]>
Commit | Line | Data |
---|---|---|
609d99a3 MCC |
1 | .. _codingstyle: |
2 | ||
d8dbbbc5 MCC |
3 | Linux kernel coding style |
4 | ========================= | |
1da177e4 LT |
5 | |
6 | This is a short document describing the preferred coding style for the | |
5d628b45 | 7 | linux kernel. Coding style is very personal, and I won't **force** my |
1da177e4 LT |
8 | views on anybody, but this is what goes for anything that I have to be |
9 | able to maintain, and I'd prefer it for most other things too. Please | |
10 | at least consider the points made here. | |
11 | ||
12 | First off, I'd suggest printing out a copy of the GNU coding standards, | |
13 | and NOT read it. Burn them, it's a great symbolic gesture. | |
14 | ||
15 | Anyway, here goes: | |
16 | ||
17 | ||
d8dbbbc5 MCC |
18 | 1) Indentation |
19 | -------------- | |
1da177e4 LT |
20 | |
21 | Tabs are 8 characters, and thus indentations are also 8 characters. | |
22 | There are heretic movements that try to make indentations 4 (or even 2!) | |
23 | characters deep, and that is akin to trying to define the value of PI to | |
24 | be 3. | |
25 | ||
26 | Rationale: The whole idea behind indentation is to clearly define where | |
27 | a block of control starts and ends. Especially when you've been looking | |
28 | at your screen for 20 straight hours, you'll find it a lot easier to see | |
29 | how the indentation works if you have large indentations. | |
30 | ||
31 | Now, some people will claim that having 8-character indentations makes | |
32 | the code move too far to the right, and makes it hard to read on a | |
33 | 80-character terminal screen. The answer to that is that if you need | |
34 | more than 3 levels of indentation, you're screwed anyway, and should fix | |
35 | your program. | |
36 | ||
37 | In short, 8-char indents make things easier to read, and have the added | |
38 | benefit of warning you when you're nesting your functions too deep. | |
39 | Heed that warning. | |
40 | ||
b3fc9941 | 41 | The preferred way to ease multiple indentation levels in a switch statement is |
b1a3459b MCC |
42 | to align the ``switch`` and its subordinate ``case`` labels in the same column |
43 | instead of ``double-indenting`` the ``case`` labels. E.g.: | |
b3fc9941 | 44 | |
d8dbbbc5 MCC |
45 | .. code-block:: c |
46 | ||
b3fc9941 RD |
47 | switch (suffix) { |
48 | case 'G': | |
49 | case 'g': | |
50 | mem <<= 30; | |
51 | break; | |
52 | case 'M': | |
53 | case 'm': | |
54 | mem <<= 20; | |
55 | break; | |
56 | case 'K': | |
57 | case 'k': | |
58 | mem <<= 10; | |
59 | /* fall through */ | |
60 | default: | |
61 | break; | |
62 | } | |
63 | ||
1da177e4 LT |
64 | Don't put multiple statements on a single line unless you have |
65 | something to hide: | |
66 | ||
d8dbbbc5 MCC |
67 | .. code-block:: c |
68 | ||
1da177e4 LT |
69 | if (condition) do_this; |
70 | do_something_everytime; | |
71 | ||
b3fc9941 RD |
72 | Don't put multiple assignments on a single line either. Kernel coding style |
73 | is super simple. Avoid tricky expressions. | |
74 | ||
1da177e4 LT |
75 | Outside of comments, documentation and except in Kconfig, spaces are never |
76 | used for indentation, and the above example is deliberately broken. | |
77 | ||
78 | Get a decent editor and don't leave whitespace at the end of lines. | |
79 | ||
80 | ||
d8dbbbc5 MCC |
81 | 2) Breaking long lines and strings |
82 | ---------------------------------- | |
1da177e4 LT |
83 | |
84 | Coding style is all about readability and maintainability using commonly | |
85 | available tools. | |
86 | ||
dff4982f AC |
87 | The limit on the length of lines is 80 columns and this is a strongly |
88 | preferred limit. | |
1da177e4 | 89 | |
6f76b6fc JT |
90 | Statements longer than 80 columns will be broken into sensible chunks, unless |
91 | exceeding 80 columns significantly increases readability and does not hide | |
92 | information. Descendants are always substantially shorter than the parent and | |
93 | are placed substantially to the right. The same applies to function headers | |
94 | with a long argument list. However, never break user-visible strings such as | |
95 | printk messages, because that breaks the ability to grep for them. | |
96 | ||
1da177e4 | 97 | |
d8dbbbc5 MCC |
98 | 3) Placing Braces and Spaces |
99 | ---------------------------- | |
1da177e4 LT |
100 | |
101 | The other issue that always comes up in C styling is the placement of | |
102 | braces. Unlike the indent size, there are few technical reasons to | |
103 | choose one placement strategy over the other, but the preferred way, as | |
104 | shown to us by the prophets Kernighan and Ritchie, is to put the opening | |
105 | brace last on the line, and put the closing brace first, thusly: | |
106 | ||
d8dbbbc5 MCC |
107 | .. code-block:: c |
108 | ||
1da177e4 LT |
109 | if (x is true) { |
110 | we do y | |
111 | } | |
112 | ||
b3fc9941 RD |
113 | This applies to all non-function statement blocks (if, switch, for, |
114 | while, do). E.g.: | |
115 | ||
d8dbbbc5 MCC |
116 | .. code-block:: c |
117 | ||
b3fc9941 RD |
118 | switch (action) { |
119 | case KOBJ_ADD: | |
120 | return "add"; | |
121 | case KOBJ_REMOVE: | |
122 | return "remove"; | |
123 | case KOBJ_CHANGE: | |
124 | return "change"; | |
125 | default: | |
126 | return NULL; | |
127 | } | |
128 | ||
1da177e4 LT |
129 | However, there is one special case, namely functions: they have the |
130 | opening brace at the beginning of the next line, thus: | |
131 | ||
d8dbbbc5 MCC |
132 | .. code-block:: c |
133 | ||
1da177e4 LT |
134 | int function(int x) |
135 | { | |
136 | body of function | |
137 | } | |
138 | ||
139 | Heretic people all over the world have claimed that this inconsistency | |
140 | is ... well ... inconsistent, but all right-thinking people know that | |
5d628b45 | 141 | (a) K&R are **right** and (b) K&R are right. Besides, functions are |
1da177e4 LT |
142 | special anyway (you can't nest them in C). |
143 | ||
5d628b45 | 144 | Note that the closing brace is empty on a line of its own, **except** in |
1da177e4 | 145 | the cases where it is followed by a continuation of the same statement, |
b1a3459b | 146 | ie a ``while`` in a do-statement or an ``else`` in an if-statement, like |
1da177e4 LT |
147 | this: |
148 | ||
