Filter DNS results by common ASN
[dnsseed-rust] / src / datastore.rs
index 827726e9aae9dae21a1d8c71840c7e8dc2325881..c8ef0e0aed812f52599ac6d1ca47dcde0692e48b 100644 (file)
@@ -1,7 +1,7 @@
 use std::{cmp, mem};
 use std::collections::{HashSet, HashMap, hash_map};
 use std::sync::{Arc, RwLock};
-use std::net::SocketAddr;
+use std::net::{IpAddr, SocketAddr};
 use std::time::{Duration, Instant};
 use std::io::{BufRead, BufReader};
 
@@ -16,6 +16,8 @@ use tokio::io::write_all;
 
 use regex::Regex;
 
+use crate::bgp_client::BGPClient;
+
 #[derive(Clone, Copy, Hash, PartialEq, Eq)]
 pub enum AddressState {
        Untested,
@@ -27,8 +29,12 @@ pub enum AddressState {
        ProtocolViolation,
        Timeout,
        TimeoutDuringRequest,
+       TimeoutAwaitingPong,
+       TimeoutAwaitingAddr,
+       TimeoutAwaitingBlock,
        Good,
        WasGood,
+       EvilNode,
 }
 
 impl AddressState {
@@ -43,8 +49,12 @@ impl AddressState {
                        0x6 => Some(AddressState::ProtocolViolation),
                        0x7 => Some(AddressState::Timeout),
                        0x8 => Some(AddressState::TimeoutDuringRequest),
-                       0x9 => Some(AddressState::Good),
-                       0xa => Some(AddressState::WasGood),
+                       0x9 => Some(AddressState::TimeoutAwaitingPong),
+                       0xa => Some(AddressState::TimeoutAwaitingAddr),
+                       0xb => Some(AddressState::TimeoutAwaitingBlock),
+                       0xc => Some(AddressState::Good),
+                       0xd => Some(AddressState::WasGood),
+                       0xe => Some(AddressState::EvilNode),
                        _   => None,
                }
        }
@@ -60,8 +70,12 @@ impl AddressState {
                        AddressState::ProtocolViolation => 6,
                        AddressState::Timeout => 7,
                        AddressState::TimeoutDuringRequest => 8,
-                       AddressState::Good => 9,
-                       AddressState::WasGood => 10,
+                       AddressState::TimeoutAwaitingPong => 9,
+                       AddressState::TimeoutAwaitingAddr => 10,
+                       AddressState::TimeoutAwaitingBlock => 11,
+                       AddressState::Good => 12,
+                       AddressState::WasGood => 13,
+                       AddressState::EvilNode => 14,
                }
        }
 
@@ -76,13 +90,17 @@ impl AddressState {
                        AddressState::ProtocolViolation => "Protocol Violation",
                        AddressState::Timeout => "Timeout",
                        AddressState::TimeoutDuringRequest => "Timeout During Request",
+                       AddressState::TimeoutAwaitingPong => "Timeout Awaiting Pong",
+                       AddressState::TimeoutAwaitingAddr => "Timeout Awaiting Addr",
+                       AddressState::TimeoutAwaitingBlock => "Timeout Awaiting Block",
                        AddressState::Good => "Good",
                        AddressState::WasGood => "Was Good",
+                       AddressState::EvilNode => "Evil Node",
                }
        }
 
-       pub fn get_count() -> u8 {
-               11
+       pub const fn get_count() -> u8 {
+               15
        }
 }
 
