[mLib] / bits.h

/* -*-c-*-
 *
 * $Id$
 *
 * Portable bit-level manipulation macros
 *
 * (c) 1998 Straylight/Edgeware
 */

/*----- Licensing notice --------------------------------------------------* 
 *
 * This file is part of the mLib utilities library.
 *
 * mLib is free software; you can redistribute it and/or modify
 * it under the terms of the GNU Library General Public License as
 * published by the Free Software Foundation; either version 2 of the
 * License, or (at your option) any later version.
 * 
 * mLib is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU Library General Public License for more details.
 * 
 * You should have received a copy of the GNU Library General Public
 * License along with mLib; if not, write to the Free
 * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
 * MA 02111-1307, USA.
 */

#ifndef MLIB_BITS_H
#define MLIB_BITS_H

#ifdef __cplusplus
  extern "C" {
#endif

/*----- Header files ------------------------------------------------------*/

#include <limits.h>
#include <stddef.h>
#if __STDC_VERSION__ >= 199900l
#  include <stdint.h>
#endif

/*----- Decide on some types ----------------------------------------------*/

/* --- Make GNU C shut up --- */

#if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 91)
#  define MLIB_BITS_EXTENSION __extension__
#else
#  define MLIB_BITS_EXTENSION
#endif

/* --- Decide on a 32-bit type --- *
 *
 * I want a type which is capable of expressing 32-bit numbers.  Because some
 * implementations have 64-bit @long@s (infinitely preferable to the abortion
 * that is @long long@), using @unsigned long@ regardless is wasteful.  So,
 * if @int@ appears to be good enough, then I'll go with that.
 */

#if UINT_MAX >= 0xffffffffu
  typedef unsigned int uint32;
#else
  typedef unsigned long uint32;
#endif

/* --- Decide on a 64-bit type --- *
 *
 * The test is quite subtle.  Think about it.  Note that (at least on my
 * machine), the 32-bit macros are *much* faster than GCC's @long long@
 * support.
 */

#if defined(ULONG_LONG_MAX) && !defined(ULLONG_MAX)
#  define ULLONG_MAX ULONG_LONG_MAX
#endif

#if UINT_MAX >> 31 > 0xffffffff
#  define HAVE_UINT64
   typedef unsigned int uint64;
#elif ULONG_MAX >> 31 > 0xffffffff
#  define HAVE_UINT64
   typedef unsigned long uint64;
#elif defined(ULLONG_MAX)
#  define HAVE_UINT64
   MLIB_BITS_EXTENSION typedef unsigned long long uint64;
#endif

#ifdef DEBUG64
#  undef HAVE_UINT64
#endif

#ifdef HAVE_UINT64
  typedef struct { uint64 i; } kludge64;
#else
  typedef struct { uint32 hi, lo; } kludge64;
#endif

/* --- Decide on a 24-bit type --- */

#if UINT_MAX >= 0x00ffffffu
  typedef unsigned int uint24;
#else
  typedef unsigned long uint24;
#endif

/* --- Decide on 16-bit and 8-bit types --- *
 *
 * This is more for brevity than anything else.
 */

typedef unsigned short uint16;
typedef unsigned char octet;

/* --- WARNING! --- *
 *
 * Never lose sight of the fact that the above types may be wider than the
 * names suggest.  Some architectures have 32-bit @short@s for example.
 */

/*----- Macros ------------------------------------------------------------*/

/* --- Useful masks --- */

#define MASK8 0xffu
#define MASK16 0xffffu
#define MASK16_L 0xffffu
#define MASK16_B 0xffffu
#define MASK24 0xffffffu
#define MASK24_L 0xffffffu
#define MASK24_B 0xffffffu
#define MASK32 0xffffffffu
#define MASK32_L 0xffffffffu
#define MASK32_B 0xffffffffu

#ifdef HAVE_UINT64
#  define MASK64 MLIB_BITS_EXTENSION 0xffffffffffffffffu
#endif

/* --- Sizes --- */

#define SZ_8 1
#define SZ_16 2
#define SZ_16_L 2
#define SZ_16_B 2
#define SZ_24 3
#define SZ_24_L 3
#define SZ_24_B 3
#define SZ_32 4
#define SZ_32_L 4
#define SZ_32_B 4

/* --- Type coercions --- */

#define U8(x) ((octet)((x) & MASK8))
#define U16(x) ((uint16)((x) & MASK16))
#define U24(x) ((uint24)((x) & MASK24))
#define U32(x) ((uint32)((x) & MASK32))

#ifdef HAVE_UINT64
#  define U64(x) ((uint64)(x) & MASK64)
#  define U64_(d, x) ((d).i = U64(x).i)
#else
#  define U64_(d, x) ((d).hi = U32((x).hi), (d).lo = U32((x).lo))
#endif

/* --- Safe shifting macros --- */

#define LSL8(v, s) U8(U8(v) << ((s) & 7u))
#define LSR8(v, s) U8(U8(v) >> ((s) & 7u))
#define LSL16(v, s) U16(U16(v) << ((s) & 15u))
#define LSR16(v, s) U16(U16(v) >> ((s) & 15u))
#define LSL24(v, s) U24(U24(v) << ((s) % 24u))
#define LSR24(v, s) U24(U24(v) >> ((s) % 24u))
#define LSL32(v, s) U32(U32(v) << ((s) & 31u))
#define LSR32(v, s) U32(U32(v) >> ((s) & 31u))

