* rpc, node: refactor request validation and add jwt validation
* node, rpc: fix error message, ignore engine api in RegisterAPIs
* node: make authenticated port configurable
* eth/catalyst: enable unauthenticated version of engine api
* node: rework obtainjwtsecret (backport later)
* cmd/geth: added auth port flag
* node: happy lint, happy life
* node: refactor authenticated api
Modifies the authentication mechanism to use default values
* node: trim spaces and newline away from secret
Co-authored-by: Marius van der Wijden <m.vanderwijden@live.de>
Previously, Ctrl-C (SIGINT) was ignored during JS execution, so it was not
possible to get out of infinite loops in the console. With this change,
Ctrl-C now interrupts JS.
Fixes#23344
Co-authored-by: Sina Mahmoodi <itz.s1na@gmail.com>
Co-authored-by: Felix Lange <fjl@twurst.com>
This PR reduces the amount of work we do when answering header queries, e.g. when a peer
is syncing from us.
For some items, e.g block bodies, when we read the rlp-data from database, we plug it
directly into the response package. We didn't do that for headers, but instead read
headers-rlp, decode to types.Header, and re-encode to rlp. This PR changes that to keep it
in RLP-form as much as possible. When a node is syncing from us, it typically requests 192
contiguous headers. On master it has the following effect:
- For headers not in ancient: 2 db lookups. One for translating hash->number (even though
the request is by number), and another for reading by hash (this latter one is sometimes
cached).
- For headers in ancient: 1 file lookup/syscall for translating hash->number (even though
the request is by number), and another for reading the header itself. After this, it
also performes a hashing of the header, to ensure that the hash is what it expected. In
this PR, I instead move the logic for "give me a sequence of blocks" into the lower
layers, where the database can determine how and what to read from leveldb and/or
ancients.
There are basically four types of requests; three of them are improved this way. The
fourth, by hash going backwards, is more tricky to optimize. However, since we know that
the gap is 0, we can look up by the parentHash, and stlil shave off all the number->hash
lookups.
The gapped collection can be optimized similarly, as a follow-up, at least in three out of
four cases.
Co-authored-by: Felix Lange <fjl@twurst.com>
This PR fixes two problems in devp2p tests (and through them, hive).
- Make the output more detailed about what is returned (always print packet kind).
- Allow Ping response to unsolicited findnode.
Without this PR, nethermind fails a hive protocol test, and I misinterpreted the result (NethermindEth/nethermind#3617). Ergo, the output was not fool-proof.
Debugging recent geth failures in hive, it took a while to realize that it's because
geth doesn't support eth/65 any longer. This PR makes such failures a bit more
easy to figure out.
This PR fixes a problem which arises on clique networks when there is a network stall. Previously, the worker packages were tracked, even if the sealing engine decided not to seal the block (due to clique rules about recent signing). These tracked-but-not-sealed blocks kept building up in memory.
This PR changes the situation so the sealing engine instead returns an error, and the worker can thus un-track the package.
When we map a file for generating the DAG, we do a simple truncate to e.g. 1Gb. This is fine, even if we have nowhere near 1Gb disk available, as the actual file doesn't take up the full 1Gb, merely a few bytes. When we start generating into it, however, it eventually crashes with a unexpected fault address .
This change fixes it (on linux systems) by using the Fallocate syscall, which preallocates suffcient space on disk to avoid that situation.
Co-authored-by: Felix Lange <fjl@twurst.com>
Some benchmarks in eth/filters were not good: they weren't reproducible, relying on geth chaindata to be present.
Another one was rejected because the receipt was lacking a backing transcation.
The p2p simulation benchmark had a lot of the warnings below, due to the framework calling both
Stop() and Close(). Apparently, the simulated adapter is the only implementation which has a Close(),
and there is no need to call both Stop and Close on it.
This PR adds a new accessor method to the freezer database. This new view offers a consistent interface, guaranteeing that all individual tables (headers, bodies etc) are all on the same number, and that this number is not changes (added/truncated) while the operation is performing.
This removes some code:
- The clique engine calculated the snapshot twice when verifying headers/blocks.
- The method GetBlockHashesFromHash in Header/Block/Lightchain was only used by tests. It
is now removed from the API.
- The method GetTdByHash internally looked up the number before calling GetTd(hash, num).
In many cases, callers already had the number, and used this method just because it has a
shorter name. I have removed the method to make the API surface smaller.
