src/test_cache_digest.cc

   1
   2 /*
   3  * $Id$
   4  *
   5  * AUTHOR: Alex Rousskov
   6  *
   7  * SQUID Web Proxy Cache          http://www.squid-cache.org/
   8  * ----------------------------------------------------------
   9  *
  10  *  Squid is the result of efforts by numerous individuals from
  11  *  the Internet community; see the CONTRIBUTORS file for full
  12  *  details.   Many organizations have provided support for Squid's
  13  *  development; see the SPONSORS file for full details.  Squid is
  14  *  Copyrighted (C) 2001 by the Regents of the University of
  15  *  California; see the COPYRIGHT file for full details.  Squid
  16  *  incorporates software developed and/or copyrighted by other
  17  *  sources; see the CREDITS file for full details.
  18  *
  19  *  This program is free software; you can redistribute it and/or modify
  20  *  it under the terms of the GNU General Public License as published by
  21  *  the Free Software Foundation; either version 2 of the License, or
  22  *  (at your option) any later version.
  23  *
  24  *  This program is distributed in the hope that it will be useful,
  25  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  26  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  27  *  GNU General Public License for more details.
  28  *
  29  *  You should have received a copy of the GNU General Public License
  30  *  along with this program; if not, write to the Free Software
  31  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  32  *
  33  */
  34
  35 /*
  36  * Test-suite for playing with cache digests
  37  */
  38
  39 #include "squid.h"
  40
  41 #if HAVE_ERRNO_H
  42 #include <errno.h>
  43 #endif
  44
  45 typedef struct {
  46     int query_count;
  47     int true_hit_count;
  48     int true_miss_count;
  49     int false_hit_count;
  50     int false_miss_count;
  51 } CacheQueryStats;
  52
  53 typedef struct _Cache Cache;
  54
  55 struct _Cache {
  56     const char *name;
  57     hash_table *hash;
  58     CacheDigest *digest;
  59     Cache *peer;
  60     CacheQueryStats qstats;
  61     int count;                  /* #currently cached entries */
  62     int req_count;              /* #requests to this cache */
  63     int bad_add_count;          /* #duplicate adds */
  64     int bad_del_count;          /* #dels with no prior add */
  65 };
  66
  67
  68 typedef struct _CacheEntry {
  69     const cache_key *key;
  70
  71     struct _CacheEntry *next;
  72     unsigned char key_arr[SQUID_MD5_DIGEST_LENGTH];
  73     /* storeSwapLogData s; */
  74 } CacheEntry;
  75
  76 /* parsed access log entry */
  77
  78 typedef struct {
  79     cache_key key[SQUID_MD5_DIGEST_LENGTH];
  80     time_t timestamp;
  81     short int use_icp;          /* true/false */
  82 } RawAccessLogEntry;
  83
  84 typedef enum {
  85     frError = -2, frMore = -1, frEof = 0, frOk = 1
  86 } fr_result;
  87
  88 typedef struct _FileIterator FileIterator;
  89 typedef fr_result(*FI_READER) (FileIterator * fi);
  90
  91 struct _FileIterator {
  92     const char *fname;
  93     FILE *file;
  94     time_t inner_time;          /* timestamp of the current entry */
  95     time_t time_offset;         /* to adjust time set by reader */
  96     int line_count;             /* number of lines scanned */
  97     int bad_line_count;         /* number of parsing errors */
  98     int time_warp_count;        /* number of out-of-order entries in the file */
