When MatcherSession encounters an error, it attempts to close the session.
Closing waits for all goroutines to finish, including the 'distributor'. However, the
distributor will not exit until all requests have returned.
This patch fixes the issue by delivering the (empty) result to the distributor
before calling Close().
* cmd/evm: improve flags handling
This fixes some issues with flags in cmd/evm. The supported flags did not
actually show up in help output because they weren't categorized. I'm also
adding the VM-related flags to the run command here so they can be given
after the subcommand name. So it can be run like this now:
./evm run --code 6001 --debug
* cmd/evm: enable all forks by default in run command
The default genesis was just empty with no forks at all, which is annoying because
contracts will be relying on opcodes introduced in a fork. So this changes the default to
have all forks enabled.
* core/asm: fix some issues in the assembler
This fixes minor bugs in the old assembler:
- It is now possible to have comments on the same line as an instruction.
- Errors for invalid numbers in the jump instruction are reported better
- Line numbers in errors were off by one
* rlp/rlpgen: remove build tag
This tag was supposed to prevent unstable output when types reference each other. Imagine
there are two struct types A and B, where a reference to type B is in A. If I run rlpgen
on type B first, and then on type A, the generator will see the B.EncodeRLP method and
call it. However, if I run rlpgen on type A first, it will inline the encoding of B.
The solution I chose for the initial release of rlpgen was to just ignore methods
generated by rlpgen using a build tag. But there is a problem with this: if any code in
the package calls EncodeRLP explicitly, the package can't be loaded without errors anymore
in rlpgen, because the loader ignores it. Would be nice if there was a way to just make it
ignore invalid functions during type checking (they're not necessary for rlpgen), but
golang.org/x/tools/go/packages does not provide a way of ignoring them.
Luckily, the types we use rlpgen with do not reference each other right now, so we can
just remove the build tags for now.
This change includes a lot of things, listed below.
### Split up interfaces, write vs read
The interfaces have been split up into one write-interface and one read-interface, with `Snapshot` being the gateway from write to read. This simplifies the semantics _a lot_.
Example of splitting up an interface into one readonly 'snapshot' part, and one updatable writeonly part:
```golang
type MeterSnapshot interface {
Count() int64
Rate1() float64
Rate5() float64
Rate15() float64
RateMean() float64
}
// Meters count events to produce exponentially-weighted moving average rates
// at one-, five-, and fifteen-minutes and a mean rate.
type Meter interface {
Mark(int64)
Snapshot() MeterSnapshot
Stop()
}
```
### A note about concurrency
This PR makes the concurrency model clearer. We have actual meters and snapshot of meters. The `meter` is the thing which can be accessed from the registry, and updates can be made to it.
- For all `meters`, (`Gauge`, `Timer` etc), it is assumed that they are accessed by different threads, making updates. Therefore, all `meters` update-methods (`Inc`, `Add`, `Update`, `Clear` etc) need to be concurrency-safe.
- All `meters` have a `Snapshot()` method. This method is _usually_ called from one thread, a backend-exporter. But it's fully possible to have several exporters simultaneously: therefore this method should also be concurrency-safe.
TLDR: `meter`s are accessible via registry, all their methods must be concurrency-safe.
For all `Snapshot`s, it is assumed that an individual exporter-thread has obtained a `meter` from the registry, and called the `Snapshot` method to obtain a readonly snapshot. This snapshot is _not_ guaranteed to be concurrency-safe. There's no need for a snapshot to be concurrency-safe, since exporters should not share snapshots.
Note, though: that by happenstance a lot of the snapshots _are_ concurrency-safe, being unmutable minimal representations of a value. Only the more complex ones are _not_ threadsafe, those that lazily calculate things like `Variance()`, `Mean()`.
Example of how a background exporter typically works, obtaining the snapshot and sequentially accessing the non-threadsafe methods in it:
```golang
ms := metric.Snapshot()
...
fields := map[string]interface{}{
"count": ms.Count(),
"max": ms.Max(),
"mean": ms.Mean(),
"min": ms.Min(),
"stddev": ms.StdDev(),
"variance": ms.Variance(),
```
TLDR: `snapshots` are not guaranteed to be concurrency-safe (but often are).
### Sample changes
I also changed the `Sample` type: previously, it iterated the samples fully every time `Mean()`,`Sum()`, `Min()` or `Max()` was invoked. Since we now have readonly base data, we can just iterate it once, in the constructor, and set all four values at once.
The same thing has been done for runtimehistogram.
### ResettingTimer API
Back when ResettingTImer was implemented, as part of https://github.com/ethereum/go-ethereum/pull/15910, Anton implemented a `Percentiles` on the new type. However, the method did not conform to the other existing types which also had a `Percentiles`.
1. The existing ones, on input, took `0.5` to mean `50%`. Anton used `50` to mean `50%`.
2. The existing ones returned `float64` outputs, thus interpolating between values. A value-set of `0, 10`, at `50%` would return `5`, whereas Anton's would return either `0` or `10`.
This PR removes the 'new' version, and uses only the 'legacy' percentiles, also for the ResettingTimer type.
