… | |
… | |
38 | typedef signed int int32_t; |
38 | typedef signed int int32_t; |
39 | typedef unsigned int uint32_t; |
39 | typedef unsigned int uint32_t; |
40 | #if __GNUC__ |
40 | #if __GNUC__ |
41 | typedef signed long long int64_t; |
41 | typedef signed long long int64_t; |
42 | typedef unsigned long long uint64_t; |
42 | typedef unsigned long long uint64_t; |
43 | #else |
43 | #else /* _MSC_VER || __BORLANDC__ */ |
44 | typedef signed __int64 int64_t; |
44 | typedef signed __int64 int64_t; |
45 | typedef unsigned __int64 uint64_t; |
45 | typedef unsigned __int64 uint64_t; |
46 | #endif |
46 | #endif |
47 | #else |
47 | #else |
48 | #include <inttypes.h> |
48 | #include <inttypes.h> |
… | |
… | |
136 | #define ecb_likely(expr) ecb_expect_true (expr) |
136 | #define ecb_likely(expr) ecb_expect_true (expr) |
137 | #define ecb_unlikely(expr) ecb_expect_false (expr) |
137 | #define ecb_unlikely(expr) ecb_expect_false (expr) |
138 | |
138 | |
139 | /* count trailing zero bits and count # of one bits */ |
139 | /* count trailing zero bits and count # of one bits */ |
140 | #if ECB_GCC_VERSION(3,4) |
140 | #if ECB_GCC_VERSION(3,4) |
|
|
141 | /* we assume int == 32 bit, long == 32 or 64 bit and long long == 64 bit */ |
|
|
142 | #define ecb_ld32(x) (__builtin_clz (x) ^ 31) |
|
|
143 | #define ecb_ld64(x) (__builtin_clzll (x) ^ 63) |
141 | #define ecb_ctz32(x) __builtin_ctz (x) |
144 | #define ecb_ctz32(x) __builtin_ctz (x) |
|
|
145 | #define ecb_ctz64(x) __builtin_ctzll (x) |
142 | #define ecb_popcount32(x) __builtin_popcount (x) |
146 | #define ecb_popcount32(x) __builtin_popcount (x) |
|
|
147 | /* no popcountll */ |
143 | #else |
148 | #else |
144 | ecb_function_ int ecb_ctz32 (uint32_t x) ecb_const; |
149 | ecb_function_ int ecb_ctz32 (uint32_t x) ecb_const; |
145 | ecb_function_ int |
150 | ecb_function_ int |
146 | ecb_ctz32 (uint32_t x) |
151 | ecb_ctz32 (uint32_t x) |
147 | { |
152 | { |
148 | int r = 0; |
153 | int r = 0; |
149 | |
154 | |
150 | x &= (uint32_t)-(int32_t)x; /* this isolates the lowest bit */ |
155 | x &= ~x + 1; /* this isolates the lowest bit */ |
151 | |
156 | |
|
|
157 | #if ECB_branchless_on_i386 |
|
|
158 | r += !!(x & 0xaaaaaaaa) << 0; |
|
|
159 | r += !!(x & 0xcccccccc) << 1; |
|
|
160 | r += !!(x & 0xf0f0f0f0) << 2; |
|
|
161 | r += !!(x & 0xff00ff00) << 3; |
|
|
162 | r += !!(x & 0xffff0000) << 4; |
|
|
163 | #else |
152 | if (x & 0xaaaaaaaa) r += 1; |
164 | if (x & 0xaaaaaaaa) r += 1; |
153 | if (x & 0xcccccccc) r += 2; |
165 | if (x & 0xcccccccc) r += 2; |
154 | if (x & 0xf0f0f0f0) r += 4; |
166 | if (x & 0xf0f0f0f0) r += 4; |
155 | if (x & 0xff00ff00) r += 8; |
167 | if (x & 0xff00ff00) r += 8; |
156 | if (x & 0xffff0000) r += 16; |
168 | if (x & 0xffff0000) r += 16; |
|
|
169 | #endif |
157 | |
170 | |
158 | return r; |
171 | return r; |
|
|
172 | } |
|
|
173 | |
|
|
174 | ecb_function_ int ecb_ctz64 (uint64_t x) ecb_const; |
|
|
175 | ecb_function_ int |
|
|
176 | ecb_ctz64 (uint64_t x) |
|
|
177 | { |
|
|
178 | int shift = x & 0xffffffffU ? 0 : 32; |
|
|
179 | return ecb_ctz32 (x >> shift) + shift; |
159 | } |
