… | |
… | |
57 | refers to any kind of boolean value, not a specific type. |
57 | refers to any kind of boolean value, not a specific type. |
58 | |
58 | |
59 | =head2 GCC ATTRIBUTES |
59 | =head2 GCC ATTRIBUTES |
60 | |
60 | |
61 | A major part of libecb deals with GCC attributes. These are additional |
61 | A major part of libecb deals with GCC attributes. These are additional |
62 | attributes that you cna assign to functions, variables and sometimes even |
62 | attributes that you can assign to functions, variables and sometimes even |
63 | types - much like C<const> or C<volatile> in C. |
63 | types - much like C<const> or C<volatile> in C. |
64 | |
64 | |
65 | While GCC allows declarations to show up in many surprising places, |
65 | While GCC allows declarations to show up in many surprising places, |
66 | but not in many expeted places, the safest way is to put attribute |
66 | but not in many expected places, the safest way is to put attribute |
67 | declarations before the whole declaration: |
67 | declarations before the whole declaration: |
68 | |
68 | |
69 | ecb_const int mysqrt (int a); |
69 | ecb_const int mysqrt (int a); |
70 | ecb_unused int i; |
70 | ecb_unused int i; |
71 | |
71 | |
… | |
… | |
101 | #else |
101 | #else |
102 | return 0; |
102 | return 0; |
103 | #endif |
103 | #endif |
104 | } |
104 | } |
105 | |
105 | |
|
|
106 | =item ecb_inline |
|
|
107 | |
|
|
108 | This is not actually an attribute, but you use it like one. It expands |
|
|
109 | either to C<static inline> or to just C<static>, if inline isn't |
|
|
110 | supported. It should be used to declare functions that should be inlined, |
|
|
111 | for code size or speed reasons. |
|
|
112 | |
|
|
113 | Example: inline this function, it surely will reduce codesize. |
|
|
114 | |
|
|
115 | ecb_inline int |
|
|
116 | negmul (int a, int b) |
|
|
117 | { |
|
|
118 | return - (a * b); |
|
|
119 | } |
|
|
120 | |
106 | =item ecb_noinline |
121 | =item ecb_noinline |
107 | |
122 | |
108 | Prevent a function from being inlined - it might be optimised away, but |
123 | Prevent a function from being inlined - it might be optimised away, but |
109 | not inlined into other functions. This is useful if you know your function |
124 | not inlined into other functions. This is useful if you know your function |
110 | is rarely called and large enough for inlining not to be helpful. |
125 | is rarely called and large enough for inlining not to be helpful. |
… | |
… | |
184 | |
199 | |
185 | In addition to placing cold functions together (or at least away from hot |
200 | In addition to placing cold functions together (or at least away from hot |
186 | functions), this knowledge can be used in other ways, for example, the |
201 | functions), this knowledge can be used in other ways, for example, the |
187 | function will be optimised for size, as opposed to speed, and codepaths |
202 | function will be optimised for size, as opposed to speed, and codepaths |
188 | leading to calls to those functions can automatically be marked as if |
203 | leading to calls to those functions can automatically be marked as if |
189 | C<ecb_unlikely> had been used to reach them. |
204 | C<ecb_expect_false> had been used to reach them. |
190 | |
205 | |
191 | Good examples for such functions would be error reporting functions, or |
206 | Good examples for such functions would be error reporting functions, or |
192 | functions only called in exceptional or rare cases. |
207 | functions only called in exceptional or rare cases. |
193 | |
208 | |
