keeping %s (status=%x)",
fe->fe_hostname, fe->conn.status);
}
}
if (wanted) return;
/* If we get this far, we're hopeless. Log as error. */
pk_log(PK_LOG_MANAGER_ERROR, "No front-end wanted! We are lame.");
}
void pkb_check_kites_dns(struct pk_manager* pkm)
{
int i, j, rv;
int in_dns = 0;
int recently_in_dns = 0;
time_t ddns_window;
struct pk_frontend* fe;
struct pk_frontend* dns_fe;
struct pk_pagekite* kite;
struct addrinfo hints;
struct addrinfo *result, *rp;
char buffer[128];
PK_TRACE_FUNCTION;
memset(&hints, 0, sizeof(struct addrinfo));
hints.ai_family = AF_UNSPEC;
hints.ai_socktype = SOCK_STREAM;
/* Clear DNS flag... */
for (j = 0, fe = pkm->frontends; j < pkm->frontend_max; j++, fe++) {
fe->conn.status &= ~FE_STATUS_IN_DNS;
}
/* Walk through kite list, look each up in DNS and update the
* frontend flags as appropriate.
*/
for (i = 0, kite = pkm->kites; i < pkm->kite_max; i++, kite++) {
rv = getaddrinfo(kite->public_domain, NULL, &hints, &result);
if (rv == 0) {
for (rp = result; rp != NULL; rp = rp->ai_next) {
for (j = 0, fe = pkm->frontends; j < pkm->frontend_max; j++, fe++) {
if (fe->ai) {
if (0 == addrcmp(fe->ai->ai_addr, rp->ai_addr)) {
pk_log(PK_LOG_MANAGER_DEBUG, "In DNS for %s: %s",
kite->public_domain,
in_ipaddr_to_str(fe->ai->ai_addr,
buffer, 128));
fe->conn.status |= FE_STATUS_IN_DNS;
fe->last_ddnsup = time(0);
in_dns++;
}
}
}
}
freeaddrinfo(result);
}
}
/* FIXME: We should really get this from the TTL of the DNS record itself,
* not from a hard coded magic number.
*/
ddns_window = time(0) - PK_DDNS_UPDATE_INTERVAL_MIN;
/* Walk through the list of frontends and rewnew the FE_STATUS_IN_DNS
* if they were either last updated within our window.
*/
dns_fe = NULL;
for (j = 0, fe = pkm->frontends; j < pkm->frontend_max; j++, fe++) {
if (fe->ai) {
if (fe->last_ddnsup > ddns_window) {
fe->conn.status |= FE_STATUS_IN_DNS;
in_dns++;
}
/* Figure out which FE was most recently seen in DNS, for use below */
if (fe->last_ddnsup > recently_in_dns) {
recently_in_dns = fe->last_ddnsup;
dns_fe = fe;
}
}
}
/* If nothing was found in DNS, but we know there was stuff in DNS
* before, then DNS is probably broken for us and the data in DNS is
* unchanged. Keep the most recent one active! This is incomplete if
* we are using many frontends at once, but still better than nothing.
*/
if (in_dns < 1 && dns_fe) {
dns_fe->conn.status |= FE_STATUS_IN_DNS;
}
}
void* pkb_frontend_ping(void* void_fe) {
struct pk_frontend* fe = (struct pk_frontend*) void_fe;
struct timeval tv1, tv2;
char buffer[1024], printip[1024];
int sockfd, bytes, want;
PK_TRACE_FUNCTION;
fe->priority = 0;
in_addr_to_str(fe->ai->ai_addr, printip, 1024);
if (pk_state.fake_ping) {
fe->priority = rand() % 500;
}
else {
gettimeofday(&tv1, NULL);
if ((0 > (sockfd = socket(fe->ai->ai_family, fe->ai->ai_socktype,
fe->ai->ai_protocol))) ||
(0 > connect(sockfd, fe->ai->ai_addr, fe->ai->ai_addrlen)) ||
(0 > write(sockfd, PK_FRONTEND_PING, strlen(PK_FRONTEND_PING))))
{
if (sockfd >= 0)
close(sockfd);
if (fe->error_count < 999)
fe->error_count += 1;
pk_log(PK_LOG_MANAGER_DEBUG, "Ping %s failed! (connect)", printip);
sleep(2); /* We don't want to return first! */
return NULL;
}
want = strlen(PK_FRONTEND_PONG);
bytes = timed_read(sockfd, buffer, want, 1000);
if ((bytes != want) ||