@@ -153,7 +171,7 @@ impl Store {
                                        }
                                } }
                        }
-                       let mut u64s = HashMap::with_capacity(15);
+                       let mut u64s = HashMap::with_capacity(AddressState::get_count() as usize + 4);
                        u64s.insert(U64Setting::ConnsPerSec, try_read!(l, u64));
                        u64s.insert(U64Setting::RunTimeout, try_read!(l, u64));
                        u64s.insert(U64Setting::WasGoodTimeout, try_read!(l, u64));
@@ -167,8 +185,12 @@ impl Store {
                        u64s.insert(U64Setting::RescanInterval(AddressState::ProtocolViolation), try_read!(l, u64));
                        u64s.insert(U64Setting::RescanInterval(AddressState::Timeout), try_read!(l, u64));
                        u64s.insert(U64Setting::RescanInterval(AddressState::TimeoutDuringRequest), try_read!(l, u64));
+                       u64s.insert(U64Setting::RescanInterval(AddressState::TimeoutAwaitingPong), try_read!(l, u64));
+                       u64s.insert(U64Setting::RescanInterval(AddressState::TimeoutAwaitingAddr), try_read!(l, u64));
+                       u64s.insert(U64Setting::RescanInterval(AddressState::TimeoutAwaitingBlock), try_read!(l, u64));
                        u64s.insert(U64Setting::RescanInterval(AddressState::Good), try_read!(l, u64));
                        u64s.insert(U64Setting::RescanInterval(AddressState::WasGood), try_read!(l, u64));
+                       u64s.insert(U64Setting::RescanInterval(AddressState::EvilNode), try_read!(l, u64));
                        future::ok((u64s, try_read!(l, Regex)))
                }).or_else(|_| -> future::FutureResult<(HashMap<U64Setting, u64>, Regex), ()> {
                        let mut u64s = HashMap::with_capacity(15);
@@ -184,9 +206,13 @@ impl Store {
                        u64s.insert(U64Setting::RescanInterval(AddressState::ProtocolViolation), 86400);
                        u64s.insert(U64Setting::RescanInterval(AddressState::Timeout), 86400);
                        u64s.insert(U64Setting::RescanInterval(AddressState::TimeoutDuringRequest), 21600);
+                       u64s.insert(U64Setting::RescanInterval(AddressState::TimeoutAwaitingPong), 3600);
+                       u64s.insert(U64Setting::RescanInterval(AddressState::TimeoutAwaitingAddr), 1800);
+                       u64s.insert(U64Setting::RescanInterval(AddressState::TimeoutAwaitingBlock), 3600);
                        u64s.insert(U64Setting::RescanInterval(AddressState::Good), 1800);
                        u64s.insert(U64Setting::RescanInterval(AddressState::WasGood), 1800);
-                       u64s.insert(U64Setting::MinProtocolVersion, 10000); //XXX
+                       u64s.insert(U64Setting::RescanInterval(AddressState::EvilNode), 315360000);
+                       u64s.insert(U64Setting::MinProtocolVersion, 70002);
                        future::ok((u64s, Regex::new(".*").unwrap()))
                });
 
@@ -360,7 +386,7 @@ impl Store {
        pub fn save_data(&'static self) -> impl Future<Item=(), Error=()> {
                let settings_file = self.store.clone() + "/settings";
                let settings_future = File::create(settings_file.clone() + ".tmp").and_then(move |f| {
-                       let settings_string = format!("{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}",
+                       let settings_string = format!("{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}\n{}",
                                self.get_u64(U64Setting::ConnsPerSec),
                                self.get_u64(U64Setting::RunTimeout),
                                self.get_u64(U64Setting::WasGoodTimeout),
@@ -374,8 +400,12 @@ impl Store {
                                self.get_u64(U64Setting::RescanInterval(AddressState::ProtocolViolation)),
                                self.get_u64(U64Setting::RescanInterval(AddressState::Timeout)),
                                self.get_u64(U64Setting::RescanInterval(AddressState::TimeoutDuringRequest)),
+                               self.get_u64(U64Setting::RescanInterval(AddressState::TimeoutAwaitingPong)),
+                               self.get_u64(U64Setting::RescanInterval(AddressState::TimeoutAwaitingAddr)),
+                               self.get_u64(U64Setting::RescanInterval(AddressState::TimeoutAwaitingBlock)),
                                self.get_u64(U64Setting::RescanInterval(AddressState::Good)),
                                self.get_u64(U64Setting::RescanInterval(AddressState::WasGood)),
+                               self.get_u64(U64Setting::RescanInterval(AddressState::EvilNode)),
                                self.get_regex(RegexSetting::SubverRegex).as_str());
                        write_all(f, settings_string).and_then(|(mut f, _)| {
                                f.poll_sync_all()
@@ -406,72 +436,74 @@ impl Store {
                        tokio::fs::rename(nodes_file.clone() + ".tmp", nodes_file)
                });
 