  99     FI_READER reader;           /* reads next entry and updates inner_time */
 100     void *entry;                /* buffer for the current entry, freed with xfree() */
 101 };
 102
 103 /* globals */
 104 static time_t cur_time = -1;    /* timestamp of the current log entry */
 105
 106 /* copied from url.c */
 107 const char *RequestMethodStr[] = {
 108     "NONE",
 109     "GET",
 110     "POST",
 111     "PUT",
 112     "HEAD",
 113     "CONNECT",
 114     "TRACE",
 115     "PURGE"
 116 };
 117
 118 /* copied from url.c */
 119 static HttpRequestMethod
 120 methodStrToId(const char *s)
 121 {
 122     if (strcasecmp(s, "GET") == 0) {
 123         return METHOD_GET;
 124     } else if (strcasecmp(s, "POST") == 0) {
 125         return METHOD_POST;
 126     } else if (strcasecmp(s, "PUT") == 0) {
 127         return METHOD_PUT;
 128     } else if (strcasecmp(s, "HEAD") == 0) {
 129         return METHOD_HEAD;
 130     } else if (strcasecmp(s, "CONNECT") == 0) {
 131         return METHOD_CONNECT;
 132     } else if (strcasecmp(s, "TRACE") == 0) {
 133         return METHOD_TRACE;
 134     } else if (strcasecmp(s, "PURGE") == 0) {
 135         return METHOD_PURGE;
 136     }
 137
 138     return METHOD_NONE;
 139 }
 140
 141 /* FileIterator */
 142
 143 static void fileIteratorAdvance(FileIterator * fi);
 144
 145 static FileIterator *
 146 fileIteratorCreate(const char *fname, FI_READER reader)
 147 {
 148     FileIterator *fi = (FileIterator *)xcalloc(1, sizeof(FileIterator));
 149     assert(fname && reader);
 150     fi->fname = fname;
 151     fi->reader = reader;
 152     fi->file = fopen(fname, "r");
 153
 154     if (!fi->file) {
 155         fprintf(stderr, "cannot open %s: %s\n", fname, strerror(errno));
 156         return NULL;
 157     } else
 158         fprintf(stderr, "opened %s\n", fname);
 159
 160     fileIteratorAdvance(fi);
 161
 162     return fi;
 163 }
 164
 165 static void
 166 fileIteratorDestroy(FileIterator * fi)
 167 {
 168     assert(fi);
 169
 170     if (fi->file) {
 171         fclose(fi->file);
 172         fprintf(stderr, "closed %s\n", fi->fname);
 173     }
 174
 175     xfree(fi->entry);
 176     xfree(fi);
 177 }
 178
 179 static void
 180 fileIteratorSetCurTime(FileIterator * fi, time_t ct)
 181 {
 182     assert(fi);
 183     assert(fi->inner_time > 0);
 184     fi->time_offset = ct - fi->inner_time;
 185 }
 186
 187 static void
 188 fileIteratorAdvance(FileIterator * fi)
 189 {
 190     int res;
 191     assert(fi);
 192
 193     do {
 194         const time_t last_time = fi->inner_time;
 195         fi->inner_time = -1;
 196         res = fi->reader(fi);
 197         ++ fi->line_count;
 198
 199         if (fi->inner_time < 0)
 200             fi->inner_time = last_time;
 201         else
 202             fi->inner_time += fi->time_offset;
 203
 204         if (res == frError)
 205             ++ fi->bad_line_count;
 206         else if (res == frEof) {
 207             fprintf(stderr, "exhausted %s (%d entries) at %s",
 208                     fi->fname, fi->line_count, ctime(&fi->inner_time));
 209             fi->inner_time = -1;
 210         } else if (fi->inner_time < last_time) {
 211             assert(last_time >= 0);
 212             ++ fi->time_warp_count;
 213             fi->inner_time = last_time;
 214         }
 215
 216         /* report progress */
 217         if (!(fi->line_count % 50000))
 218             fprintf(stderr, "%s scanned %d K entries (%d bad) at %s",
