zlib 1.2.2.2

1 files changed, 84 insertions, 31 deletions

diff --git a/adler32.c b/adler32.c
index 94f1021..007ba26 100644
--- a/adler32.c
+++ b/adler32.c
@@ -12,12 +12,13 @@
 #define NMAX 5552
 /* NMAX is the largest n such that 255n(n+1)/2 + (n+1)(BASE-1) <= 2^32-1 */
 
-#define DO1(buf,i)  {s1 += buf[i]; s2 += s1;}
+#define DO1(buf,i)  {adler += (buf)[i]; sum2 += adler;}
 #define DO2(buf,i)  DO1(buf,i); DO1(buf,i+1);
 #define DO4(buf,i)  DO2(buf,i); DO2(buf,i+2);
 #define DO8(buf,i)  DO4(buf,i); DO4(buf,i+4);
 #define DO16(buf)   DO8(buf,0); DO8(buf,8);
 
+/* use NO_DIVIDE if your processor does not do division in hardware */
 #ifdef NO_DIVIDE
 #  define MOD(a) \
     do { \
@@ -39,8 +40,17 @@
         if (a >= (BASE << 1)) a -= (BASE << 1); \
         if (a >= BASE) a -= BASE; \
     } while (0)
+#  define MOD4(a) \
+    do { \
+        if (a >= (BASE << 4)) a -= (BASE << 4); \
+        if (a >= (BASE << 3)) a -= (BASE << 3); \
+        if (a >= (BASE << 2)) a -= (BASE << 2); \
+        if (a >= (BASE << 1)) a -= (BASE << 1); \
+        if (a >= BASE) a -= BASE; \
+    } while (0)
 #else
 #  define MOD(a) a %= BASE
+#  define MOD4(a) a %= BASE
 #endif
 
 /* ========================================================================= */
@@ -49,48 +59,91 @@ uLong ZEXPORT adler32(adler, buf, len)
     const Bytef *buf;
     uInt len;
 {
-    unsigned long s1 = adler & 0xffff;
-    unsigned long s2 = (adler >> 16) & 0xffff;
-    int k;
+    unsigned long sum2;
+    unsigned n;
+
+    /* split Adler-32 into component sums */
+    sum2 = (adler >> 16) & 0xffff;
+    adler &= 0xffff;
+
+    /* in case user likes doing a byte at a time, keep it fast */
+    if (len == 1) {
+        adler += buf[0];
+        if (adler >= BASE)
+            adler -= BASE;
+        sum2 += adler;
+        if (sum2 >= BASE)
+            sum2 -= BASE;
+        return adler | (sum2 << 16);
+    }
 
-    if (buf == Z_NULL) return 1L;
+    /* initial Adler-32 value (deferred check for len == 1 speed) */
+    if (buf == Z_NULL)
+        return 1L;
 
-    while (len > 0) {
-        k = len < NMAX ? (int)len : NMAX;
-        len -= k;
-        while (k >= 16) {
+    /* in case short lengths are provided, keep it somewhat fast */
+    if (len < 16) {
+        while (len--) {
+            adler += *buf++;
+            sum2 += adler;
+        }
+        if (adler >= BASE)
+            adler -= BASE;
+        MOD4(sum2);             /* only added so many BASE's */
+        return adler | (sum2 << 16);
+    }
+
+    /* do length NMAX blocks -- requires just one modulo operation */
+    while (len >= NMAX) {
+        len -= NMAX;
+        n = NMAX / 16;          /* NMAX is divisible by 16 */
+        do {
+            DO16(buf);          /* 16 sums unrolled */
+            buf += 16;
+        } while (--n);
+        MOD(adler);
+        MOD(sum2);
+    }
+
+    /* do remaining bytes (less than NMAX, still just one modulo) */
+    if (len) {                  /* avoid modulos if none remaining */
+        while (len >= 16) {
+            len -= 16;
             DO16(buf);
             buf += 16;
-            k -= 16;
         }
-        if (k != 0) do {
-            s1 += *buf++;
-            s2 += s1;
-        } while (--k);
-        MOD(s1);
-        MOD(s2);
+        while (len--) {
+            adler += *buf++;
+            sum2 += adler;
+        }
+        MOD(adler);
+        MOD(sum2);
     }
-    return (s2 << 16) | s1;
+
+    /* return recombined sums */
+    return adler | (sum2 << 16);
 }
 
 /* ========================================================================= */
 uLong ZEXPORT adler32_combine(adler1, adler2, len2)
     uLong adler1;
     uLong adler2;
-    uLong len2;
+    z_off_t len2;
 {
-    unsigned long s1;
-    unsigned long s2;
+    unsigned long sum1;
+    unsigned long sum2;
+    unsigned rem;
 
-    len2 %= BASE;
-    s1 = adler1 & 0xffff;
-    s2 = len2 * s1;
-    MOD(s2);
-    s1 += (adler2 & 0xffff) + BASE - 1;
-    s2 += ((adler1 >> 16) & 0xffff) + ((adler2 >> 16) & 0xffff) + BASE - len2;
-    if (s1 > BASE) s1 -= BASE;
-    if (s1 > BASE) s1 -= BASE;
-    if (s2 > (BASE << 1)) s2 -= (BASE << 1);
-    if (s2 > BASE) s2 -= BASE;
-    return (s2 << 16) | s1;
+    /* the derivation of this formula is left as an exercise for the reader */
+    rem = (unsigned)(len2 % BASE);
+    sum1 = adler1 & 0xffff;
+    sum2 = rem * sum1;
+    MOD(sum2);
+    sum1 += (adler2 & 0xffff) + BASE - 1;
+    sum2 += ((adler1 >> 16) & 0xffff) + ((adler2 >> 16) & 0xffff) + BASE - rem;
+    if (sum1 > BASE) sum1 -= BASE;
+    if (sum1 > BASE) sum1 -= BASE;
+    if (sum2 > (BASE << 1)) sum2 -= (BASE << 1);
+    if (sum2 > BASE) sum2 -= BASE;
+    return sum1 | (sum2 << 16);
 }