Merge remote-tracking branch 'public/bug23817_031' into maint-0.3.1

8 files changed, 280 insertions, 40 deletions

diff --git a/changes/bug23817 b/changes/bug23817
new file mode 100644
index 0000000000..4740942799
--- /dev/null
+++ b/changes/bug23817
@@ -0,0 +1,3 @@
+  o Minor bugfixes (descriptors):
+    - Don't try fetching microdescriptors from relays that have failed to
+      deliver them in the past. Fixes bug 23817; bugfix on 0.3.0.1-alpha.
diff --git a/src/or/directory.c b/src/or/directory.c
index f54bf176ce..5ceea2fb32 100644
--- a/src/or/directory.c
+++ b/src/or/directory.c
@@ -117,7 +117,8 @@ static void dir_routerdesc_download_failed(smartlist_t *failed,
                                            int was_extrainfo,
                                            int was_descriptor_digests);
 static void dir_microdesc_download_failed(smartlist_t *failed,
-                                          int status_code);
+                                          int status_code,
+                                          const char *dir_id);
 static int client_likes_consensus(const struct consensus_cache_entry_t *ent,
                                   const char *want_url);
 
@@ -463,7 +464,7 @@ directory_pick_generic_dirserver(dirinfo_type_t type, int pds_flags,
     log_warn(LD_BUG, "Called when we have UseBridges set.");
 
   if (should_use_directory_guards(options)) {
-    const node_t *node = guards_choose_dirguard(guard_state_out);
+    const node_t *node = guards_choose_dirguard(dir_purpose, guard_state_out);
     if (node)
       rs = node->rs;
   } else {
@@ -597,7 +598,7 @@ directory_get_from_dirserver,(
        * sort of dir fetch we'll be doing, so it won't return a bridge
        * that can't answer our question.
        */
-      const node_t *node = guards_choose_dirguard(&guard_state);
+      const node_t *node = guards_choose_dirguard(dir_purpose, &guard_state);
       if (node && node->ri) {
         /* every bridge has a routerinfo. */
         routerinfo_t *ri = node->ri;
@@ -2178,8 +2179,6 @@ static int handle_response_fetch_detached_signatures(dir_connection_t *,
                                              const response_handler_args_t *);
 static int handle_response_fetch_desc(dir_connection_t *,
                                              const response_handler_args_t *);
-static int handle_response_fetch_microdesc(dir_connection_t *,
-                                           const response_handler_args_t *);
 static int handle_response_upload_dir(dir_connection_t *,
                                       const response_handler_args_t *);
 static int handle_response_upload_vote(dir_connection_t *,
@@ -2839,7 +2838,7 @@ handle_response_fetch_desc(dir_connection_t *conn,
  * Handler function: processes a response to a request for a group of
  * microdescriptors
  **/
-static int
+STATIC int
 handle_response_fetch_microdesc(dir_connection_t *conn,
                                 const response_handler_args_t *args)
 {
@@ -2856,6 +2855,7 @@ handle_response_fetch_microdesc(dir_connection_t *conn,
            conn->base_.port);
   tor_assert(conn->requested_resource &&
              !strcmpstart(conn->requested_resource, "d/"));
+  tor_assert_nonfatal(!tor_mem_is_zero(conn->identity_digest, DIGEST_LEN));
   which = smartlist_new();
   dir_split_resource_into_fingerprints(conn->requested_resource+2,
                                        which, NULL,
@@ -2866,7 +2866,7 @@ handle_response_fetch_microdesc(dir_connection_t *conn,
              "soon.",
              status_code, escaped(reason), conn->base_.address,
              (int)conn->base_.port, conn->requested_resource);
-    dir_microdesc_download_failed(which, status_code);
+    dir_microdesc_download_failed(which, status_code, conn->identity_digest);
     SMARTLIST_FOREACH(which, char *, cp, tor_free(cp));
     smartlist_free(which);
     return 0;
@@ -2878,7 +2878,7 @@ handle_response_fetch_microdesc(dir_connection_t *conn,
                                   now, which);
     if (smartlist_len(which)) {
       /* Mark remaining ones as failed. */
-      dir_microdesc_download_failed(which, status_code);
+      dir_microdesc_download_failed(which, status_code, conn->identity_digest);
     }
     if (mds && smartlist_len(mds)) {
       control_event_bootstrap(BOOTSTRAP_STATUS_LOADING_DESCRIPTORS,
@@ -5546,13 +5546,14 @@ dir_routerdesc_download_failed(smartlist_t *failed, int status_code,
    * every 10 or 60 seconds (FOO_DESCRIPTOR_RETRY_INTERVAL) in main.c. */
 }
 
