Commit Graph

165 Commits

Author SHA1 Message Date
Jacek Sieka 9a2f3176f7 guard node.state (#237)
* fix node.state being used unsafely across async callbacks (fixes #230)
* node.state -> node.stateCache
* fix attestation validation not to use asserts
* re-enable attestation pool tests
* prepare for resizing validator registry
2019-04-06 09:46:07 +02:00
Dustin Brody c53de3e550 Caching updates/refactoring & state_sim defaults updates (#235)
* rm now-superceded shuffling cache (shuffling is only called from get_crosslinks), which was badly architected due to trying to exist in state; rm one more vestige of previous light-client regime (one more to go, from datatypes)

* fix wrong shuffling list size (active validator size, not validator size) to make consistent with 0.5.1 (will be inconsistent with testnet0); fix typo and change defaults in state_sim

* doAssert a couple of constant relationships necessary to avoid underflow; rm non-spec, unused helper function get_new_recent_block_roots

* refactor separate crosslink_committee_cache and winning_root_participants_cache(s) into StateData object; remove last vestige of previous shuffling cache

* separate out caching parts of StateData to new StateCache object
2019-04-05 08:18:13 -06:00
Jacek Sieka 605dd0a0e9 Some speedups (#226)
* ssz: avoid memory allocations

* a bit fishy with the 32-item stack.. this should be a smallvector

* digest: avoid another burnmem

* avoid a few allocations by using iterator
2019-04-03 15:46:22 +00:00
cheatfate 3117a4481e Remove unneeded initialization. 2019-03-28 23:24:50 +02:00
cheatfate 6f4c7b775c Remove one-line hash to avoid burnMem.
Remove unnecessary copyMem()s.
2019-03-28 23:24:50 +02:00
Jacek Sieka 201b079fe3 avoid bad blocks by filtering included attestations
* fix epoch violation during initial attestation validation
* fill in attestation fields (finalizes again!)
2019-03-28 21:10:08 +02:00
Dustin Brody 2c566df411 rm debug message 2019-03-28 10:40:20 +02:00
Dustin Brody ced4dbe859 4x speedup on epoch processing for 64 validators 2019-03-28 10:40:20 +02:00
Jacek Sieka 3ad6053298 Use block, slot tuple to idenfiy fork
this is the beginning of tracking block-slots more precisely, so we can
the justification epoch slot bug.

* avoid asyncDiscard (swallows assertions)
* fix attestation delay
* fix several state root cache bugs
* introduce workaround for genesis epoch spec issue
2019-03-28 10:32:09 +02:00
Zahary Karadjov 406df9d44f More chatty syncing 2019-03-27 18:13:39 +02:00
Jacek Sieka 81c66fbce0
use signed_root as canonical block root (#211)
* no need to pass prev_block_root when updating state
* some Slot fixes
* fix `hash_tree_root` for Slot, Epoch
* detect missing hash_tree_root type support
* remove some <0.5 state checks
2019-03-26 19:32:35 -06:00
Mamy André-Ratsimbazafy 8273345a2c fix comment 2019-03-26 17:35:09 +02:00
Mamy André-Ratsimbazafy 8b5e18b7a4 Magic fix to #209 2019-03-26 17:35:09 +02:00
Mamy André-Ratsimbazafy 57c693c88f Only check for epoch underflow just before peer communication
During state processing epochs can underflow (hence why we have a non-zero arbitrary genesis epoch)
2019-03-26 17:35:09 +02:00
Mamy André-Ratsimbazafy e189c1d3cd Crash before making attestation when replaying state pass genesis 2019-03-26 17:35:09 +02:00
Mamy André-Ratsimbazafy e8160c2126 more logs to investigate epoch underflow/overflow https://github.com/status-im/nim-beacon-chain/issues/207 2019-03-26 17:35:09 +02:00
Jacek Sieka 1b0e67c88c
ssz: update to 0.5.1:ish (#202)
* ssz: update to 0.5.1:ish
* slightly fewer seq allocations
* still a lot of potential for optimization
* fixes #174
* ssz: avoid reallocating leaves (logN merkle impl)
2019-03-25 10:46:31 -06:00
Dustin Brody 6b1527ccd5 Almost final 0.5.1 spec updates (#201)
* update processExits to 0.5.1

* mark process_deposit, get_empty_block as 0.5.1; update get_genesis_beacon_state to 0.5.1; implement processDeposits

* hotfix via https://github.com/ethereum/eth2.0-specs/pull/821
2019-03-24 12:21:23 -06:00
Dustin Brody 9f55b4646c
More 0.5.1 spec updates (#195)
* rm gone-in-0.5.0 Proposal, verifyBlockSignature, and slot check which moved to spec function processBlockHeader

