Trie tracer is an auxiliary tool to capture all deleted nodes
which can't be captured by trie.Committer. The deleted nodes
can be removed from the disk later.
* trie: fix memory leak in trie iterator
In the trie iterator, live nodes are tracked in a stack while iterating.
Popped node states should be explictly set to nil in order to get
garbage-collected.
* trie: fix empty trie iterator
This change speeds up trie hashing and all other activities that require
RLP encoding of trie nodes by approximately 20%. The speedup is achieved by
avoiding reflection overhead during node encoding.
The interface type trie.node now contains a method 'encode' that works with
rlp.EncoderBuffer. Management of EncoderBuffers is left to calling code.
trie.hasher, which is pooled to avoid allocations, now maintains an
EncoderBuffer. This means memory resources related to trie node encoding
are tied to the hasher pool.
Co-authored-by: Felix Lange <fjl@twurst.com>
This PR adds an addtional API called `NewBatchWithSize` for db
batcher. It turns out that leveldb batch memory allocation is
super inefficient. The main reason is the allocation step of
leveldb Batch is too small when the batch size is large. It can
take a few second to build a leveldb batch with 100MB size.
Luckily, leveldb also offers another API called MakeBatch which can
pre-allocate the memory area. So if the approximate size of batch is
known in advance, this API can be used in this case.
It's needed in new state scheme PR which needs to commit a batch of
trie nodes in a single batch. Implement the feature in a seperate PR.
This functionality is needed in new path-based storage scheme, but
can be implemented in a seperate PR though.
When an account is deleted, then all the storage slots should be
nuked out from the disk as well. In hash-based storage scheme they
are still left in the disk but in new scheme, they will be iterated
and marked as deleted.
But why the NodeBlob API is needed in this scenario? Because when
the node is marked deleted, the previous value is also required to
be recorded to construct the reverse diff.
* trie/proof: edge case for VerifyRangeProof
* more consistency with other tests in the file
* trie: fix test todo
Co-authored-by: Martin Holst Swende <martin@swende.se>
Trim the search key from head as it's being pushed deeper into the trie. Previously the search key was never modified but each node kept information how to slice and compare it in keyOffset. Now the keyOffset is not needed as this information is included in the slice of the search key. This way the keyOffset can be removed and key manipulation
simplified.
Some tests take quite some time during exit, which I think causes
some appveyor fails like this:
https://ci.appveyor.com/project/ethereum/go-ethereum/builds/39511210/job/xhom84eg2e4uulq3
One of the things that seem to take time during exit is waiting
(up to 100ms) for the syncbloom to close. This PR changes it to use
a channel, instead of looping with a 100ms wait.
This also includes some unrelated changes improving the reliability of
eth/fetcher tests, which fail a lot because they are time-dependent.
When deleting in fullNode, and the new child node nn is not nil, there is no need
to check the number of non-nil entries in the node. This is because the fullNode
must've contained at least two children before deletion, so there must be another
child node other than nn.
Co-authored-by: Felix Lange <fjl@twurst.com>
* eth/protocols/snap: generate storage trie from full dirty snap data
* eth/protocols/snap: get rid of some more dead code
* eth/protocols/snap: less frequent logs, also log during trie generation
* eth/protocols/snap: implement dirty account range stack-hashing
* eth/protocols/snap: don't loop on account trie generation
* eth/protocols/snap: fix account format in trie
* core, eth, ethdb: glue snap packets together, but not chunks
* eth/protocols/snap: print completion log for snap phase
* eth/protocols/snap: extended tests
* eth/protocols/snap: make testcase pass
* eth/protocols/snap: fix account stacktrie commit without defer
* ethdb: fix key counts on reset
* eth/protocols: fix typos
* eth/protocols/snap: make better use of delivered data (#44)
* eth/protocols/snap: make better use of delivered data
* squashme
* eth/protocols/snap: reduce chunking
* squashme
* eth/protocols/snap: reduce chunking further
* eth/protocols/snap: break out hash range calculations
* eth/protocols/snap: use sort.Search instead of looping
* eth/protocols/snap: prevent crash on storage response with no keys
* eth/protocols/snap: nitpicks all around
* eth/protocols/snap: clear heal need on 1-chunk storage completion
* eth/protocols/snap: fix range chunker, add tests
Co-authored-by: Péter Szilágyi <peterke@gmail.com>
* trie: fix test API error
* eth/protocols/snap: fix some further liter issues
* eth/protocols/snap: fix accidental batch reuse
Co-authored-by: Martin Holst Swende <martin@swende.se>
* core/state/snapshot: reuse memory data instead of hitting disk when generating
* trie: minor nitpicks wrt the resolver optimization
* core/state/snapshot, trie: use key/value store for resolver
* trie: fix linter
Co-authored-by: Péter Szilágyi <peterke@gmail.com>
This change improves the efficiency of the nodeIterator seek
operation. Previously, seek essentially ran the iterator forward
until it found the matching node. With this change, it skips
over fullnode children and avoids resolving them from the database.
