Turn addr_policy_compare from a tristate to a quadstate; this should help address our "Ah, you allow 1.2.3.4:80. You are a good choice for google.com" problem.

svn:r3786

7 files changed, 66 insertions, 42 deletions

diff --git a/src/or/circuitbuild.c b/src/or/circuitbuild.c
index 83fc79586..7e01fc3d3 100644
--- a/src/or/circuitbuild.c
+++ b/src/or/circuitbuild.c
@@ -858,10 +858,11 @@ router_handles_some_port(routerinfo_t *router, smartlist_t *needed_ports) {
   uint16_t port;
 
   for (i = 0; i < smartlist_len(needed_ports); ++i) {
+    addr_policy_result_t r;
     port = *(uint16_t *)smartlist_get(needed_ports, i);
     tor_assert(port);
-    if (router_compare_addr_to_addr_policy(0, port, router->exit_policy) !=
-          ADDR_POLICY_REJECTED)
+    r = router_compare_addr_to_addr_policy(0, port, router->exit_policy);
+    if (r != ADDR_POLICY_REJECTED && r != ADDR_POLICY_PROBABLY_REJECTED)
       return 1;
   }
   return 0;
diff --git a/src/or/circuituse.c b/src/or/circuituse.c
index b7e2f914d..836bb66b7 100644
--- a/src/or/circuituse.c
+++ b/src/or/circuituse.c
@@ -291,13 +291,19 @@ int circuit_stream_is_being_handled(connection_t *conn, uint16_t port, int min)
          circ->timestamp_dirty + get_options()->MaxCircuitDirtiness < now)) {
       exitrouter = router_get_by_digest(circ->build_state->chosen_exit_digest);
       if (exitrouter &&
-          (!need_uptime || circ->build_state->need_uptime) &&
-          ((conn && connection_ap_can_use_exit(conn, exitrouter)) ||
-           (!conn &&
-            router_compare_addr_to_addr_policy(0, port, exitrouter->exit_policy) !=
-              ADDR_POLICY_REJECTED))) {
-        if (++num >= min)
-          return 1;
+          (!need_uptime || circ->build_state->need_uptime)) {
+        int ok;
+        if (conn) {
+          ok = connection_ap_can_use_exit(conn, exitrouter);
+        } else {
+          addr_policy_result_t r =
+           router_compare_addr_to_addr_policy(0, port, exitrouter->exit_policy);
+          ok = r != ADDR_POLICY_REJECTED && r != ADDR_POLICY_PROBABLY_REJECTED;
+        }
+        if (ok) {
+          if (++num >= min)
+            return 1;
+        }
       }
     }
   }
diff --git a/src/or/connection_edge.c b/src/or/connection_edge.c
index 8bce01ed3..5550e5551 100644
--- a/src/or/connection_edge.c
+++ b/src/or/connection_edge.c
@@ -1471,8 +1471,8 @@ int connection_edge_is_rendezvous_stream(connection_t *conn) {
   return 0;
 }
 