* mark get_attestation_participants and get_epoch_committee_count as 0.5.1; finish updating processAttestations to 0.5.1; add kludgy workaround for bug relating to get_winning_roots_etc using crosslink_data_root as index when we have that as ZERO_HASH for all leading to it confusing attesters on different shards; rm BeaconState.latest_attestations, which splits into previous_epoch_attestations and current_epoch_attestations

* Fix CI due to removed latest_attestations field
2019-03-22 18:33:12 +00:00
Zahary Karadjov 0a027e410a Address review concerns and make some steps towards json logging 2019-03-22 17:04:17 +02:00
Dustin Brody 94c3b005fa
https://github.com/ethereum/eth2.0-specs/pull/816 hotfix (#194) 2019-03-21 16:02:50 +00:00
Dustin Brody 1bb2859aca
implement bulk of 0.5.0 state transition function (#192) 2019-03-21 15:31:36 +00:00
Jacek Sieka ee89ef1c79
cleanups (#189)
* add simple bitfield type (fixes #19)
* fix bit endianess to match spec
* consolidate attestation and block logging
* cruft cleanup
* run optimizer on tests, speeds up build
2019-03-20 14:01:48 -06:00
Dustin Brody 13c7f7fcec begin refactoring 0.5.0 epoch transition (#185)
* begin refactoring 0.5.0 epoch transition at justification and finalization

* update processTransfers to 0.5.0 spec

* update get_crosslink_committees_at_slot to 0.5.0 except for noted workaround for pre-0.5.0 spec bug related to epoch boundary conditions, to be flipped in more coordinated way with other epoch processing changes

* mark ejection processing as 0.5.0

* 0.5.0 epoch update finalization

* rm BEACON_CHAIN_SHARD_NUMBER to complete updating miscellaneous constants to 0.5.0

* mark verify_slashable_attestation as 0.5.0

* update process_attester_slashing to 0.5.0

* update process_slashings(...) to 0.5.0 and mark process_exit_queue as 0.5.0

* refactor epoch processing by implementing 0.5.0's update_registry_and_shuffling_data(...)
2019-03-19 20:19:46 -06:00
Dustin Brody 2bc632c29b
AttestationData.justified epoch -> AttestationData.source_epoch and AttestationData.latest_crosslink -> AttestationData.previous_crosslink; implement verify_merkle_branch and update process_deposit to use it; rm BeaconState.batched_block_roots; Transfer.{from, to} -> Transfer.{sender, recipient}; update Fork data type and code to use bytes4; implement int_to_bytes8 (close to metaprogramming time); adjust signed_root and all callers to use 0.5.0 calling signature; processDepositRoot -> processEth1Data (#182) 2019-03-18 15:42:42 +00:00
Dustin Brody 798197879b
Revert "AttestationData.justified epoch -> AttestationData.source_epoch and AttestationData.latest_crosslink -> AttestationData.previous_crosslink; implement verify_merkle_branch and update process_deposit to use it; rm BeaconState.batched_block_roots; Transfer.{from, to} -> Transfer.{sender, recipient}; update Fork data type and code to use bytes4; implement int_to_bytes8 (close to metaprogramming time); adjust signed_root and all callers to use 0.5.0 calling signature; processDepositRoot -> processEth1Data"
This reverts commit 18f1d000a1.
2019-03-17 15:24:50 -07:00
Dustin Brody 18f1d000a1
AttestationData.justified epoch -> AttestationData.source_epoch and AttestationData.latest_crosslink -> AttestationData.previous_crosslink; implement verify_merkle_branch and update process_deposit to use it; rm BeaconState.batched_block_roots; Transfer.{from, to} -> Transfer.{sender, recipient}; update Fork data type and code to use bytes4; implement int_to_bytes8 (close to metaprogramming time); adjust signed_root and all callers to use 0.5.0 calling signature; processDepositRoot -> processEth1Data 2019-03-17 15:17:40 -07:00
Dustin Brody f36c2d86dc begin 0.5.0 spec update (#179)
* begin 0.5.0 spec update: parent_root -> previous_block_root, BeaconState.justified_epoch -> BeaconState.current_justified_epoch, DOMAIN_PROPOSAL -> DOMAIN_BEACON_BLOCK, temporarily rename BeaconBlockHeader to BeaconBlockHeaderRLP to allow for gradual re-merging without disrupting RLP; mark a few unchanged functions and data types, implement get_temporary_block_header/get_empty_block/should_update_validator_registry/processBlockHeader/cacheState; update a few others

* a dozen or so more trivial mostly comment changes, finding more unchanged parts of 0.5.0