This change removes misuses of sync.WaitGroup in BlockChain. Before this change,
block insertion modified the WaitGroup counter in order to ensure that Stop would wait
for pending operations to complete. This was racy and could even lead to crashes
if Stop was called at an unfortunate time. The issue is resolved by adding a specialized
'closable' mutex, which prevents chain modifications after stopping while also
synchronizing writers with each other.
Co-authored-by: Felix Lange <fjl@twurst.com>
This PR fixes an issue in traceChain, where the statedb Commit operation was performed asynchronously with dereference-operations agains the underlying trie.Database instance. Due to how the reference counting works within the trie database (where parent count is recursively updated when new parents are added), doing dereferencing in the middle of Commit can cause the refcount to become wrong, leading to an inconsistent state.
This was fixed by doing Commit/Deref from the same routine.
This PR implements a new debug method, which I've talked briefly about to some other client developers. It allows the caller to obtain the intermediate state roots for a block (which might be either a canon block or a 'bad' block).
* ethclient/gethclient: fix flaky test (due to map key ordering)
* accounts/keystore: fix test failing due to rand collision due to low time resolution on windows
This change is a rewrite of the freezer code.
When writing ancient chain data to the freezer, the previous version first encoded each
individual item to a temporary buffer, then wrote the buffer. For small item sizes (for
example, in the block hash freezer table), this strategy causes a lot of system calls for
writing tiny chunks of data. It also allocated a lot of temporary []byte buffers.
In the new version, we instead encode multiple items into a re-useable batch buffer, which
is then written to the file all at once. This avoids performing a system call for every
inserted item.
To make the internal batching work, the ancient database API had to be changed. While
integrating this new API in BlockChain.InsertReceiptChain, additional optimizations were
also added there.
Co-authored-by: Felix Lange <fjl@twurst.com>
This PR adds functionality to the evm t8n to calculate ethash difficulty. If the caller does not provide a currentDifficulty, but instead provides the parentTimestamp (well, semi-optional, will default to 0 if not given), and parentDifficulty, we can calculate it for him.
The caller can also provide a parentUncleHash. In most, but not all cases, the parent uncle hash also affects the formula. If no such hash is provided (or, if the empty all-zero hash is provided), it's assumed that there were no uncles.
* core/rawdb: implement sequential reads in freezer_table
* core/rawdb, ethdb: add sequential reader to db interface
* core/rawdb: lint nitpicks
* core/rawdb: fix some nitpicks
* core/rawdb: fix flaw with deferred reads not being performed
* core/rawdb: better documentation
In many cases, it's desireable to use already-signed transactions as input to the state transition, instead of having the evm sign them internally (for example to use malformed or not-yet-valid transactions). This PR adds support + docs for that feature.
This adds a check to verify that a sender-account does not have code, which means that the codehash is either `emptyCodeHash` _OR_ not present. The latter occurs IFF the sender did not previously exist, a situation which can only occur with zero cost gasprices.
This PR modifies the post-PING-send expectations to both be laxer and stricter: it doesn't care what order the packets arrive, but also verifies that exactly one PING and one PONG is returned.
Ticket #23273 found a flaw where we were unable to sign legacy-transactions
using the external signer, even if we're still on non-london network. That's
fixed in this PR.
Additionally, I found that even when supplying all parameters, it was impossible
to sign a london-transaction on an unsynched node. It's a pretty common usecase
that someone wants to sign a transaction using an unsynced 'vanilla' node,
providing all necessary data. Our setDefaults, however, insisted on checking the
current block against the config. This PR therefore adds a case, so that if both
MaxPriorityFeePerGas and MaxFeePerGas are provided, we accept them as given.
OBS This PR fixes a regression -- on current master, we are unable to sign a
london-transaction unless the node is synched, which may break scenarios where
geth (or clef) is used as a cold wallet.
Fixes#23273
This PR fixes a false positive PONG 'to' endpoint mismatch seen in hive tests:
got {IP:172.17.0.7 UDP:44025 TCP:44025}, want {IP:172.17.0.7 UDP:44025 TCP:0}
Co-authored-by: Felix Lange <fjl@twurst.com>
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.
This change increases the cache size from 64 to 256 Mb for block bodies.
Benchmarks have shown this to be one bottleneck when trying to achieve
higher download speeds.
The commit also includes a minor optimization for header inserts in package
core: previously, the presence of headers in the database was checked for
every header before writing it. With the change, if one header fails the
presence check, all subsequent headers are also assumed to be missing.
This is an improvement because in practice, the headers are almost always
missing during sync.
Martin Holst SwendeandGitHub