194 | =item ecb_artificial |
209 | =item ecb_artificial |
… | |
… | |
256 | |
271 | |
257 | Evaluates C<expr> and returns it. In addition, it tells the compiler that |
272 | Evaluates C<expr> and returns it. In addition, it tells the compiler that |
258 | the C<expr> evaluates to C<value> a lot, which can be used for static |
273 | the C<expr> evaluates to C<value> a lot, which can be used for static |
259 | branch optimisations. |
274 | branch optimisations. |
260 | |
275 | |
261 | Usually, you want to use the more intuitive C<ecb_likely> and |
276 | Usually, you want to use the more intuitive C<ecb_expect_true> and |
262 | C<ecb_unlikely> functions instead. |
277 | C<ecb_expect_false> functions instead. |
263 | |
278 | |
|
|
279 | =item bool ecb_expect_true (cond) |
|
|
280 | |
264 | =item bool ecb_likely (cond) |
281 | =item bool ecb_expect_false (cond) |
265 | |
|
|
266 | =item bool ecb_unlikely (cond) |
|
|
267 | |
282 | |
268 | These two functions expect a expression that is true or false and return |
283 | These two functions expect a expression that is true or false and return |
269 | C<1> or C<0>, respectively, so when used in the condition of an C<if> or |
284 | C<1> or C<0>, respectively, so when used in the condition of an C<if> or |
270 | other conditional statement, it will not change the program: |
285 | other conditional statement, it will not change the program: |
271 | |
286 | |
272 | /* these two do the same thing */ |
287 | /* these two do the same thing */ |
273 | if (some_condition) ...; |
288 | if (some_condition) ...; |
274 | if (ecb_likely (some_condition)) ...; |
289 | if (ecb_expect_true (some_condition)) ...; |
275 | |
290 | |
276 | However, by using C<ecb_likely>, you tell the compiler that the condition |
291 | However, by using C<ecb_expect_true>, you tell the compiler that the |
277 | is likely to be true (and for C<ecb_unlikely>, that it is unlikely to be |
292 | condition is likely to be true (and for C<ecb_expect_false>, that it is |
278 | true). |
293 | unlikely to be true). |
279 | |
294 | |
280 | For example, when you check for a null pointer and expect this to be a |
295 | For example, when you check for a null pointer and expect this to be a |
281 | rare, exceptional, case, then use C<ecb_unlikely>: |
296 | rare, exceptional, case, then use C<ecb_expect_false>: |
282 | |
297 | |
283 | void my_free (void *ptr) |
298 | void my_free (void *ptr) |
284 | { |
299 | { |
285 | if (ecb_unlikely (ptr == 0)) |
300 | if (ecb_expect_false (ptr == 0)) |
286 | return; |
301 | return; |
287 | } |
302 | } |
288 | |
303 | |
289 | Consequent use of these functions to mark away exceptional cases or to |
304 | Consequent use of these functions to mark away exceptional cases or to |
290 | tell the compiler what the hot path through a function is can increase |
305 | tell the compiler what the hot path through a function is can increase |
291 | performance considerably. |
306 | performance considerably. |
|
|
307 | |
|
|
308 | You might know these functions under the name C<likely> and C<unlikely> |
|
|
309 | - while these are common aliases, we find that the expect name is easier |
|
|
310 | to understand when quickly skimming code. If you wish, you can use |
|
|
311 | C<ecb_likely> instead of C<ecb_expect_true> and C<ecb_unlikely> instead of |
|
|
312 | C<ecb_expect_false> - these are simply aliases. |
292 | |
313 | |
293 | A very good example is in a function that reserves more space for some |
314 | A very good example is in a function that reserves more space for some |