* several more trivial changes; goal is to reduce noise around the upcoming substantial changes (epoch processing, etc)
2019-03-16 13:52:37 -06:00
Jacek Sieka 1cb8ae9004
state processing fixes (#177)
* remove some redundant state updates
* when attesting late, use correct state / head
* don't send out obsolete attestations
* don't propose obsolete blocks
* remove some more resundant state updates :)
* simplify block logging (experimental)
* document fork choice division
* fix some Slot / Epoch conversion warnings
2019-03-14 07:33:56 -06:00
Ștefan Talpalaru de295619be
assert() -> doAssert() 2019-03-14 00:04:43 +01:00
Zahary Karadjov 562eafe124 Finish the Spec/Epoch types refactoring 2019-03-13 23:23:01 +02:00
Dustin Brody 1479bae22f implement distinct Epoch type and continue 0.4.0 spec update (#173)
* implement distinct Epoch type
* update two more spec 0.4.0 markers, leaving just deposit processing
2019-03-12 19:46:44 -06:00
Dustin Brody 50a0948bc0 Mostly convert to using distinct Slot type (#172) 2019-03-12 16:21:32 -06:00
Dustin Brody 434ba5727f finish all but 3 items for 0.4.0 spec update (#167) 2019-03-12 10:59:43 +01:00
Dustin Brody 13de015038 more 0.4.0 spec updates: renamings, continue Proposal updates, and note several more functions which were verified not to have changed in 0.4.0 (#163) 2019-03-11 09:33:24 -06:00
Dustin Brody 89aa3ec53f
fix keccak256 digest comments 2019-03-09 12:34:08 -08:00
Jacek Sieka aec31b1126 Updates for finalizing chain! (#162)
* set epoch_boundary_root - chain finalizes!
* fix slotStart to offset GENESIS_SLOT
* work around bug that will be fixed by
https://github.com/ethereum/eth2.0-specs/pull/732
* compile with debug info (there was a GC-related crash in C land)
2019-03-09 04:23:14 +00:00
Dustin Brody 0b47acae37
some comment changes noting further 0.4.0-spec-update progress, but no meaningful semantic changes (#161) 2019-03-08 18:23:42 +00:00
Dustin Brody 1afdcda62d
spec version 0.4.0 update: shard_block_root -> crosslink_data_root; switch from validator.slashed_epoch -> validator.slashed approach; a couple assert -> doAssert; note fix for underflow-prone spec changes in checkAttestation; GENESIS_SLOT change from 2^63 to 2^32; refactor Proposer and ProposalSlashing structure; add signed_root; remove duplicate slashValidator (#159)
* spec version 0.4.0 update: shard_block_root -> crosslink_data_root; switch from validator.slashed_epoch -> validator.slashed approach; a couple assert -> doAssert; note fix for underflow-prone spec changes in checkAttestation; GENESIS_SLOT change from 2^63 to 2^32; refactor Proposer and ProposalSlashing structure; add signed_root; remove duplicate slashValidator

* re-apply shard_block_root -> crosslink_data_root

* remove incorrect humaneSlotNum

* add (run-time only, alas) sanity check on signed_root
2019-03-08 17:44:31 +00:00
Dustin Brody 72749f4d04
processEpoch 5x faster, total 2x faster, with 16k validators (#157)
* processEpoch 5x faster, total 2x faster, with 16k validators

* don't recompute active_validator_indices as much

* remove duplicate process_ejections function
2019-03-07 16:53:44 +00:00
Dustin Brody 5d0de00168 transform a few quadratic algorithms from processEpoch into linear or otherwise subquadratic algorithms (#155) 2019-03-06 08:21:13 -06:00
Dustin Brody b8402e9809
speed up epoch processing by caching shuffling results (#153)
* speed up epoch processing by caching shuffling results
2019-03-05 19:10:36 +00:00
Yuriy Glukhov 21e4debda0
Merge pull request #117 from status-im/sync2
Initial full sync impl
2019-03-04 21:05:24 +02:00
Yuriy Glukhov 9ea0bf4326
Comments addressed 2019-03-04 13:50:26 +02:00
Dustin Brody 81d5becc7b Switch hash functions, prevent underflow & verify that shuffling works (#149)
* prevent balance underflow, per spec implementation note

* verify that permutation shuffing works: add YAML dependency to be used for ingesting test cases; switch from blake2 to keccak256 to match EF test cases; remove inefficient naive-spec-version of shuffling algorithm now that protolambda's can be tested directly
2019-03-01 17:50:01 -06:00
Yuriy Glukhov bed6510da3
Initial full sync impl 2019-03-01 19:39:38 +02:00
Zahary Karadjov 68a39a21be Mysterious fix for the mysterious build issue 2019-03-01 18:51:37 +02:00
Dustin Brody 794f6aed89
fix https://github.com/status-im/nim-beacon-chain/issues/145 (#147) 2019-03-01 14:39:17 +00:00
Jacek Sieka ddb3a9cd94
Revert "Add some missing RLP serialization routines (#144)" (#146)
This reverts commit 859aa327cb.
2019-03-01 07:34:37 -06:00
zah 859aa327cb Add some missing RLP serialization routines (#144)
* Add some missing RLP serialization routines
2019-02-28 15:50:21 -06:00