0g-storage-node/version-meld/eth2_ssz/src/decode.rs
2024-01-04 18:05:32 +08:00

373 lines
13 KiB
Rust

use super::*;
use smallvec::{smallvec, SmallVec};
use std::cmp::Ordering;
type SmallVec8<T> = SmallVec<[T; 8]>;
pub mod impls;
/// Returned when SSZ decoding fails.
#[derive(Debug, PartialEq, Clone)]
pub enum DecodeError {
/// The bytes supplied were too short to be decoded into the specified type.
InvalidByteLength { len: usize, expected: usize },
/// The given bytes were too short to be read as a length prefix.
InvalidLengthPrefix { len: usize, expected: usize },
/// A length offset pointed to a byte that was out-of-bounds (OOB).
///
/// A bytes may be OOB for the following reasons:
///
/// - It is `>= bytes.len()`.
/// - When decoding variable length items, the 1st offset points "backwards" into the fixed
/// length items (i.e., `length[0] < BYTES_PER_LENGTH_OFFSET`).
/// - When decoding variable-length items, the `n`'th offset was less than the `n-1`'th offset.
OutOfBoundsByte { i: usize },
/// An offset points “backwards” into the fixed-bytes portion of the message, essentially
/// double-decoding bytes that will also be decoded as fixed-length.
///
/// https://notes.ethereum.org/ruKvDXl6QOW3gnqVYb8ezA?view#1-Offset-into-fixed-portion
OffsetIntoFixedPortion(usize),
/// The first offset does not point to the byte that follows the fixed byte portion,
/// essentially skipping a variable-length byte.
///
/// https://notes.ethereum.org/ruKvDXl6QOW3gnqVYb8ezA?view#2-Skip-first-variable-byte
OffsetSkipsVariableBytes(usize),
/// An offset points to bytes prior to the previous offset. Depending on how you look at it,
/// this either double-decodes bytes or makes the first offset a negative-length.
///
/// https://notes.ethereum.org/ruKvDXl6QOW3gnqVYb8ezA?view#3-Offsets-are-decreasing
OffsetsAreDecreasing(usize),
/// An offset references byte indices that do not exist in the source bytes.
///
/// https://notes.ethereum.org/ruKvDXl6QOW3gnqVYb8ezA?view#4-Offsets-are-out-of-bounds
OffsetOutOfBounds(usize),
/// A variable-length list does not have a fixed portion that is cleanly divisible by
/// `BYTES_PER_LENGTH_OFFSET`.
InvalidListFixedBytesLen(usize),
/// Some item has a `ssz_fixed_len` of zero. This is illegal.
ZeroLengthItem,
/// The given bytes were invalid for some application-level reason.
BytesInvalid(String),
/// The given union selector is out of bounds.
UnionSelectorInvalid(u8),
}
/// Performs checks on the `offset` based upon the other parameters provided.
///
/// ## Detail
///
/// - `offset`: the offset bytes (e.g., result of `read_offset(..)`).
/// - `previous_offset`: unless this is the first offset in the SSZ object, the value of the
/// previously-read offset. Used to ensure offsets are not decreasing.
/// - `num_bytes`: the total number of bytes in the SSZ object. Used to ensure the offset is not
/// out of bounds.
/// - `num_fixed_bytes`: the number of fixed-bytes in the struct, if it is known. Used to ensure
/// that the first offset doesn't skip any variable bytes.
///
/// ## References
///
/// The checks here are derived from this document:
///
/// https://notes.ethereum.org/ruKvDXl6QOW3gnqVYb8ezA?view
pub fn sanitize_offset(
offset: usize,
previous_offset: Option<usize>,
num_bytes: usize,
num_fixed_bytes: Option<usize>,
) -> Result<usize, DecodeError> {
if num_fixed_bytes.map_or(false, |fixed_bytes| offset < fixed_bytes) {
Err(DecodeError::OffsetIntoFixedPortion(offset))
} else if previous_offset.is_none()
&& num_fixed_bytes.map_or(false, |fixed_bytes| offset != fixed_bytes)
{
Err(DecodeError::OffsetSkipsVariableBytes(offset))
} else if offset > num_bytes {
Err(DecodeError::OffsetOutOfBounds(offset))
} else if previous_offset.map_or(false, |prev| prev > offset) {
Err(DecodeError::OffsetsAreDecreasing(offset))
} else {
Ok(offset)
}
}
/// Provides SSZ decoding (de-serialization) via the `from_ssz_bytes(&bytes)` method.
