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 <arpa/inet.h>
  20 #include <netdb.h>
  21 #include <stdlib.h>
  22 #include <string.h>
  23 #include <sys/socket.h>
  24
  25 #include <common/common.h>
  26 #include <common/defaults.h>
  27 #include <common/utils.h>
  28
  29 #include "uri.h"
  30
  31 enum uri_proto_code {
  32         P_NET, P_NET6, P_FILE, P_TCP, P_TCP6,
  33 };
  34
  35 struct uri_proto {
  36         const char *name;
  37         const char *leading_string;
  38         enum uri_proto_code code;
  39         enum lttng_proto_type type;
  40         enum lttng_dst_type dtype;
  41 };
  42
  43 /* Supported protocols */
  44 static const struct uri_proto proto_uri[] = {
  45         { .name = "file", .leading_string = "file://", .code = P_FILE, .type = 0, .dtype = LTTNG_DST_PATH },
  46         { .name = "net", .leading_string = "net://", .code = P_NET, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV4 },
  47         { .name = "net6", .leading_string = "net6://", .code = P_NET6, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV6 },
  48         { .name = "tcp", .leading_string = "tcp://", .code = P_TCP, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV4 },
  49         { .name = "tcp6", .leading_string = "tcp6://", .code = P_TCP6, .type = LTTNG_TCP, .dtype = LTTNG_DST_IPV6 },
  50         /* Invalid proto marking the end of the array. */
  51         { NULL, NULL, 0, 0, 0 }
  52 };
  53
  54 /*
  55  * Return pointer to the character in s matching one of the characters in
  56  * accept. If nothing is found, return pointer to the end of string (eos).
  57  */
  58 const inline char *strpbrk_or_eos(const char *s, const char *accept)
  59 {
  60         char *p = strpbrk(s, accept);
  61         if (p == NULL) {
  62                 p = strchr(s, '\0');
  63         }
  64
  65         return p;
  66 }
  67
  68
  69 /*
  70  * Validate if proto is a supported protocol from proto_uri array.
  71  */
  72 static const struct uri_proto *get_uri_proto(const char *uri_str)
  73 {
  74         const struct uri_proto *supported = NULL;
  75
  76         /* Safety net */
  77         if (uri_str == NULL) {
  78                 goto end;
  79         }
  80
  81         for (supported = &proto_uri[0];
  82                         supported->leading_string != NULL; ++supported) {
  83                 if (strncasecmp(uri_str, supported->leading_string,
  84                                         strlen(supported->leading_string)) == 0) {
  85                         goto end;
  86                 }
  87         }
  88
  89         /* Proto not found */
  90         return NULL;
  91
  92 end:
  93         return supported;
  94 }
  95
  96 /*
  97  * Set network address from string into dst. Supports both IP string and
  98  * hostname.
  99  */
 100 static int set_ip_address(const char *addr, int af, char *dst, size_t size)
 101 {
 102         int ret;
 103         unsigned char buf[sizeof(struct in6_addr)];
 104         struct hostent *record;
 105
 106         /* Network protocol */
 107         ret = inet_pton(af, addr, buf);
 108         if (ret < 1) {
 109                 /* We consider the dst to be an hostname or an invalid IP char */
 110                 record = gethostbyname2(addr, af);
 111                 if (record == NULL) {
 112                         /* At this point, the IP or the hostname is bad */
 113                         ERR("URI parse bad hostname %s for af %d", addr, af);
 114                         goto error;
 115                 }
 116
 117                 /* Translate IP to string */
 118                 (void) inet_ntop(af, record->h_addr_list[0], dst, size);
 119         } else {
 120                 memcpy(dst, addr, size);
 121         }
 122
 123         DBG2("IP address resolved to %s", dst);
 124
 125         return 0;
 126
 127 error:
 128         return -1;
 129 }
 130
 131 /*
 132  * Compare two URIs.
 133  *
 134  * Return 0 if equal else 1.
 135  */
 136 int uri_compare(struct lttng_uri *uri1, struct lttng_uri *uri2)
 137 {
 138         return memcmp(uri1, uri2, sizeof(struct lttng_uri));
 139 }
 140
 141 /*
 142  * Free URI memory.
 143  */
 144 void uri_free(struct lttng_uri *uri)
 145 {
 146         /* Safety check */
 147         if (uri != NULL) {
 148                 free(uri);
 149         }
 150 }
 151
 152 /*
 153  * Return an allocated URI.
 154  */
 155 struct lttng_uri *uri_create(void)
 156 {
 157         struct lttng_uri *uri;
 158
 159         uri = zmalloc(sizeof(struct lttng_uri));
 160         if (uri == NULL) {
 161                 PERROR("zmalloc uri");
 162         }
 163
 164         return uri;
 165 }
 166
 167 /*
 168  * Parses a string URI to a lttng_uri. This function can potentially return
 169  * more than one URI in uris so the size of the array is returned and uris is
 170  * allocated and populated. Caller must free(3) the array.