#ifdef HAVE_UINT64
#  define LSL64(v, s) U64(U64(v) << ((s) & 63u))
#  define LSR64(v, s) U64(U64(v) >> ((s) & 63u))
#  define LSL64_(d, v, s) ((d).i = LSL64((v).i, (s)))
#  define LSR64_(d, v, s) ((d).i = LSR64((v).i, (s)))
#else
#  define LSL64_(d, v, s) do {						\
     unsigned _s = (s) & 63u;						\
     uint32 _l = (v).lo, _h = (v).hi;					\
     kludge64 *_d = &(d);						\
     if (_s >= 32) {							\
       _d->hi = LSL32(_l, _s - 32u);					\
       _d->lo = 0;							\
     } else if (!_s) {							\
       _d->lo = _l;							\
       _d->hi = _h;							\
     } else {								\
       _d->hi = LSL32(_h, _s) | LSR32(_l, 32u - _s);			\
       _d->lo = LSL32(_l, _s);						\
     }									\
   } while (0)
#  define LSR64_(d, v, s) do {						\
     unsigned _s = (s) & 63u;						\
     uint32 _l = (v).lo, _h = (v).hi;					\
     kludge64 *_d = &(d);						\
     if (_s >= 32) {							\
       _d->lo = LSR32(_h, _s - 32u);					\
       _d->hi = 0;							\
     } else if (!_s) {							\
       _d->lo = _l;							\
       _d->hi = _h;							\
     } else {								\
       _d->lo = LSR32(_l, _s) | LSL32(_h, 32u - _s);			\
       _d->hi = LSR32(_h, _s);						\
     }									\
   } while (0)
#endif

/* --- Rotation macros --- */

#define ROL8(v, s) (LSL8((v), (s)) | (LSR8((v), 8u - (s))))
#define ROR8(v, s) (LSR8((v), (s)) | (LSL8((v), 8u - (s))))
#define ROL16(v, s) (LSL16((v), (s)) | (LSR16((v), 16u - (s))))
#define ROR16(v, s) (LSR16((v), (s)) | (LSL16((v), 16u - (s))))
#define ROL24(v, s) (LSL24((v), (s)) | (LSR24((v), 24u - (s))))
#define ROR24(v, s) (LSR24((v), (s)) | (LSL24((v), 24u - (s))))
#define ROL32(v, s) (LSL32((v), (s)) | (LSR32((v), 32u - (s))))
#define ROR32(v, s) (LSR32((v), (s)) | (LSL32((v), 32u - (s))))

#ifdef HAVE_UINT64
#  define ROL64(v, s) (LSL64((v), (s)) | (LSR64((v), 64u - (s))))
#  define ROR64(v, s) (LSR64((v), (s)) | (LSL64((v), 64u - (s))))
#  define ROL64_(d, v, s) ((d).i = ROL64((v).i, (s)))
#  define ROR64_(d, v, s) ((d).i = ROR64((v).i, (s)))
#else
#  define ROL64_(d, v, s) do {						\
     unsigned _s = (s) & 63u;						\
     uint32 _l = (v).lo, _h = (v).hi;					\
     kludge64 *_d = &(d);						\
     if (_s > 32) {							\
       _d->hi = LSL32(_l, _s - 32u) | LSR32(_h, 64u - _s);		\
       _d->lo = LSL32(_h, _s - 32u) | LSR32(_l, 64u - _s);		\
     } else if (!_s) {							\
       _d->lo = _l;							\
       _d->hi = _h;							\
     } else if (_s == 32) {						\
       _d->lo = _h;							\
       _d->hi = _l;							\
     } else {								\
       _d->hi = LSL32(_h, _s) | LSR32(_l, 32u - _s);			\
       _d->lo = LSL32(_l, _s) | LSR32(_h, 32u - _s);			\
     }									\
   } while (0)
#  define ROR64_(d, v, s) do {						\
     unsigned _s = (s) & 63u;						\
     uint32 _l = (v).lo, _h = (v).hi;					\
     kludge64 *_d = &(d);						\
     if (_s > 32) {							\
       _d->hi = LSR32(_l, _s - 32u) | LSL32(_h, 64u - _s);		\
       _d->lo = LSR32(_h, _s - 32u) | LSL32(_l, 64u - _s);		\
     } else if (!_s) {							\
       _d->lo = _l;							\
       _d->hi = _h;							\
     } else if (_s == 32) {						\
       _d->lo = _h;							\
       _d->hi = _l;							\
     } else {								\
       _d->hi = LSR32(_h, _s) | LSL32(_l, 32u - _s);			\
       _d->lo = LSR32(_l, _s) | LSL32(_h, 32u - _s);			\
     }									\
   } while (0)
#endif

/* --- Storage and retrieval --- */

#define GETBYTE(p, o) (((octet *)(p))[o] & MASK8)
#define PUTBYTE(p, o, v) (((octet *)(p))[o] = U8((v)))

#define LOAD8(p) (GETBYTE((p), 0))
#define STORE8(p, v) (PUTBYTE((p), 0, (v)))

#define LOAD16_B(p)							\
  (((uint16)GETBYTE((p), 0) << 8) |					\
   ((uint16)GETBYTE((p), 1) << 0))
#define LOAD16_L(p)							\
  (((uint16)GETBYTE((p), 0) << 0) |					\
   ((uint16)GETBYTE((p), 1) << 8))
#define LOAD16(p) LOAD16_B((p))

#define STORE16_B(p, v)							\
  (PUTBYTE((p), 0, (uint16)(v) >> 8),					\
   PUTBYTE((p), 1, (uint16)(v) >> 0))
#define STORE16_L(p, v)							\
  (PUTBYTE((p), 0, (uint16)(v) >> 0),					\
   PUTBYTE((p), 1, (uint16)(v) >> 8))
#define STORE16(p, v) STORE16_B((p), (v))

