Files
nanoreth/crates/storage/db/src/tables/mod.rs
2024-11-08 11:57:01 +00:00

561 lines
18 KiB
Rust

//! Tables and data models.
//!
//! # Overview
//!
//! This module defines the tables in reth, as well as some table-related abstractions:
//!
//! - [`codecs`] integrates different codecs into [`Encode`] and [`Decode`]
//! - [`models`](reth_db_api::models) defines the values written to tables
//!
//! # Database Tour
//!
//! TODO(onbjerg): Find appropriate format for this...
pub mod codecs;
mod raw;
pub use raw::{RawDupSort, RawKey, RawTable, RawValue, TableRawRow};
#[cfg(feature = "mdbx")]
pub(crate) mod utils;
use alloy_primitives::{Address, BlockHash, BlockNumber, TxHash, TxNumber, B256};
use reth_db_api::{
models::{
accounts::BlockNumberAddress,
blocks::{HeaderHash, StoredBlockOmmers},
storage_sharded_key::StorageShardedKey,
AccountBeforeTx, ClientVersion, CompactU256, ShardedKey, StoredBlockBodyIndices,
StoredBlockWithdrawals,
},
table::{Decode, DupSort, Encode, Table},
};
use reth_primitives::{Account, Bytecode, Header, Receipt, StorageEntry, TransactionSignedNoHash};
use reth_primitives_traits::IntegerList;
use reth_prune_types::{PruneCheckpoint, PruneSegment};
use reth_stages_types::StageCheckpoint;
use reth_trie_common::{BranchNodeCompact, StorageTrieEntry, StoredNibbles, StoredNibblesSubKey};
use serde::{Deserialize, Serialize};
use std::fmt;
/// Enum for the types of tables present in libmdbx.
#[derive(Debug, PartialEq, Eq, Copy, Clone)]
pub enum TableType {
/// key value table
Table,
/// Duplicate key value table
DupSort,
}
/// The general purpose of this is to use with a combination of Tables enum,
/// by implementing a `TableViewer` trait you can operate on db tables in an abstract way.
///
/// # Example
///
/// ```
/// use reth_db::{TableViewer, Tables};
/// use reth_db_api::table::{DupSort, Table};
///
/// struct MyTableViewer;
///
/// impl TableViewer<()> for MyTableViewer {
/// type Error = &'static str;
///
/// fn view<T: Table>(&self) -> Result<(), Self::Error> {
/// // operate on table in a generic way
/// Ok(())
/// }
///
/// fn view_dupsort<T: DupSort>(&self) -> Result<(), Self::Error> {
/// // operate on a dupsort table in a generic way
/// Ok(())
/// }
/// }
///
/// let viewer = MyTableViewer {};
///
/// let _ = Tables::Headers.view(&viewer);
/// let _ = Tables::Transactions.view(&viewer);
/// ```
pub trait TableViewer<R> {
/// The error type returned by the viewer.
type Error;
/// Calls `view` with the correct table type.
fn view_rt(&self, table: Tables) -> Result<R, Self::Error> {
table.view(self)
}
/// Operate on the table in a generic way.
fn view<T: Table>(&self) -> Result<R, Self::Error>;
/// Operate on the dupsort table in a generic way.
///
/// By default, the `view` function is invoked unless overridden.
fn view_dupsort<T: DupSort>(&self) -> Result<R, Self::Error> {
self.view::<T>()
}
}
/// Defines all the tables in the database.
#[macro_export]
macro_rules! tables {
(@bool) => { false };
(@bool $($t:tt)+) => { true };
(@view $name:ident $v:ident) => { $v.view::<$name>() };
(@view $name:ident $v:ident $_subkey:ty) => { $v.view_dupsort::<$name>() };
(@value_doc $key:ty, $value:ty) => {
concat!("[`", stringify!($value), "`]")
};
// Don't generate links if we have generics
(@value_doc $key:ty, $value:ty, $($generic:ident),*) => {
concat!("`", stringify!($value), "`")
};
($($(#[$attr:meta])* table $name:ident$(<$($generic:ident $(= $default:ty)?),*>)? { type Key = $key:ty; type Value = $value:ty; $(type SubKey = $subkey:ty;)? } )*) => {
// Table marker types.
