src/format/Token.cc

   1 #include "squid.h"
   2 #include "format/Config.h"
   3 #include "format/Token.h"
   4 #include "format/TokenTableEntry.h"
   5 #include "globals.h"
   6 #include "SquidConfig.h"
   7 #include "Store.h"
   8
   9 // Due to token overlaps between 1 and 2 letter tokens (Bug 3310)
  10 // We split the token table into sets determined by the token length
  11 namespace Format
  12 {
  13
  14 /// 1-char tokens.
  15 static TokenTableEntry TokenTable1C[] = {
  16
  17     {">a", LFT_CLIENT_IP_ADDRESS},
  18     {">p", LFT_CLIENT_PORT},
  19     {">A", LFT_CLIENT_FQDN},
  20
  21     {"<a", LFT_SERVER_IP_ADDRESS},
  22     {"<p", LFT_SERVER_PORT},
  23     {"<A", LFT_SERVER_FQDN_OR_PEER_NAME},
  24
  25     {">h", LFT_REQUEST_HEADER},
  26     {">h", LFT_REQUEST_ALL_HEADERS},
  27     {"<h", LFT_REPLY_HEADER},
  28     {"<h", LFT_REPLY_ALL_HEADERS},
  29
  30     {">v", LFT_REQUEST_VERSION_OLD_2X},
  31
  32     {"%", LFT_PERCENT},
  33
  34     {NULL, LFT_NONE}            /* this must be last */
  35 };
  36
  37 /// 2-char tokens
  38 static TokenTableEntry TokenTable2C[] = {
  39
  40     {">la", LFT_CLIENT_LOCAL_IP},
  41     {"la", LFT_LOCAL_LISTENING_IP},
  42     {">lp", LFT_CLIENT_LOCAL_PORT},
  43     {"lp", LFT_LOCAL_LISTENING_PORT},
  44     /*{ "lA", LFT_LOCAL_NAME }, */
  45
  46     {"<la", LFT_SERVER_LOCAL_IP},
  47     {"oa", LFT_SERVER_LOCAL_IP_OLD_27},
  48     {"<lp", LFT_SERVER_LOCAL_PORT},
  49
  50     {"ts", LFT_TIME_SECONDS_SINCE_EPOCH},
  51     {"tu", LFT_TIME_SUBSECOND},
  52     {"tl", LFT_TIME_LOCALTIME},
  53     {"tg", LFT_TIME_GMT},
  54     {"tS", LFT_TIME_START},
  55     {"tr", LFT_TIME_TO_HANDLE_REQUEST},
  56
  57     {"<pt", LFT_PEER_RESPONSE_TIME},
  58     {"<tt", LFT_TOTAL_SERVER_SIDE_RESPONSE_TIME},
  59     {"dt", LFT_DNS_WAIT_TIME},
  60
  61     {">ha", LFT_ADAPTED_REQUEST_HEADER},
  62     {">ha", LFT_ADAPTED_REQUEST_ALL_HEADERS},
  63
  64     {"un", LFT_USER_NAME},
  65     {"ul", LFT_USER_LOGIN},
  66     /*{ "ur", LFT_USER_REALM }, */
  67     /*{ "us", LFT_USER_SCHEME }, */
  68     {"ui", LFT_USER_IDENT},
  69     {"ue", LFT_USER_EXTERNAL},
  70
  71     {"Hs", LFT_HTTP_SENT_STATUS_CODE_OLD_30},
  72     {">Hs", LFT_HTTP_SENT_STATUS_CODE},
  73     {"<Hs", LFT_HTTP_RECEIVED_STATUS_CODE},
  74     /*{ "Ht", LFT_HTTP_STATUS }, */
  75     {"<bs", LFT_HTTP_BODY_BYTES_READ},
  76
  77     {"Ss", LFT_SQUID_STATUS},
  78     {"Sh", LFT_SQUID_HIERARCHY},
  79
  80     {"mt", LFT_MIME_TYPE},
  81
  82     {">rm", LFT_CLIENT_REQ_METHOD},
  83     {">ru", LFT_CLIENT_REQ_URI},
  84     {">rs", LFT_CLIENT_REQ_URLSCHEME},
  85     {">rd", LFT_CLIENT_REQ_URLDOMAIN},
