Optimize pack()

Instead of using lookup tables, we can use a combination of shifts and
byte swapping to achieve the same thing in less cycles and with less
code.

Benchmark files
---------------

pack1.php:
```php
for ($i = 0; $i < 10_000_000; ++$i) {
    pack("J", 0x7FFFFFFFFFFFFFFF);
}
```

pack2.php:
```php
for ($i = 0; $i < 4000000; ++$i) {
    pack("nvc*", 0x1234, 0x5678, 65, 66);
}
```

On an i7-4790:
```
Benchmark 1: ./sapi/cli/php pack1.php
  Time (mean ± σ):     408.8 ms ±   3.4 ms    [User: 406.1 ms, System: 1.6 ms]
  Range (min … max):   403.6 ms … 413.6 ms    10 runs

Benchmark 2: ./sapi/cli/php_old pack1.php
  Time (mean ± σ):     451.7 ms ±   7.7 ms    [User: 448.5 ms, System: 2.0 ms]
  Range (min … max):   442.8 ms … 461.2 ms    10 runs

Summary
  ./sapi/cli/php pack1.php ran
    1.11 ± 0.02 times faster than ./sapi/cli/php_old pack1.php

Benchmark 1: ./sapi/cli/php pack2.php
  Time (mean ± σ):     239.3 ms ±   6.0 ms    [User: 236.2 ms, System: 2.3 ms]
  Range (min … max):   233.2 ms … 256.8 ms    12 runs

Benchmark 2: ./sapi/cli/php_old pack2.php
  Time (mean ± σ):     271.9 ms ±   3.3 ms    [User: 269.7 ms, System: 1.3 ms]
  Range (min … max):   267.4 ms … 279.0 ms    11 runs

Summary
  ./sapi/cli/php pack2.php ran
    1.14 ± 0.03 times faster than ./sapi/cli/php_old pack2.php
```

On an i7-1185G7:
```
Benchmark 1: ./sapi/cli/php pack1.php
  Time (mean ± σ):     263.7 ms ±   1.8 ms    [User: 262.6 ms, System: 0.9 ms]
  Range (min … max):   261.5 ms … 268.2 ms    11 runs

Benchmark 2: ./sapi/cli/php_old pack1.php
  Time (mean ± σ):     303.3 ms ±   6.5 ms    [User: 300.7 ms, System: 2.3 ms]
  Range (min … max):   297.4 ms … 318.1 ms    10 runs

Summary
  ./sapi/cli/php pack1.php ran
    1.15 ± 0.03 times faster than ./sapi/cli/php_old pack1.php

Benchmark 1: ./sapi/cli/php pack2.php
  Time (mean ± σ):     156.7 ms ±   2.9 ms    [User: 154.7 ms, System: 1.7 ms]
  Range (min … max):   151.6 ms … 164.7 ms    19 runs

Benchmark 2: ./sapi/cli/php_old pack2.php
  Time (mean ± σ):     174.6 ms ±   3.3 ms    [User: 171.9 ms, System: 2.3 ms]
  Range (min … max):   170.7 ms … 180.4 ms    17 runs

Summary
  ./sapi/cli/php pack2.php ran
    1.11 ± 0.03 times faster than ./sapi/cli/php_old pack2.php
```

Co-authored-by: [email protected]

Author: nielsdos

ext/standard/basic_functions.c

@@ -299,7 +299,6 @@ PHP_MINIT_FUNCTION(basic) /* {{{ */
 
 	BASIC_MINIT_SUBMODULE(var)
 	BASIC_MINIT_SUBMODULE(file)
-	BASIC_MINIT_SUBMODULE(pack)
 	BASIC_MINIT_SUBMODULE(browscap)
 	BASIC_MINIT_SUBMODULE(standard_filters)
 	BASIC_MINIT_SUBMODULE(user_filters)

ext/standard/pack.c

@@ -28,7 +28,6 @@
 #else
 #include <sys/param.h>
 #endif
-#include "pack.h"
 #ifdef HAVE_PWD_H
 #ifdef PHP_WIN32
 #include "win32/pwd.h"
@@ -50,10 +49,23 @@
 	} \
 	outputpos += (a)*(b);
 
