Commit Graph

231 Commits

Author SHA1 Message Date
gary rong
93da0cf8a1
cmd, core, eth, light, trie: dump clean cache periodically (#20391)
* cmd, core, eth, light, trie: dump clean cache periodically

* eth: update config

* trie: minor fix

* core, trie: address comments

* eth: remove useless

* trie: print clean cache dump start too

Co-authored-by: Péter Szilágyi <peterke@gmail.com>
2020-07-28 16:30:31 +03:00
Martin Holst Swende
295693759e
core/vm: less allocations for various call variants (#21222)
* core/vm/runtime/tests: add more benchmarks

* core/vm: initial work on improving alloc count for calls to precompiles

name                                  old time/op    new time/op    delta
SimpleLoop/identity-precompile-10M-6     117ms ±75%      43ms ± 1%  -63.09%  (p=0.008 n=5+5)
SimpleLoop/loop-10M-6                   79.6ms ± 4%    70.5ms ± 1%  -11.42%  (p=0.008 n=5+5)

name                                  old alloc/op   new alloc/op   delta
SimpleLoop/identity-precompile-10M-6    24.4MB ± 0%     4.9MB ± 0%  -79.94%  (p=0.008 n=5+5)
SimpleLoop/loop-10M-6                   13.2kB ± 0%    13.2kB ± 0%     ~     (p=0.357 n=5+5)

name                                  old allocs/op  new allocs/op  delta
SimpleLoop/identity-precompile-10M-6      382k ± 0%      153k ± 0%  -59.99%  (p=0.000 n=5+4)
SimpleLoop/loop-10M-6                     40.0 ± 0%      40.0 ± 0%     ~     (all equal)

* core/vm: don't allocate big.int for touch

name                                  old time/op    new time/op    delta
SimpleLoop/identity-precompile-10M-6    43.3ms ± 1%    42.4ms ± 7%     ~     (p=0.151 n=5+5)
SimpleLoop/loop-10M-6                   70.5ms ± 1%    76.7ms ± 1%   +8.67%  (p=0.008 n=5+5)

name                                  old alloc/op   new alloc/op   delta
SimpleLoop/identity-precompile-10M-6    4.90MB ± 0%    2.46MB ± 0%  -49.83%  (p=0.008 n=5+5)
SimpleLoop/loop-10M-6                   13.2kB ± 0%    13.2kB ± 1%     ~     (p=0.571 n=5+5)

name                                  old allocs/op  new allocs/op  delta
SimpleLoop/identity-precompile-10M-6      153k ± 0%       76k ± 0%  -49.98%  (p=0.029 n=4+4)
SimpleLoop/loop-10M-6                     40.0 ± 0%      40.0 ± 0%     ~     (all equal)

* core/vm: reduce allocs in staticcall

name                                  old time/op    new time/op    delta
SimpleLoop/identity-precompile-10M-6    42.4ms ± 7%    37.5ms ± 6%  -11.68%  (p=0.008 n=5+5)
SimpleLoop/loop-10M-6                   76.7ms ± 1%    69.1ms ± 1%   -9.82%  (p=0.008 n=5+5)

name                                  old alloc/op   new alloc/op   delta
SimpleLoop/identity-precompile-10M-6    2.46MB ± 0%    0.02MB ± 0%  -99.35%  (p=0.008 n=5+5)
SimpleLoop/loop-10M-6                   13.2kB ± 1%    13.2kB ± 0%     ~     (p=0.143 n=5+5)

name                                  old allocs/op  new allocs/op  delta
SimpleLoop/identity-precompile-10M-6     76.4k ± 0%      0.1k ± 0%     ~     (p=0.079 n=4+5)
SimpleLoop/loop-10M-6                     40.0 ± 0%      40.0 ± 0%     ~     (all equal)

* trie: better use of hasher keccakState

* core/state/statedb: reduce allocations in getDeletedStateObject

* core/vm: reduce allocations in all call derivates

* core/vm: reduce allocations in call variants

- Make returnstack `uint32`
- Use a `sync.Pool` of `stack`s

* core/vm: fix tests

* core/vm: goimports

* core/vm: tracer fix + staticcall gas fix

* core/vm: add back snapshot to staticcall

* core/vm: review concerns + make returnstack pooled + enable returndata in traces

