src/libcore/hash/sip.rs

   1 // Copyright 2012-2015 The Rust Project Developers. See the COPYRIGHT
   2 // file at the top-level directory of this distribution and at
   3 // http://rust-lang.org/COPYRIGHT.
   4 //
   5 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
   6 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
   7 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
   8 // option. This file may not be copied, modified, or distributed
   9 // except according to those terms.
  10
  11 //! An implementation of SipHash.
  12
  13 #![allow(deprecated)]
  14
  15 use marker::PhantomData;
  16 use ptr;
  17 use cmp;
  18 use mem;
  19
  20 /// An implementation of SipHash 1-3.
  21 ///
  22 /// This is currently the default hashing function used by standard library
  23 /// (eg. `collections::HashMap` uses it by default).
  24 ///
  25 /// See: <https://131002.net/siphash>
  26 #[unstable(feature = "sip_hash_13", issue = "34767")]
  27 #[rustc_deprecated(since = "1.13.0",
  28                    reason = "use `std::collections::hash_map::DefaultHasher` instead")]
  29 #[derive(Debug, Clone, Default)]
  30 pub struct SipHasher13 {
  31     hasher: Hasher<Sip13Rounds>,
  32 }
  33
  34 /// An implementation of SipHash 2-4.
  35 ///
  36 /// See: <https://131002.net/siphash/>
  37 #[unstable(feature = "sip_hash_13", issue = "34767")]
  38 #[rustc_deprecated(since = "1.13.0",
  39                    reason = "use `std::collections::hash_map::DefaultHasher` instead")]
  40 #[derive(Debug, Clone, Default)]
  41 pub struct SipHasher24 {
  42     hasher: Hasher<Sip24Rounds>,
  43 }
  44
  45 /// An implementation of SipHash 2-4.
  46 ///
  47 /// See: <https://131002.net/siphash/>
  48 ///
  49 /// SipHash is a general-purpose hashing function: it runs at a good
  50 /// speed (competitive with Spooky and City) and permits strong _keyed_
  51 /// hashing. This lets you key your hashtables from a strong RNG, such as
  52 /// [`rand::os::OsRng`](https://doc.rust-lang.org/rand/rand/os/struct.OsRng.html).
  53 ///
  54 /// Although the SipHash algorithm is considered to be generally strong,
  55 /// it is not intended for cryptographic purposes. As such, all
  56 /// cryptographic uses of this implementation are _strongly discouraged_.
  57 #[stable(feature = "rust1", since = "1.0.0")]
  58 #[rustc_deprecated(since = "1.13.0",
  59                    reason = "use `std::collections::hash_map::DefaultHasher` instead")]
  60 #[derive(Debug, Clone, Default)]
  61 pub struct SipHasher(SipHasher24);
  62
  63 #[derive(Debug)]
  64 struct Hasher<S: Sip> {
  65     k0: u64,
  66     k1: u64,
  67     length: usize, // how many bytes we've processed
  68     state: State, // hash State
  69     tail: u64, // unprocessed bytes le
  70     ntail: usize, // how many bytes in tail are valid
  71     _marker: PhantomData<S>,
  72 }
  73
  74 #[derive(Debug, Clone, Copy)]
  75 #[repr(C)]
  76 struct State {
  77     // v0, v2 and v1, v3 show up in pairs in the algorithm,
  78     // and simd implementations of SipHash will use vectors
  79     // of v02 and v13. By placing them in this order in the struct,
  80     // the compiler can pick up on just a few simd optimizations by itself.
  81     v0: u64,
  82     v2: u64,
  83     v1: u64,
  84     v3: u64,
  85 }
  86
  87 macro_rules! compress {
  88     ($state:expr) => ({
  89         compress!($state.v0, $state.v1, $state.v2, $state.v3)
  90     });
  91     ($v0:expr, $v1:expr, $v2:expr, $v3:expr) =>
  92     ({
  93         $v0 = $v0.wrapping_add($v1); $v1 = $v1.rotate_left(13); $v1 ^= $v0;
  94         $v0 = $v0.rotate_left(32);
  95         $v2 = $v2.wrapping_add($v3); $v3 = $v3.rotate_left(16); $v3 ^= $v2;
  96         $v0 = $v0.wrapping_add($v3); $v3 = $v3.rotate_left(21); $v3 ^= $v0;
  97         $v2 = $v2.wrapping_add($v1); $v1 = $v1.rotate_left(17); $v1 ^= $v2;
  98         $v2 = $v2.rotate_left(32);
  99     });
 100 }
 101
 102 /// Load an integer of the desired type from a byte stream, in LE order. Uses
 103 /// `copy_nonoverlapping` to let the compiler generate the most efficient way
 104 /// to load it from a possibly unaligned address.
