ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libecb/ecb.pod
Revision: 1.58
Committed: Thu Oct 16 15:01:54 2014 UTC (9 years, 8 months ago) by root
Branch: MAIN
Changes since 1.57: +6 -4 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 1.43 =head2 LANGUAGE/COMPILER VERSIONS
74    
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 1.1 =head2 GCC ATTRIBUTES
169    
170 root 1.20 A major part of libecb deals with GCC attributes. These are additional
171 sf-exg 1.26 attributes that you can assign to functions, variables and sometimes even
172 root 1.20 types - much like C<const> or C<volatile> in C.
173    
174     While GCC allows declarations to show up in many surprising places,
175 sf-exg 1.26 but not in many expected places, the safest way is to put attribute
176 root 1.20 declarations before the whole declaration:
177    
178     ecb_const int mysqrt (int a);
179     ecb_unused int i;
180    
181     For variables, it is often nicer to put the attribute after the name, and
182     avoid multiple declarations using commas:
183    
184     int i ecb_unused;
185 root 1.3
186 root 1.1 =over 4
187    
188 root 1.2 =item ecb_attribute ((attrs...))
189 root 1.1
190 root 1.57 A simple wrapper that expands to C<__attribute__((attrs))> on GCC 3.1+ and
191     Clang 2.8+, and to nothing on other compilers, so the effect is that only
192     GCC and Clang see these.
193 root 1.15
194     Example: use the C<deprecated> attribute on a function.
195    
196     ecb_attribute((__deprecated__)) void
197     do_not_use_me_anymore (void);
198 root 1.2
199 root 1.3 =item ecb_unused
200    
201     Marks a function or a variable as "unused", which simply suppresses a
202     warning by GCC when it detects it as unused. This is useful when you e.g.
203     declare a variable but do not always use it:
204    
205 root 1.15 {
206     int var ecb_unused;
207 root 1.3
208 root 1.15 #ifdef SOMECONDITION
209     var = ...;
210     return var;
211     #else
212     return 0;
213     #endif
214     }
215 root 1.3
216 root 1.56 =item ecb_deprecated
217    
218     Similar to C<ecb_unused>, but marks a function, variable or type as
219     deprecated. This makes some compilers warn when the type is used.
220    
221 root 1.31 =item ecb_inline
222 root 1.29
223     This is not actually an attribute, but you use it like one. It expands
224     either to C<static inline> or to just C<static>, if inline isn't
225     supported. It should be used to declare functions that should be inlined,
226     for code size or speed reasons.
227    
228     Example: inline this function, it surely will reduce codesize.
229    
230 root 1.31 ecb_inline int
231 root 1.29 negmul (int a, int b)
232     {
233     return - (a * b);
234     }
235    
236 root 1.2 =item ecb_noinline
237    
238 root 1.9 Prevent a function from being inlined - it might be optimised away, but
239 root 1.3 not inlined into other functions. This is useful if you know your function
240     is rarely called and large enough for inlining not to be helpful.
241    
242 root 1.2 =item ecb_noreturn
243    
244 root 1.17 Marks a function as "not returning, ever". Some typical functions that
245     don't return are C<exit> or C<abort> (which really works hard to not
246     return), and now you can make your own:
247    
248     ecb_noreturn void
249     my_abort (const char *errline)
250     {
251     puts (errline);
252     abort ();
253     }
254    
255 sf-exg 1.19 In this case, the compiler would probably be smart enough to deduce it on
256     its own, so this is mainly useful for declarations.
257 root 1.17
258 root 1.53 =item ecb_restrict
259    
260     Expands to the C<restrict> keyword or equivalent on compilers that support
261     them, and to nothing on others. Must be specified on a pointer type or
262     an array index to indicate that the memory doesn't alias with any other
263     restricted pointer in the same scope.
264    
265     Example: multiply a vector, and allow the compiler to parallelise the
266     loop, because it knows it doesn't overwrite input values.
267    
268     void
269     multiply (float *ecb_restrict src,
270     float *ecb_restrict dst,
271     int len, float factor)
272     {
273     int i;
274    
275     for (i = 0; i < len; ++i)
276     dst [i] = src [i] * factor;
277     }
278    
279 root 1.2 =item ecb_const
280    
281 sf-exg 1.19 Declares that the function only depends on the values of its arguments,
282 root 1.17 much like a mathematical function. It specifically does not read or write
283     any memory any arguments might point to, global variables, or call any
284     non-const functions. It also must not have any side effects.
