src/urn.cc

   1
   2 /*
   3  * $Id$
   4  *
   5  * DEBUG: section 52    URN Parsing
   6  * AUTHOR: Kostas Anagnostakis
   7  *
   8  * SQUID Web Proxy Cache          http://www.squid-cache.org/
   9  * ----------------------------------------------------------
  10  *
  11  *  Squid is the result of efforts by numerous individuals from
  12  *  the Internet community; see the CONTRIBUTORS file for full
  13  *  details.   Many organizations have provided support for Squid's
  14  *  development; see the SPONSORS file for full details.  Squid is
  15  *  Copyrighted (C) 2001 by the Regents of the University of
  16  *  California; see the COPYRIGHT file for full details.  Squid
  17  *  incorporates software developed and/or copyrighted by other
  18  *  sources; see the CREDITS file for full details.
  19  *
  20  *  This program is free software; you can redistribute it and/or modify
  21  *  it under the terms of the GNU General Public License as published by
  22  *  the Free Software Foundation; either version 2 of the License, or
  23  *  (at your option) any later version.
  24  *
  25  *  This program is distributed in the hope that it will be useful,
  26  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  27  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  28  *  GNU General Public License for more details.
  29  *
  30  *  You should have received a copy of the GNU General Public License
  31  *  along with this program; if not, write to the Free Software
  32  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  33  *
  34  */
  35
  36 #include "squid.h"
  37 #include "errorpage.h"
  38 #include "StoreClient.h"
  39 #include "Store.h"
  40 #include "HttpReply.h"
  41 #include "HttpRequest.h"
  42 #include "MemBuf.h"
  43 #include "forward.h"
  44 #include "globals.h"
  45 #include "SquidTime.h"
  46 #include "icmp/net_db.h"
  47 #include "mime_header.h"
  48 #include "protos.h"
  49 #include "URL.h"
  50
  51 #define URN_REQBUF_SZ   4096
  52
  53 class UrnState : public StoreClient
  54 {
  55
  56 public:
  57     void created (StoreEntry *newEntry);
  58     void *operator new (size_t byteCount);
  59     void operator delete (void *address);
  60     void start (HttpRequest *, StoreEntry *);
  61     char *getHost (String &urlpath);
  62     void setUriResFromRequest(HttpRequest *);
  63     bool RequestNeedsMenu(HttpRequest *r);
  64     void updateRequestURL(HttpRequest *r, char const *newPath, const size_t newPath_len);
  65     void createUriResRequest (String &uri);
  66
  67     virtual ~UrnState();
  68
  69     StoreEntry *entry;
  70     store_client *sc;
  71     StoreEntry *urlres_e;
  72     HttpRequest *request;
  73     HttpRequest *urlres_r;
  74
  75     struct {
  76         unsigned int force_menu:1;
  77     } flags;
  78     char reqbuf[URN_REQBUF_SZ];
  79     int reqofs;
  80
  81 private:
  82     char *urlres;
  83 };
  84
  85 typedef struct {
  86     char *url;
  87     char *host;
  88     int rtt;
  89
  90     struct {
  91         int cached;
  92     } flags;
  93 } url_entry;
  94
  95 static STCB urnHandleReply;
  96 static url_entry *urnParseReply(const char *inbuf, const HttpRequestMethod&);
  97 static const char *const crlf = "\r\n";
  98 static QS url_entry_sort;
  99
 100 CBDATA_TYPE(UrnState);
 101 void *
 102 UrnState::operator new (size_t byteCount)
 103 {
 104     /* derived classes with different sizes must implement their own new */
 105     assert (byteCount == sizeof (UrnState));
 106     CBDATA_INIT_TYPE(UrnState);
 107     return cbdataAlloc(UrnState);
 108
 109 }
 110
 111 void
 112 UrnState::operator delete (void *address)
 113 {
 114     UrnState * tmp = (UrnState *)address;
 115     cbdataFree (tmp);
 116 }
 117
 118 UrnState::~UrnState ()
 119 {
 120     safe_free(urlres);
 121 }
 122
 123 static url_entry *
 124 urnFindMinRtt(url_entry * urls, const HttpRequestMethod& m, int *rtt_ret)
 125 {
 126     int min_rtt = 0;
 127     url_entry *u = NULL;
 128     url_entry *min_u = NULL;
 129     int i;
 130     int urlcnt = 0;