#define LOAD24_B(p)							\
  (((uint24)GETBYTE((p), 0) << 16) |					\
   ((uint24)GETBYTE((p), 1) <<  8) |					\
   ((uint24)GETBYTE((p), 2) <<  0))
#define LOAD24_L(p)							\
  (((uint24)GETBYTE((p), 0) <<  0) |					\
   ((uint24)GETBYTE((p), 1) <<  8) |					\
   ((uint24)GETBYTE((p), 2) << 16))
#define LOAD24(p) LOAD24_B((p))

#define STORE24_B(p, v)							\
  (PUTBYTE((p), 0, (uint24)(v) >> 16),					\
   PUTBYTE((p), 1, (uint24)(v) >>  8),					\
   PUTBYTE((p), 2, (uint24)(v) >>  0))
#define STORE24_L(p, v)							\
  (PUTBYTE((p), 0, (uint24)(v) >>  0),					\
   PUTBYTE((p), 1, (uint24)(v) >>  8),					\
   PUTBYTE((p), 2, (uint24)(v) >> 16))
#define STORE24(p, v) STORE24_B((p), (v))

#define LOAD32_B(p)							\
  (((uint32)GETBYTE((p), 0) << 24) |					\
   ((uint32)GETBYTE((p), 1) << 16) |					\
   ((uint32)GETBYTE((p), 2) <<  8) |					\
   ((uint32)GETBYTE((p), 3) <<  0))
#define LOAD32_L(p)							\
  (((uint32)GETBYTE((p), 0) <<  0) |					\
   ((uint32)GETBYTE((p), 1) <<  8) |					\
   ((uint32)GETBYTE((p), 2) << 16) |					\
   ((uint32)GETBYTE((p), 3) << 24))
#define LOAD32(p) LOAD32_B((p))

#define STORE32_B(p, v)							\
  (PUTBYTE((p), 0, (uint32)(v) >> 24),					\
   PUTBYTE((p), 1, (uint32)(v) >> 16),					\
   PUTBYTE((p), 2, (uint32)(v) >>  8),					\
   PUTBYTE((p), 3, (uint32)(v) >>  0))
#define STORE32_L(p, v)							\
  (PUTBYTE((p), 0, (uint32)(v) >>  0),					\
   PUTBYTE((p), 1, (uint32)(v) >>  8),					\
   PUTBYTE((p), 2, (uint32)(v) >> 16),					\
   PUTBYTE((p), 3, (uint32)(v) >> 24))
#define STORE32(p, v) STORE32_B((p), (v))

#ifdef HAVE_UINT64

#  define LOAD64_B(p)							\
    (((uint64)GETBYTE((p), 0) << 56) |					\
     ((uint64)GETBYTE((p), 1) << 48) |					\
     ((uint64)GETBYTE((p), 2) << 40) |					\
     ((uint64)GETBYTE((p), 3) << 32) |					\
     ((uint64)GETBYTE((p), 4) << 24) |					\
     ((uint64)GETBYTE((p), 5) << 16) |					\
     ((uint64)GETBYTE((p), 6) <<  8) |					\
     ((uint64)GETBYTE((p), 7) <<  0))
#  define LOAD64_L(p)							\
    (((uint64)GETBYTE((p), 0) <<  0) |					\
     ((uint64)GETBYTE((p), 1) <<  8) |					\
     ((uint64)GETBYTE((p), 2) << 16) |					\
     ((uint64)GETBYTE((p), 3) << 24) |					\
     ((uint64)GETBYTE((p), 4) << 32) |					\
     ((uint64)GETBYTE((p), 5) << 40) |					\
     ((uint64)GETBYTE((p), 6) << 48) |					\
     ((uint64)GETBYTE((p), 7) << 56))
#  define LOAD64(p) LOAD64_B((p))
#  define LOAD64_B_(d, p) ((d).i = LOAD64_B((p)))
#  define LOAD64_L_(d, p) ((d).i = LOAD64_L((p)))
#  define LOAD64_(d, p) LOAD64_B_((d), (p))

#  define STORE64_B(p, v)						\
    (PUTBYTE((p), 0, (uint64)(v) >> 56),				\
     PUTBYTE((p), 1, (uint64)(v) >> 48),				\
     PUTBYTE((p), 2, (uint64)(v) >> 40),				\
     PUTBYTE((p), 3, (uint64)(v) >> 32),				\
     PUTBYTE((p), 4, (uint64)(v) >> 24),				\
     PUTBYTE((p), 5, (uint64)(v) >> 16),				\
     PUTBYTE((p), 6, (uint64)(v) >>  8),				\
     PUTBYTE((p), 7, (uint64)(v) >>  0))
#  define STORE64_L(p, v)						\
    (PUTBYTE((p), 0, (uint64)(v) >>  0),				\
     PUTBYTE((p), 1, (uint64)(v) >>  8),				\
     PUTBYTE((p), 2, (uint64)(v) >> 16),				\
     PUTBYTE((p), 3, (uint64)(v) >> 24),				\
     PUTBYTE((p), 4, (uint64)(v) >> 32),				\
     PUTBYTE((p), 5, (uint64)(v) >> 40),				\
     PUTBYTE((p), 6, (uint64)(v) >> 48),				\
     PUTBYTE((p), 7, (uint64)(v) >> 56))
#  define STORE64(p, v) STORE64_B((p), (v))
#  define STORE64_B_(p, v) STORE64_B((p), (v).i)
#  define STORE64_L_(p, v) STORE64_L((p), (v).i)
#  define STORE64_(p, v) STORE64_B_((p), (v))