  86     {">rP", LFT_CLIENT_REQ_URLPORT},
  87     {">rp", LFT_CLIENT_REQ_URLPATH},
  88     /*{">rq", LFT_CLIENT_REQ_QUERY},*/
  89     {">rv", LFT_CLIENT_REQ_VERSION},
  90
  91     {"rm", LFT_REQUEST_METHOD},
  92     {"ru", LFT_REQUEST_URI},    /* doesn't include the query-string */
  93     {"rp", LFT_REQUEST_URLPATH_OLD_31},
  94     /* { "rq", LFT_REQUEST_QUERY }, * /     / * the query-string, INCLUDING the leading ? */
  95     {"rv", LFT_REQUEST_VERSION},
  96     {"rG", LFT_REQUEST_URLGROUP_OLD_2X},
  97
  98     {"<rm", LFT_SERVER_REQ_METHOD},
  99     {"<ru", LFT_SERVER_REQ_URI},
 100     {"<rs", LFT_SERVER_REQ_URLSCHEME},
 101     {"<rd", LFT_SERVER_REQ_URLDOMAIN},
 102     {"<rP", LFT_SERVER_REQ_URLPORT},
 103     {"<rp", LFT_SERVER_REQ_URLPATH},
 104     /*{"<rq", LFT_SERVER_REQ_QUERY},*/
 105     {"<rv", LFT_SERVER_REQ_VERSION},
 106
 107     {">st", LFT_CLIENT_REQUEST_SIZE_TOTAL },
 108     {">sh", LFT_CLIENT_REQUEST_SIZE_HEADERS },
 109     /*{ ">sb", LFT_REQUEST_SIZE_BODY }, */
 110     /*{ ">sB", LFT_REQUEST_SIZE_BODY_NO_TE }, */
 111
 112     {"<st", LFT_ADAPTED_REPLY_SIZE_TOTAL}, // XXX: adapted should be code: <sta
 113     {"<sH", LFT_REPLY_HIGHOFFSET},
 114     {"<sS", LFT_REPLY_OBJECTSIZE},
 115     {"<sh", LFT_ADAPTED_REPLY_SIZE_HEADERS }, // XXX: adapted should be code: <sha
 116     /*{ "<sb", LFT_REPLY_SIZE_BODY }, */
 117     /*{ "<sB", LFT_REPLY_SIZE_BODY_NO_TE }, */
 118
 119     {"st", LFT_CLIENT_IO_SIZE_TOTAL}, // XXX: total from client should be stC ??
 120     /*{"stP", LFT_SERVER_IO_SIZE_TOTAL},*/
 121
 122     {"et", LFT_TAG},
 123     {"ea", LFT_EXT_LOG},
 124     {"sn", LFT_SEQUENCE_NUMBER},
 125
 126     {NULL, LFT_NONE}            /* this must be last */
 127 };
 128
 129 /// Miscellaneous >2 byte tokens
 130 static TokenTableEntry TokenTableMisc[] = {
 131     {">eui", LFT_CLIENT_EUI},
 132     {">qos", LFT_CLIENT_LOCAL_TOS},
 133     {"<qos", LFT_SERVER_LOCAL_TOS},
 134     {">nfmark", LFT_CLIENT_LOCAL_NFMARK},
 135     {"<nfmark", LFT_SERVER_LOCAL_NFMARK},
 136     {"err_code", LFT_SQUID_ERROR },
 137     {"err_detail", LFT_SQUID_ERROR_DETAIL },
 138     {"note", LFT_NOTE },
 139     {"credentials", LFT_CREDENTIALS},
 140     {NULL, LFT_NONE}            /* this must be last */
 141 };
 142
 143 #if USE_ADAPTATION
 144 static TokenTableEntry TokenTableAdapt[] = {
 145     {"all_trs", LFT_ADAPTATION_ALL_XACT_TIMES},
 146     {"sum_trs", LFT_ADAPTATION_SUM_XACT_TIMES},
 147     {"<last_h", LFT_ADAPTATION_LAST_HEADER},
 148     {NULL, LFT_NONE}           /* this must be last */
 149 };
 150 #endif
 151
 152 #if ICAP_CLIENT
 153 /// ICAP (icap::) tokens
 154 static TokenTableEntry TokenTableIcap[] = {