///
/// See `examples/` for manual implementations or the crate root for implementations using
/// `#[derive(Decode)]`.
pub trait Decode: Sized {
/// Returns `true` if this object has a fixed-length.
///
/// I.e., there are no variable length items in this object or any of it's contained objects.
fn is_ssz_fixed_len() -> bool;
/// The number of bytes this object occupies in the fixed-length portion of the SSZ bytes.
///
/// By default, this is set to `BYTES_PER_LENGTH_OFFSET` which is suitable for variable length
/// objects, but not fixed-length objects. Fixed-length objects _must_ return a value which
/// represents their length.
fn ssz_fixed_len() -> usize {
BYTES_PER_LENGTH_OFFSET
}
/// Attempts to decode `Self` from `bytes`, returning a `DecodeError` on failure.
///
/// The supplied bytes must be the exact length required to decode `Self`, excess bytes will
/// result in an error.
fn from_ssz_bytes(bytes: &[u8]) -> Result<Self, DecodeError>;
}
#[derive(Copy, Clone, Debug)]
pub struct Offset {
position: usize,
offset: usize,
}
/// Builds an `SszDecoder`.
///
/// The purpose of this struct is to split some SSZ bytes into individual slices. The builder is
/// then converted into a `SszDecoder` which decodes those values into object instances.
///
/// See [`SszDecoder`](struct.SszDecoder.html) for usage examples.
pub struct SszDecoderBuilder<'a> {
bytes: &'a [u8],
items: SmallVec8<&'a [u8]>,
offsets: SmallVec8<Offset>,
items_index: usize,
}
impl<'a> SszDecoderBuilder<'a> {
/// Instantiate a new builder that should build a `SszDecoder` over the given `bytes` which
/// are assumed to be the SSZ encoding of some object.
pub fn new(bytes: &'a [u8]) -> Self {
Self {
bytes,
items: smallvec![],
offsets: smallvec![],
items_index: 0,
}
}
/// Registers a variable-length object as the next item in `bytes`, without specifying the
/// actual type.
///
/// ## Notes
///
/// Use of this function is generally discouraged since it cannot detect if some type changes
/// from variable to fixed length.
///
/// Use `Self::register_type` wherever possible.
pub fn register_anonymous_variable_length_item(&mut self) -> Result<(), DecodeError> {
struct Anonymous;
impl Decode for Anonymous {
fn is_ssz_fixed_len() -> bool {
false
}
fn from_ssz_bytes(_bytes: &[u8]) -> Result<Self, DecodeError> {
unreachable!("Anonymous should never be decoded")
}
}
self.register_type::<Anonymous>()
}
/// Declares that some type `T` is the next item in `bytes`.
pub fn register_type<T: Decode>(&mut self) -> Result<(), DecodeError> {
self.register_type_parameterized(T::is_ssz_fixed_len(), T::ssz_fixed_len())
}
/// Declares that a type with the given parameters is the next item in `bytes`.
pub fn register_type_parameterized(
&mut self,
is_ssz_fixed_len: bool,
ssz_fixed_len: usize,
) -> Result<(), DecodeError> {
if is_ssz_fixed_len {
let start = self.items_index;
self.items_index += ssz_fixed_len;
let slice = self.bytes.get(start..self.items_index).ok_or_else(|| {
DecodeError::InvalidByteLength {
len: self.bytes.len(),
expected: self.items_index,
}
})?;
self.items.push(slice);
} else {
self.offsets.push(Offset {
position: self.items.len(),
offset: sanitize_offset(
read_offset(&self.bytes[self.items_index..])?,
self.offsets.last().map(|o| o.offset),
self.bytes.len(),
None,
)?,
});
// Push an empty slice into items; it will be replaced later.
self.items.push(&[]);
self.items_index += BYTES_PER_LENGTH_OFFSET;
}
Ok(())
}
fn finalize(&mut self) -> Result<(), DecodeError> {
if let Some(first_offset) = self.offsets.first().map(|o| o.offset) {
// Check to ensure the first offset points to the byte immediately following the
// fixed-length bytes.
match first_offset.cmp(&self.items_index) {
Ordering::Less => return Err(DecodeError::OffsetIntoFixedPortion(first_offset)),
Ordering::Greater => {
return Err(DecodeError::OffsetSkipsVariableBytes(first_offset))
}
Ordering::Equal => (),
}
// Iterate through each pair of offsets, grabbing the slice between each of the offsets.
for pair in self.offsets.windows(2) {
let a = pair[0];
let b = pair[1];
self.items[a.position] = &self.bytes[a.offset..b.offset];
}
// Handle the last offset, pushing a slice from it's start through to the end of
// `self.bytes`.
if let Some(last) = self.offsets.last() {
self.items[last.position] = &self.bytes[last.offset..]