 219                     fi->fname, fi->line_count / 1000, fi->bad_line_count,
 220                     ctime(&fi->inner_time));
 221     } while (res < 0);
 222 }
 223
 224 /* CacheEntry */
 225
 226 static CacheEntry *
 227 cacheEntryCreate(const storeSwapLogData * s)
 228 {
 229     CacheEntry *e = (CacheEntry *)xcalloc(1, sizeof(CacheEntry));
 230     assert(s);
 231     /* e->s = *s; */
 232     memcpy(e->key_arr, s->key, SQUID_MD5_DIGEST_LENGTH);
 233     e->key = &e->key_arr[0];
 234     return e;
 235 }
 236
 237 static void
 238 cacheEntryDestroy(CacheEntry * e)
 239 {
 240     assert(e);
 241     xfree(e);
 242 }
 243
 244
 245 /* Cache */
 246
 247 static Cache *
 248 cacheCreate(const char *name)
 249 {
 250     Cache *c;
 251     assert(name && strlen(name));
 252     c = (Cache *)xcalloc(1, sizeof(Cache));
 253     c->name = name;
 254     c->hash = hash_create(storeKeyHashCmp, (int)2e6, storeKeyHashHash);
 255     return c;
 256 }
 257
 258 static void
 259 cacheDestroy(Cache * cache)
 260 {
 261     CacheEntry *e = NULL;
 262     hash_table *hash;
 263     assert(cache);
 264     hash = cache->hash;
 265     /* destroy hash table contents */
 266     hash_first(hash);
 267
 268     while ((e = (CacheEntry *)hash_next(hash))) {
 269         hash_remove_link(hash, (hash_link *) e);
 270         cacheEntryDestroy(e);
 271     }
 272
 273     /* destroy the hash table itself */
 274     hashFreeMemory(hash);
 275
 276     if (cache->digest)
 277         cacheDigestDestroy(cache->digest);
 278
 279     xfree(cache);
 280 }
 281
 282 /* re-digests currently hashed entries */
 283 static void
 284 cacheResetDigest(Cache * cache)
 285 {
 286     CacheEntry *e = NULL;
 287     hash_table *hash;
 288
 289     struct timeval t_start, t_end;
 290
 291     assert(cache);
 292     fprintf(stderr, "%s: init-ing digest with %d entries\n", cache->name, cache->count);
 293
 294     if (cache->digest)
 295         cacheDigestDestroy(cache->digest);
 296
 297     hash = cache->hash;
 298
 299     cache->digest = cacheDigestCreate(cache->count + 1, 6);
 300
 301     if (!cache->count)
 302         return;
 303
 304     gettimeofday(&t_start, NULL);
 305
 306     hash_first(hash);
 307
 308     while ((e = (CacheEntry *)hash_next(hash))) {
 309         cacheDigestAdd(cache->digest, e->key);
 310     }
 311
 312     gettimeofday(&t_end, NULL);
 313     assert(cache->digest->count == cache->count);
 314     fprintf(stderr, "%s: init-ed  digest with %d entries\n",
 315             cache->name, cache->digest->count);
 316     fprintf(stderr, "%s: init took: %f sec, %f sec/M\n",
 317             cache->name,
 318             tvSubDsec(t_start, t_end),
 319             (double) 1e6 * tvSubDsec(t_start, t_end) / cache->count);
 320     /* check how long it takes to traverse the hash */
 321     gettimeofday(&t_start, NULL);
 322     hash_first(hash);
 323
 324     for (e = (CacheEntry *)hash_next(hash); e; e = (CacheEntry *)hash_next(hash)) {}
 325
 326     gettimeofday(&t_end, NULL);
 327     fprintf(stderr, "%s: hash scan took: %f sec, %f sec/M\n",
 328             cache->name,
 329             tvSubDsec(t_start, t_end),
 330             (double) 1e6 * tvSubDsec(t_start, t_end) / cache->count);
 331 }
 332
 333 static void
 334 cacheQueryPeer(Cache * cache, const cache_key * key)
 335 {
 336     const int peer_has_it = hash_lookup(cache->peer->hash, key) != NULL;
 337     const int we_think_we_have_it = cacheDigestTest(cache->digest, key);
