Skip to content

Commit fada46c

Browse files
committed
Add core::hash containing SipHash-2-4 implementation. Re: #1616 and #859.
1 parent 0737f1b commit fada46c

File tree

4 files changed

+171
-13
lines changed

4 files changed

+171
-13
lines changed

src/libcore/core.rc

+2
Original file line numberDiff line numberDiff line change
@@ -49,6 +49,7 @@ export tuple;
4949
export to_str, to_bytes;
5050
export dvec, dvec_iter;
5151
export dlist, dlist_iter;
52+
export hash;
5253
export cmp;
5354
export num;
5455

@@ -162,6 +163,7 @@ mod tuple;
162163

163164
mod cmp;
164165
mod num;
166+
mod hash;
165167
mod either;
166168
mod iter;
167169
mod logging;

src/libcore/hash.rs

+163
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,163 @@
1+
/*!
2+
* Implementation of SipHash 2-4
3+
*
4+
* See: http://131002.net/siphash/
5+
*
6+
* Consider this as a main "general-purpose" hash for all hashtables: it
7+
* runs at good speed (competitive with spooky and city) and permits
8+
* cryptographically strong _keyed_ hashing. Key your hashtables from a
9+
* CPRNG like rand::rng.
10+
*/
11+
12+
pure fn hash_bytes(buf: &[const u8]) -> u64 {
13+
ret hash_bytes_keyed(buf, 0u64, 0u64);
14+
}
15+
16+
pure fn hash_bytes_keyed(buf: &[const u8], k0: u64, k1: u64) -> u64 {
17+
18+
let mut v0 : u64 = k0 ^ 0x736f_6d65_7073_6575;
19+
let mut v1 : u64 = k1 ^ 0x646f_7261_6e64_6f6d;
20+
let mut v2 : u64 = k0 ^ 0x6c79_6765_6e65_7261;
21+
let mut v3 : u64 = k1 ^ 0x7465_6462_7974_6573;
22+
23+
#macro([#u8to64_le(buf,i),
24+
(buf[0+i] as u64 |
25+
buf[1+i] as u64 << 8 |
26+
buf[2+i] as u64 << 16 |
27+
buf[3+i] as u64 << 24 |
28+
buf[4+i] as u64 << 32 |
29+
buf[5+i] as u64 << 40 |
30+
buf[6+i] as u64 << 48 |
31+
buf[7+i] as u64 << 56)]);
32+
33+
#macro([#rotl(x,b), (x << b) | (x >> (64 - b))]);
34+
35+
#macro([#compress(), {
36+
v0 += v1; v1 = #rotl(v1, 13); v1 ^= v0; v0 = #rotl(v0, 32);
37+
v2 += v3; v3 = #rotl(v3, 16); v3 ^= v2;
38+
v0 += v3; v3 = #rotl(v3, 21); v3 ^= v0;
39+
v2 += v1; v1 = #rotl(v1, 17); v1 ^= v2; v2 = #rotl(v2, 32);
40+
}]);
41+
42+
let len = vec::len(buf);
43+
let end = len & (!0x7);
44+
let left = len & 0x7;
45+
46+
let mut i = 0;
47+
while i < end {
48+
let m = #u8to64_le(buf, i);
49+
v3 ^= m;
50+
#compress();
51+
#compress();
52+
v0 ^= m;
53+
i += 8;
54+
}
55+
56+
let mut b : u64 = (len as u64 & 0xff) << 56;
57+
58+
if left > 0 { b |= buf[i] as u64; }
59+