294 | memory block (for example, inside an implementation of a string stream) - |
315 | memory block (for example, inside an implementation of a string stream) - |
295 | each time something is added, you have to check for a buffer overrun, but |
316 | each time something is added, you have to check for a buffer overrun, but |
296 | you expect that most checks will turn out to be false: |
317 | you expect that most checks will turn out to be false: |
297 | |
318 | |
298 | /* make sure we have "size" extra room in our buffer */ |
319 | /* make sure we have "size" extra room in our buffer */ |
299 | ecb_inline void |
320 | ecb_inline void |
300 | reserve (int size) |
321 | reserve (int size) |
301 | { |
322 | { |
302 | if (ecb_unlikely (current + size > end)) |
323 | if (ecb_expect_false (current + size > end)) |
303 | real_reserve_method (size); /* presumably noinline */ |
324 | real_reserve_method (size); /* presumably noinline */ |
304 | } |
325 | } |
305 | |
326 | |
306 | =item bool ecb_assume (cond) |
327 | =item bool ecb_assume (cond) |
307 | |
328 | |
… | |
… | |
310 | |
331 | |
311 | This can be used to teach the compiler about invariants or other |
332 | This can be used to teach the compiler about invariants or other |
312 | conditions that might improve code generation, but which are impossible to |
333 | conditions that might improve code generation, but which are impossible to |
313 | deduce form the code itself. |
334 | deduce form the code itself. |
314 | |
335 | |
315 | For example, the example reservation function from the C<ecb_unlikely> |
336 | For example, the example reservation function from the C<ecb_expect_false> |
316 | description could be written thus (only C<ecb_assume> was added): |
337 | description could be written thus (only C<ecb_assume> was added): |
317 | |
338 | |
318 | ecb_inline void |
339 | ecb_inline void |
319 | reserve (int size) |
340 | reserve (int size) |
320 | { |
341 | { |
321 | if (ecb_unlikely (current + size > end)) |
342 | if (ecb_expect_false (current + size > end)) |
322 | real_reserve_method (size); /* presumably noinline */ |
343 | real_reserve_method (size); /* presumably noinline */ |
323 | |
344 | |
324 | ecb_assume (current + size <= end); |
345 | ecb_assume (current + size <= end); |
325 | } |
346 | } |
326 | |
347 | |
… | |
… | |
375 | After processing the node, (part of) the next node might already be in |
396 | After processing the node, (part of) the next node might already be in |
376 | cache. |
397 | cache. |
377 | |
398 | |
378 | =back |
399 | =back |
379 | |
400 | |
380 | =head2 BIT FIDDLING / BITSTUFFS |
401 | =head2 BIT FIDDLING / BIT WIZARDRY |
381 | |
402 | |
382 | =over 4 |
403 | =over 4 |
383 | |
404 | |
384 | =item bool ecb_big_endian () |
405 | =item bool ecb_big_endian () |
385 | |
406 | |
… | |
… | |
391 | |
412 | |
392 | On systems that are neither, their return values are unspecified. |
413 | On systems that are neither, their return values are unspecified. |
393 | |
414 | |
394 | =item int ecb_ctz32 (uint32_t x) |
415 | =item int ecb_ctz32 (uint32_t x) |
395 | |
416 | |
|
|
417 | =item int ecb_ctz64 (uint64_t x) |
|
|
418 | |
396 | Returns the index of the least significant bit set in C<x> (or |
419 | Returns the index of the least significant bit set in C<x> (or |
397 | equivalently the number of bits set to 0 before the least significant bit |
420 | equivalently the number of bits set to 0 before the least significant bit |
398 | set), starting from 0. If C<x> is 0 the result is undefined. A common use |
421 | set), starting from 0. If C<x> is 0 the result is undefined. |
