You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

884 lines
24 KiB

abci: localClient improvements & bugfixes & pubsub Unsubscribe issues (#2748) * use READ lock/unlock in ConsensusState#GetLastHeight Refs #2721 * do not use defers when there's no need * fix peer formatting (output its address instead of the pointer) ``` [54310]: E[11-02|11:59:39.851] Connection failed @ sendRoutine module=p2p peer=0xb78f00 conn=MConn{74.207.236.148:26656} err="pong timeout" ``` https://github.com/tendermint/tendermint/issues/2721#issuecomment-435326581 * panic if peer has no state https://github.com/tendermint/tendermint/issues/2721#issuecomment-435347165 It's confusing that sometimes we check if peer has a state, but most of the times we expect it to be there 1. https://github.com/tendermint/tendermint/blob/add79700b5fe84417538202b6c927c8cc5383672/mempool/reactor.go#L138 2. https://github.com/tendermint/tendermint/blob/add79700b5fe84417538202b6c927c8cc5383672/rpc/core/consensus.go#L196 (edited) I will change everything to always assume peer has a state and panic otherwise that should help identify issues earlier * abci/localclient: extend lock on app callback App callback should be protected by lock as well (note this was already done for InitChainAsync, why not for others???). Otherwise, when we execute the block, tx might come in and call the callback in the same time we're updating it in execBlockOnProxyApp => DATA RACE Fixes #2721 Consensus state is locked ``` goroutine 113333 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00180009c, 0xc0000c7e00) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*RWMutex).RLock(0xc001800090) /usr/local/go/src/sync/rwmutex.go:50 +0x4e github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).GetRoundState(0xc001800000, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:218 +0x46 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusReactor).queryMaj23Routine(0xc0017def80, 0x11104a0, 0xc0072488f0, 0xc007248 9c0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/reactor.go:735 +0x16d created by github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusReactor).AddPeer /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/reactor.go:172 +0x236 ``` because localClient is locked ``` goroutine 1899 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00003363c, 0xc0000cb500) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*Mutex).Lock(0xc000033638) /usr/local/go/src/sync/mutex.go:134 +0xff github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client.(*localClient).SetResponseCallback(0xc0001fb560, 0xc007868540) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client/local_client.go:32 +0x33 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy.(*appConnConsensus).SetResponseCallback(0xc00002f750, 0xc007868540) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy/app_conn.go:57 +0x40 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state.execBlockOnProxyApp(0x1104e20, 0xc002ca0ba0, 0x11092a0, 0xc00002f750, 0xc0001fe960, 0xc000bfc660, 0x110cfe0, 0xc000090330, 0xc9d12, 0xc000d9d5a0, ...) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state/execution.go:230 +0x1fd github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state.(*BlockExecutor).ApplyBlock(0xc002c2a230, 0x7, 0x0, 0xc000eae880, 0x6, 0xc002e52c60, 0x16, 0x1f927, 0xc9d12, 0xc000d9d5a0, ...) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state/execution.go:96 +0x142 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).finalizeCommit(0xc001800000, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1339 +0xa3e github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).tryFinalizeCommit(0xc001800000, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1270 +0x451 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).enterCommit.func1(0xc001800000, 0x0, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1218 +0x90 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).enterCommit(0xc001800000, 0x1f928, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1247 +0x6b8 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).addVote(0xc001800000, 0xc003d8dea0, 0xc000cf4cc0, 0x28, 0xf1, 0xc003bc7ad0, 0xc003bc7b10) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1659 +0xbad github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).tryAddVote(0xc001800000, 0xc003d8dea0, 0xc000cf4cc0, 0x28, 0xf1, 0xf1, 0xf1) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1517 +0x59 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).handleMsg(0xc001800000, 0xd98200, 0xc0070dbed0, 0xc000cf4cc0, 0x28) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:660 +0x64b github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).receiveRoutine(0xc001800000, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:617 +0x670 created by github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).OnStart /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:311 +0x132 ``` tx comes in and CheckTx is executed right when we execute the block ``` goroutine 111044 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00003363c, 0x0) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*Mutex).Lock(0xc000033638) /usr/local/go/src/sync/mutex.go:134 +0xff github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client.(*localClient).CheckTxAsync(0xc0001fb0e0, 0xc002d94500, 0x13f, 0x280, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client/local_client.go:85 +0x47 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy.(*appConnMempool).CheckTxAsync(0xc00002f720, 0xc002d94500, 0x13f, 0x280, 0x1) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy/app_conn.go:114 +0x51 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/mempool.(*Mempool).CheckTx(0xc002d3a320, 0xc002d94500, 0x13f, 0x280, 0xc0072355f0, 0x0, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/mempool/mempool.go:316 +0x17b github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/core.BroadcastTxSync(0xc002d94500, 0x13f, 0x280, 0x0, 0x0, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/core/mempool.go:93 +0xb8 reflect.Value.call(0xd85560, 0x10326c0, 0x13, 0xec7b8b, 0x4, 0xc00663f180, 0x1, 0x1, 0xc00663f180, 0xc00663f188, ...) /usr/local/go/src/reflect/value.go:447 +0x449 reflect.Value.Call(0xd85560, 0x10326c0, 0x13, 0xc00663f180, 0x1, 0x1, 0x0, 0x0, 0xc005cc9344) /usr/local/go/src/reflect/value.go:308 +0xa4 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.makeHTTPHandler.func2(0x1102060, 0xc00663f100, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/handlers.go:269 +0x188 net/http.HandlerFunc.ServeHTTP(0xc002c81f20, 0x1102060, 0xc00663f100, 0xc0082d7900) /usr/local/go/src/net/http/server.go:1964 +0x44 net/http.(*ServeMux).ServeHTTP(0xc002c81b60, 0x1102060, 0xc00663f100, 0xc0082d7900) /usr/local/go/src/net/http/server.go:2361 +0x127 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.maxBytesHandler.ServeHTTP(0x10f8a40, 0xc002c81b60, 0xf4240, 0x1102060, 0xc00663f100, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/http_server.go:219 +0xcf github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.RecoverAndLogHandler.func1(0x1103220, 0xc00121e620, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/http_server.go:192 +0x394 net/http.HandlerFunc.ServeHTTP(0xc002c06ea0, 0x1103220, 0xc00121e620, 0xc0082d7900) /usr/local/go/src/net/http/server.go:1964 +0x44 net/http.serverHandler.ServeHTTP(0xc001a1aa90, 0x1103220, 0xc00121e620, 0xc0082d7900) /usr/local/go/src/net/http/server.go:2741 +0xab net/http.(*conn).serve(0xc00785a3c0, 0x11041a0, 0xc000f844c0) /usr/local/go/src/net/http/server.go:1847 +0x646 created by net/http.(*Server).Serve /usr/local/go/src/net/http/server.go:2851 +0x2f5 ``` * consensus: use read lock in Receive#VoteMessage * use defer to unlock mutex because application might panic * use defer in every method of the localClient * add a changelog entry * drain channels before Unsubscribe(All) Read https://github.com/tendermint/tendermint/blob/55362ed76630f3e1ebec159a598f6a9fb5892cb1/libs/pubsub/pubsub.go#L13 for the detailed explanation of the issue. We'll need to fix it someday. Make sure to keep an eye on https://github.com/tendermint/tendermint/blob/master/docs/architecture/adr-033-pubsub.md * retry instead of panic when peer has no state in reactors other than consensus in /dump_consensus_state RPC endpoint, skip a peer with no state * rpc/core/mempool: simplify error messages * rpc/core/mempool: use time.After instead of timer also, do not log DeliverTx result (to be consistent with other memthods) * unlock before calling the callback in reqRes#SetCallback
6 years ago