if left > 1 { b |= buf[i + 1] as u64 << 8; }
60+
if left > 2 { b |= buf[i + 2] as u64 << 16; }
61+
if left > 3 { b |= buf[i + 3] as u64 << 24; }
62+
if left > 4 { b |= buf[i + 4] as u64 << 32; }
63+
if left > 5 { b |= buf[i + 5] as u64 << 40; }
64+
if left > 6 { b |= buf[i + 6] as u64 << 48; }
65+
66+
v3 ^= b;
67+
#compress();
68+
#compress();
69+
v0 ^= b;
70+
71+
v2 ^= 0xff;
72+
#compress();
73+
#compress();
74+
#compress();
75+
#compress();
76+
77+
ret v0 ^ v1 ^ v2 ^ v3;
78+
}
79+
80+
#[test]
81+
fn test_siphash() {
82+
let vecs : [[u8]/8]/64 = [
83+
[ 0x31, 0x0e, 0x0e, 0xdd, 0x47, 0xdb, 0x6f, 0x72, ]/_,
84+
[ 0xfd, 0x67, 0xdc, 0x93, 0xc5, 0x39, 0xf8, 0x74, ]/_,
85+
[ 0x5a, 0x4f, 0xa9, 0xd9, 0x09, 0x80, 0x6c, 0x0d, ]/_,
86+
[ 0x2d, 0x7e, 0xfb, 0xd7, 0x96, 0x66, 0x67, 0x85, ]/_,
87+
[ 0xb7, 0x87, 0x71, 0x27, 0xe0, 0x94, 0x27, 0xcf, ]/_,
88+
[ 0x8d, 0xa6, 0x99, 0xcd, 0x64, 0x55, 0x76, 0x18, ]/_,
89+
[ 0xce, 0xe3, 0xfe, 0x58, 0x6e, 0x46, 0xc9, 0xcb, ]/_,
90+
[ 0x37, 0xd1, 0x01, 0x8b, 0xf5, 0x00, 0x02, 0xab, ]/_,
91+
[ 0x62, 0x24, 0x93, 0x9a, 0x79, 0xf5, 0xf5, 0x93, ]/_,
92+
[ 0xb0, 0xe4, 0xa9, 0x0b, 0xdf, 0x82, 0x00, 0x9e, ]/_,
93+
[ 0xf3, 0xb9, 0xdd, 0x94, 0xc5, 0xbb, 0x5d, 0x7a, ]/_,
94+
[ 0xa7, 0xad, 0x6b, 0x22, 0x46, 0x2f, 0xb3, 0xf4, ]/_,
95+
[ 0xfb, 0xe5, 0x0e, 0x86, 0xbc, 0x8f, 0x1e, 0x75, ]/_,
96+
[ 0x90, 0x3d, 0x84, 0xc0, 0x27, 0x56, 0xea, 0x14, ]/_,
97+
[ 0xee, 0xf2, 0x7a, 0x8e, 0x90, 0xca, 0x23, 0xf7, ]/_,
98+
[ 0xe5, 0x45, 0xbe, 0x49, 0x61, 0xca, 0x29, 0xa1, ]/_,
99+
[ 0xdb, 0x9b, 0xc2, 0x57, 0x7f, 0xcc, 0x2a, 0x3f, ]/_,
100+
[ 0x94, 0x47, 0xbe, 0x2c, 0xf5, 0xe9, 0x9a, 0x69, ]/_,
101+
[ 0x9c, 0xd3, 0x8d, 0x96, 0xf0, 0xb3, 0xc1, 0x4b, ]/_,
102+
[ 0xbd, 0x61, 0x79, 0xa7, 0x1d, 0xc9, 0x6d, 0xbb, ]/_,
103+
[ 0x98, 0xee, 0xa2, 0x1a, 0xf2, 0x5c, 0xd6, 0xbe, ]/_,
104+
[ 0xc7, 0x67, 0x3b, 0x2e, 0xb0, 0xcb, 0xf2, 0xd0, ]/_,
105+
[ 0x88, 0x3e, 0xa3, 0xe3, 0x95, 0x67, 0x53, 0x93, ]/_,
106+
[ 0xc8, 0xce, 0x5c, 0xcd, 0x8c, 0x03, 0x0c, 0xa8, ]/_,
107+
[ 0x94, 0xaf, 0x49, 0xf6, 0xc6, 0x50, 0xad, 0xb8, ]/_,
108+
[ 0xea, 0xb8, 0x85, 0x8a, 0xde, 0x92, 0xe1, 0xbc, ]/_,
109+
[ 0xf3, 0x15, 0xbb, 0x5b, 0xb8, 0x35, 0xd8, 0x17, ]/_,
110+
[ 0xad, 0xcf, 0x6b, 0x07, 0x63, 0x61, 0x2e, 0x2f, ]/_,
111+
[ 0xa5, 0xc9, 0x1d, 0xa7, 0xac, 0xaa, 0x4d, 0xde, ]/_,
112+
[ 0x71, 0x65, 0x95, 0x87, 0x66, 0x50, 0xa2, 0xa6, ]/_,
113+
[ 0x28, 0xef, 0x49, 0x5c, 0x53, 0xa3, 0x87, 0xad, ]/_,
114+
