3 * Portable bit-level manipulation macros
5 * (c) 1998 Straylight/Edgeware
8 /*----- Licensing notice --------------------------------------------------*
10 * This file is part of the mLib utilities library.
12 * mLib is free software; you can redistribute it and/or modify
13 * it under the terms of the GNU Library General Public License as
14 * published by the Free Software Foundation; either version 2 of the
15 * License, or (at your option) any later version.
17 * mLib is distributed in the hope that it will be useful,
18 * but WITHOUT ANY WARRANTY; without even the implied warranty of
19 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 * GNU Library General Public License for more details.
22 * You should have received a copy of the GNU Library General Public
23 * License along with mLib; if not, write to the Free
24 * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
35 /*----- Header files ------------------------------------------------------*/
39 #if __STDC_VERSION__ >= 199900l
43 /*----- Decide on some types ----------------------------------------------*/
45 /* --- Make GNU C shut up --- */
47 #if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 91)
48 # define MLIB_BITS_EXTENSION __extension__
50 # define MLIB_BITS_EXTENSION
53 /* --- Decide on a 32-bit type --- *
55 * I want a type which is capable of expressing 32-bit numbers. Because some
56 * implementations have 64-bit @long@s (infinitely preferable to the abortion
57 * that is @long long@), using @unsigned long@ regardless is wasteful. So,
58 * if @int@ appears to be good enough, then I'll go with that.
61 #if UINT_MAX >= 0xffffffffu
62 typedef unsigned int uint32;
64 typedef unsigned long uint32;
67 /* --- Decide on a 64-bit type --- *
69 * The test is quite subtle. Think about it. Note that (at least on my
70 * machine), the 32-bit macros are *much* faster than GCC's @long long@
74 #if defined(ULONG_LONG_MAX) && !defined(ULLONG_MAX)
75 # define ULLONG_MAX ULONG_LONG_MAX
78 #if UINT_MAX >> 31 > 0xffffffff
80 typedef unsigned int uint64;
81 #elif ULONG_MAX >> 31 > 0xffffffff
83 typedef unsigned long uint64;
84 #elif defined(ULLONG_MAX)
86 MLIB_BITS_EXTENSION typedef unsigned long long uint64;
94 typedef struct { uint64 i; } kludge64;
96 typedef struct { uint32 hi, lo; } kludge64;
99 /* --- Decide on a 24-bit type --- */
101 #if UINT_MAX >= 0x00ffffffu
102 typedef unsigned int uint24;
104 typedef unsigned long uint24;
107 /* --- Decide on 16-bit and 8-bit types --- *
109 * This is more for brevity than anything else.
112 typedef unsigned short uint16;
113 typedef unsigned char octet, uint8;
115 /* --- WARNING! --- *
117 * Never lose sight of the fact that the above types may be wider than the
118 * names suggest. Some architectures have 32-bit @short@s for example.
121 /*----- Macros ------------------------------------------------------------*/
123 /* --- Useful masks --- */
126 #define MASK16 0xffffu
127 #define MASK16_L MASK16
128 #define MASK16_B MASK16
129 #define MASK24 0xffffffu
130 #define MASK24_L MASK24