abci: localClient improvements & bugfixes & pubsub Unsubscribe issues (#2748) * use READ lock/unlock in ConsensusState#GetLastHeight Refs #2721 * do not use defers when there's no need * fix peer formatting (output its address instead of the pointer) ``` [54310]: E[11-02|11:59:39.851] Connection failed @ sendRoutine module=p2p peer=0xb78f00 conn=MConn{74.207.236.148:26656} err="pong timeout" ``` https://github.com/tendermint/tendermint/issues/2721#issuecomment-435326581 * panic if peer has no state https://github.com/tendermint/tendermint/issues/2721#issuecomment-435347165 It's confusing that sometimes we check if peer has a state, but most of the times we expect it to be there 1. https://github.com/tendermint/tendermint/blob/add79700b5fe84417538202b6c927c8cc5383672/mempool/reactor.go#L138 2. https://github.com/tendermint/tendermint/blob/add79700b5fe84417538202b6c927c8cc5383672/rpc/core/consensus.go#L196 (edited) I will change everything to always assume peer has a state and panic otherwise that should help identify issues earlier * abci/localclient: extend lock on app callback App callback should be protected by lock as well (note this was already done for InitChainAsync, why not for others???). Otherwise, when we execute the block, tx might come in and call the callback in the same time we're updating it in execBlockOnProxyApp => DATA RACE Fixes #2721 Consensus state is locked ``` goroutine 113333 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00180009c, 0xc0000c7e00) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*RWMutex).RLock(0xc001800090) /usr/local/go/src/sync/rwmutex.go:50 +0x4e github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).GetRoundState(0xc001800000, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:218 +0x46 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusReactor).queryMaj23Routine(0xc0017def80, 0x11104a0, 0xc0072488f0, 0xc007248 9c0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/reactor.go:735 +0x16d created by github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusReactor).AddPeer /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/reactor.go:172 +0x236 ``` because localClient is locked ``` goroutine 1899 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00003363c, 0xc0000cb500) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*Mutex).Lock(0xc000033638) /usr/local/go/src/sync/mutex.go:134 +0xff github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client.(*localClient).SetResponseCallback(0xc0001fb560, 0xc007868540) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client/local_client.go:32 +0x33 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy.(*appConnConsensus).SetResponseCallback(0xc00002f750, 0xc007868540) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy/app_conn.go:57 +0x40 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state.execBlockOnProxyApp(0x1104e20, 0xc002ca0ba0, 0x11092a0, 0xc00002f750, 0xc0001fe960, 0xc000bfc660, 0x110cfe0, 0xc000090330, 0xc9d12, 0xc000d9d5a0, ...) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state/execution.go:230 +0x1fd github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state.(*BlockExecutor).ApplyBlock(0xc002c2a230, 0x7, 0x0, 0xc000eae880, 0x6, 0xc002e52c60, 0x16, 0x1f927, 0xc9d12, 0xc000d9d5a0, ...) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state/execution.go:96 +0x142 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).finalizeCommit(0xc001800000, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1339 +0xa3e github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).tryFinalizeCommit(0xc001800000, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1270 +0x451 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).enterCommit.func1(0xc001800000, 0x0, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1218 +0x90 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).enterCommit(0xc001800000, 0x1f928, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1247 +0x6b8 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).addVote(0xc001800000, 0xc003d8dea0, 0xc000cf4cc0, 0x28, 0xf1, 0xc003bc7ad0, 0xc003bc7b10) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1659 +0xbad github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).tryAddVote(0xc001800000, 0xc003d8dea0, 0xc000cf4cc0, 0x28, 0xf1, 0xf1, 0xf1) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1517 +0x59 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).handleMsg(0xc001800000, 0xd98200, 0xc0070dbed0, 0xc000cf4cc0, 0x28) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:660 +0x64b github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).receiveRoutine(0xc001800000, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:617 +0x670 created by github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).OnStart /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:311 +0x132 ``` tx comes in and CheckTx is executed right when we execute the block ``` goroutine 111044 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00003363c, 0x0) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*Mutex).Lock(0xc000033638) /usr/local/go/src/sync/mutex.go:134 +0xff github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client.(*localClient).CheckTxAsync(0xc0001fb0e0, 0xc002d94500, 0x13f, 0x280, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client/local_client.go:85 +0x47 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy.(*appConnMempool).CheckTxAsync(0xc00002f720, 0xc002d94500, 0x13f, 0x280, 0x1) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy/app_conn.go:114 +0x51 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/mempool.(*Mempool).CheckTx(0xc002d3a320, 0xc002d94500, 0x13f, 0x280, 0xc0072355f0, 0x0, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/mempool/mempool.go:316 +0x17b github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/core.BroadcastTxSync(0xc002d94500, 0x13f, 0x280, 0x0, 0x0, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/core/mempool.go:93 +0xb8 reflect.Value.call(0xd85560, 0x10326c0, 0x13, 0xec7b8b, 0x4, 0xc00663f180, 0x1, 0x1, 0xc00663f180, 0xc00663f188, ...) /usr/local/go/src/reflect/value.go:447 +0x449 reflect.Value.Call(0xd85560, 0x10326c0, 0x13, 0xc00663f180, 0x1, 0x1, 0x0, 0x0, 0xc005cc9344) /usr/local/go/src/reflect/value.go:308 +0xa4 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.makeHTTPHandler.func2(0x1102060, 0xc00663f100, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/handlers.go:269 +0x188 net/http.HandlerFunc.ServeHTTP(0xc002c81f20, 0x1102060, 0xc00663f100, 0xc0082d7900) /usr/local/go/src/net/http/server.go:1964 +0x44 net/http.(*ServeMux).ServeHTTP(0xc002c81b60, 0x1102060, 0xc00663f100, 0xc0082d7900) /usr/local/go/src/net/http/server.go:2361 +0x127 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.maxBytesHandler.ServeHTTP(0x10f8a40, 0xc002c81b60, 0xf4240, 0x1102060, 0xc00663f100, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/http_server.go:219 +0xcf github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.RecoverAndLogHandler.func1(0x1103220, 0xc00121e620, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/http_server.go:192 +0x394 net/http.HandlerFunc.ServeHTTP(0xc002c06ea0, 0x1103220, 0xc00121e620, 0xc0082d7900) /usr/local/go/src/net/http/server.go:1964 +0x44 net/http.serverHandler.ServeHTTP(0xc001a1aa90, 0x1103220, 0xc00121e620, 0xc0082d7900) /usr/local/go/src/net/http/server.go:2741 +0xab net/http.(*conn).serve(0xc00785a3c0, 0x11041a0, 0xc000f844c0) /usr/local/go/src/net/http/server.go:1847 +0x646 created by net/http.(*Server).Serve /usr/local/go/src/net/http/server.go:2851 +0x2f5 ``` * consensus: use read lock in Receive#VoteMessage * use defer to unlock mutex because application might panic * use defer in every method of the localClient * add a changelog entry * drain channels before Unsubscribe(All) Read https://github.com/tendermint/tendermint/blob/55362ed76630f3e1ebec159a598f6a9fb5892cb1/libs/pubsub/pubsub.go#L13 for the detailed explanation of the issue. We'll need to fix it someday. Make sure to keep an eye on https://github.com/tendermint/tendermint/blob/master/docs/architecture/adr-033-pubsub.md * retry instead of panic when peer has no state in reactors other than consensus in /dump_consensus_state RPC endpoint, skip a peer with no state * rpc/core/mempool: simplify error messages * rpc/core/mempool: use time.After instead of timer also, do not log DeliverTx result (to be consistent with other memthods) * unlock before calling the callback in reqRes#SetCallback
6 years ago