#else

#  define LOAD64_B_(d, p)						\
    ((d).hi = LOAD32_B((octet *)(p) + 0),				\
     (d).lo = LOAD32_B((octet *)(p) + 4))
#  define LOAD64_L_(d, p)						\
    ((d).lo = LOAD32_L((octet *)(p) + 0),				\
     (d).hi = LOAD32_L((octet *)(p) + 4))
#  define LOAD64_(d, p) LOAD64_B_((d), (p))

#  define STORE64_B_(p, v)						\
    (STORE32_B((octet *)(p) + 0, (v).hi),				\
     STORE32_B((octet *)(p) + 4, (v).lo))
#  define STORE64_L_(p, v)						\
    (STORE32_L((octet *)(p) + 0, (v).lo),				\
     STORE32_L((octet *)(p) + 4, (v).hi))
#  define STORE64_(p, v) STORE64_B_((p), (v))

#endif

/* --- Other operations on 64-bit integers --- */

#ifdef HAVE_UINT64
#  define SET64(d, h, l) ((d).i = (U64((h)) << 32) | U64((l)))
#  define ASSIGN64(d, x) ((d).i = U64((x)))
#  define HI64(x) U32((x).i >> 32)
#  define LO64(x) U32((x).i)
#  define GET64(t, x) ((t)(x).i)
#else
#  define SET64(d, h, l) ((d).hi = U32(h), (d).lo = U32(l))
#  define ASSIGN64(d, x)						\
    ((d).hi = ((x & ~MASK32) >> 16) >> 16, (d).lo = U32(x))
#  define HI64(x) U32((x).hi)
#  define LO64(x) U32((x).lo)
#  define GET64(t, x) (((((t)HI64(x) << 16) << 16) & ~MASK32) | (t)LO64(x))
#endif

#ifdef HAVE_UINT64
#  define AND64(d, x, y) ((d).i = (x).i & (y).i)
#  define OR64(d, x, y) ((d).i = (x).i | (y).i)
#  define XOR64(d, x, y) ((d).i = (x).i ^ (y).i)
#  define CPL64(d, x) ((d).i = ~(x).i)
#  define ADD64(d, x, y) ((d).i = (x).i + (y).i)
#  define SUB64(d, x, y) ((d).i = (x).i - (y).i)
#  define CMP64(x, op, y) ((x).i op (y).i)
#  define ZERO64(x) ((x) == 0)
#else
#  define AND64(d, x, y) ((d).lo = (x).lo & (y).lo, (d).hi = (x).hi & (y).hi)
#  define OR64(d, x, y) ((d).lo = (x).lo | (y).lo, (d).hi = (x).hi | (y).hi)
#  define XOR64(d, x, y) ((d).lo = (x).lo ^ (y).lo, (d).hi = (x).hi ^ (y).hi)
#  define CPL64(d, x) ((d).lo = ~(x).lo, (d).hi = ~(x).hi)
#  define ADD64(d, x, y) do {						\
     uint32 _x = U32((x).lo + (y).lo);					\
     (d).hi = (x).hi + (y).hi + (_x < (x).lo);				\
     (d).lo = _x;							\
   } while (0)
#  define SUB64(d, x, y) do {						\
     uint32 _x = U32((x).lo - (y).lo);					\
     (d).hi = (x).hi - (y).hi - (_x > (x).lo);				\
     (d).lo = _x;							\
   } while (0)
#  define CMP64(x, op, y)						\
    ((x).hi == (y).hi ? (x).lo op (y).lo : (x).hi op (y).hi)
#  define ZERO64(x) ((x).lo == 0 && (x).hi == 0)
#endif

/* --- Storing integers in tables --- */

#ifdef HAVE_UINT64
#  define X64(x, y) { 0x##x##y }
#else
#  define X64(x, y) { 0x##x, 0x##y }
#endif
    