(0 != strncmp(buffer, PK_FRONTEND_PONG, want))) {
if (fe->error_count < 999)
fe->error_count += 1;
pk_log(PK_LOG_MANAGER_DEBUG, "Ping %s failed! (read=%d)", printip, bytes);
sleep(2); /* We don't want to return first! */
return NULL;
}
close(sockfd);
gettimeofday(&tv2, NULL);
fe->priority = (tv2.tv_sec - tv1.tv_sec) * 1000
+ (tv2.tv_usec - tv1.tv_usec) / 1000;
}
if (fe->conn.status & (FE_STATUS_WANTED|FE_STATUS_IS_FAST))
{
/* Bias ping time to make old decisions a bit more sticky. We ignore
* DNS though, to allow a bit of churn to spread the load around and
* make sure new frontends don't stay ignored forever. */
fe->priority /= 10;
fe->priority *= 9;
pk_log(PK_LOG_MANAGER_DEBUG,
"Ping %s: %dms (biased)", printip, fe->priority);
}
else {
/* Add artificial +/-5% jitter to ping results */
fe->priority *= ((rand() % 11) + 95);
fe->priority /= 100;
pk_log(PK_LOG_MANAGER_DEBUG, "Ping %s: %dms", printip, fe->priority);
}
return NULL;
}
void pkb_check_frontend_pingtimes(struct pk_manager* pkm)
{
int j;
struct pk_frontend* fe;
PK_TRACE_FUNCTION;
pthread_t first = 0;
pthread_t pt = 0;
for (j = 0, fe = pkm->frontends; j < pkm->frontend_max; j++, fe++) {
if (fe->ai) {
if (0 == pthread_create(&pt, NULL, pkb_frontend_ping, (void *) fe)) {
if (first)
pthread_detach(pt);
else
first = pt;
}
}
}
if (first) {
/* Sleep, but only wait for the first one - usually we only care about the
* fastest anyway. The others will return in their own good time.
*/
sleep(1);
pthread_join(first, NULL);
}
}
int pkb_update_dns(struct pk_manager* pkm)
{
int j, len, bogus, rlen;
struct pk_frontend* fe_list[1024]; /* Magic, bounded by address_list[] below */
struct pk_frontend** fes;
struct pk_frontend* fe;
struct pk_pagekite* kite;
char printip[128], get_result[10240], *result;
char address_list[1024], payload[2048], signature[2048], url[2048], *alp;
PK_TRACE_FUNCTION;
if (time(0) < pkm->last_dns_update + PK_DDNS_UPDATE_INTERVAL_MIN)
return 0;
address_list[0] = '\0';
alp = address_list;
fes = fe_list;
*fes = NULL;
bogus = 0;
for (j = 0, fe = pkm->frontends; j < pkm->frontend_max; j++, fe++) {
if ((fe->ai) && (fe->conn.sockfd >= 0)) {
if (fe->conn.status & FE_STATUS_WANTED) {
if (NULL != in_ipaddr_to_str(fe->ai->ai_addr, printip, 128)) {
len = strlen(printip);
if (len < 1000-(alp-address_list)) {
if (alp != address_list) *alp++ = ',';
strcpy(alp, printip);
alp += len;
*fes++ = fe;
*fes = NULL;
}
}
if (!(fe->conn.status & FE_STATUS_IN_DNS) || pk_state.force_update)
bogus++;
}
else /* Stuff in DNS that shouldn't be also triggers updates */
if (fe->conn.status & FE_STATUS_IN_DNS) bogus++;
}
}
if (!bogus) return 0;
if (!address_list[0]) return 0;
bogus = 0;
for (j = 0, kite = pkm->kites; j < pkm->kite_max; kite++, j++) {
if (kite->protocol[0] != '\0') {
PKS_STATE(pkm->status = PK_STATUS_DYNDNS);
sprintf(payload, "%s:%s", kite->public_domain, address_list);
pk_sign(NULL, kite->auth_secret, payload, 100, signature);
sprintf(url, pkm->dynamic_dns_url,
kite->public_domain, address_list, signature);
rlen = http_get(url, get_result, 10240);
if (rlen < 1) {
pk_log(PK_LOG_MANAGER_ERROR, "DDNS: No response from %s", url);
bogus++;
}
else {
result = skip_http_header(rlen, get_result);
if ((strncasecmp(result, "nochg", 5) == 0) ||
(strncasecmp(result, "good", 4) == 0)) {
