Merge branch 'maint-0.2.1' into release-0.2.1

2024-12-06 19:41:15 +01:00 · 2011-02-11 00:59:34 -05:00
parent a2fe17bae2 50c259d763
commit 49cda713f8
19 changed files with 3552 additions and 1887 deletions
@@ -0,0 +1,6 @@
+  o Major bugfixes
+    - Ignore and warn about "PublishServerDescriptor hidserv" torrc
+      options.  The 'hidserv' argument never controlled publication
+      of hidden service descriptors.  Bugfix on 0.2.0.1-alpha.
+
+
@@ -0,0 +1,5 @@
+  o Major bugfixes:
+    - If relays set RelayBandwidthBurst but not RelayBandwidthRate,
+      Tor would ignore their RelayBandwidthBurst setting,
+      potentially using more bandwidth than expected. Bugfix on
+      0.2.0.1-alpha. Reported by Paul Wouters. Fixes bug 2470.
@@ -0,0 +1,3 @@
+  o Minor features
+    - Adjust our TLS Diffie-Hellman parameters to match those used by
+      Apache's mod_ssl.
@@ -0,0 +1,3 @@
+  o Minor features:
+    - Update to the February 1 2011 Maxmind GeoLite Country database.
+
@@ -0,0 +1,6 @@
+  o Major bugfixes (security)
+    - Fix a bounds-checking error that could allow an attacker to
+      remotely crash a directory authority.  Found by piebeer.
+      Bugfix on 0.2.1.5-alpha.
+
+
@@ -0,0 +1,4 @@
+  o Minor bugfixes
+    - Check for and reject overly long directory certificates and
+      directory tokens before they have a chance to hit any
+      assertions. Bugfix on 0.2.1.28. Found by doorss.
@@ -1,5 +1,9 @@
-<html>
+<?xml version="1.0"?>
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
+    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
+<html xmlns="http://www.w3.org/1999/xhtml">
 <head>
+<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
 <title>This is a Tor Exit Router</title>

 <!--
@@ -19,36 +23,38 @@ They are marked with FIXME.
 -->

 </head>
-<body bgcolor=white text=black>
+<body>

-<center><h1>This is a Tor Exit Router</h1></center>
-
-<p>Most likely you are accessing this website because you had some issue with
-the traffic coming from this IP. This router is part of the <a
-href="https://www.torproject.org/">Tor Anonymity Network</a>, which is
-dedicated to <a href="https://www.torproject.org/30seconds.html.en">providing
-privacy</a> to people who need it most: average computer users. This
-router IP should be generating no other traffic, unless it has been
-compromised.
+<p style="text-align:center; font-size:xx-large; font-weight:bold">This is a
+Tor Exit Router</p>

 <p>
+Most likely you are accessing this website because you had some issue with
+the traffic coming from this IP. This router is part of the <a
+href="https://www.torproject.org/">Tor Anonymity Network</a>, which is
+dedicated to <a href="https://www.torproject.org/about/overview">providing
+privacy</a> to people who need it most: average computer users. This
+router IP should be generating no other traffic, unless it has been
+compromised.</p>


 <!-- FIXME: you should probably grab your own copy of how_tor_works_thumb.png
-and serve it locally -->
-<center><a href="https://www.torproject.org/overview.html">
-<img src="https://www.torproject.org/images/how_tor_works_thumb.png"></a></center>
+     and serve it locally -->
+
+<p style="text-align:center">
+<a href="https://www.torproject.org/about/overview">
+<img src="https://www.torproject.org/images/how_tor_works_thumb.png" alt="How Tor works" style="border-style:none"/>
+</a></p>

 <p>
-
-Tor sees use by <a href="https://www.torproject.org/torusers.html.en">many
+Tor sees use by <a href="https://www.torproject.org/about/torusers">many
 important segments of the population</a>, including whistle blowers,
 journalists, Chinese dissidents skirting the Great Firewall and oppressive
 censorship, abuse victims, stalker targets, the US military, and law
 enforcement, just to name a few.  While Tor is not designed for malicious
 computer users, it is true that they can use the network for malicious ends.
 In reality however, the actual amount of <a
-href="https://www.torproject.org/faq-abuse.html">abuse</a> is quite low. This
+href="https://www.torproject.org/docs/faq-abuse">abuse</a> is quite low. This
 is largely because criminals and hackers have significantly better access to
 privacy and anonymity than do the regular users whom they prey upon. Criminals
 can and do <a
@@ -59,44 +65,41 @@ powerful networks</a> than Tor on a daily basis. Thus, in the mind of this
 operator, the social need for easily accessible censorship-resistant private,
 anonymous communication trumps the risk of unskilled bad actors, who are
 almost always more easily uncovered by traditional police work than by
-extensive monitoring and surveillance anyway.
+extensive monitoring and surveillance anyway.</p>