p2p: fix infinite loop in addrbook (#3232) * failing test * fix infinite loop in addrbook There are cases where we only have a small number of addresses marked good ("old"), but the selection mechanism keeps trying to select more of these addresses, and hence ends up in an infinite loop. Here we fix this to only try and select such "old" addresses if we have enough of them. Note this means, if we don't have enough of them, we may return more "new" addresses than otherwise expected by the newSelectionBias. This whole GetSelectionWithBias method probably needs to be rewritten, but this is a quick fix for the issue. * changelog * fix infinite loop if not enough new addrs * fix another potential infinite loop if a.nNew == 0 -> pickFromOldBucket=true, but we don't have enough items (a.nOld > len(oldBucketToAddrsMap) false) * Revert "fix another potential infinite loop" This reverts commit 146540c1125597162bd89820d611f6531f5e5e4b. * check num addresses instead of buckets, new test * fixed the int division * add slack to bias % in test, lint fixes * Added checks for selection content in test * test cleanup * Apply suggestions from code review Co-Authored-By: ebuchman <ethan@coinculture.info> * address review comments * change after Anton's review comments * use the same docker image we use for testing when building a binary for localnet * switch back to circleci classic * more review comments * more review comments * refactor addrbook_test * build linux binary inside docker in attempt to fix ``` --> Running dep + make build-linux GOOS=linux GOARCH=amd64 make build make[1]: Entering directory `/home/circleci/.go_workspace/src/github.com/tendermint/tendermint' CGO_ENABLED=0 go build -ldflags "-X github.com/tendermint/tendermint/version.GitCommit=`git rev-parse --short=8 HEAD`" -tags 'tendermint' -o build/tendermint ./cmd/tendermint/ p2p/pex/addrbook.go:373:13: undefined: math.Round ``` * change dir from /usr to /go * use concrete Go version for localnet binary * check for nil addresses just to be sure
6 years ago
p2p: fix infinite loop in addrbook (#3232) * failing test * fix infinite loop in addrbook There are cases where we only have a small number of addresses marked good ("old"), but the selection mechanism keeps trying to select more of these addresses, and hence ends up in an infinite loop. Here we fix this to only try and select such "old" addresses if we have enough of them. Note this means, if we don't have enough of them, we may return more "new" addresses than otherwise expected by the newSelectionBias. This whole GetSelectionWithBias method probably needs to be rewritten, but this is a quick fix for the issue. * changelog * fix infinite loop if not enough new addrs * fix another potential infinite loop if a.nNew == 0 -> pickFromOldBucket=true, but we don't have enough items (a.nOld > len(oldBucketToAddrsMap) false) * Revert "fix another potential infinite loop" This reverts commit 146540c1125597162bd89820d611f6531f5e5e4b. * check num addresses instead of buckets, new test * fixed the int division * add slack to bias % in test, lint fixes * Added checks for selection content in test * test cleanup * Apply suggestions from code review Co-Authored-By: ebuchman <ethan@coinculture.info> * address review comments * change after Anton's review comments * use the same docker image we use for testing when building a binary for localnet * switch back to circleci classic * more review comments * more review comments * refactor addrbook_test * build linux binary inside docker in attempt to fix ``` --> Running dep + make build-linux GOOS=linux GOARCH=amd64 make build make[1]: Entering directory `/home/circleci/.go_workspace/src/github.com/tendermint/tendermint' CGO_ENABLED=0 go build -ldflags "-X github.com/tendermint/tendermint/version.GitCommit=`git rev-parse --short=8 HEAD`" -tags 'tendermint' -o build/tendermint ./cmd/tendermint/ p2p/pex/addrbook.go:373:13: undefined: math.Round ``` * change dir from /usr to /go * use concrete Go version for localnet binary * check for nil addresses just to be sure
6 years ago
p2p: fix infinite loop in addrbook (#3232) * failing test * fix infinite loop in addrbook There are cases where we only have a small number of addresses marked good ("old"), but the selection mechanism keeps trying to select more of these addresses, and hence ends up in an infinite loop. Here we fix this to only try and select such "old" addresses if we have enough of them. Note this means, if we don't have enough of them, we may return more "new" addresses than otherwise expected by the newSelectionBias. This whole GetSelectionWithBias method probably needs to be rewritten, but this is a quick fix for the issue. * changelog * fix infinite loop if not enough new addrs * fix another potential infinite loop if a.nNew == 0 -> pickFromOldBucket=true, but we don't have enough items (a.nOld > len(oldBucketToAddrsMap) false) * Revert "fix another potential infinite loop" This reverts commit 146540c1125597162bd89820d611f6531f5e5e4b. * check num addresses instead of buckets, new test * fixed the int division * add slack to bias % in test, lint fixes * Added checks for selection content in test * test cleanup * Apply suggestions from code review Co-Authored-By: ebuchman <ethan@coinculture.info> * address review comments * change after Anton's review comments * use the same docker image we use for testing when building a binary for localnet * switch back to circleci classic * more review comments * more review comments * refactor addrbook_test * build linux binary inside docker in attempt to fix ``` --> Running dep + make build-linux GOOS=linux GOARCH=amd64 make build make[1]: Entering directory `/home/circleci/.go_workspace/src/github.com/tendermint/tendermint' CGO_ENABLED=0 go build -ldflags "-X github.com/tendermint/tendermint/version.GitCommit=`git rev-parse --short=8 HEAD`" -tags 'tendermint' -o build/tendermint ./cmd/tendermint/ p2p/pex/addrbook.go:373:13: undefined: math.Round ``` * change dir from /usr to /go * use concrete Go version for localnet binary * check for nil addresses just to be sure
6 years ago
p2p: fix infinite loop in addrbook (#3232) * failing test * fix infinite loop in addrbook There are cases where we only have a small number of addresses marked good ("old"), but the selection mechanism keeps trying to select more of these addresses, and hence ends up in an infinite loop. Here we fix this to only try and select such "old" addresses if we have enough of them. Note this means, if we don't have enough of them, we may return more "new" addresses than otherwise expected by the newSelectionBias. This whole GetSelectionWithBias method probably needs to be rewritten, but this is a quick fix for the issue. * changelog * fix infinite loop if not enough new addrs * fix another potential infinite loop if a.nNew == 0 -> pickFromOldBucket=true, but we don't have enough items (a.nOld > len(oldBucketToAddrsMap) false) * Revert "fix another potential infinite loop" This reverts commit 146540c1125597162bd89820d611f6531f5e5e4b. * check num addresses instead of buckets, new test * fixed the int division * add slack to bias % in test, lint fixes * Added checks for selection content in test * test cleanup * Apply suggestions from code review Co-Authored-By: ebuchman <ethan@coinculture.info> * address review comments * change after Anton's review comments * use the same docker image we use for testing when building a binary for localnet * switch back to circleci classic * more review comments * more review comments * refactor addrbook_test * build linux binary inside docker in attempt to fix ``` --> Running dep + make build-linux GOOS=linux GOARCH=amd64 make build make[1]: Entering directory `/home/circleci/.go_workspace/src/github.com/tendermint/tendermint' CGO_ENABLED=0 go build -ldflags "-X github.com/tendermint/tendermint/version.GitCommit=`git rev-parse --short=8 HEAD`" -tags 'tendermint' -o build/tendermint ./cmd/tendermint/ p2p/pex/addrbook.go:373:13: undefined: math.Round ``` * change dir from /usr to /go * use concrete Go version for localnet binary * check for nil addresses just to be sure
6 years ago
p2p: fix infinite loop in addrbook (#3232) * failing test * fix infinite loop in addrbook There are cases where we only have a small number of addresses marked good ("old"), but the selection mechanism keeps trying to select more of these addresses, and hence ends up in an infinite loop. Here we fix this to only try and select such "old" addresses if we have enough of them. Note this means, if we don't have enough of them, we may return more "new" addresses than otherwise expected by the newSelectionBias. This whole GetSelectionWithBias method probably needs to be rewritten, but this is a quick fix for the issue. * changelog * fix infinite loop if not enough new addrs * fix another potential infinite loop if a.nNew == 0 -> pickFromOldBucket=true, but we don't have enough items (a.nOld > len(oldBucketToAddrsMap) false) * Revert "fix another potential infinite loop" This reverts commit 146540c1125597162bd89820d611f6531f5e5e4b. * check num addresses instead of buckets, new test * fixed the int division * add slack to bias % in test, lint fixes * Added checks for selection content in test * test cleanup * Apply suggestions from code review Co-Authored-By: ebuchman <ethan@coinculture.info> * address review comments * change after Anton's review comments * use the same docker image we use for testing when building a binary for localnet * switch back to circleci classic * more review comments * more review comments * refactor addrbook_test * build linux binary inside docker in attempt to fix ``` --> Running dep + make build-linux GOOS=linux GOARCH=amd64 make build make[1]: Entering directory `/home/circleci/.go_workspace/src/github.com/tendermint/tendermint' CGO_ENABLED=0 go build -ldflags "-X github.com/tendermint/tendermint/version.GitCommit=`git rev-parse --short=8 HEAD`" -tags 'tendermint' -o build/tendermint ./cmd/tendermint/ p2p/pex/addrbook.go:373:13: undefined: math.Round ``` * change dir from /usr to /go * use concrete Go version for localnet binary * check for nil addresses just to be sure