+typedef enum {
+	PHP_LITTLE_ENDIAN,
+	PHP_BIG_ENDIAN,
+} php_pack_endianness;
+
 #ifdef WORDS_BIGENDIAN
-#define MACHINE_LITTLE_ENDIAN 0
+# define MACHINE_LITTLE_ENDIAN 0
+# define PHP_MACHINE_ENDIAN PHP_BIG_ENDIAN
 #else
-#define MACHINE_LITTLE_ENDIAN 1
+# define MACHINE_LITTLE_ENDIAN 1
+# define PHP_MACHINE_ENDIAN PHP_LITTLE_ENDIAN
+#endif
+
+#ifdef ZEND_ENABLE_ZVAL_LONG64
+# define PHP_LONG_BSWAP(u) ZEND_BYTES_SWAP64(u)
+#else
+# define PHP_LONG_BSWAP(u) ZEND_BYTES_SWAP32(u)
 #endif
 
 typedef ZEND_SET_ALIGNED(1, uint16_t unaligned_uint16_t);
@@ -62,41 +74,17 @@ typedef ZEND_SET_ALIGNED(1, uint64_t unaligned_uint64_t);
 typedef ZEND_SET_ALIGNED(1, unsigned int unaligned_uint);
 typedef ZEND_SET_ALIGNED(1, int unaligned_int);
 
-/* Mapping of byte from char (8bit) to long for machine endian */
-static int byte_map[1];
-
-/* Mappings of bytes from int (machine dependent) to int for machine endian */
-static int int_map[sizeof(int)];
-
-/* Mappings of bytes from shorts (16bit) for all endian environments */
-static int machine_endian_short_map[2];
-static int big_endian_short_map[2];
-static int little_endian_short_map[2];
-
-/* Mappings of bytes from longs (32bit) for all endian environments */
-static int machine_endian_long_map[4];
-static int big_endian_long_map[4];
-static int little_endian_long_map[4];
-
-#if SIZEOF_ZEND_LONG > 4
-/* Mappings of bytes from quads (64bit) for all endian environments */
-static int machine_endian_longlong_map[8];
-static int big_endian_longlong_map[8];
-static int little_endian_longlong_map[8];
-#endif
-
 /* {{{ php_pack */
-static void php_pack(zval *val, size_t size, int *map, char *output)
+static void php_pack(const zval *val, size_t size, php_pack_endianness endianness, char *output)
 {
-	size_t i;
-	char *v;
+	zend_ulong zl = zval_get_long(val);
 
-	convert_to_long(val);
-	v = (char *) &Z_LVAL_P(val);
-
-	for (i = 0; i < size; i++) {
-		*output++ = v[map[i]];
+	if ((endianness == PHP_LITTLE_ENDIAN) != MACHINE_LITTLE_ENDIAN) {
+		zl = PHP_LONG_BSWAP(zl);
+		zl >>= (sizeof(zl) - size) * 8;
 	}
+
+	memcpy(output, (const char *) &zl, size);
 }
 /* }}} */
 
@@ -529,7 +517,7 @@ PHP_FUNCTION(pack)
 			case 'c':
 			case 'C':
 				while (arg-- > 0) {
-					php_pack(&argv[currentarg++], 1, byte_map, &ZSTR_VAL(output)[outputpos]);
+					php_pack(&argv[currentarg++], 1, PHP_MACHINE_ENDIAN, &ZSTR_VAL(output)[outputpos]);
 					outputpos++;
 				}
 				break;
@@ -538,16 +526,16 @@ PHP_FUNCTION(pack)
 			case 'S':
 			case 'n':
 			case 'v': {
-				int *map = machine_endian_short_map;
+				php_pack_endianness endianness = PHP_MACHINE_ENDIAN;
 
 				if (code == 'n') {
-					map = big_endian_short_map;
+					endianness = PHP_BIG_ENDIAN;
 				} else if (code == 'v') {
-					map = little_endian_short_map;
+					endianness = PHP_LITTLE_ENDIAN;
 				}
 