 338
 339     ++ cache->qstats.query_count;
 340
 341     if (peer_has_it) {
 342         if (we_think_we_have_it)
 343             ++ cache->qstats.true_hit_count;
 344         else
 345             ++ cache->qstats.false_miss_count;
 346     } else {
 347         if (we_think_we_have_it)
 348             ++ cache->qstats.false_hit_count;
 349         else
 350             ++ cache->qstats.true_miss_count;
 351     }
 352 }
 353
 354 static void
 355 cacheQueryReport(Cache * cache, CacheQueryStats * stats)
 356 {
 357     fprintf(stdout, "%s: peer queries: %d (%d%%)\n",
 358             cache->name,
 359             stats->query_count, xpercentInt(stats->query_count, cache->req_count)
 360            );
 361     fprintf(stdout, "%s: t-hit: %d (%d%%) t-miss: %d (%d%%) t-*: %d (%d%%)\n",
 362             cache->name,
 363             stats->true_hit_count, xpercentInt(stats->true_hit_count, stats->query_count),
 364             stats->true_miss_count, xpercentInt(stats->true_miss_count, stats->query_count),
 365             stats->true_hit_count + stats->true_miss_count,
 366             xpercentInt(stats->true_hit_count + stats->true_miss_count, stats->query_count)
 367            );
 368     fprintf(stdout, "%s: f-hit: %d (%d%%) f-miss: %d (%d%%) f-*: %d (%d%%)\n",
 369             cache->name,
 370             stats->false_hit_count, xpercentInt(stats->false_hit_count, stats->query_count),
 371             stats->false_miss_count, xpercentInt(stats->false_miss_count, stats->query_count),
 372             stats->false_hit_count + stats->false_miss_count,
 373             xpercentInt(stats->false_hit_count + stats->false_miss_count, stats->query_count)
 374            );
 375 }
 376
 377 static void
 378 cacheReport(Cache * cache)
 379 {
 380     fprintf(stdout, "%s: entries: %d reqs: %d bad-add: %d bad-del: %d\n",
 381             cache->name, cache->count, cache->req_count,
 382             cache->bad_add_count, cache->bad_del_count);
 383
 384 }
 385
 386 static void
 387 cacheFetch(Cache * cache, const RawAccessLogEntry * e)
 388 {
 389     assert(e);
 390     ++ cache->req_count;
 391
 392     if (e->use_icp)
 393         cacheQueryPeer(cache, e->key);
 394 }
 395
 396 static fr_result
 397 swapStateReader(FileIterator * fi)
 398 {
 399     storeSwapLogData *entry;
 400
 401     if (!fi->entry)
 402         fi->entry = xcalloc(1, sizeof(storeSwapLogData));
 403
 404     entry = (storeSwapLogData *)fi->entry;
 405
 406     if (fread(entry, sizeof(*entry), 1, fi->file) != 1)
 407         return frEof;
 408
 409     fi->inner_time = entry->lastref;
 410
 411     if (entry->op != SWAP_LOG_ADD && entry->op != SWAP_LOG_DEL) {
 412         fprintf(stderr, "%s:%d: unknown swap log action\n", fi->fname, fi->line_count);
 413         exit(-3);
 414     }
 415
 416     return frOk;
 417 }
 418
 419 static fr_result
 420 accessLogReader(FileIterator * fi)
 421 {
 422     static char buf[4096];
 423     RawAccessLogEntry *entry;
 424     char *url;
 425     char *method;
 426     HttpRequestMethod method_id = METHOD_NONE;
 427     char *hier = NULL;
 428
 429     assert(fi);
 430
 431     if (!fi->entry)
 432         fi->entry = xcalloc(1, sizeof(RawAccessLogEntry));
 433     else
 434         memset(fi->entry, 0, sizeof(RawAccessLogEntry));
 435
 436     entry = (RawAccessLogEntry*)fi->entry;
 437
 438     if (!fgets(buf, sizeof(buf), fi->file))
 439         return frEof;           /* eof */
 440