* core/vm: fix some test tracer method signatures

* core/vm: run gencodec, minor comment polish

Co-authored-by: Péter Szilágyi <peterke@gmail.com>
2020-07-16 15:06:19 +03:00
Guillaume Ballet
240d1851db
trie: quell linter in commiter.go (#21329) 2020-07-15 11:00:04 +03:00
gary rong
6eef141aef
les: historical data garbage collection (#19570)
This change introduces garbage collection for the light client. Historical
chain data is deleted periodically. If you want to disable the GC, use
the --light.nopruning flag.
2020-07-13 11:02:54 +02:00
Marius van der Wijden
ddeea1e0c6
core: types: less allocations when hashing and tx handling (#21265)
* core, crypto: various allocation savings regarding tx handling

* core: reduce allocs for gas price comparison

This change reduces the allocations needed for comparing different transactions to each other.
A call to `tx.GasPrice()` copies the gas price as it has to be safe against modifications and
also needs to be threadsafe. For comparing and ordering different transactions we don't need
these guarantees

* core: added tx.GasPriceIntCmp for comparison without allocation

adds a method to remove unneeded allocation in comparison to tx.gasPrice

* core/types: pool legacykeccak256 objects in rlpHash

rlpHash is by far the most used function in core that allocates a legacyKeccak256 object on each call.
Since it is so widely used it makes sense to add pooling here so we relieve the GC.
On my machine these changes result in > 100 MILLION less allocations and > 30 GB less allocated memory.

* reverted some changes

* reverted some changes

* trie: use crypto.KeccakState instead of replicating code

Co-authored-by: Martin Holst Swende <martin@swende.se>
2020-06-30 11:59:06 +02:00
Marius van der Wijden
b482423e61
trie: reduce allocs in insertPreimage (#21261) 2020-06-24 21:56:27 +02:00
gary rong
389da6aa48
trie: enforce monotonic range in prover and return end marker (#21130)
* trie: add hasRightElement indicator

* trie: ensure the range is monotonic increasing

* trie: address comment and fix lint

* trie: address comment

* trie: make linter happy

Co-authored-by: Péter Szilágyi <peterke@gmail.com>
2020-05-27 17:37:37 +03:00
gary rong
070a5e1252
trie: fix for range proof (#21107)
* trie: fix for range proof

* trie: fix typo
2020-05-26 13:11:29 +03:00
Martin Michlmayr
4f2784b38f
all: fix typos in comments (#21118) 2020-05-25 10:21:28 +02:00
gary rong
65ce550b37
trie: extend range proofs with non-existence (#21000)
* trie: implement range proof with non-existent edge proof

* trie: fix cornercase

* trie: consider empty range

* trie: add singleSide test

* trie: support all-elements range proof

* trie: fix typo

* trie: tiny typos and formulations

Co-authored-by: Péter Szilágyi <peterke@gmail.com>
2020-05-20 15:45:38 +03:00
Guillaume Ballet
234cc8e77f
eth/downloader: minor typo fixes in comments (#21035) 2020-05-06 15:35:04 +02:00
gary rong
5cdc2dffda
trie: fix TestBadRangeProof unit test (#21034) 2020-05-06 15:33:57 +02:00
gary rong
44ff3f3dc9
trie: initial implementation for range proof (#20908)
* trie: initial implementation for range proof

* trie: add benchmark

* trie: fix lint

* trie: fix minor issue

* trie: unset the edge valuenode as well

* trie: unset the edge valuenode as nilValuenode
2020-04-24 14:37:56 +03:00
Péter Szilágyi
79b68dd78d
Merge pull request #20923 from holiman/fix_seckeybuf
trie: fix concurrent usage of secKeyBuf, ref #20920
2020-04-20 16:52:04 +03:00
Martin Holst Swende
6402c42b67
all: simplify and fix database iteration with prefix/start (#20808)
* core/state/snapshot: start fixing disk iterator seek