131 #define MASK24_B MASK24
132 #define MASK32 0xffffffffu
133 #define MASK32_L MASK32
134 #define MASK32_B MASK32
137 # define MASK64 MLIB_BITS_EXTENSION 0xffffffffffffffffu
138 # define MASK64_L MASK64
139 # define MASK64_B MASK64
161 /* --- Type aliases --- */
164 #define TY_U16 uint16
165 #define TY_U16_L uint16
166 #define TY_U16_B uint16
167 #define TY_U24 uint24
168 #define TY_U24_L uint24
169 #define TY_U24_B uint24
170 #define TY_U32 uint32
171 #define TY_U32_L uint32
172 #define TY_U32_B uint32
175 # define TY_U64 uint64
176 # define TY_U64_L uint64
177 # define TY_U64_B uint64
180 /* --- List macros --- */
183 # define DOUINTCONV(_) \
185 _(16, 16, 16) _(16, 16_L, 16l) _(16, 16_B, 16b) \
186 _(24, 24, 24) _(24, 24_L, 24l) _(24, 24_B, 24b) \
187 _(32, 32, 32) _(32, 32_L, 32l) _(32, 32_B, 32b) \
188 _(64, 64, 64) _(64, 64_L, 64l) _(64, 64_B, 64b)
189 # define DOUINTSZ(_) _(8) _(16) _(24) _(32) _(64)
191 # define DOUINTCONV(_) \
193 _(16, 16, 16) _(16, 16_L, 16l) _(16, 16_B, 16b) \
194 _(24, 24, 24) _(24, 24_L, 24l) _(24, 24_B, 24b) \
195 _(32, 32, 32) _(32, 32_L, 32l) _(32, 32_B, 32b)
196 # define DOUINTSZ(_) _(8) _(16) _(24) _(32)
199 /* --- Type coercions --- */
201 #define U8(x) ((octet)((x) & MASK8))
202 #define U16(x) ((uint16)((x) & MASK16))
203 #define U24(x) ((uint24)((x) & MASK24))
204 #define U32(x) ((uint32)((x) & MASK32))
207 # define U64(x) ((uint64)(x) & MASK64)
208 # define U64_(d, x) ((d).i = U64(x).i)
210 # define U64_(d, x) ((d).hi = U32((x).hi), (d).lo = U32((x).lo))
213 /* --- Safe shifting macros --- */
215 #define LSL8(v, s) U8(U8(v) << ((s) & 7u))
216 #define LSR8(v, s) U8(U8(v) >> ((s) & 7u))
217 #define LSL16(v, s) U16(U16(v) << ((s) & 15u))
218 #define LSR16(v, s) U16(U16(v) >> ((s) & 15u))
219 #define LSL24(v, s) U24(U24(v) << ((s) % 24u))
220 #define LSR24(v, s) U24(U24(v) >> ((s) % 24u))
221 #define LSL32(v, s) U32(U32(v) << ((s) & 31u))
222 #define LSR32(v, s) U32(U32(v) >> ((s) & 31u))
225 # define LSL64(v, s) U64(U64(v) << ((s) & 63u))
226 # define LSR64(v, s) U64(U64(v) >> ((s) & 63u))
227 # define LSL64_(d, v, s) ((d).i = LSL64((v).i, (s)))
228 # define LSR64_(d, v, s) ((d).i = LSR64((v).i, (s)))
230 # define LSL64_(d, v, s) do { \
231 unsigned _s = (s) & 63u; \
232 uint32 _l = (v).lo, _h = (v).hi; \
233 kludge64 *_d = &(d); \
235 _d->hi = LSL32(_l, _s - 32u); \
241 _d->hi = LSL32(_h, _s) | LSR32(_l, 32u - _s); \
242 _d->lo = LSL32(_l, _s); \
245 # define LSR64_(d, v, s) do { \
246 unsigned _s = (s) & 63u; \
247 uint32 _l = (v).lo, _h = (v).hi; \
248 kludge64 *_d = &(d); \
250 _d->lo = LSR32(_h, _s - 32u); \
256 _d->lo = LSR32(_l, _s) | LSL32(_h, 32u - _s); \
257 _d->hi = LSR32(_h, _s); \
262 /* --- Rotation macros --- */
264 #define ROL8(v, s) (LSL8((v), (s)) | (LSR8((v), 8u - (s))))
265 #define ROR8(v, s) (LSR8((v), (s)) | (LSL8((v), 8u - (s))))
266 #define ROL16(v, s) (LSL16((v), (s)) | (LSR16((v), 16u - (s))))
267 #define ROR16(v, s) (LSR16((v), (s)) | (LSL16((v), 16u - (s))))
268 #define ROL24(v, s) (LSL24((v), (s)) | (LSR24((v), 24u - (s))))
