path: root/src/feature/nodelist/routerset.c

/* Copyright (c) 2001 Matej Pfajfar.
n * Copyright (c) 2001-2004, Roger Dingledine.
 * Copyright (c) 2004-2006, Roger Dingledine, Nick Mathewson.
 * Copyright (c) 2007-2019, The Tor Project, Inc. */
/* See LICENSE for licensing information */

/**
 * \file routerset.c
 *
 * \brief Functions and structures to handle set-type selection of routers
 *  by name, ID, address, etc.
 *
 * This module implements the routerset_t data structure, whose purpose
 * is to specify a set of relays based on a list of their identities or
 * properties.  Routersets can restrict relays by IP address mask,
 * identity fingerprint, country codes, and nicknames (deprecated).
 *
 * Routersets are typically used for user-specified restrictions, and
 * are created by invoking routerset_new and routerset_parse from
 * config.c and confparse.c.  To use a routerset, invoke one of
 * routerset_contains_...() functions , or use
 * routerstatus_get_all_nodes() / routerstatus_subtract_nodes() to
 * manipulate a smartlist of node_t pointers.
 *
 * Country-code restrictions are implemented in geoip.c.
 */

#define ROUTERSET_PRIVATE

#include "core/or/or.h"
#include "core/or/policies.h"
#include "feature/client/bridges.h"
#include "feature/dirparse/policy_parse.h"
#include "feature/nodelist/nickname.h"
#include "feature/nodelist/nodelist.h"
#include "feature/nodelist/routerset.h"
#include "lib/geoip/geoip.h"

#include "core/or/addr_policy_st.h"
#include "core/or/extend_info_st.h"
#include "feature/nodelist/node_st.h"
#include "feature/nodelist/routerinfo_st.h"
#include "feature/nodelist/routerstatus_st.h"

/** Return a new empty routerset. */
routerset_t *
routerset_new(void)
{
  routerset_t *result = tor_malloc_zero(sizeof(routerset_t));
  result->list = smartlist_new();
  result->names = strmap_new();
  result->digests = digestmap_new();
  result->policies = smartlist_new();
  result->country_names = smartlist_new();
  return result;
}

/** If <b>c</b> is a country code in the form {cc}, return a newly allocated
 * string holding the "cc" part.  Else, return NULL. */
STATIC char *
routerset_get_countryname(const char *c)
{
  char *country;

  if (strlen(c) < 4 || c[0] !='{' || c[3] !='}')
    return NULL;

  country = tor_strndup(c+1, 2);
  tor_strlower(country);
  return country;
}

/** Update the routerset's <b>countries</b> bitarray_t. Called whenever
 * the GeoIP IPv4 database is reloaded.
 */
void
routerset_refresh_countries(routerset_t *target)
{
  int cc;
  bitarray_free(target->countries);

  if (!geoip_is_loaded(AF_INET)) {
    target->countries = NULL;
    target->n_countries = 0;
    return;
  }
  target->n_countries = geoip_get_n_countries();
  target->countries = bitarray_init_zero(target->n_countries);
  SMARTLIST_FOREACH_BEGIN(target->country_names, const char *, country) {
    cc = geoip_get_country(country);
    if (cc >= 0) {
      tor_assert(cc < target->n_countries);
      bitarray_set(target->countries, cc);
    } else {
      log_warn(LD_CONFIG, "Country code '%s' is not recognized.",
          country);
    }
  } SMARTLIST_FOREACH_END(country);
}