[ 0x42, 0xc3, 0x41, 0xd8, 0xfa, 0x92, 0xd8, 0x32, ]/_,
115+
[ 0xce, 0x7c, 0xf2, 0x72, 0x2f, 0x51, 0x27, 0x71, ]/_,
116+
[ 0xe3, 0x78, 0x59, 0xf9, 0x46, 0x23, 0xf3, 0xa7, ]/_,
117+
[ 0x38, 0x12, 0x05, 0xbb, 0x1a, 0xb0, 0xe0, 0x12, ]/_,
118+
[ 0xae, 0x97, 0xa1, 0x0f, 0xd4, 0x34, 0xe0, 0x15, ]/_,
119+
[ 0xb4, 0xa3, 0x15, 0x08, 0xbe, 0xff, 0x4d, 0x31, ]/_,
120+
[ 0x81, 0x39, 0x62, 0x29, 0xf0, 0x90, 0x79, 0x02, ]/_,
121+
[ 0x4d, 0x0c, 0xf4, 0x9e, 0xe5, 0xd4, 0xdc, 0xca, ]/_,
122+
[ 0x5c, 0x73, 0x33, 0x6a, 0x76, 0xd8, 0xbf, 0x9a, ]/_,
123+
[ 0xd0, 0xa7, 0x04, 0x53, 0x6b, 0xa9, 0x3e, 0x0e, ]/_,
124+
[ 0x92, 0x59, 0x58, 0xfc, 0xd6, 0x42, 0x0c, 0xad, ]/_,
125+
[ 0xa9, 0x15, 0xc2, 0x9b, 0xc8, 0x06, 0x73, 0x18, ]/_,
126+
[ 0x95, 0x2b, 0x79, 0xf3, 0xbc, 0x0a, 0xa6, 0xd4, ]/_,
127+
[ 0xf2, 0x1d, 0xf2, 0xe4, 0x1d, 0x45, 0x35, 0xf9, ]/_,
128+
[ 0x87, 0x57, 0x75, 0x19, 0x04, 0x8f, 0x53, 0xa9, ]/_,
129+
[ 0x10, 0xa5, 0x6c, 0xf5, 0xdf, 0xcd, 0x9a, 0xdb, ]/_,
130+
[ 0xeb, 0x75, 0x09, 0x5c, 0xcd, 0x98, 0x6c, 0xd0, ]/_,
131+
[ 0x51, 0xa9, 0xcb, 0x9e, 0xcb, 0xa3, 0x12, 0xe6, ]/_,
132+
[ 0x96, 0xaf, 0xad, 0xfc, 0x2c, 0xe6, 0x66, 0xc7, ]/_,
133+
[ 0x72, 0xfe, 0x52, 0x97, 0x5a, 0x43, 0x64, 0xee, ]/_,
134+
[ 0x5a, 0x16, 0x45, 0xb2, 0x76, 0xd5, 0x92, 0xa1, ]/_,
135+
[ 0xb2, 0x74, 0xcb, 0x8e, 0xbf, 0x87, 0x87, 0x0a, ]/_,
136+
[ 0x6f, 0x9b, 0xb4, 0x20, 0x3d, 0xe7, 0xb3, 0x81, ]/_,
137+
[ 0xea, 0xec, 0xb2, 0xa3, 0x0b, 0x22, 0xa8, 0x7f, ]/_,
138+
[ 0x99, 0x24, 0xa4, 0x3c, 0xc1, 0x31, 0x57, 0x24, ]/_,
139+
[ 0xbd, 0x83, 0x8d, 0x3a, 0xaf, 0xbf, 0x8d, 0xb7, ]/_,
140+
[ 0x0b, 0x1a, 0x2a, 0x32, 0x65, 0xd5, 0x1a, 0xea, ]/_,
141+
[ 0x13, 0x50, 0x79, 0xa3, 0x23, 0x1c, 0xe6, 0x60, ]/_,
142+
[ 0x93, 0x2b, 0x28, 0x46, 0xe4, 0xd7, 0x06, 0x66, ]/_,
143+
[ 0xe1, 0x91, 0x5f, 0x5c, 0xb1, 0xec, 0xa4, 0x6c, ]/_,
144+
[ 0xf3, 0x25, 0x96, 0x5c, 0xa1, 0x6d, 0x62, 0x9f, ]/_,
145+
[ 0x57, 0x5f, 0xf2, 0x8e, 0x60, 0x38, 0x1b, 0xe5, ]/_,
146+
[ 0x72, 0x45, 0x06, 0xeb, 0x4c, 0x32, 0x8a, 0x95, ]/_
147+
]/_;
148+
149+
let k0 = 0x_07_06_05_04_03_02_01_00_u64;
150+
let k1 = 0x_0f_0e_0d_0c_0b_0a_09_08_u64;
151+
let mut buf : ~[u8] = ~[];
152+
let mut t = 0;
153+
while t < 64 {
154+
#debug("siphash test %?", t);
155+
let vec = #u8to64_le(vecs[t], 0);
156+
let out = hash_bytes_keyed(buf, k0, k1);
157+
#debug("got %?, expected %?", out, vec);
158+
assert vec == out;
159+
buf += ~[t as u8];
160+
t += 1;
161+
}
162+
163+
}

