Alex Sharov
7f4876880c
e3: more aggressive prune ( #780 )
2022-12-14 09:50:31 +07:00
Alex Sharov
b20afcad3c
e3: auto-remove useless files ( #779 )
2022-12-14 09:03:09 +07:00
Max Revitt
c9c1b15547
Revert "Withdrawals - Database ( #758 )" ( #778 )
...
This reverts commit 183856e788
2022-12-12 12:23:21 +00:00
hexoscott
aa99053b78
reset yielded state in txpool ( #777 )
2022-12-12 10:14:56 +00:00
ledgerwatch
d70b55bc1a
Add PeerUseless function to instruct sentry that peer should not be used for a while ( #776 )
...
Co-authored-by: Alexey Sharp <alexeysharp@Alexeys-iMac.local>
2022-12-11 22:20:33 +00:00
ledgerwatch
874c497dda
Add passive_peers parameter to the sentry to better take advantage of non-POS block gossip ( #775 )
...
Co-authored-by: Alex Sharp <alexsharp@Alexs-MacBook-Pro-2.local>
2022-12-11 14:53:19 +00:00
Alex Sharov
2f641bdeb4
e3: earlier flush ( #774 )
2022-12-10 15:53:50 +07:00
ledgerwatch
dfd0e6cbfa
[erigon3] Add iterators and tables for incremental reconstitution ( #742 )
...
Co-authored-by: Alex Sharp <alexsharp@Alexs-MacBook-Pro-2.local>
Co-authored-by: Alexey Sharp <alexeysharp@Alexeys-iMac.local>
2022-12-10 08:38:02 +00:00
hexoscott
c9a7ae152a
move to PeekBest and YieldBest methods in txpool ( #773 )
...
Following up on comments from the first iteration of this change. How
does this look?
2022-12-09 10:59:00 +00:00
hexoscott
9f72de1e9d
txpool to mark candidate txs for mining ( #772 )
...
as we're now visiting the txpool multiple times during mining we need a
way to ensure we're not trying to include the same transactions twice
and to only get new data each time. This seems a simple way to achieve
this, but if there would be a better way let me know.
2022-12-09 09:51:50 +00:00
Alex Sharov
3cf9b451e1
mdbx: runtime options are changeable in Ascedee mode ( #771 )
2022-12-07 17:10:32 +07:00
Alex Sharov
263a9e3583
mdbx: dirtySpace option expose ( #770 )
2022-12-07 16:35:55 +07:00
Alex Sharov
0c5a850177
mdbx bindings: remove txn.RawRead field ( #769 )
2022-12-07 13:08:51 +07:00
Alex Sharov
c9972dbc8f
e3: no wal support ( #768 )
2022-12-07 12:15:09 +07:00
Alex Sharov
2dcec83222
mdbx: add BeginRwAsync method ( #767 )
2022-12-07 11:19:08 +07:00
alex.sharov
bf912d4eaa
save
2022-12-07 10:31:14 +07:00
alex.sharov
3cd34cc900
Revert "save"
...
This reverts commit 00bb99d66d
2022-12-06 12:29:12 +07:00
alex.sharov
00bb99d66d
save
2022-12-06 12:28:53 +07:00
Max Revitt
183856e788
Withdrawals - Database ( #758 )
...
Adds withdrawals bucket to db.
2022-12-05 17:44:07 +00:00
Giulio rebuffo
c5f3972276
added getStatus ( #762 )
2022-12-05 03:55:18 +01:00
Alex Sharov
81be3e5d74
e3: buildFiles: return error when can ( #763 )
2022-12-04 16:19:49 +07:00
Håvard Anda Estensen
5a1726b107
Avoid leaking the popped item ( #760 )
...
Set the deleted item to its zero val so it can be garbage collected.
2022-12-04 16:19:09 +07:00
Håvard Anda Estensen
9d5f02ce76
Use hex package to convert bytes to string ( #761 )
...
From my
[benchmarks](https://github.com/estensen/benchmarks/blob/master/encoding/encoding.go )
it's a lot faster. Especially for few bytes.
```
cpu: Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
BenchmarkByteToHex/encode_bytes_to_hex_string_with_fmt.Sprintf_for_1_byte-12 12689306 89.64 ns/op 26 B/op 2 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_hex.EncodeToString_for_1_byte-12 80529291 14.74 ns/op 2 B/op 1 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_fmt.Sprintf_for_5_bytes-12 10836607 108.2 ns/op 40 B/op 2 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_hex.EncodeToString_for_5_bytes-12 42891154 28.82 ns/op 16 B/op 1 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_fmt.Sprintf_for_10_bytes-12 9706875 119.8 ns/op 48 B/op 2 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_hex.EncodeToString_for_10_bytes-12 31689613 33.88 ns/op 24 B/op 1 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_fmt.Sprintf_for_50_bytes-12 5202006 210.4 ns/op 136 B/op 2 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_hex.EncodeToString_for_50_bytes-12 10616301 109.5 ns/op 224 B/op 2 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_fmt.Sprintf_for_100_bytes-12 5614594 210.9 ns/op 136 B/op 2 allocs/op
BenchmarkByteToHex/encode_bytes_to_hex_string_with_hex.EncodeToString_for_100_bytes-12 10426776 108.6 ns/op 224 B/op 2 allocs/op
```
2022-12-04 16:17:37 +07:00
Alex Sharov
bee082f614
log experiment enabling ( #759 )
2022-12-03 12:56:12 +07:00
Alex Sharov
8a6090bcb7
e3: optimize incremental hashstate ( #756 )
2022-12-03 12:22:38 +07:00
hexoscott
6ec0a1ba7c
txpool to consider available gas in Best call ( #755 )
2022-12-02 15:32:56 +00:00
Alex Sharov
0613545530
mdbx: reproducibility of benchmarks ( #757 )
2022-12-02 18:42:35 +07:00
Alex Sharov
14ceff26c3
tests speedup ( #751 )
2022-12-01 14:45:03 +07:00
Andrew Ashikhmin
f32d6541e3
Add withdrawal operations to Engine API ( #734 )
...
