ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libecb/ecb.pod
Revision: 1.64
Committed: Wed Feb 18 20:48:59 2015 UTC (9 years, 3 months ago) by root
Branch: MAIN
Changes since 1.63: +16 -0 lines
Log Message:
*** empty log message ***

File Contents

# User Rev Content
1 root 1.14 =head1 LIBECB - e-C-Builtins
2 root 1.3
3 root 1.14 =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 sf-exg 1.16 It's part of the e-suite of libraries, other members of which include
9 root 1.14 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 sf-exg 1.16 it provides a number of other lowlevel C utilities, such as endianness
18 root 1.14 detection, byte swapping or bit rotations.
19    
20 root 1.24 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 root 1.17
24 root 1.14 More might come.
25 root 1.3
26     =head2 ABOUT THE HEADER
27    
28 root 1.14 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 sf-exg 1.16 There are currently no object files to link to - future versions might
37 root 1.14 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 root 1.1
59 root 1.40 =head2 TYPES / TYPE SUPPORT
60    
61     ecb.h makes sure that the following types are defined (in the expected way):
62    
63 root 1.42 int8_t uint8_t int16_t uint16_t
64     int32_t uint32_t int64_t uint64_t
65 root 1.49 intptr_t uintptr_t
66 root 1.40
67     The macro C<ECB_PTRSIZE> is defined to the size of a pointer on this
68 root 1.45 platform (currently C<4> or C<8>) and can be used in preprocessor
69     expressions.
70 root 1.40
71 root 1.49 For C<ptrdiff_t> and C<size_t> use C<stddef.h>.
72    
73 root 1.62 =head2 LANGUAGE/ENVIRONMENT/COMPILER VERSIONS
74 root 1.43
75 sf-exg 1.46 All the following symbols expand to an expression that can be tested in
76 root 1.44 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 root 1.43 =over 4
80    
81 root 1.44 =item ECB_C
82    
83     True if the implementation defines the C<__STDC__> macro to a true value,
84 root 1.55 while not claiming to be C++.
85 root 1.44
86 root 1.43 =item ECB_C99
87    
88 root 1.47 True if the implementation claims to be compliant to C99 (ISO/IEC
89 root 1.55 9899:1999) or any later version, while not claiming to be C++.
90 root 1.47
91     Note that later versions (ECB_C11) remove core features again (for
92     example, variable length arrays).
93 root 1.43
94     =item ECB_C11
95    
96 root 1.47 True if the implementation claims to be compliant to C11 (ISO/IEC
97 root 1.55 9899:2011) or any later version, while not claiming to be C++.
98 root 1.44
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 sf-exg 1.46 (C++11) or any later version.
108 root 1.43
109 root 1.57 =item ECB_GCC_VERSION (major, minor)
110 root 1.43
111     Expands to a true value (suitable for testing in by the preprocessor)
112 sf-exg 1.46 if the compiler used is GNU C and the version is the given version, or
113 root 1.43 higher.
114    
115     This macro tries to return false on compilers that claim to be GCC
116     compatible but aren't.
117    
118 root 1.50 =item ECB_EXTERN_C
119    
120     Expands to C<extern "C"> in C++, and a simple C<extern> in C.
121    
122     This can be used to declare a single external C function:
123    
124     ECB_EXTERN_C int printf (const char *format, ...);
125    
126     =item ECB_EXTERN_C_BEG / ECB_EXTERN_C_END
127    
128     These two macros can be used to wrap multiple C<extern "C"> definitions -
129     they expand to nothing in C.
130    
131     They are most useful in header files:
132    
133     ECB_EXTERN_C_BEG
134    
135     int mycfun1 (int x);
136     int mycfun2 (int x);
137    
138     ECB_EXTERN_C_END
139    
140     =item ECB_STDFP
141    
142     If this evaluates to a true value (suitable for testing in by the
143     preprocessor), then C<float> and C<double> use IEEE 754 single/binary32
144     and double/binary64 representations internally I<and> the endianness of
145     both types match the endianness of C<uint32_t> and C<uint64_t>.