pk_log(PK_LOG_MANAGER_INFO, "DDNS: Update OK, %s=%s",
kite->public_domain, address_list);
for (fes = fe_list; *fes; fes++) {
(*fes)->last_ddnsup = time(0);
(*fes)->conn.status |= FE_STATUS_IN_DNS;
}
}
else {
result[7] = '\0';
pk_log(PK_LOG_MANAGER_ERROR, "DDNS: Update failed for %s (%s -> %s)",
kite->public_domain, url, result);
bogus++;
}
}
}
}
pkm->last_dns_update = time(0);
return bogus;
}
void pkb_log_fe_status(struct pk_manager* pkm)
{
int j, ddnsup_ago;
struct pk_frontend* fe;
char printip[128];
char ddnsinfo[128];
PK_TRACE_FUNCTION;
for (j = 0, fe = pkm->frontends; j < pkm->frontend_max; j++, fe++) {
if (fe->ai) {
if (NULL != in_addr_to_str(fe->ai->ai_addr, printip, 128)) {
ddnsinfo[0] = '\0';
if (fe->last_ddnsup) {
ddnsup_ago = time(0) - fe->last_ddnsup;
sprintf(ddnsinfo, " (in dns %us ago)", ddnsup_ago);
}
pk_log(PK_LOG_MANAGER_DEBUG, "0x%8.8x E:%d %s%s%s",
fe->conn.status,
fe->error_count,
printip,
(fe->conn.sockfd > 0) ? " live" : "",
ddnsinfo);
}
}
}
}
void pkb_check_world(struct pk_manager* pkm)
{
PK_TRACE_FUNCTION;
if (pkm->status == PK_STATUS_NO_NETWORK) return;
pk_log(PK_LOG_MANAGER_DEBUG, "Checking state of world...");
pkb_clear_transient_flags(pkm);
pkb_check_kites_dns(pkm);
pkb_check_frontend_pingtimes(pkm);
pkb_log_fe_status(pkm);
pkm->last_world_update = time(0) + pkm->interval_fudge_factor;
}
void pkb_check_frontends(struct pk_manager* pkm)
{
int problems = 0;
PK_TRACE_FUNCTION;
if (pkm->status == PK_STATUS_NO_NETWORK) return;
pk_log(PK_LOG_MANAGER_DEBUG, "Checking frontends...");
pkb_check_kites_dns(pkm);
pkb_choose_frontends(pkm);
pkb_log_fe_status(pkm);
problems += pkm_reconnect_all(pkm);
if (!problems) pkm_disconnect_unused(pkm);
if (pkm->dynamic_dns_url && (pkm->status != PK_STATUS_REJECTED)) {
problems += pkb_update_dns(pkm);
}
/* An update has happened, clear this flag. */
pk_state.force_update = 0;
if (problems == 0 && pk_state.live_frontends > 0) {
PKS_STATE(pkm->status = PK_STATUS_FLYING);
}
else if (pkm->status != PK_STATUS_REJECTED) {
PKS_STATE(pkm->status = PK_STATUS_PROBLEMS);
}
}
void* pkb_run_blocker(void *void_pkm)
{
time_t last_check_world = 0;
time_t last_check_frontends = 0;
struct pk_job job;
struct pk_manager* pkm = (struct pk_manager*) void_pkm;
pk_log(PK_LOG_MANAGER_DEBUG, "Started blocking thread.");
while (1) {
pkb_get_job(&(pkm->blocking_jobs), &job);
switch (job.job) {
case PK_NO_JOB:
break;
case PK_CHECK_WORLD:
if (time(0) >= last_check_world + pkm->housekeeping_interval_min) {
pkb_check_world((struct pk_manager*) job.data);
pkb_check_frontends((struct pk_manager*) job.data);
last_check_world = last_check_frontends = time(0);
}
break;
case PK_CHECK_FRONTENDS:
if (time(0) >= last_check_frontends + pkm->housekeeping_interval_min) {
pkb_check_frontends((struct pk_manager*) job.data);
last_check_frontends = time(0);
}
break;
case PK_QUIT:
pk_log(PK_LOG_MANAGER_DEBUG, "Exiting blocking thread.");
return NULL;
}
}
}
int pkb_start_blocker(struct pk_manager *pkm)
{
if (0 > pthread_create(&(pkm->blocking_thread), NULL,
pkb_run_blocker, (void *) pkm)) {
pk_log(PK_LOG_MANAGER_ERROR, "Failed to start blocking thread.");
return (pk_error = ERR_NO_THREAD);
}
return 0;
}
void pkb_stop_blocker(struct pk_manager *pkm)
{
pkb_add_job(&(pkm->blocking_jobs), PK_QUIT, NULL);
pthread_join(pkm->blocking_thread, NULL);
}