-/** Called when a connection to download microdescriptors has failed in whole
- * or in part. <b>failed</b> is a list of every microdesc digest we didn't
- * get. <b>status_code</b> is the http status code we received. Reschedule the
- * microdesc downloads as appropriate. */
+/** Called when a connection to download microdescriptors from relay with
+ * <b>dir_id</b> has failed in whole or in part. <b>failed</b> is a list
+ * of every microdesc digest we didn't get. <b>status_code</b> is the http
+ * status code we received. Reschedule the microdesc downloads as
+ * appropriate. */
 static void
 dir_microdesc_download_failed(smartlist_t *failed,
-                              int status_code)
+                              int status_code, const char *dir_id)
 {
   networkstatus_t *consensus
     = networkstatus_get_latest_consensus_by_flavor(FLAV_MICRODESC);
@@ -5563,17 +5564,26 @@ dir_microdesc_download_failed(smartlist_t *failed,
 
   if (! consensus)
     return;
+
+  /* We failed to fetch a microdescriptor from 'dir_id', note it down
+   * so that we don't try the same relay next time... */
+  microdesc_note_outdated_dirserver(dir_id);
+
   SMARTLIST_FOREACH_BEGIN(failed, const char *, d) {
     rs = router_get_mutable_consensus_status_by_descriptor_digest(consensus,d);
     if (!rs)
       continue;
     dls = &rs->dl_status;
     if (dls->n_download_failures >=
-        get_options()->TestingMicrodescMaxDownloadTries)
+        get_options()->TestingMicrodescMaxDownloadTries) {
       continue;
-    {
+    }
+
+    { /* Increment the failure count for this md fetch */
       char buf[BASE64_DIGEST256_LEN+1];
       digest256_to_base64(buf, d);
+      log_info(LD_DIR, "Failed to download md %s from %s",
+               buf, hex_str(dir_id, DIGEST_LEN));
       download_status_increment_failure(dls, status_code, buf,
                                         server, now);
     }
diff --git a/src/or/directory.h b/src/or/directory.h
index 14d5ae9ef4..571c30a0fc 100644
--- a/src/or/directory.h
+++ b/src/or/directory.h
@@ -166,7 +166,12 @@ STATIC char *accept_encoding_header(void);
 STATIC int allowed_anonymous_connection_compression_method(compress_method_t);
 STATIC void warn_disallowed_anonymous_compression_method(compress_method_t);
 
-#endif
+struct response_handler_args_t;
+
+STATIC int handle_response_fetch_microdesc(dir_connection_t *conn,
+                                 const struct response_handler_args_t *args);
+
+#endif /* defined(DIRECTORY_PRIVATE) */
 
 #ifdef TOR_UNIT_TESTS
 /* Used only by test_dir.c */
diff --git a/src/or/entrynodes.c b/src/or/entrynodes.c
index d6cd58cedc..2cbc8019d4 100644
--- a/src/or/entrynodes.c
+++ b/src/or/entrynodes.c
@@ -967,7 +967,7 @@ entry_guard_learned_bridge_identity(const tor_addr_port_t *addrport,
  * violate it.
  */
 STATIC int
-num_reachable_filtered_guards(guard_selection_t *gs,
+num_reachable_filtered_guards(const guard_selection_t *gs,
                               const entry_guard_restriction_t *rst)
 {
   int n_reachable_filtered_guards = 0;
@@ -1461,6 +1461,94 @@ guard_in_node_family(const entry_guard_t *guard, const node_t *node)
   }
 }
 