 <p>
-
 In terms of applicable law, the best way to understand Tor is to consider it a
 network of routers operating as common carriers, much like the Internet
 backbone. However, unlike the Internet backbone routers, Tor routers
 explicitly do not contain identifiable routing information about the source of
 a packet, and no single Tor node can determine both the origin and destination
-of a given transmission.
+of a given transmission.</p>

 <p>
-
 As such, there is little the operator of this router can do to help you track
 the connection further. This router maintains no logs of any of the Tor
 traffic, so there is little that can be done to trace either legitimate or
 illegitimate traffic (or to filter one from the other).  Attempts to
-seize this router will accomplish nothing.
+seize this router will accomplish nothing.</p>
+
+<!-- FIXME: US-Only section. Remove if you are a non-US operator -->
+
 <p>
-
-<!--- FIXME: US-Only section. Remove if you are a non-US operator -->
-
 Furthermore, this machine also serves as a carrier of email, which means that
 its contents are further protected under the ECPA. <a
 href="http://www4.law.cornell.edu/uscode/html/uscode18/usc_sec_18_00002707----000-.html">18
 USC 2707</a> explicitly allows for civil remedies ($1000/account
-<i><b><u>plus</u></b></i>  legal fees)
+<i><b>plus</b></i>  legal fees)
 in the event of a seizure executed without good faith or probable cause (it
 should be clear at this point that traffic with an originating IP address of
 FIXME_DNS_NAME should not constitute probable cause to seize the
 machine). Similar considerations exist for 1st amendment content on this
-machine.
-
-<p>
+machine.</p>

 <!-- FIXME: May or may not be US-only. Some non-US tor nodes have in
-fact reported DMCA harassment... -->
+     fact reported DMCA harassment... -->

+<p>
 If you are a representative of a company who feels that this router is being
 used to violate the DMCA, please be aware that this machine does not host or
 contain any illegal content. Also be aware that network infrastructure
@@ -105,36 +108,37 @@ equipment, in accordance with <a
 href="http://www4.law.cornell.edu/uscode/html/uscode17/usc_sec_17_00000512----000-.html">DMCA
 "safe harbor" provisions</a>. In other words, you will have just as much luck
 sending a takedown notice to the Internet backbone providers. Please consult
-<a href="https://www.torproject.org/eff/tor-dmca-response.html">EFF's prepared
-response</a> for more information on this matter.
+<a href="https://www.torproject.org/eff/tor-dmca-response">EFF's prepared
+response</a> for more information on this matter.</p>

-<p>For more information, please consult the following documentation:
+<p>For more information, please consult the following documentation:</p>

 <ol>
-<li><a href="https://www.torproject.org/overview.html">Tor Overview</a></li>
-<li><a href="https://www.torproject.org/faq-abuse.html">Tor Abuse FAQ</a></li>
-<li><a href="https://www.torproject.org/eff/tor-legal-faq.html">Tor Legal FAQ</a></li>
+<li><a href="https://www.torproject.org/about/overview">Tor Overview</a></li>
+<li><a href="https://www.torproject.org/docs/faq-abuse">Tor Abuse FAQ</a></li>
+<li><a href="https://www.torproject.org/eff/tor-legal-faq">Tor Legal FAQ</a></li>
 </ol>
-<p>

+<p>
 That being said, if you still have a complaint about the router,  you may
 email the <a href="mailto:FIXME_YOUR_EMAIL_ADDRESS">maintainer</a>. If
 complaints are related to a particular service that is being abused, I will
 consider removing that service from my exit policy, which would prevent my
 router from allowing that traffic to exit through it. I can only do this on an
 IP+destination port basis, however. Common P2P ports are
-already blocked.
+already blocked.</p>

