ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libecb/ecb.h
Revision: 1.72
Committed: Wed Jan 18 12:13:00 2012 UTC (12 years, 4 months ago) by root
Content type: text/plain
Branch: MAIN
Changes since 1.71: +11 -2 lines
Log Message:
sun studio and sparc support

File Contents

# Content
1 /*
2 * libecb - http://software.schmorp.de/pkg/libecb
3 *
4 * Copyright (©) 2009-2012 Marc Alexander Lehmann <libecb@schmorp.de>
5 * Copyright (©) 2011 Emanuele Giaquinta
6 * All rights reserved.
7 *
8 * Redistribution and use in source and binary forms, with or without modifica-
9 * tion, are permitted provided that the following conditions are met:
10 *
11 * 1. Redistributions of source code must retain the above copyright notice,
12 * this list of conditions and the following disclaimer.
13 *
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 *
18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
19 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
20 * CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
21 * EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE-
22 * CIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
23 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
24 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
25 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH-
26 * ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
27 * OF THE POSSIBILITY OF SUCH DAMAGE.
28 */
29
30 #ifndef ECB_H
31 #define ECB_H
32
33 #ifdef _WIN32
34 typedef signed char int8_t;
35 typedef unsigned char uint8_t;
36 typedef signed short int16_t;
37 typedef unsigned short uint16_t;
38 typedef signed int int32_t;
39 typedef unsigned int uint32_t;
40 #if __GNUC__
41 typedef signed long long int64_t;
42 typedef unsigned long long uint64_t;
43 #else /* _MSC_VER || __BORLANDC__ */
44 typedef signed __int64 int64_t;
45 typedef unsigned __int64 uint64_t;
46 #endif
47 #else
48 #include <inttypes.h>
49 #endif
50
51 /* many compilers define _GNUC_ to some versions but then only implement
52 * what their idiot authors think are the "more important" extensions,
53 * causing enormous grief in return for some better fake benchmark numbers.
54 * or so.
55 * we try to detect these and simply assume they are not gcc - if they have
56 * an issue with that they should have done it right in the first place.
57 */
58 #ifndef ECB_GCC_VERSION
59 #if !defined(__GNUC_MINOR__) || defined(__INTEL_COMPILER) || defined(__SUNPRO_C) || defined(__SUNPRO_CC) || defined(__llvm__) || defined(__clang__)
60 #define ECB_GCC_VERSION(major,minor) 0
61 #else
62 #define ECB_GCC_VERSION(major,minor) (__GNUC__ > (major) || (__GNUC__ == (major) && __GNUC_MINOR__ >= (minor)))
63 #endif
64 #endif
65
66 /*****************************************************************************/
67
68 /* ECB_NO_THREADS - ecb is not used by multiple threads, ever */
69 /* ECB_NO_SMP - ecb might be used in multiple threads, but only on a single cpu */
70
71 #if ECB_NO_THREADS || ECB_NO_SMP
72 #define ECB_MEMORY_FENCE do { } while (0)
73 #endif
74
75 #ifndef ECB_MEMORY_FENCE
76 #if ECB_GCC_VERSION(2,5) || defined(__INTEL_COMPILER) || defined(__clang__) || __SUNPRO_C >= 0x5110 || __SUNPRO_CC >= 0x5110
77 #if __i386__
78 #define ECB_MEMORY_FENCE __asm__ __volatile__ ("lock; orb $0, -1(%%esp)" : : : "memory")
79 #define ECB_MEMORY_FENCE_ACQUIRE ECB_MEMORY_FENCE /* non-lock xchg might be enough */
