nimbus-eth1/fluffy/eth_data/era1.nim

512 lines
16 KiB
Nim
Raw Normal View History

# fluffy
# Copyright (c) 2024 Status Research & Development GmbH
# Licensed and distributed under either of
# * MIT license (license terms in the root directory or at https://opensource.org/licenses/MIT).
# * Apache v2 license (license terms in the root directory or at https://www.apache.org/licenses/LICENSE-2.0).
# at your option. This file may not be copied, modified, or distributed except according to those terms.
{.push raises: [].}
import
std/[strformat, typetraits],
results,
stew/[endians2, io2, byteutils, arrayops],
stint,
snappy,
eth/common/eth_types_rlp,
beacon_chain/spec/beacon_time,
ssz_serialization,
ncli/e2store,
../network/history/accumulator
from nimcrypto/hash import fromHex
from ../../nimbus/utils/utils import calcTxRoot, calcReceiptsRoot
export e2store.readRecord
# Implementation of era1 file format as current described in:
# https://github.com/ethereum/go-ethereum/pull/26621
# era1 := Version | block-tuple* | other-entries* | Accumulator | BlockIndex
# block-tuple := CompressedHeader | CompressedBody | CompressedReceipts | TotalDifficulty
# block-index := starting-number | index | index | index ... | count
# CompressedHeader = { type: 0x03, data: snappyFramed(rlp(header)) }
# CompressedBody = { type: 0x04, data: snappyFramed(rlp(body)) }
# CompressedReceipts = { type: 0x05, data: snappyFramed(rlp(receipts)) }
# TotalDifficulty = { type: 0x06, data: uint256(header.total_difficulty) }
# AccumulatorRoot = { type: 0x07, data: hash_tree_root(List(HeaderRecord, 8192)) }
# BlockIndex = { type: 0x3266, data: block-index }
# Important note:
# Snappy does not give the same compression result as the implementation used
# by go-ethereum for some block headers and block bodies. This means that we
# cannot rely on the secondary verification mechanism that is based on doing the
# sha256sum of the full era1 files.
#
const
# Note: When specification is more official, these could go with the other
# E2S types.
CompressedHeader* = [byte 0x03, 0x00]
CompressedBody* = [byte 0x04, 0x00]
CompressedReceipts* = [byte 0x05, 0x00]
TotalDifficulty* = [byte 0x06, 0x00]
AccumulatorRoot* = [byte 0x07, 0x00]
E2BlockIndex* = [byte 0x66, 0x32]
MaxEra1Size* = 8192
type
BlockIndex* = object
startNumber*: uint64
offsets*: seq[int64] # Absolute positions in file
Era1* = distinct uint64 # Period of 8192 blocks (not an exact time unit)
Era1Group* = object
blockIndex*: BlockIndex
# As stated, not really a time unit but nevertheless, need the borrows
ethTimeUnit Era1
template lenu64(x: untyped): untyped =
uint64(len(x))
# Note: appendIndex, appendRecord and readIndex for BlockIndex are very similar
# to its consensus layer counter parts. The difference lies in the naming of
# slots vs block numbers and there is different behavior for the first era
# (first slot) and the last era (era1 ends at merge block).
proc appendIndex*(
f: IoHandle, startNumber: uint64, offsets: openArray[int64]
): Result[int64, string] =
let
len = offsets.len() * sizeof(int64) + 16
pos = ?f.appendHeader(E2BlockIndex, len)
?f.append(startNumber.uint64.toBytesLE())
for v in offsets:
?f.append(cast[uint64](v - pos).toBytesLE())
?f.append(offsets.lenu64().toBytesLE())
ok(pos)
proc appendRecord(f: IoHandle, index: BlockIndex): Result[int64, string] =
f.appendIndex(index.startNumber, index.offsets)
proc readBlockIndex*(f: IoHandle): Result[BlockIndex, string] =
2024-05-22 07:27:23 +00:00
var
buf: seq[byte]
pos: int
let
startPos = ?f.getFilePos().mapErr(toString)
fileSize = ?f.getFileSize().mapErr(toString)
2024-05-22 07:27:23 +00:00
header = ?f.readRecord(buf)
if header.typ != E2BlockIndex:
return err("not an index")
2024-05-22 07:27:23 +00:00
if buf.len < 16:
return err("index entry too small")
2024-05-22 07:27:23 +00:00
if buf.len mod 8 != 0:
return err("index length invalid")
let
2024-05-22 07:27:23 +00:00
blockNumber = uint64.fromBytesLE(buf.toOpenArray(pos, pos + 7))
