aboutsummaryrefslogtreecommitdiffstats
path: root/eth/downloader
Commit message (Collapse)AuthorAgeFilesLines
* eth/downloader: fix occasional fast sync critical section test failsPéter Szilágyi2016-06-131-1/+3
|
* eth/downloader: adaptive quality of service tuningPéter Szilágyi2016-06-063-48/+235
|
* eth/downloader: make fast sync resilient to critical section failsPéter Szilágyi2016-06-022-26/+109
|
* Merge pull request #2637 from karalabe/downloader-always-cancelPéter Szilágyi2016-05-312-1/+13
|\ | | | | eth/downloader: ensure cancel channel is closed post sync
| * eth/downloader: ensure cancel channel is closed post syncPéter Szilágyi2016-05-302-1/+13
| |
* | eth/downloader, trie: pull head state concurrently with chainPéter Szilágyi2016-05-272-29/+35
|/
* eth/downloader: fix reviewer commentsPéter Szilágyi2016-05-202-9/+32
|
* eth/downloader: stream partial skeleton filling to processorPéter Szilágyi2016-05-173-19/+55
|
* eth/downloader: implement concurrent header downloadsPéter Szilágyi2016-05-174-211/+584
|
* eth/downloader: bound fork ancestry and allow heavy short forksPéter Szilágyi2016-05-132-43/+201
|
* rpc: remove NotifierContextKeyFelix Lange2016-04-161-1/+1
| | | | | | | Context keys must have a unique type in order to prevent any unintented clashes. The code used int(1) as key. Fix it by implementing the pattern recommended by package context.
* all: fix go vet warningsFelix Lange2016-04-152-5/+3
|
* all: update license informationFelix Lange2016-04-151-7/+7
|
* rpc: add pub/sub supportBas van Kervel2016-04-021-16/+59
|
* core, eth: replace reorganiz with reorganisLeif Jurvetson2016-03-161-2/+2
|
* eth: various typosLeif Jurvetson2016-03-164-10/+10
|
* eth/downloader: fix a throughput estimation data racePéter Szilágyi2016-03-101-2/+2
|
* eth/downloader: fix premature exit before notifying all part fetchersPéter Szilágyi2016-02-291-1/+3
|
* eth/downloader: fix header download limitingFelix Lange2016-02-261-1/+1
| | | | Fixes #2201
* Merge pull request #2251 from karalabe/increase-downlaoder-queuePéter Szilágyi2016-02-242-1/+6
|\ | | | | eth/downloader: bump the download queue size to prevent starvation
| * eth/downloader: bump the download queue size to prevent starvationPéter Szilágyi2016-02-232-1/+6
| |
* | Merge pull request #2242 from jimenezrick/upstream-cryptoJeffrey Wilcke2016-02-241-1/+1
|\ \ | |/ |/| Closes #2241: Use Keccak-256 from golang.org/x/crypto/sha3 and mention explicitly
| * all: Rename crypto.Sha3{,Hash}() to crypto.Keccak256{,Hash}()Ricardo Catalinas Jiménez2016-02-221-1/+1
| | | | | | | | As we aren't really using the standarized SHA-3
* | eth/downloader: fix partial rollback and ancestor lookupPéter Szilágyi2016-02-191-2/+28
|/
* eth, eth/downloader, jsre: surface state sync progress through the APIPéter Szilágyi2016-02-103-22/+31
|
* eth/downloader: raise pending state limit that prevented concurrencyPéter Szilágyi2016-02-082-3/+3
|
* rpc: migrated the RPC insterface to a new reflection based RPC layerBas van Kervel2016-01-261-1/+1
|
* core, eth/downloader: ensure state presence in ancestor lookupPéter Szilágyi2016-01-043-57/+68
|
* eth/downloader: throttling tests are time-sensitive, don't run parallelPéter Szilágyi2015-12-301-3/+1
|
* eth/downloader: fix negative balance issue in testsFelix Lange2015-12-181-2/+5
| | | | | | | | | | The test chain generated by makeChainFork included invalid uncle headers, crashing the generator during the state commit. The headers were invalid because they used the iteration counter as the block number, even though makeChainFork uses a block with number > 0 as the parent. Fix this by introducing BlockGen.Number, which allows accessing the actual number of the block being generated.
