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  * Build a string URL from a lttng_uri object.
 142  */
 143 __attribute__((visibility("hidden")))
 144 int uri_to_str_url(struct lttng_uri *uri, char *dst, size_t size)
 145 {
 146         int ipver, ret;
 147         const char *addr;
 148         char proto[4], port[7];
 149
 150         assert(uri);
 151         assert(dst);
 152
 153         if (uri->dtype == LTTNG_DST_PATH) {
 154                 ipver = 0;
 155                 addr = uri->dst.path;
 156                 (void) snprintf(proto, sizeof(proto), "file");
 157                 (void) snprintf(port, sizeof(port), "%s", "");
 158         } else {
 159                 ipver = (uri->dtype == LTTNG_DST_IPV4) ? 4 : 6;
 160                 addr = (ipver == 4) ?  uri->dst.ipv4 : uri->dst.ipv6;
 161                 (void) snprintf(proto, sizeof(proto), "net%d", ipver);
 162                 (void) snprintf(port, sizeof(port), ":%d", uri->port);
 163         }
 164
 165         ret = snprintf(dst, size, "%s://%s%s%s%s/%s", proto,
 166                         (ipver == 6) ? "[" : "", addr, (ipver == 6) ? "]" : "",
 167                         port, uri->subdir);
 168         if (ret < 0) {
 169                 PERROR("snprintf uri to url");
 170         }
 171
 172         return ret;
 173 }
 174
 175 /*
 176  * Compare two URIs.
 177  *
 178  * Return 0 if equal else 1.
 179  */
 180 __attribute__((visibility("hidden")))
 181 int uri_compare(struct lttng_uri *uri1, struct lttng_uri *uri2)
 182 {
 183         return memcmp(uri1, uri2, sizeof(struct lttng_uri));
 184 }
 185
 186 /*
 187  * Free URI memory.
 188  */
 189 __attribute__((visibility("hidden")))
 190 void uri_free(struct lttng_uri *uri)
 191 {
 192         /* Safety check */
 193         if (uri != NULL) {
 194                 free(uri);
 195         }
 196 }
 197
 198 /*
 199  * Return an allocated URI.
 200  */
 201 __attribute__((visibility("hidden")))
 202 struct lttng_uri *uri_create(void)
 203 {
 204         struct lttng_uri *uri;
 205
 206         uri = zmalloc(sizeof(struct lttng_uri));
 207         if (uri == NULL) {
 208                 PERROR("zmalloc uri");
 209         }
 210
 211         return uri;
 212 }
 213
 214 /*
 215  * Parses a string URI to a lttng_uri. This function can potentially return
 216  * more than one URI in uris so the size of the array is returned and uris is
 217  * allocated and populated. Caller must free(3) the array.
 218  *
 219  * This function can not detect the stream type of the URI so the caller has to
 220  * make sure the correct type (stype) is set on the return URI(s). The default
 221  * port must also be set by the caller if the returned URI has its port set to
 222  * zero.
 223  *
 224  * NOTE: A good part of the following code was inspired from the "wget" source
 225  * tree from the src/url.c file and url_parse() function. Also, the
 226  * strpbrk_or_eos() function found above is also inspired by the same code.
 227  * This code was originally licensed GPLv2 so we acknolwedge the Free Software
 228  * Foundation here for the work and to make sure we are compliant with it.
 229  */
 230 __attribute__((visibility("hidden")))
 231 ssize_t uri_parse(const char *str_uri, struct lttng_uri **uris)
 232 {
 233         int ret, i = 0;
 234         /* Size of the uris array. Default is 1 */
 235         ssize_t size = 1;
 236         char subdir[PATH_MAX];
 237         unsigned int ctrl_port = 0;
 238         unsigned int data_port = 0;
 239         struct lttng_uri *tmp_uris;
 240         char *addr_f = NULL;
 241         const struct uri_proto *proto;
 242         const char *purl, *addr_e, *addr_b, *subdir_b = NULL;
 243         const char *seps = ":/\0";
 244
 245         /*
 246          * The first part is the protocol portion of a maximum of 5 bytes for now.
