src/test_cache_digest.cc

   1
   2 /*
   3  * $Id: test_cache_digest.cc,v 1.10 1998/04/01 07:07:17 rousskov Exp $
   4  *
   5  * AUTHOR: Alex Rousskov
   6  *
   7  * SQUID Internet Object Cache  http://squid.nlanr.net/Squid/
   8  * --------------------------------------------------------
   9  *
  10  *  Squid is the result of efforts by numerous individuals from the
  11  *  Internet community.  Development is led by Duane Wessels of the
  12  *  National Laboratory for Applied Network Research and funded by
  13  *  the National Science Foundation.
  14  *
  15  *  This program is free software; you can redistribute it and/or modify
  16  *  it under the terms of the GNU General Public License as published by
  17  *  the Free Software Foundation; either version 2 of the License, or
  18  *  (at your option) any later version.
  19  *
  20  *  This program is distributed in the hope that it will be useful,
  21  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  22  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  23  *  GNU General Public License for more details.
  24  *
  25  *  You should have received a copy of the GNU General Public License
  26  *  along with this program; if not, write to the Free Software
  27  *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  28  *
  29  */
  30
  31 /*
  32  * Test-suite for playing with cache digests
  33  */
  34
  35 #include "squid.h"
  36
  37 typedef struct {
  38     int query_count;
  39     int true_hit_count;
  40     int true_miss_count;
  41     int false_hit_count;
  42     int false_miss_count;
  43 } CacheQueryStats;
  44
  45 typedef struct _Cache Cache;
  46 struct _Cache {
  47     const char *name;
  48     hash_table *hash;
  49     CacheDigest *digest;
  50     Cache *peer;
  51     CacheQueryStats qstats;
  52     int count;                  /* #currently cached entries */
  53     int req_count;              /* #requests to this cache */
  54     int bad_add_count;          /* #duplicate adds */
  55     int bad_del_count;          /* #dels with no prior add */
  56 };
  57
  58
  59 typedef struct _CacheEntry {
  60     const cache_key *key;
  61     struct _CacheEntry *next;
  62     unsigned char key_arr[MD5_DIGEST_CHARS];
  63     /* storeSwapLogData s; */
  64 } CacheEntry;
  65
  66 /* parsed access log entry */
  67 typedef struct {
  68     cache_key key[MD5_DIGEST_CHARS];
  69     time_t timestamp;
  70     short int use_icp; /* true/false */
  71 } RawAccessLogEntry;
  72
  73 typedef enum { frError = -2, frMore = -1, frEof = 0, frOk = 1 } fr_result;
  74 typedef struct _FileIterator FileIterator;
  75 typedef fr_result (*FI_READER)(FileIterator *fi);
  76
  77 struct _FileIterator {
  78     const char *fname;
  79     FILE *file;
  80     time_t inner_time;  /* timestamp of the current entry */
  81     time_t time_offset;  /* to adjust time set by reader */
  82     int line_count;     /* number of lines scanned */
  83     int bad_line_count; /* number of parsing errors */
  84     int time_warp_count;/* number of out-of-order entries in the file */
  85     FI_READER reader;   /* reads next entry and updates inner_time */
  86     void *entry;        /* buffer for the current entry, freed with xfree() */
  87 };
  88
  89 /* globals */
  90 static time_t cur_time = -1; /* timestamp of the current log entry */
  91
  92 #if 0
  93
  94 static int cacheIndexScanCleanPrefix(CacheIndex * idx, const char *fname, FILE * file);
  95 static int cacheIndexScanAccessLog(CacheIndex * idx, const char *fname, FILE * file);
  96
  97 #endif
  98
  99 /* copied from url.c */
 100 const char *RequestMethodStr[] =
 101 {
 102     "NONE",
 103     "GET",
 104     "POST",
 105     "PUT",
 106     "HEAD",
 107     "CONNECT",
 108     "TRACE",
 109     "PURGE"
 110 };
 111 /* copied from url.c */
 112 static method_t
 113 methodStrToId(const char *s)
 114 {
 115     if (strcasecmp(s, "GET") == 0) {
 116         return METHOD_GET;
 117     } else if (strcasecmp(s, "POST") == 0) {
