1 |
siliconforks |
332 |
/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*- */ |
2 |
|
|
/* ***** BEGIN LICENSE BLOCK ***** |
3 |
|
|
* Version: MPL 1.1/GPL 2.0/LGPL 2.1 |
4 |
|
|
* |
5 |
|
|
* The contents of this file are subject to the Mozilla Public License Version |
6 |
|
|
* 1.1 (the "License"); you may not use this file except in compliance with |
7 |
|
|
* the License. You may obtain a copy of the License at |
8 |
|
|
* http://www.mozilla.org/MPL/ |
9 |
|
|
* |
10 |
|
|
* Software distributed under the License is distributed on an "AS IS" basis, |
11 |
|
|
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License |
12 |
|
|
* for the specific language governing rights and limitations under the |
13 |
|
|
* License. |
14 |
|
|
* |
15 |
|
|
* The Original Code is Mozilla Communicator client code, released |
16 |
|
|
* March 31, 1998. |
17 |
|
|
* |
18 |
|
|
* The Initial Developer of the Original Code is |
19 |
|
|
* Netscape Communications Corporation. |
20 |
|
|
* Portions created by the Initial Developer are Copyright (C) 1998 |
21 |
|
|
* the Initial Developer. All Rights Reserved. |
22 |
|
|
* |
23 |
|
|
* Contributor(s): |
24 |
|
|
* |
25 |
|
|
* Alternatively, the contents of this file may be used under the terms of |
26 |
|
|
* either of the GNU General Public License Version 2 or later (the "GPL"), |
27 |
|
|
* or the GNU Lesser General Public License Version 2.1 or later (the "LGPL"), |
28 |
|
|
* in which case the provisions of the GPL or the LGPL are applicable instead |
29 |
|
|
* of those above. If you wish to allow use of your version of this file only |
30 |
|
|
* under the terms of either the GPL or the LGPL, and not to allow others to |
31 |
|
|
* use your version of this file under the terms of the MPL, indicate your |
32 |
|
|
* decision by deleting the provisions above and replace them with the notice |
33 |
|
|
* and other provisions required by the GPL or the LGPL. If you do not delete |
34 |
|
|
* the provisions above, a recipient may use your version of this file under |
35 |
|
|
* the terms of any one of the MPL, the GPL or the LGPL. |
36 |
|
|
* |
37 |
|
|
* ***** END LICENSE BLOCK ***** */ |
38 |
|
|
|
39 |
|
|
#ifndef jsbit_h___ |
40 |
|
|
#define jsbit_h___ |
41 |
|
|
|
42 |
|
|
#include "jstypes.h" |
43 |
|
|
#include "jsutil.h" |
44 |
|
|
|
45 |
|
|
JS_BEGIN_EXTERN_C |
46 |
|
|
|
47 |
|
|
/* |
48 |
|
|
** A jsbitmap_t is a long integer that can be used for bitmaps |
49 |
|
|
*/ |
50 |
|
|
typedef JSUword jsbitmap_t; /* NSPR name, a la Unix system types */ |
51 |
|
|
typedef jsbitmap_t jsbitmap; /* JS-style scalar typedef name */ |
52 |
|
|
|
53 |
|
|
#define JS_BITMAP_SIZE(bits) (JS_HOWMANY(bits, JS_BITS_PER_WORD) * \ |
54 |
|
|
sizeof(jsbitmap)) |
55 |
|
|
|
56 |
|
|
#define JS_TEST_BIT(_map,_bit) ((_map)[(_bit)>>JS_BITS_PER_WORD_LOG2] & \ |
57 |
|
|
((jsbitmap)1<<((_bit)&(JS_BITS_PER_WORD-1)))) |
58 |
|
|