399 | case is to compute the integer binary logarithm, i.e., C<floor (log2 |
422 | |
|
|
423 | For smaller types than C<uint32_t> you can safely use C<ecb_ctz32>. |
|
|
424 | |
400 | (n))>. For example: |
425 | For example: |
401 | |
426 | |
402 | ecb_ctz32 (3) = 0 |
427 | ecb_ctz32 (3) = 0 |
403 | ecb_ctz32 (6) = 1 |
428 | ecb_ctz32 (6) = 1 |
404 | |
429 | |
|
|
430 | =item int ecb_ld32 (uint32_t x) |
|
|
431 | |
|
|
432 | =item int ecb_ld64 (uint64_t x) |
|
|
433 | |
|
|
434 | Returns the index of the most significant bit set in C<x>, or the number |
|
|
435 | of digits the number requires in binary (so that C<< 2**ld <= x < |
|
|
436 | 2**(ld+1) >>). If C<x> is 0 the result is undefined. A common use case is |
|
|
437 | to compute the integer binary logarithm, i.e. C<floor (log2 (n))>, for |
|
|
438 | example to see how many bits a certain number requires to be encoded. |
|
|
439 | |
|
|
440 | This function is similar to the "count leading zero bits" function, except |
|
|
441 | that that one returns how many zero bits are "in front" of the number (in |
|
|
442 | the given data type), while C<ecb_ld> returns how many bits the number |
|
|
443 | itself requires. |
|
|
444 | |
|
|
445 | For smaller types than C<uint32_t> you can safely use C<ecb_ld32>. |
|
|
446 | |
405 | =item int ecb_popcount32 (uint32_t x) |
447 | =item int ecb_popcount32 (uint32_t x) |
406 | |
448 | |
|
|
449 | =item int ecb_popcount64 (uint64_t x) |
|
|
450 | |
407 | Returns the number of bits set to 1 in C<x>. For example: |
451 | Returns the number of bits set to 1 in C<x>. |
|
|
452 | |
|
|
453 | For smaller types than C<uint32_t> you can safely use C<ecb_popcount32>. |
|
|
454 | |
|
|
455 | For example: |
408 | |
456 | |
409 | ecb_popcount32 (7) = 3 |
457 | ecb_popcount32 (7) = 3 |
410 | ecb_popcount32 (255) = 8 |
458 | ecb_popcount32 (255) = 8 |
411 | |
459 | |
412 | =item uint32_t ecb_bswap16 (uint32_t x) |
460 | =item uint32_t ecb_bswap16 (uint32_t x) |
413 | |
461 | |
414 | =item uint32_t ecb_bswap32 (uint32_t x) |
462 | =item uint32_t ecb_bswap32 (uint32_t x) |
415 | |
463 | |
|
|
464 | =item uint64_t ecb_bswap64 (uint64_t x) |
|
|
465 | |
416 | These two functions return the value of the 16-bit (32-bit) value C<x> |
466 | These functions return the value of the 16-bit (32-bit, 64-bit) value |
417 | after reversing the order of bytes (0x11223344 becomes 0x44332211). |
467 | C<x> after reversing the order of bytes (0x11223344 becomes 0x44332211 in |
|
|
468 | C<ecb_bswap32>). |
|
|
469 | |
|
|
470 | =item uint8_t ecb_rotl8 (uint8_t x, unsigned int count) |
|
|
471 | |
|
|
472 | =item uint16_t ecb_rotl16 (uint16_t x, unsigned int count) |
|
|
473 | |
|
|
474 | =item uint32_t ecb_rotl32 (uint32_t x, unsigned int count) |
|
|
475 | |
|
|
476 | =item uint64_t ecb_rotl64 (uint64_t x, unsigned int count) |
|
|
477 | |
|
|
478 | =item uint8_t ecb_rotr8 (uint8_t x, unsigned int count) |
|
|
479 | |
|
|
480 | =item uint16_t ecb_rotr16 (uint16_t x, unsigned int count) |
418 | |
481 | |
419 | =item uint32_t ecb_rotr32 (uint32_t x, unsigned int count) |
482 | =item uint32_t ecb_rotr32 (uint32_t x, unsigned int count) |
420 | |
483 | |
421 | =item uint32_t ecb_rotl32 (uint32_t x, unsigned int count) |
484 | =item uint64_t ecb_rotr64 (uint64_t x, unsigned int count) |
422 | |
485 | |
423 | These two functions return the value of C<x> after rotating all the bits |
486 | These two families of functions return the value of C<x> after rotating |
424 | by C<count> positions to the right or left respectively. |