 118         return METHOD_POST;
 119     } else if (strcasecmp(s, "PUT") == 0) {
 120         return METHOD_PUT;
 121     } else if (strcasecmp(s, "HEAD") == 0) {
 122         return METHOD_HEAD;
 123     } else if (strcasecmp(s, "CONNECT") == 0) {
 124         return METHOD_CONNECT;
 125     } else if (strcasecmp(s, "TRACE") == 0) {
 126         return METHOD_TRACE;
 127     } else if (strcasecmp(s, "PURGE") == 0) {
 128         return METHOD_PURGE;
 129     }
 130     return METHOD_NONE;
 131 }
 132
 133 /* FileIterator */
 134
 135 static void fileIteratorAdvance(FileIterator *fi);
 136
 137 static FileIterator *
 138 fileIteratorCreate(const char *fname, FI_READER reader)
 139 {
 140     FileIterator *fi = xcalloc(1, sizeof(FileIterator));
 141     assert(fname && reader);
 142     fi->fname = fname;
 143     fi->reader = reader;
 144     fi->file = fopen(fname, "r");
 145     if (!fi->file) {
 146         fprintf(stderr, "cannot open %s: %s\n", fname, strerror(errno));
 147         return NULL;
 148     } else
 149         fprintf(stderr, "opened %s\n", fname);
 150     fileIteratorAdvance(fi);
 151     return fi;
 152 }
 153
 154 static void
 155 fileIteratorDestroy(FileIterator *fi)
 156 {
 157     assert(fi);
 158     if (fi->file) {
 159         fclose(fi->file);
 160         fprintf(stderr, "closed %s\n", fi->fname);
 161     }
 162     xfree(fi->entry);
 163     xfree(fi);
 164 }
 165
 166 static void
 167 fileIteratorSetCurTime(FileIterator *fi, time_t ct)
 168 {
 169     assert(fi);
 170     assert(fi->inner_time > 0);
 171     fi->time_offset = ct - fi->inner_time;
 172 }
 173
 174 static void
 175 fileIteratorAdvance(FileIterator *fi)
 176 {
 177     int res;
 178     assert(fi);
 179     do {
 180         const time_t last_time = fi->inner_time;
 181         fi->inner_time = -1;
 182         res = fi->reader(fi);
 183         fi->line_count++;
 184         if (fi->inner_time < 0)
 185             fi->inner_time = last_time;
 186         else
 187             fi->inner_time += fi->time_offset;
 188         if (res == frError)
 189             fi->bad_line_count++;
 190         else
 191         if (res == frEof) {
 192             fprintf(stderr, "exhausted %s (%d entries) at %s",
 193                 fi->fname, fi->line_count, ctime(&fi->inner_time));
 194             fi->inner_time = -1;
 195         } else
 196         if (fi->inner_time < last_time) {
 197             assert(last_time >= 0);
 198             fi->time_warp_count++;
 199             fi->inner_time = last_time;
 200         }
 201         /* report progress */
 202         if (!(fi->line_count % 50000))
 203             fprintf(stderr, "%s scanned %d K entries (%d bad) at %s",
 204                 fi->fname, fi->line_count / 1000, fi->bad_line_count,
 205                 ctime(&fi->inner_time));
 206     } while (res < 0);
 207 }
 208
 209 /* CacheEntry */
 210
 211 static CacheEntry *
 212 cacheEntryCreate(const storeSwapLogData * s)
 213 {
 214     CacheEntry *e = xcalloc(1, sizeof(CacheEntry));
 215     assert(s);
 216     /* e->s = *s; */
 217     xmemcpy(e->key_arr, s->key, MD5_DIGEST_CHARS);
 218     e->key = &e->key_arr[0];
 219     return e;
 220 }
 221
 222 static void
 223 cacheEntryDestroy(CacheEntry * e)
 224 {
 225     assert(e);
 226     xfree(e);
 227 }
 228
 229
 230 /* Cache */
 231
 232 static Cache *
 233 cacheCreate(const char *name)
 234 {
 235     Cache *c;
 236     assert(name && strlen(name));
 237     c = xcalloc(1, sizeof(Cache));
 238     c->name = name;
 239     c->hash = hash_create(storeKeyHashCmp, 2e6, storeKeyHashHash);
 240     return c;
 241 }
 242
 243 static void
 244 cacheDestroy(Cache * cache)
 245 {
 246     CacheEntry *e = NULL;
 247     hash_table *hash;
 248     assert(cache);
 249     hash = cache->hash;
 250     /* destroy hash table contents */
 251     for (e = hash_first(hash); e; e = hash_next(hash)) {
 252         hash_remove_link(hash, (hash_link*)e);
 253         cacheEntryDestroy(e);
 254     }
 255     /* destroy the hash table itself */
