1 | /* adler32.c -- compute the Adler-32 checksum of a data stream |
2 | * Copyright (C) 1995-2011, 2016 Mark Adler |
3 | * For conditions of distribution and use, see copyright notice in zlib.h |
4 | */ |
5 | |
6 | /* @(#) $Id$ */ |
7 | |
8 | #include "zutil.h" |
9 | |
10 | #define local static |
11 | |
12 | local uLong adler32_combine_ OF((uLong adler1, uLong adler2, z_off64_t len2)); |
13 | |
14 | #define BASE 65521U /* largest prime smaller than 65536 */ |
15 | #define NMAX 5552 |
16 | /* NMAX is the largest n such that 255n(n+1)/2 + (n+1)(BASE-1) <= 2^32-1 */ |
17 | |
18 | #define DO1(buf,i) {adler += (buf)[i]; sum2 += adler;} |
19 | #define DO2(buf,i) DO1(buf,i); DO1(buf,i+1); |
20 | #define DO4(buf,i) DO2(buf,i); DO2(buf,i+2); |
21 | #define DO8(buf,i) DO4(buf,i); DO4(buf,i+4); |
22 | #define DO16(buf) DO8(buf,0); DO8(buf,8); |
23 | |
24 | /* use NO_DIVIDE if your processor does not do division in hardware -- |
25 | try it both ways to see which is faster */ |
26 | #ifdef NO_DIVIDE |
27 | /* note that this assumes BASE is 65521, where 65536 % 65521 == 15 |
28 | (thank you to John Reiser for pointing this out) */ |
29 | # define CHOP(a) \ |
30 | do { \ |
31 | unsigned long tmp = a >> 16; \ |
32 | a &= 0xffffUL; \ |
33 | a += (tmp << 4) - tmp; \ |
34 | } while (0) |
35 | # define MOD28(a) \ |
36 | do { \ |
37 | CHOP(a); \ |
38 | if (a >= BASE) a -= BASE; \ |
39 | } while (0) |
40 | # define MOD(a) \ |
41 | do { \ |
42 | CHOP(a); \ |
43 | MOD28(a); \ |
44 | } while (0) |
45 | # define MOD63(a) \ |
46 | do { /* this assumes a is not negative */ \ |
47 | z_off64_t tmp = a >> 32; \ |
48 | a &= 0xffffffffL; \ |
49 | a += (tmp << 8) - (tmp << 5) + tmp; \ |
50 | tmp = a >> 16; \ |
51 | a &= 0xffffL; \ |
52 | a += (tmp << 4) - tmp; \ |
53 | tmp = a >> 16; \ |
54 | a &= 0xffffL; \ |
55 | a += (tmp << 4) - tmp; \ |
56 | if (a >= BASE) a -= BASE; \ |
57 | } while (0) |
58 | #else |
59 | # define MOD(a) a %= BASE |
60 | # define MOD28(a) a %= BASE |
61 | # define MOD63(a) a %= BASE |
62 | #endif |
63 | |
64 | /* ========================================================================= */ |
65 | uLong ZEXPORT adler32_z(adler, buf, len) |
66 | uLong adler; |
67 | const Bytef *buf; |
68 | z_size_t len; |
69 | { |
70 | unsigned long sum2; |
71 | unsigned n; |
72 | |
73 | /* split Adler-32 into component sums */ |
74 | sum2 = (adler >> 16) & 0xffff; |
75 | adler &= 0xffff; |
76 | |
77 | /* in case user likes doing a byte at a time, keep it fast */ |
78 | if (len == 1) { |
79 | adler += buf[0]; |
80 | if (adler >= BASE) |
81 | adler -= BASE; |
82 | sum2 += adler; |
83 | if (sum2 >= BASE) |
84 | sum2 -= BASE; |
85 | return adler | (sum2 << 16); |
86 | } |
87 | |
88 | /* initial Adler-32 value (deferred check for len == 1 speed) */ |
89 | if (buf == Z_NULL) |
90 | return 1L; |
91 | |
92 | /* in case short lengths are provided, keep it somewhat fast */ |