+               settings_future.join(nodes_future).then(|_| { future::ok(()) })
+       }
+
+       pub fn write_dns(&'static self, bgp_client: Arc<BGPClient>) -> impl Future<Item=(), Error=()> {
                let dns_file = self.store.clone() + "/nodes.dump";
-               let dns_future = File::create(dns_file.clone() + ".tmp").and_then(move |f| {
+               File::create(dns_file.clone() + ".tmp").and_then(move |f| {
                        let mut dns_buff = String::new();
                        {
-                               let nodes = self.nodes.read().unwrap();
                                let mut rng = thread_rng();
                                for i in &[1u64, 4, 5, 8, 9, 12, 13, 1024, 1025, 1028, 1029, 1032, 1033, 1036, 1037] {
-                                       let mut v6_set = Vec::new();
-                                       let mut v4_set = Vec::new();
-                                       if i.count_ones() == 1 {
-                                               for j in 0..64 {
-                                                       if i & (1 << j) != 0 {
-                                                               let set_ref = &nodes.good_node_services[j];
-                                                               v4_set = set_ref.iter().filter(|e| e.is_ipv4() && e.port() == 8333)
-                                                                       .choose_multiple(&mut rng, 21).iter().map(|e| e.ip()).collect();
-                                                               v6_set = set_ref.iter().filter(|e| e.is_ipv6() && e.port() == 8333)
-                                                                       .choose_multiple(&mut rng, 12).iter().map(|e| e.ip()).collect();
-                                                               break;
-                                                       }
-                                               }
-                                       } else if i.count_ones() == 2 {
-                                               let mut first_set = None;
-                                               let mut second_set = None;
-                                               for j in 0..64 {
-                                                       if i & (1 << j) != 0 {
-                                                               if first_set == None {
-                                                                       first_set = Some(&nodes.good_node_services[j]);
-                                                               } else {
-                                                                       second_set = Some(&nodes.good_node_services[j]);
+                                       let mut v6_set: Vec<IpAddr> = Vec::new();
+                                       let mut v4_set: Vec<IpAddr> = Vec::new();
+                                       {
+                                               let nodes = self.nodes.read().unwrap();
+                                               if i.count_ones() == 1 {
+                                                       for j in 0..64 {
+                                                               if i & (1 << j) != 0 {
+                                                                       let set_ref = &nodes.good_node_services[j];
+                                                                       v4_set = set_ref.iter().filter(|e| e.is_ipv4() && e.port() == 8333).map(|e| e.ip()).collect();
+                                                                       v6_set = set_ref.iter().filter(|e| e.is_ipv6() && e.port() == 8333).map(|e| e.ip()).collect();
                                                                        break;
                                                                }
                                                        }
-                                               }
-                                               v4_set = first_set.unwrap().intersection(&second_set.unwrap())
-                                                       .filter(|e| e.is_ipv4() && e.port() == 8333)
-                                                       .choose_multiple(&mut rng, 21).iter().map(|e| e.ip()).collect();
-                                               v6_set = first_set.unwrap().intersection(&second_set.unwrap())
-                                                       .filter(|e| e.is_ipv6() && e.port() == 8333)
-                                                       .choose_multiple(&mut rng, 12).iter().map(|e| e.ip()).collect();
-                                       } else {
-                                               //TODO: Could optimize this one a bit
-                                               let mut intersection;
-                                               let mut intersection_set_ref = None;
-                                               for j in 0..64 {
-                                                       if i & (1 << j) != 0 {
-                                                               if intersection_set_ref == None {
-                                                                       intersection_set_ref = Some(&nodes.good_node_services[j]);
-                                                               } else {
-                                                                       let new_intersection = intersection_set_ref.unwrap()
-                                                                               .intersection(&nodes.good_node_services[j]).map(|e| (*e).clone()).collect();
-                                                                       intersection = Some(new_intersection);
-                                                                       intersection_set_ref = Some(intersection.as_ref().unwrap());
+                                               } else if i.count_ones() == 2 {
+                                                       let mut first_set = None;
+                                                       let mut second_set = None;
+                                                       for j in 0..64 {
+                                                               if i & (1 << j) != 0 {