* ethdb, rawdb, leveldb, memorydb: implement iterators with prefix and start

* les, core/state/snapshot: iterator fixes

* all: remove two iterator methods

* all: rename Iteratee.NewIteratorWith -> NewIterator

* ethdb: fix review concerns
2020-04-15 14:08:53 +03:00
Martin Holst Swende
af4080b4b7
trie: fix concurrent usage of secKeyBuf, ref #20920 2020-04-15 11:07:29 +02:00
Martin Holst Swende
a1313b5b1e
trie: make hasher parallel when number of changes are large (#20488)
* trie: make hasher parallel when number of changes are large

* trie: remove unused field dirtyCount

* trie: rename unhashedCount/unhashed
2020-02-04 14:02:38 +02:00
Martin Holst Swende
5a9c96454e
trie: separate hashes and committer, collapse on commit
* trie:  make db insert use size instead of full data

* core/state: minor optimization in state onleaf allocation

* trie: implement dedicated committer and hasher

* trie: use dedicated committer/hasher

* trie: linter nitpicks

* core/state, trie: avoid unnecessary storage trie load+commit

* trie: review feedback, mainly docs + minor changes

* trie: start deprecating old hasher

* trie: fix misspell+lint

* trie: deprecate hasher.go, make proof framework use new hasher

* trie: rename pure_committer/hasher to committer/hasher

* trie, core/state: fix review concerns

* trie: more review concerns

* trie: make commit collapse into hashnode, don't touch dirtyness

* trie: goimports fixes

* trie: remove panics
2020-02-03 17:28:30 +02:00
Martin Holst Swende
9b09c0fc83 * trie: utilize callbacks instead of amassing lists in ref/unref (#20529)
* trie/tests: add benchmarks and update trie tests

* trie: update benchmark tests

* trie: utilize callbacks instead of amassing lists of hashes in database ref/unref

* trie: replace remaining non-callback based accesses
2020-01-17 13:59:45 +02:00
me020523
4f7b7f84ae add node.go unit test file node_test.go (#20028)
* add node.go unit test file node_test.go

* add node_test.go file license and rollback trie_test.go

* fix unused variable v

* trie: fix license year

Co-authored-by: Péter Szilágyi <peterke@gmail.com>
2020-01-07 10:31:20 +01:00
Péter Szilágyi
13ccf6016e
trie: track dirty cache metrics, track clean writes on commit 2019-12-02 12:23:35 +02:00
gary rong
08611cfd75 trie: remove dead code (#20405) 2019-11-28 12:47:35 +02:00
gary rong
9d67222f4e trie: replace bigcache with fastcache (#19971) 2019-11-25 10:58:15 +02:00
Felix Lange
0abcf03fde trie: remove unused code (#20366) 2019-11-22 17:24:48 +02:00
gary rong
ecdbb402ee trie: remove node ordering slice in sync batch (#19929)
When we flush a batch of trie nodes into database during the state
sync, we should guarantee that all children should be flushed before
parent.

Actually the trie nodes commit order is strict by: children -> parent.
But when we flush all ready nodes into db, we don't need the order
anymore since

    (1) they are all ready nodes (no more dependency)
    (2) underlying database provides write atomicity
2019-10-28 18:50:11 +01:00
gary rong
df6c08a485 core, trie: decode the value for storage dump (#19943)
* core, trie: decode the value for storage dump

* core/state: address comment
2019-08-12 17:14:40 +03:00
Christian Muehlhaeuser
5183483c53 core/state, p2p/discover, trie, whisper: avoid unnecessary conversions (#19870)
No need to convert these types.
2019-07-22 10:30:09 +03:00
Guillaume Ballet
dcc4adfcd7 cmd/geth: wrong memory size sanitizing on OpenBSD (#19793) 2019-07-05 13:13:21 +03:00
Péter Szilágyi
536b3b416c
cosensus, core, eth, params, trie: fixes + clique history cap 2019-05-16 10:39:35 +03:00
gary rong
80469bea0c
all: integrate the freezer with fast sync
* all: freezer style syncing

core, eth, les, light: clean up freezer relative APIs

core, eth, les, trie, ethdb, light: clean a bit

core, eth, les, light: add unit tests

core, light: rewrite setHead function

core, eth: fix downloader unit tests

core: add receipt chain insertion test

core: use constant instead of hardcoding table name

core: fix rollback

core: fix setHead

core/rawdb: remove canonical block first and then iterate side chain

core/rawdb, ethdb: add hasAncient interface

eth/downloader: calculate ancient limit via cht first

core, eth, ethdb: lots of fixes

* eth/downloader: print ancient disable log only for fast sync
2019-05-16 10:39:32 +03:00
Péter Szilágyi
9effd64290
core, eth, trie: bloom filter for trie node dedup during fast sync (#19489)
* core, eth, trie: bloom filter for trie node dedup during fast sync