 105 ///
 106 /// Unsafe because: unchecked indexing at i..i+size_of(int_ty)
 107 macro_rules! load_int_le {
 108     ($buf:expr, $i:expr, $int_ty:ident) =>
 109     ({
 110        debug_assert!($i + mem::size_of::<$int_ty>() <= $buf.len());
 111        let mut data = 0 as $int_ty;
 112        ptr::copy_nonoverlapping($buf.get_unchecked($i),
 113                                 &mut data as *mut _ as *mut u8,
 114                                 mem::size_of::<$int_ty>());
 115        data.to_le()
 116     });
 117 }
 118
 119 /// Load an u64 using up to 7 bytes of a byte slice.
 120 ///
 121 /// Unsafe because: unchecked indexing at start..start+len
 122 #[inline]
 123 unsafe fn u8to64_le(buf: &[u8], start: usize, len: usize) -> u64 {
 124     debug_assert!(len < 8);
 125     let mut i = 0; // current byte index (from LSB) in the output u64
 126     let mut out = 0;
 127     if i + 3 < len {
 128         out = load_int_le!(buf, start + i, u32) as u64;
 129         i += 4;
 130     }
 131     if i + 1 < len {
 132         out |= (load_int_le!(buf, start + i, u16) as u64) << (i * 8);
 133         i += 2
 134     }
 135     if i < len {
 136         out |= (*buf.get_unchecked(start + i) as u64) << (i * 8);
 137         i += 1;
 138     }
 139     debug_assert_eq!(i, len);
 140     out
 141 }
 142
 143 impl SipHasher {
 144     /// Creates a new `SipHasher` with the two initial keys set to 0.
 145     #[inline]
 146     #[stable(feature = "rust1", since = "1.0.0")]
 147     #[rustc_deprecated(since = "1.13.0",
 148                        reason = "use `std::collections::hash_map::DefaultHasher` instead")]
 149     pub fn new() -> SipHasher {
 150         SipHasher::new_with_keys(0, 0)
 151     }
 152
 153     /// Creates a `SipHasher` that is keyed off the provided keys.
 154     #[inline]
 155     #[stable(feature = "rust1", since = "1.0.0")]
 156     #[rustc_deprecated(since = "1.13.0",
 157                        reason = "use `std::collections::hash_map::DefaultHasher` instead")]
 158     pub fn new_with_keys(key0: u64, key1: u64) -> SipHasher {
 159         SipHasher(SipHasher24::new_with_keys(key0, key1))
 160     }
 161 }
 162
 163 impl SipHasher13 {
 164     /// Creates a new `SipHasher13` with the two initial keys set to 0.
 165     #[inline]
 166     #[unstable(feature = "sip_hash_13", issue = "34767")]
 167     #[rustc_deprecated(since = "1.13.0",
 168                        reason = "use `std::collections::hash_map::DefaultHasher` instead")]
 169     pub fn new() -> SipHasher13 {
 170         SipHasher13::new_with_keys(0, 0)
 171     }
 172
 173     /// Creates a `SipHasher13` that is keyed off the provided keys.
 174     #[inline]
 175     #[unstable(feature = "sip_hash_13", issue = "34767")]
 176     #[rustc_deprecated(since = "1.13.0",
 177                        reason = "use `std::collections::hash_map::DefaultHasher` instead")]
 178     pub fn new_with_keys(key0: u64, key1: u64) -> SipHasher13 {
 179         SipHasher13 {
 180             hasher: Hasher::new_with_keys(key0, key1)
 181         }
 182     }
 183 }
 184
 185 impl SipHasher24 {
 186     /// Creates a new `SipHasher24` with the two initial keys set to 0.