See https://github.com/ethereum/execution-apis/pull/195
Prerequisite: https://github.com/ledgerwatch/interfaces/pull/134
2022-12-01 08:32:55 +01:00
Giulio rebuffo
fd11ffca3d
Added execution .proto ( #749 )
2022-11-30 11:20:21 +01:00
Alex Sharov
c63b54e470
e3: option to discard all history writes. useful to test execution performance. ( #753 )
2022-11-30 16:06:51 +07:00
Alex Sharov
7e661f184b
e3: option to discard all history writes. useful to test execution performance. ( #752 )
2022-11-30 15:33:43 +07:00
Alex Sharov
c793890fdd
grpc version up ( #750 )
2022-11-30 08:53:22 +07:00
Giulio rebuffo
2d86bff102
Sentinel interfaces ( #748 )
2022-11-29 00:01:03 +01:00
hexoscott
1e24a738bf
fix for windows tests ( #747 )
2022-11-28 16:17:27 +07:00
Giulio rebuffo
f776665b21
added BeaconBlocks table ( #746 )
2022-11-26 23:05:59 +01:00
Max Revitt
3833e2503b
hotfix(downloader): detect ipv6 only on linux #6117 ( #745 )
...
Prevent port scan warnings on Hetzner and potentially other hosting
providers by only detecting ipv6 for linux by reading from
/sys/module/ipv6/parameters/disable
2022-11-24 20:43:41 +07:00
Alex Sharov
661ef494cf
mdbx: remove deprecated cmp_suffix32 feature ( #744 )
2022-11-24 09:10:09 +07:00
Andrew Ashikhmin
f2e3c2eea3
sync.Mutex instead of RWMutex in kvcache ( #743 )
...
Follow-up to PR #730
2022-11-22 14:08:32 +01:00
awskii
229b08146b
[WIP] E4 commitment keys replace ( #684 )
...
basic implementation of replacing account and storage keys encoded in
commitment branches
2022-11-21 07:39:30 +00:00
Alex Sharov
0eb4f8388a
Downloader extract, step2 ( #739 )
2022-11-20 10:41:20 +07:00
Alex Sharov
fb0238b6e6
up some deps ( #738 )
2022-11-16 14:22:28 +03:00
Manav Darji
07fa94278f
txpool: honour nonce while sorting txs ( #737 )
...
In context of https://github.com/ledgerwatch/erigon/issues/5694 , this PR
fixes the tx ordering logic in the txpool. Also, a relevant change in
erigon is made here: https://github.com/ledgerwatch/erigon/pull/6051 .
When the mining module used to query the best transactions, the response
of transactions weren't sorted according to nonce (for same sender
address). This caused the execution to fail (except 1). The txpool
orders the transactions using this underlying `better()` and `worse()`
functions. The `better()` function didn't have a nonce check for the
pending pool, while the `worse()` function had. This PR adds a check
which would compare and put a transaction with lower nonce ahead.
Moreover, on further investigation, these checks were applied for all
types of transactions. It only makes sense compare the nonce for
transactions which has same sender account. Hence, this PR also adds an
additional check along with.
2022-11-16 10:57:17 +03:00
Giulio rebuffo
b0cda9f301
Added light client database tables ( #736 )
2022-11-15 19:59:55 +01:00
Andrew Ashikhmin
10e842853b
txpool: sync.Mutex instead of RWMutex ( #735 )
...
[simplelru](https://github.com/hashicorp/golang-lru/tree/master/simplelru )
is not thread safe. During the `Get` operation, the recentness of the
accessed item is updated, so it is not a pure read-operation. Therefore,
the mutex we need to protect the LRUs in txpool is a full mutex, not
`RLock`. See
https://github.com/ledgerwatch/erigon/issues/4679#issuecomment-1312064790
and https://github.com/ethereum/go-ethereum/pull/26164 .
Also, RWMutex has a performance overhead compared with the vanilla one
(see, for example, https://github.com/golang/go/issues/38813 ).
Kudos to Martin Swende for pointing to the issue.
2022-11-14 11:59:02 +01:00
Andrew Ashikhmin
b810bb6ee0
Update interfaces ( #733 )
...
Pick up https://github.com/ledgerwatch/interfaces/pull/131
2022-11-11 16:36:22 +01:00
Alex Sharov
9e1963e40d
e3: wal ram configure ( #732 )
2022-11-11 11:28:18 +07:00
hexoscott
875b19b2b4
switch to standard lock to help cache perf ( #730 )
2022-11-08 16:20:47 +00:00
Alex Sharov
19804fd971
e3: tmpdir ( #729 )
2022-11-07 12:16:56 +07:00
Alex Sharov
bfd3996fa4
e3: parallel miss idx ( #728 )
2022-11-07 11:42:03 +07:00