The stacktrie is a bit un-untuitive, API-wise: since it mutates input values.
Such behaviour is dangerous, and easy to get wrong if the calling code 'forgets' this quirk. The behaviour is fixed by this PR, so that the input values are not modified by the stacktrie.
Note: just as with the Trie, the stacktrie still references the live input objects, so it's still _not_ safe to mutate the values form the callsite.
This adds support for EIP-2718 typed transactions as well as EIP-2930
access list transactions (tx type 1). These EIPs are scheduled for the
Berlin fork.
There very few changes to existing APIs in core/types, and several new APIs
to deal with access list transactions. In particular, there are two new
constructor functions for transactions: types.NewTx and types.SignNewTx.
Since the canonical encoding of typed transactions is not RLP-compatible,
Transaction now has new methods for encoding and decoding: MarshalBinary
and UnmarshalBinary.
The existing EIP-155 signer does not support the new transaction types.
All code dealing with transaction signatures should be updated to use the
newer EIP-2930 signer. To make this easier for future updates, we have
added new constructor functions for types.Signer: types.LatestSigner and
types.LatestSignerForChainID.
This change also adds support for the YoloV3 testnet.
Co-authored-by: Martin Holst Swende <martin@swende.se>
Co-authored-by: Felix Lange <fjl@twurst.com>
Co-authored-by: Ryan Schneider <ryanleeschneider@gmail.com>
* remove uneeded convertion type
* remove redundant type in composite literal
* omit explicit type where implicit
* remove unused redundant parenthesis
* remove redundant import alias duktape
This commit splits the eth package, separating the handling of eth and snap protocols. It also includes the capability to run snap sync (https://github.com/ethereum/devp2p/blob/master/caps/snap.md) , but does not enable it by default.
Co-authored-by: Marius van der Wijden <m.vanderwijden@live.de>
Co-authored-by: Martin Holst Swende <martin@swende.se>
This commit fixes a flaw in two testcases, and brings down the exec-time from ~40s to ~8s for trie/TestIncompleteSync.
The checkConsistency was performed over and over again on the complete set of nodes, not just the recently added, turning it into a quadratic runtime.
* trie: fix tests to work on 32-bit systems
* les: make test work on 32-bit platform
* cmd/geth: fix windows-issues on tests
* trie: improve balance
* cmd/geth: make account tests less verbose + less mem intense
* rpc: make debug-level log output less verbose
* cmd/geth: lint
* trie: fix error in stacktrie not committing small roots
* fuzzers: make trie-fuzzer use correct returnvalues
* trie: improved tests
* tests/fuzzers: fuzzer for stacktrie vs regular trie
* test/fuzzers: make stacktrie fuzzer use 32-byte keys
* trie: fix error in stacktrie with small nodes
* trie: add (skipped) testcase for stacktrie
* tests/fuzzers: address review comments for stacktrie fuzzer
* trie: fix docs in stacktrie
* trie: update tests to check commit integrity
* trie: polish committer
* trie: fix typo
* trie: remove hasvalue notion
According to the benchmarks, type assertion between the pointer and
interface is extremely fast.