 441     entry->timestamp = fi->inner_time = (time_t) atoi(buf);
 442
 443     url = strstr(buf, "://");
 444
 445     hier = url ? strstr(url, " - ") : NULL;
 446
 447     if (!url || !hier) {
 448         /*fprintf(stderr, "%s:%d: strange access log entry '%s'\n",
 449          * fname, scanned_count, buf); */
 450         return frError;
 451     }
 452
 453     method = url;
 454
 455     while (!xisdigit(*method)) {
 456         if (*method == ' ')
 457             *method = '\0';
 458
 459         --method;
 460     }
 461
 462     method += 2;
 463     method_id = methodStrToId(method);
 464
 465     if (method_id == METHOD_NONE) {
 466         /*fprintf(stderr, "%s:%d: invalid method %s in '%s'\n",
 467          * fname, scanned_count, method, buf); */
 468         return frError;
 469     }
 470
 471     while (*url)
 472         --url;
 473
 474     ++url;
 475
 476     *hier = '\0';
 477
 478     hier += 3;
 479
 480     *strchr(hier, '/') = '\0';
 481
 482     /*fprintf(stdout, "%s:%d: %s %s %s\n",
 483      * fname, count, method, url, hier); */
 484     entry->use_icp = strcmp(hier, "NONE");
 485
 486     /* no ICP lookup for these status codes */
 487     /*      strcmp(hier, "NONE") &&
 488      * strcmp(hier, "DIRECT") &&
 489      * strcmp(hier, "FIREWALL_IP_DIRECT") &&
 490      * strcmp(hier, "LOCAL_IP_DIRECT") &&
 491      * strcmp(hier, "NO_DIRECT_FAIL") &&
 492      * strcmp(hier, "NO_PARENT_DIRECT") &&
 493      * strcmp(hier, "SINGLE_PARENT") &&
 494      * strcmp(hier, "PASSTHROUGH_PARENT") &&
 495      * strcmp(hier, "SSL_PARENT_MISS") &&
 496      * strcmp(hier, "DEFAULT_PARENT");
 497      */
 498     memcpy(entry->key, storeKeyPublic(url, method_id), sizeof(entry->key));
 499
 500     /*fprintf(stdout, "%s:%d: %s %s %s %s\n",
 501      * fname, count, method, storeKeyText(entry->key), url, hier); */
 502     return frOk;
 503 }
 504
 505
 506 static void
 507 cachePurge(Cache * cache, storeSwapLogData * s, int update_digest)
 508 {
 509     CacheEntry *olde = (CacheEntry *) hash_lookup(cache->hash, s->key);
 510
 511     if (!olde) {
 512         ++ cache->bad_del_count;
 513     } else {
 514         assert(cache->count);
 515         hash_remove_link(cache->hash, (hash_link *) olde);
 516
 517         if (update_digest)
 518             cacheDigestDel(cache->digest, s->key);
 519
 520         cacheEntryDestroy(olde);
 521
 522         -- cache->count;
 523     }
 524 }
 525
 526 static void
 527 cacheStore(Cache * cache, storeSwapLogData * s, int update_digest)
 528 {
 529     CacheEntry *olde = (CacheEntry *) hash_lookup(cache->hash, s->key);
 530
 531     if (olde) {
 532         ++ cache->bad_add_count;
 533     } else {
 534         CacheEntry *e = cacheEntryCreate(s);
 535         hash_join(cache->hash, (hash_link *)&e->key);
 536         ++ cache->count;
 537
 538         if (update_digest)
 539             cacheDigestAdd(cache->digest, e->key);
 540     }
 541 }
 542
 543 static void
 544 cacheUpdateStore(Cache * cache, storeSwapLogData * s, int update_digest)
 545 {
 546     switch (s->op) {
 547
 548     case SWAP_LOG_ADD:
 549         cacheStore(cache, s, update_digest);
 550         break;
 551
 552     case SWAP_LOG_DEL:
 553         cachePurge(cache, s, update_digest);
 554         break;
 555
 556     default:
 557         assert(0);
 558     }
 559 }
 560
 561 static int
 562 usage(const char *prg_name)
 563 {
 564     fprintf(stderr, "usage: %s <access_log> <swap_state> ...\n",