285    
286     Such a function can be optimised much more aggressively by the compiler -
287     for example, multiple calls with the same arguments can be optimised into
288     a single call, which wouldn't be possible if the compiler would have to
289     expect any side effects.
290    
291     It is best suited for functions in the sense of mathematical functions,
292 sf-exg 1.19 such as a function returning the square root of its input argument.
293 root 1.17
294     Not suited would be a function that calculates the hash of some memory
295     area you pass in, prints some messages or looks at a global variable to
296     decide on rounding.
297    
298     See C<ecb_pure> for a slightly less restrictive class of functions.
299    
300 root 1.2 =item ecb_pure
301    
302 root 1.17 Similar to C<ecb_const>, declares a function that has no side
303     effects. Unlike C<ecb_const>, the function is allowed to examine global
304     variables and any other memory areas (such as the ones passed to it via
305     pointers).
306    
307     While these functions cannot be optimised as aggressively as C<ecb_const>
308     functions, they can still be optimised away in many occasions, and the
309     compiler has more freedom in moving calls to them around.
310    
311     Typical examples for such functions would be C<strlen> or C<memcmp>. A
312     function that calculates the MD5 sum of some input and updates some MD5
313     state passed as argument would I<NOT> be pure, however, as it would modify
314     some memory area that is not the return value.
315    
316 root 1.2 =item ecb_hot
317    
318 root 1.17 This declares a function as "hot" with regards to the cache - the function
319     is used so often, that it is very beneficial to keep it in the cache if
320     possible.
321    
322     The compiler reacts by trying to place hot functions near to each other in
323     memory.
324    
325 sf-exg 1.19 Whether a function is hot or not often depends on the whole program,
326 root 1.17 and less on the function itself. C<ecb_cold> is likely more useful in
327     practise.
328    
329 root 1.2 =item ecb_cold
330    
331 root 1.17 The opposite of C<ecb_hot> - declares a function as "cold" with regards to
332     the cache, or in other words, this function is not called often, or not at
333     speed-critical times, and keeping it in the cache might be a waste of said
334     cache.
335    
336     In addition to placing cold functions together (or at least away from hot
337     functions), this knowledge can be used in other ways, for example, the
338     function will be optimised for size, as opposed to speed, and codepaths
339     leading to calls to those functions can automatically be marked as if
340 root 1.27 C<ecb_expect_false> had been used to reach them.
341 root 1.17
342     Good examples for such functions would be error reporting functions, or
343     functions only called in exceptional or rare cases.
344    
345 root 1.2 =item ecb_artificial
346    
347 root 1.17 Declares the function as "artificial", in this case meaning that this
348 root 1.52 function is not really meant to be a function, but more like an accessor
349 root 1.17 - many methods in C++ classes are mere accessor functions, and having a
350     crash reported in such a method, or single-stepping through them, is not
351     usually so helpful, especially when it's inlined to just a few instructions.
352    
353     Marking them as artificial will instruct the debugger about just this,
354     leading to happier debugging and thus happier lives.
355    
356     Example: in some kind of smart-pointer class, mark the pointer accessor as
357     artificial, so that the whole class acts more like a pointer and less like
358     some C++ abstraction monster.
359    
360     template<typename T>
361     struct my_smart_ptr
362     {
363     T *value;
364    
365     ecb_artificial
366     operator T *()
367     {
368     return value;
369     }
370     };
371    
372 root 1.2 =back
373 root 1.1
374     =head2 OPTIMISATION HINTS
375    
376     =over 4
377    
378 root 1.58 =item bool ecb_is_constant (expr)
379 root 1.1
380 root 1.3 Returns true iff the expression can be deduced to be a compile-time
381     constant, and false otherwise.