 				while (arg-- > 0) {
-					php_pack(&argv[currentarg++], 2, map, &ZSTR_VAL(output)[outputpos]);
+					php_pack(&argv[currentarg++], 2, endianness, &ZSTR_VAL(output)[outputpos]);
 					outputpos += 2;
 				}
 				break;
@@ -556,7 +544,7 @@ PHP_FUNCTION(pack)
 			case 'i':
 			case 'I':
 				while (arg-- > 0) {
-					php_pack(&argv[currentarg++], sizeof(int), int_map, &ZSTR_VAL(output)[outputpos]);
+					php_pack(&argv[currentarg++], sizeof(int), PHP_MACHINE_ENDIAN, &ZSTR_VAL(output)[outputpos]);
 					outputpos += sizeof(int);
 				}
 				break;
@@ -565,16 +553,16 @@ PHP_FUNCTION(pack)
 			case 'L':
 			case 'N':
 			case 'V': {
-				int *map = machine_endian_long_map;
+				php_pack_endianness endianness = PHP_MACHINE_ENDIAN;
 
 				if (code == 'N') {
-					map = big_endian_long_map;
+					endianness = PHP_BIG_ENDIAN;
 				} else if (code == 'V') {
-					map = little_endian_long_map;
+					endianness = PHP_LITTLE_ENDIAN;
 				}
 
 				while (arg-- > 0) {
-					php_pack(&argv[currentarg++], 4, map, &ZSTR_VAL(output)[outputpos]);
+					php_pack(&argv[currentarg++], 4, endianness, &ZSTR_VAL(output)[outputpos]);
 					outputpos += 4;
 				}
 				break;
@@ -585,16 +573,16 @@ PHP_FUNCTION(pack)
 			case 'Q':
 			case 'J':
 			case 'P': {
-				int *map = machine_endian_longlong_map;
+				php_pack_endianness endianness = PHP_MACHINE_ENDIAN;
 
 				if (code == 'J') {
-					map = big_endian_longlong_map;
+					endianness = PHP_BIG_ENDIAN;
 				} else if (code == 'P') {
-					map = little_endian_longlong_map;
+					endianness = PHP_LITTLE_ENDIAN;
 				}
 