* eth/downloader, trie: address review comments

* core, ethdb, trie: restart fast-sync bloom construction now and again

* eth/downloader: initialize fast sync bloom on startup

* eth: reenable eth/62 until we properly remove it
2019-05-13 15:28:01 +03:00
gary rong
14868a37fb trie: clarify why verifyProof doesn't check hashes (#19530)
* trie: fix merkle proof

* trie: use hasher instead of allocate keccack256 every time

* trie: add comments
2019-05-07 15:06:07 +03:00
Péter Szilágyi
4a4abc41d4
trie: approximate the wasted cache metaspace closer 2019-04-12 11:43:16 +03:00
Péter Szilágyi
4bf0d11e7c
trie: there's no point in retrieving the metaroot 2019-04-05 13:09:28 +03:00
Martin Holst Swende
59e1953246 core, ethdb, trie: mode dirty data to clean cache on flush (#19307)
This PR is a more advanced form of the dirty-to-clean cacher (#18995),
where we reuse previous database write batches as datasets to uncache,
saving a dirty-trie-iteration and a dirty-trie-rlp-reencoding per block.
2019-03-26 15:48:31 +01:00
Martin Holst Swende
876f357364 trie: disable fnv64a hashing of hashes for bigcache (#19314)
* trie: disable fnv64a hashing of hashes for bigcache

* trie/database: add very important period
2019-03-22 17:13:28 +02:00
Péter Szilágyi
91eec1251c
cmd, core, eth, trie: get rid of trie cache generations (#19262)
* cmd, core, eth, trie: get rid of trie cache generations

* core, trie: get rid of remainder of cache gen boilerplate
2019-03-14 15:25:12 +02:00
Péter Szilágyi
8111b9dda5
ethdb, trie: tiny API tidy-up from the database rework pr 2019-03-12 12:32:02 +02:00
Péter Szilágyi
054412e335
all: clean up and proerly abstract database access 2019-03-06 13:35:03 +02:00
Matthew Halpern
514a9472ad trie: prefer nil slices over zero-length slices (#19084) 2019-02-19 14:50:11 +01:00
Martin Holst Swende
4f85c2b88b
trie: fix error in node decoding (#19111) 2019-02-16 16:16:12 +01:00
HackyMiner
bb7c786b09 trie: add missing unlock call in error case (#18985) 2019-02-04 12:42:46 +01:00
Dave McGregor
33d233d3e1
vendor, crypto, swarm: switch over to upstream sha3 package 2019-01-04 09:26:07 +02:00
Martin Holst Swende
2843001ac2 trie: fix overflow in write cache parent tracking (#18165)
trie/database: fix overflow in parent tracking
2018-11-22 15:14:31 +02:00
Péter Szilágyi
d136e985e8
trie: go fmt package 2018-11-16 16:35:39 +02:00
Łukasz Kurowski
68be45e5f8 trie: return hasher to pool (#18116)
* trie: return hasher to pool

* trie: minor code formatting fix
2018-11-16 11:50:48 +02:00
Péter Szilágyi
434dd5bc00
cmd, core, eth, light, trie: add trie read caching layer 2018-11-15 12:22:13 +02:00
Péter Szilágyi
53634f1e04 trie: remove unused originalRoot field (#17862) 2018-10-08 13:16:16 +02:00
Wenbiao Zheng
6a33954731 core, eth, trie: use common/prque (#17508) 2018-09-03 17:33:21 +02:00
Mymskmkt
70398d300d trie: fix typo (#17498) 2018-08-24 21:08:48 +03:00
Péter Szilágyi
11bbc66082
eth, trie: fix tracer GC which accidentally pruned the metaroot 2018-08-09 12:33:30 +03:00
Mymskmkt
8051a0768a trie: fix comment typo (#17350) 2018-08-08 16:08:40 +03:00
Oleg Kovalov
cf05ef9106 p2p, swarm, trie: avoid copying slices in loops (#17265) 2018-08-07 13:56:40 +03:00
Péter Szilágyi
8a9c31a307
trie: handle removing the freshest node too 2018-07-30 16:31:17 +03:00
Péter Szilágyi
319098cc1c
trie: fix a temporary memory leak in the memcache 2018-07-02 15:47:33 +03:00
Péter Szilágyi
d926bf2c7e trie: cache collapsed tries node, not rlp blobs (#16876)
The current trie memory database/cache that we do pruning on stores
trie nodes as binary rlp encoded blobs, and also stores the node
relationships/references for GC purposes. However, most of the trie
nodes (everything apart from a value node) is in essence just a
collection of references.