#define JS_SET_BIT(_map,_bit) ((_map)[(_bit)>>JS_BITS_PER_WORD_LOG2] |= \ |
59 |
|
|
((jsbitmap)1<<((_bit)&(JS_BITS_PER_WORD-1)))) |
60 |
|
|
#define JS_CLEAR_BIT(_map,_bit) ((_map)[(_bit)>>JS_BITS_PER_WORD_LOG2] &= \ |
61 |
|
|
~((jsbitmap)1<<((_bit)&(JS_BITS_PER_WORD-1)))) |
62 |
|
|
|
63 |
|
|
/* |
64 |
|
|
** Compute the log of the least power of 2 greater than or equal to n |
65 |
|
|
*/ |
66 |
|
|
extern JS_PUBLIC_API(JSIntn) JS_CeilingLog2(JSUint32 i); |
67 |
|
|
|
68 |
|
|
/* |
69 |
|
|
** Compute the log of the greatest power of 2 less than or equal to n |
70 |
|
|
*/ |
71 |
|
|
extern JS_PUBLIC_API(JSIntn) JS_FloorLog2(JSUint32 i); |
72 |
|
|
|
73 |
|
|
/* |
74 |
|
|
* Replace bit-scanning code sequences with CPU-specific instructions to |
75 |
|
|
* speedup calculations of ceiling/floor log2. |
76 |
|
|
* |
77 |
|
|
* With GCC 3.4 or later we can use __builtin_clz for that, see bug 327129. |
78 |
|
|
* |
79 |
|
|
* SWS: Added MSVC intrinsic bitscan support. See bugs 349364 and 356856. |
80 |
|
|
*/ |
81 |
siliconforks |
507 |
#if defined(_WIN32) && (_MSC_VER >= 1300) && (defined(_M_IX86) || defined(_M_AMD64) || defined(_M_X64)) |
82 |
siliconforks |
332 |
|
83 |
|
|
unsigned char _BitScanForward(unsigned long * Index, unsigned long Mask); |
84 |
|
|
unsigned char _BitScanReverse(unsigned long * Index, unsigned long Mask); |
85 |
|
|
# pragma intrinsic(_BitScanForward,_BitScanReverse) |
86 |
|
|
|
87 |
|
|
__forceinline static int |
88 |
|
|
__BitScanForward32(unsigned int val) |
89 |
|
|
{ |
90 |
|
|
unsigned long idx; |
91 |
|
|
|
92 |
|
|
_BitScanForward(&idx, (unsigned long)val); |
93 |
|
|
return (int)idx; |
94 |
|
|
} |
95 |
|
|
__forceinline static int |
96 |
|
|
__BitScanReverse32(unsigned int val) |
97 |
|
|
{ |
98 |
|
|
unsigned long idx; |
99 |
|
|
|
100 |
|
|
_BitScanReverse(&idx, (unsigned long)val); |
101 |
|
|
return (int)(31-idx); |
102 |
|
|
} |
103 |
|
|
# define js_bitscan_ctz32(val) __BitScanForward32(val) |
104 |
|
|
# define js_bitscan_clz32(val) __BitScanReverse32(val) |
105 |
|
|
# define JS_HAS_BUILTIN_BITSCAN32 |
106 |
|
|
|
107 |
siliconforks |
507 |
#if defined(_M_AMD64) || defined(_M_X64) |
108 |
|
|
unsigned char _BitScanForward64(unsigned long * Index, unsigned __int64 Mask); |
109 |
|
|
unsigned char _BitScanReverse64(unsigned long * Index, unsigned __int64 Mask); |
110 |
|
|
# pragma intrinsic(_BitScanForward64,_BitScanReverse64) |
111 |
|
|
|
112 |
|
|
__forceinline static int |
113 |
|
|
__BitScanForward64(unsigned __int64 val) |
114 |
|
|
{ |
115 |
|
|
unsigned long idx; |
116 |
|
|
|
117 |
|
|
_BitScanForward64(&idx, val); |
118 |
|
|
return (int)idx; |
119 |
|
|
} |
120 |
|
|
__forceinline static int |
121 |
|
|
__BitScanReverse64(unsigned __int64 val) |
122 |
|
|
{ |
123 |
|
|
unsigned long idx; |
124 |
|
|
|
125 |
|
|
_BitScanReverse64(&idx, val); |
126 |
|
|
return (int)(63-idx); |
127 |
|
|
} |
128 |
|
|