* rpc: new RPC implementation with pub/sub supportBas van Kervel2015-12-141-0/+64
|
* eth/downloader: fetch data proportionally to peer capacityPéter Szilágyi2015-11-193-246/+258
|
* eth/downloader: run tests in parallelFelix Lange2015-11-191-0/+29
|
* eth/downloader: also drain stateCh, receiptCh in eth/61 modeFelix Lange2015-11-191-71/+77
| | | | | | | | | State and receipt deliveries from a previous eth/62+ sync can hang if the downloader has moved on to syncing with eth/61. Fix this by also draining the eth/63 channels while waiting for eth/61 data. A nicer solution would be to take care of the channels in a central place, but that would involve a major rewrite.
* eth/downloader: don't hang for spurious deliveriesFelix Lange2015-11-193-236/+230
| | | | | | | | | | | | | | | | Unexpected deliveries could block indefinitely if they arrived at the right time. The fix is to ensure that the cancellation channel is always closed when the sync ends, unblocking any deliveries. Also remove the atomic check for whether a sync is currently running because it doesn't help and can be misleading. Cancelling always seems to break the tests though. The downloader spawned d.process whenever new data arrived, making it somewhat hard to track when block processing was actually done. Fix this by running d.process in a dedicated goroutine that is tied to the lifecycle of the sync. d.process gets notified of new work by the queue instead of being invoked all the time. This removes a ton of weird workaround code, including a hairy use of atomic CAS.
* eth/downloader: fix dysfunctional ignore list hidden by generic setPéter Szilágyi2015-11-042-11/+46
|
* eth: don't block sync goroutines that short circuitPéter Szilágyi2015-10-281-2/+4
|
* core, eth, trie: fix data races and merge/review issuesPéter Szilágyi2015-10-215-347/+518
|
* eth: clean out light node notions from ethPéter Szilágyi2015-10-191-4/+6
|
* core, eth: roll back uncertain headers in failed fast syncsPéter Szilágyi2015-10-193-11/+130
|
* eth/downloader: concurrent receipt and state processingPéter Szilágyi2015-10-194-72/+158
|
* core, eth, trie: direct state trie synchronizationPéter Szilágyi2015-10-196-451/+760
|
* core, eth: receipt chain reconstructionPéter Szilágyi2015-10-192-114/+166
|
* eth/downloader: add fast and light sync strategiesPéter Szilágyi2015-10-196-775/+1388
|
* eth/downloader: match capabilities when querying idle peersPéter Szilágyi2015-10-023-8/+53
|
* eth/downloader: always send termination wakes, clean leftoverPéter Szilágyi2015-09-231-20/+36
|
* eth/downloader: header-chain order and ancestry checkPéter Szilágyi2015-09-153-7/+77
|
* eth, rpc: standardize the chain sync progress countersPéter Szilágyi2015-09-152-51/+476
|
* core: split out TD from database and all internalsPéter Szilágyi2015-09-112-16/+39
|
* eth/downloader: fix race causing occasional test failurePéter Szilágyi2015-09-011-1/+3
|
* eth: update metrics collection to handle eth/62 algosPéter Szilágyi2015-08-253-4/+97
|
* eth: port the synchronisation algo to eth/62Péter Szilágyi2015-08-254-217/+1125
|
* eth: kill off protocol eth/60 in preparation for eth/62Péter Szilágyi2015-08-242-772/+5
|
* eth, eth/downloader: handle header requests, table driven proto testsPéter Szilágyi2015-08-241-5/+7
|
* eth, metrics, p2p: prepare metrics and net packets to eth/62Péter Szilágyi2015-08-211-1/+2
|
* eth, eth/downloader: don't report stall if fetcher filled the blockPéter Szilágyi2015-07-292-42/+64
|
* all: fix license headers one more timeFelix Lange2015-07-245-5/+5
| | | | I forgot to update one instance of "go-ethereum" in commit 3f047be5a.
* all: update license headers to distiguish GPL/LGPLFelix Lange2015-07-235-20/+20
| | | | | All code outside of cmd/ is licensed as LGPL. The headers now reflect this by calling the whole work "the go-ethereum library".
