Merge bitcoin/bitcoin#22577: Close minor startup race between main and scheduler threads

703b1e612a Close minor startup race between main and scheduler threads (Larry Ruane)

Pull request description:

  This is a low-priority bug fix. The scheduler thread runs `CheckForStaleTipAndEvictPeers()` every 45 seconds (EXTRA_PEER_CHECK_INTERVAL). If its first run happens before the active chain is set up (`CChain::SetTip()`), `bitcoind` will assert:
  ```
  (...)
  2021-07-28T22:16:49Z init message: Loading block index…
  bitcoind: validation.cpp:4968: CChainState& ChainstateManager::ActiveChainstate() const: Assertion `m_active_chainstate' failed.
  Aborted (core dumped)
  ```
  I ran into this while using the debugger to investigate an unrelated problem. Single-stepping through threads with a debugger can cause the relative thread execution timing to be very different than usual. I don't think any automated tests are needed for this PR. I'll give reproduction steps in the next PR comment.

ACKs for top commit:
  MarcoFalke:
    cr ACK 703b1e612a
  tryphe:
    tested ACK 703b1e612a
  0xB10C:
    ACK 703b1e612a
  glozow:
    code review ACK 703b1e612a - it makes sense to me to start peerman's background tasks here, after `chainstate->LoadChainTip()` and `node.connman->Start()` have been called.

Tree-SHA512: 9316ad768cba3b171f62e2eb400e3790af66c47d1886d7965edb38d9710fc8c8f8e4fb38232811c9346732ce311d39f740c5c2aaf5f6ca390ddc48c51a8d633b

src/init.cpp

@@ -1185,7 +1185,7 @@ bool AppInitMain(NodeContext& node, interfaces::BlockAndHeaderTipInfo* tip_info)
 
     assert(!node.peerman);
     node.peerman = PeerManager::make(chainparams, *node.connman, *node.addrman, node.banman.get(),
-                                     *node.scheduler, chainman, *node.mempool, ignores_incoming_txs);
+                                     chainman, *node.mempool, ignores_incoming_txs);
     RegisterValidationInterface(node.peerman.get());
 
     // sanitize comments per BIP-0014, format user agent and check total size
@@ -1794,6 +1794,8 @@ bool AppInitMain(NodeContext& node, interfaces::BlockAndHeaderTipInfo* tip_info)
         banman->DumpBanlist();
     }, DUMP_BANS_INTERVAL);
 
+    if (node.peerman) node.peerman->StartScheduledTasks(*node.scheduler);
+
 #if HAVE_SYSTEM
     StartupNotify(args);
 #endif

src/net_processing.cpp

@@ -292,7 +292,7 @@ class PeerManagerImpl final : public PeerManager
 {
 public:
     PeerManagerImpl(const CChainParams& chainparams, CConnman& connman, CAddrMan& addrman,
-                    BanMan* banman, CScheduler& scheduler, ChainstateManager& chainman,
+                    BanMan* banman, ChainstateManager& chainman,
                     CTxMemPool& pool, bool ignore_incoming_txs);
 
     /** Overridden from CValidationInterface. */
@@ -309,6 +309,7 @@ public:
     bool SendMessages(CNode* pto) override EXCLUSIVE_LOCKS_REQUIRED(pto->cs_sendProcessing);
 
     /** Implement PeerManager */
+    void StartScheduledTasks(CScheduler& scheduler) override;
     void CheckForStaleTipAndEvictPeers() override;
     bool GetNodeStateStats(NodeId nodeid, CNodeStateStats& stats) const override;
     bool IgnoresIncomingTxs() override { return m_ignore_incoming_txs; }
@@ -1419,14 +1420,14 @@ bool PeerManagerImpl::BlockRequestAllowed(const CBlockIndex* pindex)
 }
 
 std::unique_ptr<PeerManager> PeerManager::make(const CChainParams& chainparams, CConnman& connman, CAddrMan& addrman,
-                                               BanMan* banman, CScheduler& scheduler, ChainstateManager& chainman,
+                                               BanMan* banman, ChainstateManager& chainman,
                                                CTxMemPool& pool, bool ignore_incoming_txs)
 {
-    return std::make_unique<PeerManagerImpl>(chainparams, connman, addrman, banman, scheduler, chainman, pool, ignore_incoming_txs);
+    return std::make_unique<PeerManagerImpl>(chainparams, connman, addrman, banman, chainman, pool, ignore_incoming_txs);
 }
 