d8dbbbc5 MCC |
149 | .. code-block:: c |
150 | ||
1da177e4 LT |
151 | do { |
152 | body of do-loop | |
153 | } while (condition); | |
154 | ||
155 | and | |
156 | ||
d8dbbbc5 MCC |
157 | .. code-block:: c |
158 | ||
1da177e4 LT |
159 | if (x == y) { |
160 | .. | |
161 | } else if (x > y) { | |
162 | ... | |
163 | } else { | |
164 | .... | |
165 | } | |
166 | ||
167 | Rationale: K&R. | |
168 | ||
169 | Also, note that this brace-placement also minimizes the number of empty | |
170 | (or almost empty) lines, without any loss of readability. Thus, as the | |
171 | supply of new-lines on your screen is not a renewable resource (think | |
172 | 25-line terminal screens here), you have more empty lines to put | |
173 | comments on. | |
174 | ||
e659ba4a ON |
175 | Do not unnecessarily use braces where a single statement will do. |
176 | ||
d8dbbbc5 MCC |
177 | .. code-block:: c |
178 | ||
09677e0f PK |
179 | if (condition) |
180 | action(); | |
e659ba4a | 181 | |
38829dc9 HW |
182 | and |
183 | ||
d8dbbbc5 MCC |
184 | .. code-block:: none |
185 | ||
09677e0f PK |
186 | if (condition) |
187 | do_this(); | |
188 | else | |
189 | do_that(); | |
38829dc9 | 190 | |
b218ab0a AO |
191 | This does not apply if only one branch of a conditional statement is a single |
192 | statement; in the latter case use braces in both branches: | |
e659ba4a | 193 | |
d8dbbbc5 MCC |
194 | .. code-block:: c |
195 | ||
09677e0f PK |
196 | if (condition) { |
197 | do_this(); | |
198 | do_that(); | |
199 | } else { | |
200 | otherwise(); | |
201 | } | |
e659ba4a | 202 | |
1dbba2cf GH |
203 | Also, use braces when a loop contains more than a single simple statement: |
204 | ||
205 | .. code-block:: c | |
206 | ||
207 | while (condition) { | |
208 | if (test) | |
209 | do_something(); | |
210 | } | |
211 | ||
d8dbbbc5 MCC |
212 | 3.1) Spaces |
213 | *********** | |
b3fc9941 RD |
214 | |
215 | Linux kernel style for use of spaces depends (mostly) on | |
216 | function-versus-keyword usage. Use a space after (most) keywords. The | |
217 | notable exceptions are sizeof, typeof, alignof, and __attribute__, which look | |
218 | somewhat like functions (and are usually used with parentheses in Linux, | |
d8dbbbc5 MCC |
219 | although they are not required in the language, as in: ``sizeof info`` after |
220 | ``struct fileinfo info;`` is declared). | |
b3fc9941 | 221 | |
d8dbbbc5 | 222 | So use a space after these keywords:: |
09677e0f | 223 | |
b3fc9941 | 224 | if, switch, case, for, do, while |
09677e0f | 225 | |
b3fc9941 | 226 | but not with sizeof, typeof, alignof, or __attribute__. E.g., |
09677e0f | 227 | |
d8dbbbc5 MCC |
228 | .. code-block:: c |
229 | ||
230 | ||
b3fc9941 RD |
231 | s = sizeof(struct file); |
232 | ||
233 | Do not add spaces around (inside) parenthesized expressions. This example is | |
d8dbbbc5 MCC |
234 | **bad**: |
235 | ||
236 | .. code-block:: c | |
237 | ||
b3fc9941 RD |
238 | |
239 | s = sizeof( struct file ); | |
240 | ||
241 | When declaring pointer data or a function that returns a pointer type, the | |
b1a3459b | 242 | preferred use of ``*`` is adjacent to the data name or function name and not |
b3fc9941 RD |
243 | adjacent to the type name. Examples: |
244 | ||
d8dbbbc5 MCC |
245 | .. code-block:: c |
246 | ||
247 | ||
b3fc9941 RD |
248 | char *linux_banner; |
249 | unsigned long long memparse(char *ptr, char **retptr); | |
250 | char *match_strdup(substring_t *s); | |
251 | ||
252 | Use one space around (on each side of) most binary and ternary operators, | |
d8dbbbc5 | 253 | such as any of these:: |
b3fc9941 RD |
254 | |
255 | = + - < > * / % | & ^ <= >= == != ? : | |
256 | ||
d8dbbbc5 | 257 | but no space after unary operators:: |
09677e0f | 258 | |
b3fc9941 RD |
259 | & * + - ~ ! sizeof typeof alignof __attribute__ defined |
260 | ||
d8dbbbc5 | 261 | no space before the postfix increment & decrement unary operators:: |
09677e0f | 262 | |
b3fc9941 RD |
263 | ++ -- |
264 | ||
d8dbbbc5 | 265 | no space after the prefix increment & decrement unary operators:: |
09677e0f | 266 | |
b3fc9941 RD |
267 | ++ -- |
268 | ||
b1a3459b | 269 | and no space around the ``.`` and ``->`` structure member operators. |
b3fc9941 | 270 | |
a923fd6a | 271 | Do not leave trailing whitespace at the ends of lines. Some editors with |
b1a3459b | 272 | ``smart`` indentation will insert whitespace at the beginning of new lines as |
a923fd6a JT |
273 | appropriate, so you can start typing the next line of code right away. |
274 | However, some such editors do not remove the whitespace if you end up not | |
275 | putting a line of code there, such as if you leave a blank line. As a result, | |
276 | you end up with lines containing trailing whitespace. | |
277 | ||
278 | Git will warn you about patches that introduce trailing whitespace, and can | |
279 | optionally strip the trailing whitespace for you; however, if applying a series | |
280 | of patches, this may make later patches in the series fail by changing their | |
281 | context lines. | |
282 | ||
1da177e4 | 283 | |
d8dbbbc5 MCC |
284 | 4) Naming |
285 | --------- | |
1da177e4 LT |
286 | |
287 | C is a Spartan language, and so should your naming be. Unlike Modula-2 | |
288 | and Pascal programmers, C programmers do not use cute names like | |
289 | ThisVariableIsATemporaryCounter. A C programmer would call that | |
b1a3459b | 290 | variable ``tmp``, which is much easier to write, and not the least more |
1da177e4 LT |
291 | difficult to understand. |
292 | ||
293 | HOWEVER, while mixed-case names are frowned upon, descriptive names for | |
b1a3459b | 294 | global variables are a must. To call a global function ``foo`` is a |
1da177e4 LT |
295 | shooting offense. |
296 | ||
5d628b45 | 297 | GLOBAL variables (to be used only if you **really** need them) need to |
1da177e4 LT |
298 | have descriptive names, as do global functions. If you have a function |
299 | that counts the number of active users, you should call that | |
5d628b45 | 300 | ``count_active_users()`` or similar, you should **not** call it ``cntusr()``. |
1da177e4 LT |
301 | |
302 | Encoding the type of a function into the name (so-called Hungarian | |
303 | notation) is brain damaged - the compiler knows the types anyway and can | |
