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