146    
147     This means you can just copy the bits of a C<float> (or C<double>) to an
148     C<uint32_t> (or C<uint64_t>) and get the raw IEEE 754 bit representation
149     without having to think about format or endianness.
150    
151     This is true for basically all modern platforms, although F<ecb.h> might
152     not be able to deduce this correctly everywhere and might err on the safe
153     side.
154    
155 root 1.54 =item ECB_AMD64, ECB_AMD64_X32
156    
157     These two macros are defined to C<1> on the x86_64/amd64 ABI and the X32
158     ABI, respectively, and undefined elsewhere.
159    
160     The designers of the new X32 ABI for some inexplicable reason decided to
161     make it look exactly like amd64, even though it's completely incompatible
162     to that ABI, breaking about every piece of software that assumed that
163     C<__x86_64> stands for, well, the x86-64 ABI, making these macros
164     necessary.
165    
166 root 1.43 =back
167    
168 root 1.62 =head2 MACRO TRICKERY
169    
170     =over 4
171    
172     =item ECB_CONCAT (a, b)
173    
174     Expands any macros in C<a> and C<b>, then concatenates the result to form
175     a single token. This is mainly useful to form identifiers from components,
176     e.g.:
177    
178     #define S1 str
179     #define S2 cpy
180    
181     ECB_CONCAT (S1, S2)(dst, src); // == strcpy (dst, src);
182    
183     =item ECB_STRINGIFY (arg)
184    
185     Expands any macros in C<arg> and returns the stringified version of
186     it. This is mainly useful to get the contents of a macro in string form,
187     e.g.:
188    
189     #define SQL_LIMIT 100
190     sql_exec ("select * from table limit " ECB_STRINGIFY (SQL_LIMIT));
191    
192 root 1.64 =item ECB_STRINGIFY_EXPR (expr)
193    
194     Like C<ECB_STRINGIFY>, but additionally evaluates C<expr> to make sure it
195     is a valid expression. This is useful to catch typos or cases where the
196     macro isn't available:
197    
198     #include <errno.h>
199    
200     ECB_STRINGIFY (EDOM); // "33" (on my system at least)
201     ECB_STRINGIFY_EXPR (EDOM); // "33"
202    
203     // now imagine we had a typo:
204    
205     ECB_STRINGIFY (EDAM); // "EDAM"
206     ECB_STRINGIFY_EXPR (EDAM); // error: EDAM undefined
207    
208 root 1.62 =back
209    
210 sf-exg 1.60 =head2 ATTRIBUTES
211 root 1.1
212 sf-exg 1.60 A major part of libecb deals with additional attributes that can be
213     assigned to functions, variables and sometimes even types - much like
214     C<const> or C<volatile> in C. They are implemented using either GCC
215     attributes or other compiler/language specific features. Attributes
216     declarations must be put before the whole declaration:
217 root 1.20
218     ecb_const int mysqrt (int a);
219     ecb_unused int i;
220    
221 root 1.1 =over 4
222    
223 root 1.3 =item ecb_unused
224    
225     Marks a function or a variable as "unused", which simply suppresses a
226     warning by GCC when it detects it as unused. This is useful when you e.g.
227     declare a variable but do not always use it:
228    
229 root 1.15 {
230 sf-exg 1.61 ecb_unused int var;
231 root 1.3
232 root 1.15 #ifdef SOMECONDITION
233     var = ...;
234     return var;
235     #else
236     return 0;
237     #endif
238     }
239 root 1.3
240 root 1.56 =item ecb_deprecated
241    
242     Similar to C<ecb_unused>, but marks a function, variable or type as
243     deprecated. This makes some compilers warn when the type is used.
244    
245 root 1.62 =item ecb_deprecated_message (message)
246    
247     Same as C<ecb_deprecated>, but if possible, supply a diagnostic that is
248     used instead of a generic depreciation message when the object is being
249     used.