304 | check those, and it only confuses the programmer. No wonder MicroSoft | |
305 | makes buggy programs. | |
306 | ||
307 | LOCAL variable names should be short, and to the point. If you have | |
b1a3459b MCC |
308 | some random integer loop counter, it should probably be called ``i``. |
309 | Calling it ``loop_counter`` is non-productive, if there is no chance of it | |
310 | being mis-understood. Similarly, ``tmp`` can be just about any type of | |
1da177e4 LT |
311 | variable that is used to hold a temporary value. |
312 | ||
313 | If you are afraid to mix up your local variable names, you have another | |
314 | problem, which is called the function-growth-hormone-imbalance syndrome. | |
b3fc9941 | 315 | See chapter 6 (Functions). |
1da177e4 LT |
316 | |
317 | ||
d8dbbbc5 MCC |
318 | 5) Typedefs |
319 | ----------- | |
226a6b84 | 320 | |
b1a3459b | 321 | Please don't use things like ``vps_t``. |
5d628b45 | 322 | It's a **mistake** to use typedef for structures and pointers. When you see a |
226a6b84 | 323 | |
d8dbbbc5 MCC |
324 | .. code-block:: c |
325 | ||
326 | ||
226a6b84 RD |
327 | vps_t a; |
328 | ||
329 | in the source, what does it mean? | |
226a6b84 RD |
330 | In contrast, if it says |
331 | ||
d8dbbbc5 MCC |
332 | .. code-block:: c |
333 | ||
226a6b84 RD |
334 | struct virtual_container *a; |
335 | ||
b1a3459b | 336 | you can actually tell what ``a`` is. |
226a6b84 | 337 | |
b1a3459b | 338 | Lots of people think that typedefs ``help readability``. Not so. They are |
226a6b84 RD |
339 | useful only for: |
340 | ||
5d628b45 | 341 | (a) totally opaque objects (where the typedef is actively used to **hide** |
226a6b84 RD |
342 | what the object is). |
343 | ||
b1a3459b | 344 | Example: ``pte_t`` etc. opaque objects that you can only access using |
226a6b84 RD |
345 | the proper accessor functions. |
346 | ||
3772ec4a MCC |
347 | .. note:: |
348 | ||
349 | Opaqueness and ``accessor functions`` are not good in themselves. | |
350 | The reason we have them for things like pte_t etc. is that there | |
351 | really is absolutely **zero** portably accessible information there. | |
226a6b84 | 352 | |
5d628b45 | 353 | (b) Clear integer types, where the abstraction **helps** avoid confusion |
b1a3459b | 354 | whether it is ``int`` or ``long``. |
226a6b84 RD |
355 | |
356 | u8/u16/u32 are perfectly fine typedefs, although they fit into | |
357 | category (d) better than here. | |
358 | ||
3772ec4a MCC |
359 | .. note:: |
360 | ||
361 | Again - there needs to be a **reason** for this. If something is | |
362 | ``unsigned long``, then there's no reason to do | |
226a6b84 RD |
363 | |
364 | typedef unsigned long myflags_t; | |
365 | ||
366 | but if there is a clear reason for why it under certain circumstances | |
b1a3459b MCC |
367 | might be an ``unsigned int`` and under other configurations might be |
368 | ``unsigned long``, then by all means go ahead and use a typedef. | |
226a6b84 | 369 | |
5d628b45 | 370 | (c) when you use sparse to literally create a **new** type for |
226a6b84 RD |
371 | type-checking. |
372 | ||
373 | (d) New types which are identical to standard C99 types, in certain | |
374 | exceptional circumstances. | |
375 | ||
376 | Although it would only take a short amount of time for the eyes and | |
b1a3459b | 377 | brain to become accustomed to the standard types like ``uint32_t``, |
226a6b84 RD |
378 | some people object to their use anyway. |
379 | ||
b1a3459b | 380 | Therefore, the Linux-specific ``u8/u16/u32/u64`` types and their |
226a6b84 RD |
381 | signed equivalents which are identical to standard types are |
382 | permitted -- although they are not mandatory in new code of your | |
383 | own. | |
384 | ||
385 | When editing existing code which already uses one or the other set | |
386 | of types, you should conform to the existing choices in that code. | |
387 | ||
388 | (e) Types safe for use in userspace. | |
389 | ||
390 | In certain structures which are visible to userspace, we cannot | |
b1a3459b | 391 | require C99 types and cannot use the ``u32`` form above. Thus, we |
226a6b84 RD |
392 | use __u32 and similar types in all structures which are shared |
393 | with userspace. | |
394 | ||
395 | Maybe there are other cases too, but the rule should basically be to NEVER | |
396 | EVER use a typedef unless you can clearly match one of those rules. | |
397 | ||
398 | In general, a pointer, or a struct that has elements that can reasonably | |
5d628b45 | 399 | be directly accessed should **never** be a typedef. |
226a6b84 RD |
400 | |
401 | ||
d8dbbbc5 MCC |
402 | 6) Functions |
403 | ------------ | |
1da177e4 LT |
404 | |
405 | Functions should be short and sweet, and do just one thing. They should | |
406 | fit on one or two screenfuls of text (the ISO/ANSI screen size is 80x24, | |
407 | as we all know), and do one thing and do that well. | |
408 | ||
409 | The maximum length of a function is inversely proportional to the | |
410 | complexity and indentation level of that function. So, if you have a | |
411 | conceptually simple function that is just one long (but simple) | |
412 | case-statement, where you have to do lots of small things for a lot of | |
413 | different cases, it's OK to have a longer function. | |
414 | ||
415 | However, if you have a complex function, and you suspect that a | |
416 | less-than-gifted first-year high-school student might not even | |
417 | understand what the function is all about, you should adhere to the | |
418 | maximum limits all the more closely. Use helper functions with | |
419 | descriptive names (you can ask the compiler to in-line them if you think | |
420 | it's performance-critical, and it will probably do a better job of it | |
421 | than you would have done). | |
422 | ||
423 | Another measure of the function is the number of local variables. They | |
424 | shouldn't exceed 5-10, or you're doing something wrong. Re-think the | |
425 | function, and split it into smaller pieces. A human brain can | |
426 | generally easily keep track of about 7 different things, anything more | |
427 | and it gets confused. You know you're brilliant, but maybe you'd like | |
428 | to understand what you did 2 weeks from now. | |
429 | ||
b3fc9941 | 430 | In source files, separate functions with one blank line. If the function is |