 256         hashFreeMemory(hash);
 257     if (cache->digest)
 258         cacheDigestDestroy(cache->digest);
 259     xfree(cache);
 260 }
 261
 262 /* re-digests currently hashed entries */
 263 static void
 264 cacheResetDigest(Cache * cache)
 265 {
 266     CacheEntry *e = NULL;
 267     hash_table *hash;
 268     struct timeval t_start, t_end;
 269
 270     assert(cache);
 271     fprintf(stderr, "%s: init-ing digest with %d entries\n", cache->name, cache->count);
 272     if (cache->digest)
 273         cacheDigestDestroy(cache->digest);
 274     hash = cache->hash;
 275     cache->digest = cacheDigestCreate(2 * cache->count + 1); /* 50% utilization */
 276     if (!cache->count)
 277         return;
 278     gettimeofday(&t_start, NULL);
 279     for (e = hash_first(hash); e; e = hash_next(hash)) {
 280         cacheDigestAdd(cache->digest, e->key);
 281     }
 282     gettimeofday(&t_end, NULL);
 283     assert(cache->digest->count == cache->count);
 284     fprintf(stderr, "%s: init-ed  digest with %d entries\n",
 285         cache->name, cache->digest->count);
 286     fprintf(stderr, "%s: init took: %f sec, %f sec/M\n",
 287         cache->name,
 288         tvSubDsec(t_start, t_end),
 289         (double) 1e6 * tvSubDsec(t_start, t_end) / cache->count);
 290     /* check how long it takes to traverse the hash */
 291     gettimeofday(&t_start, NULL);
 292     for (e = hash_first(hash); e; e = hash_next(hash)) {
 293     }
 294     gettimeofday(&t_end, NULL);
 295     fprintf(stderr, "%s: hash scan took: %f sec, %f sec/M\n",
 296         cache->name,
 297         tvSubDsec(t_start, t_end),
 298         (double) 1e6 * tvSubDsec(t_start, t_end) / cache->count);
 299 }
 300
 301 static void
 302 cacheQueryPeer(Cache * cache, const cache_key * key)
 303 {
 304     const int peer_has_it = hash_lookup(cache->peer->hash, key) != NULL;
 305     const int we_think_we_have_it = cacheDigestTest(cache->digest, key);
 306
 307     cache->qstats.query_count++;
 308     if (peer_has_it) {
 309         if (we_think_we_have_it)
 310             cache->qstats.true_hit_count++;
 311         else
 312             cache->qstats.false_miss_count++;
 313     } else {
 314         if (we_think_we_have_it)
 315             cache->qstats.false_hit_count++;
 316         else
 317             cache->qstats.true_miss_count++;
 318     }
 319 }
 320
 321 static void
 322 cacheQueryReport(Cache * cache, CacheQueryStats *stats)
 323 {
 324     fprintf(stdout, "%s: peer queries: %d (%d%%)\n",
 325         cache->name,
 326         stats->query_count, xpercentInt(stats->query_count, cache->req_count)
 327         );
 328     fprintf(stdout, "%s: t-hit: %d (%d%%) t-miss: %d (%d%%) t-*: %d (%d%%)\n",
 329         cache->name,
 330         stats->true_hit_count, xpercentInt(stats->true_hit_count, stats->query_count),
 331         stats->true_miss_count, xpercentInt(stats->true_miss_count, stats->query_count),
 332         stats->true_hit_count + stats->true_miss_count,
 333         xpercentInt(stats->true_hit_count + stats->true_miss_count, stats->query_count)
 334         );
 335     fprintf(stdout, "%s: f-hit: %d (%d%%) f-miss: %d (%d%%) f-*: %d (%d%%)\n",
 336         cache->name,
 337         stats->false_hit_count, xpercentInt(stats->false_hit_count, stats->query_count),
 338         stats->false_miss_count, xpercentInt(stats->false_miss_count, stats->query_count),
 339         stats->false_hit_count + stats->false_miss_count,
 340         xpercentInt(stats->false_hit_count + stats->false_miss_count, stats->query_count)
 341         );
 342 }
 343
 344 static void
 345 cacheReport(Cache * cache)
 346 {
 347     fprintf(stdout, "%s: entries: %d reqs: %d bad-add: %d bad-del: %d\n",
 348         cache->name, cache->count, cache->req_count,
 349         cache->bad_add_count, cache->bad_del_count);
 350
 351     if (cache->digest) {
 352         int bit_count, on_count;
 353         cacheDigestUtil(cache->digest, &bit_count, &on_count);