250    
251 root 1.31 =item ecb_inline
252 root 1.29
253 sf-exg 1.60 Expands either to C<static inline> or to just C<static>, if inline
254     isn't supported. It should be used to declare functions that should be
255     inlined, for code size or speed reasons.
256 root 1.29
257     Example: inline this function, it surely will reduce codesize.
258    
259 root 1.31 ecb_inline int
260 root 1.29 negmul (int a, int b)
261     {
262     return - (a * b);
263     }
264    
265 root 1.2 =item ecb_noinline
266    
267 root 1.9 Prevent a function from being inlined - it might be optimised away, but
268 root 1.3 not inlined into other functions. This is useful if you know your function
269     is rarely called and large enough for inlining not to be helpful.
270    
271 root 1.2 =item ecb_noreturn
272    
273 root 1.17 Marks a function as "not returning, ever". Some typical functions that
274     don't return are C<exit> or C<abort> (which really works hard to not
275     return), and now you can make your own:
276    
277     ecb_noreturn void
278     my_abort (const char *errline)
279     {
280     puts (errline);
281     abort ();
282     }
283    
284 sf-exg 1.19 In this case, the compiler would probably be smart enough to deduce it on
285     its own, so this is mainly useful for declarations.
286 root 1.17
287 root 1.53 =item ecb_restrict
288    
289     Expands to the C<restrict> keyword or equivalent on compilers that support
290     them, and to nothing on others. Must be specified on a pointer type or
291     an array index to indicate that the memory doesn't alias with any other
292     restricted pointer in the same scope.
293    
294     Example: multiply a vector, and allow the compiler to parallelise the
295     loop, because it knows it doesn't overwrite input values.
296    
297     void
298 sf-exg 1.61 multiply (ecb_restrict float *src,
299     ecb_restrict float *dst,
300 root 1.53 int len, float factor)
301     {
302     int i;
303    
304     for (i = 0; i < len; ++i)
305     dst [i] = src [i] * factor;
306     }
307    
308 root 1.2 =item ecb_const
309    
310 sf-exg 1.19 Declares that the function only depends on the values of its arguments,
311 root 1.17 much like a mathematical function. It specifically does not read or write
312     any memory any arguments might point to, global variables, or call any
313     non-const functions. It also must not have any side effects.
314    
315     Such a function can be optimised much more aggressively by the compiler -
316     for example, multiple calls with the same arguments can be optimised into
317     a single call, which wouldn't be possible if the compiler would have to
318     expect any side effects.
319    
320     It is best suited for functions in the sense of mathematical functions,
321 sf-exg 1.19 such as a function returning the square root of its input argument.
322 root 1.17
323     Not suited would be a function that calculates the hash of some memory
324     area you pass in, prints some messages or looks at a global variable to
325     decide on rounding.
326    
327     See C<ecb_pure> for a slightly less restrictive class of functions.
328    
329 root 1.2 =item ecb_pure
330    
331 root 1.17 Similar to C<ecb_const>, declares a function that has no side
332     effects. Unlike C<ecb_const>, the function is allowed to examine global
333     variables and any other memory areas (such as the ones passed to it via
334     pointers).
335    
336     While these functions cannot be optimised as aggressively as C<ecb_const>
337     functions, they can still be optimised away in many occasions, and the
338     compiler has more freedom in moving calls to them around.
339    
340     Typical examples for such functions would be C<strlen> or C<memcmp>. A
341     function that calculates the MD5 sum of some input and updates some MD5
342     state passed as argument would I<NOT> be pure, however, as it would modify
343     some memory area that is not the return value.
344    
345 root 1.2 =item ecb_hot
346    
347 root 1.17 This declares a function as "hot" with regards to the cache - the function
348     is used so often, that it is very beneficial to keep it in the cache if
349     possible.
350    
351     The compiler reacts by trying to place hot functions near to each other in
352     memory.
353    
354 sf-exg 1.19 Whether a function is hot or not often depends on the whole program,
355 root 1.17 and less on the function itself. C<ecb_cold> is likely more useful in
356     practise.