 				while (arg-- > 0) {
-					php_pack(&argv[currentarg++], 8, map, &ZSTR_VAL(output)[outputpos]);
+					php_pack(&argv[currentarg++], 8, endianness, &ZSTR_VAL(output)[outputpos]);
 					outputpos += 8;
 				}
 				break;
@@ -1199,127 +1187,3 @@ PHP_FUNCTION(unpack)
 	}
 }
 /* }}} */
-
-/* {{{ PHP_MINIT_FUNCTION */
-PHP_MINIT_FUNCTION(pack)
-{
-	int i;
-
-	if (MACHINE_LITTLE_ENDIAN) {
-		/* Where to get lo to hi bytes from */
-		byte_map[0] = 0;
-
-		for (i = 0; i < (int)sizeof(int); i++) {
-			int_map[i] = i;
-		}
-
-		machine_endian_short_map[0] = 0;
-		machine_endian_short_map[1] = 1;
-		big_endian_short_map[0] = 1;
-		big_endian_short_map[1] = 0;
-		little_endian_short_map[0] = 0;
-		little_endian_short_map[1] = 1;
-
-		machine_endian_long_map[0] = 0;
-		machine_endian_long_map[1] = 1;
-		machine_endian_long_map[2] = 2;
-		machine_endian_long_map[3] = 3;
-		big_endian_long_map[0] = 3;
-		big_endian_long_map[1] = 2;
-		big_endian_long_map[2] = 1;
-		big_endian_long_map[3] = 0;
-		little_endian_long_map[0] = 0;
-		little_endian_long_map[1] = 1;
-		little_endian_long_map[2] = 2;
-		little_endian_long_map[3] = 3;
-
-#if SIZEOF_ZEND_LONG > 4
-		machine_endian_longlong_map[0] = 0;
-		machine_endian_longlong_map[1] = 1;
-		machine_endian_longlong_map[2] = 2;
-		machine_endian_longlong_map[3] = 3;
-		machine_endian_longlong_map[4] = 4;
-		machine_endian_longlong_map[5] = 5;
-		machine_endian_longlong_map[6] = 6;
-		machine_endian_longlong_map[7] = 7;
-		big_endian_longlong_map[0] = 7;
-		big_endian_longlong_map[1] = 6;
-		big_endian_longlong_map[2] = 5;
-		big_endian_longlong_map[3] = 4;
-		big_endian_longlong_map[4] = 3;
-		big_endian_longlong_map[5] = 2;
-		big_endian_longlong_map[6] = 1;
-		big_endian_longlong_map[7] = 0;
-		little_endian_longlong_map[0] = 0;
-		little_endian_longlong_map[1] = 1;
-		little_endian_longlong_map[2] = 2;
-		little_endian_longlong_map[3] = 3;
-		little_endian_longlong_map[4] = 4;
-		little_endian_longlong_map[5] = 5;
-		little_endian_longlong_map[6] = 6;
-		little_endian_longlong_map[7] = 7;
-#endif
-	}
-	else {
-		zval val;
-		int size = sizeof(Z_LVAL(val));
-		Z_LVAL(val)=0; /*silence a warning*/
-
-		/* Where to get hi to lo bytes from */
-		byte_map[0] = size - 1;
-
-		for (i = 0; i < (int)sizeof(int); i++) {
-			int_map[i] = size - (sizeof(int) - i);
-		}
-
-		machine_endian_short_map[0] = size - 2;
-		machine_endian_short_map[1] = size - 1;
-		big_endian_short_map[0] = size - 2;
-		big_endian_short_map[1] = size - 1;
-		little_endian_short_map[0] = size - 1;
-		little_endian_short_map[1] = size - 2;
-
-		machine_endian_long_map[0] = size - 4;
-		machine_endian_long_map[1] = size - 3;
-		machine_endian_long_map[2] = size - 2;
-		machine_endian_long_map[3] = size - 1;
-		big_endian_long_map[0] = size - 4;
-		big_endian_long_map[1] = size - 3;
-		big_endian_long_map[2] = size - 2;
-		big_endian_long_map[3] = size - 1;
-		little_endian_long_map[0] = size - 1;
-		little_endian_long_map[1] = size - 2;
-		little_endian_long_map[2] = size - 3;
-		little_endian_long_map[3] = size - 4;
-
-#if SIZEOF_ZEND_LONG > 4
-		machine_endian_longlong_map[0] = size - 8;
-		machine_endian_longlong_map[1] = size - 7;
-		machine_endian_longlong_map[2] = size - 6;
-		machine_endian_longlong_map[3] = size - 5;
-		machine_endian_longlong_map[4] = size - 4;
-		machine_endian_longlong_map[5] = size - 3;
-		machine_endian_longlong_map[6] = size - 2;
-		machine_endian_longlong_map[7] = size - 1;
-		big_endian_longlong_map[0] = size - 8;
-		big_endian_longlong_map[1] = size - 7;
-		big_endian_longlong_map[2] = size - 6;
-		big_endian_longlong_map[3] = size - 5;
-		big_endian_longlong_map[4] = size - 4;
-		big_endian_longlong_map[5] = size - 3;
-		big_endian_longlong_map[6] = size - 2;
-		big_endian_longlong_map[7] = size - 1;
-		little_endian_longlong_map[0] = size - 1;
-		little_endian_longlong_map[1] = size - 2;
-		little_endian_longlong_map[2] = size - 3;
-		little_endian_longlong_map[3] = size - 4;
-		little_endian_longlong_map[4] = size - 5;
-		little_endian_longlong_map[5] = size - 6;
-		little_endian_longlong_map[6] = size - 7;
-		little_endian_longlong_map[7] = size - 8;
-#endif
-	}
-
-	return SUCCESS;
-}
-/* }}} */

ext/standard/pack.h

@@ -29,7 +29,6 @@
 #include "php_ext_syslog.h"
 #include "php_filestat.h"
 #include "php_browscap.h"
-#include "pack.h"
 #include "url.h"
 #include "pageinfo.h"
 #include "fsock.h"