count = buf.len div 8 - 2
pos += 8
# technically not an error, but we'll throw this sanity check in here..
if blockNumber > int32.high().uint64:
return err("fishy block number")
var offsets = newSeqUninitialized[int64](count)
for i in 0 ..< count:
let
2024-05-22 07:27:23 +00:00
offset = uint64.fromBytesLE(buf.toOpenArray(pos, pos + 7))
absolute =
if offset == 0:
0'i64
else:
# Wrapping math is actually convenient here
cast[int64](cast[uint64](startPos) + offset)
if absolute < 0 or absolute > fileSize:
2024-05-22 07:27:23 +00:00
return err("invalid offset")
offsets[i] = absolute
2024-05-22 07:27:23 +00:00
pos += 8
2024-05-22 07:27:23 +00:00
if uint64(count) != uint64.fromBytesLE(buf.toOpenArray(pos, pos + 7)):
return err("invalid count")
ok(BlockIndex(startNumber: blockNumber, offsets: offsets))
proc skipRecord*(f: IoHandle): Result[void, string] =
let header = ?readHeader(f)
if header.len > 0:
?f.setFilePos(header.len, SeekPosition.SeekCurrent).mapErr(ioErrorMsg)
ok()
func startNumber*(era: Era1): uint64 =
era * MaxEra1Size
func endNumber*(era: Era1): uint64 =
if (era + 1) * MaxEra1Size - 1'u64 >= mergeBlockNumber:
# The incomplete era just before the merge
mergeBlockNumber - 1'u64
else:
(era + 1) * MaxEra1Size - 1'u64
func endNumber*(blockIdx: BlockIndex): uint64 =
blockIdx.startNumber + blockIdx.offsets.lenu64() - 1
func era*(blockNumber: uint64): Era1 =
Era1(blockNumber div MaxEra1Size)
func offsetsLen(startNumber: uint64): int =
# For the era where the merge happens the era files only holds the blocks
# until the merge block so the offsets length needs to be adapted too.
if startNumber.era() >= mergeBlockNumber.era():
int((mergeBlockNumber) mod MaxEra1Size)
else:
MaxEra1Size
proc toCompressedRlpBytes(item: auto): seq[byte] =
snappy.encodeFramed(rlp.encode(item))
proc fromCompressedRlpBytes(bytes: openArray[byte], T: type): Result[T, string] =
try:
ok(rlp.decode(decodeFramed(bytes, checkIntegrity = false), T))
except RlpError as e:
err("Invalid Compressed RLP data" & e.msg)
proc init*(T: type Era1Group, f: IoHandle, startNumber: uint64): Result[T, string] =
discard ?f.appendHeader(E2Version, 0)
ok(
Era1Group(
blockIndex: BlockIndex(
startNumber: startNumber, offsets: newSeq[int64](startNumber.offsetsLen())
)
)
)
proc update*(
g: var Era1Group,
f: IoHandle,
blockNumber: uint64,
header, body, receipts, totalDifficulty: openArray[byte],
): Result[void, string] =
doAssert blockNumber >= g.blockIndex.startNumber
g.blockIndex.offsets[int(blockNumber - g.blockIndex.startNumber)] =
?f.appendRecord(CompressedHeader, header)
discard ?f.appendRecord(CompressedBody, body)
discard ?f.appendRecord(CompressedReceipts, receipts)
discard ?f.appendRecord(TotalDifficulty, totalDifficulty)
ok()
proc update*(
g: var Era1Group,
f: IoHandle,
blockNumber: uint64,
header: BlockHeader,
body: BlockBody,
receipts: seq[Receipt],
totalDifficulty: UInt256,
): Result[void, string] =
g.update(
f,
blockNumber,
toCompressedRlpBytes(header),
toCompressedRlpBytes(body),
toCompressedRlpBytes(receipts),
totalDifficulty.toBytesLE(),
)
proc finish*(
g: var Era1Group, f: IoHandle, accumulatorRoot: Digest, lastBlockNumber: uint64
): Result[void, string] =
let accumulatorRootPos = ?f.appendRecord(AccumulatorRoot, accumulatorRoot.data)
if lastBlockNumber > 0:
discard ?f.appendRecord(g.blockIndex)