* eth/downloader: drop peer if advertised TD but won't delvierPéter Szilágyi2015-07-092-99/+44
|
* all: update license informationFelix Lange2015-07-075-0/+80
|
* eth/downloader: fix a rare test race on the OSX CIPéter Szilágyi2015-07-031-1/+7
|
* eth, eth/downloader: fix #1231, DOS vulnerability in hash queueingPéter Szilágyi2015-07-011-4/+14
|
* eth/downloader: reduce hash fetches in prep for eth/61Péter Szilágyi2015-07-011-3/+3
|
* eth, eth/downloader: transition to eth 61Péter Szilágyi2015-07-014-81/+606
|
* eth, eth/downloader: pass the eth protocol version throughPéter Szilágyi2015-07-013-35/+44
|
* eth/downloader, eth/fetcher: use core.GenerateChain in testsFelix Lange2015-06-301-156/+97
| | | | | | TestMadeupParentBlockChainAttack has been deleted because it was too hard to port and the attack that it checks the prevention of is being averted in a different way (through a protocol change).
* core/types: make blocks immutableFelix Lange2015-06-301-1/+4
|
* eth, eth/fetcher: move propagated block import into fetcherPéter Szilágyi2015-06-181-1/+1
|
* eth/fetcher: fix timer reset bug, add initial testsPéter Szilágyi2015-06-181-3/+5
|
* eth, eth/fetcher: separate notification sync mechanismPéter Szilágyi2015-06-181-0/+1
|
* eth/downloader: extend slow test to fix even slower CI server...Péter Szilágyi2015-06-181-1/+1
|
* eth/downloader: always reenter processing if not exitingPéter Szilágyi2015-06-182-39/+37
|
* eth, eth/downloader: fix processing interrupt caused by temp cancelPéter Szilágyi2015-06-182-23/+21
|
* eth/downloader: fix #1280, overlapping (good/bad) delivery hangPéter Szilágyi2015-06-172-0/+35
|
* eth/downloader: detailed comment for the race corner casePéter Szilágyi2015-06-151-4/+15
|
* eth/downloader: fix import statistic reset, fetch hashes asyncPéter Szilágyi2015-06-151-11/+14
|
* eth, eth/downloader: do async block fetches, add dl testsPéter Szilágyi2015-06-152-4/+46
|
* eth/downloader: sanity test for multi peer syncsPéter Szilágyi2015-06-151-0/+31
|
* eth, eth/downloader: move block processing into the downlaoderPéter Szilágyi2015-06-152-180/+249
|
* eth/downloader: add valid peer during attacks (check interference)Péter Szilágyi2015-06-151-93/+140
|
* eth/downloader: instreument and test the sync peer dropPéter Szilágyi2015-06-152-9/+66
|
* eth/downloader: support individual peers in the test suitePéter Szilágyi2015-06-151-112/+113
|
* eth/downloader: remove uneeded testing functionsPéter Szilágyi2015-06-152-37/+2
|
* eth, eth/downloader: move peer removal into downloaderPéter Szilágyi2015-06-153-45/+76
|
* eth/downloader: fetch the block hashes on the fly, when neededPéter Szilágyi2015-06-111-9/+5
|
* cmd/geth, eth/downloader: rough guess at the import etaPéter Szilágyi2015-06-101-2/+13
|
* cmd/geth, eth/downloader: collect and report import progress tooPéter Szilágyi2015-06-101-4/+35
|
* eth/downloader: fix hash fetch timeout handlingFelix Lange2015-06-091-12/+18
| | | | Fixes #1206
* eth/downloader: test registration rejection on head banPéter Szilágyi2015-06-081-2/+10
|
* eth/downloader: reject peer registration if head is bannedPéter Szilágyi2015-06-081-4/+15
|
* eth/downloader: cap the hash ban set, add test for itPéter Szilágyi2015-06-084-9/+61
|
* eth/downloader: fix throttling test to be less timing dependentPéter Szilágyi2015-06-081-3/+7
|
* eth/downloader: fix another rebase errorPéter Szilágyi2015-06-081-2/+1
|
* eth/downloader: preallocate the block cachePéter Szilágyi2015-06-083-21/+14
|
* eth/downloader: fix merge compile errorPéter Szilágyi2015-06-081-1/+1
|
* eth/downloader: don't block hash deliveries while pulling blocksPéter Szilágyi2015-06-081-17/+14
|
* eth/downloader: update test for shitty travisPéter Szilágyi2015-06-081-1/+1
|
* eth/downloader: clean up and simplify the code a bitPéter Szilágyi2015-06-081-13/+3