-/** Return 1 if router <b>exit</b> might allow stream <b>conn</b>
- * to exit from it, or 0 if it definitely will not allow it.
+/** Return 1 if router <b>exit</b> is likely to allow stream <b>conn</b>
+ * to exit from it, or 0 if it probably will not allow it.
  * (We might be uncertain if conn's destination address has not yet been
  * resolved.)
  */
@@ -1502,10 +1502,12 @@ int connection_ap_can_use_exit(connection_t *conn, routerinfo_t *exit)
   if (conn->socks_request->command != SOCKS_COMMAND_RESOLVE) {
     struct in_addr in;
     uint32_t addr = 0;
+    addr_policy_result_t r;
     if (tor_inet_aton(conn->socks_request->address, &in))
       addr = ntohl(in.s_addr);
-    if (router_compare_addr_to_addr_policy(addr, conn->socks_request->port,
-          exit->exit_policy) == ADDR_POLICY_REJECTED)
+    r = router_compare_addr_to_addr_policy(addr, conn->socks_request->port,
+                                           exit->exit_policy);
+    if (r == ADDR_POLICY_REJECTED || r == ADDR_POLICY_PROBABLY_REJECTED)
       return 0;
   }
   return 1;
@@ -1550,11 +1552,10 @@ int socks_policy_permits_address(uint32_t addr)
   if (!socks_policy) /* 'no socks policy' means 'accept' */
     return 1;
   a = router_compare_addr_to_addr_policy(addr, 1, socks_policy);
-  if (a==-1)
+  if (a==ADDR_POLICY_REJECTED)
     return 0;
-  else if (a==0)
+  else if (a==ADDR_POLICY_ACCEPTED)
     return 1;
-  tor_assert(a==1);
   log_fn(LOG_WARN, "Bug: Got unexpected 'maybe' answer from socks policy");
   return 0;
 }
diff --git a/src/or/directory.c b/src/or/directory.c
index fd9fc6371..2b5c84b15 100644
--- a/src/or/directory.c
+++ b/src/or/directory.c
@@ -100,11 +100,10 @@ int dir_policy_permits_address(uint32_t addr)
   if (!dir_policy) /* 'no dir policy' means 'accept' */
     return 1;
   a = router_compare_addr_to_addr_policy(addr, 1, dir_policy);
-  if (a==-1)
+  if (a==ADDR_POLICY_REJECTED)
     return 0;
-  else if (a==0)
+  else if (a==ADDR_POLICY_ACCEPTED)
     return 1;
-  tor_assert(a==1);
   log_fn(LOG_WARN, "Bug: got unexpected 'maybe' answer from dir policy");
   return 0;
 }
diff --git a/src/or/or.h b/src/or/or.h
index c80240196..0a8d6610f 100644
--- a/src/or/or.h
+++ b/src/or/or.h
@@ -1661,6 +1661,12 @@ int rend_mid_establish_rendezvous(circuit_t *circ, const char *request, size_t r
 int rend_mid_rendezvous(circuit_t *circ, const char *request, size_t request_len);
 
 /********************************* router.c ***************************/
+typedef enum {
+  ADDR_POLICY_ACCEPTED=0,
+  ADDR_POLICY_REJECTED=-1,
+  ADDR_POLICY_PROBABLY_ACCEPTED=1,
+  ADDR_POLICY_PROBABLY_REJECTED=2
+} addr_policy_result_t;
 
 void set_onion_key(crypto_pk_env_t *k);
 crypto_pk_env_t *get_onion_key(void);
@@ -1760,11 +1766,9 @@ void routerlist_remove_old_routers(int age);
 int router_load_single_router(const char *s);
 int router_load_routerlist_from_directory(const char *s,crypto_pk_env_t *pkey,
                                         int dir_is_recent, int dir_is_cached);
-int router_compare_addr_to_addr_policy(uint32_t addr, uint16_t port,
-                                       addr_policy_t *policy);
-#define ADDR_POLICY_ACCEPTED 0
-#define ADDR_POLICY_REJECTED -1
-#define ADDR_POLICY_UNKNOWN 1
+addr_policy_result_t router_compare_addr_to_addr_policy(uint32_t addr,
+                              uint16_t port, addr_policy_t *policy);
+
 int router_exit_policy_all_routers_reject(uint32_t addr, uint16_t port,
                                           int need_uptime);
 
diff --git a/src/or/router.c b/src/or/router.c
index e0b500f86..e2263eef8 100644
--- a/src/or/router.c
+++ b/src/or/router.c
@@ -581,8 +581,8 @@ static void router_add_exit_policy_from_config(routerinfo_t *router) {
   config_parse_addr_policy(&default_policy, &router->exit_policy);
 }
 
-/** OR only: Return false if my exit policy says to allow connection to
- * conn.  Else return true.
+/** OR only: Check whether my exit policy says to allow connection to
+ * conn.  Return false if we accept; true if we reject.
  */
 int router_compare_to_my_exit_policy(connection_t *conn)
 {
@@ -594,10 +594,11 @@ int router_compare_to_my_exit_policy(connection_t *conn)
     return -1;
 
   return router_compare_addr_to_addr_policy(conn->addr, conn->port,
-                   desc_routerinfo->exit_policy);
+                   desc_routerinfo->exit_policy) != ADDR_POLICY_ACCEPTED;
 