269 #define ROR24(v, s) (LSR24((v), (s)) | (LSL24((v), 24u - (s))))
270 #define ROL32(v, s) (LSL32((v), (s)) | (LSR32((v), 32u - (s))))
271 #define ROR32(v, s) (LSR32((v), (s)) | (LSL32((v), 32u - (s))))
274 # define ROL64(v, s) (LSL64((v), (s)) | (LSR64((v), 64u - (s))))
275 # define ROR64(v, s) (LSR64((v), (s)) | (LSL64((v), 64u - (s))))
276 # define ROL64_(d, v, s) ((d).i = ROL64((v).i, (s)))
277 # define ROR64_(d, v, s) ((d).i = ROR64((v).i, (s)))
279 # define ROL64_(d, v, s) do { \
280 unsigned _s = (s) & 63u; \
281 uint32 _l = (v).lo, _h = (v).hi; \
282 kludge64 *_d = &(d); \
284 _d->hi = LSL32(_l, _s - 32u) | LSR32(_h, 64u - _s); \
285 _d->lo = LSL32(_h, _s - 32u) | LSR32(_l, 64u - _s); \
289 } else if (_s == 32) { \
293 _d->hi = LSL32(_h, _s) | LSR32(_l, 32u - _s); \
294 _d->lo = LSL32(_l, _s) | LSR32(_h, 32u - _s); \
297 # define ROR64_(d, v, s) do { \
298 unsigned _s = (s) & 63u; \
299 uint32 _l = (v).lo, _h = (v).hi; \
300 kludge64 *_d = &(d); \
302 _d->hi = LSR32(_l, _s - 32u) | LSL32(_h, 64u - _s); \
303 _d->lo = LSR32(_h, _s - 32u) | LSL32(_l, 64u - _s); \
307 } else if (_s == 32) { \
311 _d->hi = LSR32(_h, _s) | LSL32(_l, 32u - _s); \
312 _d->lo = LSR32(_l, _s) | LSL32(_h, 32u - _s); \
317 /* --- Endianness swapping --- */
320 # define ENDSWAP8(x) U8(x)
323 # define ENDSWAP16(x) \
324 ((((uint16)(x) >> 8)&0xff) | \
325 (((uint16)(x)&0xff) << 8))
328 # define ENDSWAP24(x) \
329 ((((uint24)(x) >> 16)&0xff) | \
330 ((uint24)(x)&0xff00) | \
331 ((uint24)((x)&0xff) << 16))
334 # define ENDSWAP32(x) \
335 (ENDSWAP16(((uint32)(x) >> 16)&0xffff) | \
336 ((uint32)ENDSWAP16((x)&0xffff) << 16))
338 #if defined(HAVE_UINT64) && !defined(ENDSWAP64)
339 # define ENDSWAP64(x) \
340 (ENDSWAP32(((uint64)(x) >> 32)&0xffffffff) | \
341 ((uint64)ENDSWAP32((x)&0xffffffff) << 32))
344 # define ENDSWAP64_(z, x) \
345 ((z).i = ENDSWAP64((x).i))
347 # define ENDSWAP64_(z, x) \
348 ((z).lo = ENDSWAP32((x).hi), \
349 (z).hi = ENDSWAP32((x).lo))
352 #define MLIB_LITTLE_ENDIAN 1234
353 #define MLIB_BIG_ENDIAN 4321
354 #if defined(__ORDER_LITTLE_ENDIAN__) && \
355 __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
356 # define MLIB_BYTE_ORDER MLIB_LITTLE_ENDIAN
357 #elif defined(__ORDER_BIG_ENDIAN__) && \
358 __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
359 # define MLIB_BYTE_ORDER MLIB_BIG_ENDIAN
362 #if MLIB_BYTE_ORDER == MLIB_LITTLE_ENDIAN
363 # define HTOL16(x) (x)
364 # define LTOH16(x) (x)
365 # define HTOB16(x) ENDSWAP16(x)
366 # define BTOH16(x) ENDSWAP16(x)
367 # define HTOL24(x) (x)
368 # define LTOH24(x) (x)
369 # define HTOB24(x) ENDSWAP24(x)
370 # define BTOH24(x) ENDSWAP24(x)
371 # define HTOL32(x) (x)
372 # define LTOH32(x) (x)
373 # define HTOB32(x) ENDSWAP32(x)
374 # define BTOH32(x) ENDSWAP32(x)
376 # define HTOL64(x) (x)
377 # define LTOH64(x) (x)
378 # define HTOB64(x) ENDSWAP64(x)
379 # define BTOH64(x) ENDSWAP64(x)
381 # define HTOL64_(z, x) ASSIGN64(z, x)
382 # define LTOH64_(z, x) ASSIGN64(z, x)
383 # define HTOB64_(z, x) ENDSWAP64_(z, x)
384 # define BTOH64_(z, x) ENDSWAP64_(z, x)