|
* eth/downloader: short circuit sync if head hash is bannedPéter Szilágyi2015-06-081-0/+4
|
* eth/downloader: accumulating hash bans for reconnecting attackersPéter Szilágyi2015-06-082-5/+126
|
* eth/downloader: clean up tests and unused variablesPéter Szilágyi2015-06-082-120/+79
|
* Merge pull request #1189 from karalabe/downloader-polishesJeffrey Wilcke2015-06-053-19/+50
|\ | | | | eth/downloader: handle timeouts more gracefully
| * eth/downloader: differentiate stale and nonexistent deliveriesPéter Szilágyi2015-06-052-19/+46
| |
| * eth/downloader: log hard timeouts and reset capacityPéter Szilágyi2015-06-052-0/+4
| |
* | eth, eth/downloader: fix #1098, elevate empty hash errors to peer dropsPéter Szilágyi2015-06-051-2/+2
|/
* eth/downloader: drop log entry from peer, it's covered alreadyPéter Szilágyi2015-06-041-5/+0
|
* eth/downloader: log after state updates, easier to debugPéter Szilágyi2015-06-041-5/+3
|
* eth/downloader: fix #1178, don't request blocks beyond the cache boundsPéter Szilágyi2015-06-043-15/+35
|
* eth/downloader: demote peers if they exceed the soft limits at 1 blocks alreadyPéter Szilágyi2015-06-031-3/+8
|
* eth/downloader: add a basic block download congestion controlPéter Szilágyi2015-06-035-55/+67
|
* core, eth/downloader: expose the bad hashes, check in downloaderPéter Szilágyi2015-05-281-0/+16
|
* eth/downloader: silence "Added N blocks from..." if N == 0Péter Szilágyi2015-05-261-1/+1
|
* eth, eth/downloader: surface downloaded block origin, drop on errorPéter Szilágyi2015-05-263-14/+22
|
* Merge pull request #1064 from karalabe/downloader-attacksJeffrey Wilcke2015-05-223-21/+93
|\ | | | | Fix two additional download vulnerabilities
| * eth, eth/downloader: remove duplicate consts, bump hash fetch to 2KPéter Szilágyi2015-05-213-6/+8
| |
| * eth/downloader: circumvent a forged block chain with known parent attackPéter Szilágyi2015-05-212-13/+56
| |
| * eth/downloader: prevent a peer from dripping bad hashesPéter Szilágyi2015-05-212-9/+36
| |
* | eth/downloader: fix test to it doesn't time out on a slow machinePéter Szilágyi2015-05-201-1/+1
|/
* eth/downloader: moved start eventobscuren2015-05-161-2/+2
| | | | Start event has moved because it could possibly could stall the miner
* Merge pull request #996 from karalabe/fix-potential-crosscheck-raceJeffrey Wilcke2015-05-161-5/+7
|\ | | | | eth/downloader: circumvent download race between crosscheck and hashes
| * eth/downloader: circumvent download race between crosscheck and hashesPéter Szilágyi2015-05-161-5/+7
| |
* | eth/downloader: fix #992, where tests may time out on a slow machinePéter Szilágyi2015-05-161-0/+6
|/
* eth/downloader: fix cancel channel double closePéter Szilágyi2015-05-161-3/+8
|
* eth/downloader: circumvent a fake blockchain attackPéter Szilágyi2015-05-152-27/+58
|
* eth/downloader: don't penalize for stale cross checksPéter Szilágyi2015-05-151-7/+17
|
* eth, eth/downloader: detect and handle madeup hash attacksPéter Szilágyi2015-05-153-42/+109
|
* eth/downloader: circumvent hash reordering attacksPéter Szilágyi2015-05-153-49/+88
|
* eth/downloader: check sync after failed attacksPéter Szilágyi2015-05-152-7/+23
|
* eth/downloader: add a test for detecting missing blocksPéter Szilágyi2015-05-152-8/+24
|
* eth, eth/downloader: prevent hash repeater attackPéter Szilágyi2015-05-153-13/+46
|
* eth/downloader: fixed testobscuren2015-05-151-1/+3
|
* eth, eth/downloader, miner: use download events to check miner startobscuren2015-05-152-1/+21
|
* Merge pull request #974 from karalabe/downloader-fix-unknown-parent-attackJeffrey Wilcke2015-05-143-34/+82
|\ | | | | eth, eth/downloader: handle a potential unknown parent attack
| * eth/downloader: add a small additional check to the testPéter Szilágyi2015-05-141-0/+3