 187     #[inline]
 188     #[unstable(feature = "sip_hash_13", issue = "34767")]
 189     #[rustc_deprecated(since = "1.13.0",
 190                        reason = "use `std::collections::hash_map::DefaultHasher` instead")]
 191     pub fn new() -> SipHasher24 {
 192         SipHasher24::new_with_keys(0, 0)
 193     }
 194
 195     /// Creates a `SipHasher24` that is keyed off the provided keys.
 196     #[inline]
 197     #[unstable(feature = "sip_hash_13", issue = "34767")]
 198     #[rustc_deprecated(since = "1.13.0",
 199                        reason = "use `std::collections::hash_map::DefaultHasher` instead")]
 200     pub fn new_with_keys(key0: u64, key1: u64) -> SipHasher24 {
 201         SipHasher24 {
 202             hasher: Hasher::new_with_keys(key0, key1)
 203         }
 204     }
 205 }
 206
 207 impl<S: Sip> Hasher<S> {
 208     #[inline]
 209     fn new_with_keys(key0: u64, key1: u64) -> Hasher<S> {
 210         let mut state = Hasher {
 211             k0: key0,
 212             k1: key1,
 213             length: 0,
 214             state: State {
 215                 v0: 0,
 216                 v1: 0,
 217                 v2: 0,
 218                 v3: 0,
 219             },
 220             tail: 0,
 221             ntail: 0,
 222             _marker: PhantomData,
 223         };
 224         state.reset();
 225         state
 226     }
 227
 228     #[inline]
 229     fn reset(&mut self) {
 230         self.length = 0;
 231         self.state.v0 = self.k0 ^ 0x736f6d6570736575;
 232         self.state.v1 = self.k1 ^ 0x646f72616e646f6d;
 233         self.state.v2 = self.k0 ^ 0x6c7967656e657261;
 234         self.state.v3 = self.k1 ^ 0x7465646279746573;
 235         self.ntail = 0;
 236     }
 237
 238     // Specialized write function that is only valid for buffers with len <= 8.
 239     // It's used to force inlining of write_u8 and write_usize, those would normally be inlined
 240     // except for composite types (that includes slices and str hashing because of delimiter).
 241     // Without this extra push the compiler is very reluctant to inline delimiter writes,
 242     // degrading performance substantially for the most common use cases.
 243     #[inline]
 244     fn short_write(&mut self, msg: &[u8]) {
 245         debug_assert!(msg.len() <= 8);
 246         let length = msg.len();
 247         self.length += length;
 248
 249         let needed = 8 - self.ntail;
 250         let fill = cmp::min(length, needed);
 251         if fill == 8 {
 252             self.tail = unsafe { load_int_le!(msg, 0, u64) };
 253         } else {
 254             self.tail |= unsafe { u8to64_le(msg, 0, fill) } << (8 * self.ntail);
 255             if length < needed {
 256                 self.ntail += length;
 257                 return;
 258             }
 259         }
 260         self.state.v3 ^= self.tail;
 261         S::c_rounds(&mut self.state);
 262         self.state.v0 ^= self.tail;
 263
 264         // Buffered tail is now flushed, process new input.
