ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libecb/ecb.pod
Revision: 1.49
Committed: Thu Jun 28 14:50:01 2012 UTC (11 years, 11 months ago) by root
Branch: MAIN
Changes since 1.48: +3 -1 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 into any standard C system,
21 but aren't, implemented as efficient as possible with GCC, and still
22 correct with other compilers.
23
24 More might come.
25
26 =head2 ABOUT THE HEADER
27
28 At the moment, all you have to do is copy F<ecb.h> somewhere where your
29 compiler can find it and include it:
30
31 #include <ecb.h>
32
33 The header should work fine for both C and C++ compilation, and gives you
34 all of F<inttypes.h> in addition to the ECB symbols.
35
36 There are currently no object files to link to - future versions might
37 come with an (optional) object code library to link against, to reduce
38 code size or gain access to additional features.
39
40 It also currently includes everything from F<inttypes.h>.
41
42 =head2 ABOUT THIS MANUAL / CONVENTIONS
43
44 This manual mainly describes each (public) function available after
45 including the F<ecb.h> header. The header might define other symbols than
46 these, but these are not part of the public API, and not supported in any
47 way.
48
49 When the manual mentions a "function" then this could be defined either as
50 as inline function, a macro, or an external symbol.
51
52 When functions use a concrete standard type, such as C<int> or
53 C<uint32_t>, then the corresponding function works only with that type. If
54 only a generic name is used (C<expr>, C<cond>, C<value> and so on), then
55 the corresponding function relies on C to implement the correct types, and
56 is usually implemented as a macro. Specifically, a "bool" in this manual
57 refers to any kind of boolean value, not a specific type.
58
59 =head2 TYPES / TYPE SUPPORT
60
61 ecb.h makes sure that the following types are defined (in the expected way):
62
63 int8_t uint8_t int16_t uint16_t
64 int32_t uint32_t int64_t uint64_t
65 intptr_t uintptr_t
66
67 The macro C<ECB_PTRSIZE> is defined to the size of a pointer on this
68 platform (currently C<4> or C<8>) and can be used in preprocessor
69 expressions.
70
71 For C<ptrdiff_t> and C<size_t> use C<stddef.h>.
72
73 =head2 LANGUAGE/COMPILER VERSIONS
74
75 All the following symbols expand to an expression that can be tested in
76 preprocessor instructions as well as treated as a boolean (use C<!!> to
77 ensure it's either C<0> or C<1> if you need that).
78
79 =over 4
80
81 =item ECB_C
82
83 True if the implementation defines the C<__STDC__> macro to a true value,
84 which is typically true for both C and C++ compilers.
85
86 =item ECB_C99
87
88 True if the implementation claims to be compliant to C99 (ISO/IEC
89 9899:1999) or any later version.
90
91 Note that later versions (ECB_C11) remove core features again (for
92 example, variable length arrays).
93
94 =item ECB_C11
95
96 True if the implementation claims to be compliant to C11 (ISO/IEC
97 9899:2011) or any later version.
98
99 =item ECB_CPP
100
101 True if the implementation defines the C<__cplusplus__> macro to a true
102 value, which is typically true for C++ compilers.
103
104 =item ECB_CPP11
105
106 True if the implementation claims to be compliant to ISO/IEC 14882:2011
107 (C++11) or any later version.
108
109 =item ECB_GCC_VERSION(major,minor)
110
111 Expands to a true value (suitable for testing in by the preprocessor)
112 if the compiler used is GNU C and the version is the given version, or
113 higher.
114
115 This macro tries to return false on compilers that claim to be GCC
116 compatible but aren't.
117
118 =back
119
120 =head2 GCC ATTRIBUTES
121
122 A major part of libecb deals with GCC attributes. These are additional
123 attributes that you can assign to functions, variables and sometimes even
124 types - much like C<const> or C<volatile> in C.
125
126 While GCC allows declarations to show up in many surprising places,
127 but not in many expected places, the safest way is to put attribute
128 declarations before the whole declaration:
129
130 ecb_const int mysqrt (int a);
131 ecb_unused int i;
132
133 For variables, it is often nicer to put the attribute after the name, and
134 avoid multiple declarations using commas:
135
136 int i ecb_unused;
137
138 =over 4
139
140 =item ecb_attribute ((attrs...))
141
142 A simple wrapper that expands to C<__attribute__((attrs))> on GCC, and to
143 nothing on other compilers, so the effect is that only GCC sees these.