6 years ago
abci: localClient improvements & bugfixes & pubsub Unsubscribe issues (#2748) * use READ lock/unlock in ConsensusState#GetLastHeight Refs #2721 * do not use defers when there's no need * fix peer formatting (output its address instead of the pointer) ``` [54310]: E[11-02|11:59:39.851] Connection failed @ sendRoutine module=p2p peer=0xb78f00 conn=MConn{74.207.236.148:26656} err="pong timeout" ``` https://github.com/tendermint/tendermint/issues/2721#issuecomment-435326581 * panic if peer has no state https://github.com/tendermint/tendermint/issues/2721#issuecomment-435347165 It's confusing that sometimes we check if peer has a state, but most of the times we expect it to be there 1. https://github.com/tendermint/tendermint/blob/add79700b5fe84417538202b6c927c8cc5383672/mempool/reactor.go#L138 2. https://github.com/tendermint/tendermint/blob/add79700b5fe84417538202b6c927c8cc5383672/rpc/core/consensus.go#L196 (edited) I will change everything to always assume peer has a state and panic otherwise that should help identify issues earlier * abci/localclient: extend lock on app callback App callback should be protected by lock as well (note this was already done for InitChainAsync, why not for others???). Otherwise, when we execute the block, tx might come in and call the callback in the same time we're updating it in execBlockOnProxyApp => DATA RACE Fixes #2721 Consensus state is locked ``` goroutine 113333 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00180009c, 0xc0000c7e00) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*RWMutex).RLock(0xc001800090) /usr/local/go/src/sync/rwmutex.go:50 +0x4e github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).GetRoundState(0xc001800000, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:218 +0x46 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusReactor).queryMaj23Routine(0xc0017def80, 0x11104a0, 0xc0072488f0, 0xc007248 9c0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/reactor.go:735 +0x16d created by github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusReactor).AddPeer /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/reactor.go:172 +0x236 ``` because localClient is locked ``` goroutine 1899 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00003363c, 0xc0000cb500) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*Mutex).Lock(0xc000033638) /usr/local/go/src/sync/mutex.go:134 +0xff github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client.(*localClient).SetResponseCallback(0xc0001fb560, 0xc007868540) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client/local_client.go:32 +0x33 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy.(*appConnConsensus).SetResponseCallback(0xc00002f750, 0xc007868540) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy/app_conn.go:57 +0x40 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state.execBlockOnProxyApp(0x1104e20, 0xc002ca0ba0, 0x11092a0, 0xc00002f750, 0xc0001fe960, 0xc000bfc660, 0x110cfe0, 0xc000090330, 0xc9d12, 0xc000d9d5a0, ...) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state/execution.go:230 +0x1fd github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state.(*BlockExecutor).ApplyBlock(0xc002c2a230, 0x7, 0x0, 0xc000eae880, 0x6, 0xc002e52c60, 0x16, 0x1f927, 0xc9d12, 0xc000d9d5a0, ...) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/state/execution.go:96 +0x142 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).finalizeCommit(0xc001800000, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1339 +0xa3e github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).tryFinalizeCommit(0xc001800000, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1270 +0x451 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).enterCommit.func1(0xc001800000, 0x0, 0x1f928) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1218 +0x90 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).enterCommit(0xc001800000, 0x1f928, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1247 +0x6b8 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).addVote(0xc001800000, 0xc003d8dea0, 0xc000cf4cc0, 0x28, 0xf1, 0xc003bc7ad0, 0xc003bc7b10) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1659 +0xbad github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).tryAddVote(0xc001800000, 0xc003d8dea0, 0xc000cf4cc0, 0x28, 0xf1, 0xf1, 0xf1) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:1517 +0x59 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).handleMsg(0xc001800000, 0xd98200, 0xc0070dbed0, 0xc000cf4cc0, 0x28) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:660 +0x64b github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).receiveRoutine(0xc001800000, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:617 +0x670 created by github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus.(*ConsensusState).OnStart /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/consensus/state.go:311 +0x132 ``` tx comes in and CheckTx is executed right when we execute the block ``` goroutine 111044 [semacquire, 309 minutes]: sync.runtime_SemacquireMutex(0xc00003363c, 0x0) /usr/local/go/src/runtime/sema.go:71 +0x3d sync.(*Mutex).Lock(0xc000033638) /usr/local/go/src/sync/mutex.go:134 +0xff github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client.(*localClient).CheckTxAsync(0xc0001fb0e0, 0xc002d94500, 0x13f, 0x280, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/abci/client/local_client.go:85 +0x47 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy.(*appConnMempool).CheckTxAsync(0xc00002f720, 0xc002d94500, 0x13f, 0x280, 0x1) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/proxy/app_conn.go:114 +0x51 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/mempool.(*Mempool).CheckTx(0xc002d3a320, 0xc002d94500, 0x13f, 0x280, 0xc0072355f0, 0x0, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/mempool/mempool.go:316 +0x17b github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/core.BroadcastTxSync(0xc002d94500, 0x13f, 0x280, 0x0, 0x0, 0x0) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/core/mempool.go:93 +0xb8 reflect.Value.call(0xd85560, 0x10326c0, 0x13, 0xec7b8b, 0x4, 0xc00663f180, 0x1, 0x1, 0xc00663f180, 0xc00663f188, ...) /usr/local/go/src/reflect/value.go:447 +0x449 reflect.Value.Call(0xd85560, 0x10326c0, 0x13, 0xc00663f180, 0x1, 0x1, 0x0, 0x0, 0xc005cc9344) /usr/local/go/src/reflect/value.go:308 +0xa4 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.makeHTTPHandler.func2(0x1102060, 0xc00663f100, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/handlers.go:269 +0x188 net/http.HandlerFunc.ServeHTTP(0xc002c81f20, 0x1102060, 0xc00663f100, 0xc0082d7900) /usr/local/go/src/net/http/server.go:1964 +0x44 net/http.(*ServeMux).ServeHTTP(0xc002c81b60, 0x1102060, 0xc00663f100, 0xc0082d7900) /usr/local/go/src/net/http/server.go:2361 +0x127 github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.maxBytesHandler.ServeHTTP(0x10f8a40, 0xc002c81b60, 0xf4240, 0x1102060, 0xc00663f100, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/http_server.go:219 +0xcf github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server.RecoverAndLogHandler.func1(0x1103220, 0xc00121e620, 0xc0082d7900) /root/go/src/github.com/MinterTeam/minter-go-node/vendor/github.com/tendermint/tendermint/rpc/lib/server/http_server.go:192 +0x394 net/http.HandlerFunc.ServeHTTP(0xc002c06ea0, 0x1103220, 0xc00121e620, 0xc0082d7900) /usr/local/go/src/net/http/server.go:1964 +0x44 net/http.serverHandler.ServeHTTP(0xc001a1aa90, 0x1103220, 0xc00121e620, 0xc0082d7900) /usr/local/go/src/net/http/server.go:2741 +0xab net/http.(*conn).serve(0xc00785a3c0, 0x11041a0, 0xc000f844c0) /usr/local/go/src/net/http/server.go:1847 +0x646 created by net/http.(*Server).Serve /usr/local/go/src/net/http/server.go:2851 +0x2f5 ``` * consensus: use read lock in Receive#VoteMessage * use defer to unlock mutex because application might panic * use defer in every method of the localClient * add a changelog entry * drain channels before Unsubscribe(All) Read https://github.com/tendermint/tendermint/blob/55362ed76630f3e1ebec159a598f6a9fb5892cb1/libs/pubsub/pubsub.go#L13 for the detailed explanation of the issue. We'll need to fix it someday. Make sure to keep an eye on https://github.com/tendermint/tendermint/blob/master/docs/architecture/adr-033-pubsub.md * retry instead of panic when peer has no state in reactors other than consensus in /dump_consensus_state RPC endpoint, skip a peer with no state * rpc/core/mempool: simplify error messages * rpc/core/mempool: use time.After instead of timer also, do not log DeliverTx result (to be consistent with other memthods) * unlock before calling the callback in reqRes#SetCallback
6 years ago
  1. // Modified for Tendermint
  2. // Originally Copyright (c) 2013-2014 Conformal Systems LLC.