 247          * The second part is the hostname or IP address. The 255 bytes size is the
 248          * limit found in the RFC 1035 for the total length of a domain name
 249          * (https://www.ietf.org/rfc/rfc1035.txt). Finally, for the net://
 250          * protocol, two ports CAN be specified.
 251          */
 252
 253         DBG3("URI string: %s", str_uri);
 254
 255         proto = get_uri_proto(str_uri);
 256         if (proto == NULL) {
 257                 ERR("URI parse unknown protocol %s", str_uri);
 258                 goto error;
 259         }
 260
 261         purl = str_uri;
 262
 263         if (proto->code == P_NET || proto->code == P_NET6) {
 264                 /* Special case for net:// which requires two URI objects */
 265                 size = 2;
 266         }
 267
 268         /* Allocate URI array */
 269         tmp_uris = zmalloc(sizeof(struct lttng_uri) * size);
 270         if (tmp_uris == NULL) {
 271                 PERROR("zmalloc uri");
 272                 goto error;
 273         }
 274
 275         memset(subdir, 0, sizeof(subdir));
 276         purl += strlen(proto->leading_string);
 277
 278         /* Copy known value to the first URI. */
 279         tmp_uris[0].dtype = proto->dtype;
 280         tmp_uris[0].proto = proto->type;
 281
 282         if (proto->code == P_FILE) {
 283                 if (*purl != '/') {
 284                         ERR("Missing destination full path.");
 285                         goto free_error;
 286                 }
 287
 288                 strncpy(tmp_uris[0].dst.path, purl, sizeof(tmp_uris[0].dst.path));
 289                 tmp_uris[0].dst.path[sizeof(tmp_uris[0].dst.path) - 1] = '\0';
 290                 DBG3("URI file destination: %s", purl);
 291                 goto end;
 292         }
 293
 294         /* Assume we are at the beginning of an address or host of some sort. */
 295         addr_b = purl;
 296
 297         /*
 298          * Handle IPv6 address inside square brackets as mention by RFC 2732. IPv6
 299          * address that does not start AND end with brackets will be rejected even
 300          * if valid.
 301          *
 302          * proto://[<addr>]...
 303          *         ^
 304          */
 305         if (*purl == '[') {
 306                 /* Address begins after '[' */
 307                 addr_b = purl + 1;
 308                 addr_e = strchr(addr_b, ']');
 309                 if (addr_e == NULL || addr_b == addr_e) {
 310                         ERR("Broken IPv6 address %s", addr_b);
 311                         goto free_error;
 312                 }
 313
 314                 /* Moving parsed URL pointer after the final bracket ']' */
 315                 purl = addr_e + 1;
 316
 317                 /*
 318                  * The closing bracket must be followed by a seperator or NULL char.
 319                  */
 320                 if (strchr(seps, *purl) == NULL) {
 321                         ERR("Unknown symbol after IPv6 address: %s", purl);
 322                         goto free_error;
 323                 }
 324         } else {
 325                 purl = strpbrk_or_eos(purl, seps);
 326                 addr_e = purl;
 327         }
 328
 329         /* Check if we at least have a char for the addr or hostname. */
 330         if (addr_b == addr_e) {
 331                 ERR("No address or hostname detected.");
 332                 goto free_error;
 333         }
 334
 335         addr_f = utils_strdupdelim(addr_b, addr_e);
 336         if (addr_f == NULL) {
 337                 goto free_error;
 338         }
 339
 340         /*
 341          * Detect PORT after address. The net/net6 protocol allows up to two port
 342          * so we can define the control and data port.
 343          */
 344         while (*purl == ':') {
 345                 const char *port_b, *port_e;
 346                 char *port_f;
 347
 348                 /* Update pass counter */
 349                 i++;
 350
 351                 /*
 352                  * Maximum of two ports is possible if P_NET/NET6. Bigger than that,
 353                  * two much stuff.
 354                  */
 355                 if ((i == 2 && (proto->code != P_NET && proto->code != P_NET6))
 356                                 || i > 2) {
 357                         break;
 358                 }
 359
 360                 /*
 361                  * Move parsed URL to port value.
