src/common/uri.c

   1 /*
   2  * Copyright (C) 2012 - David Goulet <dgoulet@efficios.com>
   3  *
   4  * This program is free software; you can redistribute it and/or modify it
   5  * under the terms of the GNU General Public License, version 2 only, as
   6  * published by the Free Software Foundation.
   7  *
   8  * This program is distributed in the hope that it will be useful, but WITHOUT
   9  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  10  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
  11  * more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along with
  14  * this program; if not, write to the Free Software Foundation, Inc., 51
  15  * Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  16  */
  17
  18 #define _GNU_SOURCE
  19 #define _LGPL_SOURCE
  20 #include <assert.h>
  21 #include <arpa/inet.h>
  22 #include <netdb.h>
  23 #include <stdlib.h>
  24 #include <string.h>
  25 #include <sys/socket.h>
  26
  27 #include <common/common.h>
  28 #include <common/defaults.h>
  29 #include <common/utils.h>
  30
  31 #include "uri.h"
  32
  33 enum uri_proto_code {
  34         P_NET, P_NET6, P_FILE, P_TCP, P_TCP6,
  35 };
  36
  37 struct uri_proto {
  38         const char *name;
  39         const char *leading_string;
  40         enum uri_proto_code code;
  41         enum lttng_proto_type type;
  42         enum lttng_dst_type dtype;
  43 };
  44
  45 /* Supported protocols */
  46 static const struct uri_proto proto_uri[] = {
  47         { .name = "file", .leading_string = "file://", .code = P_FILE, .type = 0, .dtype = LTTNG_DST_PATH },
  48         { .name = "net", .leading_string = "net://", .code = P_NET, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV4 },
  49         { .name = "net4", .leading_string = "net4://", .code = P_NET, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV4 },
  50         { .name = "net6", .leading_string = "net6://", .code = P_NET6, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV6 },
  51         { .name = "tcp", .leading_string = "tcp://", .code = P_TCP, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV4 },
  52         { .name = "tcp4", .leading_string = "tcp4://", .code = P_TCP, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV4 },
  53         { .name = "tcp6", .leading_string = "tcp6://", .code = P_TCP6, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV6 },
  54         /* Invalid proto marking the end of the array. */
  55         { NULL, NULL, 0, 0, 0 }
  56 };
  57
  58 /*
  59  * Return pointer to the character in s matching one of the characters in
  60  * accept. If nothing is found, return pointer to the end of string (eos).
  61  */
  62 static const inline char *strpbrk_or_eos(const char *s, const char *accept)
  63 {
  64         char *p = strpbrk(s, accept);
  65         if (p == NULL) {
  66                 p = strchr(s, '\0');
  67         }
  68
  69         return p;
  70 }
  71
  72
  73 /*
  74  * Validate if proto is a supported protocol from proto_uri array.
  75  */
  76 static const struct uri_proto *get_uri_proto(const char *uri_str)
  77 {
  78         const struct uri_proto *supported = NULL;
  79
  80         /* Safety net */
  81         if (uri_str == NULL) {
  82                 goto end;
  83         }
  84
  85         for (supported = &proto_uri[0];
  86                         supported->leading_string != NULL; ++supported) {
  87                 if (strncasecmp(uri_str, supported->leading_string,
  88                                         strlen(supported->leading_string)) == 0) {
  89                         goto end;
  90                 }
  91         }
  92
  93         /* Proto not found */
  94         return NULL;
  95
  96 end:
  97         return supported;
  98 }
  99
 100 /*
 101  * Set network address from string into dst. Supports both IP string and
 102  * hostname.