 171  *
 172  * This function can not detect the stream type of the URI so the caller has to
 173  * make sure the correct type (stype) is set on the return URI(s). The default
 174  * port must also be set by the caller if the returned URI has its port set to
 175  * zero.
 176  *
 177  * NOTE: A good part of the following code was inspired from the "wget" source
 178  * tree from the src/url.c file and url_parse() function. Also, the
 179  * strpbrk_or_eos() function found above is also inspired by the same code.
 180  * This code was originally licensed GPLv2 so we acknolwedge the Free Software
 181  * Foundation here for the work and to make sure we are compliant with it.
 182  */
 183 ssize_t uri_parse(const char *str_uri, struct lttng_uri **uris)
 184 {
 185         int ret, i = 0;
 186         /* Size of the uris array. Default is 1 */
 187         ssize_t size = 1;
 188         char subdir[PATH_MAX];
 189         unsigned int ctrl_port = 0;
 190         unsigned int data_port = 0;
 191         struct lttng_uri *tmp_uris;
 192         char *addr_f = NULL;
 193         const struct uri_proto *proto;
 194         const char *purl, *addr_e, *addr_b, *subdir_b = NULL;
 195         const char *seps = ":/\0";
 196
 197         /*
 198          * The first part is the protocol portion of a maximum of 5 bytes for now.
 199          * The second part is the hostname or IP address. The 255 bytes size is the
 200          * limit found in the RFC 1035 for the total length of a domain name
 201          * (https://www.ietf.org/rfc/rfc1035.txt). Finally, for the net://
 202          * protocol, two ports CAN be specified.
 203          */
 204
 205         DBG3("URI string: %s", str_uri);
 206
 207         proto = get_uri_proto(str_uri);
 208         if (proto == NULL) {
 209                 ERR("URI parse unknown protocol %s", str_uri);
 210                 goto error;
 211         }
 212
 213         purl = str_uri;
 214
 215         if (proto->code == P_NET || proto->code == P_NET6) {
 216                 /* Special case for net:// which requires two URI objects */
 217                 size = 2;
 218         }
 219
 220         /* Allocate URI array */
 221         tmp_uris = zmalloc(sizeof(struct lttng_uri) * size);
 222         if (tmp_uris == NULL) {
 223                 PERROR("zmalloc uri");
 224                 goto error;
 225         }
 226
 227         memset(subdir, 0, sizeof(subdir));
 228         purl += strlen(proto->leading_string);
 229
 230         /* Copy known value to the first URI. */
 231         tmp_uris[0].dtype = proto->dtype;
 232         tmp_uris[0].proto = proto->type;
 233
 234         if (proto->code == P_FILE) {
 235                 if (*purl != '/') {
 236                         ERR("Missing destination full path.");
 237                         goto free_error;
 238                 }
 239
 240                 strncpy(tmp_uris[0].dst.path, purl, sizeof(tmp_uris[0].dst.path));
 241                 tmp_uris[0].dst.path[sizeof(tmp_uris[0].dst.path) - 1] = '\0';
 242                 DBG3("URI file destination: %s", purl);
 243                 goto end;
 244         }
 245
 246         /* Assume we are at the beginning of an address or host of some sort. */
 247         addr_b = purl;
 248
 249         /*
 250          * Handle IPv6 address inside square brackets as mention by RFC 2732. IPv6
 251          * address that does not start AND end with brackets will be rejected even
 252          * if valid.
 253          *
 254          * proto://[<addr>]...
 255          *         ^
 256          */
 257         if (*purl == '[') {
 258                 /* Address begins after '[' */
 259                 addr_b = purl + 1;
 260                 addr_e = strchr(addr_b, ']');
 261                 if (addr_e == NULL || addr_b == addr_e) {
 262                         ERR("Broken IPv6 address %s", addr_b);
 263                         goto free_error;
 264                 }
 265
 266                 /* Moving parsed URL pointer after the final bracket ']' */
 267                 purl = addr_e + 1;
 268
 269                 /*
 270                  * The closing bracket must be followed by a seperator or NULL char.
 271                  */
 272                 if (strchr(seps, *purl) == NULL) {
 273                         ERR("Unknown symbol after IPv6 address: %s", purl);
 274                         goto free_error;
 275                 }
 276         } else {
 277                 purl = strpbrk_or_eos(purl, seps);
 278                 addr_e = purl;
 279         }
 280
 281         /* Check if we at least have a char for the addr or hostname. */
 282         if (addr_b == addr_e) {
 283                 ERR("No address or hostname detected.");
 284                 goto free_error;
 285         }
 286
 287         addr_f = utils_strdupdelim(addr_b, addr_e);
 288         if (addr_f == NULL) {
 289                 goto free_error;
 290         }
 291
 292         /*
 293          * Detect PORT after address. The net/net6 protocol allows up to two port
 294          * so we can define the control and data port.