# TODO:
# This is not something added in current specification of era1.
# But perhaps we want to be able to quickly jump to acummulator root.
# discard ? f.appendIndex(lastBlockNumber, [accumulatorRootPos])
discard accumulatorRootPos
ok()
func shortLog*(x: Digest): string =
x.data.toOpenArray(0, 3).toHex()
func era1FileName*(network: string, era: Era1, eraRoot: Digest): string =
try:
&"{network}-{era.uint64:05}-{shortLog(eraRoot)}.era1"
except ValueError as exc:
raiseAssert exc.msg
# Helpers to directly read objects from era1 files
# TODO: Might want to var parameters to avoid copying as is done for era files.
type
Era1File* = ref object
handle: Opt[IoHandle]
blockIdx*: BlockIndex
BlockTuple* =
tuple[header: BlockHeader, body: BlockBody, receipts: seq[Receipt], td: UInt256]
proc open*(_: type Era1File, name: string): Result[Era1File, string] =
var f = Opt[IoHandle].ok(?openFile(name, {OpenFlags.Read}).mapErr(ioErrorMsg))
defer:
if f.isSome():
discard closeFile(f[])
# Indices can be found at the end of each era file - we only support
# single-era files for now
?f[].setFilePos(0, SeekPosition.SeekEnd).mapErr(ioErrorMsg)
# Last in the file is the block index
let blockIdxPos = ?f[].findIndexStartOffset()
?f[].setFilePos(blockIdxPos, SeekPosition.SeekCurrent).mapErr(ioErrorMsg)
let blockIdx = ?f[].readBlockIndex()
if blockIdx.offsets.len() != blockIdx.startNumber.offsetsLen():
return err("Block index length invalid")
let res = Era1File(handle: f, blockIdx: blockIdx)
reset(f)
ok res
proc close*(f: Era1File) =
if f.handle.isSome():
discard closeFile(f.handle.get())
reset(f.handle)
proc skipRecord*(f: Era1File): Result[void, string] =
doAssert f[].handle.isSome()
f[].handle.get().skipRecord()
proc getBlockHeader(f: Era1File): Result[BlockHeader, string] =
var bytes: seq[byte]
let header = ?f[].handle.get().readRecord(bytes)
if header.typ != CompressedHeader:
return err("Invalid era file: didn't find block header at index position")
fromCompressedRlpBytes(bytes, BlockHeader)
proc getBlockBody(f: Era1File): Result[BlockBody, string] =
var bytes: seq[byte]
let header = ?f[].handle.get().readRecord(bytes)
if header.typ != CompressedBody:
return err("Invalid era file: didn't find block body at index position")
fromCompressedRlpBytes(bytes, BlockBody)
proc getReceipts(f: Era1File): Result[seq[Receipt], string] =
var bytes: seq[byte]
let header = ?f[].handle.get().readRecord(bytes)
if header.typ != CompressedReceipts:
return err("Invalid era file: didn't find receipts at index position")
fromCompressedRlpBytes(bytes, seq[Receipt])
proc getTotalDifficulty(f: Era1File): Result[UInt256, string] =
var bytes: seq[byte]
let header = ?f[].handle.get().readRecord(bytes)
if header.typ != TotalDifficulty:
return err("Invalid era file: didn't find total difficulty at index position")
if bytes.len != 32:
return err("Invalid total difficulty length")
ok(UInt256.fromBytesLE(bytes))
Consolidate block type for block processing (#2325) This PR consolidates the split header-body sequences into a single EthBlock sequence and cleans up the fallout from that which significantly reduces block processing overhead during import thanks to less garbage collection and fewer copies of things all around. Notably, since the number of headers must always match the number of bodies, we also get rid of a pointless degree of freedom that in the future could introduce unnecessary bugs. * only read header and body from era file * avoid several unnecessary copies along the block processing way * simplify signatures, cleaning up unused arguemnts