 155     {"tt", LFT_ICAP_TOTAL_TIME},
 156     {"<last_h", LFT_ADAPTATION_LAST_HEADER}, // deprecated
 157
 158     {"<A",  LFT_ICAP_ADDR},
 159     {"<service_name",  LFT_ICAP_SERV_NAME},
 160     {"ru",  LFT_ICAP_REQUEST_URI},
 161     {"rm",  LFT_ICAP_REQUEST_METHOD},
 162     {">st", LFT_ICAP_BYTES_SENT},
 163     {"<st", LFT_ICAP_BYTES_READ},
 164     {"<bs", LFT_ICAP_BODY_BYTES_READ},
 165
 166     {">h",  LFT_ICAP_REQ_HEADER},
 167     {"<h",  LFT_ICAP_REP_HEADER},
 168
 169     {"tr",  LFT_ICAP_TR_RESPONSE_TIME},
 170     {"tio", LFT_ICAP_IO_TIME},
 171     {"to",  LFT_ICAP_OUTCOME},
 172     {"Hs",  LFT_ICAP_STATUS_CODE},
 173
 174     {NULL, LFT_NONE}           /* this must be last */
 175 };
 176 #endif
 177
 178 #if USE_OPENSSL
 179 // SSL (ssl::) tokens
 180 static TokenTableEntry TokenTableSsl[] = {
 181     {"bump_mode", LFT_SSL_BUMP_MODE},
 182     {">cert_subject", LFT_SSL_USER_CERT_SUBJECT},
 183     {">cert_issuer", LFT_SSL_USER_CERT_ISSUER},
 184     {">sni", LFT_SSL_CLIENT_SNI},
 185     /*{"<cert_subject", LFT_SSL_SERVER_CERT_SUBJECT}, */
 186     /*{"<cert_issuer", LFT_SSL_SERVER_CERT_ISSUER}, */
 187     {NULL, LFT_NONE}
 188 };
 189 #endif
 190 } // namespace Format
 191
 192 /// Register all components custom format tokens
 193 void
 194 Format::Token::Init()
 195 {
 196     // TODO standard log tokens
 197     // TODO external ACL fmt tokens
 198
 199 #if USE_ADAPTATION
 200     TheConfig.registerTokens(String("adapt"),::Format::TokenTableAdapt);
 201 #endif
 202 #if ICAP_CLIENT
 203     TheConfig.registerTokens(String("icap"),::Format::TokenTableIcap);
 204 #endif
 205 #if USE_OPENSSL
 206     TheConfig.registerTokens(String("ssl"),::Format::TokenTableSsl);
 207 #endif
 208 }
 209
 210 /// Scans a token table to see if the next token exists there
 211 /// returns a pointer to next unparsed byte and updates type member if found
 212 const char *
 213 Format::Token::scanForToken(TokenTableEntry const table[], const char *cur)
 214 {
 215     for (TokenTableEntry const *lte = table; lte->configTag != NULL; ++lte) {
 216         debugs(46, 8, HERE << "compare tokens '" << lte->configTag << "' with '" << cur << "'");
 217         if (strncmp(lte->configTag, cur, strlen(lte->configTag)) == 0) {
 218             type = lte->tokenType;
 219             label = lte->configTag;
 220             debugs(46, 7, HERE << "Found token '" << label << "'");
 221             return cur + strlen(lte->configTag);
 222         }
 223     }
 224     return cur;
 225 }
 226
 227 /* parses a single token. Returns the token length in characters,
 228  * and fills in the lt item with the token information.