  3. // https://github.com/conformal/btcd/blob/master/LICENSE
  4. package pex
  5. import (
  6. "crypto/sha256"
  7. "encoding/binary"
  8. "fmt"
  9. "math"
  10. "net"
  11. "sync"
  12. "time"
  13. "github.com/tendermint/tendermint/crypto"
  14. cmn "github.com/tendermint/tendermint/libs/common"
  15. "github.com/tendermint/tendermint/p2p"
  16. )
  17. const (
  18. bucketTypeNew = 0x01
  19. bucketTypeOld = 0x02
  20. )
  21. // AddrBook is an address book used for tracking peers
  22. // so we can gossip about them to others and select
  23. // peers to dial.
  24. // TODO: break this up?
  25. type AddrBook interface {
  26. cmn.Service
  27. // Add our own addresses so we don't later add ourselves
  28. AddOurAddress(*p2p.NetAddress)
  29. // Check if it is our address
  30. OurAddress(*p2p.NetAddress) bool
  31. AddPrivateIDs([]string)
  32. // Add and remove an address
  33. AddAddress(addr *p2p.NetAddress, src *p2p.NetAddress) error
  34. RemoveAddress(*p2p.NetAddress)
  35. // Check if the address is in the book
  36. HasAddress(*p2p.NetAddress) bool
  37. // Do we need more peers?
  38. NeedMoreAddrs() bool
  39. // Is Address Book Empty? Answer should not depend on being in your own
  40. // address book, or private peers
  41. Empty() bool
  42. // Pick an address to dial
  43. PickAddress(biasTowardsNewAddrs int) *p2p.NetAddress
  44. // Mark address
  45. MarkGood(*p2p.NetAddress)
  46. MarkAttempt(*p2p.NetAddress)
  47. MarkBad(*p2p.NetAddress)
  48. IsGood(*p2p.NetAddress) bool
  49. // Send a selection of addresses to peers
  50. GetSelection() []*p2p.NetAddress
  51. // Send a selection of addresses with bias
  52. GetSelectionWithBias(biasTowardsNewAddrs int) []*p2p.NetAddress
  53. // TODO: remove
  54. ListOfKnownAddresses() []*knownAddress
  55. // Persist to disk
  56. Save()
  57. }
  58. var _ AddrBook = (*addrBook)(nil)
  59. // addrBook - concurrency safe peer address manager.
  60. // Implements AddrBook.
  61. type addrBook struct {
  62. cmn.BaseService
  63. // immutable after creation
  64. filePath string
  65. routabilityStrict bool
  66. key string // random prefix for bucket placement
  67. // accessed concurrently
  68. mtx sync.Mutex
  69. rand *cmn.Rand
  70. ourAddrs map[string]struct{}
  71. privateIDs map[p2p.ID]struct{}
  72. addrLookup map[p2p.ID]*knownAddress // new & old
  73. bucketsOld []map[string]*knownAddress
  74. bucketsNew []map[string]*knownAddress
  75. nOld int
  76. nNew int
  77. wg sync.WaitGroup
  78. }
  79. // NewAddrBook creates a new address book.
  80. // Use Start to begin processing asynchronous address updates.
  81. func NewAddrBook(filePath string, routabilityStrict bool) *addrBook {
  82. am := &addrBook{
  83. rand: cmn.NewRand(),
  84. ourAddrs: make(map[string]struct{}),
  85. privateIDs: make(map[p2p.ID]struct{}),
  86. addrLookup: make(map[p2p.ID]*knownAddress),
  87. filePath: filePath,
  88. routabilityStrict: routabilityStrict,
  89. }
  90. am.init()
  91. am.BaseService = *cmn.NewBaseService(nil, "AddrBook", am)
  92. return am
  93. }
  94. // Initialize the buckets.
  95. // When modifying this, don't forget to update loadFromFile()
  96. func (a *addrBook) init() {
  97. a.key = crypto.CRandHex(24) // 24/2 * 8 = 96 bits
  98. // New addr buckets
  99. a.bucketsNew = make([]map[string]*knownAddress, newBucketCount)
  100. for i := range a.bucketsNew {
  101. a.bucketsNew[i] = make(map[string]*knownAddress)
  102. }
  103. // Old addr buckets
  104. a.bucketsOld = make([]map[string]*knownAddress, oldBucketCount)
  105. for i := range a.bucketsOld {
  106. a.bucketsOld[i] = make(map[string]*knownAddress)
  107. }
  108. }
  109. // OnStart implements Service.
  110. func (a *addrBook) OnStart() error {
  111. if err := a.BaseService.OnStart(); err != nil {
  112. return err
  113. }
  114. a.loadFromFile(a.filePath)
  115. // wg.Add to ensure that any invocation of .Wait()
  116. // later on will wait for saveRoutine to terminate.
  117. a.wg.Add(1)
  118. go a.saveRoutine()
  119. return nil
  120. }
  121. // OnStop implements Service.
  122. func (a *addrBook) OnStop() {
  123. a.BaseService.OnStop()
  124. }
  125. func (a *addrBook) Wait() {
  126. a.wg.Wait()
  127. }
  128. func (a *addrBook) FilePath() string {
  129. return a.filePath
  130. }
  131. //-------------------------------------------------------
  132. // AddOurAddress one of our addresses.
  133. func (a *addrBook) AddOurAddress(addr *p2p.NetAddress) {
  134. a.mtx.Lock()
  135. defer a.mtx.Unlock()
  136. a.Logger.Info("Add our address to book", "addr", addr)
  137. a.ourAddrs[addr.String()] = struct{}{}
  138. }
  139. // OurAddress returns true if it is our address.
  140. func (a *addrBook) OurAddress(addr *p2p.NetAddress) bool {
  141. a.mtx.Lock()
  142. defer a.mtx.Unlock()
  143. _, ok := a.ourAddrs[addr.String()]
  144. return ok
  145. }
  146. func (a *addrBook) AddPrivateIDs(IDs []string) {
  147. a.mtx.Lock()
  148. defer a.mtx.Unlock()
  149. for _, id := range IDs {
  150. a.privateIDs[p2p.ID(id)] = struct{}{}
  151. }
  152. }
  153. // AddAddress implements AddrBook
  154. // Add address to a "new" bucket. If it's already in one, only add it probabilistically.
  155. // Returns error if the addr is non-routable. Does not add self.
  156. // NOTE: addr must not be nil
  157. func (a *addrBook) AddAddress(addr *p2p.NetAddress, src *p2p.NetAddress) error {
  158. a.mtx.Lock()
  159. defer a.mtx.Unlock()
  160. return a.addAddress(addr, src)
  161. }
  162. // RemoveAddress implements AddrBook - removes the address from the book.
  163. func (a *addrBook) RemoveAddress(addr *p2p.NetAddress) {
  164. a.mtx.Lock()
  165. defer a.mtx.Unlock()
  166. ka := a.addrLookup[addr.ID]
  167. if ka == nil {
  168. return
  169. }
  170. a.Logger.Info("Remove address from book", "addr", addr)
  171. a.removeFromAllBuckets(ka)
  172. }
  173. // IsGood returns true if peer was ever marked as good and haven't
  174. // done anything wrong since then.
  175. func (a *addrBook) IsGood(addr *p2p.NetAddress) bool {
  176. a.mtx.Lock()
  177. defer a.mtx.Unlock()
  178. return a.addrLookup[addr.ID].isOld()
  179. }
  180. // HasAddress returns true if the address is in the book.
  181. func (a *addrBook) HasAddress(addr *p2p.NetAddress) bool {
  182. a.mtx.Lock()
  183. defer a.mtx.Unlock()
  184. ka := a.addrLookup[addr.ID]
  185. return ka != nil
  186. }
  187. // NeedMoreAddrs implements AddrBook - returns true if there are not have enough addresses in the book.
  188. func (a *addrBook) NeedMoreAddrs() bool {
  189. return a.Size() < needAddressThreshold
  190. }
  191. // Empty implements AddrBook - returns true if there are no addresses in the address book.
  192. // Does not count the peer appearing in its own address book, or private peers.
  193. func (a *addrBook) Empty() bool {
  194. return a.Size() == 0
  195. }
  196. // PickAddress implements AddrBook. It picks an address to connect to.
  197. // The address is picked randomly from an old or new bucket according
  198. // to the biasTowardsNewAddrs argument, which must be between [0, 100] (or else is truncated to that range)
  199. // and determines how biased we are to pick an address from a new bucket.
  200. // PickAddress returns nil if the AddrBook is empty or if we try to pick
  201. // from an empty bucket.