 354         fprintf(stdout, "%s: digest entries: cnt: %d cap: %d util: %d%% size: %d b\n",
 355             cache->name,
 356             cache->digest->count, cache->digest->capacity,
 357             xpercentInt(cache->digest->count, cache->digest->capacity),
 358             bit_count/8
 359         );
 360         fprintf(stdout, "%s: digest bits: on: %d cap: %d util: %d%%\n",
 361             cache->name,
 362             on_count, bit_count,
 363             xpercentInt(on_count, bit_count)
 364         );
 365     }
 366 }
 367
 368 static void
 369 cacheFetch(Cache *cache, const RawAccessLogEntry *e)
 370 {
 371     assert(e);
 372     cache->req_count++;
 373     if (e->use_icp)
 374         cacheQueryPeer(cache, e->key);
 375 }
 376
 377 static fr_result
 378 swapStateReader(FileIterator * fi)
 379 {
 380     storeSwapLogData *entry;
 381     if (!fi->entry)
 382         fi->entry = xcalloc(1, sizeof(storeSwapLogData));
 383     entry = fi->entry;
 384     if (fread(entry, sizeof(*entry), 1, fi->file) != 1)
 385         return frEof;
 386     fi->inner_time = entry->lastref;
 387     if (entry->op != SWAP_LOG_ADD && entry->op != SWAP_LOG_DEL) {
 388         fprintf(stderr, "%s:%d: unknown swap log action\n", fi->fname, fi->line_count);
 389         exit(-3);
 390     }
 391     return frOk;
 392 }
 393
 394 static fr_result
 395 accessLogReader(FileIterator * fi)
 396 {
 397     static char buf[4096];
 398     RawAccessLogEntry *entry;
 399     char *url;
 400     char *method;
 401     int method_id = METHOD_NONE;
 402     char *hier = NULL;
 403
 404     assert(fi);
 405     if (!fi->entry)
 406         fi->entry = xcalloc(1, sizeof(RawAccessLogEntry));
 407     else
 408         memset(fi->entry, 0, sizeof(RawAccessLogEntry));
 409     entry = fi->entry;
 410     if (!fgets(buf, sizeof(buf), fi->file))
 411         return frEof; /* eof */
 412     entry->timestamp = fi->inner_time = (time_t)atoi(buf);
 413     url = strstr(buf, "://");
 414     hier = url ? strstr(url, " - ") : NULL;
 415
 416     if (!url || !hier) {
 417         /*fprintf(stderr, "%s:%d: strange access log entry '%s'\n",
 418          * fname, scanned_count, buf); */
 419         return frError;
 420     }
 421     method = url;
 422     while (!isdigit(*method)) {
 423         if (*method == ' ')
 424             *method = '\0';
 425         --method;
 426     }
 427     method += 2;
 428     method_id = methodStrToId(method);
 429     if (method_id == METHOD_NONE) {
 430         /*fprintf(stderr, "%s:%d: invalid method %s in '%s'\n",
 431          * fname, scanned_count, method, buf); */
 432         return frError;
 433     }
 434     while (*url) url--;
 435     url++;
 436     *hier = '\0';
 437     hier += 3;
 438     *strchr(hier, '/') = '\0';
 439     /*fprintf(stdout, "%s:%d: %s %s %s\n",
 440      * fname, count, method, url, hier); */
 441     entry->use_icp =  /* no ICP lookup for these status codes */
 442         strcmp(hier, "NONE") &&
 443         strcmp(hier, "DIRECT") &&
 444         strcmp(hier, "FIREWALL_IP_DIRECT") &&
 445         strcmp(hier, "LOCAL_IP_DIRECT") &&
 446         strcmp(hier, "NO_DIRECT_FAIL") &&
 447         strcmp(hier, "NO_PARENT_DIRECT") &&
 448         strcmp(hier, "SINGLE_PARENT") &&
 449         strcmp(hier, "PASSTHROUGH_PARENT") &&
 450         strcmp(hier, "SSL_PARENT_MISS") &&
 451         strcmp(hier, "DEFAULT_PARENT");
 452     memcpy(entry->key, storeKeyPublic(url, method_id), sizeof(entry->key));
 453     /*fprintf(stdout, "%s:%d: %s %s %s %s\n",
 454         fname, count, method, storeKeyText(entry->key), url, hier); */
 455     return frOk;
 456 }
 457
 458
 459 static void
 460 cachePurge(Cache *cache, storeSwapLogData *s)
 461 {
 462     CacheEntry *olde = (CacheEntry *) hash_lookup(cache->hash, s->key);
 463     if (!olde) {
 464         cache->bad_del_count++;
 465     } else {
 466         assert(cache->count);
 467         hash_remove_link(cache->hash, (hash_link *) olde);
 468         cacheEntryDestroy(olde);
 469         cache->count--;