+                                                                       if first_set == None {
+                                                                               first_set = Some(&nodes.good_node_services[j]);
+                                                                       } else {
+                                                                               second_set = Some(&nodes.good_node_services[j]);
+                                                                               break;
+                                                                       }
+                                                               }
+                                                       }
+                                                       v4_set = first_set.unwrap().intersection(&second_set.unwrap())
+                                                               .filter(|e| e.is_ipv4() && e.port() == 8333).map(|e| e.ip()).collect();
+                                                       v6_set = first_set.unwrap().intersection(&second_set.unwrap())
+                                                               .filter(|e| e.is_ipv6() && e.port() == 8333).map(|e| e.ip()).collect();
+                                               } else {
+                                                       //TODO: Could optimize this one a bit
+                                                       let mut intersection;
+                                                       let mut intersection_set_ref = None;
+                                                       for j in 0..64 {
+                                                               if i & (1 << j) != 0 {
+                                                                       if intersection_set_ref == None {
+                                                                               intersection_set_ref = Some(&nodes.good_node_services[j]);
+                                                                       } else {
+                                                                               let new_intersection = intersection_set_ref.unwrap()
+                                                                                       .intersection(&nodes.good_node_services[j]).map(|e| (*e).clone()).collect();
+                                                                               intersection = Some(new_intersection);
+                                                                               intersection_set_ref = Some(intersection.as_ref().unwrap());
+                                                                       }
                                                                }
                                                        }
+                                                       v4_set = intersection_set_ref.unwrap().iter()
+                                                               .filter(|e| e.is_ipv4() && e.port() == 8333).map(|e| e.ip()).collect();
+                                                       v6_set = intersection_set_ref.unwrap().iter()
+                                                               .filter(|e| e.is_ipv6() && e.port() == 8333).map(|e| e.ip()).collect();
                                                }
-                                               v4_set = intersection_set_ref.unwrap().iter()
-                                                       .filter(|e| e.is_ipv4() && e.port() == 8333)
-                                                       .choose_multiple(&mut rng, 21).iter().map(|e| e.ip()).collect();
-                                               v6_set = intersection_set_ref.unwrap().iter()
-                                                       .filter(|e| e.is_ipv6() && e.port() == 8333)
-                                                       .choose_multiple(&mut rng, 12).iter().map(|e| e.ip()).collect();
                                        }
-                                       for a in v4_set {
+                                       let mut asn_set = HashSet::with_capacity(cmp::max(v4_set.len(), v6_set.len()));
+                                       for a in v4_set.iter().filter(|a| asn_set.insert(bgp_client.get_asn(**a))).choose_multiple(&mut rng, 21) {
                                                dns_buff += &format!("x{:x}.dnsseed\tIN\tA\t{}\n", i, a);
                                        }
-                                       for a in v6_set {
+                                       asn_set.clear();
+                                       for a in v6_set.iter().filter(|a| asn_set.insert(bgp_client.get_asn(**a))).choose_multiple(&mut rng, 12) {
                                                dns_buff += &format!("x{:x}.dnsseed\tIN\tAAAA\t{}\n", i, a);
                                        }
                                }
@@ -481,9 +513,7 @@ impl Store {
                        f.poll_sync_all()
                }).and_then(|_| {
                        tokio::fs::rename(dns_file.clone() + ".tmp", dns_file)
-               });
-
-               settings_future.join3(nodes_future, dns_future).then(|_| { future::ok(()) })
+               }).then(|_| { future::ok(()) })
        }
 
        pub fn get_next_scan_nodes(&self) -> Vec<SocketAddr> {
@@ -496,7 +526,7 @@ impl Store {
                        let mut nodes = self.nodes.write().unwrap();
                        for (idx, state_nodes) in nodes.state_next_scan.iter_mut().enumerate() {
                                let cmp_time = cur_time - Duration::from_secs(self.get_u64(U64Setting::RescanInterval(AddressState::from_num(idx as u8).unwrap())));
-                               let split_point = cmp::min(cmp::min(results - res.len(), results - (per_bucket_results * (AddressState::get_count() as usize - idx))),
+                               let split_point = cmp::min(cmp::min(results - res.len(), (per_bucket_results * (idx + 1)) - res.len()),
                                                state_nodes.binary_search_by(|a| a.0.cmp(&cmp_time)).unwrap_or_else(|idx| idx));
                                let mut new_nodes = state_nodes.split_off(split_point);
                                mem::swap(&mut new_nodes, state_nodes);