  202. func (a *addrBook) PickAddress(biasTowardsNewAddrs int) *p2p.NetAddress {
  203. a.mtx.Lock()
  204. defer a.mtx.Unlock()
  205. bookSize := a.size()
  206. if bookSize <= 0 {
  207. if bookSize < 0 {
  208. a.Logger.Error("Addrbook size less than 0", "nNew", a.nNew, "nOld", a.nOld)
  209. }
  210. return nil
  211. }
  212. if biasTowardsNewAddrs > 100 {
  213. biasTowardsNewAddrs = 100
  214. }
  215. if biasTowardsNewAddrs < 0 {
  216. biasTowardsNewAddrs = 0
  217. }
  218. // Bias between new and old addresses.
  219. oldCorrelation := math.Sqrt(float64(a.nOld)) * (100.0 - float64(biasTowardsNewAddrs))
  220. newCorrelation := math.Sqrt(float64(a.nNew)) * float64(biasTowardsNewAddrs)
  221. // pick a random peer from a random bucket
  222. var bucket map[string]*knownAddress
  223. pickFromOldBucket := (newCorrelation+oldCorrelation)*a.rand.Float64() < oldCorrelation
  224. if (pickFromOldBucket && a.nOld == 0) ||
  225. (!pickFromOldBucket && a.nNew == 0) {
  226. return nil
  227. }
  228. // loop until we pick a random non-empty bucket
  229. for len(bucket) == 0 {
  230. if pickFromOldBucket {
  231. bucket = a.bucketsOld[a.rand.Intn(len(a.bucketsOld))]
  232. } else {
  233. bucket = a.bucketsNew[a.rand.Intn(len(a.bucketsNew))]
  234. }
  235. }
  236. // pick a random index and loop over the map to return that index
  237. randIndex := a.rand.Intn(len(bucket))
  238. for _, ka := range bucket {
  239. if randIndex == 0 {
  240. return ka.Addr
  241. }
  242. randIndex--
  243. }
  244. return nil
  245. }
  246. // MarkGood implements AddrBook - it marks the peer as good and
  247. // moves it into an "old" bucket.
  248. func (a *addrBook) MarkGood(addr *p2p.NetAddress) {
  249. a.mtx.Lock()
  250. defer a.mtx.Unlock()
  251. ka := a.addrLookup[addr.ID]
  252. if ka == nil {
  253. return
  254. }
  255. ka.markGood()
  256. if ka.isNew() {
  257. a.moveToOld(ka)
  258. }
  259. }
  260. // MarkAttempt implements AddrBook - it marks that an attempt was made to connect to the address.
  261. func (a *addrBook) MarkAttempt(addr *p2p.NetAddress) {
  262. a.mtx.Lock()
  263. defer a.mtx.Unlock()
  264. ka := a.addrLookup[addr.ID]
  265. if ka == nil {
  266. return
  267. }
  268. ka.markAttempt()
  269. }
  270. // MarkBad implements AddrBook. Currently it just ejects the address.
  271. // TODO: black list for some amount of time
  272. func (a *addrBook) MarkBad(addr *p2p.NetAddress) {
  273. a.RemoveAddress(addr)
  274. }
  275. // GetSelection implements AddrBook.
  276. // It randomly selects some addresses (old & new). Suitable for peer-exchange protocols.
  277. // Must never return a nil address.
  278. func (a *addrBook) GetSelection() []*p2p.NetAddress {
  279. a.mtx.Lock()
  280. defer a.mtx.Unlock()
  281. bookSize := a.size()
  282. if bookSize <= 0 {
  283. if bookSize < 0 {
  284. a.Logger.Error("Addrbook size less than 0", "nNew", a.nNew, "nOld", a.nOld)
  285. }
  286. return nil
  287. }
  288. numAddresses := cmn.MaxInt(
  289. cmn.MinInt(minGetSelection, bookSize),
  290. bookSize*getSelectionPercent/100)
  291. numAddresses = cmn.MinInt(maxGetSelection, numAddresses)
  292. // XXX: instead of making a list of all addresses, shuffling, and slicing a random chunk,
  293. // could we just select a random numAddresses of indexes?
  294. allAddr := make([]*p2p.NetAddress, bookSize)
  295. i := 0
  296. for _, ka := range a.addrLookup {
  297. allAddr[i] = ka.Addr
  298. i++
  299. }
  300. // Fisher-Yates shuffle the array. We only need to do the first
  301. // `numAddresses' since we are throwing the rest.
  302. for i := 0; i < numAddresses; i++ {
  303. // pick a number between current index and the end
  304. j := cmn.RandIntn(len(allAddr)-i) + i
  305. allAddr[i], allAddr[j] = allAddr[j], allAddr[i]
  306. }
  307. // slice off the limit we are willing to share.
  308. return allAddr[:numAddresses]
  309. }
  310. func percentageOfNum(p, n int) int {
  311. return int(math.Round((float64(p) / float64(100)) * float64(n)))
  312. }
  313. // GetSelectionWithBias implements AddrBook.
  314. // It randomly selects some addresses (old & new). Suitable for peer-exchange protocols.
  315. // Must never return a nil address.
  316. //
  317. // Each address is picked randomly from an old or new bucket according to the
  318. // biasTowardsNewAddrs argument, which must be between [0, 100] (or else is truncated to
  319. // that range) and determines how biased we are to pick an address from a new
  320. // bucket.
  321. func (a *addrBook) GetSelectionWithBias(biasTowardsNewAddrs int) []*p2p.NetAddress {
  322. a.mtx.Lock()
  323. defer a.mtx.Unlock()
  324. bookSize := a.size()
  325. if bookSize <= 0 {
  326. if bookSize < 0 {
  327. a.Logger.Error("Addrbook size less than 0", "nNew", a.nNew, "nOld", a.nOld)
  328. }
  329. return nil
  330. }
  331. if biasTowardsNewAddrs > 100 {
  332. biasTowardsNewAddrs = 100
  333. }
  334. if biasTowardsNewAddrs < 0 {
  335. biasTowardsNewAddrs = 0
  336. }
  337. numAddresses := cmn.MaxInt(
  338. cmn.MinInt(minGetSelection, bookSize),
  339. bookSize*getSelectionPercent/100)
  340. numAddresses = cmn.MinInt(maxGetSelection, numAddresses)
  341. selection := make([]*p2p.NetAddress, numAddresses)
  342. oldBucketToAddrsMap := make(map[int]map[string]struct{})
  343. var oldIndex int
  344. newBucketToAddrsMap := make(map[int]map[string]struct{})
  345. var newIndex int
  346. // initialize counters used to count old and new added addresses.
  347. // len(oldBucketToAddrsMap) cannot be used as multiple addresses can endup in the same bucket.
  348. var oldAddressesAdded int
  349. var newAddressesAdded int
  350. // number of new addresses that, if possible, should be in the beginning of the selection
  351. numRequiredNewAdd := percentageOfNum(biasTowardsNewAddrs, numAddresses)
  352. selectionIndex := 0
  353. ADDRS_LOOP:
  354. for selectionIndex < numAddresses {
  355. // biasedTowardsOldAddrs indicates if the selection can switch to old addresses
  356. biasedTowardsOldAddrs := selectionIndex >= numRequiredNewAdd
  357. // An old addresses is selected if:
  358. // - the bias is for old and old addressees are still available or,
  359. // - there are no new addresses or all new addresses have been selected.