}
} else {
// If the container is fixed-length, ensure there are no excess bytes.
if self.items_index != self.bytes.len() {
return Err(DecodeError::InvalidByteLength {
len: self.bytes.len(),
expected: self.items_index,
});
}
}
Ok(())
}
/// Finalizes the builder, returning a `SszDecoder` that may be used to instantiate objects.
pub fn build(mut self) -> Result<SszDecoder<'a>, DecodeError> {
self.finalize()?;
Ok(SszDecoder { items: self.items })
}
}
/// Decodes some slices of SSZ into object instances. Should be instantiated using
/// [`SszDecoderBuilder`](struct.SszDecoderBuilder.html).
///
/// ## Example
///
/// ```rust
/// use ssz_derive::{Encode, Decode};
/// use ssz::{Decode, Encode, SszDecoder, SszDecoderBuilder};
///
/// #[derive(PartialEq, Debug, Encode, Decode)]
/// struct Foo {
/// a: u64,
/// b: Vec<u16>,
/// }
///
/// fn ssz_decoding_example() {
/// let foo = Foo {
/// a: 42,
/// b: vec![1, 3, 3, 7]
/// };
///
/// let bytes = foo.as_ssz_bytes();
///
/// let mut builder = SszDecoderBuilder::new(&bytes);
///
/// builder.register_type::<u64>().unwrap();
/// builder.register_type::<Vec<u16>>().unwrap();
///
/// let mut decoder = builder.build().unwrap();
///
/// let decoded_foo = Foo {
/// a: decoder.decode_next().unwrap(),
/// b: decoder.decode_next().unwrap(),
/// };
///
/// assert_eq!(foo, decoded_foo);
/// }
///
/// ```
pub struct SszDecoder<'a> {
items: SmallVec8<&'a [u8]>,
}
impl<'a> SszDecoder<'a> {
/// Decodes the next item.
///
/// # Panics
///
/// Panics when attempting to decode more items than actually exist.
pub fn decode_next<T: Decode>(&mut self) -> Result<T, DecodeError> {
self.decode_next_with(|slice| T::from_ssz_bytes(slice))
}
/// Decodes the next item using the provided function.
pub fn decode_next_with<T, F>(&mut self, f: F) -> Result<T, DecodeError>
where
F: FnOnce(&'a [u8]) -> Result<T, DecodeError>,
{
f(self.items.remove(0))
}
}
/// Takes `bytes`, assuming it is the encoding for a SSZ union, and returns the union-selector and
/// the body (trailing bytes).
///
/// ## Errors
///
/// Returns an error if:
///
/// - `bytes` is empty.
/// - the union selector is not a valid value (i.e., larger than the maximum number of variants.
pub fn split_union_bytes(bytes: &[u8]) -> Result<(UnionSelector, &[u8]), DecodeError> {
let selector = bytes
.first()
.copied()
.ok_or(DecodeError::OutOfBoundsByte { i: 0 })
.and_then(UnionSelector::new)?;
let body = bytes
.get(1..)
.ok_or(DecodeError::OutOfBoundsByte { i: 1 })?;
Ok((selector, body))
}
/// Reads a `BYTES_PER_LENGTH_OFFSET`-byte length from `bytes`, where `bytes.len() >=
/// BYTES_PER_LENGTH_OFFSET`.
pub fn read_offset(bytes: &[u8]) -> Result<usize, DecodeError> {
decode_offset(bytes.get(0..BYTES_PER_LENGTH_OFFSET).ok_or_else(|| {
DecodeError::InvalidLengthPrefix {
len: bytes.len(),
expected: BYTES_PER_LENGTH_OFFSET,
}
})?)
}
/// Decode bytes as a little-endian usize, returning an `Err` if `bytes.len() !=
/// BYTES_PER_LENGTH_OFFSET`.
fn decode_offset(bytes: &[u8]) -> Result<usize, DecodeError> {
let len = bytes.len();
let expected = BYTES_PER_LENGTH_OFFSET;
if len != expected {
Err(DecodeError::InvalidLengthPrefix { len, expected })
} else {
let mut array: [u8; BYTES_PER_LENGTH_OFFSET] = std::default::Default::default();
array.clone_from_slice(bytes);
Ok(u32::from_le_bytes(array) as usize)
}
}