/*----- That's all, folks -------------------------------------------------*/

#ifdef __cplusplus
  }
#endif

#endif
Commit	Line	Data
6b033bc4	1	/* --c--
6b033bc4	2	*
5892fd39	3	* $Id$
6b033bc4	4	*
	5	* Portable bit-level manipulation macros
	6	*
	7	* (c) 1998 Straylight/Edgeware
	8	*/
	9
	10	/----- Licensing notice --------------------------------------------------
	11	*
	12	* This file is part of the mLib utilities library.
	13	*
	14	* mLib is free software; you can redistribute it and/or modify
	15	* it under the terms of the GNU Library General Public License as
	16	* published by the Free Software Foundation; either version 2 of the
	17	* License, or (at your option) any later version.
	18	*
	19	* mLib is distributed in the hope that it will be useful,
	20	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	21	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	22	* GNU Library General Public License for more details.
	23	*
	24	* You should have received a copy of the GNU Library General Public
	25	* License along with mLib; if not, write to the Free
	26	* Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
	27	* MA 02111-1307, USA.
	28	*/
	29
c6e0eaf0	30	#ifndef MLIB_BITS_H
c6e0eaf0	31	#define MLIB_BITS_H
6b033bc4	32
	33	#ifdef __cplusplus
	34	extern "C" {
	35	#endif
	36
	37	/----- Header files ------------------------------------------------------/
	38
	39	#include <limits.h>
	40	#include <stddef.h>
a6f4a484	41	#if __STDC_VERSION__ >= 199900l
	42	# include <stdint.h>
	43	#endif
6b033bc4	44
	45	/----- Decide on some types ----------------------------------------------/
	46
1f175471	47	/* --- Make GNU C shut up --- */
	48
	49	#if __GNUC__ > 2 \|\| (__GNUC__ == 2 && __GNUC_MINOR__ >= 91)
	50	# define MLIB_BITS_EXTENSION __extension__
	51	#else
	52	# define MLIB_BITS_EXTENSION
	53	#endif
	54
6b033bc4	55	/* --- Decide on a 32-bit type --- *
	56	*
	57	* I want a type which is capable of expressing 32-bit numbers. Because some
	58	* implementations have 64-bit @long@s (infinitely preferable to the abortion
	59	* that is @long long@), using @unsigned long@ regardless is wasteful. So,
	60	* if @int@ appears to be good enough, then I'll go with that.
	61	*/
	62
	63	#if UINT_MAX >= 0xffffffffu
	64	typedef unsigned int uint32;
	65	#else
	66	typedef unsigned long uint32;
	67	#endif
	68
a6f4a484	69	/* --- Decide on a 64-bit type --- *
	70	*
	71	* The test is quite subtle. Think about it. Note that (at least on my
	72	* machine), the 32-bit macros are much faster than GCC's @long long@
	73	* support.
	74	*/
	75
	76	#if defined(ULONG_LONG_MAX) && !defined(ULLONG_MAX)
	77	# define ULLONG_MAX ULONG_LONG_MAX
	78	#endif
	79
	80	#if UINT_MAX >> 31 > 0xffffffff
1f175471	81	# define HAVE_UINT64
1f175471	82	typedef unsigned int uint64;
a6f4a484	83	#elif ULONG_MAX >> 31 > 0xffffffff
1f175471	84	# define HAVE_UINT64
1f175471	85	typedef unsigned long uint64;
a6f4a484	86	#elif defined(ULLONG_MAX)
1f175471	87	# define HAVE_UINT64
1f175471	88	MLIB_BITS_EXTENSION typedef unsigned long long uint64;
a6f4a484	89	#endif
	90
	91	#ifdef DEBUG64
	92	# undef HAVE_UINT64
	93	#endif
	94
	95	#ifdef HAVE_UINT64
	96	typedef struct { uint64 i; } kludge64;
	97	#else
	98	typedef struct { uint32 hi, lo; } kludge64;
	99	#endif
	100
6a0129ea	101	/* --- Decide on a 24-bit type --- */
	102
	103	#if UINT_MAX >= 0x00ffffffu
	104	typedef unsigned int uint24;
	105	#else
	106	typedef unsigned long uint24;
	107	#endif
	108
6b033bc4	109	/* --- Decide on 16-bit and 8-bit types --- *
	110	*
	111	* This is more for brevity than anything else.
	112	*/
	113
	114	typedef unsigned short uint16;
	115	typedef unsigned char octet;
	116
	117	/* --- WARNING! --- *
	118	*
	119	* Never lose sight of the fact that the above types may be wider than the
	120	* names suggest. Some architectures have 32-bit @short@s for example.
	121	*/
	122
	123	/----- Macros ------------------------------------------------------------/
	124
	125	/* --- Useful masks --- */
	126
	127	#define MASK8 0xffu
	128	#define MASK16 0xffffu
76a7638e	129	#define MASK16_L 0xffffu
