Commit Graph

281 Commits

Author SHA1 Message Date
Alexander Block
3f7553a711
Disable new connection handling and concentrated recovery for large LLMQs (#3548)
* Rename CollectSigSharesToSend to CollectSigSharesToSendConcentrated

* Allow to disable concentrated recovery and new connection handling for large LLMQs
2020-06-25 09:47:24 +02:00
UdjinM6
955e3868fb
Fix CSigningManager::VerifyRecoveredSig (#3546) 2020-06-24 10:04:10 +02:00
Alexander Block
de099fe7a3
Use exponential backoff timeouts for recovery (#3535)
Starting with 2 seconds.
2020-06-24 10:02:57 +02:00
PastaPastaPasta
3f8a7b2068
Merge #12193: RPC: Consistently use UniValue.pushKV instead of push_back(Pair()) (karel-3d) (#3532)
* Begin Merge 12193 Squashed 'src/univalue/' changes from 07947ff2da..51d3ab34ba

51d3ab34ba Merge #10: Add pushKV(key, boolean) function (replaces #5)
129bad96d5 [tests] test pushKV for boolean values
b3c44c947f Pushing boolean value to univalue correctly

git-subtree-dir: src/univalue
git-subtree-split: 51d3ab34ba2857f0d03dc07250cb4a2b5e712e67

* scripted-diff: Use UniValue.pushKV instead of push_back(Pair()) (end #12193)

-BEGIN VERIFY SCRIPT-
git grep -l "push_back(Pair" | xargs sed -i "s/push_back(Pair(\(.*\)));/pushKV(\1);/g"
-END VERIFY SCRIPT-

Signed-off-by: pasta <pasta@dashboost.org>

Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-06-18 12:17:23 +03:00
Alexander Block
737ed479bc
Don'd send SENDXXX messages to fMasternode connections (#3537)
And respect a received QSENDRECSIGS no matter if fMasternode or not. We
assume that fMasternode peers won't send QSENDRECSIGS from now on.
2020-06-18 11:41:18 +03:00
Alexander Block
0fa058deb5
Only relay DKG messages to intra quorum connection members (#3536)
This reuses the old intra-quorum connections algorithm.
2020-06-18 11:40:33 +03:00
Pasta
ec7c7dfc8d
fix lint failure
Signed-off-by: Pasta <pasta@dashboost.org>
2020-06-12 01:17:37 -05:00
pasta
518ccbf986
scripted-diff: Convert 11 enums into scoped enums (C++11) (merge #10742)
-BEGIN VERIFY SCRIPT-

sed -i 's/enum DBErrors/enum class DBErrors/g' src/wallet/walletdb.h
git grep -l DB_ | xargs sed -i 's/DB_\(LOAD_OK\|CORRUPT\|NONCRITICAL_ERROR\|TOO_NEW\|LOAD_FAIL\|NEED_REWRITE\)/DBErrors::\1/g'
sed -i 's/^    DBErrors::/    /g' src/wallet/walletdb.h

sed -i 's/enum VerifyResult/enum class VerifyResult/g' src/wallet/db.h
sed -i 's/\(VERIFY_OK\|RECOVER_OK\|RECOVER_FAIL\)/VerifyResult::\1/g' src/wallet/db.cpp

sed -i 's/enum ThresholdState/enum class ThresholdState/g' src/versionbits.h
git grep -l THRESHOLD_ | xargs sed -i 's/THRESHOLD_\(DEFINED\|STARTED\|LOCKED_IN\|ACTIVE\|FAILED\)/ThresholdState::\1/g'
sed -i 's/^    ThresholdState::/    /g' src/versionbits.h

sed -i 's/enum SigVersion/enum class SigVersion/g' src/script/interpreter.h
git grep -l SIGVERSION_ | xargs sed -i 's/SIGVERSION_\(BASE\|WITNESS_V0\)/SigVersion::\1/g'
sed -i 's/^    SigVersion::/    /g' src/script/interpreter.h

