ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libecb/ecb.pod
Revision: 1.17
Committed: Thu May 26 23:55:58 2011 UTC (13 years ago) by root
Branch: MAIN
Changes since 1.16: +100 -0 lines
Log Message:
*** empty log message ***

File Contents

# Content
1 =head1 LIBECB - e-C-Builtins
2
3 =head2 ABOUT LIBECB
4
5 Libecb is currently a simple header file that doesn't require any
6 configuration to use or include in your project.
7
8 It's part of the e-suite of libraries, other members of which include
9 libev and libeio.
10
11 Its homepage can be found here:
12
13 http://software.schmorp.de/pkg/libecb
14
15 It mainly provides a number of wrappers around GCC built-ins, together
16 with replacement functions for other compilers. In addition to this,
17 it provides a number of other lowlevel C utilities, such as endianness
18 detection, byte swapping or bit rotations.
19
20 Or in other words, things that should be built-in into any standard C
21 system, but isn't.
22
23 More might come.
24
25 =head2 ABOUT THE HEADER
26
27 At the moment, all you have to do is copy F<ecb.h> somewhere where your
28 compiler can find it and include it:
29
30 #include <ecb.h>
31
32 The header should work fine for both C and C++ compilation, and gives you
33 all of F<inttypes.h> in addition to the ECB symbols.
34
35 There are currently no object files to link to - future versions might
36 come with an (optional) object code library to link against, to reduce
37 code size or gain access to additional features.
38
39 It also currently includes everything from F<inttypes.h>.
40
41 =head2 ABOUT THIS MANUAL / CONVENTIONS
42
43 This manual mainly describes each (public) function available after
44 including the F<ecb.h> header. The header might define other symbols than
45 these, but these are not part of the public API, and not supported in any
46 way.
47
48 When the manual mentions a "function" then this could be defined either as
49 as inline function, a macro, or an external symbol.
50
51 When functions use a concrete standard type, such as C<int> or
52 C<uint32_t>, then the corresponding function works only with that type. If
53 only a generic name is used (C<expr>, C<cond>, C<value> and so on), then
54 the corresponding function relies on C to implement the correct types, and
55 is usually implemented as a macro. Specifically, a "bool" in this manual
56 refers to any kind of boolean value, not a specific type.
57
58 =head2 GCC ATTRIBUTES
59
60 blabla where to put, what others
61
62 =over 4
63
64 =item ecb_attribute ((attrs...))
65
66 A simple wrapper that expands to C<__attribute__((attrs))> on GCC, and to
67 nothing on other compilers, so the effect is that only GCC sees these.
68
69 Example: use the C<deprecated> attribute on a function.
70
71 ecb_attribute((__deprecated__)) void
72 do_not_use_me_anymore (void);
73
74 =item ecb_unused
75
76 Marks a function or a variable as "unused", which simply suppresses a
77 warning by GCC when it detects it as unused. This is useful when you e.g.
78 declare a variable but do not always use it:
79
80 {
81 int var ecb_unused;
82
83 #ifdef SOMECONDITION
84 var = ...;
85 return var;
86 #else
87 return 0;
88 #endif
89 }
90
91 =item ecb_noinline
92
93 Prevent a function from being inlined - it might be optimised away, but
94 not inlined into other functions. This is useful if you know your function
95 is rarely called and large enough for inlining not to be helpful.
96
97 =item ecb_noreturn
98
99 Marks a function as "not returning, ever". Some typical functions that
100 don't return are C<exit> or C<abort> (which really works hard to not
101 return), and now you can make your own:
102
103 ecb_noreturn void
104 my_abort (const char *errline)
105 {
106 puts (errline);
107 abort ();
108 }
109
110 In this case, the compiler would probbaly be smart enough to decude it on
111 it's own, so this is mainly useful for declarations.
112
113 =item ecb_const
114
115 Declares that the function only depends on the values of it's arguments,
116 much like a mathematical function. It specifically does not read or write
117 any memory any arguments might point to, global variables, or call any
118 non-const functions. It also must not have any side effects.
119
120 Such a function can be optimised much more aggressively by the compiler -
121 for example, multiple calls with the same arguments can be optimised into
122 a single call, which wouldn't be possible if the compiler would have to
123 expect any side effects.
124
125 It is best suited for functions in the sense of mathematical functions,
126 such as a function return the square root of its input argument.
127
128 Not suited would be a function that calculates the hash of some memory
129 area you pass in, prints some messages or looks at a global variable to
130 decide on rounding.
131
132 See C<ecb_pure> for a slightly less restrictive class of functions.
133
134 =item ecb_pure
135
136 Similar to C<ecb_const>, declares a function that has no side
137 effects. Unlike C<ecb_const>, the function is allowed to examine global
138 variables and any other memory areas (such as the ones passed to it via
139 pointers).
140
141 While these functions cannot be optimised as aggressively as C<ecb_const>
142 functions, they can still be optimised away in many occasions, and the
143 compiler has more freedom in moving calls to them around.