 }
 
+
 /** Return true iff <b>router</b> has the same nickname as this OR.  (For an
  * OP, always returns false.)
  */
diff --git a/src/or/routerlist.c b/src/or/routerlist.c
index 5f7df4fb5..0ae62ad30 100644
--- a/src/or/routerlist.c
+++ b/src/or/routerlist.c
@@ -1001,16 +1001,26 @@ router_resolve_routerlist(routerlist_t *rl)
   return 0;
 }
 
-/** Decide whether a given addr:port is definitely accepted, definitely
- * rejected, or neither by a given policy.  If <b>addr</b> is 0, we
- * don't know the IP of the target address. If <b>port</b> is 0, we
- * don't know the port of the target address.
+/** Decide whether a given addr:port is definitely accepted,
+ * definitely rejected, probably accepted, or probably rejected by a
+ * given policy.  If <b>addr</b> is 0, we don't know the IP of the
+ * target address. If <b>port</b> is 0, we don't know the port of the
+ * target address.
  *
- * Returns -1 for "rejected", 0 for "accepted", 1 for "maybe" (since IP or
- * port is unknown).
+ * For now, the algorithm is pretty simple: we look for definite and
+ * uncertain matches.  The first definite match is what we guess; if
+ * it was proceded by no uncertain matches of the opposite policy,
+ * then the guess is definite; otherwise it is probable.  (If we
+ * have a known addr and port, all matches are definite; if we have an
+ * unknown addr/port, any address/port ranges other than "all" are
+ * uncertain.)
+ *
+ * We could do better by assuming that some ranges never match typical
+ * addresses (127.0.0.1, and so on).  But we'll try this for now.
  */
-int router_compare_addr_to_addr_policy(uint32_t addr, uint16_t port,
-                                       addr_policy_t *policy)
+addr_policy_result_t
+router_compare_addr_to_addr_policy(uint32_t addr, uint16_t port,
+                                   addr_policy_t *policy)
 {
   int maybe_reject = 0;
   int maybe_accept = 0;
@@ -1060,14 +1070,14 @@ int router_compare_addr_to_addr_policy(uint32_t addr, uint16_t port,
       if (tmpe->policy_type == ADDR_POLICY_ACCEPT) {
         /* If we already hit a clause that might trigger a 'reject', than we
          * can't be sure of this certain 'accept'.*/
-        return maybe_reject ? ADDR_POLICY_UNKNOWN : ADDR_POLICY_ACCEPTED;
+        return maybe_reject ? ADDR_POLICY_PROBABLY_ACCEPTED : ADDR_POLICY_ACCEPTED;
       } else {
-        return maybe_accept ? ADDR_POLICY_UNKNOWN : ADDR_POLICY_REJECTED;
+        return maybe_accept ? ADDR_POLICY_PROBABLY_REJECTED : ADDR_POLICY_REJECTED;
       }
     }
   }
   /* accept all by default. */
-  return maybe_reject ? ADDR_POLICY_UNKNOWN : ADDR_POLICY_ACCEPTED;
+  return maybe_reject ? ADDR_POLICY_PROBABLY_ACCEPTED : ADDR_POLICY_ACCEPTED;
 }
 
 /** Return 1 if all running sufficiently-stable routers will reject
@@ -1076,15 +1086,17 @@ int router_exit_policy_all_routers_reject(uint32_t addr, uint16_t port,
                                           int need_uptime) {
   int i;
   routerinfo_t *router;
+  addr_policy_result_t r;
   if (!routerlist) return 1;
 
   for (i=0;i<smartlist_len(routerlist->routers);i++) {
     router = smartlist_get(routerlist->routers, i);
     if (router->is_running &&
-        !router_is_unreliable(router, need_uptime, 0) &&
-        router_compare_addr_to_addr_policy(
-             addr, port, router->exit_policy) != ADDR_POLICY_REJECTED)
-      return 0; /* this one could be ok. good enough. */
+        !router_is_unreliable(router, need_uptime, 0)) {
+      r = router_compare_addr_to_addr_policy(addr, port, router->exit_policy);
+      if (r != ADDR_POLICY_REJECTED && r != ADDR_POLICY_PROBABLY_REJECTED)
+        return 0; /* this one could be ok. good enough. */
+    }
   }
   return 1; /* all will reject. */
 }