 470     }
 471 }
 472
 473 static void
 474 cacheStore(Cache *cache, storeSwapLogData *s, int update_digest)
 475 {
 476     CacheEntry *olde = (CacheEntry *) hash_lookup(cache->hash, s->key);
 477     if (olde) {
 478         cache->bad_add_count++;
 479     } else {
 480         CacheEntry *e = cacheEntryCreate(s);
 481         hash_join(cache->hash, (hash_link *) e);
 482         cache->count++;
 483         if (update_digest)
 484             cacheDigestAdd(cache->digest, e->key);
 485     }
 486 }
 487
 488 static void
 489 cacheUpdateStore(Cache *cache, storeSwapLogData *s, int update_digest)
 490 {
 491     switch (s->op) {
 492         case SWAP_LOG_ADD:
 493             cacheStore(cache, s, update_digest);
 494             break;
 495         case SWAP_LOG_DEL:
 496             cachePurge(cache, s);
 497             break;
 498         default:
 499             assert(0);
 500     }
 501 }
 502
 503 static int
 504 usage(const char *prg_name)
 505 {
 506     fprintf(stderr, "usage: %s <access_log> <swap_state> ...\n",
 507         prg_name);
 508     return -1;
 509 }
 510
 511 int
 512 main(int argc, char *argv[])
 513 {
 514     FileIterator **fis = NULL;
 515     const int fi_count = argc-1;
 516     int active_fi_count = 0;
 517     time_t ready_time;
 518     Cache *them, *us;
 519     int i;
 520
 521     if (argc < 3)
 522         return usage(argv[0]);
 523
 524     them = cacheCreate("them");
 525     us = cacheCreate("us");
 526     them->peer = us;
 527     us->peer = them;
 528
 529     fis = xcalloc(fi_count, sizeof(FileIterator *));
 530     /* init iterators with files */
 531     fis[0] = fileIteratorCreate(argv[1], accessLogReader);
 532     for (i = 2; i < argc; ++i) {
 533         fis[i-1] = fileIteratorCreate(argv[i], swapStateReader);
 534         if (!fis[i-1]) return -2;
 535     }
 536     /* read prefix to get start-up contents of the peer cache */
 537     ready_time = -1;
 538     for (i = 1; i < fi_count; ++i) {
 539         FileIterator *fi = fis[i];
 540         while (fi->inner_time > 0) {
 541             ready_time = fi->inner_time;
 542             if (((storeSwapLogData*)fi->entry)->op != SWAP_LOG_ADD) {
 543                 break;
 544             } else {
 545                 cacheStore(them, fi->entry, 0);
 546                 fileIteratorAdvance(fi);
 547             }
 548         }
 549     }
 550     /* digest peer cache content */
 551     cacheResetDigest(them);
 552     us->digest = cacheDigestClone(them->digest); /* @netw@ */
 553
 554     /* shift the time in access log to match ready_time */
 555     fileIteratorSetCurTime(fis[0], ready_time);
 556
 557     /* iterate, use the iterator with the smallest positive inner_time */
 558     cur_time = -1;
 559     do {
 560         int next_i = -1;
 561         time_t next_time = -1;
 562         active_fi_count = 0;
 563         for (i = 0; i < fi_count; ++i) {
 564             if (fis[i]->inner_time >= 0) {
 565                 if (!active_fi_count || fis[i]->inner_time < next_time) {
 566                     next_i = i;
 567                     next_time = fis[i]->inner_time;
 568                 }
 569                 active_fi_count++;
 570             }
 571         }
 572         if (next_i >= 0) {
 573             cur_time = next_time;
 574             /*fprintf(stderr, "%2d time: %d %s", next_i, (int)cur_time, ctime(&cur_time));*/
 575             if (next_i == 0)
 576                 cacheFetch(us, fis[next_i]->entry);
 577             else
 578                 cacheUpdateStore(them, fis[next_i]->entry, 1);
 579             fileIteratorAdvance(fis[next_i]);
 580         }
 581     } while (active_fi_count);
 582
 583     /* report */
 584     cacheReport(them);
 585     cacheReport(us);
 586     cacheQueryReport(us, &us->qstats);
 587
 588     /* clean */
 589     for (i = 0; i < argc-1; ++i) {
 590         fileIteratorDestroy(fis[i]);
 591     }
 592     xfree(fis);
 593     cacheDestroy(them);
 594     cacheDestroy(us);
 595     return 0;
 596 }