357    
358 root 1.2 =item ecb_cold
359    
360 root 1.17 The opposite of C<ecb_hot> - declares a function as "cold" with regards to
361     the cache, or in other words, this function is not called often, or not at
362     speed-critical times, and keeping it in the cache might be a waste of said
363     cache.
364    
365     In addition to placing cold functions together (or at least away from hot
366     functions), this knowledge can be used in other ways, for example, the
367     function will be optimised for size, as opposed to speed, and codepaths
368     leading to calls to those functions can automatically be marked as if
369 root 1.27 C<ecb_expect_false> had been used to reach them.
370 root 1.17
371     Good examples for such functions would be error reporting functions, or
372     functions only called in exceptional or rare cases.
373    
374 root 1.2 =item ecb_artificial
375    
376 root 1.17 Declares the function as "artificial", in this case meaning that this
377 root 1.52 function is not really meant to be a function, but more like an accessor
378 root 1.17 - many methods in C++ classes are mere accessor functions, and having a
379     crash reported in such a method, or single-stepping through them, is not
380     usually so helpful, especially when it's inlined to just a few instructions.
381    
382     Marking them as artificial will instruct the debugger about just this,
383     leading to happier debugging and thus happier lives.
384    
385     Example: in some kind of smart-pointer class, mark the pointer accessor as
386     artificial, so that the whole class acts more like a pointer and less like
387     some C++ abstraction monster.
388    
389     template<typename T>
390     struct my_smart_ptr
391     {
392     T *value;
393    
394     ecb_artificial
395     operator T *()
396     {
397     return value;
398     }
399     };
400    
401 root 1.2 =back
402 root 1.1
403     =head2 OPTIMISATION HINTS
404    
405     =over 4
406    
407 root 1.58 =item bool ecb_is_constant (expr)
408 root 1.1
409 root 1.3 Returns true iff the expression can be deduced to be a compile-time
410     constant, and false otherwise.
411    
412     For example, when you have a C<rndm16> function that returns a 16 bit
413     random number, and you have a function that maps this to a range from
414 root 1.5 0..n-1, then you could use this inline function in a header file:
415 root 1.3
416     ecb_inline uint32_t
417     rndm (uint32_t n)
418     {
419 root 1.6 return (n * (uint32_t)rndm16 ()) >> 16;
420 root 1.3 }
421    
422     However, for powers of two, you could use a normal mask, but that is only
423     worth it if, at compile time, you can detect this case. This is the case
424     when the passed number is a constant and also a power of two (C<n & (n -
425     1) == 0>):
426    
427     ecb_inline uint32_t
428     rndm (uint32_t n)
429     {
430     return is_constant (n) && !(n & (n - 1))
431     ? rndm16 () & (num - 1)
432 root 1.6 : (n * (uint32_t)rndm16 ()) >> 16;
433 root 1.3 }
434    
435 root 1.62 =item ecb_expect (expr, value)
436 root 1.1
437 root 1.7 Evaluates C<expr> and returns it. In addition, it tells the compiler that
438     the C<expr> evaluates to C<value> a lot, which can be used for static
439     branch optimisations.
440 root 1.1
441 root 1.27 Usually, you want to use the more intuitive C<ecb_expect_true> and
442     C<ecb_expect_false> functions instead.
443 root 1.1
444 root 1.27 =item bool ecb_expect_true (cond)
445 root 1.1
446 root 1.27 =item bool ecb_expect_false (cond)
447 root 1.1
448 root 1.7 These two functions expect a expression that is true or false and return
449     C<1> or C<0>, respectively, so when used in the condition of an C<if> or
450     other conditional statement, it will not change the program:
451    
452     /* these two do the same thing */
453     if (some_condition) ...;
454 root 1.27 if (ecb_expect_true (some_condition)) ...;
455 root 1.7
456 root 1.27 However, by using C<ecb_expect_true>, you tell the compiler that the
457     condition is likely to be true (and for C<ecb_expect_false>, that it is
458     unlikely to be true).