d8dbbbc5 MCC |
431 | exported, the **EXPORT** macro for it should follow immediately after the |
432 | closing function brace line. E.g.: | |
433 | ||
434 | .. code-block:: c | |
b3fc9941 | 435 | |
09677e0f PK |
436 | int system_is_up(void) |
437 | { | |
438 | return system_state == SYSTEM_RUNNING; | |
439 | } | |
440 | EXPORT_SYMBOL(system_is_up); | |
b3fc9941 RD |
441 | |
442 | In function prototypes, include parameter names with their data types. | |
443 | Although this is not required by the C language, it is preferred in Linux | |
444 | because it is a simple way to add valuable information for the reader. | |
445 | ||
1da177e4 | 446 | |
d8dbbbc5 MCC |
447 | 7) Centralized exiting of functions |
448 | ----------------------------------- | |
1da177e4 LT |
449 | |
450 | Albeit deprecated by some people, the equivalent of the goto statement is | |
451 | used frequently by compilers in form of the unconditional jump instruction. | |
452 | ||
453 | The goto statement comes in handy when a function exits from multiple | |
b57a0505 DC |
454 | locations and some common work such as cleanup has to be done. If there is no |
455 | cleanup needed then just return directly. | |
1da177e4 | 456 | |
ea040360 | 457 | Choose label names which say what the goto does or why the goto exists. An |
b1a3459b MCC |
458 | example of a good name could be ``out_free_buffer:`` if the goto frees ``buffer``. |
459 | Avoid using GW-BASIC names like ``err1:`` and ``err2:``, as you would have to | |
865a1caa JD |
460 | renumber them if you ever add or remove exit paths, and they make correctness |
461 | difficult to verify anyway. | |
462 | ||
ea040360 | 463 | The rationale for using gotos is: |
1da177e4 LT |
464 | |
465 | - unconditional statements are easier to understand and follow | |
466 | - nesting is reduced | |
467 | - errors by not updating individual exit points when making | |
d8dbbbc5 | 468 | modifications are prevented |
1da177e4 LT |
469 | - saves the compiler work to optimize redundant code away ;) |
470 | ||
d8dbbbc5 MCC |
471 | .. code-block:: c |
472 | ||
09677e0f PK |
473 | int fun(int a) |
474 | { | |
475 | int result = 0; | |
476 | char *buffer; | |
477 | ||
478 | buffer = kmalloc(SIZE, GFP_KERNEL); | |
479 | if (!buffer) | |
480 | return -ENOMEM; | |
481 | ||
482 | if (condition1) { | |
483 | while (loop1) { | |
484 | ... | |
485 | } | |
486 | result = 1; | |
beab6cb2 | 487 | goto out_free_buffer; |
1da177e4 | 488 | } |
09677e0f | 489 | ... |
79c70c30 | 490 | out_free_buffer: |
09677e0f PK |
491 | kfree(buffer); |
492 | return result; | |
1da177e4 | 493 | } |
1da177e4 | 494 | |
b1a3459b | 495 | A common type of bug to be aware of is ``one err bugs`` which look like this: |
ea040360 | 496 | |
d8dbbbc5 MCC |
497 | .. code-block:: c |
498 | ||
79c70c30 | 499 | err: |
09677e0f PK |
500 | kfree(foo->bar); |
501 | kfree(foo); | |
502 | return ret; | |
ea040360 | 503 | |
b1a3459b MCC |
504 | The bug in this code is that on some exit paths ``foo`` is NULL. Normally the |
505 | fix for this is to split it up into two error labels ``err_free_bar:`` and | |
506 | ``err_free_foo:``: | |
865a1caa | 507 | |
d8dbbbc5 MCC |
508 | .. code-block:: c |
509 | ||
865a1caa JD |
510 | err_free_bar: |
511 | kfree(foo->bar); | |
512 | err_free_foo: | |
513 | kfree(foo); | |
514 | return ret; | |
515 | ||
516 | Ideally you should simulate errors to test all exit paths. | |
ea040360 DC |
517 | |
518 | ||
d8dbbbc5 MCC |
519 | 8) Commenting |
520 | ------------- | |
1da177e4 LT |
521 | |
522 | Comments are good, but there is also a danger of over-commenting. NEVER | |
523 | try to explain HOW your code works in a comment: it's much better to | |
5d628b45 | 524 | write the code so that the **working** is obvious, and it's a waste of |
1da177e4 LT |
525 | time to explain badly written code. |
526 | ||
527 | Generally, you want your comments to tell WHAT your code does, not HOW. | |
528 | Also, try to avoid putting comments inside a function body: if the | |
529 | function is so complex that you need to separately comment parts of it, | |
b3fc9941 | 530 | you should probably go back to chapter 6 for a while. You can make |
1da177e4 LT |
531 | small comments to note or warn about something particularly clever (or |
532 | ugly), but try to avoid excess. Instead, put the comments at the head | |
533 | of the function, telling people what it does, and possibly WHY it does | |
534 | it. | |
535 | ||
b3fc9941 | 536 | When commenting the kernel API functions, please use the kernel-doc format. |
1dc4bbf0 MCC |
537 | See the files at :ref:`Documentation/doc-guide/ <doc_guide>` and |
538 | ``scripts/kernel-doc`` for details. | |
1da177e4 | 539 | |
b3fc9941 RD |
540 | The preferred style for long (multi-line) comments is: |
541 | ||
d8dbbbc5 MCC |
542 | .. code-block:: c |
543 | ||
b3fc9941 RD |
544 | /* |
545 | * This is the preferred style for multi-line | |
546 | * comments in the Linux kernel source code. | |
547 | * Please use it consistently. | |
548 | * | |
549 | * Description: A column of asterisks on the left side, | |
550 | * with beginning and ending almost-blank lines. | |
551 | */ | |
552 | ||
c4ff1b5f JP |
553 | For files in net/ and drivers/net/ the preferred style for long (multi-line) |
554 | comments is a little different. | |
555 | ||
d8dbbbc5 MCC |
556 | .. code-block:: c |
557 | ||
c4ff1b5f JP |
558 | /* The preferred comment style for files in net/ and drivers/net |
559 | * looks like this. | |
560 | * | |
561 | * It is nearly the same as the generally preferred comment style, | |
562 | * but there is no initial almost-blank line. | |
563 | */ | |
564 | ||
b3fc9941 RD |
565 | It's also important to comment data, whether they are basic types or derived |
566 | types. To this end, use just one data declaration per line (no commas for | |
567 | multiple data declarations). This leaves you room for a small comment on each | |
568 | item, explaining its use. | |
569 | ||
570 | ||
d8dbbbc5 MCC |
571 | 9) You've made a mess of it |
572 | --------------------------- | |
1da177e4 LT |
573 | |
574 | That's OK, we all do. You've probably been told by your long-time Unix | |
b1a3459b | 575 | user helper that ``GNU emacs`` automatically formats the C sources for |
1da177e4 LT |