$(
$(#[$attr])*
///
#[doc = concat!("Marker type representing a database table mapping [`", stringify!($key), "`] to ", tables!(@value_doc $key, $value, $($($generic),*)?), ".")]
$(
#[doc = concat!("\n\nThis table's `DUPSORT` subkey is [`", stringify!($subkey), "`].")]
)?
pub struct $name$(<$($generic $( = $default)?),*>)? {
_private: std::marker::PhantomData<($($($generic,)*)?)>,
}
// Ideally this implementation wouldn't exist, but it is necessary to derive `Debug`
// when a type is generic over `T: Table`. See: https://github.com/rust-lang/rust/issues/26925
impl$(<$($generic),*>)? fmt::Debug for $name$(<$($generic),*>)? {
fn fmt(&self, _: &mut fmt::Formatter<'_>) -> fmt::Result {
unreachable!("this type cannot be instantiated")
}
}
impl$(<$($generic),*>)? reth_db_api::table::Table for $name$(<$($generic),*>)?
where
$value: reth_db_api::table::Value + 'static
{
const NAME: &'static str = table_names::$name;
type Key = $key;
type Value = $value;
}
$(
impl DupSort for $name {
type SubKey = $subkey;
}
)?
)*
// Tables enum.
// NOTE: the ordering of the enum does not matter, but it is assumed that the discriminants
// start at 0 and increment by 1 for each variant (the default behavior).
// See for example `reth_db::implementation::mdbx::tx::Tx::db_handles`.
/// A table in the database.
#[derive(Clone, Copy, PartialEq, Eq, Hash)]
pub enum Tables {
$(
#[doc = concat!("The [`", stringify!($name), "`] database table.")]
$name,
)*
}
impl Tables {
/// All the tables in the database.
pub const ALL: &'static [Self] = &[$(Self::$name,)*];
/// The number of tables in the database.
pub const COUNT: usize = Self::ALL.len();
/// Returns the name of the table as a string.
pub const fn name(&self) -> &'static str {
match self {
$(
Self::$name => table_names::$name,
)*
}
}
/// Returns `true` if the table is a `DUPSORT` table.
pub const fn is_dupsort(&self) -> bool {
match self {
$(
Self::$name => tables!(@bool $($subkey)?),
)*
}
}
/// The type of the given table in database.
pub const fn table_type(&self) -> TableType {
if self.is_dupsort() {
TableType::DupSort
} else {
TableType::Table
}
}
/// Allows to operate on specific table type
pub fn view<T, R>(&self, visitor: &T) -> Result<R, T::Error>
where
T: ?Sized + TableViewer<R>,
{
match self {
$(
Self::$name => tables!(@view $name visitor $($subkey)?),
)*
}
}
}
impl fmt::Debug for Tables {
#[inline]
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str(self.name())
}
}
impl fmt::Display for Tables {
#[inline]
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
self.name().fmt(f)
}
}
impl std::str::FromStr for Tables {
type Err = String;
fn from_str(s: &str) -> Result<Self, Self::Err> {
match s {
$(
table_names::$name => Ok(Self::$name),
)*
s => Err(format!("unknown table: {s:?}")),
}
}
}
// Need constants to match on in the `FromStr` implementation.
#[allow(non_upper_case_globals)]
mod table_names {
$(
pub(super) const $name: &'static str = stringify!($name);
)*
}
/// Maps a run-time [`Tables`] enum value to its corresponding compile-time [`Table`] type.
///
/// This is a simpler alternative to [`TableViewer`].