144
145 Typical examples for such functions would be C<strlen> or C<memcmp>. A
146 function that calculates the MD5 sum of some input and updates some MD5
147 state passed as argument would I<NOT> be pure, however, as it would modify
148 some memory area that is not the return value.
149
150 =item ecb_hot
151
152 This declares a function as "hot" with regards to the cache - the function
153 is used so often, that it is very beneficial to keep it in the cache if
154 possible.
155
156 The compiler reacts by trying to place hot functions near to each other in
157 memory.
158
159 Whether a function is hot or not often depend son the whole program,
160 and less on the function itself. C<ecb_cold> is likely more useful in
161 practise.
162
163 =item ecb_cold
164
165 The opposite of C<ecb_hot> - declares a function as "cold" with regards to
166 the cache, or in other words, this function is not called often, or not at
167 speed-critical times, and keeping it in the cache might be a waste of said
168 cache.
169
170 In addition to placing cold functions together (or at least away from hot
171 functions), this knowledge can be used in other ways, for example, the
172 function will be optimised for size, as opposed to speed, and codepaths
173 leading to calls to those functions can automatically be marked as if
174 C<ecb_unlikel> had been used to reach them.
175
176 Good examples for such functions would be error reporting functions, or
177 functions only called in exceptional or rare cases.
178
179 =item ecb_artificial
180
181 Declares the function as "artificial", in this case meaning that this
182 function is not really mean to be a function, but more like an accessor
183 - many methods in C++ classes are mere accessor functions, and having a
184 crash reported in such a method, or single-stepping through them, is not
185 usually so helpful, especially when it's inlined to just a few instructions.
186
187 Marking them as artificial will instruct the debugger about just this,
188 leading to happier debugging and thus happier lives.
189
190 Example: in some kind of smart-pointer class, mark the pointer accessor as
191 artificial, so that the whole class acts more like a pointer and less like
192 some C++ abstraction monster.
193
194 template<typename T>
195 struct my_smart_ptr
196 {
197 T *value;
198
199 ecb_artificial
200 operator T *()
201 {
202 return value;
203 }
204 };
205
206 =back
207
208 =head2 OPTIMISATION HINTS
209
210 =over 4
211
212 =item bool ecb_is_constant(expr)
213
214 Returns true iff the expression can be deduced to be a compile-time
215 constant, and false otherwise.
216
217 For example, when you have a C<rndm16> function that returns a 16 bit
218 random number, and you have a function that maps this to a range from
219 0..n-1, then you could use this inline function in a header file:
220
221 ecb_inline uint32_t
222 rndm (uint32_t n)
223 {
224 return (n * (uint32_t)rndm16 ()) >> 16;
225 }
226
227 However, for powers of two, you could use a normal mask, but that is only
228 worth it if, at compile time, you can detect this case. This is the case
229 when the passed number is a constant and also a power of two (C<n & (n -
230 1) == 0>):
231
232 ecb_inline uint32_t
233 rndm (uint32_t n)
234 {
235 return is_constant (n) && !(n & (n - 1))
236 ? rndm16 () & (num - 1)
237 : (n * (uint32_t)rndm16 ()) >> 16;
238 }
239
240 =item bool ecb_expect (expr, value)
241
242 Evaluates C<expr> and returns it. In addition, it tells the compiler that
243 the C<expr> evaluates to C<value> a lot, which can be used for static
244 branch optimisations.
245
246 Usually, you want to use the more intuitive C<ecb_likely> and
247 C<ecb_unlikely> functions instead.
248
249 =item bool ecb_likely (cond)
250
251 =item bool ecb_unlikely (cond)
252
253 These two functions expect a expression that is true or false and return
254 C<1> or C<0>, respectively, so when used in the condition of an C<if> or
255 other conditional statement, it will not change the program:
256
257 /* these two do the same thing */
258 if (some_condition) ...;
259 if (ecb_likely (some_condition)) ...;
260
261 However, by using C<ecb_likely>, you tell the compiler that the condition
262 is likely to be true (and for C<ecb_unlikely>, that it is unlikely to be
263 true).
264
265 For example, when you check for a null pointer and expect this to be a
266 rare, exceptional, case, then use C<ecb_unlikely>:
267
268 void my_free (void *ptr)
269 {
270 if (ecb_unlikely (ptr == 0))
271 return;
272 }
273
274 Consequent use of these functions to mark away exceptional cases or to
275 tell the compiler what the hot path through a function is can increase
276 performance considerably.
277
278 A very good example is in a function that reserves more space for some
279 memory block (for example, inside an implementation of a string stream) -
280 each time something is added, you have to check for a buffer overrun, but
281 you expect that most checks will turn out to be false:
282
283 /* make sure we have "size" extra room in our buffer */
284 ecb_inline void
285 reserve (int size)
286 {
287 if (ecb_unlikely (current + size > end))
288 real_reserve_method (size); /* presumably noinline */
289 }
290
291 =item bool ecb_assume (cond)
292
293 Try to tell the compiler that some condition is true, even if it's not
294 obvious.