This PR switches out the RLP encoded trie blobs with the
collapsed-but-not-serialized trie nodes. This permits most of the
references to be recovered from within the node data structure,
avoiding the need to track them a second time (expensive memory wise).
2018-06-21 11:28:05 +02:00
Péter Szilágyi
a3267ed929
trie: don't report the root flushlist as an alloc 2018-06-11 14:32:13 +03:00
Sarlor
ea06da0892 trie: avoid unnecessary slicing on shortnode decoding (#16917)
optimization code
2018-06-07 11:48:36 +03:00
Felix Lange
e8ea5aa0d5 trie: reduce hasher allocations (#16896)
* trie: reduce hasher allocations

name    old time/op    new time/op    delta
Hash-8    4.05µs ±12%    3.56µs ± 9%  -12.13%  (p=0.000 n=20+19)

name    old alloc/op   new alloc/op   delta
Hash-8    1.30kB ± 0%    0.66kB ± 0%  -49.15%  (p=0.000 n=20+20)

name    old allocs/op  new allocs/op  delta
Hash-8      11.0 ± 0%       8.0 ± 0%  -27.27%  (p=0.000 n=20+20)

* trie: bump initial buffer cap in hasher
2018-06-05 15:06:29 +03:00
Péter Szilágyi
143c4341d8
core, eth, trie: streaming GC for the trie cache (#16810)
* core, eth, trie: streaming GC for the trie cache

* trie: track memcache statistics
2018-06-04 10:47:43 +03:00
Wenbiao Zheng
38c7eb0f26 trie: rename TrieSync to Sync and improve hexToKeybytes (#16804)
This removes a golint warning: type name will be used as trie.TrieSync by
other packages, and that stutters; consider calling this Sync.

In hexToKeybytes len(hex) is even and (even+1)/2 == even/2, remove the +1.
2018-05-29 17:48:43 +02:00
Wenbiao Zheng
d6ed2f67a8 eth, node, trie: fix minor typos (#16802) 2018-05-24 15:55:20 +03:00
Péter Szilágyi
54294b45b1
Merge pull request #16803 from karalabe/trie-avoid-funccall
trie: cleaner logic, one less func call
2018-05-24 15:54:00 +03:00
Péter Szilágyi
d31802312a
trie: cleaner logic, one less func call 2018-05-24 13:46:45 +03:00
Péter Szilágyi
c934c06cc1
trie: support proof generation from the iterator 2018-05-23 13:02:20 +03:00
kiel barry
0fe47e98c4 trie: fixes to comply with golint (#16771) 2018-05-21 23:41:31 +03:00
gary rong
7beccb29be all: get rid of error when creating memory database (#16716)
* all: get rid of error when create mdb

* core: clean up variables definition

* all: inline mdb definition
2018-05-09 15:24:25 +03:00
Eli
4a8d5d2b1e trie: golint iterator fixes (#16639) 2018-05-02 11:24:34 +03:00
xincaosu
cfe8f5fd94 trie: remove unused buf parameter (#16583) 2018-04-27 12:45:02 +03:00
gary rong
1100e8ba63 eth/downloader: flush state sync data before exit (#16280) 2018-04-09 14:46:27 +02:00
Anton Evangelatov
ae9f97221a metrics: pull library and introduce ResettingTimer and InfluxDB reporter (#15910)
* go-metrics: fork library and introduce ResettingTimer and InfluxDB reporter.

* vendor: change nonsense/go-metrics to ethersphere/go-metrics

* go-metrics: add tests. move ResettingTimer logic from reporter to type.