 103  */
 104 static int set_ip_address(const char *addr, int af, char *dst, size_t size)
 105 {
 106         int ret;
 107         unsigned char buf[sizeof(struct in6_addr)];
 108         struct hostent *record;
 109
 110         assert(addr);
 111         assert(dst);
 112
 113         memset(dst, 0, size);
 114
 115         /* Network protocol */
 116         ret = inet_pton(af, addr, buf);
 117         if (ret < 1) {
 118                 /* We consider the dst to be an hostname or an invalid IP char */
 119                 record = gethostbyname2(addr, af);
 120                 if (record == NULL) {
 121                         /* At this point, the IP or the hostname is bad */
 122                         ERR("URI parse bad hostname %s for af %d", addr, af);
 123                         goto error;
 124                 }
 125
 126                 /* Translate IP to string */
 127                 (void) inet_ntop(af, record->h_addr_list[0], dst, size);
 128         } else {
 129                 if (size > 0) {
 130                         strncpy(dst, addr, size);
 131                         dst[size - 1] = '\0';
 132                 }
 133         }
 134
 135         DBG2("IP address resolved to %s", dst);
 136
 137         return 0;
 138
 139 error:
 140         return -1;
 141 }
 142
 143 /*
 144  * Set default URI attribute which is basically the given stream type and the
 145  * default port if none is set in the URI.
 146  */
 147 static void set_default_uri_attr(struct lttng_uri *uri,
 148                 enum lttng_stream_type stype)
 149 {
 150         uri->stype = stype;
 151         if (uri->dtype != LTTNG_DST_PATH && uri->port == 0) {
 152                 uri->port = (stype == LTTNG_STREAM_CONTROL) ?
 153                         DEFAULT_NETWORK_CONTROL_PORT : DEFAULT_NETWORK_DATA_PORT;
 154         }
 155 }
 156
 157 /*
 158  * Compare two URL destination.
 159  *
 160  * Return 0 is equal else is not equal.
 161  */
 162 static int compare_destination(struct lttng_uri *ctrl, struct lttng_uri *data)
 163 {
 164         int ret;
 165
 166         assert(ctrl);
 167         assert(data);
 168
 169         switch (ctrl->dtype) {
 170         case LTTNG_DST_IPV4:
 171                 ret = strncmp(ctrl->dst.ipv4, data->dst.ipv4, sizeof(ctrl->dst.ipv4));
 172                 break;
 173         case LTTNG_DST_IPV6:
 174                 ret = strncmp(ctrl->dst.ipv6, data->dst.ipv6, sizeof(ctrl->dst.ipv6));
 175                 break;
 176         default:
 177                 ret = -1;
 178                 break;
 179         }
 180
 181         return ret;
 182 }
 183
 184 /*
 185  * Build a string URL from a lttng_uri object.
 186  */
 187 LTTNG_HIDDEN
 188 int uri_to_str_url(struct lttng_uri *uri, char *dst, size_t size)
 189 {
 190         int ipver, ret;
 191         const char *addr;
 192         char proto[5], port[7];
 193
 194         assert(uri);
 195         assert(dst);
 196
 197         if (uri->dtype == LTTNG_DST_PATH) {
 198                 ipver = 0;
 199                 addr = uri->dst.path;
 200                 (void) snprintf(proto, sizeof(proto), "file");
 201                 (void) snprintf(port, sizeof(port), "%s", "");
 202         } else {
 203                 ipver = (uri->dtype == LTTNG_DST_IPV4) ? 4 : 6;
 204                 addr = (ipver == 4) ?  uri->dst.ipv4 : uri->dst.ipv6;
 205                 (void) snprintf(proto, sizeof(proto), "tcp%d", ipver);
 206                 (void) snprintf(port, sizeof(port), ":%d", uri->port);
 207         }
 208
 209         ret = snprintf(dst, size, "%s://%s%s%s%s/%s", proto,
 210                         (ipver == 6) ? "[" : "", addr, (ipver == 6) ? "]" : "",
 211                         port, uri->subdir);
 212         if (ret < 0) {
 213                 PERROR("snprintf uri to url");
 214         }
 215
 216         return ret;
 217 }
 218
 219 /*
 220  * Compare two URIs.
 221  *
 222  * Return 0 if equal else 1.