459 root 1.7
460 root 1.9 For example, when you check for a null pointer and expect this to be a
461 root 1.27 rare, exceptional, case, then use C<ecb_expect_false>:
462 root 1.7
463     void my_free (void *ptr)
464     {
465 root 1.27 if (ecb_expect_false (ptr == 0))
466 root 1.7 return;
467     }
468    
469     Consequent use of these functions to mark away exceptional cases or to
470     tell the compiler what the hot path through a function is can increase
471     performance considerably.
472    
473 root 1.27 You might know these functions under the name C<likely> and C<unlikely>
474     - while these are common aliases, we find that the expect name is easier
475     to understand when quickly skimming code. If you wish, you can use
476     C<ecb_likely> instead of C<ecb_expect_true> and C<ecb_unlikely> instead of
477     C<ecb_expect_false> - these are simply aliases.
478    
479 root 1.7 A very good example is in a function that reserves more space for some
480     memory block (for example, inside an implementation of a string stream) -
481 root 1.9 each time something is added, you have to check for a buffer overrun, but
482 root 1.7 you expect that most checks will turn out to be false:
483    
484     /* make sure we have "size" extra room in our buffer */
485     ecb_inline void
486     reserve (int size)
487     {
488 root 1.27 if (ecb_expect_false (current + size > end))
489 root 1.7 real_reserve_method (size); /* presumably noinline */
490     }
491    
492 root 1.62 =item ecb_assume (cond)
493 root 1.7
494     Try to tell the compiler that some condition is true, even if it's not
495     obvious.
496    
497     This can be used to teach the compiler about invariants or other
498     conditions that might improve code generation, but which are impossible to
499     deduce form the code itself.
500    
501 root 1.27 For example, the example reservation function from the C<ecb_expect_false>
502 root 1.7 description could be written thus (only C<ecb_assume> was added):
503    
504     ecb_inline void
505     reserve (int size)
506     {
507 root 1.27 if (ecb_expect_false (current + size > end))
508 root 1.7 real_reserve_method (size); /* presumably noinline */
509    
510     ecb_assume (current + size <= end);
511     }
512    
513     If you then call this function twice, like this:
514    
515     reserve (10);
516     reserve (1);
517    
518     Then the compiler I<might> be able to optimise out the second call
519     completely, as it knows that C<< current + 1 > end >> is false and the
520     call will never be executed.
521    
522 root 1.62 =item ecb_unreachable ()
523 root 1.7
524     This function does nothing itself, except tell the compiler that it will
525 root 1.9 never be executed. Apart from suppressing a warning in some cases, this
526 root 1.7 function can be used to implement C<ecb_assume> or similar functions.
527    
528 root 1.62 =item ecb_prefetch (addr, rw, locality)
529 root 1.7
530     Tells the compiler to try to prefetch memory at the given C<addr>ess
531 root 1.10 for either reading (C<rw> = 0) or writing (C<rw> = 1). A C<locality> of
532 root 1.7 C<0> means that there will only be one access later, C<3> means that
533     the data will likely be accessed very often, and values in between mean
534     something... in between. The memory pointed to by the address does not
535     need to be accessible (it could be a null pointer for example), but C<rw>
536     and C<locality> must be compile-time constants.
537    
538     An obvious way to use this is to prefetch some data far away, in a big
539 root 1.9 array you loop over. This prefetches memory some 128 array elements later,
540 root 1.7 in the hope that it will be ready when the CPU arrives at that location.
541    
542     int sum = 0;
543    
544     for (i = 0; i < N; ++i)
545     {
546     sum += arr [i]
547     ecb_prefetch (arr + i + 128, 0, 0);
548     }
549    
550     It's hard to predict how far to prefetch, and most CPUs that can prefetch
551     are often good enough to predict this kind of behaviour themselves. It
552     gets more interesting with linked lists, especially when you do some fair
553     processing on each list element:
554    
555     for (node *n = start; n; n = n->next)
556     {
557     ecb_prefetch (n->next, 0, 0);
558     ... do medium amount of work with *n
559     }
560    
561     After processing the node, (part of) the next node might already be in
562     cache.