* all, metrics: pull in metrics package in go-ethereum

* metrics/test: make sure metrics are enabled for tests

* metrics: apply gosimple rules

* metrics/exp, internal/debug: init expvar endpoint when starting pprof server

* internal/debug: tiny comment formatting fix
2018-02-23 11:56:08 +02:00
Felix Lange
a5c0bbb4f4
all: update license information (#16089) 2018-02-14 13:49:11 +01:00
Péter Szilágyi
55599ee95d core, trie: intermediate mempool between trie and database (#15857)
This commit reduces database I/O by not writing every state trie to disk.
2018-02-05 17:40:32 +01:00
Péter Szilágyi
216e584899
Revert "trie: make fullnode children hash calculation concurrently (#15131)" (#15889)
This reverts commit 0f7fbb85d6.
2018-01-15 15:32:14 +02:00
Furkan KAMACI
b8caba9709 various: remove redundant parentheses (#15793) 2018-01-03 14:14:47 +02:00
gary rong
0f7fbb85d6 trie: make fullnode children hash calculation concurrently (#15131)
* trie: make fullnode children hash calculation concurrently

* trie: thread out only on topmost fullnode

* trie: clean up full node children hash calculation

* trie: minor code fixups
2017-11-27 13:34:17 +02:00
Felföldi Zsolt
ca376ead88 les, light: LES/2 protocol version (#14970)
This PR implements the new LES protocol version extensions:

* new and more efficient Merkle proofs reply format (when replying to
  a multiple Merkle proofs request, we just send a single set of trie
  nodes containing all necessary nodes)
* BBT (BloomBitsTrie) works similarly to the existing CHT and contains
  the bloombits search data to speed up log searches
* GetTxStatusMsg returns the inclusion position or the
  pending/queued/unknown state of a transaction referenced by hash
* an optional signature of new block data (number/hash/td) can be
  included in AnnounceMsg to provide an option for "very light
  clients" (mobile/embedded devices) to skip expensive Ethash check
  and accept multiple signatures of somewhat trusted servers (still a
  lot better than trusting a single server completely and retrieving
  everything through RPC). The new client mode is not implemented in
  this PR, just the protocol extension.
2017-10-24 15:19:09 +02:00
Péter Szilágyi
a5fcaa55ac
trie: make hasher benchmark meaningful post-caches 2017-10-13 11:22:38 +03:00
Paul Litvak
a92d8a2654 trie: fix typo (#15152) 2017-09-18 23:07:19 +02:00
Felix Lange
10181b57a9 core, eth/downloader: commit block data using batches (#15115)
* ethdb: add Putter interface and Has method

* ethdb: improve docs and add IdealBatchSize

* ethdb: remove memory batch lock

Batches are not safe for concurrent use.

* core: use ethdb.Putter for Write* functions

This covers the easy cases.

* core/state: simplify StateSync

* trie: optimize local node check

* ethdb: add ValueSize to Batch

* core: optimize HasHeader check

This avoids one random database read get the block number. For many uses
of HasHeader, the expectation is that it's actually there. Using Has
avoids a load + decode of the value.

* core: write fast sync block data in batches

Collect writes into batches up to the ideal size instead of issuing many
small, concurrent writes.

* eth/downloader: commit larger state batches

Collect nodes into a batch up to the ideal size instead of committing
whenever a node is received.

* core: optimize HasBlock check

This avoids a random database read to get the number.

* core: use numberCache in HasHeader

numberCache has higher capacity, increasing the odds of finding the
header without a database lookup.

* core: write imported block data using a batch

Restore batch writes of state and add blocks, tx entries, receipts to
the same batch. The change also simplifies the miner.

This commit also removes posting of logs when a forked block is imported.

* core: fix DB write error handling

* ethdb: use RLock for Has

* core: fix HasBlock comment
2017-09-09 19:03:07 +03:00
Petr Mikusek
99da85c895 trie: typo in comment 2017-07-25 19:02:40 +02:00
Felix Lange
9e5f03b6c4 core/state: access trie through Database interface, track errors (#14589)
With this commit, core/state's access to the underlying key/value database is
mediated through an interface. Database errors are tracked in StateDB and
returned by CommitTo or the new Error method.