382    
383     For example, when you have a C<rndm16> function that returns a 16 bit
384     random number, and you have a function that maps this to a range from
385 root 1.5 0..n-1, then you could use this inline function in a header file:
386 root 1.3
387     ecb_inline uint32_t
388     rndm (uint32_t n)
389     {
390 root 1.6 return (n * (uint32_t)rndm16 ()) >> 16;
391 root 1.3 }
392    
393     However, for powers of two, you could use a normal mask, but that is only
394     worth it if, at compile time, you can detect this case. This is the case
395     when the passed number is a constant and also a power of two (C<n & (n -
396     1) == 0>):
397    
398     ecb_inline uint32_t
399     rndm (uint32_t n)
400     {
401     return is_constant (n) && !(n & (n - 1))
402     ? rndm16 () & (num - 1)
403 root 1.6 : (n * (uint32_t)rndm16 ()) >> 16;
404 root 1.3 }
405    
406 root 1.14 =item bool ecb_expect (expr, value)
407 root 1.1
408 root 1.7 Evaluates C<expr> and returns it. In addition, it tells the compiler that
409     the C<expr> evaluates to C<value> a lot, which can be used for static
410     branch optimisations.
411 root 1.1
412 root 1.27 Usually, you want to use the more intuitive C<ecb_expect_true> and
413     C<ecb_expect_false> functions instead.
414 root 1.1
415 root 1.27 =item bool ecb_expect_true (cond)
416 root 1.1
417 root 1.27 =item bool ecb_expect_false (cond)
418 root 1.1
419 root 1.7 These two functions expect a expression that is true or false and return
420     C<1> or C<0>, respectively, so when used in the condition of an C<if> or
421     other conditional statement, it will not change the program:
422    
423     /* these two do the same thing */
424     if (some_condition) ...;
425 root 1.27 if (ecb_expect_true (some_condition)) ...;
426 root 1.7
427 root 1.27 However, by using C<ecb_expect_true>, you tell the compiler that the
428     condition is likely to be true (and for C<ecb_expect_false>, that it is
429     unlikely to be true).
430 root 1.7
431 root 1.9 For example, when you check for a null pointer and expect this to be a
432 root 1.27 rare, exceptional, case, then use C<ecb_expect_false>:
433 root 1.7
434     void my_free (void *ptr)
435     {
436 root 1.27 if (ecb_expect_false (ptr == 0))
437 root 1.7 return;
438     }
439    
440     Consequent use of these functions to mark away exceptional cases or to
441     tell the compiler what the hot path through a function is can increase
442     performance considerably.
443    
444 root 1.27 You might know these functions under the name C<likely> and C<unlikely>
445     - while these are common aliases, we find that the expect name is easier
446     to understand when quickly skimming code. If you wish, you can use
447     C<ecb_likely> instead of C<ecb_expect_true> and C<ecb_unlikely> instead of
448     C<ecb_expect_false> - these are simply aliases.
449    
450 root 1.7 A very good example is in a function that reserves more space for some
451     memory block (for example, inside an implementation of a string stream) -
452 root 1.9 each time something is added, you have to check for a buffer overrun, but
453 root 1.7 you expect that most checks will turn out to be false:
454    
455     /* make sure we have "size" extra room in our buffer */
456     ecb_inline void
457     reserve (int size)
458     {
459 root 1.27 if (ecb_expect_false (current + size > end))
460 root 1.7 real_reserve_method (size); /* presumably noinline */
461     }
462    
463 root 1.14 =item bool ecb_assume (cond)
464 root 1.7
465     Try to tell the compiler that some condition is true, even if it's not
466     obvious.
467    
468     This can be used to teach the compiler about invariants or other
469     conditions that might improve code generation, but which are impossible to
470     deduce form the code itself.
471    
472 root 1.27 For example, the example reservation function from the C<ecb_expect_false>
473 root 1.7 description could be written thus (only C<ecb_assume> was added):
474    
475     ecb_inline void
476     reserve (int size)
477     {
478 root 1.27 if (ecb_expect_false (current + size > end))
479 root 1.7 real_reserve_method (size); /* presumably noinline */
480    
481     ecb_assume (current + size <= end);
482     }
483    
484     If you then call this function twice, like this:
485    
486     reserve (10);
487     reserve (1);
488    
489     Then the compiler I<might> be able to optimise out the second call
490     completely, as it knows that C<< current + 1 > end >> is false and the
491     call will never be executed.
492    
493     =item bool ecb_unreachable ()
494    
495     This function does nothing itself, except tell the compiler that it will
496 root 1.9 never be executed. Apart from suppressing a warning in some cases, this
497 root 1.7 function can be used to implement C<ecb_assume> or similar functions.