The resetting timer snapshot was also defined so that it would expose the internal values. This has been removed, and getters for `Max, Min, Mean` have been added instead.
### Unexport types
A lot of types were exported, but do not need to be. This PR unexports quite a lot of them.
On ACD 163, it was agreed to bump the target and max blob values from `2/4` to `3/6` for future devnets until we could decide on final mainnet number. This change contains said update, making master pass all the hive tests. The final decision for mainnet cancun is still to be made.
---------
Co-authored-by: Felix Lange <fjl@twurst.com>
* core/forkid: skip genesis forks by time
* core/forkid: add comment about skipping non-zero fork times
* core/forkid: skip all time based forks in genesis using loop
* core/forkid: simplify logic for dropping time-based forks
This chang creates a GaugeInfo metrics type for registering informational (textual) metrics, e.g. geth version number. It also improves the testing for backend-exporters, and uses a shared subpackage in 'internal' to provide sample datasets and ordered registry.
Implements #21783
---------
Co-authored-by: Martin Holst Swende <martin@swende.se>
This changes implements faster post-selfdestruct iteration of storage slots for deletion, by using snapshot-storage+stacktrie to recover the trienodes to be deleted. This mechanism is only implemented for path-based schema.
For hash-based schema, the entire post-selfdestruct storage iteration is skipped, with this change, since hash-based does not actually perform deletion anyway.
---------
Co-authored-by: Martin Holst Swende <martin@swende.se>
This PR makes EIP-4788 work in the engine API and miner. It also fixes some bugs related to
EIP-4844 block processing and mining. Changes in detail:
- Header.BeaconRoot has been renamed to ParentBeaconRoot.
- The engine API now implements forkchoiceUpdatedV3
- newPayloadV3 method has been updated with the parentBeaconBlockRoot parameter
- beacon root is now applied to new blocks in miner
- For EIP-4844, block creation now updates the blobGasUsed field of the header
Just some minor optimizations I figured out a while ago. By using ReadBytes instead of
Bytes on the rlp stream, we can save the allocation of a temporary buffer for the typed tx
payload.
If kind == rlp.Byte, the size reported by Stream.Kind will be zero, but we need a buffer
of size 1 for ReadBytes. Since typed txs always have to be longer than 1 byte, we can just
return an error for kind == rlp.Byte.
There is a also a small change for Log: since the first three fields of Log are the ones that
should appear in the canon encoding, we can simply ignore the remaining fields via
struct tag. Doing this removes an indirection through the rlpLog type.
---------
Co-authored-by: Martin Holst Swende <martin@swende.se>
This change implements "EIP 4788 : Beacon block root in the EVM". It implements version-2 of EPI-4788, main difference being that the contract is an actual contract rather than a precompile, as in #27289.
Currently, we trigger the logic to (un)index transactions when the node receives a new
block. However, in some cases the node may not receive new blocks (eg, when the Geth node
is configured without peer discovery, or when it acts as an RPC node for historical-only
data).
In these situations, the Geth node user may not have previously configured txlookuplimit
(i.e. the default of around one year), but later realizes they need to index all
historical blocks. However, adding txlookuplimit=0 and restarting geth has no effect. This
change makes it check for required indexing work once, on startup, to fix the issue.
Co-authored-by: Martin Holst Swende <martin@swende.se>
This changes the forkID calculation to ignore time-based forks that occurred before the
genesis block. It's supposed to be done this way because the spec says:
> If a chain is configured to start with a non-Frontier ruleset already in its genesis, that is NOT considered a fork.
This PR removes the newly added txpool.Transaction wrapper type, and instead adds a way
of keeping the blob sidecar within types.Transaction. It's better this way because most
code in go-ethereum does not care about blob transactions, and probably never will. This
will start mattering especially on the client side of RPC, where all APIs are based on
types.Transaction. Users need to be able to use the same signing flows they already
have.
However, since blobs are only allowed in some places but not others, we will now need to
add checks to avoid creating invalid blocks. I'm still trying to figure out the best place
to do some of these. The way I have it currently is as follows:
- In block validation (import), txs are verified not to have a blob sidecar.
- In miner, we strip off the sidecar when committing the transaction into the block.
- In TxPool validation, txs must have a sidecar to be added into the blobpool.
- Note there is a special case here: when transactions are re-added because of a chain
reorg, we cannot use the transactions gathered from the old chain blocks as-is,
because they will be missing their blobs. This was previously handled by storing the
blobs into the 'blobpool limbo'. The code has now changed to store the full
transaction in the limbo instead, but it might be confusing for code readers why we're
not simply adding the types.Transaction we already have.
Code changes summary:
- txpool.Transaction removed and all uses replaced by types.Transaction again
- blobpool now stores types.Transaction instead of defining its own blobTx format for storage
- the blobpool limbo now stores types.Transaction instead of storing only the blobs
- checks to validate the presence/absence of the blob sidecar added in certain critical places
The Go authors updated golang/x/ext to change the function signature of the slices sort method.
It's an entire shitshow now because x/ext is not tagged, so everyone's codebase just
picked a new version that some other dep depends on, causing our code to fail building.
This PR updates the dep on our code too and does all the refactorings to follow upstream...