 223  */
 224 LTTNG_HIDDEN
 225 int uri_compare(struct lttng_uri *uri1, struct lttng_uri *uri2)
 226 {
 227         return memcmp(uri1, uri2, sizeof(struct lttng_uri));
 228 }
 229
 230 /*
 231  * Free URI memory.
 232  */
 233 LTTNG_HIDDEN
 234 void uri_free(struct lttng_uri *uri)
 235 {
 236         free(uri);
 237 }
 238
 239 /*
 240  * Return an allocated URI.
 241  */
 242 LTTNG_HIDDEN
 243 struct lttng_uri *uri_create(void)
 244 {
 245         struct lttng_uri *uri;
 246
 247         uri = zmalloc(sizeof(struct lttng_uri));
 248         if (uri == NULL) {
 249                 PERROR("zmalloc uri");
 250         }
 251
 252         return uri;
 253 }
 254
 255 /*
 256  * Parses a string URI to a lttng_uri. This function can potentially return
 257  * more than one URI in uris so the size of the array is returned and uris is
 258  * allocated and populated. Caller must free(3) the array.
 259  *
 260  * This function can not detect the stream type of the URI so the caller has to
 261  * make sure the correct type (stype) is set on the return URI(s). The default
 262  * port must also be set by the caller if the returned URI has its port set to
 263  * zero.
 264  *
 265  * NOTE: A good part of the following code was inspired from the "wget" source
 266  * tree from the src/url.c file and url_parse() function. Also, the
 267  * strpbrk_or_eos() function found above is also inspired by the same code.
 268  * This code was originally licensed GPLv2 so we acknolwedge the Free Software
 269  * Foundation here for the work and to make sure we are compliant with it.
 270  */
 271 LTTNG_HIDDEN
 272 ssize_t uri_parse(const char *str_uri, struct lttng_uri **uris)
 273 {
 274         int ret, i = 0;
 275         /* Size of the uris array. Default is 1 */
 276         ssize_t size = 1;
 277         char subdir[PATH_MAX];
 278         unsigned int ctrl_port = 0;
 279         unsigned int data_port = 0;
 280         struct lttng_uri *tmp_uris;
 281         char *addr_f = NULL;
 282         const struct uri_proto *proto;
 283         const char *purl, *addr_e, *addr_b, *subdir_b = NULL;
 284         const char *seps = ":/\0";
 285
 286         /*
 287          * The first part is the protocol portion of a maximum of 5 bytes for now.
 288          * The second part is the hostname or IP address. The 255 bytes size is the
 289          * limit found in the RFC 1035 for the total length of a domain name
 290          * (https://www.ietf.org/rfc/rfc1035.txt). Finally, for the net://
 291          * protocol, two ports CAN be specified.
 292          */
 293
 294         DBG3("URI string: %s", str_uri);
 295
 296         proto = get_uri_proto(str_uri);
 297         if (proto == NULL) {
 298                 ERR("URI parse unknown protocol %s", str_uri);
 299                 goto error;
 300         }
 301
 302         purl = str_uri;
 303
 304         if (proto->code == P_NET || proto->code == P_NET6) {
 305                 /* Special case for net:// which requires two URI objects */
 306                 size = 2;
 307         }
 308
 309         /* Allocate URI array */
 310         tmp_uris = zmalloc(sizeof(struct lttng_uri) * size);
 311         if (tmp_uris == NULL) {
 312                 PERROR("zmalloc uri");
 313                 goto error;
 314         }
 315
 316         memset(subdir, 0, sizeof(subdir));
 317         purl += strlen(proto->leading_string);
 318
 319         /* Copy known value to the first URI. */
 320         tmp_uris[0].dtype = proto->dtype;
 321         tmp_uris[0].proto = proto->type;
 322
 323         if (proto->code == P_FILE) {
 324                 if (*purl != '/') {
 325                         ERR("Missing destination full path.");
 326                         goto free_error;
 327                 }
 328
 329                 strncpy(tmp_uris[0].dst.path, purl, sizeof(tmp_uris[0].dst.path));
 330                 tmp_uris[0].dst.path[sizeof(tmp_uris[0].dst.path) - 1] = '\0';
 331                 DBG3("URI file destination: %s", purl);
 332                 goto end;
 333         }
 334
 335         /* Assume we are at the beginning of an address or host of some sort. */
 336         addr_b = purl;
 337
 338         /*
 339          * Handle IPv6 address inside square brackets as mention by RFC 2732. IPv6
 340          * address that does not start AND end with brackets will be rejected even
 341          * if valid.