-<p>You also have the option of blocking this IP address and others on
+<p>
+You also have the option of blocking this IP address and others on
 the Tor network if you so desire. The Tor project provides a <a
-href="https://tor-svn.freehaven.net/svn/tor/trunk/contrib/exitlist">python script</a> to
-extract all IP addresses of Tor exit nodes, and an official <a
-href="https://www.torproject.org/tordnsel/">DNSRBL</a> is also available to
+href="https://check.torproject.org/cgi-bin/TorBulkExitList.py">web service</a>
+to fetch a list of all IP addresses of Tor exit nodes that allow exiting to a
+specified IP:port combination, and an official <a
+href="https://www.torproject.org/tordnsel/dist/">DNSRBL</a> is also available to
 determine if a given IP address is actually a Tor exit server. Please
 be considerate
 when using these options. It would be unfortunate to deny all Tor users access
-to your site indefinitely simply because of a few bad apples.
+to your site indefinitely simply because of a few bad apples.</p>

 </body>
 </html>
-
@@ -907,7 +907,7 @@ This directive can be specified multiple times to bind to multiple
 addresses/ports.
 .LP
 .TP
-\fBPublishServerDescriptor \fR\fB0\fR|\fB1\fR|\fBv1\fR|\fBv2\fR|\fBv3\fR|\fBbridge\fR|\fBhidserv\fR, ...\fP
+\fBPublishServerDescriptor \fR\fB0\fR|\fB1\fR|\fBv1\fR|\fBv2\fR|\fBv3\fR|\fBbridge\fR, ...\fP
 This option is only considered if you have an ORPort defined. You can
 choose multiple arguments, separated by commas.

@@ -1505,8 +1505,10 @@ crypto_hmac_sha1(char *hmac_out,

 /* DH */

-/** Shared P parameter for our DH key exchanged. */
+/** Shared P parameter for our circuit-crypto DH key exchanges. */
 static BIGNUM *dh_param_p = NULL;
+/** Shared P parameter for our TLS DH key exchanges. */
+static BIGNUM *dh_param_p_tls = NULL;
 /** Shared G parameter for our DH key exchanges. */
 static BIGNUM *dh_param_g = NULL;

@@ -1515,14 +1517,16 @@ static BIGNUM *dh_param_g = NULL;
 static void
 init_dh_param(void)
 {
-  BIGNUM *p, *g;
+  BIGNUM *p, *p2, *g;
  int r;
-  if (dh_param_p && dh_param_g)
+  if (dh_param_p && dh_param_g && dh_param_p_tls)
    return;

  p = BN_new();
+  p2 = BN_new();
  g = BN_new();
  tor_assert(p);
+  tor_assert(p2);
  tor_assert(g);

  /* This is from rfc2409, section 6.2.  It's a safe prime, and
@@ -1536,10 +1540,20 @@ init_dh_param(void)
                "A637ED6B0BFF5CB6F406B7EDEE386BFB5A899FA5AE9F24117C4B1FE6"
                "49286651ECE65381FFFFFFFFFFFFFFFF");
  tor_assert(r);
+  /* This is the 1024-bit safe prime that Apache uses for its DH stuff; see
+   * modules/ssl/ssl_engine_dh.c */
+  r = BN_hex2bn(&p2,
+                  "D67DE440CBBBDC1936D693D34AFD0AD50C84D239A45F520BB88174CB98"
+                "BCE951849F912E639C72FB13B4B4D7177E16D55AC179BA420B2A29FE324A"
+                "467A635E81FF5901377BEDDCFD33168A461AAD3B72DAE8860078045B07A7"
+                "DBCA7874087D1510EA9FCC9DDD330507DD62DB88AEAA747DE0F4D6E2BD68"
+                "B0E7393E0F24218EB3");
+  tor_assert(r);

  r = BN_set_word(g, 2);
  tor_assert(r);
  dh_param_p = p;
+  dh_param_p_tls = p2;
  dh_param_g = g;
 }

@@ -1548,18 +1562,26 @@ init_dh_param(void)
 /** Allocate and return a new DH object for a key exchange.
 */
 crypto_dh_env_t *
-crypto_dh_new(void)
+crypto_dh_new(int dh_type)
 {
  crypto_dh_env_t *res = tor_malloc_zero(sizeof(crypto_dh_env_t));

+  tor_assert(dh_type == DH_TYPE_CIRCUIT || dh_type == DH_TYPE_TLS ||
+             dh_type == DH_TYPE_REND);
+
  if (!dh_param_p)
    init_dh_param();

  if (!(res->dh = DH_new()))
    goto err;