# define js_bitscan_ctz64(val) __BitScanForward64(val) |
129 |
|
|
# define js_bitscan_clz64(val) __BitScanReverse64(val) |
130 |
|
|
# define JS_HAS_BUILTIN_BITSCAN64 |
131 |
|
|
#endif |
132 |
siliconforks |
332 |
#elif (__GNUC__ >= 4) || (__GNUC__ == 3 && __GNUC_MINOR__ >= 4) |
133 |
|
|
|
134 |
|
|
# define js_bitscan_ctz32(val) __builtin_ctz(val) |
135 |
|
|
# define js_bitscan_clz32(val) __builtin_clz(val) |
136 |
|
|
# define JS_HAS_BUILTIN_BITSCAN32 |
137 |
|
|
# if (JS_BYTES_PER_WORD == 8) |
138 |
|
|
# define js_bitscan_ctz64(val) __builtin_ctzll(val) |
139 |
|
|
# define js_bitscan_clz64(val) __builtin_clzll(val) |
140 |
|
|
# define JS_HAS_BUILTIN_BITSCAN64 |
141 |
|
|
# endif |
142 |
|
|
|
143 |
|
|
#endif |
144 |
|
|
|
145 |
|
|
/* |
146 |
|
|
** Macro version of JS_CeilingLog2: Compute the log of the least power of |
147 |
|
|
** 2 greater than or equal to _n. The result is returned in _log2. |
148 |
|
|
*/ |
149 |
|
|
#ifdef JS_HAS_BUILTIN_BITSCAN32 |
150 |
|
|
/* |
151 |
|
|
* Use intrinsic function or count-leading-zeros to calculate ceil(log2(_n)). |
152 |
|
|
* The macro checks for "n <= 1" and not "n != 0" as js_bitscan_clz32(0) is |
153 |
|
|
* undefined. |
154 |
|
|
*/ |
155 |
|
|
# define JS_CEILING_LOG2(_log2,_n) \ |
156 |
|
|
JS_BEGIN_MACRO \ |
157 |
|
|
unsigned int j_ = (unsigned int)(_n); \ |
158 |
|
|
(_log2) = (j_ <= 1 ? 0 : 32 - js_bitscan_clz32(j_ - 1)); \ |
159 |
|
|
JS_END_MACRO |
160 |
|
|
#else |
161 |
|
|
# define JS_CEILING_LOG2(_log2,_n) \ |
162 |
|
|
JS_BEGIN_MACRO \ |
163 |
|
|
JSUint32 j_ = (JSUint32)(_n); \ |
164 |
|
|
(_log2) = 0; \ |
165 |
|
|
if ((j_) & ((j_)-1)) \ |
166 |
|
|
(_log2) += 1; \ |
167 |
|
|
if ((j_) >> 16) \ |
168 |
|
|
(_log2) += 16, (j_) >>= 16; \ |
169 |
|
|
if ((j_) >> 8) \ |
170 |
|
|
(_log2) += 8, (j_) >>= 8; \ |
171 |
|
|
if ((j_) >> 4) \ |
172 |
|
|
(_log2) += 4, (j_) >>= 4; \ |
173 |
|
|
if ((j_) >> 2) \ |
174 |
|
|
(_log2) += 2, (j_) >>= 2; \ |
175 |
|
|
if ((j_) >> 1) \ |
176 |
|
|
(_log2) += 1; \ |
177 |
|
|
JS_END_MACRO |
178 |
|
|
#endif |
179 |
|
|
|
180 |
|
|
/* |
181 |
|
|
** Macro version of JS_FloorLog2: Compute the log of the greatest power of |
182 |
|
|
** 2 less than or equal to _n. The result is returned in _log2. |
183 |
|
|
** |
184 |
|
|
** This is equivalent to finding the highest set bit in the word. |
185 |
|
|
*/ |
186 |
|
|
#ifdef JS_HAS_BUILTIN_BITSCAN32 |
187 |
|
|
/* |
188 |
|
|
* Use js_bitscan_clz32 or count-leading-zeros to calculate floor(log2(_n)). |
189 |
|
|
* Since js_bitscan_clz32(0) is undefined, the macro set the loweset bit to 1 |
190 |
|
|
* to ensure 0 result when _n == 0. |
191 |
|
|
*/ |
192 |
|
|
# define JS_FLOOR_LOG2(_log2,_n) \ |
193 |
|
|
JS_BEGIN_MACRO \ |
194 |
|
|
(_log2) = 31 - js_bitscan_clz32(((unsigned int)(_n)) | 1); \ |
195 |
|
|
JS_END_MACRO |
196 |
|
|
#else |
197 |
|
|
# define JS_FLOOR_LOG2(_log2,_n) \ |
198 |
|
|
JS_BEGIN_MACRO \ |
199 |
|
|
JSUint32 j_ = (JSUint32)(_n); \ |