 PeerManagerImpl::PeerManagerImpl(const CChainParams& chainparams, CConnman& connman, CAddrMan& addrman,
-                                 BanMan* banman, CScheduler& scheduler, ChainstateManager& chainman,
+                                 BanMan* banman, ChainstateManager& chainman,
                                  CTxMemPool& pool, bool ignore_incoming_txs)
     : m_chainparams(chainparams),
       m_connman(connman),
@@ -1435,6 +1436,10 @@ PeerManagerImpl::PeerManagerImpl(const CChainParams& chainparams, CConnman& conn
       m_chainman(chainman),
       m_mempool(pool),
       m_ignore_incoming_txs(ignore_incoming_txs)
+{
+}
+
+void PeerManagerImpl::StartScheduledTasks(CScheduler& scheduler)
 {
     // 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

src/net_processing.h

@@ -38,10 +38,13 @@ class PeerManager : public CValidationInterface, public NetEventsInterface
 {
 public:
     static std::unique_ptr<PeerManager> make(const CChainParams& chainparams, CConnman& connman, CAddrMan& addrman,
-                                             BanMan* banman, CScheduler& scheduler, ChainstateManager& chainman,
+                                             BanMan* banman, ChainstateManager& chainman,
                                              CTxMemPool& pool, bool ignore_incoming_txs);
     virtual ~PeerManager() { }
 
+    /** Begin running background tasks, should only be called once */
+    virtual void StartScheduledTasks(CScheduler& scheduler) = 0;
+
     /** Get statistics from node state */
     virtual bool GetNodeStateStats(NodeId nodeid, CNodeStateStats& stats) const = 0;
 

src/test/denialofservice_tests.cpp

@@ -53,7 +53,7 @@ BOOST_AUTO_TEST_CASE(outbound_slow_chain_eviction)
     const CChainParams& chainparams = Params();
     auto connman = std::make_unique<CConnman>(0x1337, 0x1337, *m_node.addrman);
     auto peerLogic = PeerManager::make(chainparams, *connman, *m_node.addrman, nullptr,
-                                       *m_node.scheduler, *m_node.chainman, *m_node.mempool, false);
+                                       *m_node.chainman, *m_node.mempool, false);
 
     // Mock an outbound peer
     CAddress addr1(ip(0xa0b0c001), NODE_NONE);
@@ -121,7 +121,7 @@ BOOST_AUTO_TEST_CASE(stale_tip_peer_management)
     const CChainParams& chainparams = Params();
     auto connman = std::make_unique<ConnmanTestMsg>(0x1337, 0x1337, *m_node.addrman);
     auto peerLogic = PeerManager::make(chainparams, *connman, *m_node.addrman, nullptr,
-                                       *m_node.scheduler, *m_node.chainman, *m_node.mempool, false);
+                                       *m_node.chainman, *m_node.mempool, false);
 
     constexpr int max_outbound_full_relay = MAX_OUTBOUND_FULL_RELAY_CONNECTIONS;
     CConnman::Options options;
@@ -194,7 +194,7 @@ BOOST_AUTO_TEST_CASE(peer_discouragement)
     auto banman = std::make_unique<BanMan>(m_args.GetDataDirBase() / "banlist", nullptr, DEFAULT_MISBEHAVING_BANTIME);
     auto connman = std::make_unique<ConnmanTestMsg>(0x1337, 0x1337, *m_node.addrman);
     auto peerLogic = PeerManager::make(chainparams, *connman, *m_node.addrman, banman.get(),
-                                       *m_node.scheduler, *m_node.chainman, *m_node.mempool, false);
+                                       *m_node.chainman, *m_node.mempool, false);
 
     CNetAddr tor_netaddr;
     BOOST_REQUIRE(
@@ -288,7 +288,7 @@ BOOST_AUTO_TEST_CASE(DoS_bantime)
     auto banman = std::make_unique<BanMan>(m_args.GetDataDirBase() / "banlist", nullptr, DEFAULT_MISBEHAVING_BANTIME);
     auto connman = std::make_unique<CConnman>(0x1337, 0x1337, *m_node.addrman);
     auto peerLogic = PeerManager::make(chainparams, *connman, *m_node.addrman, banman.get(),
-                                       *m_node.scheduler, *m_node.chainman, *m_node.mempool, false);
+                                       *m_node.chainman, *m_node.mempool, false);
 
     banman->ClearBanned();
     int64_t nStartTime = GetTime();

src/test/util/setup_common.cpp

@@ -197,7 +197,7 @@ TestingSetup::TestingSetup(const std::string& chainName, const std::vector<const
     m_node.banman = std::make_unique<BanMan>(m_args.GetDataDirBase() / "banlist", nullptr, DEFAULT_MISBEHAVING_BANTIME);
     m_node.connman = std::make_unique<CConnman>(0x1337, 0x1337, *m_node.addrman); // Deterministic randomness for tests.
     m_node.peerman = PeerManager::make(chainparams, *m_node.connman, *m_node.addrman,
-                                       m_node.banman.get(), *m_node.scheduler, *m_node.chainman,
+                                       m_node.banman.get(), *m_node.chainman,
                                        *m_node.mempool, false);
     {
         CConnman::Options options;