1 |
/* |
2 |
* libecb - http://software.schmorp.de/pkg/libecb |
3 |
* |
4 |
* Copyright (©) 2009-2011 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 |
#define ECB_MEMORY_FENCE_ACQUIRE do { } while (0) |
74 |
#define ECB_MEMORY_FENCE_RELEASE do { } while (0) |
75 |
#endif |
76 |
|
77 |
#ifndef ECB_MEMORY_FENCE |
78 |
#if ECB_GCC_VERSION(2,5) |
79 |
#if __x86 |
80 |
#define ECB_MEMORY_FENCE __asm__ __volatile__ ("lock; orb $0, -1(%%esp)" : : : "memory") |
81 |
#define ECB_MEMORY_FENCE_ACQUIRE ECB_MEMORY_FENCE /* non-lock xchg might be enough */ |
82 |
#define ECB_MEMORY_FENCE_RELEASE do { } while (0) /* unlikely to change in future cpus */ |
83 |
#elif __amd64 |
84 |
#define ECB_MEMORY_FENCE __asm__ __volatile__ ("mfence" : : : "memory") |
85 |
#define ECB_MEMORY_FENCE_ACQUIRE __asm__ __volatile__ ("lfence" : : : "memory") |
86 |
#define ECB_MEMORY_FENCE_RELEASE __asm__ __volatile__ ("sfence") /* play safe - not needed in any current cpu */ |
87 |
#endif |
88 |
#endif |
89 |
#endif |
90 |
|
91 |
#ifndef ECB_MEMORY_FENCE |
92 |
#if ECB_GCC_VERSION(4,4) |
93 |
#define ECB_MEMORY_FENCE __sync_synchronize () |
94 |
#define ECB_MEMORY_FENCE_ACQUIRE ({ char dummy = 0; __sync_lock_test_and_set (&dummy, 1); }) |
95 |
#define ECB_MEMORY_FENCE_RELEASE ({ char dummy = 1; __sync_lock_release (&dummy ); }) |
96 |
#elif _MSC_VER >= 1400 /* VC++ 2005 */ |
97 |
#pragma intrinsic(_ReadBarrier,_WriteBarrier,_ReadWriteBarrier) |
98 |
#define ECB_MEMORY_FENCE _ReadWriteBarrier () |
99 |
#define ECB_MEMORY_FENCE_ACQUIRE _ReadWriteBarrier () /* according to msdn, _ReadBarrier is not a load fence */ |
100 |
#define ECB_MEMORY_FENCE_RELEASE _WriteBarrier () |
101 |
#elif defined(_WIN32) |
102 |
#include <WinNT.h> |
103 |
#define ECB_MEMORY_FENCE MemoryBarrier () /* actually just xchg on x86... scary */ |
104 |
#define ECB_MEMORY_FENCE_ACQUIRE ECB_MEMORY_FENCE |
105 |
#define ECB_MEMORY_FENCE_RELEASE ECB_MEMORY_FENCE |
106 |
#endif |
107 |
#endif |
108 |
|
109 |
#ifndef ECB_MEMORY_FENCE |
110 |
/* |
111 |
* if you get undefined symbol references to pthread_mutex_lock, |
112 |
* or failure to find pthread.h, then you should implement |
113 |
* the ECB_MEMORY_FENCE operations for your cpu/compiler |
114 |
* OR provide pthread.h and link against the posix thread library |
115 |
* of your system. |
116 |
*/ |
117 |
#include <pthread.h> |
118 |
|
119 |
static pthread_mutex_t ecb_mf_lock = PTHREAD_MUTEX_INITIALIZER; |
120 |
#define ECB_MEMORY_FENCE do { pthread_mutex_lock (&ecb_mf_lock); pthread_mutex_unlock (&ecb_mf_lock); } while (0) |
121 |
#define ECB_MEMORY_FENCE_ACQUIRE ECB_MEMORY_FENCE |
122 |
#define ECB_MEMORY_FENCE_RELEASE ECB_MEMORY_FENCE |
123 |
#endif |
124 |
|
125 |
/*****************************************************************************/ |
126 |
|
127 |
#define ECB_C99 (__STDC_VERSION__ >= 199901L) |
128 |
|
129 |
#if __cplusplus |
130 |
#define ecb_inline static inline |
131 |
#elif ECB_GCC_VERSION(2,5) |
132 |
#define ecb_inline static __inline__ |
133 |
#elif ECB_C99 |
134 |
#define ecb_inline static inline |
135 |
#else |
136 |
#define ecb_inline static |
137 |
#endif |