 362                  * proto://addr_host:PORT1:PORT2/foo/bar
 363                  *                   ^
 364                  */
 365                 ++purl;
 366                 port_b = purl;
 367                 purl = strpbrk_or_eos(purl, seps);
 368                 port_e = purl;
 369
 370                 if (port_b != port_e) {
 371                         int port;
 372
 373                         port_f = utils_strdupdelim(port_b, port_e);
 374                         if (port_f == NULL) {
 375                                 goto free_error;
 376                         }
 377
 378                         port = atoi(port_f);
 379                         if (port > 0xffff || port <= 0x0) {
 380                                 ERR("Invalid port number %d", port);
 381                                 free(port_f);
 382                                 goto free_error;
 383                         }
 384                         free(port_f);
 385
 386                         if (i == 1) {
 387                                 ctrl_port = port;
 388                         } else {
 389                                 data_port = port;
 390                         }
 391                 }
 392         };
 393
 394         /* Check for a valid subdir or trailing garbage */
 395         if (*purl == '/') {
 396                 /*
 397                  * Move to subdir value.
 398                  * proto://addr_host:PORT1:PORT2/foo/bar
 399                  *                               ^
 400                  */
 401                 ++purl;
 402                 subdir_b = purl;
 403         } else if (*purl != '\0') {
 404                 ERR("Trailing characters not recognized: %s", purl);
 405                 goto free_error;
 406         }
 407
 408         /* We have enough valid information to create URI(s) object */
 409
 410         /* Copy generic information */
 411         tmp_uris[0].port = ctrl_port;
 412
 413         /* Copy subdirectory if one. */
 414         if (subdir_b) {
 415                 strncpy(tmp_uris[0].subdir, subdir_b, sizeof(tmp_uris[0].subdir));
 416                 tmp_uris[0].subdir[sizeof(tmp_uris[0].subdir) - 1] = '\0';
 417         }
 418
 419         switch (proto->code) {
 420         case P_NET:
 421                 ret = set_ip_address(addr_f, AF_INET, tmp_uris[0].dst.ipv4,
 422                                 sizeof(tmp_uris[0].dst.ipv4));
 423                 if (ret < 0) {
 424                         goto free_error;
 425                 }
 426
 427                 memcpy(tmp_uris[1].dst.ipv4, tmp_uris[0].dst.ipv4, sizeof(tmp_uris[1].dst.ipv4));
 428
 429                 tmp_uris[1].dtype = proto->dtype;
 430                 tmp_uris[1].proto = proto->type;
 431                 tmp_uris[1].port = data_port;
 432                 break;
 433         case P_NET6:
 434                 ret = set_ip_address(addr_f, AF_INET6, tmp_uris[0].dst.ipv6,
 435                                 sizeof(tmp_uris[0].dst.ipv6));
 436                 if (ret < 0) {
 437                         goto free_error;
 438                 }
 439
 440                 memcpy(tmp_uris[1].dst.ipv6, tmp_uris[0].dst.ipv6, sizeof(tmp_uris[1].dst.ipv6));
 441
 442                 tmp_uris[1].dtype = proto->dtype;
 443                 tmp_uris[1].proto = proto->type;
 444                 tmp_uris[1].port = data_port;
 445                 break;
 446         case P_TCP:
 447                 ret = set_ip_address(addr_f, AF_INET, tmp_uris[0].dst.ipv4,
 448                                 sizeof(tmp_uris[0].dst.ipv4));
 449                 if (ret < 0) {
 450                         goto free_error;
 451                 }
 452                 break;
 453         case P_TCP6:
 454                 ret = set_ip_address(addr_f, AF_INET6, tmp_uris[0].dst.ipv6,
 455                                 sizeof(tmp_uris[0].dst.ipv6));
 456                 if (ret < 0) {
 457                         goto free_error;
 458                 }
 459                 break;
 460         default:
 461                 goto free_error;
 462         }
 463
 464 end:
 465         DBG3("URI dtype: %d, proto: %d, host: %s, subdir: %s, ctrl: %d, data: %d",
 466                         proto->dtype, proto->type, (addr_f == NULL) ? "" : addr_f,
 467                         (subdir_b == NULL) ? "" : subdir_b, ctrl_port, data_port);
 468
 469         free(addr_f);
 470
 471         *uris = tmp_uris;
 472         return size;
 473
 474 free_error:
 475         free(addr_f);
 476         free(tmp_uris);
 477 error:
 478         return -1;
 479 }