zlib: add cloudflare/zlib

configure.py

@@ -376,6 +376,18 @@
     dest='shared_openssl_libpath',
     help='a directory to search for the shared OpenSSL DLLs')
 
+shared_optgroup.add_argument('--cloudflare-zlib',
+    action='store_true',
+    dest='use_cl_zlib',
+    default=None,
+    help='use cloudflare/zlib')
+
+shared_optgroup.add_argument('--no-cloudflare-zlib',
+    action='store_false',
+    dest='use_cl_zlib',
+    default=None,
+    help='do not use use cloudflare/zlib')
+
 shared_optgroup.add_argument('--shared-zlib',
     action='store_true',
     dest='shared_zlib',
@@ -1522,6 +1534,13 @@ def configure_library(lib, output, pkgname=None):
     elif pkg_libs:
       output['libraries'] += pkg_libs.split()
 
+def configure_zlib(o):
+  configure_library('zlib', o)
+  if (options.use_cl_zlib is None and o['variables']['target_arch'] == 'arm64'
+      or options.use_cl_zlib):
+    o['variables']['zlib_root'] = 'deps/cl-zlib'
+  else:
+    o['variables']['zlib_root'] = 'deps/zlib'
 
 def configure_v8(o):
   o['variables']['v8_enable_webassembly'] = 0 if options.v8_lite_mode else 1
@@ -2067,7 +2086,7 @@ def make_bin_override():
 configure_node(output)
 configure_node_lib_files(output)
 configure_napi(output)
-configure_library('zlib', output)
+configure_zlib(output)
 configure_library('http_parser', output)
 configure_library('libuv', output)
 configure_library('brotli', output, pkgname=['libbrotlidec', 'libbrotlienc'])

deps/cl-zlib/adler32.c

@@ -0,0 +1,188 @@
+/* adler32.c -- compute the Adler-32 checksum of a data stream
+ * Copyright (C) 1995-2011 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* @(#) $Id$ */
+
+#include "zutil.h"
+
+#define local static
+
+local uLong adler32_combine_ OF((uLong adler1, uLong adler2, z_off64_t len2));
+
+#define BASE 65521      /* largest prime smaller than 65536 */
+#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)  {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 --
+   try it both ways to see which is faster */
+#ifdef NO_DIVIDE
+/* note that this assumes BASE is 65521, where 65536 % 65521 == 15
+   (thank you to John Reiser for pointing this out) */
+#  define CHOP(a) \
+    do { \
+        unsigned long tmp = a >> 16; \
+        a &= 0xffffUL; \
+        a += (tmp << 4) - tmp; \
+    } while (0)
+#  define MOD28(a) \
+    do { \
+        CHOP(a); \
+        if (a >= BASE) a -= BASE; \
+    } while (0)
+#  define MOD(a) \
+    do { \
+        CHOP(a); \
+        MOD28(a); \
+    } while (0)
+#  define MOD63(a) \
+    do { /* this assumes a is not negative */ \
+        z_off64_t tmp = a >> 32; \
+        a &= 0xffffffffL; \
+        a += (tmp << 8) - (tmp << 5) + tmp; \
+        tmp = a >> 16; \
+        a &= 0xffffL; \
+        a += (tmp << 4) - tmp; \
+        tmp = a >> 16; \
+        a &= 0xffffL; \
+        a += (tmp << 4) - tmp; \
+        if (a >= BASE) a -= BASE; \
+    } while (0)
+#else
+#  define MOD(a) a %= BASE
+#  define MOD28(a) a %= BASE
+#  define MOD63(a) a %= BASE
+#endif
+
+#if defined(ADLER32_SIMD_NEON) || defined (ADLER32_SIMD_SSSE3)
+#include "adler32_simd.h"
+#endif
+
+/* ========================================================================= */
+uLong ZEXPORT adler32(adler, buf, len)
+    uLong adler;
+    const Bytef *buf;
+    uInt len;
+{
+    unsigned long sum2;
+    unsigned n;
+
+#if defined(ADLER32_SIMD_NEON) || defined(ADLER32_SIMD_SSSE3)
+    if (buf && len >= 64)
+        return adler32_simd_(adler, buf, len);
+#endif
+
+    /* 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);
+    }
+
+    /* initial Adler-32 value (deferred check for len == 1 speed) */
+    if (buf == Z_NULL)
+        return 1L;
+
+    /* in case short lengths are provided, keep it somewhat fast */
+    if (len < 16) {
+        while (len--) {
+            adler += *buf++;
+            sum2 += adler;
+        }
+        if (adler >= BASE)
+            adler -= BASE;
+        MOD28(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;
+        }
+        while (len--) {
+            adler += *buf++;
+            sum2 += adler;
+        }
+        MOD(adler);
+        MOD(sum2);
+    }
+
+    /* return recombined sums */
+    return adler | (sum2 << 16);
+}
+
+/* ========================================================================= */
+local uLong adler32_combine_(adler1, adler2, len2)
+    uLong adler1;
+    uLong adler2;
+    z_off64_t len2;
+{
+    unsigned long sum1;
+    unsigned long sum2;
+    unsigned rem;
+
+    /* for negative len, return invalid adler32 as a clue for debugging */
+    if (len2 < 0)
+        return 0xffffffffUL;
+
+    /* the derivation of this formula is left as an exercise for the reader */
+    MOD63(len2);                /* assumes len2 >= 0 */
+    rem = (unsigned)len2;
+    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);
+}
+
+/* ========================================================================= */
+uLong ZEXPORT adler32_combine(adler1, adler2, len2)
+    uLong adler1;
+    uLong adler2;
+    z_off_t len2;
+{
+    return adler32_combine_(adler1, adler2, len2);
+}
+
+uLong ZEXPORT adler32_combine64(adler1, adler2, len2)
+    uLong adler1;
+    uLong adler2;
+    z_off64_t len2;
+{
+    return adler32_combine_(adler1, adler2, len2);
+}