138 |
|
139 |
#if ECB_GCC_VERSION(3,3) |
140 |
#define ecb_restrict __restrict__ |
141 |
#elif ECB_C99 |
142 |
#define ecb_restrict restrict |
143 |
#else |
144 |
#define ecb_restrict |
145 |
#endif |
146 |
|
147 |
typedef int ecb_bool; |
148 |
|
149 |
#define ECB_CONCAT_(a, b) a ## b |
150 |
#define ECB_CONCAT(a, b) ECB_CONCAT_(a, b) |
151 |
#define ECB_STRINGIFY_(a) # a |
152 |
#define ECB_STRINGIFY(a) ECB_STRINGIFY_(a) |
153 |
|
154 |
#define ecb_function_ ecb_inline |
155 |
|
156 |
#if ECB_GCC_VERSION(3,1) |
157 |
#define ecb_attribute(attrlist) __attribute__(attrlist) |
158 |
#define ecb_is_constant(expr) __builtin_constant_p (expr) |
159 |
#define ecb_expect(expr,value) __builtin_expect ((expr),(value)) |
160 |
#define ecb_prefetch(addr,rw,locality) __builtin_prefetch (addr, rw, locality) |
161 |
#else |
162 |
#define ecb_attribute(attrlist) |
163 |
#define ecb_is_constant(expr) 0 |
164 |
#define ecb_expect(expr,value) (expr) |
165 |
#define ecb_prefetch(addr,rw,locality) |
166 |
#endif |
167 |
|
168 |
/* no emulation for ecb_decltype */ |
169 |
#if ECB_GCC_VERSION(4,5) |
170 |
#define ecb_decltype(x) __decltype(x) |
171 |
#elif ECB_GCC_VERSION(3,0) |
172 |
#define ecb_decltype(x) __typeof(x) |
173 |
#endif |
174 |
|
175 |
#define ecb_noinline ecb_attribute ((__noinline__)) |
176 |
#define ecb_noreturn ecb_attribute ((__noreturn__)) |
177 |
#define ecb_unused ecb_attribute ((__unused__)) |
178 |
#define ecb_const ecb_attribute ((__const__)) |
179 |
#define ecb_pure ecb_attribute ((__pure__)) |
180 |
|
181 |
#if ECB_GCC_VERSION(4,3) |
182 |
#define ecb_artificial ecb_attribute ((__artificial__)) |
183 |
#define ecb_hot ecb_attribute ((__hot__)) |
184 |
#define ecb_cold ecb_attribute ((__cold__)) |
185 |
#else |
186 |
#define ecb_artificial |
187 |
#define ecb_hot |
188 |
#define ecb_cold |
189 |
#endif |
190 |
|
191 |
/* put around conditional expressions if you are very sure that the */ |
192 |
/* expression is mostly true or mostly false. note that these return */ |
193 |
/* booleans, not the expression. */ |
194 |
#define ecb_expect_false(expr) ecb_expect (!!(expr), 0) |
195 |
#define ecb_expect_true(expr) ecb_expect (!!(expr), 1) |
196 |
/* for compatibility to the rest of the world */ |
197 |
#define ecb_likely(expr) ecb_expect_true (expr) |
198 |
#define ecb_unlikely(expr) ecb_expect_false (expr) |
199 |
|
200 |
/* count trailing zero bits and count # of one bits */ |
201 |
#if ECB_GCC_VERSION(3,4) |
202 |
/* we assume int == 32 bit, long == 32 or 64 bit and long long == 64 bit */ |
203 |
#define ecb_ld32(x) (__builtin_clz (x) ^ 31) |
204 |
#define ecb_ld64(x) (__builtin_clzll (x) ^ 63) |
205 |
#define ecb_ctz32(x) __builtin_ctz (x) |
206 |
#define ecb_ctz64(x) __builtin_ctzll (x) |
207 |
#define ecb_popcount32(x) __builtin_popcount (x) |
208 |
/* no popcountll */ |
209 |
#else |
210 |
ecb_function_ int ecb_ctz32 (uint32_t x) ecb_const; |
211 |
ecb_function_ int |
212 |
ecb_ctz32 (uint32_t x) |
213 |
{ |
214 |
int r = 0; |
215 |
|
216 |
x &= ~x + 1; /* this isolates the lowest bit */ |
217 |
|
218 |
#if ECB_branchless_on_i386 |
219 |
r += !!(x & 0xaaaaaaaa) << 0; |