 131     debugs(52, 3, "urnFindMinRtt");
 132     assert(urls != NULL);
 133
 134     for (i = 0; NULL != urls[i].url; ++i)
 135         ++urlcnt;
 136
 137     debugs(53, 3, "urnFindMinRtt: Counted " << i << " URLs");
 138
 139     if (1 == urlcnt) {
 140         debugs(52, 3, "urnFindMinRtt: Only one URL - return it!");
 141         return urls;
 142     }
 143
 144     for (i = 0; i < urlcnt; ++i) {
 145         u = &urls[i];
 146         debugs(52, 3, "urnFindMinRtt: " << u->host << " rtt=" << u->rtt);
 147
 148         if (u->rtt == 0)
 149             continue;
 150
 151         if (u->rtt > min_rtt && min_rtt != 0)
 152             continue;
 153
 154         min_rtt = u->rtt;
 155
 156         min_u = u;
 157     }
 158
 159     if (rtt_ret)
 160         *rtt_ret = min_rtt;
 161
 162     debugs(52, DBG_IMPORTANT, "urnFindMinRtt: Returning '" <<
 163            (min_u ? min_u->url : "NONE") << "' RTT " <<
 164            min_rtt  );
 165
 166     return min_u;
 167 }
 168
 169 char *
 170 UrnState::getHost (String &urlpath)
 171 {
 172     char * result;
 173     size_t p;
 174
 175     /** FIXME: this appears to be parsing the URL. *very* badly. */
 176     /*   a proper encapsulated URI/URL type needs to clear this up. */
 177     if ((p=urlpath.find(':')) != String::npos) {
 178         result=xstrndup(urlpath.rawBuf(),p-1);
 179     } else {
 180         result = xstrndup(urlpath.rawBuf(),urlpath.size());
 181     }
 182     return result;
 183 }
 184
 185 bool
 186 UrnState::RequestNeedsMenu(HttpRequest *r)
 187 {
 188     if (r->urlpath.size() < 5)
 189         return false;
 190     //now we're sure it's long enough
 191     return strncasecmp(r->urlpath.rawBuf(), "menu.", 5) == 0;
 192 }
 193
 194 void
 195 UrnState::updateRequestURL(HttpRequest *r, char const *newPath, const size_t newPath_len)
 196 {
 197     char *new_path = xstrndup (newPath, newPath_len);
 198     r->urlpath = new_path;
 199     xfree(new_path);
 200 }
 201
 202 void
 203 UrnState::createUriResRequest (String &uri)
 204 {
 205     LOCAL_ARRAY(char, local_urlres, 4096);
 206     char *host = getHost (uri);
 207     snprintf(local_urlres, 4096, "http://%s/uri-res/N2L?urn:" SQUIDSTRINGPH,
 208              host, SQUIDSTRINGPRINT(uri));
 209     safe_free (host);
 210     safe_free (urlres);
 211     urlres = xstrdup (local_urlres);
 212     urlres_r = HttpRequest::CreateFromUrl(urlres);
 213 }
 214
 215 void
 216 UrnState::setUriResFromRequest(HttpRequest *r)
 217 {
 218     if (RequestNeedsMenu(r)) {
 219         updateRequestURL(r, r->urlpath.rawBuf() + 5, r->urlpath.size() - 5 );
 220         flags.force_menu = 1;
 221     }
 222
 223     createUriResRequest (r->urlpath);
 224
 225     if (urlres_r == NULL) {
 226         debugs(52, 3, "urnStart: Bad uri-res URL " << urlres);
 227         ErrorState *err = new ErrorState(ERR_URN_RESOLVE, HTTP_NOT_FOUND, r);
 228         err->url = urlres;
 229         urlres = NULL;
 230         errorAppendEntry(entry, err);
 231         return;
 232     }
 233
 234     HTTPMSGLOCK(urlres_r);
 235     urlres_r->header.putStr(HDR_ACCEPT, "text/plain");
 236 }
 237
 238 void
 239 UrnState::start(HttpRequest * r, StoreEntry * e)
 240 {
 241     debugs(52, 3, "urnStart: '" << e->url() << "'" );
 242     entry = e;
 243     request = HTTPMSGLOCK(r);
 244
 245     entry->lock();
 246     setUriResFromRequest(r);
 247
 248     if (urlres_r == NULL)
 249         return;
 250
 251     StoreEntry::getPublic (this, urlres, METHOD_GET);
 252 }
 253
 254 void
 255 UrnState::created(StoreEntry *newEntry)
 256 {
 257     urlres_e = newEntry;
 258
 259     if (urlres_e->isNull()) {
 260         urlres_e = storeCreateEntry(urlres, urlres, request_flags(), METHOD_GET);
 261         sc = storeClientListAdd(urlres_e, this);
 262         FwdState::fwdStart(Comm::ConnectionPointer(), urlres_e, urlres_r);
 263     } else {
 264
 265         urlres_e->lock();
 266         sc = storeClientListAdd(urlres_e, this);
 267     }
 268
 269     reqofs = 0;
 270     StoreIOBuffer tempBuffer;
 271     tempBuffer.offset = reqofs;
 272     tempBuffer.length = URN_REQBUF_SZ;