-  if (!(res->dh->p = BN_dup(dh_param_p)))
-    goto err;
+  if (dh_type == DH_TYPE_TLS) {
+    if (!(res->dh->p = BN_dup(dh_param_p_tls)))
+      goto err;
+  } else {
+    if (!(res->dh->p = BN_dup(dh_param_p)))
+      goto err;
+  }

  if (!(res->dh->g = BN_dup(dh_param_g)))
    goto err;
@@ -159,7 +159,10 @@ void crypto_hmac_sha1(char *hmac_out,
                      const char *msg, size_t msg_len);

 /* Key negotiation */
-crypto_dh_env_t *crypto_dh_new(void);
+#define DH_TYPE_CIRCUIT 1
+#define DH_TYPE_REND 2
+#define DH_TYPE_TLS 3
+crypto_dh_env_t *crypto_dh_new(int dh_type);
 int crypto_dh_get_bytes(crypto_dh_env_t *dh);
 int crypto_dh_generate_public(crypto_dh_env_t *dh);
 int crypto_dh_get_public(crypto_dh_env_t *dh, char *pubkey_out,
@@ -684,7 +684,7 @@ tor_tls_context_new(crypto_pk_env_t *identity, unsigned int key_lifetime)
  if (!SSL_CTX_check_private_key(result->ctx))
    goto error;
  {
-    crypto_dh_env_t *dh = crypto_dh_new();
+    crypto_dh_env_t *dh = crypto_dh_new(DH_TYPE_TLS);
    SSL_CTX_set_tmp_dh(result->ctx, _crypto_dh_env_get_dh(dh));
    crypto_dh_free(dh);
  }
@@ -2878,7 +2878,9 @@ compute_publishserverdescriptor(or_options_t *options)
    else if (!strcasecmp(string, "bridge"))
      *auth |= BRIDGE_AUTHORITY;
    else if (!strcasecmp(string, "hidserv"))
-      *auth |= HIDSERV_AUTHORITY;
+      log_warn(LD_CONFIG,
+               "PublishServerDescriptor hidserv is invalid. See "
+               "PublishHidServDescriptors.");
    else if (!strcasecmp(string, "") || !strcmp(string, "0"))
      /* no authority */;
    else
@@ -3368,6 +3370,11 @@ options_validate(or_options_t *old_options, or_options_t *options,
                           "RelayBandwidthBurst", msg) < 0)
    return -1;

+  if (options->RelayBandwidthRate && !options->RelayBandwidthBurst)
+    options->RelayBandwidthBurst = options->RelayBandwidthRate;
+  if (options->RelayBandwidthBurst && !options->RelayBandwidthRate)
+    options->RelayBandwidthRate = options->RelayBandwidthBurst;
+
  if (server_mode(options)) {
    if (options->BandwidthRate < ROUTER_REQUIRED_MIN_BANDWIDTH) {
      r = tor_snprintf(buf, sizeof(buf),
@@ -3399,9 +3406,6 @@ options_validate(or_options_t *old_options, or_options_t *options,
    }
  }

-  if (options->RelayBandwidthRate && !options->RelayBandwidthBurst)
-    options->RelayBandwidthBurst = options->RelayBandwidthRate;
-
  if (options->RelayBandwidthRate > options->RelayBandwidthBurst)
    REJECT("RelayBandwidthBurst must be at least equal "
           "to RelayBandwidthRate.");
@@ -173,7 +173,7 @@ onion_skin_create(crypto_pk_env_t *dest_router_key,
  *handshake_state_out = NULL;
  memset(onion_skin_out, 0, ONIONSKIN_CHALLENGE_LEN);

-  if (!(dh = crypto_dh_new()))
+  if (!(dh = crypto_dh_new(DH_TYPE_CIRCUIT)))
    goto err;

  dhbytes = crypto_dh_get_bytes(dh);
@@ -247,7 +247,7 @@ onion_skin_server_handshake(const char *onion_skin, /*ONIONSKIN_CHALLENGE_LEN*/
    goto err;
  }

-  dh = crypto_dh_new();
+  dh = crypto_dh_new(DH_TYPE_CIRCUIT);
  if (crypto_dh_get_public(dh, handshake_reply_out, DH_KEY_LEN)) {
    log_info(LD_GENERAL, "crypto_dh_get_public failed.");
    goto err;
@@ -1209,8 +1209,8 @@ policy_summarize(smartlist_t *policy)
  accepts_str = smartlist_join_strings(accepts, ",", 0, &accepts_len);
  rejects_str = smartlist_join_strings(rejects, ",", 0, &rejects_len);