576 | you, and you've noticed that yes, it does do that, but the defaults it |
577 | uses are less than desirable (in fact, they are worse than random | |
578 | typing - an infinite number of monkeys typing into GNU emacs would never | |
579 | make a good program). | |
580 | ||
581 | So, you can either get rid of GNU emacs, or change it to use saner | |
582 | values. To do the latter, you can stick the following in your .emacs file: | |
583 | ||
d8dbbbc5 MCC |
584 | .. code-block:: none |
585 | ||
586 | (defun c-lineup-arglist-tabs-only (ignored) | |
587 | "Line up argument lists by tabs, not spaces" | |
588 | (let* ((anchor (c-langelem-pos c-syntactic-element)) | |
589 | (column (c-langelem-2nd-pos c-syntactic-element)) | |
590 | (offset (- (1+ column) anchor)) | |
591 | (steps (floor offset c-basic-offset))) | |
592 | (* (max steps 1) | |
593 | c-basic-offset))) | |
594 | ||
595 | (add-hook 'c-mode-common-hook | |
596 | (lambda () | |
597 | ;; Add kernel style | |
598 | (c-add-style | |
599 | "linux-tabs-only" | |
600 | '("linux" (c-offsets-alist | |
601 | (arglist-cont-nonempty | |
602 | c-lineup-gcc-asm-reg | |
603 | c-lineup-arglist-tabs-only)))))) | |
604 | ||
605 | (add-hook 'c-mode-hook | |
606 | (lambda () | |
607 | (let ((filename (buffer-file-name))) | |
608 | ;; Enable kernel mode for the appropriate files | |
609 | (when (and filename | |
610 | (string-match (expand-file-name "~/src/linux-trees") | |
611 | filename)) | |
612 | (setq indent-tabs-mode t) | |
613 | (setq show-trailing-whitespace t) | |
614 | (c-set-style "linux-tabs-only"))))) | |
a7f371e5 JW |
615 | |
616 | This will make emacs go better with the kernel coding style for C | |
d8dbbbc5 | 617 | files below ``~/src/linux-trees``. |
1da177e4 LT |
618 | |
619 | But even if you fail in getting emacs to do sane formatting, not | |
b1a3459b | 620 | everything is lost: use ``indent``. |
1da177e4 LT |
621 | |
622 | Now, again, GNU indent has the same brain-dead settings that GNU emacs | |
623 | has, which is why you need to give it a few command line options. | |
624 | However, that's not too bad, because even the makers of GNU indent | |
625 | recognize the authority of K&R (the GNU people aren't evil, they are | |
626 | just severely misguided in this matter), so you just give indent the | |
b1a3459b MCC |
627 | options ``-kr -i8`` (stands for ``K&R, 8 character indents``), or use |
628 | ``scripts/Lindent``, which indents in the latest style. | |
1da177e4 | 629 | |
b1a3459b | 630 | ``indent`` has a lot of options, and especially when it comes to comment |
1da177e4 | 631 | re-formatting you may want to take a look at the man page. But |
b1a3459b | 632 | remember: ``indent`` is not a fix for bad programming. |
1da177e4 | 633 | |
d4ef8d3f MO |
634 | Note that you can also use the ``clang-format`` tool to help you with |
635 | these rules, to quickly re-format parts of your code automatically, | |
636 | and to review full files in order to spot coding style mistakes, | |
637 | typos and possible improvements. It is also handy for sorting ``#includes``, | |
638 | for aligning variables/macros, for reflowing text and other similar tasks. | |
639 | See the file :ref:`Documentation/process/clang-format.rst <clangformat>` | |
640 | for more details. | |
641 | ||
1da177e4 | 642 | |
d8dbbbc5 MCC |
643 | 10) Kconfig configuration files |
644 | ------------------------------- | |
1da177e4 | 645 | |
6754bb4d | 646 | For all of the Kconfig* configuration files throughout the source tree, |
b1a3459b | 647 | the indentation is somewhat different. Lines under a ``config`` definition |
6754bb4d | 648 | are indented with one tab, while help text is indented an additional two |
d8dbbbc5 | 649 | spaces. Example:: |
1da177e4 | 650 | |
d8dbbbc5 | 651 | config AUDIT |
6754bb4d RD |
652 | bool "Auditing support" |
653 | depends on NET | |
1da177e4 | 654 | help |
6754bb4d RD |
655 | Enable auditing infrastructure that can be used with another |
656 | kernel subsystem, such as SELinux (which requires this for | |
657 | logging of avc messages output). Does not do system-call | |
658 | auditing without CONFIG_AUDITSYSCALL. | |
659 | ||
0335cb46 | 660 | Seriously dangerous features (such as write support for certain |
d8dbbbc5 | 661 | filesystems) should advertise this prominently in their prompt string:: |
6754bb4d | 662 | |
d8dbbbc5 | 663 | config ADFS_FS_RW |
6754bb4d RD |
664 | bool "ADFS write support (DANGEROUS)" |
665 | depends on ADFS_FS | |
666 | ... | |
1da177e4 | 667 | |
6754bb4d RD |
668 | For full documentation on the configuration files, see the file |
669 | Documentation/kbuild/kconfig-language.txt. | |
1da177e4 LT |
670 | |
671 | ||
d8dbbbc5 MCC |
672 | 11) Data structures |
673 | ------------------- | |
1da177e4 LT |
674 | |
675 | Data structures that have visibility outside the single-threaded | |
676 | environment they are created and destroyed in should always have | |
677 | reference counts. In the kernel, garbage collection doesn't exist (and | |
678 | outside the kernel garbage collection is slow and inefficient), which | |
5d628b45 | 679 | means that you absolutely **have** to reference count all your uses. |
1da177e4 LT |
680 | |
681 | Reference counting means that you can avoid locking, and allows multiple | |
682 | users to have access to the data structure in parallel - and not having | |
683 | to worry about the structure suddenly going away from under them just | |
684 | because they slept or did something else for a while. | |
685 | ||
5d628b45 | 686 | Note that locking is **not** a replacement for reference counting. |
1da177e4 LT |
687 | Locking is used to keep data structures coherent, while reference |
688 | counting is a memory management technique. Usually both are needed, and | |
689 | they are not to be confused with each other. | |
690 | ||
691 | Many data structures can indeed have two levels of reference counting, | |
b1a3459b | 692 | when there are users of different ``classes``. The subclass count counts |
1da177e4 LT |
693 | the number of subclass users, and decrements the global count just once |
694 | when the subclass count goes to zero. | |
695 | ||
b1a3459b MCC |
696 | Examples of this kind of ``multi-level-reference-counting`` can be found in |
697 | memory management (``struct mm_struct``: mm_users and mm_count), and in | |
