nimbus-eth1/nimbus/sync/snap/worker/db/hexary_import.nim

223 lines
7.0 KiB
Nim
Raw Normal View History

# nimbus-eth1
# Copyright (c) 2021 Status Research & Development GmbH
# Licensed under either of
# * Apache License, version 2.0, ([LICENSE-APACHE](LICENSE-APACHE) or
# http://www.apache.org/licenses/LICENSE-2.0)
# * MIT license ([LICENSE-MIT](LICENSE-MIT) or
# http://opensource.org/licenses/MIT)
# at your option. This file may not be copied, modified, or distributed
# except according to those terms.
{.push raises: [].}
import
std/[sets, tables],
eth/[common, trie/nibbles],
../../range_desc,
"."/[hexary_desc, hexary_error]
# ------------------------------------------------------------------------------
# Private debugging helpers
# ------------------------------------------------------------------------------
#proc pp(q: openArray[byte]): string =
# q.toSeq.mapIt(it.toHex(2)).join.toLowerAscii.pp(hex = true)
# ------------------------------------------------------------------------------
# Public
# ------------------------------------------------------------------------------
proc hexaryImport*(
db: HexaryTreeDbRef; ## Contains node table
recData: Blob; ## Node to add
unrefNodes: var HashSet[RepairKey]; ## Keep track of freestanding nodes
nodeRefs: var HashSet[RepairKey]; ## Ditto
): HexaryNodeReport
{.gcsafe, raises: [RlpError, KeyError].} =
## Decode a single trie item for adding to the table and add it to the
## database. Branch and exrension record links are collected.
if recData.len == 0:
return HexaryNodeReport(error: RlpNonEmptyBlobExpected)
let
nodeKey = recData.digestTo(NodeKey)
repairKey = nodeKey.to(RepairKey) # for repair table
var
rlp = recData.rlpFromBytes
blobs = newSeq[Blob](2) # temporary, cache
links: array[16,RepairKey] # reconstruct branch node
blob16: Blob # reconstruct branch node
top = 0 # count entries
rNode: RNodeRef # repair tree node
if not rlp.isList:
# Otherwise `rlp.items` will raise a `Defect`
return HexaryNodeReport(error: Rlp2Or17ListEntries)
# Collect lists of either 2 or 17 blob entries.
for w in rlp.items:
case top
of 0, 1:
if not w.isBlob:
return HexaryNodeReport(error: RlpBlobExpected)
blobs[top] = rlp.read(Blob)
of 2 .. 15:
var key: NodeKey
if not key.init(rlp.read(Blob)):
return HexaryNodeReport(error: RlpBranchLinkExpected)
# Update ref pool
links[top] = key.to(RepairKey)
unrefNodes.excl links[top] # is referenced, now (if any)
nodeRefs.incl links[top]
of 16:
if not w.isBlob:
return HexaryNodeReport(error: RlpBlobExpected)
blob16 = rlp.read(Blob)
else:
return HexaryNodeReport(error: Rlp2Or17ListEntries)
top.inc
# Verify extension data
case top
of 2:
if blobs[0].len == 0:
return HexaryNodeReport(error: RlpNonEmptyBlobExpected)
let (isLeaf, pathSegment) = hexPrefixDecode blobs[0]
if isLeaf:
rNode = RNodeRef(
kind: Leaf,
lPfx: pathSegment,
lData: blobs[1])
else:
var key: NodeKey
if not key.init(blobs[1]):
return HexaryNodeReport(error: RlpExtPathEncoding)
# Update ref pool
rNode = RNodeRef(
kind: Extension,
ePfx: pathSegment,
eLink: key.to(RepairKey))
unrefNodes.excl rNode.eLink # is referenced, now (if any)
nodeRefs.incl rNode.eLink
of 17:
for n in [0,1]:
var key: NodeKey
if not key.init(blobs[n]):
return HexaryNodeReport(error: RlpBranchLinkExpected)
# Update ref pool
links[n] = key.to(RepairKey)
unrefNodes.excl links[n] # is referenced, now (if any)
nodeRefs.incl links[n]
rNode = RNodeRef(
kind: Branch,
bLink: links,
bData: blob16)
else:
discard
# Add to database
if not db.tab.hasKey(repairKey):
db.tab[repairKey] = rNode
# Update unreferenced nodes list
if repairKey notin nodeRefs:
unrefNodes.incl repairKey # keep track of stray nodes
elif db.tab[repairKey].convertTo(Blob) != recData:
return HexaryNodeReport(error: DifferentNodeValueExists)
HexaryNodeReport(kind: some(rNode.kind))
proc hexaryImport*(
db: HexaryTreeDbRef; ## Contains node table
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
rec: NodeSpecs; ## Expected key and value data pair
): HexaryNodeReport
{.gcsafe, raises: [RlpError, KeyError].} =
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
## Ditto without referece checks but expected node key argument.
if rec.data.len == 0:
return HexaryNodeReport(error: RlpNonEmptyBlobExpected)
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
if rec.nodeKey != rec.data.digestTo(NodeKey):
return HexaryNodeReport(error: ExpectedNodeKeyDiffers)
let
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
repairKey = rec.nodeKey.to(RepairKey) # for repair table
var
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
rlp = rec.data.rlpFromBytes
blobs = newSeq[Blob](2) # temporary, cache
links: array[16,RepairKey] # reconstruct branch node
blob16: Blob # reconstruct branch node
top = 0 # count entries
rNode: RNodeRef # repair tree node
if not rlp.isList:
# Otherwise `rlp.items` will raise a `Defect`
return HexaryNodeReport(error: Rlp2Or17ListEntries)
# Collect lists of either 2 or 17 blob entries.
for w in rlp.items:
case top
of 0, 1:
if not w.isBlob:
return HexaryNodeReport(error: RlpBlobExpected)
blobs[top] = rlp.read(Blob)
of 2 .. 15:
var key: NodeKey
if not key.init(rlp.read(Blob)):
return HexaryNodeReport(error: RlpBranchLinkExpected)
# Update ref pool
links[top] = key.to(RepairKey)
of 16:
if not w.isBlob:
return HexaryNodeReport(error: RlpBlobExpected)
blob16 = rlp.read(Blob)
else:
return HexaryNodeReport(error: Rlp2Or17ListEntries)
top.inc
# Verify extension data
case top
of 2:
if blobs[0].len == 0:
return HexaryNodeReport(error: RlpNonEmptyBlobExpected)
let (isLeaf, pathSegment) = hexPrefixDecode blobs[0]
if isLeaf:
rNode = RNodeRef(
kind: Leaf,
lPfx: pathSegment,
lData: blobs[1])
else:
var key: NodeKey
if not key.init(blobs[1]):
return HexaryNodeReport(error: RlpExtPathEncoding)
# Update ref pool
rNode = RNodeRef(
kind: Extension,
ePfx: pathSegment,
eLink: key.to(RepairKey))
of 17:
for n in [0,1]:
var key: NodeKey
if not key.init(blobs[n]):
return HexaryNodeReport(error: RlpBranchLinkExpected)
# Update ref pool
links[n] = key.to(RepairKey)
rNode = RNodeRef(
kind: Branch,
bLink: links,
bData: blob16)
else:
discard
# Add to database
if not db.tab.hasKey(repairKey):
db.tab[repairKey] = rNode
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
elif db.tab[repairKey].convertTo(Blob) != rec.data:
return HexaryNodeReport(error: DifferentNodeValueExists)
HexaryNodeReport(kind: some(rNode.kind))
# ------------------------------------------------------------------------------
# End
# ------------------------------------------------------------------------------