 229  * def is for sure null-terminated
 230  */
 231 int
 232 Format::Token::parse(const char *def, Quoting *quoting)
 233 {
 234     const char *cur = def;
 235
 236     int l;
 237
 238     l = strcspn(cur, "%");
 239
 240     if (l > 0) {
 241         char *cp;
 242         /* it's a string for sure, until \0 or the next % */
 243         cp = (char *)xmalloc(l + 1);
 244         xstrncpy(cp, cur, l + 1);
 245         type = LFT_STRING;
 246         data.string = cp;
 247
 248         while (l > 0) {
 249             switch (*cur) {
 250
 251             case '"':
 252
 253                 if (*quoting == LOG_QUOTE_NONE)
 254                     *quoting = LOG_QUOTE_QUOTES;
 255                 else if (*quoting == LOG_QUOTE_QUOTES)
 256                     *quoting = LOG_QUOTE_NONE;
 257
 258                 break;
 259
 260             case '[':
 261                 if (*quoting == LOG_QUOTE_NONE)
 262                     *quoting = LOG_QUOTE_MIMEBLOB;
 263
 264                 break;
 265
 266             case ']':
 267                 if (*quoting == LOG_QUOTE_MIMEBLOB)
 268                     *quoting = LOG_QUOTE_NONE;
 269
 270                 break;
 271             }
 272
 273             ++cur;
 274             --l;
 275         }
 276
 277     } else if (*cur) {
 278
 279         ++cur;
 280
 281         // select quoting style for his particular token
 282         switch (*cur) {
 283
 284         case '"':
 285             quote = LOG_QUOTE_QUOTES;
 286             ++cur;
 287             break;
 288
 289         case '\'':
 290             quote = LOG_QUOTE_RAW;
 291             ++cur;
 292             break;
 293
 294         case '[':
 295             quote = LOG_QUOTE_MIMEBLOB;
 296             ++cur;
 297             break;
 298
 299         case '#':
 300             quote = LOG_QUOTE_URL;
 301             ++cur;
 302             break;
 303
 304         default:
 305             quote = *quoting;
 306             break;
 307         }
 308
 309         if (*cur == '-') {
 310             left = true;
 311             ++cur;
 312         }
 313
 314         if (*cur == '0') {
 315             zero = true;
 316             ++cur;
 317         }
 318
 319         char *endp;
 320         if (xisdigit(*cur)) {
 321             widthMin = strtol(cur, &endp, 10);
 322             cur = endp;
 323         }
 324
 325         if (*cur == '.' && xisdigit(*(++cur))) {
 326             widthMax = strtol(cur, &endp, 10);
 327             cur = endp;
 328         }
 329
 330         if (*cur == '{') {
 331             char *cp;
 332             ++cur;
 333             l = strcspn(cur, "}");
 334             cp = (char *)xmalloc(l + 1);
 335             xstrncpy(cp, cur, l + 1);
 336             data.string = cp;
 337             cur += l;
 338
 339             if (*cur == '}')
 340                 ++cur;
 341         }
 342
 343         type = LFT_NONE;
 344
 345         // Scan each registered token namespace
 346         debugs(46, 9, HERE << "check for token in " << TheConfig.tokens.size() << " namespaces.");
 347         for (std::list<TokenNamespace>::const_iterator itr = TheConfig.tokens.begin(); itr != TheConfig.tokens.end(); ++itr) {
 348             debugs(46, 7, HERE << "check for possible " << itr->prefix << ":: token");
 349             const size_t len = itr->prefix.size();
 350             if (itr->prefix.cmp(cur, len) == 0 && cur[len] == ':' && cur[len+1] == ':') {
 351                 debugs(46, 5, HERE << "check for " << itr->prefix << ":: token in '" << cur << "'");
 352                 const char *old = cur;
 353                 cur = scanForToken(itr->tokenSet, cur+len+2);
 354                 if (old != cur) // found
 355                     break;
 356                 else // reset to start of namespace
 357                     cur = cur - len - 2;
 358             }
 359         }
 360
 361         if (type == LFT_NONE) {
 362             // For upward compatibility, assume "http::" prefix as default prefix
 363             // for all log access formatting codes, except those starting with a
 364             // "%" or a known namespace. (ie "icap::", "adapt::")
 365             if (strncmp(cur,"http::", 6) == 0 && *(cur+6) != '%' )
 366                 cur += 6;
 367
 368             // NP: scan the sets of tokens in decreasing size to guarantee no
 369             //     mistakes made with overlapping names. (Bug 3310)
 370
 371             // Scan for various long tokens
 372             debugs(46, 5, HERE << "scan for possible Misc token");
 373             cur = scanForToken(TokenTableMisc, cur);
 374             // scan for 2-char tokens
 375             if (type == LFT_NONE) {
 376                 debugs(46, 5, HERE << "scan for possible 2C token");
 377                 cur = scanForToken(TokenTable2C, cur);
 378             }
 379             // finally scan for 1-char tokens.