144
145 Example: use the C<deprecated> attribute on a function.
146
147 ecb_attribute((__deprecated__)) void
148 do_not_use_me_anymore (void);
149
150 =item ecb_unused
151
152 Marks a function or a variable as "unused", which simply suppresses a
153 warning by GCC when it detects it as unused. This is useful when you e.g.
154 declare a variable but do not always use it:
155
156 {
157 int var ecb_unused;
158
159 #ifdef SOMECONDITION
160 var = ...;
161 return var;
162 #else
163 return 0;
164 #endif
165 }
166
167 =item ecb_inline
168
169 This is not actually an attribute, but you use it like one. It expands
170 either to C<static inline> or to just C<static>, if inline isn't
171 supported. It should be used to declare functions that should be inlined,
172 for code size or speed reasons.
173
174 Example: inline this function, it surely will reduce codesize.
175
176 ecb_inline int
177 negmul (int a, int b)
178 {
179 return - (a * b);
180 }
181
182 =item ecb_noinline
183
184 Prevent a function from being inlined - it might be optimised away, but
185 not inlined into other functions. This is useful if you know your function
186 is rarely called and large enough for inlining not to be helpful.
187
188 =item ecb_noreturn
189
190 Marks a function as "not returning, ever". Some typical functions that
191 don't return are C<exit> or C<abort> (which really works hard to not
192 return), and now you can make your own:
193
194 ecb_noreturn void
195 my_abort (const char *errline)
196 {
197 puts (errline);
198 abort ();
199 }
200
201 In this case, the compiler would probably be smart enough to deduce it on
202 its own, so this is mainly useful for declarations.
203
204 =item ecb_const
205
206 Declares that the function only depends on the values of its arguments,
207 much like a mathematical function. It specifically does not read or write
208 any memory any arguments might point to, global variables, or call any
209 non-const functions. It also must not have any side effects.
210
211 Such a function can be optimised much more aggressively by the compiler -
212 for example, multiple calls with the same arguments can be optimised into
213 a single call, which wouldn't be possible if the compiler would have to
214 expect any side effects.
215
216 It is best suited for functions in the sense of mathematical functions,
217 such as a function returning the square root of its input argument.
218
219 Not suited would be a function that calculates the hash of some memory
220 area you pass in, prints some messages or looks at a global variable to
221 decide on rounding.
222
223 See C<ecb_pure> for a slightly less restrictive class of functions.
224
225 =item ecb_pure
226
227 Similar to C<ecb_const>, declares a function that has no side
228 effects. Unlike C<ecb_const>, the function is allowed to examine global
229 variables and any other memory areas (such as the ones passed to it via
230 pointers).
231
232 While these functions cannot be optimised as aggressively as C<ecb_const>
233 functions, they can still be optimised away in many occasions, and the
234 compiler has more freedom in moving calls to them around.
235
236 Typical examples for such functions would be C<strlen> or C<memcmp>. A
237 function that calculates the MD5 sum of some input and updates some MD5
238 state passed as argument would I<NOT> be pure, however, as it would modify
239 some memory area that is not the return value.
240
241 =item ecb_hot
242
243 This declares a function as "hot" with regards to the cache - the function
244 is used so often, that it is very beneficial to keep it in the cache if
245 possible.
246
247 The compiler reacts by trying to place hot functions near to each other in
248 memory.
249
250 Whether a function is hot or not often depends on the whole program,
251 and less on the function itself. C<ecb_cold> is likely more useful in
252 practise.
253
254 =item ecb_cold
255
256 The opposite of C<ecb_hot> - declares a function as "cold" with regards to
257 the cache, or in other words, this function is not called often, or not at
258 speed-critical times, and keeping it in the cache might be a waste of said
259 cache.
260
261 In addition to placing cold functions together (or at least away from hot
262 functions), this knowledge can be used in other ways, for example, the
263 function will be optimised for size, as opposed to speed, and codepaths
264 leading to calls to those functions can automatically be marked as if
265 C<ecb_expect_false> had been used to reach them.
266
267 Good examples for such functions would be error reporting functions, or
268 functions only called in exceptional or rare cases.
269
270 =item ecb_artificial
271
272 Declares the function as "artificial", in this case meaning that this
273 function is not really mean to be a function, but more like an accessor
274 - many methods in C++ classes are mere accessor functions, and having a
275 crash reported in such a method, or single-stepping through them, is not
276 usually so helpful, especially when it's inlined to just a few instructions.