sed -i 's/enum RetFormat {/enum class RetFormat {/g' src/rest.cpp
sed -i 's/RF_\(UNDEF\|BINARY\|HEX\|JSON\)/RetFormat::\1/g' src/rest.cpp
sed -i 's/^    RetFormat::/    /g' src/rest.cpp

sed -i 's/enum HelpMessageMode {/enum class HelpMessageMode {/g' src/init.h
git grep -l HMM_ | xargs sed -i 's/HMM_BITCOIN/HelpMessageMode::BITCOIN/g'
sed -i 's/^    HelpMessageMode::/    /g' src/init.h

sed -i 's/enum FeeEstimateHorizon/enum class FeeEstimateHorizon/g' src/policy/fees.h

sed -i 's/enum BlockSource {/enum class BlockSource {/g' src/qt/clientmodel.h
git grep -l BLOCK_SOURCE_ | xargs sed -i 's/BLOCK_SOURCE_\(NONE\|REINDEX\|DISK\|NETWORK\)/BlockSource::\1/g'
sed -i 's/^    BlockSource::/    /g' src/qt/clientmodel.h

sed -i 's/enum FlushStateMode {/enum class FlushStateMode {/g' src/validation.cpp
sed -i 's/FLUSH_STATE_\(NONE\|IF_NEEDED\|PERIODIC\|ALWAYS\)/FlushStateMode::\1/g' src/validation.cpp
sed -i 's/^    FlushStateMode::/    /g' src/validation.cpp

-END VERIFY SCRIPT-

Signed-off-by: pasta <pasta@dashboost.org>
2020-06-11 23:20:48 -05:00
UdjinM6
89851d3361
Fix chainlock scheduler handling (#3514)
Reindexing corrupted chain always fails otherwise.
2020-06-11 11:39:23 +03:00
UdjinM6
232430fac9
Fix ProcessNewBlock vs EnforceBestChainLock deadlocks in ActivateBestChain (#3492)
* Drop dead code in DoInvalidateBlock

* Let ActivateBestChain skip SyncWithValidationInterfaceQueue when called from IS or CL threads

* Use CL's own scheduler instead of a global one

* Revert "Let ActivateBestChain skip SyncWithValidationInterfaceQueue when called from IS or CL threads"

This reverts commit 1c9f6da50a.
2020-06-09 06:53:16 +03:00
UdjinM6
ae5faf23da
Better error handling while processing special txes (#3504)
* Return invalid state for GetTxPayload failures instead of asserting

* Throw an error when internal `find` fails instead of asserting

* Unify mnlist checks for lists created from blocks and lists created via updates of older lists with diffs

* Wrap evo block processing in try..catch to make sure all earlier introduced throws are caught

* evo/llmq: Adjust reject reasons for CFinalCommitmentTxPayload

In CheckLLMQCommitment this reject string was used so i adjusted the
other one in CQuorumBlockProcessor::GetCommitmentsFromBlock and the new
introduced in PR #3506

* return state.DoS instead of asserting in cbtx.cpp

Signed-off-by: pasta <pasta@dashboost.org>

Co-authored-by: xdustinface <xdustinfacex@gmail.com>
Co-authored-by: pasta <pasta@dashboost.org>
2020-06-08 05:57:57 +03:00
thephez
970c23d6e6
Remove logic for forcing chainlocks without DIP8 activation on testnet (#3470) 2020-05-11 15:35:55 +03:00
dustinface
ae15299117
Feature request #3425 - Add information to "protx diff" (#3468)
* llmq: Add more information to the result of CFinalCommitment::ToJson