///
/// # Examples
///
/// ```
/// use reth_db::{Tables, tables_to_generic};
/// use reth_db_api::table::Table;
///
/// let table = Tables::Headers;
/// let result = tables_to_generic!(table, |GenericTable| <GenericTable as Table>::NAME);
/// assert_eq!(result, table.name());
/// ```
#[macro_export]
macro_rules! tables_to_generic {
($table:expr, |$generic_name:ident| $e:expr) => {
match $table {
$(
Tables::$name => {
use $crate::tables::$name as $generic_name;
$e
},
)*
}
};
}
};
}
tables! {
/// Stores the header hashes belonging to the canonical chain.
table CanonicalHeaders {
type Key = BlockNumber;
type Value = HeaderHash;
}
/// Stores the total difficulty from a block header.
table HeaderTerminalDifficulties {
type Key = BlockNumber;
type Value = CompactU256;
}
/// Stores the block number corresponding to a header.
table HeaderNumbers {
type Key = BlockHash;
type Value = BlockNumber;
}
/// Stores header bodies.
table Headers<H = Header> {
type Key = BlockNumber;
type Value = H;
}
/// Stores block indices that contains indexes of transaction and the count of them.
///
/// More information about stored indices can be found in the [`StoredBlockBodyIndices`] struct.
table BlockBodyIndices {
type Key = BlockNumber;
type Value = StoredBlockBodyIndices;
}
/// Stores the uncles/ommers of the block.
table BlockOmmers {
type Key = BlockNumber;
type Value = StoredBlockOmmers;
}
/// Stores the block withdrawals.
table BlockWithdrawals {
type Key = BlockNumber;
type Value = StoredBlockWithdrawals;
}
/// Canonical only Stores the transaction body for canonical transactions.
table Transactions<T = TransactionSignedNoHash> {
type Key = TxNumber;
type Value = T;
}
/// Stores the mapping of the transaction hash to the transaction number.
table TransactionHashNumbers {
type Key = TxHash;
type Value = TxNumber;
}
/// Stores the mapping of transaction number to the blocks number.
///
/// The key is the highest transaction ID in the block.
table TransactionBlocks {
type Key = TxNumber;
type Value = BlockNumber;
}
/// Canonical only Stores transaction receipts.
table Receipts {
type Key = TxNumber;
type Value = Receipt;
}
/// Stores all smart contract bytecodes.
/// There will be multiple accounts that have same bytecode
/// So we would need to introduce reference counter.
/// This will be small optimization on state.
table Bytecodes {
type Key = B256;
type Value = Bytecode;
}
/// Stores the current state of an [`Account`].
table PlainAccountState {
type Key = Address;
type Value = Account;
}
/// Stores the current value of a storage key.
table PlainStorageState {
type Key = Address;
type Value = StorageEntry;
type SubKey = B256;
}
/// Stores pointers to block changeset with changes for each account key.
///
/// Last shard key of the storage will contain `u64::MAX` `BlockNumber`,
/// this would allows us small optimization on db access when change is in plain state.
///
/// Imagine having shards as:
/// * `Address | 100`
/// * `Address | u64::MAX`
///
/// What we need to find is number that is one greater than N. Db `seek` function allows us to fetch
/// the shard that equal or more than asked. For example:
/// * For N=50 we would get first shard.
/// * for N=150 we would get second shard.
/// * If max block number is 200 and we ask for N=250 we would fetch last shard and
/// know that needed entry is in `AccountPlainState`.
/// * If there were no shard we would get `None` entry or entry of different storage key.
///
/// Code example can be found in `reth_provider::HistoricalStateProviderRef`
table AccountsHistory {
type Key = ShardedKey<Address>;
type Value = BlockNumberList;
}
/// Stores pointers to block number changeset with changes for each storage key.
///
/// Last shard key of the storage will contain `u64::MAX` `BlockNumber`,
/// this would allows us small optimization on db access when change is in plain state.
///
/// Imagine having shards as:
/// * `Address | StorageKey | 100`
/// * `Address | StorageKey | u64::MAX`
///
/// What we need to find is number that is one greater than N. Db `seek` function allows us to fetch
/// the shard that equal or more than asked. For example:
/// * For N=50 we would get first shard.
/// * for N=150 we would get second shard.
/// * If max block number is 200 and we ask for N=250 we would fetch last shard and
/// know that needed entry is in `StoragePlainState`.
/// * If there were no shard we would get `None` entry or entry of different storage key.