 273     tempBuffer.data = reqbuf;
 274     storeClientCopy(sc, urlres_e,
 275                     tempBuffer,
 276                     urnHandleReply,
 277                     this);
 278 }
 279
 280 void
 281 urnStart(HttpRequest * r, StoreEntry * e)
 282 {
 283     UrnState *anUrn = new UrnState();
 284     anUrn->start (r, e);
 285 }
 286
 287 static int
 288 url_entry_sort(const void *A, const void *B)
 289 {
 290     const url_entry *u1 = (const url_entry *)A;
 291     const url_entry *u2 = (const url_entry *)B;
 292
 293     if (u2->rtt == u1->rtt)
 294         return 0;
 295     else if (0 == u1->rtt)
 296         return 1;
 297     else if (0 == u2->rtt)
 298         return -1;
 299     else
 300         return u1->rtt - u2->rtt;
 301 }
 302
 303 static void
 304 urnHandleReplyError(UrnState *urnState, StoreEntry *urlres_e)
 305 {
 306     urlres_e->unlock();
 307     urnState->entry->unlock();
 308     HTTPMSGUNLOCK(urnState->request);
 309     HTTPMSGUNLOCK(urnState->urlres_r);
 310     delete urnState;
 311 }
 312
 313 /* TODO: use the clientStream support for this */
 314 static void
 315 urnHandleReply(void *data, StoreIOBuffer result)
 316 {
 317     UrnState *urnState = static_cast<UrnState *>(data);
 318     StoreEntry *e = urnState->entry;
 319     StoreEntry *urlres_e = urnState->urlres_e;
 320     char *s = NULL;
 321     size_t k;
 322     HttpReply *rep;
 323     url_entry *urls;
 324     url_entry *u;
 325     url_entry *min_u;
 326     MemBuf *mb = NULL;
 327     ErrorState *err;
 328     int i;
 329     int urlcnt = 0;
 330     char *buf = urnState->reqbuf;
 331     StoreIOBuffer tempBuffer;
 332
 333     debugs(52, 3, "urnHandleReply: Called with size=" << result.length << ".");
 334
 335     if (EBIT_TEST(urlres_e->flags, ENTRY_ABORTED) || result.length == 0 || result.flags.error) {
 336         urnHandleReplyError(urnState, urlres_e);
 337         return;
 338     }
 339
 340     /* Update reqofs to point to where in the buffer we'd be */
 341     urnState->reqofs += result.length;
 342
 343     /* Handle reqofs being bigger than normal */
 344     if (urnState->reqofs >= URN_REQBUF_SZ) {
 345         urnHandleReplyError(urnState, urlres_e);
 346         return;
 347     }
 348
 349     /* If we haven't received the entire object (urn), copy more */
 350     if (urlres_e->store_status == STORE_PENDING &&
 351             urnState->reqofs < URN_REQBUF_SZ) {
 352         tempBuffer.offset = urnState->reqofs;
 353         tempBuffer.length = URN_REQBUF_SZ;
 354         tempBuffer.data = urnState->reqbuf + urnState->reqofs;
 355         storeClientCopy(urnState->sc, urlres_e,
 356                         tempBuffer,
 357                         urnHandleReply,
 358                         urnState);
 359         return;
 360     }
 361
 362     /* we know its STORE_OK */
 363     k = headersEnd(buf, urnState->reqofs);
 364
 365     if (0 == k) {
 366         debugs(52, DBG_IMPORTANT, "urnHandleReply: didn't find end-of-headers for " << e->url()  );
 367         urnHandleReplyError(urnState, urlres_e);
 368         return;
 369     }
 370
 371     s = buf + k;
 372     assert(urlres_e->getReply());
 373     rep = new HttpReply;
 374     rep->parseCharBuf(buf, k);
 375     debugs(52, 3, "reply exists, code=" << rep->sline.status << ".");
 376
 377     if (rep->sline.status != HTTP_OK) {
 378         debugs(52, 3, "urnHandleReply: failed.");
 379         err = new ErrorState(ERR_URN_RESOLVE, HTTP_NOT_FOUND, urnState->request);
 380         err->url = xstrdup(e->url());
 381         errorAppendEntry(e, err);
 382         delete rep;
 383         urnHandleReplyError(urnState, urlres_e);
 384         return;
 385     }
 386
 387     delete rep;
 388
 389     while (xisspace(*s))
 390         ++s;
 391
 392     urls = urnParseReply(s, urnState->request->method);
 393
 394     for (i = 0; NULL != urls[i].url; ++i)
 395         ++urlcnt;
 396
 397     debugs(53, 3, "urnFindMinRtt: Counted " << i << " URLs");
 398
 399     if (urls == NULL) {         /* unkown URN error */