+/* Allocate and return a new exit guard restriction (where <b>exit_id</b> is of
+ * size DIGEST_LEN) */
+STATIC entry_guard_restriction_t *
+guard_create_exit_restriction(const uint8_t *exit_id)
+{
+  entry_guard_restriction_t *rst = NULL;
+  rst = tor_malloc_zero(sizeof(entry_guard_restriction_t));
+  rst->type = RST_EXIT_NODE;
+  memcpy(rst->exclude_id, exit_id, DIGEST_LEN);
+  return rst;
+}
+
+/** If we have fewer than this many possible usable guards, don't set
+ * MD-availability-based restrictions: we might blacklist all of them. */
+#define MIN_GUARDS_FOR_MD_RESTRICTION 10
+
+/** Return true if we should set md dirserver restrictions. We might not want
+ *  to set those if our guard options are too restricted, since we don't want
+ *  to blacklist all of them. */
+static int
+should_set_md_dirserver_restriction(void)
+{
+  const guard_selection_t *gs = get_guard_selection_info();
+  int num_usable_guards = num_reachable_filtered_guards(gs, NULL);
+
+  /* Don't set restriction if too few reachable filtered guards. */
+  if (num_usable_guards < MIN_GUARDS_FOR_MD_RESTRICTION) {
+    log_info(LD_GUARD, "Not setting md restriction: only %d"
+             " usable guards.", num_usable_guards);
+    return 0;
+  }
+
+  /* We have enough usable guards: set MD restriction */
+  return 1;
+}
+
+/** Allocate and return an outdated md guard restriction. Return NULL if no
+ *  such restriction is needed. */
+STATIC entry_guard_restriction_t *
+guard_create_dirserver_md_restriction(void)
+{
+  entry_guard_restriction_t *rst = NULL;
+
+  if (!should_set_md_dirserver_restriction()) {
+    log_debug(LD_GUARD, "Not setting md restriction: too few "
+              "filtered guards.");
+    return NULL;
+  }
+
+  rst = tor_malloc_zero(sizeof(entry_guard_restriction_t));
+  rst->type = RST_OUTDATED_MD_DIRSERVER;
+
+  return rst;
+}
+
+/* Return True if <b>guard</b> obeys the exit restriction <b>rst</b>. */
+static int
+guard_obeys_exit_restriction(const entry_guard_t *guard,
+                             const entry_guard_restriction_t *rst)
+{
+  tor_assert(rst->type == RST_EXIT_NODE);
+
+  // Exclude the exit ID and all of its family.
+  const node_t *node = node_get_by_id((const char*)rst->exclude_id);
+  if (node && guard_in_node_family(guard, node))
+    return 0;
+
+  return tor_memneq(guard->identity, rst->exclude_id, DIGEST_LEN);
+}
+
+/** Return True if <b>guard</b> should be used as a dirserver for fetching
+ *  microdescriptors. */
+static int
+guard_obeys_md_dirserver_restriction(const entry_guard_t *guard)
+{
+  /* If this guard is an outdated dirserver, don't use it. */
+  if (microdesc_relay_is_outdated_dirserver(guard->identity)) {
+    log_info(LD_GENERAL, "Skipping %s dirserver: outdated",
+             hex_str(guard->identity, DIGEST_LEN));
+    return 0;
+  }
+
+  log_debug(LD_GENERAL, "%s dirserver obeys md restrictions",
+            hex_str(guard->identity, DIGEST_LEN));
+
+  return 1;
+}
+
 /**
  * Return true iff <b>guard</b> obeys the restrictions defined in <b>rst</b>.
  * (If <b>rst</b> is NULL, there are no restrictions.)
@@ -1473,13 +1561,14 @@ entry_guard_obeys_restriction(const entry_guard_t *guard,
   if (! rst)
     return 1; // No restriction?  No problem.
 
