1 files changed, 97 insertions, 2 deletions

diff --git a/src/net_processing.cpp b/src/net_processing.cpp
index 92c88477c9..0eb01f36b2 100644
--- a/src/net_processing.cpp
+++ b/src/net_processing.cpp
@@ -23,6 +23,7 @@
 #include "primitives/transaction.h"
 #include "random.h"
 #include "reverse_iterator.h"
+#include "scheduler.h"
 #include "tinyformat.h"
 #include "txmempool.h"
 #include "ui_interface.h"
@@ -119,7 +120,6 @@ namespace {
     /** Number of outbound peers with m_chain_sync.m_protect. */
     int g_outbound_peers_with_protect_from_disconnect = 0;
 
-
     /** When our tip was last updated. */
     int64_t g_last_tip_update = 0;
 
@@ -428,6 +428,15 @@ void MaybeSetPeerAsAnnouncingHeaderAndIDs(NodeId nodeid, CConnman* connman) {
     }
 }
 
+bool TipMayBeStale(const Consensus::Params &consensusParams)
+{
+    AssertLockHeld(cs_main);
+    if (g_last_tip_update == 0) {
+        g_last_tip_update = GetTime();
+    }
+    return g_last_tip_update < GetTime() - consensusParams.nPowTargetSpacing * 3 && mapBlocksInFlight.empty();
+}
+
 // Requires cs_main
 bool CanDirectFetch(const Consensus::Params &consensusParams)
 {
@@ -754,9 +763,17 @@ void Misbehaving(NodeId pnode, int howmuch)
 // blockchain -> download logic notification
 //
 
-PeerLogicValidation::PeerLogicValidation(CConnman* connmanIn) : connman(connmanIn) {
+PeerLogicValidation::PeerLogicValidation(CConnman* connmanIn, CScheduler &scheduler) : connman(connmanIn), m_stale_tip_check_time(0) {
     // Initialize global variables that cannot be constructed at startup.
     recentRejects.reset(new CRollingBloomFilter(120000, 0.000001));
+
+    const Consensus::Params& consensusParams = Params().GetConsensus();
+    // Stale tip checking and peer eviction are on two different timers, but we
+    // don't want them to get out of sync due to drift in the scheduler, so we
+    // combine them in one function and schedule at the quicker (peer-eviction)
+    // timer.
+    static_assert(EXTRA_PEER_CHECK_INTERVAL < STALE_CHECK_INTERVAL, "peer eviction timer should be less than stale tip check timer");
+    scheduler.scheduleEvery(std::bind(&PeerLogicValidation::CheckForStaleTipAndEvictPeers, this, consensusParams), EXTRA_PEER_CHECK_INTERVAL * 1000);
 }
 
 void PeerLogicValidation::BlockConnected(const std::shared_ptr<const CBlock>& pblock, const CBlockIndex* pindex, const std::vector<CTransactionRef>& vtxConflicted) {
@@ -1412,6 +1429,7 @@ bool static ProcessHeadersMessage(CNode *pfrom, CConnman *connman, const std::ve
             // If this is an outbound peer, check to see if we should protect
             // it from the bad/lagging chain logic.
             if (g_outbound_peers_with_protect_from_disconnect < MAX_OUTBOUND_PEERS_TO_PROTECT_FROM_DISCONNECT && nodestate->pindexBestKnownBlock->nChainWork >= chainActive.Tip()->nChainWork && !nodestate->m_chain_sync.m_protect) {
+                LogPrint(BCLog::NET, "Protecting outbound peer=%d from eviction\n", pfrom->GetId());
                 nodestate->m_chain_sync.m_protect = true;
                 ++g_outbound_peers_with_protect_from_disconnect;
             }
@@ -2978,6 +2996,83 @@ void PeerLogicValidation::ConsiderEviction(CNode *pto, int64_t time_in_seconds)
     }
 }
 
+void PeerLogicValidation::EvictExtraOutboundPeers(int64_t time_in_seconds)
+{
+    // Check whether we have too many outbound peers
+    int extra_peers = connman->GetExtraOutboundCount();
+    if (extra_peers > 0) {
+        // If we have more outbound peers than we target, disconnect one.
+        // Pick the outbound peer that least recently announced
+        // us a new block, with ties broken by choosing the more recent
+        // connection (higher node id)
+        NodeId worst_peer = -1;
+        int64_t oldest_block_announcement = std::numeric_limits<int64_t>::max();
+
+        LOCK(cs_main);
+
+        connman->ForEachNode([&](CNode* pnode) {
+            // Ignore non-outbound peers, or nodes marked for disconnect already
+            if (!IsOutboundDisconnectionCandidate(pnode) || pnode->fDisconnect) return;
+            CNodeState *state = State(pnode->GetId());
+            if (state == nullptr) return; // shouldn't be possible, but just in case
+            // Don't evict our protected peers
+            if (state->m_chain_sync.m_protect) return;
+            if (state->m_last_block_announcement < oldest_block_announcement || (state->m_last_block_announcement == oldest_block_announcement && pnode->GetId() > worst_peer)) {
+                worst_peer = pnode->GetId();
+                oldest_block_announcement = state->m_last_block_announcement;
+            }
+        });
+        if (worst_peer != -1) {
+            bool disconnected = connman->ForNode(worst_peer, [&](CNode *pnode) {
+                // Only disconnect a peer that has been connected to us for
+                // some reasonable fraction of our check-frequency, to give
+                // it time for new information to have arrived.
+                // Also don't disconnect any peer we're trying to download a
+                // block from.
+                CNodeState &state = *State(pnode->GetId());
+                if (time_in_seconds - pnode->nTimeConnected > MINIMUM_CONNECT_TIME && state.nBlocksInFlight == 0) {
+                    LogPrint(BCLog::NET, "disconnecting extra outbound peer=%d (last block announcement received at time %d)\n", pnode->GetId(), oldest_block_announcement);
+                    pnode->fDisconnect = true;
+                    return true;
+                } else {
+                    LogPrint(BCLog::NET, "keeping outbound peer=%d chosen for eviction (connect time: %d, blocks_in_flight: %d)\n", pnode->GetId(), pnode->nTimeConnected, state.nBlocksInFlight);
+                    return false;
+                }
+            });
+            if (disconnected) {
+                // If we disconnected an extra peer, that means we successfully
+                // connected to at least one peer after the last time we
+                // detected a stale tip. Don't try any more extra peers until
+                // we next detect a stale tip, to limit the load we put on the
+                // network from these extra connections.
+                connman->SetTryNewOutboundPeer(false);
+            }
+        }
+    }
+}
+
+void PeerLogicValidation::CheckForStaleTipAndEvictPeers(const Consensus::Params &consensusParams)
+{
+    if (connman == nullptr) return;
+
+    int64_t time_in_seconds = GetTime();
+
+    EvictExtraOutboundPeers(time_in_seconds);
+
+    if (time_in_seconds > m_stale_tip_check_time) {
+        LOCK(cs_main);
+        // Check whether our tip is stale, and if so, allow using an extra
+        // outbound peer
+        if (TipMayBeStale(consensusParams)) {
+            LogPrintf("Potential stale tip detected, will try using extra outbound peer (last tip update: %d seconds ago)\n", time_in_seconds - g_last_tip_update);
+            connman->SetTryNewOutboundPeer(true);
+        } else if (connman->GetTryNewOutboundPeer()) {
+            connman->SetTryNewOutboundPeer(false);
+        }
+        m_stale_tip_check_time = time_in_seconds + STALE_CHECK_INTERVAL;
+    }
+}
+
 class CompareInvMempoolOrder
 {
     CTxMemPool *mp;