487 | all the bits by C<count> positions to the right (C<ecb_rotr>) or left |
|
|
488 | (C<ecb_rotl>). |
425 | |
489 | |
426 | Current GCC versions understand these functions and usually compile them |
490 | Current GCC versions understand these functions and usually compile them |
427 | to "optimal" code (e.g. a single C<roll> on x86). |
491 | to "optimal" code (e.g. a single C<rol> or a combination of C<shld> on |
|
|
492 | x86). |
428 | |
493 | |
429 | =back |
494 | =back |
430 | |
495 | |
431 | =head2 ARITHMETIC |
496 | =head2 ARITHMETIC |
432 | |
497 | |
433 | =over 4 |
498 | =over 4 |
434 | |
499 | |
435 | =item x = ecb_mod (m, n) |
500 | =item x = ecb_mod (m, n) |
436 | |
501 | |
437 | Returns the positive remainder of the modulo operation between C<m> and |
502 | Returns C<m> modulo C<n>, which is the same as the positive remainder |
438 | C<n>, using floored division. Unlike the C modulo operator C<%>, this |
503 | of the division operation between C<m> and C<n>, using floored |
439 | function ensures that the return value is always positive and that the two |
504 | division. Unlike the C remainder operator C<%>, this function ensures that |
|
|
505 | the return value is always positive and that the two numbers I<m> and |
440 | numbers I<m> and I<m' = m + i * n> result in the same value modulo I<n> - |
506 | I<m' = m + i * n> result in the same value modulo I<n> - in other words, |
441 | the C<%> operator usually has a behaviour change at C<m = 0>. |
507 | C<ecb_mod> implements the mathematical modulo operation, which is missing |
|
|
508 | in the language. |
442 | |
509 | |
443 | C<n> must be strictly positive (i.e. C<< >= 1 >>), while C<m> must be |
510 | C<n> must be strictly positive (i.e. C<< >= 1 >>), while C<m> must be |
444 | negatable, that is, both C<m> and C<-m> must be representable in its |
511 | negatable, that is, both C<m> and C<-m> must be representable in its |
445 | type. |
512 | type (this typically excludes the minimum signed integer value, the same |
|
|
513 | limitation as for C</> and C<%> in C). |
446 | |
514 | |
447 | Current GCC versions compile this into an efficient branchless sequence on |
515 | Current GCC versions compile this into an efficient branchless sequence on |
448 | many systems. |
516 | almost all CPUs. |
449 | |
517 | |
450 | For example, when you want to rotate forward through the members of an |
518 | For example, when you want to rotate forward through the members of an |
451 | array for increasing C<m> (which might be negative), then you should use |
519 | array for increasing C<m> (which might be negative), then you should use |
452 | C<ecb_mod>, as the C<%> operator might give either negative results, or |
520 | C<ecb_mod>, as the C<%> operator might give either negative results, or |
453 | change direction for negative values: |
521 | change direction for negative values: |
454 | |
522 | |
455 | for (m = -100; m <= 100; ++m) |
523 | for (m = -100; m <= 100; ++m) |
456 | int elem = myarray [ecb_mod (m, ecb_array_length (myarray))]; |
524 | int elem = myarray [ecb_mod (m, ecb_array_length (myarray))]; |
457 | |
525 | |
|
|
526 | =item x = ecb_div_rd (val, div) |
|
|
527 | |
|
|
528 | =item x = ecb_div_ru (val, div) |
|
|
529 | |
|
|
530 | Returns C<val> divided by C<div> rounded down or up, respectively. |
|
|
531 | C<val> and C<div> must have integer types and C<div> must be strictly |
|
|
532 | positive. |
|
|
533 | |
458 | =back |
534 | =back |
459 | |
535 | |
460 | =head2 UTILITY |
536 | =head2 UTILITY |
461 | |
537 | |
462 | =over 4 |
538 | =over 4 |