src/libcore/str.rs

+4-5
Original file line numberDiff line numberDiff line change
@@ -660,11 +660,10 @@ pure fn le(&&a: ~str, &&b: ~str) -> bool { a <= b }
660660

661661
/// String hash function
662662
pure fn hash(&&s: ~str) -> uint {
663-
// djb hash.
664-
// FIXME: replace with murmur. (see #859 and #1616)
665-
let mut u: uint = 5381u;
666-
for each(s) |c| { u *= 33u; u += c as uint; }
667-
ret u;
663+
let x = do as_bytes(s) |bytes| {
664+
hash::hash_bytes(bytes)
665+
};
666+
ret x as uint;
668667
}
669668

670669
/*

src/libcore/vec.rs

+2-8
Original file line numberDiff line numberDiff line change
@@ -1663,15 +1663,9 @@ mod u8 {
16631663
/// Bytewise greater than
16641664
pure fn gt(&&a: ~[u8], &&b: ~[u8]) -> bool { cmp(a, b) > 0 }
16651665

1666-
/// String hash function
1666+
/// Byte-vec hash function
16671667
fn hash(&&s: ~[u8]) -> uint {
1668-
/* Seems to have been tragically copy/pasted from str.rs,
1669-
or vice versa. But I couldn't figure out how to abstract
1670-
it out. -- tjc */
1671-
1672-
let mut u: uint = 5381u;
1673-
vec::iter(s, |c| {u *= 33u; u += c as uint;});
1674-
ret u;
1668+
hash::hash_bytes(s) as uint
16751669
}
16761670

16771671
/**

0 commit comments

Comments
 (0)