5 * Portable bit-level manipulation macros
7 * (c) 1998 Straylight/Edgeware
10 /*----- Licensing notice --------------------------------------------------*
12 * This file is part of the mLib utilities library.
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.
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.
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,
37 /*----- Header files ------------------------------------------------------*/
41 #if __STDC_VERSION__ >= 199900l
45 /*----- Decide on some types ----------------------------------------------*/
47 /* --- Make GNU C shut up --- */
49 #if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 91)
50 # define MLIB_BITS_EXTENSION __extension__
52 # define MLIB_BITS_EXTENSION
55 /* --- Decide on a 32-bit type --- *
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.
63 #if UINT_MAX >= 0xffffffffu
64 typedef unsigned int uint32;
66 typedef unsigned long uint32;
69 /* --- Decide on a 64-bit type --- *
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@
76 #if defined(ULONG_LONG_MAX) && !defined(ULLONG_MAX)
77 # define ULLONG_MAX ULONG_LONG_MAX
80 #if UINT_MAX >> 31 > 0xffffffff
82 typedef unsigned int uint64;
83 #elif ULONG_MAX >> 31 > 0xffffffff
85 typedef unsigned long uint64;
86 #elif defined(ULLONG_MAX)
88 MLIB_BITS_EXTENSION typedef unsigned long long uint64;
96 typedef struct { uint64 i; } kludge64;
98 typedef struct { uint32 hi, lo; } kludge64;
101 /* --- Decide on a 24-bit type --- */
103 #if UINT_MAX >= 0x00ffffffu
104 typedef unsigned int uint24;
106 typedef unsigned long uint24;
109 /* --- Decide on 16-bit and 8-bit types --- *
111 * This is more for brevity than anything else.
114 typedef unsigned short uint16;
115 typedef unsigned char octet;
117 /* --- WARNING! --- *
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.
123 /*----- Macros ------------------------------------------------------------*/
125 /* --- Useful masks --- */
128 #define MASK16 0xffffu
129 #define MASK16_L 0xffffu
130 #define MASK16_B 0xffffu
131 #define MASK24 0xffffffu
132 #define MASK24_L 0xffffffu
133 #define MASK24_B 0xffffffu
134 #define MASK32 0xffffffffu
135 #define MASK32_L 0xffffffffu
136 #define MASK32_B 0xffffffffu
139 # define MASK64 MLIB_BITS_EXTENSION 0xffffffffffffffffu
155 /* --- Type coercions --- */
157 #define U8(x) ((octet)((x) & MASK8))
158 #define U16(x) ((uint16)((x) & MASK16))
159 #define U24(x) ((uint24)((x) & MASK24))
160 #define U32(x) ((uint32)((x) & MASK32))
163 # define U64(x) ((uint64)(x) & MASK64)
164 # define U64_(d, x) ((d).i = U64(x).i)
166 # define U64_(d, x) ((d).hi = U32((x).hi), (d).lo = U32((x).lo))
169 /* --- Safe shifting macros --- */
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))
175 #define LSL24(v, s) U24(U24(v) << ((s) % 24u))
176 #define LSR24(v, s) U24(U24(v) >> ((s) % 24u))
177 #define LSL32(v, s) U32(U32(v) << ((s) & 31u))
178 #define LSR32(v, s) U32(U32(v) >> ((s) & 31u))
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)))
186 # define LSL64_(d, v, s) do { \
187 unsigned _s = (s) & 63u; \
188 uint32 _l = (v).lo, _h = (v).hi; \
189 kludge64 *_d = &(d); \
191 _d->hi = LSL32(_l, _s - 32u); \
197 _d->hi = LSL32(_h, _s) | LSR32(_l, 32u - _s); \
198 _d->lo = LSL32(_l, _s); \
201 # define LSR64_(d, v, s) do { \
202 unsigned _s = (s) & 63u; \
203 uint32 _l = (v).lo, _h = (v).hi; \
204 kludge64 *_d = &(d); \