80 #define ECB_MEMORY_FENCE_RELEASE do { } while (0) /* unlikely to change in future cpus */
81 #elif __amd64
82 #define ECB_MEMORY_FENCE __asm__ __volatile__ ("mfence" : : : "memory")
83 #define ECB_MEMORY_FENCE_ACQUIRE __asm__ __volatile__ ("lfence" : : : "memory")
84 #define ECB_MEMORY_FENCE_RELEASE __asm__ __volatile__ ("sfence") /* play safe - not needed in any current cpu */
85 #elif __powerpc__ || __ppc__ || __powerpc64__ || __ppc64__
86 #define ECB_MEMORY_FENCE __asm__ __volatile__ ("sync" : : : "memory")
87 #elif defined(__ARM_ARCH_6__ ) || defined(__ARM_ARCH_6J__ ) \
88 || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6ZK__)
89 #define ECB_MEMORY_FENCE __asm__ __volatile__ ("mcr p15,0,%0,c7,c10,5" : : "r" (0) : "memory")
90 #elif defined(__ARM_ARCH_7__ ) || defined(__ARM_ARCH_7A__ ) \
91 || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7R__ )
92 #define ECB_MEMORY_FENCE __asm__ __volatile__ ("dmb" : : : "memory")
93 #elif defined(__sparc)
94 #define ECB_MEMORY_FENCE __asm__ __volatile__ ("membar #LoadStore | #StoreLoad | #LoadLoad | #StoreStore" : : : "memory")
95 #define ECB_MEMORY_FENCE_ACQUIRE __asm__ __volatile__ ("membar #LoadLoad" : : : "memory")
96 #define ECB_MEMORY_FENCE_RELEASE __asm__ __volatile__ ("membar #StoreStore")
97 #endif
98 #endif
99 #endif
100
101 #ifndef ECB_MEMORY_FENCE
102 #if ECB_GCC_VERSION(4,4) || defined(__INTEL_COMPILER) || defined(__clang__)
103 #define ECB_MEMORY_FENCE __sync_synchronize ()
104 /*#define ECB_MEMORY_FENCE_ACQUIRE ({ char dummy = 0; __sync_lock_test_and_set (&dummy, 1); }) */
105 /*#define ECB_MEMORY_FENCE_RELEASE ({ char dummy = 1; __sync_lock_release (&dummy ); }) */
106 #elif _MSC_VER >= 1400 /* VC++ 2005 */
107 #pragma intrinsic(_ReadBarrier,_WriteBarrier,_ReadWriteBarrier)
108 #define ECB_MEMORY_FENCE _ReadWriteBarrier ()
109 #define ECB_MEMORY_FENCE_ACQUIRE _ReadWriteBarrier () /* according to msdn, _ReadBarrier is not a load fence */
110 #define ECB_MEMORY_FENCE_RELEASE _WriteBarrier ()
111 #elif defined(_WIN32)
112 #include <WinNT.h>
113 #define ECB_MEMORY_FENCE MemoryBarrier () /* actually just xchg on x86... scary */
114 #elif __SUNPRO_C >= 0x5110 || __SUNPRO_CC >= 0x5110
115 #include <mbarrier.h>
116 #define ECB_MEMORY_FENCE __machine_rw_barrier ()
117 #define ECB_MEMORY_FENCE_ACQUIRE __machine_r_barrier ()
118 #define ECB_MEMORY_FENCE_RELEASE __machine_w_barrier ()
119 #endif
120 #endif
121
122 #ifndef ECB_MEMORY_FENCE
123 #if !ECB_AVOID_PTHREADS
124 /*
125 * if you get undefined symbol references to pthread_mutex_lock,
126 * or failure to find pthread.h, then you should implement
127 * the ECB_MEMORY_FENCE operations for your cpu/compiler
128 * OR provide pthread.h and link against the posix thread library
129 * of your system.
130 */
131 #include <pthread.h>
132 #define ECB_NEEDS_PTHREADS 1
133 #define ECB_MEMORY_FENCE_NEEDS_PTHREADS 1
134
135 static pthread_mutex_t ecb_mf_lock = PTHREAD_MUTEX_INITIALIZER;
136 #define ECB_MEMORY_FENCE do { pthread_mutex_lock (&ecb_mf_lock); pthread_mutex_unlock (&ecb_mf_lock); } while (0)
137 #endif
138 #endif
139
140 #if !defined(ECB_MEMORY_FENCE_ACQUIRE) && defined(ECB_MEMORY_FENCE)
141 #define ECB_MEMORY_FENCE_ACQUIRE ECB_MEMORY_FENCE
142 #endif
143
144 #if !defined(ECB_MEMORY_FENCE_RELEASE) && defined(ECB_MEMORY_FENCE)
145 #define ECB_MEMORY_FENCE_RELEASE ECB_MEMORY_FENCE
146 #endif
147
148 /*****************************************************************************/