| |
| * eth/downloader: remove a previous leftoverPéter Szilágyi2015-05-141-1/+0
| |
| * eth, eth/downloader: remove parent verification from the downlaoderPéter Szilágyi2015-05-143-37/+13
| |
| * eth, eth/downloader: handle a potential unknown parent attackPéter Szilágyi2015-05-142-25/+95
| |
* | Merge pull request #950 from karalabe/downloader-cancel-syncJeffrey Wilcke2015-05-141-32/+37
|\ \ | | | | | | eth/downloader: fix deliveries to check for sync cancels
| * | eth/downloader: remove a redundant sync progress checkPéter Szilágyi2015-05-131-4/+0
| | |
| * | eth/downloader: add a user sync notificatonPéter Szilágyi2015-05-131-0/+5
| | |
| * | eth/downloader: sync the cancel channel during cancel tooPéter Szilágyi2015-05-131-21/+4
| | |
| * | eth/downloader: fix deliveries to check for sync cancelsPéter Szilágyi2015-05-131-10/+31
| |/
* / eth/downloader: fix nil panic caused by wrong variable usePéter Szilágyi2015-05-131-1/+1
|/
* eth/downloader: fix active peer shadowing, polish func namesPéter Szilágyi2015-05-132-22/+14
|
* eth, eth/downloader: moved pending queue error message to debugobscuren2015-05-111-2/+2
|
* eth/downloader: use count instead of peers, clearerPéter Szilágyi2015-05-112-4/+4
|
* eth/downloader: revert to demotion, use harsher penaltyPéter Szilágyi2015-05-112-8/+19
|
* eth/downloader: fix #910, thread safe peers & polishesPéter Szilágyi2015-05-114-161/+210
|
* eth: added a cancel method for the downloaderobscuren2015-05-102-2/+103
| | | | | | | | | Added a cancel method to the downloader which gracefully shuts down any active syncing process (hash fetching or block downloading) and resets the queue and remove any pending blocks. Issue with the downloader which would stall because of an active ongoing process when an invalid block was found.
* eth/downloader: fix leftover state between syncsPéter Szilágyi2015-05-081-1/+4
|
* eth, eth/downloader: handle sync errors a bit more gracefullyPéter Szilágyi2015-05-082-18/+16
|
* eth, eth/downloader: make synchronize thread safePéter Szilágyi2015-05-083-68/+16
|
* eth/downloader: fix priority queue reset, add throttling testPéter Szilágyi2015-05-071-0/+48
|
* eth/downloader: fix expiration not running while fetchingPéter Szilágyi2015-05-072-34/+32
|
* eth/downloader: prioritize block fetch based on chain position, cap memory usePéter Szilágyi2015-05-075-273/+369
|
* downloader: hash downloading recoveryobscuren2015-05-032-29/+71
| | | | | | If a peer fails to respond (disconnect, etc) during hash downloading switch to a different peer which has it's current_hash in the queue's peer set.
* eth/downloader: put back hashes on block overflow errorobscuren2015-05-031-19/+14
|
* eth/downloader: changed has to blockHashes for chain linkingobscuren2015-05-031-1/+1
|
* eth/downloader: added additional testsobscuren2015-05-032-4/+73
|
* eth, eth/downloader: Moved block processing & graceful shutdownobscuren2015-05-013-173/+221
| | | | | | | The downloader is no longer responsible for processing blocks. The eth-protocol handler now takes care of this instead. Added graceful shutdown during block processing. Closes #846
* eth/downloader: delete blocks from queueobscuren2015-04-301-2/+14
|
* eth, eth/downloader: remove bad peers from peer setobscuren2015-04-301-10/+8
| | | | Peers in the eth protocol handler are now being ignored for catch up.
* eth/downloader: drop blockobscuren2015-04-301-1/+1
|
* core: fixed tetst to reflect (int, error) return by insertChainobscuren2015-04-291-2/+2
|
* eth/downloader: ignore orphan blocks in the downloader.obscuren2015-04-291-5/+9
| | | | | | When blocks have been sorted and are being processed, orphan blocks should be ignored and thrown out. The protocol handler is responsible for downloading blocks which have missing parents.
