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