-  // Only one kind of restriction exists right now: excluding an exit
-  // ID and all of its family.
-  const node_t *node = node_get_by_id((const char*)rst->exclude_id);
-  if (node && guard_in_node_family(guard, node))
-    return 0;
+  if (rst->type == RST_EXIT_NODE) {
+    return guard_obeys_exit_restriction(guard, rst);
+  } else if (rst->type == RST_OUTDATED_MD_DIRSERVER) {
+    return guard_obeys_md_dirserver_restriction(guard);
+  }
 
-  return tor_memneq(guard->identity, rst->exclude_id, DIGEST_LEN);
+  tor_assert_nonfatal_unreached();
+  return 0;
 }
 
 /**
@@ -2106,7 +2195,7 @@ entry_guard_has_higher_priority(entry_guard_t *a, entry_guard_t *b)
 }
 
 /** Release all storage held in <b>restriction</b> */
-static void
+STATIC void
 entry_guard_restriction_free(entry_guard_restriction_t *rst)
 {
   tor_free(rst);
@@ -3359,8 +3448,8 @@ guards_choose_guard(cpath_build_state_t *state,
     /* We're building to a targeted exit node, so that node can't be
      * chosen as our guard for this circuit.  Remember that fact in a
      * restriction. */
-    rst = tor_malloc_zero(sizeof(entry_guard_restriction_t));
-    memcpy(rst->exclude_id, exit_id, DIGEST_LEN);
+    rst = guard_create_exit_restriction(exit_id);
+    tor_assert(rst);
   }
   if (entry_guard_pick_for_circuit(get_guard_selection_info(),
                                    GUARD_USAGE_TRAFFIC,
@@ -3412,12 +3501,20 @@ remove_all_entry_guards(void)
 
 /** Helper: pick a directory guard, with whatever algorithm is used. */
 const node_t *
-guards_choose_dirguard(circuit_guard_state_t **guard_state_out)
+guards_choose_dirguard(uint8_t dir_purpose,
+                       circuit_guard_state_t **guard_state_out)
 {
   const node_t *r = NULL;
+  entry_guard_restriction_t *rst = NULL;
+
+  /* If we are fetching microdescs, don't query outdated dirservers. */
+  if (dir_purpose == DIR_PURPOSE_FETCH_MICRODESC) {
+    rst = guard_create_dirserver_md_restriction();
+  }
+
   if (entry_guard_pick_for_circuit(get_guard_selection_info(),
                                    GUARD_USAGE_DIRGUARD,
-                                   NULL,
+                                   rst,
                                    &r,
                                    guard_state_out) < 0) {
     tor_assert(r == NULL);
diff --git a/src/or/entrynodes.h b/src/or/entrynodes.h
index 735c7738ba..39bff14381 100644
--- a/src/or/entrynodes.h
+++ b/src/or/entrynodes.h
@@ -272,22 +272,28 @@ struct guard_selection_s {
 
 struct entry_guard_handle_t;
 
+/** Types of restrictions we impose when picking guard nodes */
+typedef enum guard_restriction_type_t {
+  /* Don't pick the same guard node as our exit node (or its family) */
+  RST_EXIT_NODE = 0,
+  /* Don't pick dirguards that have previously shown to be outdated */
+  RST_OUTDATED_MD_DIRSERVER = 1
+} guard_restriction_type_t;
+
 /**
  * A restriction to remember which entry guards are off-limits for a given
  * circuit.
  *
- * Right now, we only use restrictions to block a single guard and its family
- * from being selected; this mechanism is designed to be more extensible in
- * the future, however.
- *
  * Note: This mechanism is NOT for recording which guards are never to be
  * used: only which guards cannot be used on <em>one particular circuit</em>.
  */
 struct entry_guard_restriction_t {
-  /**
-   * The guard's RSA identity digest must not equal this; and it must not
-   * be in the same family as any node with this digest.
-   */
+  /* What type of restriction are we imposing? */
+  guard_restriction_type_t type;
+
+  /* In case of restriction type RST_EXIT_NODE, the guard's RSA identity
+   * digest must not equal this; and it must not be in the same family as any
+   * node with this digest. */
   uint8_t exclude_id[DIGEST_LEN];
 };
 