 380             if (type == LFT_NONE) {
 381                 debugs(46, 5, HERE << "scan for possible 1C token");
 382                 cur = scanForToken(TokenTable1C, cur);
 383             }
 384         }
 385
 386         if (type == LFT_NONE) {
 387             fatalf("Can't parse configuration token: '%s'\n", def);
 388         }
 389
 390         if (*cur == ' ') {
 391             space = true;
 392             ++cur;
 393         }
 394     }
 395
 396     switch (type) {
 397
 398 #if USE_ADAPTATION
 399     case LFT_ADAPTATION_LAST_HEADER:
 400 #endif
 401
 402 #if ICAP_CLIENT
 403     case LFT_ICAP_REQ_HEADER:
 404
 405     case LFT_ICAP_REP_HEADER:
 406 #endif
 407
 408     case LFT_ADAPTED_REQUEST_HEADER:
 409
 410     case LFT_REQUEST_HEADER:
 411
 412     case LFT_REPLY_HEADER:
 413
 414     case LFT_NOTE:
 415
 416         if (data.string) {
 417             char *header = data.string;
 418             char *cp = strchr(header, ':');
 419
 420             if (cp) {
 421                 *cp = '\0';
 422                 ++cp;
 423
 424                 if (*cp == ',' || *cp == ';' || *cp == ':') {
 425                     data.header.separator = *cp;
 426                     ++cp;
 427                 } else {
 428                     data.header.separator = ',';
 429                 }
 430
 431                 data.header.element = cp;
 432
 433                 switch (type) {
 434                 case LFT_REQUEST_HEADER:
 435                     type = LFT_REQUEST_HEADER_ELEM;
 436                     break;
 437
 438                 case LFT_ADAPTED_REQUEST_HEADER:
 439                     type = LFT_ADAPTED_REQUEST_HEADER_ELEM;
 440                     break;
 441
 442                 case LFT_REPLY_HEADER:
 443                     type = LFT_REPLY_HEADER_ELEM;
 444                     break;
 445 #if USE_ADAPTATION
 446                 case LFT_ADAPTATION_LAST_HEADER:
 447                     type = LFT_ADAPTATION_LAST_HEADER_ELEM;
 448                     break;
 449 #endif
 450 #if ICAP_CLIENT
 451                 case LFT_ICAP_REQ_HEADER:
 452                     type = LFT_ICAP_REQ_HEADER_ELEM;
 453                     break;
 454                 case LFT_ICAP_REP_HEADER:
 455                     type = LFT_ICAP_REP_HEADER_ELEM;
 456                     break;
 457 #endif
 458                 default:
 459                     break;
 460                 }
 461             }
 462
 463             data.header.header = header;
 464         } else {
 465             switch (type) {
 466             case LFT_REQUEST_HEADER:
 467                 type = LFT_REQUEST_ALL_HEADERS;
 468                 break;
 469
 470             case LFT_ADAPTED_REQUEST_HEADER:
 471                 type = LFT_ADAPTED_REQUEST_ALL_HEADERS;
 472                 break;
 473
 474             case LFT_REPLY_HEADER:
 475                 type = LFT_REPLY_ALL_HEADERS;
 476                 break;
 477 #if USE_ADAPTATION
 478             case LFT_ADAPTATION_LAST_HEADER:
 479                 type = LFT_ADAPTATION_LAST_ALL_HEADERS;
 480                 break;
 481 #endif
 482 #if ICAP_CLIENT
 483             case LFT_ICAP_REQ_HEADER:
 484                 type = LFT_ICAP_REQ_ALL_HEADERS;
 485                 break;
 486             case LFT_ICAP_REP_HEADER:
 487                 type = LFT_ICAP_REP_ALL_HEADERS;
 488                 break;
 489 #endif
 490             default:
 491                 break;
 492             }
 493             Config.onoff.log_mime_hdrs = 1;
 494         }
 495
 496         break;
 497
 498     case LFT_CLIENT_FQDN:
 499         Config.onoff.log_fqdn = 1;
 500         break;
 501
 502     case LFT_TIME_START:
 503     case LFT_TIME_SUBSECOND:
 504         divisor = 1000;
 505
 506         if (widthMax > 0) {
 507             divisor = 1000000;
 508
 509             for (int i = widthMax; i > 0; --i)
 510                 divisor /= 10;
 511
 512             if (!divisor)
 513                 divisor = 1;
 514         }
 515         break;
 516
 517     case LFT_HTTP_SENT_STATUS_CODE_OLD_30:
 518         debugs(46, DBG_PARSE_NOTE(DBG_IMPORTANT), "WARNING: The \"Hs\" formatting code is deprecated. Use the \">Hs\" instead.");
 519         type = LFT_HTTP_SENT_STATUS_CODE;
 520         break;
 521
 522     case LFT_SERVER_LOCAL_IP_OLD_27:
 523         debugs(46, DBG_PARSE_NOTE(DBG_IMPORTANT), "WARNING: The \"oa\" formatting code is deprecated. Use the \"<la\" instead.");
 524         type = LFT_SERVER_LOCAL_IP;
 525         break;
 526
 527     case LFT_REQUEST_URLPATH_OLD_31:
 528         debugs(46, DBG_PARSE_NOTE(DBG_IMPORTANT), "WARNING: The \"rp\" formatting code is deprecated. Use the \">rp\" instead.");
 529         type = LFT_CLIENT_REQ_URLPATH;
 530         break;
 531
 532     case LFT_REQUEST_VERSION_OLD_2X:
 533         debugs(46, DBG_PARSE_NOTE(DBG_IMPORTANT), "WARNING: The \">v\" formatting code is deprecated. Use the \">rv\" instead.");
 534         type = LFT_REQUEST_VERSION;
 535         break;
 536
 537 #if !USE_SQUID_EUI
 538     case LFT_CLIENT_EUI:
 539         debugs(46, DBG_CRITICAL, "WARNING: The \">eui\" formatting code requires EUI features which are disabled in this Squid.");
 540         break;
 541 #endif
 542
 543     case LFT_REQUEST_URLGROUP_OLD_2X:
 544         debugs(46, DBG_PARSE_NOTE(DBG_IMPORTANT), "WARNING: The \"rG\" formatting code is deprecated. Use \"note{urlgroup}\" instead.");
 545         type = LFT_NOTE;
 546         data.header.header = xstrdup("urlgroup");
 547         break;
 548
 549     default:
 550         break;
 551     }
 552
 553     return (cur - def);
 554 }
 555
 556 Format::Token::Token() : type(LFT_NONE),
 557         label(NULL),
 558         widthMin(-1),
 559         widthMax(-1),
 560         quote(LOG_QUOTE_NONE),
 561         left(false),
 562         space(false),
 563         zero(false),
 564         divisor(1),
 565         next(NULL)
 566 {
 567     data.string = NULL;
 568     data.header.header = NULL;
 569     data.header.element = NULL;
 570     data.header.separator = ',';
 571 }
 572
 573 Format::Token::~Token()
 574 {
 575     label = NULL; // drop reference to global static.
 576     safe_free(data.string);
 577     while (next) {
 578         Token *tokens = next;
 579         next = next->next;
 580         tokens->next = NULL;
 581         delete tokens;
 582     }
 583 }
 584