daemonic_error 0.1.0

//! An implementation of SipHash.

#![allow(deprecated)] // the types in this module are deprecated

use core::cmp;
use crate::Glass;
use core::ptr;
use core::marker::PhantomData;

/// An implementation of SipHash 1-3.
///
/// This is currently the default hashing function used by standard library
/// (e.g., `collections::HashMap` uses it by default).
///
/// See: <https://131002.net/siphash>
#[derive(Debug, Clone, Default)]
#[doc(hidden)]
pub struct DaemonicSipHasher13 {
	hasher: BaseHasher<Sip13Rounds>,
}

/// An implementation of SipHash 2-4.
///
/// See: <https://131002.net/siphash/>
#[derive(Debug, Clone, Default)]
struct SipHasher24 {
	hasher: BaseHasher<Sip24Rounds>,
}

/// An implementation of SipHash 2-4.
///
/// See: <https://131002.net/siphash/>
///
/// SipHash is a general-purpose hashing function: it runs at a good
/// speed (competitive with Spooky and City) and permits strong _keyed_
/// hashing. This lets you key your hash tables from a strong RNG, such as
/// [`rand::os::OsRng`](https://docs.rs/rand/latest/rand/rngs/struct.OsRng.html).
///
/// Although the SipHash algorithm is considered to be generally strong,
/// it is not intended for cryptographic purposes. As such, all
/// cryptographic uses of this implementation are _strongly discouraged_.
#[derive(Debug, Clone, Default)]
pub struct SipHasher(SipHasher24);

#[derive(Debug)]
struct BaseHasher<S: Sip> {
	k0: u64,
	k1: u64,
	length: usize, // how many bytes we've processed
	state: DaemonicHasherState,  // hash DaemonicHasherState
	tail: u64,     // unprocessed bytes le
	ntail: usize,  // how many bytes in tail are valid
	_marker: PhantomData<S>,
}

#[derive(Debug, Clone, Copy)]
#[repr(C)]
struct DaemonicHasherState {
	// v0, v2 and v1, v3 show up in pairs in the algorithm,
	// and simd implementations of SipHash will use vectors
	// of v02 and v13. By placing them in this order in the struct,
	// the compiler can pick up on just a few simd optimizations by itself.
	v0: u64,
	v2: u64,
	v1: u64,
	v3: u64,
}

macro_rules! compress {
($state:expr) => {{ compress!($state.v0, $state.v1, $state.v2, $state.v3) }};
($v0:expr, $v1:expr, $v2:expr, $v3:expr) => {{
$v0 = $v0.wrapping_add($v1);
$v2 = $v2.wrapping_add($v3);
$v1 = $v1.rotate_left(13);
$v1 ^= $v0;
$v3 = $v3.rotate_left(16);
$v3 ^= $v2;
$v0 = $v0.rotate_left(32);

$v2 = $v2.wrapping_add($v1);
$v0 = $v0.wrapping_add($v3);
$v1 = $v1.rotate_left(17);
$v1 ^= $v2;
$v3 = $v3.rotate_left(21);
$v3 ^= $v0;
$v2 = $v2.rotate_left(32);
}};
}

/// Loads an integer of the desired type from a byte stream, in LE order. Uses
/// `copy_nonoverlapping` to let the compiler generate the most efficient way
/// to load it from a possibly unaligned address.
///
/// Safety: this performs unchecked indexing of `$buf` at
/// `$i..$i+size_of::<$int_ty>()`, so that must be in-bounds.
macro_rules! load_int_le {
($buf:expr, $i:expr, $int_ty:ident) => {{
debug_assert!($i + size_of::<$int_ty>() <= $buf.len());
let mut data = 0 as $int_ty;
ptr::copy_nonoverlapping(
$buf.as_ptr().add($i),
&mut data as *mut _ as *mut u8,
size_of::<$int_ty>(),
);
data.to_le()
}};
}