200 |
|
|
(_log2) = 0; \ |
201 |
|
|
if ((j_) >> 16) \ |
202 |
|
|
(_log2) += 16, (j_) >>= 16; \ |
203 |
|
|
if ((j_) >> 8) \ |
204 |
|
|
(_log2) += 8, (j_) >>= 8; \ |
205 |
|
|
if ((j_) >> 4) \ |
206 |
|
|
(_log2) += 4, (j_) >>= 4; \ |
207 |
|
|
if ((j_) >> 2) \ |
208 |
|
|
(_log2) += 2, (j_) >>= 2; \ |
209 |
|
|
if ((j_) >> 1) \ |
210 |
|
|
(_log2) += 1; \ |
211 |
|
|
JS_END_MACRO |
212 |
|
|
#endif |
213 |
|
|
|
214 |
|
|
/* |
215 |
|
|
* Internal function. |
216 |
|
|
* Compute the log of the least power of 2 greater than or equal to n. |
217 |
|
|
* This is a version of JS_CeilingLog2 that operates on jsuword with |
218 |
|
|
* CPU-dependant size. |
219 |
|
|
*/ |
220 |
|
|
#define JS_CEILING_LOG2W(n) ((n) <= 1 ? 0 : 1 + JS_FLOOR_LOG2W((n) - 1)) |
221 |
|
|
|
222 |
|
|
/* |
223 |
|
|
* Internal function. |
224 |
|
|
* Compute the log of the greatest power of 2 less than or equal to n. |
225 |
|
|
* This is a version of JS_FloorLog2 that operates on jsuword with |
226 |
|
|
* CPU-dependant size and requires that n != 0. |
227 |
|
|
*/ |
228 |
|
|
#define JS_FLOOR_LOG2W(n) (JS_ASSERT((n) != 0), js_FloorLog2wImpl(n)) |
229 |
|
|
|
230 |
|
|
#if JS_BYTES_PER_WORD == 4 |
231 |
|
|
|
232 |
|
|
# ifdef JS_HAS_BUILTIN_BITSCAN32 |
233 |
|
|
# define js_FloorLog2wImpl(n) \ |
234 |
|
|
((JSUword)(JS_BITS_PER_WORD - 1 - js_bitscan_clz32(n))) |
235 |
|
|
# else |
236 |
|
|
# define js_FloorLog2wImpl(n) ((JSUword)JS_FloorLog2(n)) |
237 |
|
|
#endif |
238 |
|
|
|
239 |
|
|
#elif JS_BYTES_PER_WORD == 8 |
240 |
|
|
|
241 |
|
|
# ifdef JS_HAS_BUILTIN_BITSCAN64 |
242 |
|
|
# define js_FloorLog2wImpl(n) \ |
243 |
|
|
((JSUword)(JS_BITS_PER_WORD - 1 - js_bitscan_clz64(n))) |
244 |
|
|
# else |
245 |
|
|
extern JSUword js_FloorLog2wImpl(JSUword n); |
246 |
|
|
# endif |
247 |
|
|
|
248 |
|
|
#else |
249 |
|
|
|
250 |
|
|
# error "NOT SUPPORTED" |
251 |
|
|
|
252 |
|
|
#endif |
253 |
|
|
|
254 |
|
|
/* |
255 |
|
|
* Macros for rotate left. There is no rotate operation in the C Language so |
256 |
|
|
* the construct (a << 4) | (a >> 28) is used instead. Most compilers convert |
257 |
|
|
* this to a rotate instruction but some versions of MSVC don't without a |
258 |
|
|
* little help. To get MSVC to generate a rotate instruction, we have to use |
259 |
|
|
* the _rotl intrinsic and use a pragma to make _rotl inline. |
260 |
|
|
* |
261 |
|
|
* MSVC in VS2005 will do an inline rotate instruction on the above construct. |
262 |
|
|
*/ |
263 |
|
|
|
264 |
|
|
#if defined(_MSC_VER) && (defined(_M_IX86) || defined(_M_AMD64) || \ |
265 |
|
|
defined(_M_X64)) |
266 |
|
|
#include <stdlib.h> |
267 |
|
|
#pragma intrinsic(_rotl) |
268 |
|
|
#define JS_ROTATE_LEFT32(a, bits) _rotl(a, bits) |
269 |
|
|
#else |
270 |
|
|
#define JS_ROTATE_LEFT32(a, bits) (((a) << (bits)) | ((a) >> (32 - (bits)))) |
271 |
|
|
#endif |
272 |
|
|
|
273 |
|
|
JS_END_EXTERN_C |
274 |
|
|
#endif /* jsbit_h___ */ |