* 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
* p2p/discover, p2p/discv5: add marshaling methods to Node
* p2p/netutil: make Netlist decodable from TOML
* common/math: encode nil HexOrDecimal256 as 0x0
* cmd/geth: add --config file flag
* cmd/geth: add missing license header
* eth: prettify Config again, fix tests
* eth: use gasprice.Config instead of duplicating its fields
* eth/gasprice: hide nil default from dumpconfig output
* cmd/geth: hide genesis block in dumpconfig output
* node: make tests compile
* console: fix tests
* cmd/geth: make TOML keys look exactly like Go struct fields
* p2p: use discovery by default
This makes the zero Config slightly more useful. It also fixes package
node tests because Node detects reuse of the datadir through the
NodeDatabase.
* cmd/geth: make ethstats URL settable through config file
* cmd/faucet: fix configuration
* cmd/geth: dedup attach tests
* eth: add comment for DefaultConfig
* eth: pass downloader.SyncMode in Config
This removes the FastSync, LightSync flags in favour of a more
general SyncMode flag.
* cmd/utils: remove jitvm flags
* cmd/utils: make mutually exclusive flag error prettier
It now reads:
Fatal: flags --dev, --testnet can't be used at the same time
* p2p: fix typo
* node: add DefaultConfig, use it for geth
* mobile: add missing NoDiscovery option
* cmd/utils: drop MakeNode
This exposed a couple of places that needed to be updated to use
node.DefaultConfig.
* node: fix typo
* eth: make fast sync the default mode
* cmd/utils: remove IPCApiFlag (unused)
* node: remove default IPC path
Set it in the frontends instead.
* cmd/geth: add --syncmode
* cmd/utils: make --ipcdisable and --ipcpath mutually exclusive
* cmd/utils: don't enable WS, HTTP when setting addr
* cmd/utils: fix --identity
There is no need to depend on the old context package now that the
minimum Go version is 1.7. The move to "context" eliminates our weird
vendoring setup. Some vendored code still uses golang.org/x/net/context
and it is now vendored in the normal way.
This change triggered new vet checks around context.WithTimeout which
didn't fire with golang.org/x/net/context.
* accounts, cmd, eth, ethdb: port logs over to new system
* ethdb: drop concept of cache distribution between dbs
* eth: fix some log nitpicks to make them nicer
Reworked the EVM gas instructions to use 64bit integers rather than
arbitrary size big ints. All gas operations, be it additions,
multiplications or divisions, are checked and guarded against 64 bit
integer overflows.
In additon, most of the protocol paramaters in the params package have
been converted to uint64 and are now constants rather than variables.
* common/math: added overflow check ops
* core: vmenv, env renamed to evm
* eth, internal/ethapi, les: unmetered eth_call and cancel methods
* core/vm: implemented big.Int pool for evm instructions
* core/vm: unexported intPool methods & verification methods
* core/vm: added memoryGasCost overflow check and test
Reworked the EVM gas instructions to use 64bit integers rather than
arbitrary size big ints. All gas operations, be it additions,
multiplications or divisions, are checked and guarded against 64 bit
integer overflows.
In additon, most of the protocol paramaters in the params package have
been converted to uint64 and are now constants rather than variables.
* common/math: added overflow check ops
* core: vmenv, env renamed to evm
* eth, internal/ethapi, les: unmetered eth_call and cancel methods
* core/vm: implemented big.Int pool for evm instructions
* core/vm: unexported intPool methods & verification methods
* core/vm: added memoryGasCost overflow check and test
Commit d3b751e accidentally deleted a crucial 'return' statement,
leading to a crash in case of an issue with node data. This change
improves the fix in PR #3591 by removing the lock entirely.
This commit implements EIP158 part 1, 2, 3 & 4
1. If an account is empty it's no longer written to the trie. An empty
account is defined as (balance=0, nonce=0, storage=0, code=0).
2. Delete an empty account if it's touched
3. An empty account is redefined as either non-existent or empty.
4. Zero value calls and zero value suicides no longer consume the 25k
reation costs.
params: moved core/config to params
Signed-off-by: Jeffrey Wilcke <jeffrey@ethereum.org>
* 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
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.
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.
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.
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.
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).
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.