277
278 Marking them as artificial will instruct the debugger about just this,
279 leading to happier debugging and thus happier lives.
280
281 Example: in some kind of smart-pointer class, mark the pointer accessor as
282 artificial, so that the whole class acts more like a pointer and less like
283 some C++ abstraction monster.
284
285 template<typename T>
286 struct my_smart_ptr
287 {
288 T *value;
289
290 ecb_artificial
291 operator T *()
292 {
293 return value;
294 }
295 };
296
297 =back
298
299 =head2 OPTIMISATION HINTS
300
301 =over 4
302
303 =item bool ecb_is_constant(expr)
304
305 Returns true iff the expression can be deduced to be a compile-time
306 constant, and false otherwise.
307
308 For example, when you have a C<rndm16> function that returns a 16 bit
309 random number, and you have a function that maps this to a range from
310 0..n-1, then you could use this inline function in a header file:
311
312 ecb_inline uint32_t
313 rndm (uint32_t n)
314 {
315 return (n * (uint32_t)rndm16 ()) >> 16;
316 }
317
318 However, for powers of two, you could use a normal mask, but that is only
319 worth it if, at compile time, you can detect this case. This is the case
320 when the passed number is a constant and also a power of two (C<n & (n -
321 1) == 0>):
322
323 ecb_inline uint32_t
324 rndm (uint32_t n)
325 {
326 return is_constant (n) && !(n & (n - 1))
327 ? rndm16 () & (num - 1)
328 : (n * (uint32_t)rndm16 ()) >> 16;
329 }
330
331 =item bool ecb_expect (expr, value)
332
333 Evaluates C<expr> and returns it. In addition, it tells the compiler that
334 the C<expr> evaluates to C<value> a lot, which can be used for static
335 branch optimisations.
336
337 Usually, you want to use the more intuitive C<ecb_expect_true> and
338 C<ecb_expect_false> functions instead.
339
340 =item bool ecb_expect_true (cond)
341
342 =item bool ecb_expect_false (cond)
343
344 These two functions expect a expression that is true or false and return
345 C<1> or C<0>, respectively, so when used in the condition of an C<if> or
346 other conditional statement, it will not change the program:
347
348 /* these two do the same thing */
349 if (some_condition) ...;
350 if (ecb_expect_true (some_condition)) ...;
351
352 However, by using C<ecb_expect_true>, you tell the compiler that the
353 condition is likely to be true (and for C<ecb_expect_false>, that it is
354 unlikely to be true).
355
356 For example, when you check for a null pointer and expect this to be a
357 rare, exceptional, case, then use C<ecb_expect_false>:
358
359 void my_free (void *ptr)
360 {
361 if (ecb_expect_false (ptr == 0))
362 return;
363 }
364
365 Consequent use of these functions to mark away exceptional cases or to
366 tell the compiler what the hot path through a function is can increase
367 performance considerably.
368
369 You might know these functions under the name C<likely> and C<unlikely>
370 - while these are common aliases, we find that the expect name is easier
371 to understand when quickly skimming code. If you wish, you can use
372 C<ecb_likely> instead of C<ecb_expect_true> and C<ecb_unlikely> instead of
373 C<ecb_expect_false> - these are simply aliases.
374
375 A very good example is in a function that reserves more space for some
376 memory block (for example, inside an implementation of a string stream) -
377 each time something is added, you have to check for a buffer overrun, but
378 you expect that most checks will turn out to be false:
379
380 /* make sure we have "size" extra room in our buffer */
381 ecb_inline void
382 reserve (int size)
383 {
384 if (ecb_expect_false (current + size > end))
385 real_reserve_method (size); /* presumably noinline */
386 }
387
388 =item bool ecb_assume (cond)
389
390 Try to tell the compiler that some condition is true, even if it's not
391 obvious.
392
393 This can be used to teach the compiler about invariants or other
394 conditions that might improve code generation, but which are impossible to
395 deduce form the code itself.
396
397 For example, the example reservation function from the C<ecb_expect_false>
398 description could be written thus (only C<ecb_assume> was added):
399
400 ecb_inline void
401 reserve (int size)
402 {
403 if (ecb_expect_false (current + size > end))
404 real_reserve_method (size); /* presumably noinline */
405
406 ecb_assume (current + size <= end);
407 }
408
409 If you then call this function twice, like this:
410
411 reserve (10);
412 reserve (1);
413
414 Then the compiler I<might> be able to optimise out the second call
415 completely, as it knows that C<< current + 1 > end >> is false and the
416 call will never be executed.