 400         debugs(52, 3, "urnTranslateDone: unknown URN " << e->url()  );
 401         err = new ErrorState(ERR_URN_RESOLVE, HTTP_NOT_FOUND, urnState->request);
 402         err->url = xstrdup(e->url());
 403         errorAppendEntry(e, err);
 404         urnHandleReplyError(urnState, urlres_e);
 405         return;
 406     }
 407
 408     min_u = urnFindMinRtt(urls, urnState->request->method, NULL);
 409     qsort(urls, urlcnt, sizeof(*urls), url_entry_sort);
 410     e->buffer();
 411     mb = new MemBuf;
 412     mb->init();
 413     mb->Printf( "<TITLE>Select URL for %s</TITLE>\n"
 414                 "<STYLE type=\"text/css\"><!--BODY{background-color:#ffffff;font-family:verdana,sans-serif}--></STYLE>\n"
 415                 "<H2>Select URL for %s</H2>\n"
 416                 "<TABLE BORDER=\"0\" WIDTH=\"100%%\">\n", e->url(), e->url());
 417
 418     for (i = 0; i < urlcnt; ++i) {
 419         u = &urls[i];
 420         debugs(52, 3, "URL {" << u->url << "}");
 421         mb->Printf(
 422             "<TR><TD><A HREF=\"%s\">%s</A></TD>", u->url, u->url);
 423
 424         if (urls[i].rtt > 0)
 425             mb->Printf(
 426                 "<TD align=\"right\">%4d <it>ms</it></TD>", u->rtt);
 427         else
 428             mb->Printf("<TD align=\"right\">Unknown</TD>");
 429
 430         mb->Printf(
 431             "<TD>%s</TD></TR>\n", u->flags.cached ? "    [cached]" : " ");
 432     }
 433
 434     mb->Printf(
 435         "</TABLE>"
 436         "<HR noshade size=\"1px\">\n"
 437         "<ADDRESS>\n"
 438         "Generated by %s@%s\n"
 439         "</ADDRESS>\n",
 440         APP_FULLNAME, getMyHostname());
 441     rep = new HttpReply;
 442     rep->setHeaders(HTTP_MOVED_TEMPORARILY, NULL, "text/html", mb->contentSize(), 0, squid_curtime);
 443
 444     if (urnState->flags.force_menu) {
 445         debugs(51, 3, "urnHandleReply: forcing menu");
 446     } else if (min_u) {
 447         rep->header.putStr(HDR_LOCATION, min_u->url);
 448     }
 449
 450     rep->body.setMb(mb);
 451     /* don't clean or delete mb; rep->body owns it now */
 452     e->replaceHttpReply(rep);
 453     e->complete();
 454
 455     for (i = 0; i < urlcnt; ++i) {
 456         safe_free(urls[i].url);
 457         safe_free(urls[i].host);
 458     }
 459
 460     safe_free(urls);
 461     /* mb was absorbed in httpBodySet call, so we must not clean it */
 462     storeUnregister(urnState->sc, urlres_e, urnState);
 463
 464     urnHandleReplyError(urnState, urlres_e);
 465 }
 466
 467 static url_entry *
 468 urnParseReply(const char *inbuf, const HttpRequestMethod& m)
 469 {
 470     char *buf = xstrdup(inbuf);
 471     char *token;
 472     char *url;
 473     char *host;
 474     url_entry *list;
 475     url_entry *old;
 476     int n = 32;
 477     int i = 0;
 478     debugs(52, 3, "urnParseReply");
 479     list = (url_entry *)xcalloc(n + 1, sizeof(*list));
 480
 481     for (token = strtok(buf, crlf); token; token = strtok(NULL, crlf)) {
 482         debugs(52, 3, "urnParseReply: got '" << token << "'");
 483
 484         if (i == n) {
 485             old = list;
 486             n <<= 2;
 487             list = (url_entry *)xcalloc(n + 1, sizeof(*list));
 488             memcpy(list, old, i * sizeof(*list));
 489             safe_free(old);
 490         }
 491
 492         url = xstrdup(token);
 493         host = urlHostname(url);
 494
 495         if (NULL == host)
 496             continue;
 497
 498 #if USE_ICMP
 499         list[i].rtt = netdbHostRtt(host);
 500
 501         if (0 == list[i].rtt) {
 502             debugs(52, 3, "urnParseReply: Pinging " << host);
 503             netdbPingSite(host);
 504         }
 505 #else
 506         list[i].rtt = 0;
 507 #endif
 508
 509         list[i].url = url;
 510         list[i].host = xstrdup(host);
 511         // TODO: Use storeHas() or lock/unlock entry to avoid creating unlocked
 512         // ones.
 513         list[i].flags.cached = storeGetPublic(url, m) ? 1 : 0;
 514         ++i;
 515     }
 516
 517     debugs(52, 3, "urnParseReply: Found " << i << " URLs");
 518     return list;
 519 }