498    
499 root 1.14 =item bool ecb_prefetch (addr, rw, locality)
500 root 1.7
501     Tells the compiler to try to prefetch memory at the given C<addr>ess
502 root 1.10 for either reading (C<rw> = 0) or writing (C<rw> = 1). A C<locality> of
503 root 1.7 C<0> means that there will only be one access later, C<3> means that
504     the data will likely be accessed very often, and values in between mean
505     something... in between. The memory pointed to by the address does not
506     need to be accessible (it could be a null pointer for example), but C<rw>
507     and C<locality> must be compile-time constants.
508    
509     An obvious way to use this is to prefetch some data far away, in a big
510 root 1.9 array you loop over. This prefetches memory some 128 array elements later,
511 root 1.7 in the hope that it will be ready when the CPU arrives at that location.
512    
513     int sum = 0;
514    
515     for (i = 0; i < N; ++i)
516     {
517     sum += arr [i]
518     ecb_prefetch (arr + i + 128, 0, 0);
519     }
520    
521     It's hard to predict how far to prefetch, and most CPUs that can prefetch
522     are often good enough to predict this kind of behaviour themselves. It
523     gets more interesting with linked lists, especially when you do some fair
524     processing on each list element:
525    
526     for (node *n = start; n; n = n->next)
527     {
528     ecb_prefetch (n->next, 0, 0);
529     ... do medium amount of work with *n
530     }
531    
532     After processing the node, (part of) the next node might already be in
533     cache.
534 root 1.1
535 root 1.2 =back
536 root 1.1
537 root 1.36 =head2 BIT FIDDLING / BIT WIZARDRY
538 root 1.1
539 root 1.4 =over 4
540    
541 root 1.3 =item bool ecb_big_endian ()
542    
543     =item bool ecb_little_endian ()
544    
545 sf-exg 1.11 These two functions return true if the byte order is big endian
546     (most-significant byte first) or little endian (least-significant byte
547     first) respectively.
548    
549 root 1.24 On systems that are neither, their return values are unspecified.
550    
551 root 1.3 =item int ecb_ctz32 (uint32_t x)
552    
553 root 1.35 =item int ecb_ctz64 (uint64_t x)
554    
555 sf-exg 1.11 Returns the index of the least significant bit set in C<x> (or
556 root 1.24 equivalently the number of bits set to 0 before the least significant bit
557 root 1.35 set), starting from 0. If C<x> is 0 the result is undefined.
558    
559 root 1.36 For smaller types than C<uint32_t> you can safely use C<ecb_ctz32>.
560    
561 root 1.35 For example:
562 sf-exg 1.11
563 root 1.15 ecb_ctz32 (3) = 0
564     ecb_ctz32 (6) = 1
565 sf-exg 1.11
566 root 1.41 =item bool ecb_is_pot32 (uint32_t x)
567    
568     =item bool ecb_is_pot64 (uint32_t x)
569    
570     Return true iff C<x> is a power of two or C<x == 0>.
571    
572     For smaller types then C<uint32_t> you can safely use C<ecb_is_pot32>.
573    
574 root 1.35 =item int ecb_ld32 (uint32_t x)
575    
576     =item int ecb_ld64 (uint64_t x)
577    
578     Returns the index of the most significant bit set in C<x>, or the number
579     of digits the number requires in binary (so that C<< 2**ld <= x <
580     2**(ld+1) >>). If C<x> is 0 the result is undefined. A common use case is
581     to compute the integer binary logarithm, i.e. C<floor (log2 (n))>, for
582     example to see how many bits a certain number requires to be encoded.
583    
584     This function is similar to the "count leading zero bits" function, except
585     that that one returns how many zero bits are "in front" of the number (in
586     the given data type), while C<ecb_ld> returns how many bits the number
587     itself requires.
588    
589 root 1.36 For smaller types than C<uint32_t> you can safely use C<ecb_ld32>.
590    
591 root 1.3 =item int ecb_popcount32 (uint32_t x)
592    
593 root 1.35 =item int ecb_popcount64 (uint64_t x)
594    
595 root 1.36 Returns the number of bits set to 1 in C<x>.