 342          *
 343          * proto://[<addr>]...
 344          *         ^
 345          */
 346         if (*purl == '[') {
 347                 /* Address begins after '[' */
 348                 addr_b = purl + 1;
 349                 addr_e = strchr(addr_b, ']');
 350                 if (addr_e == NULL || addr_b == addr_e) {
 351                         ERR("Broken IPv6 address %s", addr_b);
 352                         goto free_error;
 353                 }
 354
 355                 /* Moving parsed URL pointer after the final bracket ']' */
 356                 purl = addr_e + 1;
 357
 358                 /*
 359                  * The closing bracket must be followed by a seperator or NULL char.
 360                  */
 361                 if (strchr(seps, *purl) == NULL) {
 362                         ERR("Unknown symbol after IPv6 address: %s", purl);
 363                         goto free_error;
 364                 }
 365         } else {
 366                 purl = strpbrk_or_eos(purl, seps);
 367                 addr_e = purl;
 368         }
 369
 370         /* Check if we at least have a char for the addr or hostname. */
 371         if (addr_b == addr_e) {
 372                 ERR("No address or hostname detected.");
 373                 goto free_error;
 374         }
 375
 376         addr_f = utils_strdupdelim(addr_b, addr_e);
 377         if (addr_f == NULL) {
 378                 goto free_error;
 379         }
 380
 381         /*
 382          * Detect PORT after address. The net/net6 protocol allows up to two port
 383          * so we can define the control and data port.
 384          */
 385         while (*purl == ':') {
 386                 const char *port_b, *port_e;
 387                 char *port_f;
 388
 389                 /* Update pass counter */
 390                 i++;
 391
 392                 /*
 393                  * Maximum of two ports is possible if P_NET/NET6. Bigger than that,
 394                  * two much stuff.
 395                  */
 396                 if ((i == 2 && (proto->code != P_NET && proto->code != P_NET6))
 397                                 || i > 2) {
 398                         break;
 399                 }
 400
 401                 /*
 402                  * Move parsed URL to port value.
 403                  * proto://addr_host:PORT1:PORT2/foo/bar
 404                  *                   ^
 405                  */
 406                 ++purl;
 407                 port_b = purl;
 408                 purl = strpbrk_or_eos(purl, seps);
 409                 port_e = purl;
 410
 411                 if (port_b != port_e) {
 412                         int port;
 413
 414                         port_f = utils_strdupdelim(port_b, port_e);
 415                         if (port_f == NULL) {
 416                                 goto free_error;
 417                         }
 418
 419                         port = atoi(port_f);
 420                         if (port > 0xffff || port <= 0x0) {
 421                                 ERR("Invalid port number %d", port);
 422                                 free(port_f);
 423                                 goto free_error;
 424                         }
 425                         free(port_f);
 426
 427                         if (i == 1) {
 428                                 ctrl_port = port;
 429                         } else {
 430                                 data_port = port;
 431                         }
 432                 }
 433         };
 434
 435         /* Check for a valid subdir or trailing garbage */
 436         if (*purl == '/') {
 437                 /*
 438                  * Move to subdir value.