  360. // numAddresses <= a.nOld + a.nNew therefore it is guaranteed that there are enough
  361. // addresses to fill the selection
  362. pickFromOldBucket :=
  363. (biasedTowardsOldAddrs && oldAddressesAdded < a.nOld) ||
  364. a.nNew == 0 || newAddressesAdded >= a.nNew
  365. bucket := make(map[string]*knownAddress)
  366. // loop until we pick a random non-empty bucket
  367. for len(bucket) == 0 {
  368. if pickFromOldBucket {
  369. oldIndex = a.rand.Intn(len(a.bucketsOld))
  370. bucket = a.bucketsOld[oldIndex]
  371. } else {
  372. newIndex = a.rand.Intn(len(a.bucketsNew))
  373. bucket = a.bucketsNew[newIndex]
  374. }
  375. }
  376. // pick a random index
  377. randIndex := a.rand.Intn(len(bucket))
  378. // loop over the map to return that index
  379. var selectedAddr *p2p.NetAddress
  380. for _, ka := range bucket {
  381. if randIndex == 0 {
  382. selectedAddr = ka.Addr
  383. break
  384. }
  385. randIndex--
  386. }
  387. // if we have selected the address before, restart the loop
  388. // otherwise, record it and continue
  389. if pickFromOldBucket {
  390. if addrsMap, ok := oldBucketToAddrsMap[oldIndex]; ok {
  391. if _, ok = addrsMap[selectedAddr.String()]; ok {
  392. continue ADDRS_LOOP
  393. }
  394. } else {
  395. oldBucketToAddrsMap[oldIndex] = make(map[string]struct{})
  396. }
  397. oldBucketToAddrsMap[oldIndex][selectedAddr.String()] = struct{}{}
  398. oldAddressesAdded++
  399. } else {
  400. if addrsMap, ok := newBucketToAddrsMap[newIndex]; ok {
  401. if _, ok = addrsMap[selectedAddr.String()]; ok {
  402. continue ADDRS_LOOP
  403. }
  404. } else {
  405. newBucketToAddrsMap[newIndex] = make(map[string]struct{})
  406. }
  407. newBucketToAddrsMap[newIndex][selectedAddr.String()] = struct{}{}
  408. newAddressesAdded++
  409. }
  410. selection[selectionIndex] = selectedAddr
  411. selectionIndex++
  412. }
  413. return selection
  414. }
  415. // ListOfKnownAddresses returns the new and old addresses.
  416. func (a *addrBook) ListOfKnownAddresses() []*knownAddress {
  417. a.mtx.Lock()
  418. defer a.mtx.Unlock()
  419. addrs := []*knownAddress{}
  420. for _, addr := range a.addrLookup {
  421. addrs = append(addrs, addr.copy())
  422. }
  423. return addrs
  424. }
  425. //------------------------------------------------
  426. // Size returns the number of addresses in the book.
  427. func (a *addrBook) Size() int {
  428. a.mtx.Lock()
  429. defer a.mtx.Unlock()
  430. return a.size()
  431. }
  432. func (a *addrBook) size() int {
  433. return a.nNew + a.nOld
  434. }
  435. //----------------------------------------------------------
  436. // Save persists the address book to disk.
  437. func (a *addrBook) Save() {
  438. a.saveToFile(a.filePath) // thread safe
  439. }
  440. func (a *addrBook) saveRoutine() {
  441. defer a.wg.Done()
  442. saveFileTicker := time.NewTicker(dumpAddressInterval)
  443. out:
  444. for {
  445. select {
  446. case <-saveFileTicker.C:
  447. a.saveToFile(a.filePath)
  448. case <-a.Quit():
  449. break out
  450. }
  451. }
  452. saveFileTicker.Stop()
  453. a.saveToFile(a.filePath)
  454. }
  455. //----------------------------------------------------------
  456. func (a *addrBook) getBucket(bucketType byte, bucketIdx int) map[string]*knownAddress {
  457. switch bucketType {
  458. case bucketTypeNew:
  459. return a.bucketsNew[bucketIdx]
  460. case bucketTypeOld:
  461. return a.bucketsOld[bucketIdx]
  462. default:
  463. cmn.PanicSanity("Should not happen")
  464. return nil
  465. }
  466. }
  467. // Adds ka to new bucket. Returns false if it couldn't do it cuz buckets full.
  468. // NOTE: currently it always returns true.
  469. func (a *addrBook) addToNewBucket(ka *knownAddress, bucketIdx int) {
  470. // Sanity check
  471. if ka.isOld() {
  472. a.Logger.Error("Failed Sanity Check! Cant add old address to new bucket", "ka", ka, "bucket", bucketIdx)
  473. return
  474. }
  475. addrStr := ka.Addr.String()
  476. bucket := a.getBucket(bucketTypeNew, bucketIdx)
  477. // Already exists?
  478. if _, ok := bucket[addrStr]; ok {
  479. return
  480. }
  481. // Enforce max addresses.
  482. if len(bucket) > newBucketSize {
  483. a.Logger.Info("new bucket is full, expiring new")
  484. a.expireNew(bucketIdx)
  485. }
  486. // Add to bucket.
  487. bucket[addrStr] = ka
  488. // increment nNew if the peer doesnt already exist in a bucket
  489. if ka.addBucketRef(bucketIdx) == 1 {
  490. a.nNew++
  491. }
  492. // Add it to addrLookup
  493. a.addrLookup[ka.ID()] = ka
  494. }
  495. // Adds ka to old bucket. Returns false if it couldn't do it cuz buckets full.
  496. func (a *addrBook) addToOldBucket(ka *knownAddress, bucketIdx int) bool {
  497. // Sanity check
  498. if ka.isNew() {
  499. a.Logger.Error(fmt.Sprintf("Cannot add new address to old bucket: %v", ka))
  500. return false
  501. }
  502. if len(ka.Buckets) != 0 {
  503. a.Logger.Error(fmt.Sprintf("Cannot add already old address to another old bucket: %v", ka))
  504. return false
  505. }
  506. addrStr := ka.Addr.String()
  507. bucket := a.getBucket(bucketTypeOld, bucketIdx)
  508. // Already exists?
  509. if _, ok := bucket[addrStr]; ok {
  510. return true
  511. }
  512. // Enforce max addresses.
  513. if len(bucket) > oldBucketSize {
  514. return false
  515. }
  516. // Add to bucket.
  517. bucket[addrStr] = ka
  518. if ka.addBucketRef(bucketIdx) == 1 {
  519. a.nOld++
  520. }
  521. // Ensure in addrLookup
  522. a.addrLookup[ka.ID()] = ka
  523. return true
  524. }
  525. func (a *addrBook) removeFromBucket(ka *knownAddress, bucketType byte, bucketIdx int) {
  526. if ka.BucketType != bucketType {
  527. a.Logger.Error(fmt.Sprintf("Bucket type mismatch: %v", ka))
  528. return
  529. }
  530. bucket := a.getBucket(bucketType, bucketIdx)
  531. delete(bucket, ka.Addr.String())
  532. if ka.removeBucketRef(bucketIdx) == 0 {
  533. if bucketType == bucketTypeNew {
  534. a.nNew--
  535. } else {
  536. a.nOld--
  537. }
  538. delete(a.addrLookup, ka.ID())
  539. }
  540. }
  541. func (a *addrBook) removeFromAllBuckets(ka *knownAddress) {
  542. for _, bucketIdx := range ka.Buckets {
  543. bucket := a.getBucket(ka.BucketType, bucketIdx)
  544. delete(bucket, ka.Addr.String())
  545. }
  546. ka.Buckets = nil
  547. if ka.BucketType == bucketTypeNew {
  548. a.nNew--
  549. } else {
  550. a.nOld--
  551. }
  552. delete(a.addrLookup, ka.ID())
  553. }
  554. //----------------------------------------------------------
  555. func (a *addrBook) pickOldest(bucketType byte, bucketIdx int) *knownAddress {
  556. bucket := a.getBucket(bucketType, bucketIdx)
  557. var oldest *knownAddress
  558. for _, ka := range bucket {
  559. if oldest == nil || ka.LastAttempt.Before(oldest.LastAttempt) {
  560. oldest = ka
  561. }
  562. }
  563. return oldest
  564. }
  565. // adds the address to a "new" bucket. if its already in one,
  566. // it only adds it probabilistically
  567. func (a *addrBook) addAddress(addr, src *p2p.NetAddress) error {
  568. if addr == nil || src == nil {
  569. return ErrAddrBookNilAddr{addr, src}
  570. }
  571. if a.routabilityStrict && !addr.Routable() {
  572. return ErrAddrBookNonRoutable{addr}
  573. }
  574. if !addr.Valid() {
  575. return ErrAddrBookInvalidAddr{addr}
  576. }
  577. if !addr.HasID() {
  578. return ErrAddrBookInvalidAddrNoID{addr}
  579. }
  580. // TODO: we should track ourAddrs by ID and by IP:PORT and refuse both.