385 #elif MLIB_BYTE_ORDER == MLIB_BIG_ENDIAN
386 # define HTOL16(x) ENDSWAP16(x)
387 # define LTOH16(x) ENDSWAP16(x)
388 # define HTOB16(x) (x)
389 # define BTOH16(x) (x)
390 # define HTOL24(x) ENDSWAP24(x)
391 # define LTOH24(x) ENDSWAP24(x)
392 # define HTOB24(x) (x)
393 # define BTOH24(x) (x)
394 # define HTOL32(x) ENDSWAP32(x)
395 # define LTOH32(x) ENDSWAP32(x)
396 # define HTOB32(x) (x)
397 # define BTOH32(x) (x)
399 # define HTOL64(x) ENDSWAP64(x)
400 # define LTOH64(x) ENDSWAP64(x)
401 # define HTOB64(x) (x)
402 # define BTOH64(x) (x)
403 # define HTOL64_(z, x) ENDSWAP64_(z, x)
404 # define LTOH64_(z, x) ENDSWAP64_(z, x)
405 # define HTOB64_(z, x) ((z).i = (x).i)
406 # define BTOH64_(z, x) ((z).i = (x).i)
408 # define HTOL64_(z, x) ENDSWAP64_(z, x)
409 # define LTOH64_(z, x) ENDSWAP64_(z, x)
410 # define HTOB64_(z, x) ASSIGN64(z, x)
411 # define BTOH64_(z, x) ASSIGN64(z, x)
414 /* --- Storage and retrieval --- */
416 #define GETBYTE(p, o) (((octet *)(p))[o] & MASK8)
417 #define PUTBYTE(p, o, v) (((octet *)(p))[o] = U8((v)))
419 #define LOAD8(p) (GETBYTE((p), 0))
420 #define STORE8(p, v) (PUTBYTE((p), 0, (v)))
424 (((uint16)GETBYTE((p), 0) << 8) | \
425 ((uint16)GETBYTE((p), 1) << 0))
428 # define LOAD16_L(p) \
429 (((uint16)GETBYTE((p), 0) << 0) | \
430 ((uint16)GETBYTE((p), 1) << 8))
432 #define LOAD16(p) LOAD16_B((p))
435 # define STORE16_B(p, v) \
436 (PUTBYTE((p), 0, (uint16)(v) >> 8), \
437 PUTBYTE((p), 1, (uint16)(v) >> 0))
440 # define STORE16_L(p, v) \
441 (PUTBYTE((p), 0, (uint16)(v) >> 0), \
442 PUTBYTE((p), 1, (uint16)(v) >> 8))
444 #define STORE16(p, v) STORE16_B((p), (v))
447 # define LOAD24_B(p) \
448 (((uint24)GETBYTE((p), 0) << 16) | \
449 ((uint24)LOAD16_B((octet *)(p) + 1) << 0))
452 # define LOAD24_L(p) \
453 (((uint24)LOAD16_L((octet *)(p) + 0) << 0) | \
454 ((uint24)GETBYTE((p), 2) << 16))
456 #define LOAD24(p) LOAD24_B((p))
459 # define STORE24_B(p, v) \
460 (PUTBYTE((p), 0, (uint24)(v) >> 16), \
461 STORE16_B((octet *)(p) + 1, (uint24)(v) >> 0))
464 # define STORE24_L(p, v) \
465 (STORE16_L((octet *)(p) + 0, (uint24)(v) >> 0), \
466 PUTBYTE((p), 2, (uint24)(v) >> 16))
468 #define STORE24(p, v) STORE24_B((p), (v))
471 # define LOAD32_B(p) \
472 (((uint32)LOAD16_B((octet *)(p) + 0) << 16) | \
473 ((uint32)LOAD16_B((octet *)(p) + 2) << 0))
476 # define LOAD32_L(p) \
477 (((uint32)LOAD16_L((octet *)(p) + 0) << 0) | \
478 ((uint32)LOAD16_L((octet *)(p) + 2) << 16))
480 #define LOAD32(p) LOAD32_B((p))
483 # define STORE32_B(p, v) \
484 (STORE16_B((octet *)(p) + 0, (uint32)(v) >> 16), \
485 STORE16_B((octet *)(p) + 2, (uint32)(v) >> 0))
488 # define STORE32_L(p, v) \
489 (STORE16_L((octet *)(p) + 0, (uint32)(v) >> 0), \
490 STORE16_L((octet *)(p) + 2, (uint32)(v) >> 16))
492 #define STORE32(p, v) STORE32_B((p), (v))
497 # define LOAD64_B(p) \
498 (((uint64)LOAD32_B((octet *)(p) + 0) << 32) | \
499 ((uint64)LOAD32_B((octet *)(p) + 4) << 0))
502 # define LOAD64_L(p) \
503 (((uint64)LOAD32_L((octet *)(p) + 0) << 0) | \
504 ((uint64)LOAD32_L((octet *)(p) + 4) << 32))
506 # define LOAD64(p) LOAD64_B((p))
