c++: over-eager friend matching [PR109649]
[official-gcc.git] / zlib / adler32.c
blobb3d35e2a891b9dfd14ffecf79bf202c58659b986
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 */
6 /* @(#) $Id: adler32.c,v 1.1.1.2 2002/03/11 21:53:23 tromey Exp $ */
8 #include "zutil.h"
10 local uLong adler32_combine_ OF((uLong adler1, uLong adler2, z_off64_t len2));
12 #define BASE 65521U /* largest prime smaller than 65536 */
13 #define NMAX 5552
14 /* NMAX is the largest n such that 255n(n+1)/2 + (n+1)(BASE-1) <= 2^32-1 */
16 #define DO1(buf,i) {adler += (buf)[i]; sum2 += adler;}
17 #define DO2(buf,i) DO1(buf,i); DO1(buf,i+1);
18 #define DO4(buf,i) DO2(buf,i); DO2(buf,i+2);
19 #define DO8(buf,i) DO4(buf,i); DO4(buf,i+4);
20 #define DO16(buf) DO8(buf,0); DO8(buf,8);
22 /* use NO_DIVIDE if your processor does not do division in hardware --
23 try it both ways to see which is faster */
24 #ifdef NO_DIVIDE
25 /* note that this assumes BASE is 65521, where 65536 % 65521 == 15
26 (thank you to John Reiser for pointing this out) */
27 # define CHOP(a) \
28 do { \
29 unsigned long tmp = a >> 16; \
30 a &= 0xffffUL; \
31 a += (tmp << 4) - tmp; \
32 } while (0)
33 # define MOD28(a) \
34 do { \
35 CHOP(a); \
36 if (a >= BASE) a -= BASE; \
37 } while (0)
38 # define MOD(a) \
39 do { \
40 CHOP(a); \
41 MOD28(a); \
42 } while (0)
43 # define MOD63(a) \
44 do { /* this assumes a is not negative */ \
45 z_off64_t tmp = a >> 32; \
46 a &= 0xffffffffL; \
47 a += (tmp << 8) - (tmp << 5) + tmp; \
48 tmp = a >> 16; \
49 a &= 0xffffL; \
50 a += (tmp << 4) - tmp; \
51 tmp = a >> 16; \
52 a &= 0xffffL; \
53 a += (tmp << 4) - tmp; \
54 if (a >= BASE) a -= BASE; \
55 } while (0)
56 #else
57 # define MOD(a) a %= BASE
58 # define MOD28(a) a %= BASE
59 # define MOD63(a) a %= BASE
60 #endif
62 /* ========================================================================= */
63 uLong ZEXPORT adler32_z(adler, buf, len)
64 uLong adler;
65 const Bytef *buf;
66 z_size_t len;
68 unsigned long sum2;
69 unsigned n;
71 /* split Adler-32 into component sums */
72 sum2 = (adler >> 16) & 0xffff;
73 adler &= 0xffff;
75 /* in case user likes doing a byte at a time, keep it fast */
76 if (len == 1) {
77 adler += buf[0];
78 if (adler >= BASE)
79 adler -= BASE;
80 sum2 += adler;
81 if (sum2 >= BASE)
82 sum2 -= BASE;
83 return adler | (sum2 << 16);
86 /* initial Adler-32 value (deferred check for len == 1 speed) */
87 if (buf == Z_NULL)
88 return 1L;
90 /* in case short lengths are provided, keep it somewhat fast */
91 if (len < 16) {
92 while (len--) {
93 adler += *buf++;
94 sum2 += adler;
96 if (adler >= BASE)
97 adler -= BASE;
98 MOD28(sum2); /* only added so many BASE's */
99 return adler | (sum2 << 16);
102 /* do length NMAX blocks -- requires just one modulo operation */
103 while (len >= NMAX) {
104 len -= NMAX;
105 n = NMAX / 16; /* NMAX is divisible by 16 */
106 do {
107 DO16(buf); /* 16 sums unrolled */
108 buf += 16;
109 } while (--n);
110 MOD(adler);
111 MOD(sum2);
114 /* do remaining bytes (less than NMAX, still just one modulo) */
115 if (len) { /* avoid modulos if none remaining */
116 while (len >= 16) {
117 len -= 16;
118 DO16(buf);
119 buf += 16;
121 while (len--) {
122 adler += *buf++;
123 sum2 += adler;
125 MOD(adler);
126 MOD(sum2);
129 /* return recombined sums */
130 return adler | (sum2 << 16);
133 /* ========================================================================= */
134 uLong ZEXPORT adler32(adler, buf, len)
135 uLong adler;
136 const Bytef *buf;
137 uInt len;
139 return adler32_z(adler, buf, len);
142 /* ========================================================================= */
143 local uLong adler32_combine_(adler1, adler2, len2)
144 uLong adler1;
145 uLong adler2;
146 z_off64_t len2;
148 unsigned long sum1;
149 unsigned long sum2;
150 unsigned rem;
152 /* for negative len, return invalid adler32 as a clue for debugging */
153 if (len2 < 0)
154 return 0xffffffffUL;
156 /* the derivation of this formula is left as an exercise for the reader */
157 MOD63(len2); /* assumes len2 >= 0 */
158 rem = (unsigned)len2;
159 sum1 = adler1 & 0xffff;
160 sum2 = rem * sum1;
161 MOD(sum2);
162 sum1 += (adler2 & 0xffff) + BASE - 1;
163 sum2 += ((adler1 >> 16) & 0xffff) + ((adler2 >> 16) & 0xffff) + BASE - rem;
164 if (sum1 >= BASE) sum1 -= BASE;
165 if (sum1 >= BASE) sum1 -= BASE;
166 if (sum2 >= ((unsigned long)BASE << 1)) sum2 -= ((unsigned long)BASE << 1);
167 if (sum2 >= BASE) sum2 -= BASE;
168 return sum1 | (sum2 << 16);
171 /* ========================================================================= */
172 uLong ZEXPORT adler32_combine(adler1, adler2, len2)
173 uLong adler1;
174 uLong adler2;
175 z_off_t len2;
177 return adler32_combine_(adler1, adler2, len2);
180 uLong ZEXPORT adler32_combine64(adler1, adler2, len2)
181 uLong adler1;
182 uLong adler2;
183 z_off64_t len2;
185 return adler32_combine_(adler1, adler2, len2);