and returns * use `stew/assign2` in a few strategic places where the generated nim assignent is slow and add a few `move` to work around poor analysis in nim 1.6 (will need to be revisited for 2.0) ``` stats-20240607_2223-a814aa0b.csv vs stats-20240608_0714-21c1d0a9.csv bps_x bps_y tps_x tps_y bpsd tpsd timed block_number (498305, 713245] 1,540.52 1,809.73 2,361.58 2775.340189 17.63% 17.63% -14.92% (713245, 928185] 730.36 865.26 1,715.90 2028.973852 18.01% 18.01% -15.21% (928185, 1143126] 663.03 789.10 2,529.26 3032.490771 19.79% 19.79% -16.28% (1143126, 1358066] 393.46 508.05 2,152.50 2777.578119 29.13% 29.13% -22.50% (1358066, 1573007] 370.88 440.72 2,351.31 2791.896052 18.81% 18.81% -15.80% (1573007, 1787947] 283.65 335.11 2,068.93 2441.373402 17.60% 17.60% -14.91% (1787947, 2002888] 287.29 342.11 2,078.39 2474.179448 18.99% 18.99% -15.91% (2002888, 2217828] 293.38 343.16 2,208.83 2584.77457 17.16% 17.16% -14.61% (2217828, 2432769] 140.09 167.86 1,081.87 1296.336926 18.82% 18.82% -15.80% blocks: 1934464, baseline: 3h13m1s, contender: 2h43m47s bpsd (mean): 19.55% tpsd (mean): 19.55% Time (total): -29m13s, -15.14% ```
2024-06-09 14:32:20 +00:00
proc getNextEthBlock*(f: Era1File): Result[EthBlock, string] =
doAssert not isNil(f) and f[].handle.isSome
var
header = ?getBlockHeader(f)
body = ?getBlockBody(f)
?skipRecord(f) # receipts
?skipRecord(f) # totalDifficulty
ok(EthBlock.init(move(header), move(body)))
proc getEthBlock*(f: Era1File, blockNumber: uint64): Result[EthBlock, string] =
doAssert not isNil(f) and f[].handle.isSome
doAssert(
blockNumber >= f[].blockIdx.startNumber and blockNumber <= f[].blockIdx.endNumber,
"Wrong era1 file for selected block number",
)
let pos = f[].blockIdx.offsets[blockNumber - f[].blockIdx.startNumber]
?f[].handle.get().setFilePos(pos, SeekPosition.SeekBegin).mapErr(ioErrorMsg)
getNextEthBlock(f)
proc getNextBlockTuple*(f: Era1File): Result[BlockTuple, string] =
doAssert not isNil(f) and f[].handle.isSome
let
blockHeader = ?getBlockHeader(f)
blockBody = ?getBlockBody(f)
receipts = ?getReceipts(f)
totalDifficulty = ?getTotalDifficulty(f)
ok((blockHeader, blockBody, receipts, totalDifficulty))
proc getBlockTuple*(f: Era1File, blockNumber: uint64): Result[BlockTuple, string] =
doAssert not isNil(f) and f[].handle.isSome
doAssert(
blockNumber >= f[].blockIdx.startNumber and blockNumber <= f[].blockIdx.endNumber,
"Wrong era1 file for selected block number",
)
let pos = f[].blockIdx.offsets[blockNumber - f[].blockIdx.startNumber]
?f[].handle.get().setFilePos(pos, SeekPosition.SeekBegin).mapErr(ioErrorMsg)
getNextBlockTuple(f)
proc getBlockHeader*(f: Era1File, blockNumber: uint64): Result[BlockHeader, string] =
doAssert not isNil(f) and f[].handle.isSome
doAssert(
blockNumber >= f[].blockIdx.startNumber and blockNumber <= f[].blockIdx.endNumber,
"Wrong era1 file for selected block number",
)
let pos = f[].blockIdx.offsets[blockNumber - f[].blockIdx.startNumber]
?f[].handle.get().setFilePos(pos, SeekPosition.SeekBegin).mapErr(ioErrorMsg)
getBlockHeader(f)
proc getTotalDifficulty*(f: Era1File, blockNumber: uint64): Result[UInt256, string] =
doAssert not isNil(f) and f[].handle.isSome
doAssert(
blockNumber >= f[].blockIdx.startNumber and blockNumber <= f[].blockIdx.endNumber,
"Wrong era1 file for selected block number",
)
let pos = f[].blockIdx.offsets[blockNumber - f[].blockIdx.startNumber]
?f[].handle.get().setFilePos(pos, SeekPosition.SeekBegin).mapErr(ioErrorMsg)
?skipRecord(f) # BlockHeader
?skipRecord(f) # BlockBody
?skipRecord(f) # Receipts
getTotalDifficulty(f)