220 |
r += !!(x & 0xcccccccc) << 1; |
221 |
r += !!(x & 0xf0f0f0f0) << 2; |
222 |
r += !!(x & 0xff00ff00) << 3; |
223 |
r += !!(x & 0xffff0000) << 4; |
224 |
#else |
225 |
if (x & 0xaaaaaaaa) r += 1; |
226 |
if (x & 0xcccccccc) r += 2; |
227 |
if (x & 0xf0f0f0f0) r += 4; |
228 |
if (x & 0xff00ff00) r += 8; |
229 |
if (x & 0xffff0000) r += 16; |
230 |
#endif |
231 |
|
232 |
return r; |
233 |
} |
234 |
|
235 |
ecb_function_ int ecb_ctz64 (uint64_t x) ecb_const; |
236 |
ecb_function_ int |
237 |
ecb_ctz64 (uint64_t x) |
238 |
{ |
239 |
int shift = x & 0xffffffffU ? 0 : 32; |
240 |
return ecb_ctz32 (x >> shift) + shift; |
241 |
} |
242 |
|
243 |
ecb_function_ int ecb_popcount32 (uint32_t x) ecb_const; |
244 |
ecb_function_ int |
245 |
ecb_popcount32 (uint32_t x) |
246 |
{ |
247 |
x -= (x >> 1) & 0x55555555; |
248 |
x = ((x >> 2) & 0x33333333) + (x & 0x33333333); |
249 |
x = ((x >> 4) + x) & 0x0f0f0f0f; |
250 |
x *= 0x01010101; |
251 |
|
252 |
return x >> 24; |
253 |
} |
254 |
|
255 |
ecb_function_ int ecb_ld32 (uint32_t x) ecb_const; |
256 |
ecb_function_ int ecb_ld32 (uint32_t x) |
257 |
{ |
258 |
int r = 0; |
259 |
|
260 |
if (x >> 16) { x >>= 16; r += 16; } |
261 |
if (x >> 8) { x >>= 8; r += 8; } |
262 |
if (x >> 4) { x >>= 4; r += 4; } |
263 |
if (x >> 2) { x >>= 2; r += 2; } |
264 |
if (x >> 1) { r += 1; } |
265 |
|
266 |
return r; |
267 |
} |
268 |
|
269 |
ecb_function_ int ecb_ld64 (uint64_t x) ecb_const; |
270 |
ecb_function_ int ecb_ld64 (uint64_t x) |
271 |
{ |
272 |
int r = 0; |
273 |
|
274 |
if (x >> 32) { x >>= 32; r += 32; } |
275 |
|
276 |
return r + ecb_ld32 (x); |
277 |
} |
278 |
#endif |
279 |
|
280 |
/* popcount64 is only available on 64 bit cpus as gcc builtin */ |
281 |
/* so for this version we are lazy */ |
282 |
ecb_function_ int ecb_popcount64 (uint64_t x) ecb_const; |
283 |
ecb_function_ int |
284 |
ecb_popcount64 (uint64_t x) |
285 |
{ |
286 |
return ecb_popcount32 (x) + ecb_popcount32 (x >> 32); |
287 |
} |
288 |
|
289 |
ecb_inline uint8_t ecb_rotl8 (uint8_t x, unsigned int count) ecb_const; |
290 |
ecb_inline uint8_t ecb_rotr8 (uint8_t x, unsigned int count) ecb_const; |
291 |
ecb_inline uint16_t ecb_rotl16 (uint16_t x, unsigned int count) ecb_const; |
292 |
ecb_inline uint16_t ecb_rotr16 (uint16_t x, unsigned int count) ecb_const; |
293 |
ecb_inline uint32_t ecb_rotl32 (uint32_t x, unsigned int count) ecb_const; |
294 |
ecb_inline uint32_t ecb_rotr32 (uint32_t x, unsigned int count) ecb_const; |
295 |
ecb_inline uint64_t ecb_rotl64 (uint64_t x, unsigned int count) ecb_const; |
296 |
ecb_inline uint64_t ecb_rotr64 (uint64_t x, unsigned int count) ecb_const; |
297 |
|
298 |
ecb_inline uint8_t ecb_rotl8 (uint8_t x, unsigned int count) { return (x >> ( 8 - count)) | (x << count); } |
299 |
ecb_inline uint8_t ecb_rotr8 (uint8_t x, unsigned int count) { return (x << ( 8 - count)) | (x >> count); } |
300 |
ecb_inline uint16_t ecb_rotl16 (uint16_t x, unsigned int count) { return (x >> (16 - count)) | (x << count); } |
301 |
ecb_inline uint16_t ecb_rotr16 (uint16_t x, unsigned int count) { return (x << (16 - count)) | (x >> count); } |