507 # define LOAD64_B_(d, p) ((d).i = LOAD64_B((p)))
508 # define LOAD64_L_(d, p) ((d).i = LOAD64_L((p)))
509 # define LOAD64_(d, p) LOAD64_B_((d), (p))
512 # define STORE64_B(p, v) \
513 (STORE32_B((octet *)(p) + 0, (uint64)(v) >> 32), \
514 STORE32_B((octet *)(p) + 4, (uint64)(v) >> 0))
517 # define STORE64_L(p, v) \
518 (STORE32_L((octet *)(p) + 0, (uint64)(v) >> 0), \
519 STORE32_L((octet *)(p) + 4, (uint64)(v) >> 32))
521 # define STORE64(p, v) STORE64_B((p), (v))
522 # define STORE64_B_(p, v) STORE64_B((p), (v).i)
523 # define STORE64_L_(p, v) STORE64_L((p), (v).i)
524 # define STORE64_(p, v) STORE64_B_((p), (v))
528 # define LOAD64_B_(d, p) \
529 ((d).hi = LOAD32_B((octet *)(p) + 0), \
530 (d).lo = LOAD32_B((octet *)(p) + 4))
531 # define LOAD64_L_(d, p) \
532 ((d).lo = LOAD32_L((octet *)(p) + 0), \
533 (d).hi = LOAD32_L((octet *)(p) + 4))
534 # define LOAD64_(d, p) LOAD64_B_((d), (p))
536 # define STORE64_B_(p, v) \
537 (STORE32_B((octet *)(p) + 0, (v).hi), \
538 STORE32_B((octet *)(p) + 4, (v).lo))
539 # define STORE64_L_(p, v) \
540 (STORE32_L((octet *)(p) + 0, (v).lo), \
541 STORE32_L((octet *)(p) + 4, (v).hi))
542 # define STORE64_(p, v) STORE64_B_((p), (v))
546 /* --- Other operations on 64-bit integers --- */
549 # define SET64(d, h, l) ((d).i = (U64((h)) << 32) | U64((l)))
550 # define ASSIGN64(d, x) ((d).i = U64((x)))
551 # define HI64(x) U32((x).i >> 32)
552 # define LO64(x) U32((x).i)
553 # define GET64(t, x) ((t)(x).i)
555 # define SET64(d, h, l) ((d).hi = U32(h), (d).lo = U32(l))
556 # define ASSIGN64(d, x) \
557 ((d).hi = ((x & ~MASK32) >> 16) >> 16, (d).lo = U32(x))
558 # define HI64(x) U32((x).hi)
559 # define LO64(x) U32((x).lo)
560 # define GET64(t, x) (((((t)HI64(x) << 16) << 16) & ~MASK32) | (t)LO64(x))
564 # define AND64(d, x, y) ((d).i = (x).i & (y).i)
565 # define OR64(d, x, y) ((d).i = (x).i | (y).i)
566 # define XOR64(d, x, y) ((d).i = (x).i ^ (y).i)
567 # define CPL64(d, x) ((d).i = ~(x).i)
568 # define ADD64(d, x, y) ((d).i = (x).i + (y).i)
569 # define SUB64(d, x, y) ((d).i = (x).i - (y).i)
570 # define CMP64(x, op, y) ((x).i op (y).i)
571 # define ZERO64(x) ((x) == 0)
573 # define AND64(d, x, y) ((d).lo = (x).lo & (y).lo, (d).hi = (x).hi & (y).hi)
574 # define OR64(d, x, y) ((d).lo = (x).lo | (y).lo, (d).hi = (x).hi | (y).hi)
575 # define XOR64(d, x, y) ((d).lo = (x).lo ^ (y).lo, (d).hi = (x).hi ^ (y).hi)
576 # define CPL64(d, x) ((d).lo = ~(x).lo, (d).hi = ~(x).hi)
577 # define ADD64(d, x, y) do { \
578 uint32 _x = U32((x).lo + (y).lo); \
579 (d).hi = (x).hi + (y).hi + (_x < (x).lo); \
582 # define SUB64(d, x, y) do { \
583 uint32 _x = U32((x).lo - (y).lo); \
584 (d).hi = (x).hi - (y).hi - (_x > (x).lo); \
587 # define CMP64(x, op, y) \
588 ((x).hi == (y).hi ? (x).lo op (y).lo : (x).hi op (y).hi)
589 # define ZERO64(x) ((x).lo == 0 && (x).hi == 0)
592 /* --- Storing integers in tables --- */
595 # define X64(x, y) { 0x##x##y }
597 # define X64(x, y) { 0x##x, 0x##y }
600 /*----- That's all, folks -------------------------------------------------*/