76a7638e	130	#define MASK16_B 0xffffu
6a0129ea	131	#define MASK24 0xffffffu
76a7638e	132	#define MASK24_L 0xffffffu
76a7638e	133	#define MASK24_B 0xffffffu
6b033bc4	134	#define MASK32 0xffffffffu
76a7638e	135	#define MASK32_L 0xffffffffu
76a7638e	136	#define MASK32_B 0xffffffffu
6b033bc4	137
a6f4a484	138	#ifdef HAVE_UINT64
1f175471	139	# define MASK64 MLIB_BITS_EXTENSION 0xffffffffffffffffu
a6f4a484	140	#endif
a6f4a484	141
76a7638e	142	/* --- Sizes --- */
	143
	144	#define SZ_8 1
	145	#define SZ_16 2
	146	#define SZ_16_L 2
	147	#define SZ_16_B 2
	148	#define SZ_24 3
	149	#define SZ_24_L 3
	150	#define SZ_24_B 3
	151	#define SZ_32 4
	152	#define SZ_32_L 4
	153	#define SZ_32_B 4
	154
6b033bc4	155	/* --- Type coercions --- */
	156
	157	#define U8(x) ((octet)((x) & MASK8))
	158	#define U16(x) ((uint16)((x) & MASK16))
6a0129ea	159	#define U24(x) ((uint24)((x) & MASK24))
6b033bc4	160	#define U32(x) ((uint32)((x) & MASK32))
6b033bc4	161
a6f4a484	162	#ifdef HAVE_UINT64
	163	# define U64(x) ((uint64)(x) & MASK64)
	164	# define U64_(d, x) ((d).i = U64(x).i)
	165	#else
	166	# define U64_(d, x) ((d).hi = U32((x).hi), (d).lo = U32((x).lo))
	167	#endif
	168
6b033bc4	169	/* --- Safe shifting macros --- */
6b033bc4	170
bc9f60b6	171	#define LSL8(v, s) U8(U8(v) << ((s) & 7u))
	172	#define LSR8(v, s) U8(U8(v) >> ((s) & 7u))
	173	#define LSL16(v, s) U16(U16(v) << ((s) & 15u))
	174	#define LSR16(v, s) U16(U16(v) >> ((s) & 15u))
6a0129ea	175	#define LSL24(v, s) U24(U24(v) << ((s) % 24u))
6a0129ea	176	#define LSR24(v, s) U24(U24(v) >> ((s) % 24u))
bc9f60b6	177	#define LSL32(v, s) U32(U32(v) << ((s) & 31u))
bc9f60b6	178	#define LSR32(v, s) U32(U32(v) >> ((s) & 31u))
6b033bc4	179
a6f4a484	180	#ifdef HAVE_UINT64
	181	# define LSL64(v, s) U64(U64(v) << ((s) & 63u))
	182	# define LSR64(v, s) U64(U64(v) >> ((s) & 63u))
	183	# define LSL64_(d, v, s) ((d).i = LSL64((v).i, (s)))
	184	# define LSR64_(d, v, s) ((d).i = LSR64((v).i, (s)))
	185	#else
	186	# define LSL64_(d, v, s) do { \
	187	unsigned _s = (s) & 63u; \
	188	uint32 _l = (v).lo, _h = (v).hi; \
	189	kludge64 *_d = &(d); \
	190	if (_s >= 32) { \
	191	_d->hi = LSL32(_l, _s - 32u); \
	192	_d->lo = 0; \
	193	} else if (!_s) { \
	194	_d->lo = _l; \
	195	_d->hi = _h; \
	196	} else { \
	197	_d->hi = LSL32(_h, _s) \| LSR32(_l, 32u - _s); \
	198	_d->lo = LSL32(_l, _s); \
	199	} \
	200	} while (0)
	201	# define LSR64_(d, v, s) do { \
	202	unsigned _s = (s) & 63u; \
	203	uint32 _l = (v).lo, _h = (v).hi; \
	204	kludge64 *_d = &(d); \
	205	if (_s >= 32) { \
	206	_d->lo = LSR32(_h, _s - 32u); \
	207	_d->hi = 0; \
	208	} else if (!_s) { \
	209	_d->lo = _l; \
	210	_d->hi = _h; \
	211	} else { \
	212	_d->lo = LSR32(_l, _s) \| LSL32(_h, 32u - _s); \
	213	_d->hi = LSR32(_h, _s); \
	214	} \
	215	} while (0)
	216	#endif
	217
6b033bc4	218	/* --- Rotation macros --- */
6b033bc4	219
bc9f60b6	220	#define ROL8(v, s) (LSL8((v), (s)) \| (LSR8((v), 8u - (s))))
	221	#define ROR8(v, s) (LSR8((v), (s)) \| (LSL8((v), 8u - (s))))
	222	#define ROL16(v, s) (LSL16((v), (s)) \| (LSR16((v), 16u - (s))))
	223	#define ROR16(v, s) (LSR16((v), (s)) \| (LSL16((v), 16u - (s))))
6a0129ea	224	#define ROL24(v, s) (LSL24((v), (s)) \| (LSR24((v), 24u - (s))))
6a0129ea	225	#define ROR24(v, s) (LSR24((v), (s)) \| (LSL24((v), 24u - (s))))
bc9f60b6	226	#define ROL32(v, s) (LSL32((v), (s)) \| (LSR32((v), 32u - (s))))
bc9f60b6	227	#define ROR32(v, s) (LSR32((v), (s)) \| (LSL32((v), 32u - (s))))
6b033bc4	228
a6f4a484	229	#ifdef HAVE_UINT64
	230	# define ROL64(v, s) (LSL64((v), (s)) \| (LSR64((v), 64u - (s))))
	231	# define ROR64(v, s) (LSR64((v), (s)) \| (LSL64((v), 64u - (s))))
	232	# define ROL64_(d, v, s) ((d).i = ROL64((v).i, (s)))
	233	# define ROR64_(d, v, s) ((d).i = ROR64((v).i, (s)))
	234	#else
	235	# define ROL64_(d, v, s) do { \
	236	unsigned _s = (s) & 63u; \
	237	uint32 _l = (v).lo, _h = (v).hi; \
	238	kludge64 *_d = &(d); \
5892fd39	239	if (_s > 32) { \
a6f4a484	240	_d->hi = LSL32(_l, _s - 32u) \| LSR32(_h, 64u - _s); \
	241	_d->lo = LSL32(_h, _s - 32u) \| LSR32(_l, 64u - _s); \
	242	} else if (!_s) { \
	243	_d->lo = _l; \