Motivation for this change: We can remove the light client's duplicated copy of
core/state. The light client now supports node iteration, so tracing and storage
enumeration can work with the light client (not implemented in this commit).
2017-06-27 15:57:06 +02:00
Felix Lange
0042f13d47 eth/downloader: separate state sync from queue (#14460)
* eth/downloader: separate state sync from queue

Scheduling of state node downloads hogged the downloader queue lock when
new requests were scheduled. This caused timeouts for other requests.
With this change, state sync is fully independent of all other downloads
and doesn't involve the queue at all.

State sync is started and checked on in processContent. This is slightly
awkward because processContent doesn't have a select loop. Instead, the
queue is closed by an auxiliary goroutine when state sync fails. We
tried several alternatives to this but settled on the current approach
because it's the least amount of change overall.

Handling of the pivot block has changed slightly: the queue previously
prevented import of pivot block receipts before the state of the pivot
block was available. In this commit, the receipt will be imported before
the state. This causes an annoyance where the pivot block is committed
as fast block head even when state downloads fail. Stay tuned for more
updates in this area ;)

* eth/downloader: remove cancelTimeout channel

* eth/downloader: retry state requests on timeout

* eth/downloader: improve comment

* eth/downloader: mark peers idle when state sync is done

* eth/downloader: move pivot block splitting to processContent

This change also ensures that pivot block receipts aren't imported
before the pivot block itself.

* eth/downloader: limit state node retries

* eth/downloader: improve state node error handling and retry check

* eth/downloader: remove maxStateNodeRetries

It fails the sync too much.

* eth/downloader: remove last use of cancelCh in statesync.go

Fixes TestDeliverHeadersHang*Fast and (hopefully)
the weird cancellation behaviour at the end of fast sync.

* eth/downloader: fix leak in runStateSync

* eth/downloader: don't run processFullSyncContent in LightSync mode

* eth/downloader: improve comments

* eth/downloader: fix vet, megacheck

* eth/downloader: remove unrequested tasks anyway

* eth/downloader, trie: various polishes around duplicate items

This commit explicitly tracks duplicate and unexpected state
delieveries done against a trie Sync structure, also adding
there to import info logs.

The commit moves the db batch used to commit trie changes one
level deeper so its flushed after every node insertion. This
is needed to avoid a lot of duplicate retrievals caused by
inconsistencies between Sync internals and database. A better
approach is to track not-yet-written states in trie.Sync and
flush on commit, but I'm focuing on correctness first now.

The commit fixes a regression around pivot block fail count.
The counter previously was reset to 1 if and only if a sync
cycle progressed (inserted at least 1 entry to the database).
The current code reset it already if a node was delivered,
which is not stong enough, because unless it ends up written
to disk, an attacker can just loop and attack ad infinitum.

The commit also fixes a regression around state deliveries
and timeouts. The old downloader tracked if a delivery is
stale (none of the deliveries were requestedt), in which
case it didn't mark the node idle and did not send further
requests, since it signals a past timeout. The current code
did mark it idle even on stale deliveries, which eventually
caused two requests to be in flight at the same time, making
the deliveries always stale and mass duplicating retrievals
between multiple peers.

* eth/downloader: fix state request leak

This commit fixes the hang seen sometimes while doing the state
sync. The cause of the hang was a rare combination of events:
request state data from peer, peer drops and reconnects almost
immediately. This caused a new download task to be assigned to
the peer, overwriting the old one still waiting for a timeout,
which in turned leaked the requests out, never to be retried.
The fix is to ensure that a task assignment moves any pending
one back into the retry queue.

The commit also fixes a regression with peer dropping due to
stalls. The current code considered a peer stalling if they
timed out delivering 1 item. However, the downloader never
requests only one, the minimum is 2 (attempt to fine tune
estimated latency/bandwidth). The fix is simply to drop if
a timeout is detected at 2 items.

Apart from the above bugfixes, the commit contains some code
polishes I made while debugging the hang.

* core, eth, trie: support batched trie sync db writes

* trie: rename SyncMemCache to syncMemBatch
2017-06-22 15:26:03 +03:00
Felix Lange
693d9ccbfb trie: more node iterator improvements (#14615)
* ethdb: remove Set

Set deadlocks immediately and isn't part of the Database interface.