698 | filesystem code (``struct super_block``: s_count and s_active). | |
1da177e4 LT |
699 | |
700 | Remember: if another thread can find your data structure, and you don't | |
701 | have a reference count on it, you almost certainly have a bug. | |
702 | ||
703 | ||
d8dbbbc5 MCC |
704 | 12) Macros, Enums and RTL |
705 | ------------------------- | |
1da177e4 LT |
706 | |
707 | Names of macros defining constants and labels in enums are capitalized. | |
708 | ||
d8dbbbc5 MCC |
709 | .. code-block:: c |
710 | ||
09677e0f | 711 | #define CONSTANT 0x12345 |
1da177e4 LT |
712 | |
713 | Enums are preferred when defining several related constants. | |
714 | ||
715 | CAPITALIZED macro names are appreciated but macros resembling functions | |
716 | may be named in lower case. | |
717 | ||
718 | Generally, inline functions are preferable to macros resembling functions. | |
719 | ||
720 | Macros with multiple statements should be enclosed in a do - while block: | |
721 | ||
d8dbbbc5 MCC |
722 | .. code-block:: c |
723 | ||
724 | #define macrofun(a, b, c) \ | |
09677e0f PK |
725 | do { \ |
726 | if (a == 5) \ | |
727 | do_this(b, c); \ | |
728 | } while (0) | |
1da177e4 LT |
729 | |
730 | Things to avoid when using macros: | |
731 | ||
732 | 1) macros that affect control flow: | |
733 | ||
d8dbbbc5 MCC |
734 | .. code-block:: c |
735 | ||
09677e0f PK |
736 | #define FOO(x) \ |
737 | do { \ | |
738 | if (blah(x) < 0) \ | |
739 | return -EBUGGERED; \ | |
32fd52d5 | 740 | } while (0) |
1da177e4 | 741 | |
5d628b45 | 742 | is a **very** bad idea. It looks like a function call but exits the ``calling`` |
1da177e4 LT |
743 | function; don't break the internal parsers of those who will read the code. |
744 | ||
745 | 2) macros that depend on having a local variable with a magic name: | |
746 | ||
d8dbbbc5 MCC |
747 | .. code-block:: c |
748 | ||
09677e0f | 749 | #define FOO(val) bar(index, val) |
1da177e4 LT |
750 | |
751 | might look like a good thing, but it's confusing as hell when one reads the | |
752 | code and it's prone to breakage from seemingly innocent changes. | |
753 | ||
754 | 3) macros with arguments that are used as l-values: FOO(x) = y; will | |
755 | bite you if somebody e.g. turns FOO into an inline function. | |
756 | ||
757 | 4) forgetting about precedence: macros defining constants using expressions | |
758 | must enclose the expression in parentheses. Beware of similar issues with | |
759 | macros using parameters. | |
760 | ||
d8dbbbc5 MCC |
761 | .. code-block:: c |
762 | ||
09677e0f PK |
763 | #define CONSTANT 0x4000 |
764 | #define CONSTEXP (CONSTANT | 3) | |
1da177e4 | 765 | |
f2027543 BG |
766 | 5) namespace collisions when defining local variables in macros resembling |
767 | functions: | |
768 | ||
d8dbbbc5 MCC |
769 | .. code-block:: c |
770 | ||
771 | #define FOO(x) \ | |
772 | ({ \ | |
773 | typeof(x) ret; \ | |
774 | ret = calc_ret(x); \ | |
775 | (ret); \ | |
776 | }) | |
f2027543 BG |
777 | |
778 | ret is a common name for a local variable - __foo_ret is less likely | |
779 | to collide with an existing variable. | |
780 | ||
1da177e4 LT |
781 | The cpp manual deals with macros exhaustively. The gcc internals manual also |
782 | covers RTL which is used frequently with assembly language in the kernel. | |
783 | ||
784 | ||
d8dbbbc5 MCC |
785 | 13) Printing kernel messages |
786 | ---------------------------- | |
1da177e4 LT |
787 | |
788 | Kernel developers like to be seen as literate. Do mind the spelling | |
789 | of kernel messages to make a good impression. Do not use crippled | |
b1a3459b | 790 | words like ``dont``; use ``do not`` or ``don't`` instead. Make the messages |
6b09448a | 791 | concise, clear, and unambiguous. |
1da177e4 LT |
792 | |
793 | Kernel messages do not have to be terminated with a period. | |
794 | ||
795 | Printing numbers in parentheses (%d) adds no value and should be avoided. | |
796 | ||
6b09448a DB |
797 | There are a number of driver model diagnostic macros in <linux/device.h> |
798 | which you should use to make sure messages are matched to the right device | |
799 | and driver, and are tagged with the right level: dev_err(), dev_warn(), | |
800 | dev_info(), and so forth. For messages that aren't associated with a | |
6e099f55 DS |
801 | particular device, <linux/printk.h> defines pr_notice(), pr_info(), |
802 | pr_warn(), pr_err(), etc. | |
6b09448a DB |
803 | |
804 | Coming up with good debugging messages can be quite a challenge; and once | |
6e099f55 DS |
805 | you have them, they can be a huge help for remote troubleshooting. However |
806 | debug message printing is handled differently than printing other non-debug | |
807 | messages. While the other pr_XXX() functions print unconditionally, | |
808 | pr_debug() does not; it is compiled out by default, unless either DEBUG is | |
809 | defined or CONFIG_DYNAMIC_DEBUG is set. That is true for dev_dbg() also, | |
810 | and a related convention uses VERBOSE_DEBUG to add dev_vdbg() messages to | |
811 | the ones already enabled by DEBUG. | |
812 | ||
813 | Many subsystems have Kconfig debug options to turn on -DDEBUG in the | |
814 | corresponding Makefile; in other cases specific files #define DEBUG. And | |
815 | when a debug message should be unconditionally printed, such as if it is | |
7c18fd78 | 816 | already inside a debug-related #ifdef section, printk(KERN_DEBUG ...) can be |
6e099f55 | 817 | used. |
6b09448a | 818 | |
1da177e4 | 819 | |
d8dbbbc5 MCC |
820 | 14) Allocating memory |
821 | --------------------- | |
af4e5a21 PE |
822 | |
823 | The kernel provides the following general purpose memory allocators: | |
15837294 XW |
824 | kmalloc(), kzalloc(), kmalloc_array(), kcalloc(), vmalloc(), and |
825 | vzalloc(). Please refer to the API documentation for further information | |
826 | about them. | |
af4e5a21 PE |
827 | |
828 | The preferred form for passing a size of a struct is the following: | |
829 | ||
d8dbbbc5 MCC |
830 | .. code-block:: c |
831 | ||
af4e5a21 PE |
832 | p = kmalloc(sizeof(*p), ...); |
833 | ||
834 | The alternative form where struct name is spelled out hurts readability and | |
835 | introduces an opportunity for a bug when the pointer variable type is changed | |