302 |
ecb_inline uint32_t ecb_rotl32 (uint32_t x, unsigned int count) { return (x >> (32 - count)) | (x << count); } |
303 |
ecb_inline uint32_t ecb_rotr32 (uint32_t x, unsigned int count) { return (x << (32 - count)) | (x >> count); } |
304 |
ecb_inline uint64_t ecb_rotl64 (uint64_t x, unsigned int count) { return (x >> (64 - count)) | (x << count); } |
305 |
ecb_inline uint64_t ecb_rotr64 (uint64_t x, unsigned int count) { return (x << (64 - count)) | (x >> count); } |
306 |
|
307 |
#if ECB_GCC_VERSION(4,3) |
308 |
#define ecb_bswap16(x) (__builtin_bswap32 (x) >> 16) |
309 |
#define ecb_bswap32(x) __builtin_bswap32 (x) |
310 |
#define ecb_bswap64(x) __builtin_bswap64 (x) |
311 |
#else |
312 |
ecb_function_ uint16_t ecb_bswap16 (uint16_t x) ecb_const; |
313 |
ecb_function_ uint16_t |
314 |
ecb_bswap16 (uint16_t x) |
315 |
{ |
316 |
return ecb_rotl16 (x, 8); |
317 |
} |
318 |
|
319 |
ecb_function_ uint32_t ecb_bswap32 (uint32_t x) ecb_const; |
320 |
ecb_function_ uint32_t |
321 |
ecb_bswap32 (uint32_t x) |
322 |
{ |
323 |
return (((uint32_t)ecb_bswap16 (x)) << 16) | ecb_bswap16 (x >> 16); |
324 |
} |
325 |
|
326 |
ecb_function_ uint64_t ecb_bswap64 (uint64_t x) ecb_const; |
327 |
ecb_function_ uint64_t |
328 |
ecb_bswap64 (uint64_t x) |
329 |
{ |
330 |
return (((uint64_t)ecb_bswap32 (x)) << 32) | ecb_bswap32 (x >> 32); |
331 |
} |
332 |
#endif |
333 |
|
334 |
#if ECB_GCC_VERSION(4,5) |
335 |
#define ecb_unreachable() __builtin_unreachable () |
336 |
#else |
337 |
/* this seems to work fine, but gcc always emits a warning for it :/ */ |
338 |
ecb_function_ void ecb_unreachable (void) ecb_noreturn; |
339 |
ecb_function_ void ecb_unreachable (void) { } |
340 |
#endif |
341 |
|
342 |
/* try to tell the compiler that some condition is definitely true */ |
343 |
#define ecb_assume(cond) do { if (!(cond)) ecb_unreachable (); } while (0) |
344 |
|
345 |
ecb_function_ unsigned char ecb_byteorder_helper (void) ecb_const; |
346 |
ecb_function_ unsigned char |
347 |
ecb_byteorder_helper (void) |
348 |
{ |
349 |
const uint32_t u = 0x11223344; |
350 |
return *(unsigned char *)&u; |
351 |
} |
352 |
|
353 |
ecb_function_ ecb_bool ecb_big_endian (void) ecb_const; |
354 |
ecb_function_ ecb_bool ecb_big_endian (void) { return ecb_byteorder_helper () == 0x11; } |
355 |
ecb_function_ ecb_bool ecb_little_endian (void) ecb_const; |
356 |
ecb_function_ ecb_bool ecb_little_endian (void) { return ecb_byteorder_helper () == 0x44; } |
357 |
|
358 |
#if ECB_GCC_VERSION(3,0) || ECB_C99 |
359 |
#define ecb_mod(m,n) ((m) % (n) + ((m) % (n) < 0 ? (n) : 0)) |
360 |
#else |
361 |
#define ecb_mod(m,n) ((m) < 0 ? ((n) - 1 - ((-1 - (m)) % (n))) : ((m) % (n))) |
362 |
#endif |
363 |
|
364 |
#if ecb_cplusplus_does_not_suck |
365 |
/* does not work for local types (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2008/n2657.htm) */ |
366 |
template<typename T, int N> |
367 |
static inline int ecb_array_length (const T (&arr)[N]) |
368 |
{ |
369 |
return N; |
370 |
} |
371 |
#else |
372 |
#define ecb_array_length(name) (sizeof (name) / sizeof (name [0])) |
373 |
#endif |
374 |
|
375 |
#endif |
376 |
|