417
418 =item bool ecb_unreachable ()
419
420 This function does nothing itself, except tell the compiler that it will
421 never be executed. Apart from suppressing a warning in some cases, this
422 function can be used to implement C<ecb_assume> or similar functions.
423
424 =item bool ecb_prefetch (addr, rw, locality)
425
426 Tells the compiler to try to prefetch memory at the given C<addr>ess
427 for either reading (C<rw> = 0) or writing (C<rw> = 1). A C<locality> of
428 C<0> means that there will only be one access later, C<3> means that
429 the data will likely be accessed very often, and values in between mean
430 something... in between. The memory pointed to by the address does not
431 need to be accessible (it could be a null pointer for example), but C<rw>
432 and C<locality> must be compile-time constants.
433
434 An obvious way to use this is to prefetch some data far away, in a big
435 array you loop over. This prefetches memory some 128 array elements later,
436 in the hope that it will be ready when the CPU arrives at that location.
437
438 int sum = 0;
439
440 for (i = 0; i < N; ++i)
441 {
442 sum += arr [i]
443 ecb_prefetch (arr + i + 128, 0, 0);
444 }
445
446 It's hard to predict how far to prefetch, and most CPUs that can prefetch
447 are often good enough to predict this kind of behaviour themselves. It
448 gets more interesting with linked lists, especially when you do some fair
449 processing on each list element:
450
451 for (node *n = start; n; n = n->next)
452 {
453 ecb_prefetch (n->next, 0, 0);
454 ... do medium amount of work with *n
455 }
456
457 After processing the node, (part of) the next node might already be in
458 cache.
459
460 =back
461
462 =head2 BIT FIDDLING / BIT WIZARDRY
463
464 =over 4
465
466 =item bool ecb_big_endian ()
467
468 =item bool ecb_little_endian ()
469
470 These two functions return true if the byte order is big endian
471 (most-significant byte first) or little endian (least-significant byte
472 first) respectively.
473
474 On systems that are neither, their return values are unspecified.
475
476 =item int ecb_ctz32 (uint32_t x)
477
478 =item int ecb_ctz64 (uint64_t x)
479
480 Returns the index of the least significant bit set in C<x> (or
481 equivalently the number of bits set to 0 before the least significant bit
482 set), starting from 0. If C<x> is 0 the result is undefined.
483
484 For smaller types than C<uint32_t> you can safely use C<ecb_ctz32>.
485
486 For example:
487
488 ecb_ctz32 (3) = 0
489 ecb_ctz32 (6) = 1
490
491 =item bool ecb_is_pot32 (uint32_t x)
492
493 =item bool ecb_is_pot64 (uint32_t x)
494
495 Return true iff C<x> is a power of two or C<x == 0>.
496
497 For smaller types then C<uint32_t> you can safely use C<ecb_is_pot32>.
498
499 =item int ecb_ld32 (uint32_t x)
500
501 =item int ecb_ld64 (uint64_t x)
502
503 Returns the index of the most significant bit set in C<x>, or the number
504 of digits the number requires in binary (so that C<< 2**ld <= x <
505 2**(ld+1) >>). If C<x> is 0 the result is undefined. A common use case is
506 to compute the integer binary logarithm, i.e. C<floor (log2 (n))>, for
507 example to see how many bits a certain number requires to be encoded.
508
509 This function is similar to the "count leading zero bits" function, except
510 that that one returns how many zero bits are "in front" of the number (in
511 the given data type), while C<ecb_ld> returns how many bits the number
512 itself requires.
513
514 For smaller types than C<uint32_t> you can safely use C<ecb_ld32>.
515
516 =item int ecb_popcount32 (uint32_t x)
517
518 =item int ecb_popcount64 (uint64_t x)
519
520 Returns the number of bits set to 1 in C<x>.
521
522 For smaller types than C<uint32_t> you can safely use C<ecb_popcount32>.
523
524 For example:
525
526 ecb_popcount32 (7) = 3
527 ecb_popcount32 (255) = 8
528
529 =item uint8_t ecb_bitrev8 (uint8_t x)
530
531 =item uint16_t ecb_bitrev16 (uint16_t x)
532
533 =item uint32_t ecb_bitrev32 (uint32_t x)
534
535 Reverses the bits in x, i.e. the MSB becomes the LSB, MSB-1 becomes LSB+1
536 and so on.