295
296 This can be used to teach the compiler about invariants or other
297 conditions that might improve code generation, but which are impossible to
298 deduce form the code itself.
299
300 For example, the example reservation function from the C<ecb_unlikely>
301 description could be written thus (only C<ecb_assume> was added):
302
303 ecb_inline void
304 reserve (int size)
305 {
306 if (ecb_unlikely (current + size > end))
307 real_reserve_method (size); /* presumably noinline */
308
309 ecb_assume (current + size <= end);
310 }
311
312 If you then call this function twice, like this:
313
314 reserve (10);
315 reserve (1);
316
317 Then the compiler I<might> be able to optimise out the second call
318 completely, as it knows that C<< current + 1 > end >> is false and the
319 call will never be executed.
320
321 =item bool ecb_unreachable ()
322
323 This function does nothing itself, except tell the compiler that it will
324 never be executed. Apart from suppressing a warning in some cases, this
325 function can be used to implement C<ecb_assume> or similar functions.
326
327 =item bool ecb_prefetch (addr, rw, locality)
328
329 Tells the compiler to try to prefetch memory at the given C<addr>ess
330 for either reading (C<rw> = 0) or writing (C<rw> = 1). A C<locality> of
331 C<0> means that there will only be one access later, C<3> means that
332 the data will likely be accessed very often, and values in between mean
333 something... in between. The memory pointed to by the address does not
334 need to be accessible (it could be a null pointer for example), but C<rw>
335 and C<locality> must be compile-time constants.
336
337 An obvious way to use this is to prefetch some data far away, in a big
338 array you loop over. This prefetches memory some 128 array elements later,
339 in the hope that it will be ready when the CPU arrives at that location.
340
341 int sum = 0;
342
343 for (i = 0; i < N; ++i)
344 {
345 sum += arr [i]
346 ecb_prefetch (arr + i + 128, 0, 0);
347 }
348
349 It's hard to predict how far to prefetch, and most CPUs that can prefetch
350 are often good enough to predict this kind of behaviour themselves. It
351 gets more interesting with linked lists, especially when you do some fair
352 processing on each list element:
353
354 for (node *n = start; n; n = n->next)
355 {
356 ecb_prefetch (n->next, 0, 0);
357 ... do medium amount of work with *n
358 }
359
360 After processing the node, (part of) the next node might already be in
361 cache.
362
363 =back
364
365 =head2 BIT FIDDLING / BITSTUFFS
366
367 =over 4
368
369 =item bool ecb_big_endian ()
370
371 =item bool ecb_little_endian ()
372
373 These two functions return true if the byte order is big endian
374 (most-significant byte first) or little endian (least-significant byte
375 first) respectively.
376
377 =item int ecb_ctz32 (uint32_t x)
378
379 Returns the index of the least significant bit set in C<x> (or
380 equivalently the number of bits set to 0 before the least significant
381 bit set), starting from 0. If C<x> is 0 the result is undefined. A
382 common use case is to compute the integer binary logarithm, i.e.,
383 floor(log2(n)). For example:
384
385 ecb_ctz32 (3) = 0
386 ecb_ctz32 (6) = 1
387
388 =item int ecb_popcount32 (uint32_t x)
389
390 Returns the number of bits set to 1 in C<x>. For example:
391
392 ecb_popcount32 (7) = 3
393 ecb_popcount32 (255) = 8
394
395 =item uint32_t ecb_bswap16 (uint32_t x)
396
397 =item uint32_t ecb_bswap32 (uint32_t x)
398
399 These two functions return the value of the 16-bit (32-bit) variable
400 C<x> after reversing the order of bytes.
401
402 =item uint32_t ecb_rotr32 (uint32_t x, unsigned int count)
403
404 =item uint32_t ecb_rotl32 (uint32_t x, unsigned int count)
405
406 These two functions return the value of C<x> after shifting all the bits
407 by C<count> positions to the right or left respectively.
408
409 =back
410
411 =head2 ARITHMETIC
412
413 =over 4
414
415 =item x = ecb_mod (m, n)
416
417 Returns the positive remainder of the modulo operation between C<m> and
418 C<n>. Unlike the C modulo operator C<%>, this function ensures that the
419 return value is always positive).
420
421 C<n> must be strictly positive (i.e. C<< >1 >>), while C<m> must be
422 negatable, that is, both C<m> and C<-m> must be representable in its
423 type.
424
425 =back
426
427 =head2 UTILITY
428
429 =over 4
430
431 =item element_count = ecb_array_length (name) [MACRO]
432
433 Returns the number of elements in the array C<name>. For example:
434
435 int primes[] = { 2, 3, 5, 7, 11 };
436 int sum = 0;
437
438 for (i = 0; i < ecb_array_length (primes); i++)
439 sum += primes [i];
440
441 =back
442
443