2997e4fac50ec9f1364d1a4de852d0c2834c3806
3 * Simple and efficient universal hashing for hashtables
5 * (c) 2003 Straylight/Edgeware
8 /*----- Licensing notice --------------------------------------------------*
10 * This file is part of the mLib utilities library.
12 * mLib is free software; you can redistribute it and/or modify
13 * it under the terms of the GNU Library General Public License as
14 * published by the Free Software Foundation; either version 2 of the
15 * License, or (at your option) any later version.
17 * mLib is distributed in the hope that it will be useful,
18 * but WITHOUT ANY WARRANTY; without even the implied warranty of
19 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 * GNU Library General Public License for more details.
22 * You should have received a copy of the GNU Library General Public
23 * License along with mLib; if not, write to the Free
24 * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
28 /*----- Header files ------------------------------------------------------*/
35 /*----- Main code ---------------------------------------------------------*/
37 /* --- @unihash_setkey@ --- *
39 * Arguments: @unihash_info *i@ = where to store the precomputed tables
40 * @uint32 k@ = the key to set, randomly chosen
44 * Use: Calculates the tables required for efficient hashing.
47 static uint32
mul(uint32 x
, uint32 y
)
53 x
= U32(x
<< 1) ^ UNIHASH_POLY
;
61 void unihash_setkey(unihash_info
*i
, uint32 k
)
67 for (a
= 0; a
< UNIHASH_NBATCH
; a
++) {
69 for (b
= 0; b
< 256; b
++) {
70 i
->s
[a
][0][b
] = mul(x
, b
<< 0);
71 i
->s
[a
][1][b
] = mul(x
, b
<< 8);
72 i
->s
[a
][2][b
] = mul(x
, b
<< 16);
73 i
->s
[a
][3][b
] = mul(x
, b
<< 24);
78 /* --- @unihash_hash@ --- *
80 * Arguments: @const unihash_info *i@ = pointer to precomputed table
81 * @uint32 a@ = @i->[0][0][1]@ or value from previous call
82 * @const void *p@ = pointer to data to hash
83 * @size_t sz@ = size of the data
85 * Returns: Hash of data so far.
87 * Use: Hashes data. Call this as many times as needed.
90 uint32
unihash_hash(const unihash_info
*i
, uint32 a
,
91 const void *p
, size_t sz
)
95 assert(UNIHASH_NBATCH
== 4);
97 #define FULLMULT(u, x) \
98 (i->s[u][0][U8((x) >> 0)] ^ i->s[u][1][U8((x) >> 8)] ^ \
99 i->s[u][2][U8((x) >> 16)] ^ i->s[u][3][U8((x) >> 24)]);
101 #define BYTEMULT(u, x) i->s[u][0][x]
103 /* --- Do the main bulk in batches of %$n$% bytes --- *
105 * We have %$a$% and %$m_{n-1}, \ldots, m_1, m_0$%; we want
107 * %$a' = (a + m_{n-1}) k^n + m_{n-2} k^{n-1} + \cdots + m_1 k^2 + m_0 k$%
110 while (sz
>= UNIHASH_NBATCH
) {
113 a
^= BYTEMULT(2, *pp
++);
114 a
^= BYTEMULT(1, *pp
++);
115 a
^= BYTEMULT(0, *pp
++);
116 sz
-= UNIHASH_NBATCH
;
119 /* --- The tail end is a smaller batch --- */
122 case 3: a
^= *pp
++; a
= FULLMULT(2, a
); goto batch_2
;
123 case 2: a
^= *pp
++; a
= FULLMULT(1, a
); goto batch_1
;
124 case 1: a
^= *pp
++; a
= FULLMULT(0, a
); goto batch_0
;
125 batch_2
: a
^= BYTEMULT(1, *pp
++);
126 batch_1
: a
^= BYTEMULT(0, *pp
++);
133 /* --- @unihash@ --- *
135 * Arguments: @const unihash_info *i@ = precomputed tables
136 * @const void *p@ = pointer to data to hash
137 * @size_t sz@ = size of the data
139 * Returns: The hash value computed.
141 * Use: All-in-one hashing function. No faster than using the
142 * separate calls, but more convenient.
145 uint32
unihash(const unihash_info
*i
, const void *p
, size_t sz
)
147 return (UNIHASH(i
, p
, sz
));
150 /*----- Test rig ----------------------------------------------------------*/
156 static int verify(dstr
*v
)
166 static const int step
[] = { 0, 1, 5, 6, 7, 8, 23, -1 };
168 /* --- Set up for using this key --- */
170 k
= *(uint32
*)v
[0].buf
;
171 h
= *(uint32
*)v
[2].buf
;
172 unihash_setkey(&ui
, k
);
174 /* --- Hash the data a lot --- */
176 for (i
= 0; step
[i
] >= 0; i
++) {
179 hh
= unihash(&ui
, v
[1].buf
, v
[1].len
);
181 hh
= UNIHASH_INIT(&ui
);
186 hh
= unihash_hash(&ui
, hh
, p
, c
);
193 fprintf(stderr
, "\nunihash failed\n");
194 fprintf(stderr
, " key = %08lx\n", (unsigned long)k
);
195 fprintf(stderr
, " data = %s\n", v
[1].buf
);
196 fprintf(stderr
, " step = %d\n", step
[i
]);
197 fprintf(stderr
, " expected = %08lx\n", (unsigned long)h
);
198 fprintf(stderr
, " computed = %08lx\n", (unsigned long)hh
);
204 static const test_chunk tests
[] = {
205 { "hash", verify
, { &type_uint32
, &type_string
, &type_uint32
} },
209 int main(int argc
, char *argv
[])
211 test_run(argc
, argv
, tests
, "unihash.in");
217 /*----- That's all, folks -------------------------------------------------*/