596    
597     For smaller types than C<uint32_t> you can safely use C<ecb_popcount32>.
598    
599     For example:
600 sf-exg 1.11
601 root 1.15 ecb_popcount32 (7) = 3
602     ecb_popcount32 (255) = 8
603 sf-exg 1.11
604 root 1.39 =item uint8_t ecb_bitrev8 (uint8_t x)
605    
606     =item uint16_t ecb_bitrev16 (uint16_t x)
607    
608     =item uint32_t ecb_bitrev32 (uint32_t x)
609    
610     Reverses the bits in x, i.e. the MSB becomes the LSB, MSB-1 becomes LSB+1
611     and so on.
612    
613     Example:
614    
615     ecb_bitrev8 (0xa7) = 0xea
616     ecb_bitrev32 (0xffcc4411) = 0x882233ff
617    
618 root 1.8 =item uint32_t ecb_bswap16 (uint32_t x)
619    
620 root 1.3 =item uint32_t ecb_bswap32 (uint32_t x)
621    
622 root 1.34 =item uint64_t ecb_bswap64 (uint64_t x)
623 sf-exg 1.13
624 root 1.34 These functions return the value of the 16-bit (32-bit, 64-bit) value
625     C<x> after reversing the order of bytes (0x11223344 becomes 0x44332211 in
626     C<ecb_bswap32>).
627    
628     =item uint8_t ecb_rotl8 (uint8_t x, unsigned int count)
629    
630     =item uint16_t ecb_rotl16 (uint16_t x, unsigned int count)
631 root 1.3
632     =item uint32_t ecb_rotl32 (uint32_t x, unsigned int count)
633    
634 root 1.34 =item uint64_t ecb_rotl64 (uint64_t x, unsigned int count)
635    
636     =item uint8_t ecb_rotr8 (uint8_t x, unsigned int count)
637    
638     =item uint16_t ecb_rotr16 (uint16_t x, unsigned int count)
639    
640     =item uint32_t ecb_rotr32 (uint32_t x, unsigned int count)
641    
642 root 1.33 =item uint64_t ecb_rotr64 (uint64_t x, unsigned int count)
643    
644 root 1.34 These two families of functions return the value of C<x> after rotating
645     all the bits by C<count> positions to the right (C<ecb_rotr>) or left
646     (C<ecb_rotl>).
647 sf-exg 1.11
648 root 1.20 Current GCC versions understand these functions and usually compile them
649 root 1.34 to "optimal" code (e.g. a single C<rol> or a combination of C<shld> on
650     x86).
651 root 1.20
652 root 1.3 =back
653 root 1.1
654 root 1.50 =head2 FLOATING POINT FIDDLING
655    
656     =over 4
657    
658     =item uint32_t ecb_float_to_binary32 (float x) [-UECB_NO_LIBM]
659    
660     =item uint64_t ecb_double_to_binary64 (double x) [-UECB_NO_LIBM]
661    
662     These functions each take an argument in the native C<float> or C<double>
663     type and return the IEEE 754 bit representation of it.
664    
665     The bit representation is just as IEEE 754 defines it, i.e. the sign bit
666     will be the most significant bit, followed by exponent and mantissa.
667    
668     This function should work even when the native floating point format isn't
669     IEEE compliant, of course at a speed and code size penalty, and of course
670     also within reasonable limits (it tries to convert NaNs, infinities and
671     denormals, but will likely convert negative zero to positive zero).
672    
673     On all modern platforms (where C<ECB_STDFP> is true), the compiler should
674     be able to optimise away this function completely.
675    
676     These functions can be helpful when serialising floats to the network - you
677     can serialise the return value like a normal uint32_t/uint64_t.
678    
679     Another use for these functions is to manipulate floating point values
680     directly.
681    
682     Silly example: toggle the sign bit of a float.
683    
684     /* On gcc-4.7 on amd64, */
685     /* this results in a single add instruction to toggle the bit, and 4 extra */
686     /* instructions to move the float value to an integer register and back. */
687    
688     x = ecb_binary32_to_float (ecb_float_to_binary32 (x) ^ 0x80000000U)
689    
690 root 1.58 =item float ecb_binary16_to_float (uint16_t x) [-UECB_NO_LIBM]
691    
692 root 1.50 =item float ecb_binary32_to_float (uint32_t x) [-UECB_NO_LIBM]
693    
694     =item double ecb_binary32_to_double (uint64_t x) [-UECB_NO_LIBM]
695    
696 root 1.58 The reverse operation of the previos function - takes the bit
697     representation of an IEEE binary16, binary32 or binary64 number and
698     converts it to the native C<float> or C<double> format.