	244	_d->hi = _h; \
5892fd39	245	} else if (_s == 32) { \
	246	_d->lo = _h; \
	247	_d->hi = _l; \
a6f4a484	248	} else { \
	249	_d->hi = LSL32(_h, _s) \| LSR32(_l, 32u - _s); \
	250	_d->lo = LSL32(_l, _s) \| LSR32(_h, 32u - _s); \
	251	} \
	252	} while (0)
	253	# define ROR64_(d, v, s) do { \
	254	unsigned _s = (s) & 63u; \
	255	uint32 _l = (v).lo, _h = (v).hi; \
	256	kludge64 *_d = &(d); \
5892fd39	257	if (_s > 32) { \
a6f4a484	258	_d->hi = LSR32(_l, _s - 32u) \| LSL32(_h, 64u - _s); \
	259	_d->lo = LSR32(_h, _s - 32u) \| LSL32(_l, 64u - _s); \
	260	} else if (!_s) { \
	261	_d->lo = _l; \
	262	_d->hi = _h; \
5892fd39	263	} else if (_s == 32) { \
	264	_d->lo = _h; \
	265	_d->hi = _l; \
a6f4a484	266	} else { \
	267	_d->hi = LSR32(_h, _s) \| LSL32(_l, 32u - _s); \
	268	_d->lo = LSR32(_l, _s) \| LSL32(_h, 32u - _s); \
	269	} \
	270	} while (0)
	271	#endif
	272
6b033bc4	273	/* --- Storage and retrieval --- */
	274
	275	#define GETBYTE(p, o) (((octet *)(p))[o] & MASK8)
	276	#define PUTBYTE(p, o, v) (((octet *)(p))[o] = U8((v)))
	277
	278	#define LOAD8(p) (GETBYTE((p), 0))
	279	#define STORE8(p, v) (PUTBYTE((p), 0, (v)))
	280
0d4e268a	281	#define LOAD16_B(p) \
	282	(((uint16)GETBYTE((p), 0) << 8) \| \
	283	((uint16)GETBYTE((p), 1) << 0))
	284	#define LOAD16_L(p) \
	285	(((uint16)GETBYTE((p), 0) << 0) \| \
	286	((uint16)GETBYTE((p), 1) << 8))
6b033bc4	287	#define LOAD16(p) LOAD16_B((p))
6b033bc4	288
0d4e268a	289	#define STORE16_B(p, v) \
	290	(PUTBYTE((p), 0, (uint16)(v) >> 8), \
	291	PUTBYTE((p), 1, (uint16)(v) >> 0))
	292	#define STORE16_L(p, v) \
	293	(PUTBYTE((p), 0, (uint16)(v) >> 0), \
	294	PUTBYTE((p), 1, (uint16)(v) >> 8))
6b033bc4	295	#define STORE16(p, v) STORE16_B((p), (v))
6b033bc4	296
6a0129ea	297	#define LOAD24_B(p) \
	298	(((uint24)GETBYTE((p), 0) << 16) \| \
	299	((uint24)GETBYTE((p), 1) << 8) \| \
	300	((uint24)GETBYTE((p), 2) << 0))
	301	#define LOAD24_L(p) \
	302	(((uint24)GETBYTE((p), 0) << 0) \| \
	303	((uint24)GETBYTE((p), 1) << 8) \| \
	304	((uint24)GETBYTE((p), 2) << 16))
	305	#define LOAD24(p) LOAD24_B((p))
	306
	307	#define STORE24_B(p, v) \
	308	(PUTBYTE((p), 0, (uint24)(v) >> 16), \
	309	PUTBYTE((p), 1, (uint24)(v) >> 8), \
	310	PUTBYTE((p), 2, (uint24)(v) >> 0))
	311	#define STORE24_L(p, v) \
	312	(PUTBYTE((p), 0, (uint24)(v) >> 0), \
	313	PUTBYTE((p), 1, (uint24)(v) >> 8), \
	314	PUTBYTE((p), 2, (uint24)(v) >> 16))
	315	#define STORE24(p, v) STORE24_B((p), (v))
	316
6b033bc4	317	#define LOAD32_B(p) \
0d4e268a	318	(((uint32)GETBYTE((p), 0) << 24) \| \
	319	((uint32)GETBYTE((p), 1) << 16) \| \
	320	((uint32)GETBYTE((p), 2) << 8) \| \
	321	((uint32)GETBYTE((p), 3) << 0))
6b033bc4	322	#define LOAD32_L(p) \
0d4e268a	323	(((uint32)GETBYTE((p), 0) << 0) \| \
	324	((uint32)GETBYTE((p), 1) << 8) \| \
	325	((uint32)GETBYTE((p), 2) << 16) \| \
	326	((uint32)GETBYTE((p), 3) << 24))
6b033bc4	327	#define LOAD32(p) LOAD32_B((p))
	328
	329	#define STORE32_B(p, v) \
0d4e268a	330	(PUTBYTE((p), 0, (uint32)(v) >> 24), \
	331	PUTBYTE((p), 1, (uint32)(v) >> 16), \
	332	PUTBYTE((p), 2, (uint32)(v) >> 8), \
	333	PUTBYTE((p), 3, (uint32)(v) >> 0))
6b033bc4	334	#define STORE32_L(p, v) \
0d4e268a	335	(PUTBYTE((p), 0, (uint32)(v) >> 0), \
	336	PUTBYTE((p), 1, (uint32)(v) >> 8), \
	337	PUTBYTE((p), 2, (uint32)(v) >> 16), \
	338	PUTBYTE((p), 3, (uint32)(v) >> 24))
6b033bc4	339	#define STORE32(p, v) STORE32_B((p), (v))
6b033bc4	340
a6f4a484	341	#ifdef HAVE_UINT64
	342
	343	# define LOAD64_B(p) \
	344	(((uint64)GETBYTE((p), 0) << 56) \| \
	345	((uint64)GETBYTE((p), 1) << 48) \| \
	346	((uint64)GETBYTE((p), 2) << 40) \| \
	347	((uint64)GETBYTE((p), 3) << 32) \| \
	348	((uint64)GETBYTE((p), 4) << 24) \| \
	349	((uint64)GETBYTE((p), 5) << 16) \| \
	350	((uint64)GETBYTE((p), 6) << 8) \| \
	351	((uint64)GETBYTE((p), 7) << 0))
	352	# define LOAD64_L(p) \
	353	(((uint64)GETBYTE((p), 0) << 0) \| \
	354	((uint64)GETBYTE((p), 1) << 8) \| \
	355	((uint64)GETBYTE((p), 2) << 16) \| \
	356	((uint64)GETBYTE((p), 3) << 24) \| \
	357	((uint64)GETBYTE((p), 4) << 32) \| \
	358	((uint64)GETBYTE((p), 5) << 40) \| \