* trie: add Err to Iterator

This is useful for testing because the underlying NodeIterator doesn't
need to be kept in a separate variable just to get the error.

* trie: add LeafKey to iterator, panic when not at leaf

LeafKey is useful for callers that can't interpret Path.

* trie: retry failed seek/peek in iterator Next

Instead of failing iteration irrecoverably, make it so Next retries the
pending seek or peek every time.

Smaller changes in this commit make this easier to test:

* The iterator previously returned from Next on encountering a hash
  node. This caused it to visit the same path twice.
* Path returned nibbles with terminator symbol for valueNode attached
  to fullNode, but removed it for valueNode attached to shortNode. Now
  the terminator is always present. This makes Path unique to each node
  and simplifies Leaf.

* trie: add Path to MissingNodeError

The light client trie iterator needs to know the path of the node that's
missing so it can retrieve a proof for it. NodeIterator.Path is not
sufficient because it is updated when the node is resolved and actually
visited by the iterator.

Also remove unused fields. They were added a long time ago before we
knew which fields would be needed for the light client.
2017-06-20 18:26:09 +02:00
S. Matthew English
061889d4ea rlp, trie, contracts, compression, consensus: improve comments (#14580) 2017-06-12 14:45:17 +02:00
Felix Lange
4047ccad2f trie: add start key to NodeIterator constructors
The 'step' method is split into two parts, 'peek' and 'push'. peek
returns the next state but doesn't make it current.

The end of iteration was previously tracked by setting 'trie' to nil.
End of iteration is now tracked using the 'iteratorEnd' error, which is
slightly cleaner and requires less code.
2017-04-25 02:14:31 +02:00
Felix Lange
a13e920af0 trie: clean up iterator constructors
Make it so each iterator has exactly one public constructor:

- NodeIterators can be created through a method.
- Iterators can be created through NewIterator on any NodeIterator.
2017-04-25 02:14:31 +02:00
Felix Lange
f958d7d482 trie: rework and document key encoding
'encode' and 'decode' are meaningless because the code deals with three
encodings. Document the encodings and give a name to each one.
2017-04-25 02:14:31 +02:00
Felix Lange
a31d268b76 trie: remove Key in MissingNodeError
The key was constructed from nibbles, which isn't possible for all
nodes. Remove the only use of Key in LightTrie by always retrying with
the original key that was looked up.
2017-04-18 14:52:11 +02:00
Nick Johnson
b35aa21f9f trie: implement unionIterator (#14312) 2017-04-13 11:14:19 +02:00
Péter Szilágyi
09777952ee core, consensus: pluggable consensus engines (#3817)
This commit adds pluggable consensus engines to go-ethereum. In short, it
introduces a generic consensus interface, and refactors the entire codebase to
use this interface.
2017-04-05 00:16:29 +02:00
Felix Lange
f1534f5797 trie, whisper/whisperv5: use math/rand Read function 2017-03-22 20:49:15 +01:00
Péter Szilágyi
d4fd06c3dc
all: blidly swap out glog to our log15, logs need rework 2017-02-23 12:16:44 +02:00
Nick Johnson
555273495b trie: add difference iterator (#3637)
This PR implements a differenceIterator, which allows iterating over trie nodes
that exist in one trie but not in another. This is a prerequisite for most GC
strategies, in order to find obsolete nodes.
2017-02-22 23:49:34 +01:00
Felix Lange
b9b3efb09f all: fix ineffectual assignments and remove uses of crypto.Sha3
go get github.com/gordonklaus/ineffassign
ineffassign .
2017-01-09 16:24:42 +01:00
Péter Szilágyi
18c77744ff
all: fix spelling errors 2017-01-06 19:44:35 +02:00
Felix Lange
f2da6581ba all: fix issues reported by honnef.co/go/simple/cmd/gosimple 2017-01-06 18:18:07 +01:00
Felix Lange
35a7dcb162 all: gofmt -w -s 2017-01-06 15:52:03 +01:00
Felix Lange
d3b751e4d9 trie: remove dependency on ethdb
This removes the core/types -> leveldb dependency.
2017-01-06 14:15:22 +01:00
Péter Szilágyi
90b16a3e85
core/state, eth/downloader, trie: reset fast-failure on progress 2016-10-31 14:19:14 +02:00