699 root 1.50
700     This function should work even when the native floating point format isn't
701     IEEE compliant, of course at a speed and code size penalty, and of course
702     also within reasonable limits (it tries to convert normals and denormals,
703     and might be lucky for infinities, and with extraordinary luck, also for
704     negative zero).
705    
706     On all modern platforms (where C<ECB_STDFP> is true), the compiler should
707     be able to optimise away this function completely.
708    
709     =back
710    
711 root 1.1 =head2 ARITHMETIC
712    
713 root 1.3 =over 4
714    
715 root 1.14 =item x = ecb_mod (m, n)
716 root 1.3
717 root 1.25 Returns C<m> modulo C<n>, which is the same as the positive remainder
718     of the division operation between C<m> and C<n>, using floored
719     division. Unlike the C remainder operator C<%>, this function ensures that
720     the return value is always positive and that the two numbers I<m> and
721     I<m' = m + i * n> result in the same value modulo I<n> - in other words,
722     C<ecb_mod> implements the mathematical modulo operation, which is missing
723     in the language.
724 root 1.14
725 sf-exg 1.23 C<n> must be strictly positive (i.e. C<< >= 1 >>), while C<m> must be
726 root 1.14 negatable, that is, both C<m> and C<-m> must be representable in its
727 root 1.30 type (this typically excludes the minimum signed integer value, the same
728 root 1.25 limitation as for C</> and C<%> in C).
729 sf-exg 1.11
730 root 1.24 Current GCC versions compile this into an efficient branchless sequence on
731 root 1.28 almost all CPUs.
732 root 1.24
733     For example, when you want to rotate forward through the members of an
734     array for increasing C<m> (which might be negative), then you should use
735     C<ecb_mod>, as the C<%> operator might give either negative results, or
736     change direction for negative values:
737    
738     for (m = -100; m <= 100; ++m)
739     int elem = myarray [ecb_mod (m, ecb_array_length (myarray))];
740    
741 sf-exg 1.37 =item x = ecb_div_rd (val, div)
742    
743     =item x = ecb_div_ru (val, div)
744    
745     Returns C<val> divided by C<div> rounded down or up, respectively.
746     C<val> and C<div> must have integer types and C<div> must be strictly
747 sf-exg 1.38 positive. Note that these functions are implemented with macros in C
748     and with function templates in C++.
749 sf-exg 1.37
750 root 1.3 =back
751 root 1.1
752     =head2 UTILITY
753    
754 root 1.3 =over 4
755    
756 sf-exg 1.23 =item element_count = ecb_array_length (name)
757 root 1.3
758 sf-exg 1.13 Returns the number of elements in the array C<name>. For example:
759    
760     int primes[] = { 2, 3, 5, 7, 11 };
761     int sum = 0;
762    
763     for (i = 0; i < ecb_array_length (primes); i++)
764     sum += primes [i];
765    
766 root 1.3 =back
767 root 1.1
768 root 1.43 =head2 SYMBOLS GOVERNING COMPILATION OF ECB.H ITSELF
769    
770     These symbols need to be defined before including F<ecb.h> the first time.
771    
772     =over 4
773    
774 root 1.51 =item ECB_NO_THREADS
775 root 1.43
776     If F<ecb.h> is never used from multiple threads, then this symbol can
777     be defined, in which case memory fences (and similar constructs) are
778     completely removed, leading to more efficient code and fewer dependencies.
779    
780     Setting this symbol to a true value implies C<ECB_NO_SMP>.
781    
782     =item ECB_NO_SMP
783    
784     The weaker version of C<ECB_NO_THREADS> - if F<ecb.h> is used from
785     multiple threads, but never concurrently (e.g. if the system the program
786     runs on has only a single CPU with a single core, no hyperthreading and so
787     on), then this symbol can be defined, leading to more efficient code and
788     fewer dependencies.
789    
790 root 1.50 =item ECB_NO_LIBM
791    
792     When defined to C<1>, do not export any functions that might introduce
793     dependencies on the math library (usually called F<-lm>) - these are
794     marked with [-UECB_NO_LIBM].
795    
796 root 1.43 =back
797    
798 root 1.1