149
150 #define ECB_C99 (__STDC_VERSION__ >= 199901L)
151
152 #if __cplusplus
153 #define ecb_inline static inline
154 #elif ECB_GCC_VERSION(2,5)
155 #define ecb_inline static __inline__
156 #elif ECB_C99
157 #define ecb_inline static inline
158 #else
159 #define ecb_inline static
160 #endif
161
162 #if ECB_GCC_VERSION(3,3)
163 #define ecb_restrict __restrict__
164 #elif ECB_C99
165 #define ecb_restrict restrict
166 #else
167 #define ecb_restrict
168 #endif
169
170 typedef int ecb_bool;
171
172 #define ECB_CONCAT_(a, b) a ## b
173 #define ECB_CONCAT(a, b) ECB_CONCAT_(a, b)
174 #define ECB_STRINGIFY_(a) # a
175 #define ECB_STRINGIFY(a) ECB_STRINGIFY_(a)
176
177 #define ecb_function_ ecb_inline
178
179 #if ECB_GCC_VERSION(3,1)
180 #define ecb_attribute(attrlist) __attribute__(attrlist)
181 #define ecb_is_constant(expr) __builtin_constant_p (expr)
182 #define ecb_expect(expr,value) __builtin_expect ((expr),(value))
183 #define ecb_prefetch(addr,rw,locality) __builtin_prefetch (addr, rw, locality)
184 #else
185 #define ecb_attribute(attrlist)
186 #define ecb_is_constant(expr) 0
187 #define ecb_expect(expr,value) (expr)
188 #define ecb_prefetch(addr,rw,locality)
189 #endif
190
191 /* no emulation for ecb_decltype */
192 #if ECB_GCC_VERSION(4,5)
193 #define ecb_decltype(x) __decltype(x)
194 #elif ECB_GCC_VERSION(3,0)
195 #define ecb_decltype(x) __typeof(x)
196 #endif
197
198 #define ecb_noinline ecb_attribute ((__noinline__))
199 #define ecb_noreturn ecb_attribute ((__noreturn__))
200 #define ecb_unused ecb_attribute ((__unused__))
201 #define ecb_const ecb_attribute ((__const__))
202 #define ecb_pure ecb_attribute ((__pure__))
203
204 #if ECB_GCC_VERSION(4,3)
205 #define ecb_artificial ecb_attribute ((__artificial__))
206 #define ecb_hot ecb_attribute ((__hot__))
207 #define ecb_cold ecb_attribute ((__cold__))
208 #else
209 #define ecb_artificial
210 #define ecb_hot
211 #define ecb_cold
212 #endif
213
214 /* put around conditional expressions if you are very sure that the */
215 /* expression is mostly true or mostly false. note that these return */
216 /* booleans, not the expression. */
217 #define ecb_expect_false(expr) ecb_expect (!!(expr), 0)
218 #define ecb_expect_true(expr) ecb_expect (!!(expr), 1)
219 /* for compatibility to the rest of the world */
220 #define ecb_likely(expr) ecb_expect_true (expr)
221 #define ecb_unlikely(expr) ecb_expect_false (expr)
222
223 /* count trailing zero bits and count # of one bits */
224 #if ECB_GCC_VERSION(3,4)
225 /* we assume int == 32 bit, long == 32 or 64 bit and long long == 64 bit */
226 #define ecb_ld32(x) (__builtin_clz (x) ^ 31)
227 #define ecb_ld64(x) (__builtin_clzll (x) ^ 63)
228 #define ecb_ctz32(x) __builtin_ctz (x)
229 #define ecb_ctz64(x) __builtin_ctzll (x)
230 #define ecb_popcount32(x) __builtin_popcount (x)
231 /* no popcountll */
232 #else
233 ecb_function_ int ecb_ctz32 (uint32_t x) ecb_const;
234 ecb_function_ int
235 ecb_ctz32 (uint32_t x)
236 {
237 int r = 0;
238
239 x &= ~x + 1; /* this isolates the lowest bit */
240
241 #if ECB_branchless_on_i386
242 r += !!(x & 0xaaaaaaaa) << 0;
243 r += !!(x & 0xcccccccc) << 1;
244 r += !!(x & 0xf0f0f0f0) << 2;
245 r += !!(x & 0xff00ff00) << 3;
246 r += !!(x & 0xffff0000) << 4;
247 #else
248 if (x & 0xaaaaaaaa) r += 1;
249 if (x & 0xcccccccc) r += 2;
250 if (x & 0xf0f0f0f0) r += 4;
251 if (x & 0xff00ff00) r += 8;