93 | if (len < 16) { |
94 | while (len--) { |
95 | adler += *buf++; |
96 | sum2 += adler; |
97 | } |
98 | if (adler >= BASE) |
99 | adler -= BASE; |
100 | MOD28(sum2); /* only added so many BASE's */ |
101 | return adler | (sum2 << 16); |
102 | } |
103 | |
104 | /* do length NMAX blocks -- requires just one modulo operation */ |
105 | while (len >= NMAX) { |
106 | len -= NMAX; |
107 | n = NMAX / 16; /* NMAX is divisible by 16 */ |
108 | do { |
109 | DO16(buf); /* 16 sums unrolled */ |
110 | buf += 16; |
111 | } while (--n); |
112 | MOD(adler); |
113 | MOD(sum2); |
114 | } |
115 | |
116 | /* do remaining bytes (less than NMAX, still just one modulo) */ |
117 | if (len) { /* avoid modulos if none remaining */ |
118 | while (len >= 16) { |
119 | len -= 16; |
120 | DO16(buf); |
121 | buf += 16; |
122 | } |
123 | while (len--) { |
124 | adler += *buf++; |
125 | sum2 += adler; |
126 | } |
127 | MOD(adler); |
128 | MOD(sum2); |
129 | } |
130 | |
131 | /* return recombined sums */ |
132 | return adler | (sum2 << 16); |
133 | } |
134 | |
135 | /* ========================================================================= */ |
136 | uLong ZEXPORT adler32(adler, buf, len) |
137 | uLong adler; |
138 | const Bytef *buf; |
139 | uInt len; |
140 | { |
141 | return adler32_z(adler, buf, len); |
142 | } |
143 | |
144 | /* ========================================================================= */ |
145 | local uLong adler32_combine_(adler1, adler2, len2) |
146 | uLong adler1; |
147 | uLong adler2; |
148 | z_off64_t len2; |
149 | { |
150 | unsigned long sum1; |
151 | unsigned long sum2; |
152 | unsigned rem; |
153 | |
154 | /* for negative len, return invalid adler32 as a clue for debugging */ |
155 | if (len2 < 0) |
156 | return 0xffffffffUL; |
157 | |
158 | /* the derivation of this formula is left as an exercise for the reader */ |
159 | MOD63(len2); /* assumes len2 >= 0 */ |
160 | rem = (unsigned)len2; |
161 | sum1 = adler1 & 0xffff; |
162 | sum2 = rem * sum1; |
163 | MOD(sum2); |
164 | sum1 += (adler2 & 0xffff) + BASE - 1; |
165 | sum2 += ((adler1 >> 16) & 0xffff) + ((adler2 >> 16) & 0xffff) + BASE - rem; |
166 | if (sum1 >= BASE) sum1 -= BASE; |
167 | if (sum1 >= BASE) sum1 -= BASE; |
168 | if (sum2 >= ((unsigned long)BASE << 1)) sum2 -= ((unsigned long)BASE << 1); |
169 | if (sum2 >= BASE) sum2 -= BASE; |
170 | return sum1 | (sum2 << 16); |
171 | } |
172 | |
173 | /* ========================================================================= */ |
174 | uLong ZEXPORT adler32_combine(adler1, adler2, len2) |
175 | uLong adler1; |
176 | uLong adler2; |
177 | z_off_t len2; |
178 | { |
179 | return adler32_combine_(adler1, adler2, len2); |
180 | } |
181 | |
182 | uLong ZEXPORT adler32_combine64(adler1, adler2, len2) |
183 | uLong adler1; |
184 | uLong adler2; |
185 | z_off64_t len2; |
186 | { |
187 | return adler32_combine_(adler1, adler2, len2); |
188 | } |
189 | |