180 | } |
160 | |
181 | |
161 | ecb_function_ int ecb_popcount32 (uint32_t x) ecb_const; |
182 | ecb_function_ int ecb_popcount32 (uint32_t x) ecb_const; |
162 | ecb_function_ int |
183 | ecb_function_ int |
163 | ecb_popcount32 (uint32_t x) |
184 | ecb_popcount32 (uint32_t x) |
… | |
… | |
167 | x = ((x >> 4) + x) & 0x0f0f0f0f; |
188 | x = ((x >> 4) + x) & 0x0f0f0f0f; |
168 | x *= 0x01010101; |
189 | x *= 0x01010101; |
169 | |
190 | |
170 | return x >> 24; |
191 | return x >> 24; |
171 | } |
192 | } |
|
|
193 | |
|
|
194 | /* you have the choice beetween something with a table lookup, */ |
|
|
195 | /* something using lots of bit arithmetic and a simple loop */ |
|
|
196 | /* we went for the loop */ |
|
|
197 | ecb_function_ int ecb_ld32 (uint32_t x) ecb_const; |
|
|
198 | ecb_function_ int ecb_ld32 (uint32_t x) |
|
|
199 | { |
|
|
200 | int r = 0; |
|
|
201 | |
|
|
202 | if (x >> 16) { x >>= 16; r += 16; } |
|
|
203 | if (x >> 8) { x >>= 8; r += 8; } |
|
|
204 | if (x >> 4) { x >>= 4; r += 4; } |
|
|
205 | if (x >> 2) { x >>= 2; r += 2; } |
|
|
206 | if (x >> 1) { r += 1; } |
|
|
207 | |
|
|
208 | return r; |
|
|
209 | } |
|
|
210 | |
|
|
211 | ecb_function_ int ecb_ld64 (uint64_t x) ecb_const; |
|
|
212 | ecb_function_ int ecb_ld64 (uint64_t x) |
|
|
213 | { |
|
|
214 | int r = 0; |
|
|
215 | |
|
|
216 | if (x >> 32) { x >>= 32; r += 32; } |
|
|
217 | |
|
|
218 | return r + ecb_ld32 (x); |
|
|
219 | } |
172 | #endif |
220 | #endif |
|
|
221 | |
|
|
222 | /* popcount64 is only available on 64 bit cpus as gcc builtin */ |
|
|
223 | /* so for this version we are lazy */ |
|
|
224 | ecb_function_ int ecb_popcount64 (uint64_t x) ecb_const; |
|
|
225 | ecb_function_ int |
|
|
226 | ecb_popcount64 (uint64_t x) |
|
|
227 | { |
|
|
228 | return ecb_popcount32 (x) + ecb_popcount32 (x >> 32); |
|
|
229 | } |
|
|
230 | |
|
|
231 | ecb_inline uint8_t ecb_rotl8 (uint8_t x, unsigned int count) ecb_const; |
|
|
232 | ecb_inline uint8_t ecb_rotr8 (uint8_t x, unsigned int count) ecb_const; |
|
|
233 | ecb_inline uint16_t ecb_rotl16 (uint16_t x, unsigned int count) ecb_const; |
|
|
234 | ecb_inline uint16_t ecb_rotr16 (uint16_t x, unsigned int count) ecb_const; |
|
|
235 | ecb_inline uint32_t ecb_rotl32 (uint32_t x, unsigned int count) ecb_const; |
|
|
236 | ecb_inline uint32_t ecb_rotr32 (uint32_t x, unsigned int count) ecb_const; |
|
|
237 | ecb_inline uint64_t ecb_rotl64 (uint64_t x, unsigned int count) ecb_const; |
|
|
238 | ecb_inline uint64_t ecb_rotr64 (uint64_t x, unsigned int count) ecb_const; |
|
|
239 | |
|
|
240 | ecb_inline uint8_t ecb_rotl8 (uint8_t x, unsigned int count) { return (x >> ( 8 - count)) | (x << count); } |
|
|
241 | ecb_inline uint8_t ecb_rotr8 (uint8_t x, unsigned int count) { return (x << ( 8 - count)) | (x >> count); } |
|
|
242 | ecb_inline uint16_t ecb_rotl16 (uint16_t x, unsigned int count) { return (x >> (16 - count)) | (x << count); } |
|
|
243 | ecb_inline uint16_t ecb_rotr16 (uint16_t x, unsigned int count) { return (x << (16 - count)) | (x >> count); } |
|
|
244 | ecb_inline uint32_t ecb_rotl32 (uint32_t x, unsigned int count) { return (x >> (32 - count)) | (x << count); } |