252 if (x & 0xffff0000) r += 16;
253 #endif
254
255 return r;
256 }
257
258 ecb_function_ int ecb_ctz64 (uint64_t x) ecb_const;
259 ecb_function_ int
260 ecb_ctz64 (uint64_t x)
261 {
262 int shift = x & 0xffffffffU ? 0 : 32;
263 return ecb_ctz32 (x >> shift) + shift;
264 }
265
266 ecb_function_ int ecb_popcount32 (uint32_t x) ecb_const;
267 ecb_function_ int
268 ecb_popcount32 (uint32_t x)
269 {
270 x -= (x >> 1) & 0x55555555;
271 x = ((x >> 2) & 0x33333333) + (x & 0x33333333);
272 x = ((x >> 4) + x) & 0x0f0f0f0f;
273 x *= 0x01010101;
274
275 return x >> 24;
276 }
277
278 ecb_function_ int ecb_ld32 (uint32_t x) ecb_const;
279 ecb_function_ int ecb_ld32 (uint32_t x)
280 {
281 int r = 0;
282
283 if (x >> 16) { x >>= 16; r += 16; }
284 if (x >> 8) { x >>= 8; r += 8; }
285 if (x >> 4) { x >>= 4; r += 4; }
286 if (x >> 2) { x >>= 2; r += 2; }
287 if (x >> 1) { r += 1; }
288
289 return r;
290 }
291
292 ecb_function_ int ecb_ld64 (uint64_t x) ecb_const;
293 ecb_function_ int ecb_ld64 (uint64_t x)
294 {
295 int r = 0;
296
297 if (x >> 32) { x >>= 32; r += 32; }
298
299 return r + ecb_ld32 (x);
300 }
301 #endif
302
303 ecb_function_ uint8_t ecb_bitrev8 (uint8_t x) ecb_const;
304 ecb_function_ uint8_t ecb_bitrev8 (uint8_t x)
305 {
306 return ( (x * 0x0802U & 0x22110U)
307 | (x * 0x8020U & 0x88440U)) * 0x10101U >> 16;
308 }
309
310 ecb_function_ uint16_t ecb_bitrev16 (uint16_t x) ecb_const;
311 ecb_function_ uint16_t ecb_bitrev16 (uint16_t x)
312 {
313 x = ((x >> 1) & 0x5555) | ((x & 0x5555) << 1);
314 x = ((x >> 2) & 0x3333) | ((x & 0x3333) << 2);
315 x = ((x >> 4) & 0x0f0f) | ((x & 0x0f0f) << 4);
316 x = ( x >> 8 ) | ( x << 8);
317
318 return x;
319 }
320
321 ecb_function_ uint32_t ecb_bitrev32 (uint32_t x) ecb_const;
322 ecb_function_ uint32_t ecb_bitrev32 (uint32_t x)
323 {
324 x = ((x >> 1) & 0x55555555) | ((x & 0x55555555) << 1);
325 x = ((x >> 2) & 0x33333333) | ((x & 0x33333333) << 2);
326 x = ((x >> 4) & 0x0f0f0f0f) | ((x & 0x0f0f0f0f) << 4);
327 x = ((x >> 8) & 0x00ff00ff) | ((x & 0x00ff00ff) << 8);
328 x = ( x >> 16 ) | ( x << 16);
329
330 return x;
331 }
332
333 /* popcount64 is only available on 64 bit cpus as gcc builtin */
334 /* so for this version we are lazy */
335 ecb_function_ int ecb_popcount64 (uint64_t x) ecb_const;
336 ecb_function_ int
337 ecb_popcount64 (uint64_t x)
338 {
339 return ecb_popcount32 (x) + ecb_popcount32 (x >> 32);
340 }
341
342 ecb_inline uint8_t ecb_rotl8 (uint8_t x, unsigned int count) ecb_const;
343 ecb_inline uint8_t ecb_rotr8 (uint8_t x, unsigned int count) ecb_const;
344 ecb_inline uint16_t ecb_rotl16 (uint16_t x, unsigned int count) ecb_const;
345 ecb_inline uint16_t ecb_rotr16 (uint16_t x, unsigned int count) ecb_const;
346 ecb_inline uint32_t ecb_rotl32 (uint32_t x, unsigned int count) ecb_const;
347 ecb_inline uint32_t ecb_rotr32 (uint32_t x, unsigned int count) ecb_const;
348 ecb_inline uint64_t ecb_rotl64 (uint64_t x, unsigned int count) ecb_const;
349 ecb_inline uint64_t ecb_rotr64 (uint64_t x, unsigned int count) ecb_const;
350
351 ecb_inline uint8_t ecb_rotl8 (uint8_t x, unsigned int count) { return (x >> ( 8 - count)) | (x << count); }
352 ecb_inline uint8_t ecb_rotr8 (uint8_t x, unsigned int count) { return (x << ( 8 - count)) | (x >> count); }
353 ecb_inline uint16_t ecb_rotl16 (uint16_t x, unsigned int count) { return (x >> (16 - count)) | (x << count); }
354 ecb_inline uint16_t ecb_rotr16 (uint16_t x, unsigned int count) { return (x << (16 - count)) | (x >> count); }