537
538 Example:
539
540 ecb_bitrev8 (0xa7) = 0xea
541 ecb_bitrev32 (0xffcc4411) = 0x882233ff
542
543 =item uint32_t ecb_bswap16 (uint32_t x)
544
545 =item uint32_t ecb_bswap32 (uint32_t x)
546
547 =item uint64_t ecb_bswap64 (uint64_t x)
548
549 These functions return the value of the 16-bit (32-bit, 64-bit) value
550 C<x> after reversing the order of bytes (0x11223344 becomes 0x44332211 in
551 C<ecb_bswap32>).
552
553 =item uint8_t ecb_rotl8 (uint8_t x, unsigned int count)
554
555 =item uint16_t ecb_rotl16 (uint16_t x, unsigned int count)
556
557 =item uint32_t ecb_rotl32 (uint32_t x, unsigned int count)
558
559 =item uint64_t ecb_rotl64 (uint64_t x, unsigned int count)
560
561 =item uint8_t ecb_rotr8 (uint8_t x, unsigned int count)
562
563 =item uint16_t ecb_rotr16 (uint16_t x, unsigned int count)
564
565 =item uint32_t ecb_rotr32 (uint32_t x, unsigned int count)
566
567 =item uint64_t ecb_rotr64 (uint64_t x, unsigned int count)
568
569 These two families of functions return the value of C<x> after rotating
570 all the bits by C<count> positions to the right (C<ecb_rotr>) or left
571 (C<ecb_rotl>).
572
573 Current GCC versions understand these functions and usually compile them
574 to "optimal" code (e.g. a single C<rol> or a combination of C<shld> on
575 x86).
576
577 =back
578
579 =head2 ARITHMETIC
580
581 =over 4
582
583 =item x = ecb_mod (m, n)
584
585 Returns C<m> modulo C<n>, which is the same as the positive remainder
586 of the division operation between C<m> and C<n>, using floored
587 division. Unlike the C remainder operator C<%>, this function ensures that
588 the return value is always positive and that the two numbers I<m> and
589 I<m' = m + i * n> result in the same value modulo I<n> - in other words,
590 C<ecb_mod> implements the mathematical modulo operation, which is missing
591 in the language.
592
593 C<n> must be strictly positive (i.e. C<< >= 1 >>), while C<m> must be
594 negatable, that is, both C<m> and C<-m> must be representable in its
595 type (this typically excludes the minimum signed integer value, the same
596 limitation as for C</> and C<%> in C).
597
598 Current GCC versions compile this into an efficient branchless sequence on
599 almost all CPUs.
600
601 For example, when you want to rotate forward through the members of an
602 array for increasing C<m> (which might be negative), then you should use
603 C<ecb_mod>, as the C<%> operator might give either negative results, or
604 change direction for negative values:
605
606 for (m = -100; m <= 100; ++m)
607 int elem = myarray [ecb_mod (m, ecb_array_length (myarray))];
608
609 =item x = ecb_div_rd (val, div)
610
611 =item x = ecb_div_ru (val, div)
612
613 Returns C<val> divided by C<div> rounded down or up, respectively.
614 C<val> and C<div> must have integer types and C<div> must be strictly
615 positive. Note that these functions are implemented with macros in C
616 and with function templates in C++.
617
618 =back
619
620 =head2 UTILITY
621
622 =over 4
623
624 =item element_count = ecb_array_length (name)
625
626 Returns the number of elements in the array C<name>. For example:
627
628 int primes[] = { 2, 3, 5, 7, 11 };
629 int sum = 0;
630
631 for (i = 0; i < ecb_array_length (primes); i++)
632 sum += primes [i];
633
634 =back
635
636 =head2 SYMBOLS GOVERNING COMPILATION OF ECB.H ITSELF
637
638 These symbols need to be defined before including F<ecb.h> the first time.
639
640 =over 4
641
642 =item ECB_NO_THRADS
643
644 If F<ecb.h> is never used from multiple threads, then this symbol can
645 be defined, in which case memory fences (and similar constructs) are
646 completely removed, leading to more efficient code and fewer dependencies.
647
648 Setting this symbol to a true value implies C<ECB_NO_SMP>.
649
650 =item ECB_NO_SMP
651
652 The weaker version of C<ECB_NO_THREADS> - if F<ecb.h> is used from
653 multiple threads, but never concurrently (e.g. if the system the program
654 runs on has only a single CPU with a single core, no hyperthreading and so
655 on), then this symbol can be defined, leading to more efficient code and
656 fewer dependencies.
657
658 =back
659
660