563 root 1.1
564 root 1.2 =back
565 root 1.1
566 root 1.36 =head2 BIT FIDDLING / BIT WIZARDRY
567 root 1.1
568 root 1.4 =over 4
569    
570 root 1.3 =item bool ecb_big_endian ()
571    
572     =item bool ecb_little_endian ()
573    
574 sf-exg 1.11 These two functions return true if the byte order is big endian
575     (most-significant byte first) or little endian (least-significant byte
576     first) respectively.
577    
578 root 1.24 On systems that are neither, their return values are unspecified.
579    
580 root 1.3 =item int ecb_ctz32 (uint32_t x)
581    
582 root 1.35 =item int ecb_ctz64 (uint64_t x)
583    
584 sf-exg 1.11 Returns the index of the least significant bit set in C<x> (or
585 root 1.24 equivalently the number of bits set to 0 before the least significant bit
586 root 1.35 set), starting from 0. If C<x> is 0 the result is undefined.
587    
588 root 1.36 For smaller types than C<uint32_t> you can safely use C<ecb_ctz32>.
589    
590 root 1.35 For example:
591 sf-exg 1.11
592 root 1.15 ecb_ctz32 (3) = 0
593     ecb_ctz32 (6) = 1
594 sf-exg 1.11
595 root 1.41 =item bool ecb_is_pot32 (uint32_t x)
596    
597     =item bool ecb_is_pot64 (uint32_t x)
598    
599     Return true iff C<x> is a power of two or C<x == 0>.
600    
601     For smaller types then C<uint32_t> you can safely use C<ecb_is_pot32>.
602    
603 root 1.35 =item int ecb_ld32 (uint32_t x)
604    
605     =item int ecb_ld64 (uint64_t x)
606    
607     Returns the index of the most significant bit set in C<x>, or the number
608     of digits the number requires in binary (so that C<< 2**ld <= x <
609     2**(ld+1) >>). If C<x> is 0 the result is undefined. A common use case is
610     to compute the integer binary logarithm, i.e. C<floor (log2 (n))>, for
611     example to see how many bits a certain number requires to be encoded.
612    
613     This function is similar to the "count leading zero bits" function, except
614     that that one returns how many zero bits are "in front" of the number (in
615     the given data type), while C<ecb_ld> returns how many bits the number
616     itself requires.
617    
618 root 1.36 For smaller types than C<uint32_t> you can safely use C<ecb_ld32>.
619    
620 root 1.3 =item int ecb_popcount32 (uint32_t x)
621    
622 root 1.35 =item int ecb_popcount64 (uint64_t x)
623    
624 root 1.36 Returns the number of bits set to 1 in C<x>.
625    
626     For smaller types than C<uint32_t> you can safely use C<ecb_popcount32>.
627    
628     For example:
629 sf-exg 1.11
630 root 1.15 ecb_popcount32 (7) = 3
631     ecb_popcount32 (255) = 8
632 sf-exg 1.11
633 root 1.39 =item uint8_t ecb_bitrev8 (uint8_t x)
634    
635     =item uint16_t ecb_bitrev16 (uint16_t x)
636    
637     =item uint32_t ecb_bitrev32 (uint32_t x)
638    
639     Reverses the bits in x, i.e. the MSB becomes the LSB, MSB-1 becomes LSB+1
640     and so on.
641    
642     Example:
643    
644     ecb_bitrev8 (0xa7) = 0xea
645     ecb_bitrev32 (0xffcc4411) = 0x882233ff
646    
647 root 1.8 =item uint32_t ecb_bswap16 (uint32_t x)
648    
649 root 1.3 =item uint32_t ecb_bswap32 (uint32_t x)
650    
651 root 1.34 =item uint64_t ecb_bswap64 (uint64_t x)
652 sf-exg 1.13
653 root 1.34 These functions return the value of the 16-bit (32-bit, 64-bit) value