-  if (rejects_len > MAX_EXITPOLICY_SUMMARY_LEN &&
-      accepts_len > MAX_EXITPOLICY_SUMMARY_LEN) {
+  if (rejects_len > MAX_EXITPOLICY_SUMMARY_LEN-strlen("reject")-1 &&
+      accepts_len > MAX_EXITPOLICY_SUMMARY_LEN-strlen("accept")-1) {
    char *c;
    shorter_str = accepts_str;
    prefix = "accept";
@@ -130,7 +130,7 @@ rend_client_send_introduction(origin_circuit_t *introcirc,
    cpath = rendcirc->build_state->pending_final_cpath =
      tor_malloc_zero(sizeof(crypt_path_t));
    cpath->magic = CRYPT_PATH_MAGIC;
-    if (!(cpath->dh_handshake_state = crypto_dh_new())) {
+    if (!(cpath->dh_handshake_state = crypto_dh_new(DH_TYPE_REND))) {
      log_warn(LD_BUG, "Internal error: couldn't allocate DH.");
      goto err;
    }
@@ -1151,7 +1151,7 @@ rend_service_introduce(origin_circuit_t *circuit, const uint8_t *request,
  }

  /* Try DH handshake... */
-  dh = crypto_dh_new();
+  dh = crypto_dh_new(DH_TYPE_REND);
  if (!dh || crypto_dh_generate_public(dh)<0) {
    log_warn(LD_BUG,"Internal error: couldn't build DH state "
             "or generate public key.");
@@ -1583,6 +1583,10 @@ extrainfo_parse_entry_from_string(const char *s, const char *end,
 authority_cert_t *
 authority_cert_parse_from_string(const char *s, const char **end_of_string)
 {
+  /** Reject any certificate at least this big; it is probably an overflow, an
+   * attack, a bug, or some other nonsense. */
+#define MAX_CERT_SIZE (128*1024)
+
  authority_cert_t *cert = NULL, *old_cert;
  smartlist_t *tokens = NULL;
  char digest[DIGEST_LEN];
@@ -1609,6 +1613,12 @@ authority_cert_parse_from_string(const char *s, const char **end_of_string)
  ++eos;
  len = eos - s;

+  if (len > MAX_CERT_SIZE) {
+    log_warn(LD_DIR, "Certificate is far too big (at %lu bytes long); "
+             "rejecting", (unsigned long)len);
+    return NULL;
+  }
+
  tokens = smartlist_create();
  area = memarea_new();
  if (tokenize_string(area,s, eos, tokens, dir_key_certificate_table, 0) < 0) {
@@ -3024,6 +3034,9 @@ get_next_token(memarea_t *area,
  /** Reject any object at least this big; it is probably an overflow, an
   * attack, a bug, or some other nonsense. */
 #define MAX_UNPARSED_OBJECT_SIZE (128*1024)
+  /** Reject any line at least this big; it is probably an overflow, an
+   * attack, a bug, or some other nonsense. */
+#define MAX_LINE_LENGTH (128*1024)

  const char *next, *eol, *obstart;
  size_t obname_len;
@@ -3043,6 +3056,10 @@ get_next_token(memarea_t *area,
  eol = memchr(*s, '\n', eos-*s);
  if (!eol)
    eol = eos;
+  if (eol - *s > MAX_LINE_LENGTH) {
+    RET_ERR("Line far too long");
+  }
+
  next = find_whitespace_eos(*s, eol);

  if (!strcmp_len(*s, "opt", next-*s)) {
@@ -404,8 +404,8 @@ test_buffers(void)
 static void
 test_crypto_dh(void)
 {
-  crypto_dh_env_t *dh1 = crypto_dh_new();
-  crypto_dh_env_t *dh2 = crypto_dh_new();
+  crypto_dh_env_t *dh1 = crypto_dh_new(DH_TYPE_CIRCUIT);
+  crypto_dh_env_t *dh2 = crypto_dh_new(DH_TYPE_CIRCUIT);
  char p1[DH_BYTES];
  char p2[DH_BYTES];
  char s1[DH_BYTES];