355 ecb_inline uint32_t ecb_rotl32 (uint32_t x, unsigned int count) { return (x >> (32 - count)) | (x << count); }
356 ecb_inline uint32_t ecb_rotr32 (uint32_t x, unsigned int count) { return (x << (32 - count)) | (x >> count); }
357 ecb_inline uint64_t ecb_rotl64 (uint64_t x, unsigned int count) { return (x >> (64 - count)) | (x << count); }
358 ecb_inline uint64_t ecb_rotr64 (uint64_t x, unsigned int count) { return (x << (64 - count)) | (x >> count); }
359
360 #if ECB_GCC_VERSION(4,3)
361 #define ecb_bswap16(x) (__builtin_bswap32 (x) >> 16)
362 #define ecb_bswap32(x) __builtin_bswap32 (x)
363 #define ecb_bswap64(x) __builtin_bswap64 (x)
364 #else
365 ecb_function_ uint16_t ecb_bswap16 (uint16_t x) ecb_const;
366 ecb_function_ uint16_t
367 ecb_bswap16 (uint16_t x)
368 {
369 return ecb_rotl16 (x, 8);
370 }
371
372 ecb_function_ uint32_t ecb_bswap32 (uint32_t x) ecb_const;
373 ecb_function_ uint32_t
374 ecb_bswap32 (uint32_t x)
375 {
376 return (((uint32_t)ecb_bswap16 (x)) << 16) | ecb_bswap16 (x >> 16);
377 }
378
379 ecb_function_ uint64_t ecb_bswap64 (uint64_t x) ecb_const;
380 ecb_function_ uint64_t
381 ecb_bswap64 (uint64_t x)
382 {
383 return (((uint64_t)ecb_bswap32 (x)) << 32) | ecb_bswap32 (x >> 32);
384 }
385 #endif
386
387 #if ECB_GCC_VERSION(4,5)
388 #define ecb_unreachable() __builtin_unreachable ()
389 #else
390 /* this seems to work fine, but gcc always emits a warning for it :/ */
391 ecb_function_ void ecb_unreachable (void) ecb_noreturn;
392 ecb_function_ void ecb_unreachable (void) { }
393 #endif
394
395 /* try to tell the compiler that some condition is definitely true */
396 #define ecb_assume(cond) do { if (!(cond)) ecb_unreachable (); } while (0)
397
398 ecb_function_ unsigned char ecb_byteorder_helper (void) ecb_const;
399 ecb_function_ unsigned char
400 ecb_byteorder_helper (void)
401 {
402 const uint32_t u = 0x11223344;
403 return *(unsigned char *)&u;
404 }
405
406 ecb_function_ ecb_bool ecb_big_endian (void) ecb_const;
407 ecb_function_ ecb_bool ecb_big_endian (void) { return ecb_byteorder_helper () == 0x11; }
408 ecb_function_ ecb_bool ecb_little_endian (void) ecb_const;
409 ecb_function_ ecb_bool ecb_little_endian (void) { return ecb_byteorder_helper () == 0x44; }
410
411 #if ECB_GCC_VERSION(3,0) || ECB_C99
412 #define ecb_mod(m,n) ((m) % (n) + ((m) % (n) < 0 ? (n) : 0))
413 #else
414 #define ecb_mod(m,n) ((m) < 0 ? ((n) - 1 - ((-1 - (m)) % (n))) : ((m) % (n)))
415 #endif
416
417 #if __cplusplus
418 template<typename T>
419 static inline T ecb_div_rd (T val, T div)
420 {
421 return val < 0 ? - ((-val + div - 1) / div) : (val ) / div;
422 }
423 template<typename T>
424 static inline T ecb_div_ru (T val, T div)
425 {
426 return val < 0 ? - ((-val ) / div) : (val + div - 1) / div;
427 }
428 #else
429 #define ecb_div_rd(val,div) ((val) < 0 ? - ((-(val) + (div) - 1) / (div)) : ((val) ) / (div))
430 #define ecb_div_ru(val,div) ((val) < 0 ? - ((-(val) ) / (div)) : ((val) + (div) - 1) / (div))
431 #endif
432
433 #if ecb_cplusplus_does_not_suck
434 /* does not work for local types (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2008/n2657.htm) */
435 template<typename T, int N>
436 static inline int ecb_array_length (const T (&arr)[N])
437 {
438 return N;
439 }
440 #else
441 #define ecb_array_length(name) (sizeof (name) / sizeof (name [0]))
442 #endif
443
444 #endif
445