654     C<x> after reversing the order of bytes (0x11223344 becomes 0x44332211 in
655     C<ecb_bswap32>).
656    
657     =item uint8_t ecb_rotl8 (uint8_t x, unsigned int count)
658    
659     =item uint16_t ecb_rotl16 (uint16_t x, unsigned int count)
660 root 1.3
661     =item uint32_t ecb_rotl32 (uint32_t x, unsigned int count)
662    
663 root 1.34 =item uint64_t ecb_rotl64 (uint64_t x, unsigned int count)
664    
665     =item uint8_t ecb_rotr8 (uint8_t x, unsigned int count)
666    
667     =item uint16_t ecb_rotr16 (uint16_t x, unsigned int count)
668    
669     =item uint32_t ecb_rotr32 (uint32_t x, unsigned int count)
670    
671 root 1.33 =item uint64_t ecb_rotr64 (uint64_t x, unsigned int count)
672    
673 root 1.34 These two families of functions return the value of C<x> after rotating
674     all the bits by C<count> positions to the right (C<ecb_rotr>) or left
675     (C<ecb_rotl>).
676 sf-exg 1.11
677 root 1.20 Current GCC versions understand these functions and usually compile them
678 root 1.34 to "optimal" code (e.g. a single C<rol> or a combination of C<shld> on
679     x86).
680 root 1.20
681 root 1.3 =back
682 root 1.1
683 root 1.50 =head2 FLOATING POINT FIDDLING
684    
685     =over 4
686    
687 root 1.62 =item ECB_INFINITY
688    
689     Evaluates to positive infinity if supported by the platform, otherwise to
690     a truly huge number.
691    
692 root 1.63 =item ECB_NAN
693 root 1.62
694     Evaluates to a quiet NAN if supported by the platform, otherwise to
695     C<ECB_INFINITY>.
696    
697     =item float ecb_ldexpf (float x, int exp)
698    
699     Same as C<ldexpf>, but always available.
700    
701 root 1.50 =item uint32_t ecb_float_to_binary32 (float x) [-UECB_NO_LIBM]
702    
703     =item uint64_t ecb_double_to_binary64 (double x) [-UECB_NO_LIBM]
704    
705     These functions each take an argument in the native C<float> or C<double>
706     type and return the IEEE 754 bit representation of it.
707    
708     The bit representation is just as IEEE 754 defines it, i.e. the sign bit
709     will be the most significant bit, followed by exponent and mantissa.
710    
711     This function should work even when the native floating point format isn't
712     IEEE compliant, of course at a speed and code size penalty, and of course
713     also within reasonable limits (it tries to convert NaNs, infinities and
714     denormals, but will likely convert negative zero to positive zero).
715    
716     On all modern platforms (where C<ECB_STDFP> is true), the compiler should
717     be able to optimise away this function completely.
718    
719     These functions can be helpful when serialising floats to the network - you
720     can serialise the return value like a normal uint32_t/uint64_t.
721    
722     Another use for these functions is to manipulate floating point values
723     directly.
724    
725     Silly example: toggle the sign bit of a float.
726    
727     /* On gcc-4.7 on amd64, */
728     /* this results in a single add instruction to toggle the bit, and 4 extra */
729     /* instructions to move the float value to an integer register and back. */
730    
731     x = ecb_binary32_to_float (ecb_float_to_binary32 (x) ^ 0x80000000U)
732    
733 root 1.58 =item float ecb_binary16_to_float (uint16_t x) [-UECB_NO_LIBM]
734    
735 root 1.50 =item float ecb_binary32_to_float (uint32_t x) [-UECB_NO_LIBM]
736    
737     =item double ecb_binary32_to_double (uint64_t x) [-UECB_NO_LIBM]
738    
739 sf-exg 1.59 The reverse operation of the previous function - takes the bit
740 root 1.58 representation of an IEEE binary16, binary32 or binary64 number and
741     converts it to the native C<float> or C<double> format.