Example output: 

```
{
      "version": 1,
      "llmqType": 1,
      "quorumHash": 
"000002a20a35757fc467f0ebe8663f76e4c5df74411cc5f1735cfd92278071fe",
      "signersCount": 49,
      "signers": "03fbffffffffff",
      "validMembersCount": 49,
      "validMembers": "03fbffffffffff",
      "quorumPublicKey": 
"8db1108b50dd741076b577f6798f0bf382ca4d11dafd55ed14aaaf613f604aa55bf1da1f0469f75105376e440381ee23",
      "quorumVvecHash": 
"feedb15d5a7ca25a68d9c51036973d37925d8439d4769d7dd7bace7440089071",
      "quorumSig": 
"1879f6260ef427dc9542088b39e02a8e73e34c0484269cbc12108ab991d438f81e5be80a5b4b5595b6634113fdf060060c8fe190677627fb2d25f6e052c7f93598d36dd415f8f84c693c6a5943098616a1e135522852d8f572bb6e1470df4a94",
      "membersSig": 
"8ef7d881fcc1e8111dd82c556ab62dad56ea5ecb47a9ac067eb2057d92d27e39530630a7f77d19adc857d33eb485a1b601b2d78ac4c73742f2036089bfb1f25441f1dfcd3cfd1eed6e7d664c0e97bf75b7d7097ebd647791fbcef070fe3e3e91"
    }
```

* llmq: Changed the byte order in ToHexStr(const std::vector<bool>& vBits)
2020-05-11 15:35:28 +03:00
10xcryptodev
017c4779ca
Fix recovery from coin db crashes (and dbcrash.py test) (#3467)
* 🪲 improve evodb consistency recovering from dbcrash

* Adjust the fix

Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>

* Fix it

Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>

* Disable recovery from a crash during a fork and a corresponding part of dbcrash.py

Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>

* Skip some checks in CQuorumBlockProcessor when replaying blocks after the crash

Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>

* Process special txes in RollforwardBlock

Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>

* Update src/init.cpp

Co-authored-by: PastaPastaPasta <6443210+PastaPastaPasta@users.noreply.github.com>

Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
Co-authored-by: PastaPastaPasta <6443210+PastaPastaPasta@users.noreply.github.com>
2020-05-11 15:33:21 +03:00
Alexander Block
608aed3d85
Don't try to connect to itself through CLLMQUtils::GetQuorumConnections (#3448) 2020-04-21 16:43:59 +02:00
Alexander Block
96ed9fae39
Merge pull request #3429 from codablock/pr_remove_optimistic_send
Fix flushing of rejects before disconnecting
2020-04-17 21:46:36 +02:00
Alexander Block
d8bbdee7c4 Remove support for optimistic send
This was only used in only one remaining place and only to ensure that
reject messages are sent before closing sockets. This is solved by the
previous commit now.
2020-04-17 11:32:18 +02:00
Alexander Block
6ca78b601e Introduce NETCONN logging category and NET_NETCONN to log in NET and NETCONN
This allows easier debugging of connection issues, as it produces much less
chatter in the logs when only -debug=mnconn is given.
2020-04-17 08:52:57 +02:00
Alexander Block
72afaddf72 Introduce new way of deterministic selection of outbound LLMQ connections
The new way does not require the quorum hash to be included in the
calculation, which allows us to have determinism globally and not only
for single LLMQs.
2020-04-16 13:59:02 +02:00
Alexander Block
26dec64e79
Merge pull request #3397 from codablock/pr_backport_txrequests
Backport bitcoin#14897 and bitcoin#15834 and modify it to work with Dash messages
2020-04-09 00:12:44 +02:00
Alexander Block
6b32192bdc Fix compilation 2020-04-08 14:55:27 +02:00
Alexander Block
26fcd3f0bf Also remove m_tx_announced and m_tx_in_flight entries when EraseObjectRequest is called
Otherwise they'll run into false-positive timeouts.
2020-04-07 13:25:29 +02:00
Alexander Block
4bfc20cb67 Force re-requesting of IS locked TXs 2020-04-07 13:23:29 +02:00
Alexander Block
f142fff881 Skip verification of recovered sigs that were reconstructed in InstantSend
Also don't request them via getdata
2020-04-07 07:30:00 +02:00
Alexander Block
454fae3bda Only process 32 IS locks at a time in ProcessPendingInstantSendLocks 2020-04-07 07:29:51 +02:00
Alexander Block
d856fd1407 Use salted hasher for pendingInstantSendLocks 2020-04-07 07:29:32 +02:00
Alexander Block
465af48e83 Improve logging in LLMQ sig handling 2020-04-07 07:29:23 +02:00
Alexander Block
2443c8f582 Replace uses of CNode::AskFor with RequestObject 2020-04-07 07:14:53 +02:00
Alexander Block
004d9224c4 Replace uses of CConnman::RemoveAskFor with EraseObjectRequest 2020-04-07 07:14:42 +02:00
Alexander Block
c7b6eb851d
Merge pull request #3389 from codablock/pr_concentrated_recovery
Implement "concentrated recovery" of LLMQ signatures
2020-04-02 13:59:05 +02:00
Alexander Block
148bbdd5cf
Use GetTime instead of GetAdjustedTime
Co-Authored-By: UdjinM6 <UdjinM6@users.noreply.github.com>
2020-04-01 18:05:44 +02:00
Alexander Block
76b6614fe5 Use auto keyword in CollectSigSharesToSend 2020-03-31 19:58:07 +02:00
Alexander Block
900ee0f194 Use range based for loop in SelectMemberForRecovery 2020-03-31 19:58:07 +02:00
Alexander Block
b212f21c15 Implement new way of concentrated signature recovery
Instead of propagating all sig shares to all LLMQ members, this will now
make all members send their individual sig share to a single member, which
is then responsible for the recovery and propagation of the recovered
signature. This process is repeated by all members every second for another
target/recovering member, until a recovered signature appears.
2020-03-31 19:58:07 +02:00
Alexander Block
45064d8dc9 Rename sigSharesToSend to sigShareBatchesToSend 2020-03-31 19:58:07 +02:00
Alexander Block
88d4f551c9 Move spork21 condition into VerifyConnectionAndMinProtoVersions 2020-03-31 07:11:27 +02:00
Alexander Block
32c83b432b Verify min proto version and open ports of LLMQ members and vote on bad ones 2020-03-31 07:07:52 +02:00
Alexander Block
b0668028b6
Implement more randomized behavior in GetQuorumConnections (#3385)
When taking the proTxHash naively, we might end up with a few unlucky MNs
which always have to perform most of the outbound connections while other
unlucky MNs would always have to wait for inbound connections. Hashing
the proTxHash with the quorum hash makes this more random.
2020-03-28 00:59:32 +03:00
Alexander Block
a09e36106e Fix onlyOutbound handling 2020-03-27 15:12:20 +01:00
Alexander Block
f82204db95 Move intra-quorum connection calculation into local func 2020-03-27 15:12:20 +01:00
Alexander Block
dbaf13848d Include inbound connections in output of "quorum dkgstatus" 2020-03-27 15:12:20 +01:00
Alexander Block
8f644d18d7 Implement probing of public ip/port of LLMQ members 2020-03-27 15:12:20 +01:00
Alexander Block
14bb62ac8e Connect all LLMQ members to all other members 2020-03-27 15:12:20 +01:00
Alexander Block
c68b5f68aa
Hold CEvoDB lock while iterating mined commitments (#3379) 2020-03-27 15:11:42 +01:00
Alexander Block
17ece14f40
Better/more logging for DKGs (#3381)
* Better/more logging for DKGs

* Use type name instead of type number in CDKGLogger
2020-03-26 13:25:01 +01:00
Alexander Block
8211ff0a8a
Merge pull request #3368 from codablock/pr_fmasternode
Don't relay anything to inbound/outbound fMasternode connections
2020-03-24 17:40:51 +01:00
Alexander Block
7f1f1d12f5 Make EnsureQuorumConnections re-set connections in every iteration
Instead of only doing it the first time. This is a preparation for
the new intra-quorum connection system (which connects all members to all
other members)
2020-03-24 10:58:29 +01:00
Alexander Block
9ef1e7cb51 Only log new quorum connections when it's actually new 2020-03-24 10:58:16 +01:00
Alexander Block
364d6c37f7 Move and unify logic for quorum connection establishment into CLLMQUtils 2020-03-24 10:58:16 +01:00
Alexander Block
f7ddee13a1
Fix possible segfault (#3365)
When CheckActiveState is called very early, there might be no chain yet.
2020-03-21 13:31:09 +01:00