BenchmarkIntmethod-12 1000000000 1.91 ns/op
BenchmarkInterface-12 1000000000 2.13 ns/op
BenchmarkTypeSwitch-12 1000000000 1.81 ns/op
BenchmarkTypeAssertion-12 2000000000 1.78 ns/op
So the overhead for asserting whether the shortnode has "valuenode"
child is super tiny. No necessary to have another field.
* trie: linter nitpicks
Co-authored-by: Martin Holst Swende <martin@swende.se>
core/types: use stacktrie for derivesha
trie: add stacktrie file
trie: fix linter
core/types: use stacktrie for derivesha
rebased: adapt stacktrie to the newer version of DeriveSha
Co-authored-by: Martin Holst Swende <martin@swende.se>
More linter fixes
review feedback: no key offset for nodes converted to hashes
trie: use EncodeRLP for full nodes
core/types: insert txs in order in derivesha
trie: tests for derivesha with stacktrie
trie: make stacktrie use pooled hashers
trie: make stacktrie reuse tmp slice space
trie: minor polishes on stacktrie
trie/stacktrie: less rlp dancing
core/types: explain the contorsions in DeriveSha
ci: fix goimport errors
trie: clear mem on subtrie hashing
squashme: linter fix
stracktrie: use pooling, less allocs (#3)
trie: in-place hex prefix, reduce allocs and add rawNode.EncodeRLP
Reintroduce the `[]node` method, add the missing `EncodeRLP` implementation for `rawNode` and calculate the hex prefix in place.
Co-authored-by: Martin Holst Swende <martin@swende.se>
Co-authored-by: Martin Holst Swende <martin@swende.se>
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.
* 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>
* 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
* 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>
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
* 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
* 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
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.
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).
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.
* 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
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.
* 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
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).
* 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
* 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.
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.
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.
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.
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.
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.
Commit 40cdcf1183 broke the optimisation which kept nodes resolved
during Get in the trie. The decoder assigned cache generation 0
unconditionally, causing resolved nodes to get flushed on Commit.
This commit fixes it and adds two tests.
The trie sync code links subtries using pointers into node structs.
Since commit 40cdcf1183 nodes are no longer copied when unpacking from
an interface value, causing all nodes to get linked up as the sync
progresses. Fix it by breaking the pointer chain with an explicit copy.
* trie: store nodes as pointers
This avoids memory copies when unwrapping node interface values.
name old time/op new time/op delta
Get 388ns ± 8% 215ns ± 2% -44.56% (p=0.000 n=15+15)
GetDB 363ns ± 3% 202ns ± 2% -44.21% (p=0.000 n=15+15)
UpdateBE 1.57µs ± 2% 1.29µs ± 3% -17.80% (p=0.000 n=13+15)
UpdateLE 1.92µs ± 2% 1.61µs ± 2% -16.25% (p=0.000 n=14+14)
HashBE 2.16µs ± 6% 2.18µs ± 6% ~ (p=0.436 n=15+15)
HashLE 7.43µs ± 3% 7.21µs ± 3% -2.96% (p=0.000 n=15+13)
* trie: close temporary databases in GetDB benchmark
* trie: don't keep []byte from DB load around
Nodes decoded from a DB load kept hashes and values as sub-slices of
the DB value. This can be a problem because loading from leveldb often
returns []byte with a cap that's larger than necessary, increasing
memory usage.
* trie: unload old cached nodes
* trie, core/state: use cache unloading for account trie
* trie: use explicit private flags (fixes Go 1.5 reflection issue).
* trie: fixup cachegen overflow at request of nick
* core/state: rename journal size constant
Delete crashed if a fullNode contained a valueNode directly. This bug is
very unlikely to occur with SecureTrie, but can happen with regular
tries. This commit also introduces a randomised test which triggers all
trie operations, which should prevent such bugs in the future.
Credit for finding this bug goes to Github user @rjl493456442.
