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