	359	((uint64)GETBYTE((p), 6) << 48) \| \
	360	((uint64)GETBYTE((p), 7) << 56))
	361	# define LOAD64(p) LOAD64_B((p))
	362	# define LOAD64_B_(d, p) ((d).i = LOAD64_B((p)))
	363	# define LOAD64_L_(d, p) ((d).i = LOAD64_L((p)))
	364	# define LOAD64_(d, p) LOAD64_B_((d), (p))
	365
	366	# define STORE64_B(p, v) \
	367	(PUTBYTE((p), 0, (uint64)(v) >> 56), \
	368	PUTBYTE((p), 1, (uint64)(v) >> 48), \
	369	PUTBYTE((p), 2, (uint64)(v) >> 40), \
	370	PUTBYTE((p), 3, (uint64)(v) >> 32), \
	371	PUTBYTE((p), 4, (uint64)(v) >> 24), \
	372	PUTBYTE((p), 5, (uint64)(v) >> 16), \
	373	PUTBYTE((p), 6, (uint64)(v) >> 8), \
	374	PUTBYTE((p), 7, (uint64)(v) >> 0))
	375	# define STORE64_L(p, v) \
	376	(PUTBYTE((p), 0, (uint64)(v) >> 0), \
	377	PUTBYTE((p), 1, (uint64)(v) >> 8), \
	378	PUTBYTE((p), 2, (uint64)(v) >> 16), \
	379	PUTBYTE((p), 3, (uint64)(v) >> 24), \
	380	PUTBYTE((p), 4, (uint64)(v) >> 32), \
	381	PUTBYTE((p), 5, (uint64)(v) >> 40), \
	382	PUTBYTE((p), 6, (uint64)(v) >> 48), \
	383	PUTBYTE((p), 7, (uint64)(v) >> 56))
	384	# define STORE64(p, v) STORE64_B((p), (v))
	385	# define STORE64_B_(p, v) STORE64_B((p), (v).i)
	386	# define STORE64_L_(p, v) STORE64_L((p), (v).i)
	387	# define STORE64_(p, v) STORE64_B_((p), (v))
	388
	389	#else
	390
	391	# define LOAD64_B_(d, p) \
	392	((d).hi = LOAD32_B((octet *)(p) + 0), \
	393	(d).lo = LOAD32_B((octet *)(p) + 4))
	394	# define LOAD64_L_(d, p) \
	395	((d).lo = LOAD32_L((octet *)(p) + 0), \
	396	(d).hi = LOAD32_L((octet *)(p) + 4))
	397	# define LOAD64_(d, p) LOAD64_B_((d), (p))
	398
	399	# define STORE64_B_(p, v) \
	400	(STORE32_B((octet *)(p) + 0, (v).hi), \
	401	STORE32_B((octet *)(p) + 4, (v).lo))
	402	# define STORE64_L_(p, v) \
	403	(STORE32_L((octet *)(p) + 0, (v).lo), \
	404	STORE32_L((octet *)(p) + 4, (v).hi))
405	# define STORE64_(p, v) STORE64_B_((p), (v))
406
407	#endif
408
409	/* --- Other operations on 64-bit integers --- */
410
411	#ifdef HAVE_UINT64
412	# define SET64(d, h, l) ((d).i = (U64((h)) << 32) \| U64((l)))
413	# define ASSIGN64(d, x) ((d).i = U64((x)))
414	# define HI64(x) U32((x).i >> 32)
415	# define LO64(x) U32((x).i)
3b7cea93	416	# define GET64(t, x) ((t)(x).i)
a6f4a484	417	#else
	418	# define SET64(d, h, l) ((d).hi = U32(h), (d).lo = U32(l))
	419	# define ASSIGN64(d, x) \
	420	((d).hi = ((x & ~MASK32) >> 16) >> 16, (d).lo = U32(x))
	421	# define HI64(x) U32((x).hi)
	422	# define LO64(x) U32((x).lo)
3b7cea93	423	# define GET64(t, x) (((((t)HI64(x) << 16) << 16) & ~MASK32) \| (t)LO64(x))
a6f4a484	424	#endif
	425
	426	#ifdef HAVE_UINT64
	427	# define AND64(d, x, y) ((d).i = (x).i & (y).i)
	428	# define OR64(d, x, y) ((d).i = (x).i \| (y).i)
	429	# define XOR64(d, x, y) ((d).i = (x).i ^ (y).i)
	430	# define CPL64(d, x) ((d).i = ~(x).i)
	431	# define ADD64(d, x, y) ((d).i = (x).i + (y).i)
	432	# define SUB64(d, x, y) ((d).i = (x).i - (y).i)
	433	# define CMP64(x, op, y) ((x).i op (y).i)
	434	# define ZERO64(x) ((x) == 0)
	435	#else
	436	# define AND64(d, x, y) ((d).lo = (x).lo & (y).lo, (d).hi = (x).hi & (y).hi)
	437	# define OR64(d, x, y) ((d).lo = (x).lo \| (y).lo, (d).hi = (x).hi \| (y).hi)
	438	# define XOR64(d, x, y) ((d).lo = (x).lo ^ (y).lo, (d).hi = (x).hi ^ (y).hi)
	439	# define CPL64(d, x) ((d).lo = ~(x).lo, (d).hi = ~(x).hi)
	440	# define ADD64(d, x, y) do { \
	441	uint32 _x = U32((x).lo + (y).lo); \
	442	(d).hi = (x).hi + (y).hi + (_x < (x).lo); \
	443	(d).lo = _x; \
	444	} while (0)
	445	# define SUB64(d, x, y) do { \
	446	uint32 _x = U32((x).lo - (y).lo); \
	447	(d).hi = (x).hi - (y).hi - (_x > (x).lo); \
	448	(d).lo = _x; \
	449	} while (0)
	450	# define CMP64(x, op, y) \
	451	((x).hi == (y).hi ? (x).lo op (y).lo : (x).hi op (y).hi)
	452	# define ZERO64(x) ((x).lo == 0 && (x).hi == 0)
	453	#endif
	454
91928c79	455	/* --- Storing integers in tables --- */
	456
	457	#ifdef HAVE_UINT64
	458	# define X64(x, y) { 0x##x##y }
	459	#else
	460	# define X64(x, y) { 0x##x, 0x##y }
	461	#endif
	462
6b033bc4	463	/----- That's all, folks -------------------------------------------------/
	464
	465	#ifdef __cplusplus
	466	}
	467	#endif
	468
	469	#endif