///
/// Code example can be found in `reth_provider::HistoricalStateProviderRef`
table StoragesHistory {
type Key = StorageShardedKey;
type Value = BlockNumberList;
}
/// Stores the state of an account before a certain transaction changed it.
/// Change on state can be: account is created, selfdestructed, touched while empty
/// or changed balance,nonce.
table AccountChangeSets {
type Key = BlockNumber;
type Value = AccountBeforeTx;
type SubKey = Address;
}
/// Stores the state of a storage key before a certain transaction changed it.
/// If [`StorageEntry::value`] is zero, this means storage was not existing
/// and needs to be removed.
table StorageChangeSets {
type Key = BlockNumberAddress;
type Value = StorageEntry;
type SubKey = B256;
}
/// Stores the current state of an [`Account`] indexed with `keccak256Address`
/// This table is in preparation for merklization and calculation of state root.
/// We are saving whole account data as it is needed for partial update when
/// part of storage is changed. Benefit for merklization is that hashed addresses are sorted.
table HashedAccounts {
type Key = B256;
type Value = Account;
}
/// Stores the current storage values indexed with `keccak256Address` and
/// hash of storage key `keccak256key`.
/// This table is in preparation for merklization and calculation of state root.
/// Benefit for merklization is that hashed addresses/keys are sorted.
table HashedStorages {
type Key = B256;
type Value = StorageEntry;
type SubKey = B256;
}
/// Stores the current state's Merkle Patricia Tree.
table AccountsTrie {
type Key = StoredNibbles;
type Value = BranchNodeCompact;
}
/// From HashedAddress => NibblesSubKey => Intermediate value
table StoragesTrie {
type Key = B256;
type Value = StorageTrieEntry;
type SubKey = StoredNibblesSubKey;
}
/// Stores the transaction sender for each canonical transaction.
/// It is needed to speed up execution stage and allows fetching signer without doing
/// transaction signed recovery
table TransactionSenders {
type Key = TxNumber;
type Value = Address;
}
/// Stores the highest synced block number and stage-specific checkpoint of each stage.
table StageCheckpoints {
type Key = StageId;
type Value = StageCheckpoint;
}
/// Stores arbitrary data to keep track of a stage first-sync progress.
table StageCheckpointProgresses {
type Key = StageId;
type Value = Vec<u8>;
}
/// Stores the highest pruned block number and prune mode of each prune segment.
table PruneCheckpoints {
type Key = PruneSegment;
type Value = PruneCheckpoint;
}
/// Stores the history of client versions that have accessed the database with write privileges by unix timestamp in seconds.
table VersionHistory {
type Key = u64;
type Value = ClientVersion;
}
/// Stores generic chain state info, like the last finalized block.
table ChainState {
type Key = ChainStateKey;
type Value = BlockNumber;
}
}
/// Keys for the `ChainState` table.
#[derive(Ord, Clone, Eq, PartialOrd, PartialEq, Debug, Deserialize, Serialize, Hash)]
pub enum ChainStateKey {
/// Last finalized block key
LastFinalizedBlock,
/// Last finalized block key
LastSafeBlockBlock,
}
impl Encode for ChainStateKey {
type Encoded = [u8; 1];
fn encode(self) -> Self::Encoded {
match self {
Self::LastFinalizedBlock => [0],
Self::LastSafeBlockBlock => [1],
}
}
}
impl Decode for ChainStateKey {
fn decode(value: &[u8]) -> Result<Self, reth_db_api::DatabaseError> {
match value {
[0] => Ok(Self::LastFinalizedBlock),
[1] => Ok(Self::LastSafeBlockBlock),
_ => Err(reth_db_api::DatabaseError::Decode),
}
}
}
// Alias types.
/// List with transaction numbers.
pub type BlockNumberList = IntegerList;
/// Encoded stage id.
pub type StageId = String;
#[cfg(test)]
mod tests {
use super::*;
use std::str::FromStr;
#[test]
fn parse_table_from_str() {
for table in Tables::ALL {
assert_eq!(format!("{table:?}"), table.name());
assert_eq!(table.to_string(), table.name());
assert_eq!(Tables::from_str(table.name()).unwrap(), *table);
}
}
}