 439                  * proto://addr_host:PORT1:PORT2/foo/bar
 440                  *                               ^
 441                  */
 442                 ++purl;
 443                 subdir_b = purl;
 444         } else if (*purl != '\0') {
 445                 ERR("Trailing characters not recognized: %s", purl);
 446                 goto free_error;
 447         }
 448
 449         /* We have enough valid information to create URI(s) object */
 450
 451         /* Copy generic information */
 452         tmp_uris[0].port = ctrl_port;
 453
 454         /* Copy subdirectory if one. */
 455         if (subdir_b) {
 456                 strncpy(tmp_uris[0].subdir, subdir_b, sizeof(tmp_uris[0].subdir));
 457                 tmp_uris[0].subdir[sizeof(tmp_uris[0].subdir) - 1] = '\0';
 458         }
 459
 460         switch (proto->code) {
 461         case P_NET:
 462                 ret = set_ip_address(addr_f, AF_INET, tmp_uris[0].dst.ipv4,
 463                                 sizeof(tmp_uris[0].dst.ipv4));
 464                 if (ret < 0) {
 465                         goto free_error;
 466                 }
 467
 468                 memcpy(tmp_uris[1].dst.ipv4, tmp_uris[0].dst.ipv4, sizeof(tmp_uris[1].dst.ipv4));
 469
 470                 tmp_uris[1].dtype = proto->dtype;
 471                 tmp_uris[1].proto = proto->type;
 472                 tmp_uris[1].port = data_port;
 473                 break;
 474         case P_NET6:
 475                 ret = set_ip_address(addr_f, AF_INET6, tmp_uris[0].dst.ipv6,
 476                                 sizeof(tmp_uris[0].dst.ipv6));
 477                 if (ret < 0) {
 478                         goto free_error;
 479                 }
 480
 481                 memcpy(tmp_uris[1].dst.ipv6, tmp_uris[0].dst.ipv6, sizeof(tmp_uris[1].dst.ipv6));
 482
 483                 tmp_uris[1].dtype = proto->dtype;
 484                 tmp_uris[1].proto = proto->type;
 485                 tmp_uris[1].port = data_port;
 486                 break;
 487         case P_TCP:
 488                 ret = set_ip_address(addr_f, AF_INET, tmp_uris[0].dst.ipv4,
 489                                 sizeof(tmp_uris[0].dst.ipv4));
 490                 if (ret < 0) {
 491                         goto free_error;
 492                 }
 493                 break;
 494         case P_TCP6:
 495                 ret = set_ip_address(addr_f, AF_INET6, tmp_uris[0].dst.ipv6,
 496                                 sizeof(tmp_uris[0].dst.ipv6));
 497                 if (ret < 0) {
 498                         goto free_error;
 499                 }
 500                 break;
 501         default:
 502                 goto free_error;
 503         }
 504
 505 end:
 506         DBG3("URI dtype: %d, proto: %d, host: %s, subdir: %s, ctrl: %d, data: %d",
 507                         proto->dtype, proto->type, (addr_f == NULL) ? "" : addr_f,
 508                         (subdir_b == NULL) ? "" : subdir_b, ctrl_port, data_port);
 509
 510         free(addr_f);
 511
 512         *uris = tmp_uris;
 513         return size;
 514
 515 free_error:
 516         free(addr_f);
 517         free(tmp_uris);
 518 error:
 519         return -1;
 520 }
 521
 522 /*
 523  * Parse a string URL and creates URI(s) returning the size of the populated
 524  * array.
 525  */
 526 LTTNG_HIDDEN
 527 ssize_t uri_parse_str_urls(const char *ctrl_url, const char *data_url,
 528                 struct lttng_uri **uris)
 529 {
 530         unsigned int equal = 1, idx = 0;
 531         /* Add the "file://" size to the URL maximum size */
 532         char url[PATH_MAX + 7];
 533         ssize_t size_ctrl = 0, size_data = 0, size;
 534         struct lttng_uri *ctrl_uris = NULL, *data_uris = NULL;
 535         struct lttng_uri *tmp_uris = NULL;
 536
 537         /* No URL(s) is allowed. This means that the consumer will be disabled. */
 538         if (ctrl_url == NULL && data_url == NULL) {
 539                 return 0;
 540         }
 541
 542         /* Check if URLs are equal and if so, only use the control URL */
 543         if ((ctrl_url && *ctrl_url != '\0') && (data_url && *data_url != '\0')) {
 544                 equal = !strcmp(ctrl_url, data_url);
 545         }
 546
 547         /*
 548          * Since we allow the str_url to be a full local filesystem path, we are
 549          * going to create a valid file:// URL if it's the case.