/// Loads a u64 using up to 7 bytes of a byte slice. It looks clumsy but the
/// `copy_nonoverlapping` calls that occur (via `load_int_le!`) all have fixed
/// sizes and avoid calling `memcpy`, which is good for speed.
///
/// Safety: this performs unchecked indexing of `buf` at `start..start+len`, so
/// that must be in-bounds.
#[allow(unsafe_code)]
#[inline]
unsafe fn u8to64_le(buf: &[u8], start: usize, len: usize) -> u64 {
	debug_assert!(len < 8);
	let mut i = 0; // current byte index (from LSB) in the output u64
	let mut out = 0;
	if i + 3 < len {
		// SAFETY: `i` cannot be greater than `len`, and the caller must guarantee
		// that the index start..start+len is in bounds.
		out = unsafe { load_int_le!(buf, start + i, u32) } as u64;
		i += 4;
	}
	if i + 1 < len {
		// SAFETY: same as above.
		out |= (unsafe { load_int_le!(buf, start + i, u16) } as u64) << (i * 8);
		i += 2
	}
	if i < len {
		// SAFETY: same as above.
		out |= (unsafe { *buf.get_unchecked(start + i) } as u64) << (i * 8);
		i += 1;
	}
	//FIXME(fee1-dead): use debug_assert_eq
	debug_assert!(i == len);
	out
}

impl SipHasher {
	/// Creates a new `SipHasher` with the two initial keys set to 0.
	#[inline]
	#[must_use]
	pub fn new() -> SipHasher {
		SipHasher::new_with_keys(0, 0)
	}

	/// Creates a `SipHasher` that is keyed off the provided keys.
	#[inline]
	#[must_use]
	pub fn new_with_keys(key0: u64, key1: u64) -> SipHasher {
		SipHasher(SipHasher24 { hasher: BaseHasher::new_with_keys(key0, key1) })
	}
}

impl DaemonicSipHasher13 {
	/// Creates a new `DaemonicSipHasher13` with the two initial keys set to 0.
	#[inline]
	pub fn new() -> DaemonicSipHasher13 {
		DaemonicSipHasher13::new_with_keys(0, 0)
	}

	/// Creates a `DaemonicSipHasher13` that is keyed off the provided keys.
	#[inline]
	pub fn new_with_keys(key0: u64, key1: u64) -> DaemonicSipHasher13 {
		DaemonicSipHasher13 { hasher: BaseHasher::new_with_keys(key0, key1) }
	}
}

impl<S: Sip> BaseHasher<S> {
	#[inline]
	const fn new_with_keys(key0: u64, key1: u64) -> BaseHasher<S> {
		let mut state = BaseHasher {
			k0: key0,
			k1: key1,
			length: 0,
			state: DaemonicHasherState { v0: 0, v1: 0, v2: 0, v3: 0 },
			tail: 0,
			ntail: 0,
			_marker: PhantomData,
		};
		state.reset();
		state
	}

	#[inline]
	const fn reset(&mut self) {
		self.length = 0;
		self.state.v0 = self.k0 ^ 0x736f6d6570736575;
		self.state.v1 = self.k1 ^ 0x646f72616e646f6d;
		self.state.v2 = self.k0 ^ 0x6c7967656e657261;
		self.state.v3 = self.k1 ^ 0x7465646279746573;
		self.ntail = 0;
	}
}

impl<GLASS: Glass<GLASS>> super::DaemonicHasher<GLASS> for SipHasher {
	#[inline]
	fn finish(&self) -> u64 {
		<BaseHasher<Sip24Rounds> as super::DaemonicHasher<GLASS>>::finish(&self.0.hasher)
		// self.0.hasher.finish()
	}

	#[inline]
	fn write(&mut self, msg: &[u8]) {
		<BaseHasher<Sip24Rounds> as super::DaemonicHasher<GLASS>>::write(&mut self.0.hasher, msg)
		// self.0.hasher.write(msg)
	}

	#[inline]
	fn write_str(&mut self, s: &str) {
		<BaseHasher<Sip24Rounds> as super::DaemonicHasher<GLASS>>::write_str(&mut self.0.hasher, s);
	}
}

impl<GLASS: Glass<GLASS>> super::DaemonicHasher<GLASS> for DaemonicSipHasher13 {
	#[inline]
	fn finish(&self) -> u64 {
		<BaseHasher<Sip13Rounds> as super::DaemonicHasher<GLASS>>::finish(&self.hasher)
	}

	#[inline]
	fn write(&mut self, msg: &[u8]) {
		<BaseHasher<Sip13Rounds> as super::DaemonicHasher<GLASS>>::write(&mut self.hasher, msg)
	}