/** Parse the string <b>s</b> to create a set of routerset entries, and add
 * them to <b>target</b>.  In log messages, refer to the string as
 * <b>description</b>.  Return 0 on success, -1 on failure.
 *
 * Three kinds of elements are allowed in routersets: nicknames, IP address
 * patterns, and fingerprints.  They may be surrounded by optional space, and
 * must be separated by commas.
 */
int
routerset_parse(routerset_t *target, const char *s, const char *description)
{
  int r = 0;
  int added_countries = 0;
  char *countryname;
  smartlist_t *list = smartlist_new();
  int malformed_list;
  smartlist_split_string(list, s, ",",
                         SPLIT_SKIP_SPACE | SPLIT_IGNORE_BLANK, 0);
  SMARTLIST_FOREACH_BEGIN(list, char *, nick) {
      addr_policy_t *p;
      /* if it doesn't pass our validation, assume it's malformed */
      malformed_list = 1;
      if (is_legal_hexdigest(nick)) {
        char d[DIGEST_LEN];
        if (*nick == '$')
          ++nick;
        log_debug(LD_CONFIG, "Adding identity %s to %s", nick, description);
        base16_decode(d, sizeof(d), nick, HEX_DIGEST_LEN);
        digestmap_set(target->digests, d, (void*)1);
      } else if (is_legal_nickname(nick)) {
        log_debug(LD_CONFIG, "Adding nickname %s to %s", nick, description);
        strmap_set_lc(target->names, nick, (void*)1);
      } else if ((countryname = routerset_get_countryname(nick)) != NULL) {
        log_debug(LD_CONFIG, "Adding country %s to %s", nick,
                  description);
        smartlist_add(target->country_names, countryname);
        added_countries = 1;
      } else if ((strchr(nick,'.') || strchr(nick, ':') ||  strchr(nick, '*'))
                 && (p = router_parse_addr_policy_item_from_string(
                                     nick, ADDR_POLICY_REJECT,
                                     &malformed_list))) {
        /* IPv4 addresses contain '.', IPv6 addresses contain ':',
         * and wildcard addresses contain '*'. */
        log_debug(LD_CONFIG, "Adding address %s to %s", nick, description);
        smartlist_add(target->policies, p);
      } else if (malformed_list) {
        log_warn(LD_CONFIG, "Entry '%s' in %s is malformed. Discarding entire"
                 " list.", nick, description);
        r = -1;
        tor_free(nick);
        SMARTLIST_DEL_CURRENT(list, nick);
      } else {
        log_notice(LD_CONFIG, "Entry '%s' in %s is ignored. Using the"
                   " remainder of the list.", nick, description);
        tor_free(nick);
        SMARTLIST_DEL_CURRENT(list, nick);
      }
  } SMARTLIST_FOREACH_END(nick);
  policy_expand_unspec(&target->policies);
  smartlist_add_all(target->list, list);
  smartlist_free(list);
  if (added_countries)
    routerset_refresh_countries(target);
  return r;
}

/** Add all members of the set <b>source</b> to <b>target</b>. */
void
routerset_union(routerset_t *target, const routerset_t *source)
{
  char *s;
  tor_assert(target);
  if (!source || !source->list)
    return;
  s = routerset_to_string(source);
  routerset_parse(target, s, "other routerset");
  tor_free(s);
}

/** Return true iff <b>set</b> lists only nicknames and digests, and includes
 * no IP ranges or countries. */
int
routerset_is_list(const routerset_t *set)
{
  return smartlist_len(set->country_names) == 0 &&
    smartlist_len(set->policies) == 0;
}

/** Return true iff we need a GeoIP IP-to-country database to make sense of
 * <b>set</b>. */
int
routerset_needs_geoip(const routerset_t *set)
{
  return set && smartlist_len(set->country_names);
}

/** Return true iff there are no entries in <b>set</b>. */
int
routerset_is_empty(const routerset_t *set)
{
  return !set || smartlist_len(set->list) == 0;
}

/** Return the number of entries in <b>set</b>. This does NOT return a
 * negative value. */
int
routerset_len(const routerset_t *set)
{
  if (!set) {
    return 0;
  }
  return smartlist_len(set->list);
}

/** Helper.  Return true iff <b>set</b> contains a router based on the other
 * provided fields.  Return higher values for more specific subentries: a
 * single router is more specific than an address range of routers, which is
 * more specific in turn than a country code.
 *
 * (If country is -1, then we take the country
 * from addr.) */
STATIC int
routerset_contains(const routerset_t *set, const tor_addr_t *addr,
                   uint16_t orport,
                   const char *nickname, const char *id_digest,
                   country_t country)
{
  if (!set || !set->list)
    return 0;
  if (nickname && strmap_get_lc(set->names, nickname))
    return 4;
  if (id_digest && digestmap_get(set->digests, id_digest))
    return 4;
  if (addr && compare_tor_addr_to_addr_policy(addr, orport, set->policies)
      == ADDR_POLICY_REJECTED)
    return 3;
  if (set->countries) {
    if (country < 0 &&