 265         self.ntail = length - needed;
 266         self.tail = unsafe { u8to64_le(msg, needed, self.ntail) };
 267     }
 268 }
 269
 270 #[stable(feature = "rust1", since = "1.0.0")]
 271 impl super::Hasher for SipHasher {
 272     #[inline]
 273     fn write(&mut self, msg: &[u8]) {
 274         self.0.write(msg)
 275     }
 276
 277     #[inline]
 278     fn finish(&self) -> u64 {
 279         self.0.finish()
 280     }
 281 }
 282
 283 #[unstable(feature = "sip_hash_13", issue = "34767")]
 284 impl super::Hasher for SipHasher13 {
 285     #[inline]
 286     fn write(&mut self, msg: &[u8]) {
 287         self.hasher.write(msg)
 288     }
 289
 290     #[inline]
 291     fn finish(&self) -> u64 {
 292         self.hasher.finish()
 293     }
 294 }
 295
 296 #[unstable(feature = "sip_hash_13", issue = "34767")]
 297 impl super::Hasher for SipHasher24 {
 298     #[inline]
 299     fn write(&mut self, msg: &[u8]) {
 300         self.hasher.write(msg)
 301     }
 302
 303     #[inline]
 304     fn finish(&self) -> u64 {
 305         self.hasher.finish()
 306     }
 307 }
 308
 309 impl<S: Sip> super::Hasher for Hasher<S> {
 310     // see short_write comment for explanation
 311     #[inline]
 312     fn write_usize(&mut self, i: usize) {
 313         let bytes = unsafe {
 314             ::slice::from_raw_parts(&i as *const usize as *const u8, mem::size_of::<usize>())
 315         };
 316         self.short_write(bytes);
 317     }
 318
 319     // see short_write comment for explanation
 320     #[inline]
 321     fn write_u8(&mut self, i: u8) {
 322         self.short_write(&[i]);
 323     }
 324
 325     #[inline]
 326     fn write(&mut self, msg: &[u8]) {
 327         let length = msg.len();
 328         self.length += length;
 329
 330         let mut needed = 0;
 331
 332         if self.ntail != 0 {
 333             needed = 8 - self.ntail;
 334             self.tail |= unsafe { u8to64_le(msg, 0, cmp::min(length, needed)) } << 8 * self.ntail;
 335             if length < needed {
 336                 self.ntail += length;
 337                 return
 338             } else {
 339                 self.state.v3 ^= self.tail;
 340                 S::c_rounds(&mut self.state);
 341                 self.state.v0 ^= self.tail;
 342                 self.ntail = 0;
 343             }
 344         }
 345
 346         // Buffered tail is now flushed, process new input.
 347         let len = length - needed;
 348         let left = len & 0x7;
 349
 350         let mut i = needed;
 351         while i < len - left {
 352             let mi = unsafe { load_int_le!(msg, i, u64) };
 353
 354             self.state.v3 ^= mi;
 355             S::c_rounds(&mut self.state);
 356             self.state.v0 ^= mi;
 357
 358             i += 8;
 359         }
 360
 361         self.tail = unsafe { u8to64_le(msg, i, left) };
 362         self.ntail = left;
 363     }
 364
 365     #[inline]
 366     fn finish(&self) -> u64 {
 367         let mut state = self.state;
 368
 369         let b: u64 = ((self.length as u64 & 0xff) << 56) | self.tail;
 370
 371         state.v3 ^= b;
 372         S::c_rounds(&mut state);
 373         state.v0 ^= b;
 374
 375         state.v2 ^= 0xff;
 376         S::d_rounds(&mut state);
 377
 378         state.v0 ^ state.v1 ^ state.v2 ^ state.v3
 379     }
 380 }
 381
 382 impl<S: Sip> Clone for Hasher<S> {
 383     #[inline]
 384     fn clone(&self) -> Hasher<S> {
 385         Hasher {
 386             k0: self.k0,
 387             k1: self.k1,
 388             length: self.length,
 389             state: self.state,
 390             tail: self.tail,
 391             ntail: self.ntail,
 392             _marker: self._marker,
 393         }
 394     }
 395 }
 396
 397 impl<S: Sip> Default for Hasher<S> {
 398     /// Creates a `Hasher<S>` with the two initial keys set to 0.
 399     #[inline]
 400     fn default() -> Hasher<S> {
 401         Hasher::new_with_keys(0, 0)
 402     }
 403 }
 404
 405 #[doc(hidden)]
 406 trait Sip {
 407     fn c_rounds(_: &mut State);
 408     fn d_rounds(_: &mut State);
 409 }
 410
 411 #[derive(Debug, Clone, Default)]
 412 struct Sip13Rounds;
 413
 414 impl Sip for Sip13Rounds {
 415     #[inline]
 416     fn c_rounds(state: &mut State) {
 417         compress!(state);
 418     }
 419
 420     #[inline]
 421     fn d_rounds(state: &mut State) {
 422         compress!(state);
 423         compress!(state);
 424         compress!(state);
 425     }
 426 }
 427
 428 #[derive(Debug, Clone, Default)]
 429 struct Sip24Rounds;
 430
 431 impl Sip for Sip24Rounds {
 432     #[inline]
 433     fn c_rounds(state: &mut State) {
 434         compress!(state);
 435         compress!(state);
 436     }
 437
 438     #[inline]
 439     fn d_rounds(state: &mut State) {
 440         compress!(state);
 441         compress!(state);
 442         compress!(state);
 443         compress!(state);
 444     }
 445 }