	#[inline]
	fn write_str(&mut self, s: &str) {
		<BaseHasher<Sip13Rounds> as super::DaemonicHasher<GLASS>>::write_str(&mut self.hasher, s);
	}
}

impl<S: Sip, GLASS: Glass<GLASS>> super::DaemonicHasher<GLASS> for BaseHasher<S> {
	#[inline]
	fn finish(&self) -> u64 {
		let mut state = self.state;

		let b: u64 = ((self.length as u64 & 0xff) << 56) | self.tail;

		state.v3 ^= b;
		S::c_rounds(&mut state);
		state.v0 ^= b;

		state.v2 ^= 0xff;
		S::d_rounds(&mut state);

		state.v0 ^ state.v1 ^ state.v2 ^ state.v3
	}

	// Note: no integer hashing methods (`write_u*`, `write_i*`) are defined
	// for this type. We could add them, copy the `short_write` implementation
	// in librustc_data_structures/sip128.rs, and add `write_u*`/`write_i*`
	// methods to `SipHasher`, `DaemonicSipHasher13`, and `DefaultDaemonicHasher`. This would
	// greatly speed up integer hashing by those hashers, at the cost of
	// slightly slowing down compile speeds on some benchmarks. See #69152 for
	// details.
	#[allow(unsafe_code)]
	#[inline]
	fn write(&mut self, msg: &[u8]) {
		let length = msg.len();
		self.length += length;

		let mut needed = 0;
		if self.ntail != 0 {
			needed = 8 - self.ntail;
			// SAFETY: `cmp::min(length, needed)` is guaranteed to not be over `length`
			self.tail |= unsafe { u8to64_le(msg, 0, cmp::min(length, needed)) } << (8 * self.ntail);
			if length < needed {
				self.ntail += length;
				return;
			} else {
				self.state.v3 ^= self.tail;
				S::c_rounds(&mut self.state);
				self.state.v0 ^= self.tail;
				self.ntail = 0;
			}
		}

		// Buffered tail is now flushed, process new input.
		let len = length - needed;
		let left = len & 0x7; // len % 8

		let mut i = needed;
		while i < len - left {
			// SAFETY: because `len - left` is the biggest multiple of 8 under
			// `len`, and because `i` starts at `needed` where `len` is `length - needed`,
			// `i + 8` is guaranteed to be less than or equal to `length`.
			let mi = unsafe { load_int_le!(msg, i, u64) };

			self.state.v3 ^= mi;
			S::c_rounds(&mut self.state);
			self.state.v0 ^= mi;

			i += 8;
		}

		// SAFETY: `i` is now `needed + len.div_euclid(8) * 8`,
		// so `i + left` = `needed + len` = `length`, which is by
		// definition equal to `msg.len()`.
		self.tail = unsafe { u8to64_le(msg, i, left) };
		self.ntail = left;
	}

	#[inline]
	fn write_str(&mut self, s: &str) {
		// This hasher works byte-wise, and `0xFF` cannot show up in a `str`,
		// so just hashing the one extra byte is enough to be prefix-free.
		<BaseHasher<S> as super::DaemonicHasher<GLASS>>::write(self, s.as_bytes());
		<BaseHasher<S> as super::DaemonicHasher<GLASS>>::write_u8(self, 0xFF);
		// self.write_u8(0xFF);
	}
}

impl<S: Sip> Clone for BaseHasher<S> {
	#[inline]
	fn clone(&self) -> BaseHasher<S> {
		BaseHasher {
			k0: self.k0,
			k1: self.k1,
			length: self.length,
			state: self.state,
			tail: self.tail,
			ntail: self.ntail,
			_marker: self._marker,
		}
	}
}

impl<S: Sip> Default for BaseHasher<S> {
	/// Creates a `BaseHasher<S>` with the two initial keys set to 0.
	#[inline]
	fn default() -> BaseHasher<S> {
		BaseHasher::new_with_keys(0, 0)
	}
}

#[doc(hidden)]
trait Sip {
	fn c_rounds(_: &mut DaemonicHasherState);
	fn d_rounds(_: &mut DaemonicHasherState);
}

#[derive(Debug, Clone, Default)]
struct Sip13Rounds;

impl Sip for Sip13Rounds {
	#[inline]
	fn c_rounds(state: &mut DaemonicHasherState) {
		compress!(state);
	}

	#[inline]
	fn d_rounds(state: &mut DaemonicHasherState) {
		compress!(state);
		compress!(state);
		compress!(state);
	}
}

#[derive(Debug, Clone, Default)]
struct Sip24Rounds;

impl Sip for Sip24Rounds {
	#[inline]
	fn c_rounds(state: &mut DaemonicHasherState) {
		compress!(state);
		compress!(state);
	}

	#[inline]
	fn d_rounds(state: &mut DaemonicHasherState) {
		compress!(state);
		compress!(state);
		compress!(state);
		compress!(state);
	}
}