  581. if _, ok := a.ourAddrs[addr.String()]; ok {
  582. return ErrAddrBookSelf{addr}
  583. }
  584. if _, ok := a.privateIDs[addr.ID]; ok {
  585. return ErrAddrBookPrivate{addr}
  586. }
  587. if _, ok := a.privateIDs[src.ID]; ok {
  588. return ErrAddrBookPrivateSrc{src}
  589. }
  590. ka := a.addrLookup[addr.ID]
  591. if ka != nil {
  592. // If its already old and the addr is the same, ignore it.
  593. if ka.isOld() && ka.Addr.Equals(addr) {
  594. return nil
  595. }
  596. // Already in max new buckets.
  597. if len(ka.Buckets) == maxNewBucketsPerAddress {
  598. return nil
  599. }
  600. // The more entries we have, the less likely we are to add more.
  601. factor := int32(2 * len(ka.Buckets))
  602. if a.rand.Int31n(factor) != 0 {
  603. return nil
  604. }
  605. } else {
  606. ka = newKnownAddress(addr, src)
  607. }
  608. bucket := a.calcNewBucket(addr, src)
  609. a.addToNewBucket(ka, bucket)
  610. return nil
  611. }
  612. // Make space in the new buckets by expiring the really bad entries.
  613. // If no bad entries are available we remove the oldest.
  614. func (a *addrBook) expireNew(bucketIdx int) {
  615. for addrStr, ka := range a.bucketsNew[bucketIdx] {
  616. // If an entry is bad, throw it away
  617. if ka.isBad() {
  618. a.Logger.Info(fmt.Sprintf("expiring bad address %v", addrStr))
  619. a.removeFromBucket(ka, bucketTypeNew, bucketIdx)
  620. return
  621. }
  622. }
  623. // If we haven't thrown out a bad entry, throw out the oldest entry
  624. oldest := a.pickOldest(bucketTypeNew, bucketIdx)
  625. a.removeFromBucket(oldest, bucketTypeNew, bucketIdx)
  626. }
  627. // Promotes an address from new to old. If the destination bucket is full,
  628. // demote the oldest one to a "new" bucket.
  629. // TODO: Demote more probabilistically?
  630. func (a *addrBook) moveToOld(ka *knownAddress) {
  631. // Sanity check
  632. if ka.isOld() {
  633. a.Logger.Error(fmt.Sprintf("Cannot promote address that is already old %v", ka))
  634. return
  635. }
  636. if len(ka.Buckets) == 0 {
  637. a.Logger.Error(fmt.Sprintf("Cannot promote address that isn't in any new buckets %v", ka))
  638. return
  639. }
  640. // Remove from all (new) buckets.
  641. a.removeFromAllBuckets(ka)
  642. // It's officially old now.
  643. ka.BucketType = bucketTypeOld
  644. // Try to add it to its oldBucket destination.
  645. oldBucketIdx := a.calcOldBucket(ka.Addr)
  646. added := a.addToOldBucket(ka, oldBucketIdx)
  647. if !added {
  648. // No room; move the oldest to a new bucket
  649. oldest := a.pickOldest(bucketTypeOld, oldBucketIdx)
  650. a.removeFromBucket(oldest, bucketTypeOld, oldBucketIdx)
  651. newBucketIdx := a.calcNewBucket(oldest.Addr, oldest.Src)
  652. a.addToNewBucket(oldest, newBucketIdx)
  653. // Finally, add our ka to old bucket again.
  654. added = a.addToOldBucket(ka, oldBucketIdx)
  655. if !added {
  656. a.Logger.Error(fmt.Sprintf("Could not re-add ka %v to oldBucketIdx %v", ka, oldBucketIdx))
  657. }
  658. }
  659. }
  660. //---------------------------------------------------------------------
  661. // calculate bucket placements
  662. // doublesha256( key + sourcegroup +
  663. // int64(doublesha256(key + group + sourcegroup))%bucket_per_group ) % num_new_buckets
  664. func (a *addrBook) calcNewBucket(addr, src *p2p.NetAddress) int {
  665. data1 := []byte{}
  666. data1 = append(data1, []byte(a.key)...)
  667. data1 = append(data1, []byte(a.groupKey(addr))...)
  668. data1 = append(data1, []byte(a.groupKey(src))...)
  669. hash1 := doubleSha256(data1)
  670. hash64 := binary.BigEndian.Uint64(hash1)
  671. hash64 %= newBucketsPerGroup
  672. var hashbuf [8]byte
  673. binary.BigEndian.PutUint64(hashbuf[:], hash64)
  674. data2 := []byte{}
  675. data2 = append(data2, []byte(a.key)...)
  676. data2 = append(data2, a.groupKey(src)...)
  677. data2 = append(data2, hashbuf[:]...)
  678. hash2 := doubleSha256(data2)
  679. return int(binary.BigEndian.Uint64(hash2) % newBucketCount)
  680. }
  681. // doublesha256( key + group +
  682. // int64(doublesha256(key + addr))%buckets_per_group ) % num_old_buckets
  683. func (a *addrBook) calcOldBucket(addr *p2p.NetAddress) int {
  684. data1 := []byte{}
  685. data1 = append(data1, []byte(a.key)...)
  686. data1 = append(data1, []byte(addr.String())...)
  687. hash1 := doubleSha256(data1)
  688. hash64 := binary.BigEndian.Uint64(hash1)
  689. hash64 %= oldBucketsPerGroup
  690. var hashbuf [8]byte
  691. binary.BigEndian.PutUint64(hashbuf[:], hash64)
  692. data2 := []byte{}
  693. data2 = append(data2, []byte(a.key)...)
  694. data2 = append(data2, a.groupKey(addr)...)
  695. data2 = append(data2, hashbuf[:]...)
  696. hash2 := doubleSha256(data2)
  697. return int(binary.BigEndian.Uint64(hash2) % oldBucketCount)
  698. }
  699. // Return a string representing the network group of this address.
  700. // This is the /16 for IPv4, the /32 (/36 for he.net) for IPv6, the string
  701. // "local" for a local address and the string "unroutable" for an unroutable
  702. // address.
  703. func (a *addrBook) groupKey(na *p2p.NetAddress) string {
  704. if a.routabilityStrict && na.Local() {
  705. return "local"
  706. }
  707. if a.routabilityStrict && !na.Routable() {
  708. return "unroutable"
  709. }
  710. if ipv4 := na.IP.To4(); ipv4 != nil {
  711. return (&net.IPNet{IP: na.IP, Mask: net.CIDRMask(16, 32)}).String()
  712. }
  713. if na.RFC6145() || na.RFC6052() {
  714. // last four bytes are the ip address
  715. ip := net.IP(na.IP[12:16])
  716. return (&net.IPNet{IP: ip, Mask: net.CIDRMask(16, 32)}).String()
  717. }
  718. if na.RFC3964() {
  719. ip := net.IP(na.IP[2:7])
  720. return (&net.IPNet{IP: ip, Mask: net.CIDRMask(16, 32)}).String()
  721. }
  722. if na.RFC4380() {
  723. // teredo tunnels have the last 4 bytes as the v4 address XOR
  724. // 0xff.
  725. ip := net.IP(make([]byte, 4))
  726. for i, byte := range na.IP[12:16] {
  727. ip[i] = byte ^ 0xff
  728. }
  729. return (&net.IPNet{IP: ip, Mask: net.CIDRMask(16, 32)}).String()
  730. }
  731. // OK, so now we know ourselves to be a IPv6 address.
  732. // bitcoind uses /32 for everything, except for Hurricane Electric's
  733. // (he.net) IP range, which it uses /36 for.
  734. bits := 32
  735. heNet := &net.IPNet{IP: net.ParseIP("2001:470::"),
  736. Mask: net.CIDRMask(32, 128)}
  737. if heNet.Contains(na.IP) {
  738. bits = 36
  739. }
  740. return (&net.IPNet{IP: na.IP, Mask: net.CIDRMask(bits, 128)}).String()
  741. }
  742. // doubleSha256 calculates sha256(sha256(b)) and returns the resulting bytes.
  743. func doubleSha256(b []byte) []byte {
  744. hasher := sha256.New()
  745. hasher.Write(b) // nolint: errcheck, gas
  746. sum := hasher.Sum(nil)
  747. hasher.Reset()
  748. hasher.Write(sum) // nolint: errcheck, gas
  749. return hasher.Sum(nil)
  750. }