@@ -316,7 +322,8 @@ struct circuit_guard_state_t {
 int guards_update_all(void);
 const node_t *guards_choose_guard(cpath_build_state_t *state,
                                   circuit_guard_state_t **guard_state_out);
-const node_t *guards_choose_dirguard(circuit_guard_state_t **guard_state_out);
+const node_t *guards_choose_dirguard(uint8_t dir_purpose,
+                                     circuit_guard_state_t **guard_state_out);
 
 #if 1
 /* XXXX NM I would prefer that all of this stuff be private to
@@ -514,7 +521,7 @@ STATIC void entry_guard_consider_retry(entry_guard_t *guard);
 STATIC void make_guard_confirmed(guard_selection_t *gs, entry_guard_t *guard);
 STATIC void entry_guards_update_confirmed(guard_selection_t *gs);
 STATIC void entry_guards_update_primary(guard_selection_t *gs);
-STATIC int num_reachable_filtered_guards(guard_selection_t *gs,
+STATIC int num_reachable_filtered_guards(const guard_selection_t *gs,
                                          const entry_guard_restriction_t *rst);
 STATIC void sampled_guards_update_from_consensus(guard_selection_t *gs);
 /**
@@ -550,7 +557,17 @@ STATIC unsigned entry_guards_note_guard_success(guard_selection_t *gs,
                                                 unsigned old_state);
 STATIC int entry_guard_has_higher_priority(entry_guard_t *a, entry_guard_t *b);
 STATIC char *getinfo_helper_format_single_entry_guard(const entry_guard_t *e);
-#endif
+
+STATIC entry_guard_restriction_t *
+guard_create_exit_restriction(const uint8_t *exit_id);
+
+STATIC entry_guard_restriction_t *
+guard_create_dirserver_md_restriction(void);
+
+STATIC void
+entry_guard_restriction_free(entry_guard_restriction_t *rst);
+
+#endif /* defined(ENTRYNODES_PRIVATE) */
 
 void remove_all_entry_guards_for_guard_selection(guard_selection_t *gs);
 void remove_all_entry_guards(void);
diff --git a/src/or/microdesc.c b/src/or/microdesc.c
index a4e6b409c4..32242d0054 100644
--- a/src/or/microdesc.c
+++ b/src/or/microdesc.c
@@ -74,6 +74,102 @@ HT_GENERATE2(microdesc_map, microdesc_t, node,
              microdesc_hash_, microdesc_eq_, 0.6,
              tor_reallocarray_, tor_free_)
 