836 | but the corresponding sizeof that is passed to a memory allocator is not. | |
837 | ||
838 | Casting the return value which is a void pointer is redundant. The conversion | |
839 | from void pointer to any other pointer type is guaranteed by the C programming | |
840 | language. | |
841 | ||
15837294 XW |
842 | The preferred form for allocating an array is the following: |
843 | ||
d8dbbbc5 MCC |
844 | .. code-block:: c |
845 | ||
15837294 XW |
846 | p = kmalloc_array(n, sizeof(...), ...); |
847 | ||
848 | The preferred form for allocating a zeroed array is the following: | |
849 | ||
d8dbbbc5 MCC |
850 | .. code-block:: c |
851 | ||
15837294 XW |
852 | p = kcalloc(n, sizeof(...), ...); |
853 | ||
854 | Both forms check for overflow on the allocation size n * sizeof(...), | |
855 | and return NULL if that occurred. | |
856 | ||
af4e5a21 | 857 | |
d8dbbbc5 MCC |
858 | 15) The inline disease |
859 | ---------------------- | |
a771f2b8 AV |
860 | |
861 | There appears to be a common misperception that gcc has a magic "make me | |
b1a3459b | 862 | faster" speedup option called ``inline``. While the use of inlines can be |
53ab97a1 | 863 | appropriate (for example as a means of replacing macros, see Chapter 12), it |
a771f2b8 AV |
864 | very often is not. Abundant use of the inline keyword leads to a much bigger |
865 | kernel, which in turn slows the system as a whole down, due to a bigger | |
866 | icache footprint for the CPU and simply because there is less memory | |
867 | available for the pagecache. Just think about it; a pagecache miss causes a | |
19af5cdb MO |
868 | disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles |
869 | that can go into these 5 milliseconds. | |
a771f2b8 AV |
870 | |
871 | A reasonable rule of thumb is to not put inline at functions that have more | |
872 | than 3 lines of code in them. An exception to this rule are the cases where | |
873 | a parameter is known to be a compiletime constant, and as a result of this | |
874 | constantness you *know* the compiler will be able to optimize most of your | |
875 | function away at compile time. For a good example of this later case, see | |
876 | the kmalloc() inline function. | |
877 | ||
878 | Often people argue that adding inline to functions that are static and used | |
879 | only once is always a win since there is no space tradeoff. While this is | |
880 | technically correct, gcc is capable of inlining these automatically without | |
881 | help, and the maintenance issue of removing the inline when a second user | |
882 | appears outweighs the potential value of the hint that tells gcc to do | |
883 | something it would have done anyway. | |
884 | ||
885 | ||
d8dbbbc5 MCC |
886 | 16) Function return values and names |
887 | ------------------------------------ | |
c16a02d6 AS |
888 | |
889 | Functions can return values of many different kinds, and one of the | |
890 | most common is a value indicating whether the function succeeded or | |
891 | failed. Such a value can be represented as an error-code integer | |
b1a3459b | 892 | (-Exxx = failure, 0 = success) or a ``succeeded`` boolean (0 = failure, |
c16a02d6 AS |
893 | non-zero = success). |
894 | ||
895 | Mixing up these two sorts of representations is a fertile source of | |
896 | difficult-to-find bugs. If the C language included a strong distinction | |
897 | between integers and booleans then the compiler would find these mistakes | |
898 | for us... but it doesn't. To help prevent such bugs, always follow this | |
d8dbbbc5 | 899 | convention:: |
c16a02d6 AS |
900 | |
901 | If the name of a function is an action or an imperative command, | |
902 | the function should return an error-code integer. If the name | |
903 | is a predicate, the function should return a "succeeded" boolean. | |
904 | ||
b1a3459b MCC |
905 | For example, ``add work`` is a command, and the add_work() function returns 0 |
906 | for success or -EBUSY for failure. In the same way, ``PCI device present`` is | |
c16a02d6 AS |
907 | a predicate, and the pci_dev_present() function returns 1 if it succeeds in |
908 | finding a matching device or 0 if it doesn't. | |
909 | ||
910 | All EXPORTed functions must respect this convention, and so should all | |
911 | public functions. Private (static) functions need not, but it is | |
912 | recommended that they do. | |
913 | ||
914 | Functions whose return value is the actual result of a computation, rather | |
915 | than an indication of whether the computation succeeded, are not subject to | |
916 | this rule. Generally they indicate failure by returning some out-of-range | |
917 | result. Typical examples would be functions that return pointers; they use | |
918 | NULL or the ERR_PTR mechanism to report failure. | |
919 | ||
920 | ||
d8dbbbc5 MCC |
921 | 17) Don't re-invent the kernel macros |
922 | ------------------------------------- | |
58637ec9 RD |
923 | |
924 | The header file include/linux/kernel.h contains a number of macros that | |
925 | you should use, rather than explicitly coding some variant of them yourself. | |
926 | For example, if you need to calculate the length of an array, take advantage | |
927 | of the macro | |
928 | ||
d8dbbbc5 MCC |
929 | .. code-block:: c |
930 | ||
09677e0f | 931 | #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) |
58637ec9 RD |
932 | |
933 | Similarly, if you need to calculate the size of some structure member, use | |
934 | ||
d8dbbbc5 MCC |
935 | .. code-block:: c |
936 | ||
09677e0f | 937 | #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f)) |
58637ec9 RD |
938 | |
939 | There are also min() and max() macros that do strict type checking if you | |
940 | need them. Feel free to peruse that header file to see what else is already | |
941 | defined that you shouldn't reproduce in your code. | |
942 | ||
943 | ||
d8dbbbc5 MCC |
944 | 18) Editor modelines and other cruft |
945 | ------------------------------------ | |
4e7bd663 JT |
946 | |
947 | Some editors can interpret configuration information embedded in source files, | |
948 | indicated with special markers. For example, emacs interprets lines marked | |
949 | like this: | |
950 | ||
d8dbbbc5 MCC |
951 | .. code-block:: c |
952 | ||
09677e0f | 953 | -*- mode: c -*- |
4e7bd663 JT |
954 | |
955 | Or like this: | |