742 root 1.50
743     This function should work even when the native floating point format isn't
744     IEEE compliant, of course at a speed and code size penalty, and of course
745     also within reasonable limits (it tries to convert normals and denormals,
746     and might be lucky for infinities, and with extraordinary luck, also for
747     negative zero).
748    
749     On all modern platforms (where C<ECB_STDFP> is true), the compiler should
750     be able to optimise away this function completely.
751    
752     =back
753    
754 root 1.1 =head2 ARITHMETIC
755    
756 root 1.3 =over 4
757    
758 root 1.14 =item x = ecb_mod (m, n)
759 root 1.3
760 root 1.25 Returns C<m> modulo C<n>, which is the same as the positive remainder
761     of the division operation between C<m> and C<n>, using floored
762     division. Unlike the C remainder operator C<%>, this function ensures that
763     the return value is always positive and that the two numbers I<m> and
764     I<m' = m + i * n> result in the same value modulo I<n> - in other words,
765     C<ecb_mod> implements the mathematical modulo operation, which is missing
766     in the language.
767 root 1.14
768 sf-exg 1.23 C<n> must be strictly positive (i.e. C<< >= 1 >>), while C<m> must be
769 root 1.14 negatable, that is, both C<m> and C<-m> must be representable in its
770 root 1.30 type (this typically excludes the minimum signed integer value, the same
771 root 1.25 limitation as for C</> and C<%> in C).
772 sf-exg 1.11
773 root 1.24 Current GCC versions compile this into an efficient branchless sequence on
774 root 1.28 almost all CPUs.
775 root 1.24
776     For example, when you want to rotate forward through the members of an
777     array for increasing C<m> (which might be negative), then you should use
778     C<ecb_mod>, as the C<%> operator might give either negative results, or
779     change direction for negative values:
780    
781     for (m = -100; m <= 100; ++m)
782     int elem = myarray [ecb_mod (m, ecb_array_length (myarray))];
783    
784 sf-exg 1.37 =item x = ecb_div_rd (val, div)
785    
786     =item x = ecb_div_ru (val, div)
787    
788     Returns C<val> divided by C<div> rounded down or up, respectively.
789     C<val> and C<div> must have integer types and C<div> must be strictly
790 sf-exg 1.38 positive. Note that these functions are implemented with macros in C
791     and with function templates in C++.
792 sf-exg 1.37
793 root 1.3 =back
794 root 1.1
795     =head2 UTILITY
796    
797 root 1.3 =over 4
798    
799 sf-exg 1.23 =item element_count = ecb_array_length (name)
800 root 1.3
801 sf-exg 1.13 Returns the number of elements in the array C<name>. For example:
802    
803     int primes[] = { 2, 3, 5, 7, 11 };
804     int sum = 0;
805    
806     for (i = 0; i < ecb_array_length (primes); i++)
807     sum += primes [i];
808    
809 root 1.3 =back
810 root 1.1
811 root 1.43 =head2 SYMBOLS GOVERNING COMPILATION OF ECB.H ITSELF
812    
813     These symbols need to be defined before including F<ecb.h> the first time.
814    
815     =over 4
816    
817 root 1.51 =item ECB_NO_THREADS
818 root 1.43
819     If F<ecb.h> is never used from multiple threads, then this symbol can
820     be defined, in which case memory fences (and similar constructs) are
821     completely removed, leading to more efficient code and fewer dependencies.
822    
823     Setting this symbol to a true value implies C<ECB_NO_SMP>.
824    
825     =item ECB_NO_SMP
826    
827     The weaker version of C<ECB_NO_THREADS> - if F<ecb.h> is used from
828     multiple threads, but never concurrently (e.g. if the system the program
829     runs on has only a single CPU with a single core, no hyperthreading and so
830     on), then this symbol can be defined, leading to more efficient code and
831     fewer dependencies.
832    
833 root 1.50 =item ECB_NO_LIBM
834    
835     When defined to C<1>, do not export any functions that might introduce
836     dependencies on the math library (usually called F<-lm>) - these are
837     marked with [-UECB_NO_LIBM].
838    
839 root 1.43 =back
840    
841 root 1.1