# TODO: Should we add this perhaps in the Era1File object and grab it in open()?
proc getAccumulatorRoot*(f: Era1File): Result[Digest, string] =
# Get position of BlockIndex
?f[].handle.get().setFilePos(0, SeekPosition.SeekEnd).mapErr(ioErrorMsg)
let blockIdxPos = ?f[].handle.get().findIndexStartOffset()
# Accumulator root is 40 bytes before the BlockIndex
let accumulatorRootPos = blockIdxPos - 40 # 8 + 32
?f[].handle.get().setFilePos(accumulatorRootPos, SeekPosition.SeekCurrent).mapErr(
ioErrorMsg
)
var bytes: seq[byte]
let header = ?f[].handle.get().readRecord(bytes)
if header.typ != AccumulatorRoot:
return err("Invalid era file: didn't find accumulator root at index position")
if bytes.len != 32:
return err("invalid accumulator root")
ok(Digest(data: array[32, byte].initCopyFrom(bytes)))
proc buildAccumulator*(f: Era1File): Result[EpochAccumulatorCached, string] =
let
startNumber = f.blockIdx.startNumber
endNumber = f.blockIdx.endNumber()
var headerRecords: seq[HeaderRecord]
for blockNumber in startNumber .. endNumber:
let
blockHeader = ?f.getBlockHeader(blockNumber)
totalDifficulty = ?f.getTotalDifficulty(blockNumber)
headerRecords.add(
HeaderRecord(blockHash: blockHeader.blockHash(), totalDifficulty: totalDifficulty)
)
ok(EpochAccumulatorCached.init(headerRecords))
proc verify*(f: Era1File): Result[Digest, string] =
let
startNumber = f.blockIdx.startNumber
endNumber = f.blockIdx.endNumber()
var headerRecords: seq[HeaderRecord]
for blockNumber in startNumber .. endNumber:
let
(blockHeader, blockBody, receipts, totalDifficulty) =
?f.getBlockTuple(blockNumber)
txRoot = calcTxRoot(blockBody.transactions)
ommershHash = rlpHash(blockBody.uncles)
if blockHeader.txRoot != txRoot:
return err("Invalid transactions root")
if blockHeader.ommersHash != ommershHash:
return err("Invalid ommers hash")
if blockHeader.receiptsRoot != calcReceiptsRoot(receipts):
return err("Invalid receipts root")
headerRecords.add(
HeaderRecord(blockHash: blockHeader.blockHash(), totalDifficulty: totalDifficulty)
)
let expectedRoot = ?f.getAccumulatorRoot()
let accumulatorRoot = getEpochAccumulatorRoot(headerRecords)
if accumulatorRoot != expectedRoot:
err("Invalid accumulator root")
else:
ok(accumulatorRoot)
iterator era1BlockHeaders*(f: Era1File): BlockHeader =
let
startNumber = f.blockIdx.startNumber
endNumber = f.blockIdx.endNumber()
for blockNumber in startNumber .. endNumber:
let header = f.getBlockHeader(blockNumber).valueOr:
raiseAssert("Failed to read block header")
yield header
iterator era1BlockTuples*(f: Era1File): BlockTuple =
let
startNumber = f.blockIdx.startNumber
endNumber = f.blockIdx.endNumber()
for blockNumber in startNumber .. endNumber:
let blockTuple = f.getBlockTuple(blockNumber).valueOr:
raiseAssert("Failed to read block header")
yield blockTuple