+/************************* md fetch fail cache *****************************/
+
+/* If we end up with too many outdated dirservers, something probably went
+ * wrong so clean up the list. */
+#define TOO_MANY_OUTDATED_DIRSERVERS 30
+
+/** List of dirservers with outdated microdesc information. The smartlist is
+ *  filled with the hex digests of outdated dirservers. */
+static smartlist_t *outdated_dirserver_list = NULL;
+
+/** Note that we failed to fetch a microdescriptor from the relay with
+ *  <b>relay_digest</b> (of size DIGEST_LEN). */
+void
+microdesc_note_outdated_dirserver(const char *relay_digest)
+{
+  char relay_hexdigest[HEX_DIGEST_LEN+1];
+
+  /* Don't register outdated dirservers if we don't have a live consensus,
+   * since we might be trying to fetch microdescriptors that are not even
+   * currently active. */
+  if (!networkstatus_get_live_consensus(approx_time())) {
+    return;
+  }
+
+  if (!outdated_dirserver_list) {
+    outdated_dirserver_list = smartlist_new();
+  }
+
+  tor_assert(outdated_dirserver_list);
+
+  /* If the list grows too big, clean it up */
+  if (BUG(smartlist_len(outdated_dirserver_list) >
+          TOO_MANY_OUTDATED_DIRSERVERS)) {
+    microdesc_reset_outdated_dirservers_list();
+  }
+
+  /* Turn the binary relay digest to a hex since smartlists have better support
+   * for strings than digests. */
+  base16_encode(relay_hexdigest,sizeof(relay_hexdigest),
+                relay_digest, DIGEST_LEN);
+
+  /* Make sure we don't add a dirauth as an outdated dirserver */
+  if (router_get_trusteddirserver_by_digest(relay_digest)) {
+    log_info(LD_GENERAL, "Auth %s gave us outdated dirinfo.", relay_hexdigest);
+    return;
+  }
+
+  /* Don't double-add outdated dirservers */
+  if (smartlist_contains_string(outdated_dirserver_list, relay_hexdigest)) {
+    return;
+  }
+
+  /* Add it to the list of outdated dirservers */
+  smartlist_add_strdup(outdated_dirserver_list, relay_hexdigest);
+
+  log_info(LD_GENERAL, "Noted %s as outdated md dirserver", relay_hexdigest);
+}
+
+/** Return True if the relay with <b>relay_digest</b> (size DIGEST_LEN) is an
+ *  outdated dirserver */
+int
+microdesc_relay_is_outdated_dirserver(const char *relay_digest)
+{
+  char relay_hexdigest[HEX_DIGEST_LEN+1];
+
+  if (!outdated_dirserver_list) {
+    return 0;
+  }
+
+  /* Convert identity digest to hex digest */
+  base16_encode(relay_hexdigest, sizeof(relay_hexdigest),
+                relay_digest, DIGEST_LEN);
+
+  /* Last time we tried to fetch microdescs, was this directory mirror missing
+   * any mds we asked for? */
+  if (smartlist_contains_string(outdated_dirserver_list, relay_hexdigest)) {
+    return 1;
+  }
+
+  return 0;
+}
+
+/** Reset the list of outdated dirservers. */
+void
+microdesc_reset_outdated_dirservers_list(void)
+{
+  if (!outdated_dirserver_list) {
+    return;
+  }
+
+  SMARTLIST_FOREACH(outdated_dirserver_list, char *, cp, tor_free(cp));
+  smartlist_clear(outdated_dirserver_list);
+}
+
+/****************************************************************************/
+
 /** Write the body of <b>md</b> into <b>f</b>, with appropriate annotations.
  * On success, return the total number of bytes written, and set
  * *<b>annotation_len_out</b> to the number of bytes written as
@@ -789,6 +885,11 @@ microdesc_free_all(void)
     tor_free(the_microdesc_cache->journal_fname);
     tor_free(the_microdesc_cache);
   }
+
+  if (outdated_dirserver_list) {
+    SMARTLIST_FOREACH(outdated_dirserver_list, char *, cp, tor_free(cp));
+    smartlist_free(outdated_dirserver_list);
+  }
 }
 
 /** If there is a microdescriptor in <b>cache</b> whose sha256 digest is
diff --git a/src/or/microdesc.h b/src/or/microdesc.h
index 943873066e..1be12156a4 100644
--- a/src/or/microdesc.h
+++ b/src/or/microdesc.h
@@ -50,5 +50,9 @@ int we_fetch_microdescriptors(const or_options_t *options);
 int we_fetch_router_descriptors(const or_options_t *options);
 int we_use_microdescriptors_for_circuits(const or_options_t *options);
 
-#endif
+void microdesc_note_outdated_dirserver(const char *relay_digest);
+int microdesc_relay_is_outdated_dirserver(const char *relay_digest);
+void microdesc_reset_outdated_dirservers_list(void);
+
+#endif /* !defined(TOR_MICRODESC_H) */
 
diff --git a/src/or/networkstatus.c b/src/or/networkstatus.c
index 997280de52..36e62020e3 100644
--- a/src/or/networkstatus.c
+++ b/src/or/networkstatus.c
@@ -2041,6 +2041,9 @@ networkstatus_set_current_consensus(const char *consensus,
                     "CLOCK_SKEW MIN_SKEW=%ld SOURCE=CONSENSUS", delta);
   }
 
+  /* We got a new consesus. Reset our md fetch fail cache */
+  microdesc_reset_outdated_dirservers_list();
+
   router_dir_info_changed();
 
   result = 0;