 295          */
 296         while (*purl == ':') {
 297                 int port;
 298                 const char *port_b, *port_e;
 299                 char *port_f;
 300
 301                 /* Update pass counter */
 302                 i++;
 303
 304                 /*
 305                  * Maximum of two ports is possible if P_NET/NET6. Bigger than that,
 306                  * two much stuff.
 307                  */
 308                 if ((i == 2 && (proto->code != P_NET && proto->code != P_NET6))
 309                                 || i > 2) {
 310                         break;
 311                 }
 312
 313                 /*
 314                  * Move parsed URL to port value.
 315                  * proto://addr_host:PORT1:PORT2/foo/bar
 316                  *                   ^
 317                  */
 318                 ++purl;
 319                 port_b = purl;
 320                 purl = strpbrk_or_eos(purl, seps);
 321                 port_e = purl;
 322
 323                 if (port_b != port_e) {
 324                         port_f = utils_strdupdelim(port_b, port_e);
 325                         if (port_f == NULL) {
 326                                 goto free_error;
 327                         }
 328
 329                         port = atoi(port_f);
 330                         if (port > 0xffff || port <= 0x0) {
 331                                 ERR("Invalid port number %d", port);
 332                                 free(port_f);
 333                                 goto free_error;
 334                         }
 335                         free(port_f);
 336
 337                         if (i == 1) {
 338                                 ctrl_port = port;
 339                         } else {
 340                                 data_port = port;
 341                         }
 342                 }
 343         };
 344
 345         /* Check for a valid subdir or trailing garbage */
 346         if (*purl == '/') {
 347                 /*
 348                  * Move to subdir value.
 349                  * proto://addr_host:PORT1:PORT2/foo/bar
 350                  *                               ^
 351                  */
 352                 ++purl;
 353                 subdir_b = purl;
 354         } else if (*purl != '\0') {
 355                 ERR("Trailing characters not recognized: %s", purl);
 356                 goto free_error;
 357         }
 358
 359         /* We have enough valid information to create URI(s) object */
 360
 361         /* Copy generic information */
 362         tmp_uris[0].port = ctrl_port;
 363
 364         /* Copy subdirectory if one. */
 365         if (subdir_b) {
 366                 strncpy(tmp_uris[0].subdir, subdir_b, sizeof(tmp_uris[0].subdir));
 367                 tmp_uris[0].subdir[sizeof(tmp_uris[0].subdir) - 1] = '\0';
 368         }
 369
 370         switch (proto->code) {
 371         case P_NET:
 372                 ret = set_ip_address(addr_f, AF_INET, tmp_uris[0].dst.ipv4,
 373                                 sizeof(tmp_uris[0].dst.ipv4));
 374                 if (ret < 0) {
 375                         goto free_error;
 376                 }
 377
 378                 memcpy(tmp_uris[1].dst.ipv4, tmp_uris[0].dst.ipv4, sizeof(tmp_uris[1].dst.ipv4));
 379
 380                 tmp_uris[1].dtype = proto->dtype;
 381                 tmp_uris[1].proto = proto->type;
 382                 tmp_uris[1].port = data_port;
 383                 break;
 384         case P_NET6:
 385                 ret = set_ip_address(addr_f, AF_INET6, tmp_uris[0].dst.ipv6,
 386                                 sizeof(tmp_uris[0].dst.ipv6));
 387                 if (ret < 0) {
 388                         goto free_error;
 389                 }
 390
 391                 memcpy(tmp_uris[1].dst.ipv6, tmp_uris[0].dst.ipv6, sizeof(tmp_uris[1].dst.ipv6));
 392
 393                 tmp_uris[1].dtype = proto->dtype;
 394                 tmp_uris[1].proto = proto->type;
 395                 tmp_uris[1].port = data_port;
 396                 break;
 397         case P_TCP:
 398                 ret = set_ip_address(addr_f, AF_INET, tmp_uris[0].dst.ipv4,
 399                                 sizeof(tmp_uris[0].dst.ipv4));
 400                 if (ret < 0) {
 401                         goto free_error;
 402                 }
 403                 break;
 404         case P_TCP6:
 405                 ret = set_ip_address(addr_f, AF_INET6, tmp_uris[0].dst.ipv6,
 406                                 sizeof(tmp_uris[0].dst.ipv6));
 407                 if (ret < 0) {
 408                         goto free_error;
 409                 }
 410                 break;
 411         default:
 412                 goto free_error;
 413         }
 414
 415 end:
 416         DBG3("URI dtype: %d, proto: %d, host: %s, subdir: %s, ctrl: %d, data: %d",
 417                         proto->dtype, proto->type, (addr_f == NULL) ? "" : addr_f,
 418                         (subdir_b == NULL) ? "" : subdir_b, ctrl_port, data_port);
 419
 420         free(addr_f);
 421
 422         *uris = tmp_uris;
 423         return size;
 424
 425 free_error:
 426         free(addr_f);
 427         free(tmp_uris);
 428 error:
 429         return -1;
 430 }