 565             prg_name);
 566     return -1;
 567 }
 568
 569 int
 570 main(int argc, char *argv[])
 571 {
 572     FileIterator **fis = NULL;
 573     const int fi_count = argc - 1;
 574     int active_fi_count = 0;
 575     time_t ready_time;
 576     Cache *them, *us;
 577     int i;
 578
 579     if (argc < 3)
 580         return usage(argv[0]);
 581
 582     them = cacheCreate("them");
 583
 584     us = cacheCreate("us");
 585
 586     them->peer = us;
 587
 588     us->peer = them;
 589
 590     fis = (FileIterator **)xcalloc(fi_count, sizeof(FileIterator *));
 591
 592     /* init iterators with files */
 593     fis[0] = fileIteratorCreate(argv[1], accessLogReader);
 594
 595     for (i = 2; i < argc; ++i)
 596         fis[i - 1] = fileIteratorCreate(argv[i], swapStateReader);
 597
 598     /* check that all files were found */
 599     for (i = 0; i < fi_count; ++i)
 600         if (!fis[i])
 601             return -2;
 602
 603     /* read prefix to get start-up contents of the peer cache */
 604     ready_time = -1;
 605
 606     for (i = 1; i < fi_count; ++i) {
 607         FileIterator *fi = fis[i];
 608
 609         while (fi->inner_time > 0) {
 610             if (((storeSwapLogData *) fi->entry)->op == SWAP_LOG_DEL) {
 611                 cachePurge(them, (storeSwapLogData *)fi->entry, 0);
 612
 613                 if (ready_time < 0)
 614                     ready_time = fi->inner_time;
 615             } else {
 616                 if (ready_time > 0 && fi->inner_time > ready_time)
 617                     break;
 618
 619                 cacheStore(them, (storeSwapLogData *)fi->entry, 0);
 620             }
 621
 622             fileIteratorAdvance(fi);
 623         }
 624     }
 625
 626     /* digest peer cache content */
 627     cacheResetDigest(them);
 628
 629     us->digest = cacheDigestClone(them->digest);        /* @netw@ */
 630
 631     /* shift the time in access log to match ready_time */
 632     fileIteratorSetCurTime(fis[0], ready_time);
 633
 634     /* iterate, use the iterator with the smallest positive inner_time */
 635     cur_time = -1;
 636
 637     do {
 638         int next_i = -1;
 639         time_t next_time = -1;
 640         active_fi_count = 0;
 641
 642         for (i = 0; i < fi_count; ++i) {
 643             if (fis[i]->inner_time >= 0) {
 644                 if (!active_fi_count || fis[i]->inner_time < next_time) {
 645                     next_i = i;
 646                     next_time = fis[i]->inner_time;
 647                 }
 648
 649                 ++active_fi_count;
 650             }
 651         }
 652
 653         if (next_i >= 0) {
 654             cur_time = next_time;
 655             /*fprintf(stderr, "%2d time: %d %s", next_i, (int)cur_time, ctime(&cur_time)); */
 656
 657             if (next_i == 0)
 658                 cacheFetch(us, (RawAccessLogEntry *)fis[next_i]->entry);
 659             else
 660                 cacheUpdateStore(them, (storeSwapLogData *)fis[next_i]->entry, 1);
 661
 662             fileIteratorAdvance(fis[next_i]);
 663         }
 664     } while (active_fi_count);
 665
 666     /* report */
 667     cacheReport(them);
 668
 669     cacheReport(us);
 670
 671     cacheQueryReport(us, &us->qstats);
 672
 673     /* clean */
 674     for (i = 0; i < argc - 1; ++i) {
 675         fileIteratorDestroy(fis[i]);
 676     }
 677
 678     xfree(fis);
 679     cacheDestroy(them);
 680     cacheDestroy(us);
 681     return 0;
 682 }