|
|
245 | ecb_inline uint32_t ecb_rotr32 (uint32_t x, unsigned int count) { return (x << (32 - count)) | (x >> count); } |
|
|
246 | ecb_inline uint64_t ecb_rotl64 (uint64_t x, unsigned int count) { return (x >> (64 - count)) | (x << count); } |
|
|
247 | ecb_inline uint64_t ecb_rotr64 (uint64_t x, unsigned int count) { return (x << (64 - count)) | (x >> count); } |
173 | |
248 | |
174 | #if ECB_GCC_VERSION(4,3) |
249 | #if ECB_GCC_VERSION(4,3) |
175 | #define ecb_bswap32(x) __builtin_bswap32 (x) |
|
|
176 | #define ecb_bswap16(x) (__builtin_bswap32(x) >> 16) |
250 | #define ecb_bswap16(x) (__builtin_bswap32 (x) >> 16) |
|
|
251 | #define ecb_bswap32(x) __builtin_bswap32 (x) |
|
|
252 | #define ecb_bswap64(x) __builtin_bswap64 (x) |
177 | #else |
253 | #else |
|
|
254 | ecb_function_ uint16_t ecb_bswap16 (uint16_t x) ecb_const; |
|
|
255 | ecb_function_ uint16_t |
|
|
256 | ecb_bswap16 (uint16_t x) |
|
|
257 | { |
|
|
258 | return ecb_rotl16 (x, 8); |
|
|
259 | } |
|
|
260 | |
178 | ecb_function_ uint32_t ecb_bswap32 (uint32_t x) ecb_const; |
261 | ecb_function_ uint32_t ecb_bswap32 (uint32_t x) ecb_const; |
179 | ecb_function_ uint32_t |
262 | ecb_function_ uint32_t |
180 | ecb_bswap32 (uint32_t x) |
263 | ecb_bswap32 (uint32_t x) |
181 | { |
264 | { |
182 | return (x >> 24) |
265 | return (((uint32_t)ecb_bswap16 (x)) << 16) | ecb_bswap16 (x >> 16); |
183 | | ((x >> 8) & 0x0000ff00) |
|
|
184 | | ((x << 8) & 0x00ff0000) |
|
|
185 | | (x << 24); |
|
|
186 | } |
266 | } |
187 | |
267 | |
188 | ecb_function_ uint32_t ecb_bswap16 (uint32_t x) ecb_const; |
268 | ecb_function_ uint64_t ecb_bswap64 (uint64_t x) ecb_const; |
189 | ecb_function_ uint32_t |
269 | ecb_function_ uint64_t |
190 | ecb_bswap16 (uint32_t x) |
270 | ecb_bswap64 (uint64_t x) |
191 | { |
271 | { |
192 | return ((x >> 8) & 0xff) |
272 | return (((uint64_t)ecb_bswap32 (x)) << 32) | ecb_bswap32 (x >> 32); |
193 | | ((x << 8) & 0x00ff0000) |
|
|
194 | | (x << 24); |
|
|
195 | } |
273 | } |
196 | #endif |
274 | #endif |
197 | |
275 | |
198 | #if ECB_GCC_VERSION(4,5) |
276 | #if ECB_GCC_VERSION(4,5) |
199 | #define ecb_unreachable() __builtin_unreachable () |
277 | #define ecb_unreachable() __builtin_unreachable () |
… | |
… | |
234 | } |
312 | } |
235 | #else |
313 | #else |
236 | #define ecb_array_length(name) (sizeof (name) / sizeof (name [0])) |
314 | #define ecb_array_length(name) (sizeof (name) / sizeof (name [0])) |
237 | #endif |
315 | #endif |
238 | |
316 | |
239 | ecb_inline uint32_t ecb_rotr32 (uint32_t x, unsigned int count) ecb_const; |
|
|
240 | ecb_inline uint32_t |
|
|
241 | ecb_rotr32 (uint32_t x, unsigned int count) |
|
|
242 | { |
|
|
243 | return (x << (32 - count)) | (x >> count); |
|
|
244 | } |
|
|
245 | |
|
|
246 | ecb_inline uint32_t ecb_rotl32 (uint32_t x, unsigned int count) ecb_const; |
|
|
247 | ecb_inline uint32_t |
|
|
248 | ecb_rotl32 (uint32_t x, unsigned int count) |
|
|
249 | { |
|
|
250 | return (x >> (32 - count)) | (x << count); |
|
|
251 | } |
|
|
252 | |
|
|
253 | #endif |
317 | #endif |
254 | |
318 | |