mirror of https://git.tukaani.org/xz.git
Update tuklib_integer.h with bit scan functions.
Thanks to Joachim Henke for the original patch.
This commit is contained in:
parent
c74c132f7f
commit
7ac3985d89
|
@ -1,10 +1,12 @@
|
||||||
///////////////////////////////////////////////////////////////////////////////
|
///////////////////////////////////////////////////////////////////////////////
|
||||||
//
|
//
|
||||||
/// \file tuklib_integer.h
|
/// \file tuklib_integer.h
|
||||||
/// \brief Byte swapping and endianness related macros and functions
|
/// \brief Various integer and bit operations
|
||||||
///
|
///
|
||||||
/// This file provides macros or functions to do basic endianness related
|
/// This file provides macros or functions to do some basic integer and bit
|
||||||
/// integer operations (XX = 16, 32, or 64; Y = b or l):
|
/// operations.
|
||||||
|
///
|
||||||
|
/// Endianness related integer operations (XX = 16, 32, or 64; Y = b or l):
|
||||||
/// - Byte swapping: bswapXX(num)
|
/// - Byte swapping: bswapXX(num)
|
||||||
/// - Byte order conversions to/from native: convXXYe(num)
|
/// - Byte order conversions to/from native: convXXYe(num)
|
||||||
/// - Aligned reads: readXXYe(ptr)
|
/// - Aligned reads: readXXYe(ptr)
|
||||||
|
@ -18,8 +20,18 @@
|
||||||
/// \todo PowerPC and possibly some other architectures support
|
/// \todo PowerPC and possibly some other architectures support
|
||||||
/// byte swapping load and store instructions. This file
|
/// byte swapping load and store instructions. This file
|
||||||
/// doesn't take advantage of those instructions.
|
/// doesn't take advantage of those instructions.
|
||||||
|
///
|
||||||
|
/// Bit scan operations for non-zero 32-bit integers:
|
||||||
|
/// - Bit scan reverse (find highest non-zero bit): bsr32(num)
|
||||||
|
/// - Count leading zeros: clz32(num)
|
||||||
|
/// - Count trailing zeros: ctz32(num)
|
||||||
|
/// - Bit scan forward (simply an alias for ctz32()): bsf32(num)
|
||||||
|
///
|
||||||
|
/// The above bit scan operations return 0-31. If num is zero,
|
||||||
|
/// the result is undefined.
|
||||||
//
|
//
|
||||||
// Author: Lasse Collin
|
// Authors: Lasse Collin
|
||||||
|
// Joachim Henke
|
||||||
//
|
//
|
||||||
// This file has been put into the public domain.
|
// This file has been put into the public domain.
|
||||||
// You can do whatever you want with this file.
|
// You can do whatever you want with this file.
|
||||||
|
@ -213,8 +225,7 @@ read64le(const uint8_t *buf)
|
||||||
// to optimize byte swapping of constants when using glibc's or *BSD's
|
// to optimize byte swapping of constants when using glibc's or *BSD's
|
||||||
// byte swapping macros. The actual write is done in an inline function
|
// byte swapping macros. The actual write is done in an inline function
|
||||||
// to make type checking of the buf pointer possible similarly to readXXYe()
|
// to make type checking of the buf pointer possible similarly to readXXYe()
|
||||||
// functions. This also seems to silence a probably bogus GCC warning about
|
// functions.
|
||||||
// strict aliasing when buf points to the beginning of an uint8_t array.
|
|
||||||
|
|
||||||
#define write16be(buf, num) write16ne((buf), conv16be(num))
|
#define write16be(buf, num) write16ne((buf), conv16be(num))
|
||||||
#define write16le(buf, num) write16ne((buf), conv16le(num))
|
#define write16le(buf, num) write16ne((buf), conv16le(num))
|
||||||
|
@ -272,7 +283,7 @@ write64ne(uint8_t *buf, uint64_t num)
|
||||||
static inline uint16_t
|
static inline uint16_t
|
||||||
unaligned_read16be(const uint8_t *buf)
|
unaligned_read16be(const uint8_t *buf)
|
||||||
{
|
{
|
||||||
uint16_t num = ((uint16_t)buf[0] << 8) | buf[1];
|
uint16_t num = ((uint16_t)buf[0] << 8) | (uint16_t)buf[1];
|
||||||
return num;
|
return num;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -280,7 +291,7 @@ unaligned_read16be(const uint8_t *buf)
|
||||||
static inline uint16_t
|
static inline uint16_t
|
||||||
unaligned_read16le(const uint8_t *buf)
|
unaligned_read16le(const uint8_t *buf)
|
||||||
{
|
{
|
||||||
uint16_t num = ((uint32_t)buf[0]) | ((uint16_t)buf[1] << 8);
|
uint16_t num = ((uint16_t)buf[0]) | ((uint16_t)buf[1] << 8);
|
||||||
return num;
|
return num;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -347,4 +358,166 @@ unaligned_write32le(uint8_t *buf, uint32_t num)
|
||||||
}
|
}
|
||||||
|
|
||||||
#endif
|
#endif
|
||||||
|
|
||||||
|
|
||||||
|
static inline uint32_t
|
||||||
|
bsr32(uint32_t n)
|
||||||
|
{
|
||||||
|
// Check for ICC first, since it tends to define __GNUC__ too.