 550          *
 551          * Check if first character is a '/' or else reject the URL.
 552          */
 553         if (ctrl_url && ctrl_url[0] == '/') {
 554                 int ret;
 555
 556                 ret = snprintf(url, sizeof(url), "file://%s", ctrl_url);
 557                 if (ret < 0) {
 558                         PERROR("snprintf file url");
 559                         goto parse_error;
 560                 }
 561                 ctrl_url = url;
 562         }
 563
 564         /* Parse the control URL if there is one */
 565         if (ctrl_url && *ctrl_url != '\0') {
 566                 size_ctrl = uri_parse(ctrl_url, &ctrl_uris);
 567                 if (size_ctrl < 1) {
 568                         ERR("Unable to parse the URL %s", ctrl_url);
 569                         goto parse_error;
 570                 }
 571
 572                 /* At this point, we know there is at least one URI in the array */
 573                 set_default_uri_attr(&ctrl_uris[0], LTTNG_STREAM_CONTROL);
 574
 575                 if (ctrl_uris[0].dtype == LTTNG_DST_PATH &&
 576                                 (data_url && *data_url != '\0')) {
 577                         ERR("Can not have a data URL when destination is file://");
 578                         goto error;
 579                 }
 580
 581                 /* URL are not equal but the control URL uses a net:// protocol */
 582                 if (size_ctrl == 2) {
 583                         if (!equal) {
 584                                 ERR("Control URL uses the net:// protocol and the data URL is "
 585                                                 "different. Not allowed.");
 586                                 goto error;
 587                         } else {
 588                                 set_default_uri_attr(&ctrl_uris[1], LTTNG_STREAM_DATA);
 589                                 /*
 590                                  * The data_url and ctrl_url are equal and the ctrl_url
 591                                  * contains a net:// protocol so we just skip the data part.
 592                                  */
 593                                 data_url = NULL;
 594                         }
 595                 }
 596         }
 597
 598         if (data_url && *data_url != '\0') {
 599                 int ret;
 600
 601                 /* We have to parse the data URL in this case */
 602                 size_data = uri_parse(data_url, &data_uris);
 603                 if (size_data < 1) {
 604                         ERR("Unable to parse the URL %s", data_url);
 605                         goto error;
 606                 } else if (size_data == 2) {
 607                         ERR("Data URL can not be set with the net[4|6]:// protocol");
 608                         goto error;
 609                 }
 610
 611                 set_default_uri_attr(&data_uris[0], LTTNG_STREAM_DATA);
 612
 613                 ret = compare_destination(&ctrl_uris[0], &data_uris[0]);
 614                 if (ret != 0) {
 615                         ERR("Control and data destination mismatch");
 616                         goto error;
 617                 }
 618         }
 619
 620         /* Compute total size */
 621         size = size_ctrl + size_data;
 622
 623         tmp_uris = zmalloc(sizeof(struct lttng_uri) * size);
 624         if (tmp_uris == NULL) {
 625                 PERROR("zmalloc uris");
 626                 goto error;
 627         }
 628
 629         if (ctrl_uris) {
 630                 /* It's possible the control URIs array contains more than one URI */
 631                 memcpy(tmp_uris, ctrl_uris, sizeof(struct lttng_uri) * size_ctrl);
 632                 ++idx;
 633                 free(ctrl_uris);
 634         }
 635
 636         if (data_uris) {
 637                 memcpy(&tmp_uris[idx], data_uris, sizeof(struct lttng_uri));
 638                 free(data_uris);
 639         }
 640
 641         *uris = tmp_uris;
 642
 643         return size;
 644
 645 error:
 646         free(ctrl_uris);
 647         free(data_uris);
 648         free(tmp_uris);
 649 parse_error:
 650         return -1;
 651 }