956 | ||
d8dbbbc5 MCC |
957 | .. code-block:: c |
958 | ||
09677e0f PK |
959 | /* |
960 | Local Variables: | |
961 | compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" | |
962 | End: | |
963 | */ | |
4e7bd663 JT |
964 | |
965 | Vim interprets markers that look like this: | |
966 | ||
d8dbbbc5 MCC |
967 | .. code-block:: c |
968 | ||
09677e0f | 969 | /* vim:set sw=8 noet */ |
4e7bd663 JT |
970 | |
971 | Do not include any of these in source files. People have their own personal | |
972 | editor configurations, and your source files should not override them. This | |
973 | includes markers for indentation and mode configuration. People may use their | |
974 | own custom mode, or may have some other magic method for making indentation | |
975 | work correctly. | |
976 | ||
977 | ||
d8dbbbc5 MCC |
978 | 19) Inline assembly |
979 | ------------------- | |
9a7c48b7 JT |
980 | |
981 | In architecture-specific code, you may need to use inline assembly to interface | |
982 | with CPU or platform functionality. Don't hesitate to do so when necessary. | |
983 | However, don't use inline assembly gratuitously when C can do the job. You can | |
984 | and should poke hardware from C when possible. | |
985 | ||
986 | Consider writing simple helper functions that wrap common bits of inline | |
987 | assembly, rather than repeatedly writing them with slight variations. Remember | |
988 | that inline assembly can use C parameters. | |
989 | ||
990 | Large, non-trivial assembly functions should go in .S files, with corresponding | |
991 | C prototypes defined in C header files. The C prototypes for assembly | |
b1a3459b | 992 | functions should use ``asmlinkage``. |
9a7c48b7 JT |
993 | |
994 | You may need to mark your asm statement as volatile, to prevent GCC from | |
995 | removing it if GCC doesn't notice any side effects. You don't always need to | |
996 | do so, though, and doing so unnecessarily can limit optimization. | |
997 | ||
998 | When writing a single inline assembly statement containing multiple | |
999 | instructions, put each instruction on a separate line in a separate quoted | |
68f04b57 AS |
1000 | string, and end each string except the last with ``\n\t`` to properly indent |
1001 | the next instruction in the assembly output: | |
9a7c48b7 | 1002 | |
d8dbbbc5 MCC |
1003 | .. code-block:: c |
1004 | ||
9a7c48b7 JT |
1005 | asm ("magic %reg1, #42\n\t" |
1006 | "more_magic %reg2, %reg3" | |
1007 | : /* outputs */ : /* inputs */ : /* clobbers */); | |
1008 | ||
1009 | ||
d8dbbbc5 MCC |
1010 | 20) Conditional Compilation |
1011 | --------------------------- | |
21228a18 JT |
1012 | |
1013 | Wherever possible, don't use preprocessor conditionals (#if, #ifdef) in .c | |
1014 | files; doing so makes code harder to read and logic harder to follow. Instead, | |
1015 | use such conditionals in a header file defining functions for use in those .c | |
1016 | files, providing no-op stub versions in the #else case, and then call those | |
1017 | functions unconditionally from .c files. The compiler will avoid generating | |
1018 | any code for the stub calls, producing identical results, but the logic will | |
1019 | remain easy to follow. | |
1020 | ||
1021 | Prefer to compile out entire functions, rather than portions of functions or | |
1022 | portions of expressions. Rather than putting an ifdef in an expression, factor | |
1023 | out part or all of the expression into a separate helper function and apply the | |
1024 | conditional to that function. | |
1025 | ||
1026 | If you have a function or variable which may potentially go unused in a | |
1027 | particular configuration, and the compiler would warn about its definition | |
1028 | going unused, mark the definition as __maybe_unused rather than wrapping it in | |
1029 | a preprocessor conditional. (However, if a function or variable *always* goes | |
1030 | unused, delete it.) | |
1031 | ||
1032 | Within code, where possible, use the IS_ENABLED macro to convert a Kconfig | |
1033 | symbol into a C boolean expression, and use it in a normal C conditional: | |
1034 | ||
d8dbbbc5 MCC |
1035 | .. code-block:: c |
1036 | ||
21228a18 JT |
1037 | if (IS_ENABLED(CONFIG_SOMETHING)) { |
1038 | ... | |
1039 | } | |
1040 | ||
1041 | The compiler will constant-fold the conditional away, and include or exclude | |
1042 | the block of code just as with an #ifdef, so this will not add any runtime | |
1043 | overhead. However, this approach still allows the C compiler to see the code | |
1044 | inside the block, and check it for correctness (syntax, types, symbol | |
1045 | references, etc). Thus, you still have to use an #ifdef if the code inside the | |
1046 | block references symbols that will not exist if the condition is not met. | |
1047 | ||
1048 | At the end of any non-trivial #if or #ifdef block (more than a few lines), | |
1049 | place a comment after the #endif on the same line, noting the conditional | |
1050 | expression used. For instance: | |
1051 | ||
d8dbbbc5 MCC |
1052 | .. code-block:: c |
1053 | ||
09677e0f PK |
1054 | #ifdef CONFIG_SOMETHING |
1055 | ... | |
1056 | #endif /* CONFIG_SOMETHING */ | |
21228a18 | 1057 | |
a771f2b8 | 1058 | |
d8dbbbc5 MCC |
1059 | Appendix I) References |
1060 | ---------------------- | |
1da177e4 LT |
1061 | |
1062 | The C Programming Language, Second Edition | |
1063 | by Brian W. Kernighan and Dennis M. Ritchie. | |
1064 | Prentice Hall, Inc., 1988. | |
1065 | ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback). | |
1da177e4 LT |
1066 | |
1067 | The Practice of Programming | |
1068 | by Brian W. Kernighan and Rob Pike. | |
1069 | Addison-Wesley, Inc., 1999. | |
1070 | ISBN 0-201-61586-X. | |
1da177e4 LT |
1071 | |
1072 | GNU manuals - where in compliance with K&R and this text - for cpp, gcc, | |
5b0ed2c6 | 1073 | gcc internals and indent, all available from http://www.gnu.org/manual/ |
1da177e4 LT |
1074 | |
1075 | WG14 is the international standardization working group for the programming | |
5b0ed2c6 XVP |
1076 | language C, URL: http://www.open-std.org/JTC1/SC22/WG14/ |
1077 | ||
8c27ceff | 1078 | Kernel process/coding-style.rst, by greg@kroah.com at OLS 2002: |
5b0ed2c6 | 1079 | http://www.kroah.com/linux/talks/ols_2002_kernel_codingstyle_talk/html/ |