206 _d->lo = LSR32(_h, _s - 32u); \
212 _d->lo = LSR32(_l, _s) | LSL32(_h, 32u - _s); \
213 _d->hi = LSR32(_h, _s); \
218 /* --- Rotation macros --- */
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))))
224 #define ROL24(v, s) (LSL24((v), (s)) | (LSR24((v), 24u - (s))))
225 #define ROR24(v, s) (LSR24((v), (s)) | (LSL24((v), 24u - (s))))
226 #define ROL32(v, s) (LSL32((v), (s)) | (LSR32((v), 32u - (s))))
227 #define ROR32(v, s) (LSR32((v), (s)) | (LSL32((v), 32u - (s))))
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)))
235 # define ROL64_(d, v, s) do { \
236 unsigned _s = (s) & 63u; \
237 uint32 _l = (v).lo, _h = (v).hi; \
238 kludge64 *_d = &(d); \
240 _d->hi = LSL32(_l, _s - 32u) | LSR32(_h, 64u - _s); \
241 _d->lo = LSL32(_h, _s - 32u) | LSR32(_l, 64u - _s); \
245 } else if (_s == 32) { \
249 _d->hi = LSL32(_h, _s) | LSR32(_l, 32u - _s); \
250 _d->lo = LSL32(_l, _s) | LSR32(_h, 32u - _s); \
253 # define ROR64_(d, v, s) do { \
254 unsigned _s = (s) & 63u; \
255 uint32 _l = (v).lo, _h = (v).hi; \
256 kludge64 *_d = &(d); \
258 _d->hi = LSR32(_l, _s - 32u) | LSL32(_h, 64u - _s); \
259 _d->lo = LSR32(_h, _s - 32u) | LSL32(_l, 64u - _s); \
263 } else if (_s == 32) { \
267 _d->hi = LSR32(_h, _s) | LSL32(_l, 32u - _s); \
268 _d->lo = LSR32(_l, _s) | LSL32(_h, 32u - _s); \
273 /* --- Storage and retrieval --- */
275 #define GETBYTE(p, o) (((octet *)(p))[o] & MASK8)
276 #define PUTBYTE(p, o, v) (((octet *)(p))[o] = U8((v)))
278 #define LOAD8(p) (GETBYTE((p), 0))
279 #define STORE8(p, v) (PUTBYTE((p), 0, (v)))
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))
287 #define LOAD16(p) LOAD16_B((p))
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))
295 #define STORE16(p, v) STORE16_B((p), (v))
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))
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))
317 #define LOAD32_B(p) \
318 (((uint32)GETBYTE((p), 0) << 24) | \
319 ((uint32)GETBYTE((p), 1) << 16) | \
320 ((uint32)GETBYTE((p), 2) << 8) | \
321 ((uint32)GETBYTE((p), 3) << 0))
322 #define LOAD32_L(p) \
323 (((uint32)GETBYTE((p), 0) << 0) | \
324 ((uint32)GETBYTE((p), 1) << 8) | \
325 ((uint32)GETBYTE((p), 2) << 16) | \
326 ((uint32)GETBYTE((p), 3) << 24))
327 #define LOAD32(p) LOAD32_B((p))
329 #define STORE32_B(p, v) \
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))
334 #define STORE32_L(p, v) \
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))
339 #define STORE32(p, v) STORE32_B((p), (v))
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))
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))
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))
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))
409 /* --- Other operations on 64-bit integers --- */
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)
416 # define GET64(t, x) ((t)(x).i)
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)
423 # define GET64(t, x) (((((t)HI64(x) << 16) << 16) & ~MASK32) | (t)LO64(x))
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)
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); \
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); \
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)
455 /* --- Storing integers in tables --- */
458 # define X64(x, y) { 0x##x##y }
460 # define X64(x, y) { 0x##x, 0x##y }
463 /*----- That's all, folks -------------------------------------------------*/