* core: return the index of the block that failed when inserting a chainobscuren2015-04-291-5/+4
|
* eth/downloader: reset the chain when parent is missing during processingobscuren2015-04-291-0/+2
|
* eth: check if downloader is busy before showing log messageobscuren2015-04-291-0/+4
|
* eth/downloader: removed update loop and synch channelobscuren2015-04-251-25/+0
|
* eth/downloader: fixed testsobscuren2015-04-241-12/+8
|
* eth, eth/downloader: simplified synchronisation processobscuren2015-04-241-24/+13
|
* eth/downloader: removed peer td management and best peer selectionobscuren2015-04-242-16/+2
|
* eth, eth/downloader: don't require td on downloader. Fixed testsobscuren2015-04-242-21/+18
|
* eth, eth/downloader: moved peer selection to protocol handlerobscuren2015-04-242-124/+50
|
* core: added chain head reset to known blockobscuren2015-04-201-0/+2
|
* geth: admin download statusobscuren2015-04-201-11/+15
|
* downloader: fixed tests with low ttlobscuren2015-04-192-2/+3
|
* movedobscuren2015-04-191-6/+6
|
* Removed debug printlnobscuren2015-04-191-3/+0
|
* downloader: moved chunk ignoring. Fixes issue with catching upobscuren2015-04-194-31/+44
|
* downloader: defer peer reset after downloadobscuren2015-04-192-4/+8
|
* downloader: reset hashTtl on receiveobscuren2015-04-191-0/+2
|
* downloader: don't remove peers. keep them aroundobscuren2015-04-191-2/+3
|
* typoobscuren2015-04-191-1/+1
|
* downloader: free up peers from work when the downloader resetsobscuren2015-04-193-5/+22
|
* downloader: throw an error if there are no peers available for downloadobscuren2015-04-191-6/+16
| | | | | | If all peers have been tried during the block download process and some hashes are unfetchable (available peers > 0 and fetching == 0) throw an error so the process can be aborted.
* downloader: fixed a race condition for download statusobscuren2015-04-191-1/+0
|
* eth: adapted to new synchronous api of downloader's AddBlockobscuren2015-04-192-2/+2
|
* downloader: improved downloading and synchronisationobscuren2015-04-194-110/+129
| | | | | | * Downloader's peers keeps track of peer's previously requested hashes so that we don't have to re-request * Changed `AddBlock` to be fully synchronous
* downloader: make sure that hashes are only accepted from the active peerobscuren2015-04-181-18/+45
|
* downloader: reset the queue if a peer response with an empty hash setobscuren2015-04-182-14/+36
|
* downloader: all handlers check for isBusyobscuren2015-04-181-7/+7
|
* Merge branch 'develop' into downloader-protoobscuren2015-04-182-5/+117
|\ | | | | | | | | Conflicts: eth/downloader/downloader.go
| * downloader: fixed a typoobscuren2015-04-171-1/+1
| |
| * downloader: added demotion / promotion in prep. for rep. systemobscuren2015-04-162-0/+25
| |
| * downloader: updated downloader and fixed issues with catch upobscuren2015-04-161-21/+47
| | | | | | | | | | | | Properly ignore blocks coming from peers not in our peer list (blocked) and do never request anything from bad peers. Added some checks to account for blocks known when requesting hashes (missing parents).
* | downloader: updated downloader and fixed issues with catch upobscuren2015-04-161-22/+48
|/ | | | | | Properly ignore blocks coming from peers not in our peer list (blocked) and do never request anything from bad peers. Added some checks to account for blocks known when requesting hashes (missing parents).
* downloader: added missing blocks catchup functionalityobscuren2015-04-134-133/+303
| | | | | When a parent is missing in the block list an attempt should be made to fetch the missing parent and grandparents.
* downloader: return an error for peer.fetch and return chunk to queueobscuren2015-04-122-3/+16
| | | | | If a peer was somehow already fetching and somehow managed to end up in the `available` pool it should return it's work.
* downloader: renamed chunks to queueobscuren2015-04-121-0/+0
|
* downloader: implemented new downloaderobscuren2015-04-124-0/+602