|
||||||
|
#if defined(__INTEL_COMPILER)
|
||||||
|
return _bit_scan_reverse(n);
|
||||||
|
|
||||||
|
#elif TUKLIB_GNUC_REQ(3, 4) && UINT_MAX == UINT32_MAX
|
||||||
|
// GCC >= 3.4 has __builtin_clz(), which gives good results on
|
||||||
|
// multiple architectures. On x86, __builtin_clz() ^ 31U becomes
|
||||||
|
// either plain BSR (so the XOR gets optimized away) or LZCNT and
|
||||||
|
// XOR (if -march indicates that SSE4a instructions are supported).
|
||||||
|
return __builtin_clz(n) ^ 31U;
|
||||||
|
|
||||||
|
#elif defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__))
|
||||||
|
uint32_t i;
|
||||||
|
__asm__("bsrl %1, %0" : "=r" (i) : "rm" (n));
|
||||||
|
return i;
|
||||||
|
|
||||||
|
#elif defined(_MSC_VER) && _MSC_VER >= 1400
|
||||||
|
// MSVC isn't supported by tuklib, but since this code exists,
|
||||||
|
// it doesn't hurt to have it here anyway.
|
||||||
|
uint32_t i;
|
||||||
|
_BitScanReverse((DWORD *)&i, n);
|
||||||
|
return i;
|
||||||
|
|
||||||
|
#else
|
||||||
|
uint32_t i = 31;
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0xFFFF0000)) == 0) {
|
||||||
|
n <<= 16;
|
||||||
|
i = 15;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0xFF000000)) == 0) {
|
||||||
|
n <<= 8;
|
||||||
|
i -= 8;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0xF0000000)) == 0) {
|
||||||
|
n <<= 4;
|
||||||
|
i -= 4;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0xC0000000)) == 0) {
|
||||||
|
n <<= 2;
|
||||||
|
i -= 2;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0x80000000)) == 0)
|
||||||
|
--i;
|
||||||
|
|
||||||
|
return i;
|
||||||
|
#endif
|
||||||
|
}
|
||||||
|
|
||||||
|
|
||||||
|
static inline uint32_t
|
||||||
|
clz32(uint32_t n)
|
||||||
|
{
|
||||||
|
#if defined(__INTEL_COMPILER)
|
||||||
|
return _bit_scan_reverse(n) ^ 31U;
|
||||||
|
|
||||||
|
#elif TUKLIB_GNUC_REQ(3, 4) && UINT_MAX == UINT32_MAX
|
||||||
|
return __builtin_clz(n);
|
||||||
|
|
||||||
|
#elif defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__))
|
||||||
|
uint32_t i;
|
||||||
|
__asm__("bsrl %1, %0\n\t"
|
||||||
|
"xorl $31, %0"
|
||||||
|
: "=r" (i) : "rm" (n));
|
||||||
|
return i;
|
||||||
|
|
||||||
|
#elif defined(_MSC_VER) && _MSC_VER >= 1400
|
||||||
|
uint32_t i;
|
||||||
|
_BitScanReverse((DWORD *)&i, n);
|
||||||
|
return i ^ 31U;
|
||||||
|
|
||||||
|
#else
|
||||||
|
uint32_t i = 0;
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0xFFFF0000)) == 0) {
|
||||||
|
n <<= 16;
|
||||||
|
i = 16;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0xFF000000)) == 0) {
|
||||||
|
n <<= 8;
|
||||||
|
i += 8;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0xF0000000)) == 0) {
|
||||||
|
n <<= 4;
|
||||||
|
i += 4;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0xC0000000)) == 0) {
|
||||||
|
n <<= 2;
|
||||||
|
i += 2;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0x80000000)) == 0)
|
||||||
|
++i;
|
||||||
|
|
||||||
|
return i;
|
||||||
|
#endif
|
||||||
|
}
|
||||||
|
|
||||||
|
|
||||||
|
static inline uint32_t
|
||||||
|
ctz32(uint32_t n)
|
||||||
|
{
|
||||||
|
#if defined(__INTEL_COMPILER)
|
||||||
|
return _bit_scan_forward(n);
|
||||||
|
|
||||||
|
#elif TUKLIB_GNUC_REQ(3, 4) && UINT_MAX >= UINT32_MAX
|
||||||
|
return __builtin_ctz(n);
|
||||||
|
|
||||||
|
#elif defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__))
|
||||||
|
uint32_t i;
|
||||||
|
__asm__("bsfl %1, %0" : "=r" (i) : "rm" (n));
|
||||||
|
return i;
|
||||||
|
|
||||||
|
#elif defined(_MSC_VER) && _MSC_VER >= 1400
|
||||||
|
uint32_t i;
|
||||||
|
_BitScanForward((DWORD *)&i, n);
|
||||||
|
return i;
|
||||||
|
|
||||||
|
#else
|
||||||
|
uint32_t i = 0;
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0x0000FFFF)) == 0) {
|
||||||
|
n >>= 16;
|
||||||
|
i = 16;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0x000000FF)) == 0) {
|
||||||
|
n >>= 8;
|
||||||
|
i += 8;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0x0000000F)) == 0) {
|
||||||
|
n >>= 4;
|
||||||
|
i += 4;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0x00000003)) == 0) {
|
||||||
|
n >>= 2;
|
||||||
|
i += 2;
|
||||||
|
}
|
||||||
|
|
||||||
|
if ((n & UINT32_C(0x00000001)) == 0)
|
||||||
|
++i;
|
||||||
|
|
||||||
|
return i;
|
||||||
|
#endif
|
||||||
|
}
|
||||||
|
|
||||||
|
#define bsf32 ctz32
|
||||||
|
|
||||||
#endif
|
#endif
|
||||||
|
|
Loading…
Reference in New Issue