blob: 449a8a19fc21fd12582e4ab144549740bd6dd4c3 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001
2 Linux kernel coding style
3
4This is a short document describing the preferred coding style for the
5linux kernel. Coding style is very personal, and I won't _force_ my
6views on anybody, but this is what goes for anything that I have to be
7able to maintain, and I'd prefer it for most other things too. Please
8at least consider the points made here.
9
10First off, I'd suggest printing out a copy of the GNU coding standards,
11and NOT read it. Burn them, it's a great symbolic gesture.
12
13Anyway, here goes:
14
15
16 Chapter 1: Indentation
17
18Tabs are 8 characters, and thus indentations are also 8 characters.
19There are heretic movements that try to make indentations 4 (or even 2!)
20characters deep, and that is akin to trying to define the value of PI to
21be 3.
22
23Rationale: The whole idea behind indentation is to clearly define where
24a block of control starts and ends. Especially when you've been looking
25at your screen for 20 straight hours, you'll find it a lot easier to see
26how the indentation works if you have large indentations.
27
28Now, some people will claim that having 8-character indentations makes
29the code move too far to the right, and makes it hard to read on a
3080-character terminal screen. The answer to that is that if you need
31more than 3 levels of indentation, you're screwed anyway, and should fix
32your program.
33
34In short, 8-char indents make things easier to read, and have the added
35benefit of warning you when you're nesting your functions too deep.
36Heed that warning.
37
Randy Dunlapb3fc9942006-12-10 02:18:56 -080038The preferred way to ease multiple indentation levels in a switch statement is
39to align the "switch" and its subordinate "case" labels in the same column
40instead of "double-indenting" the "case" labels. E.g.:
41
42 switch (suffix) {
43 case 'G':
44 case 'g':
45 mem <<= 30;
46 break;
47 case 'M':
48 case 'm':
49 mem <<= 20;
50 break;
51 case 'K':
52 case 'k':
53 mem <<= 10;
54 /* fall through */
55 default:
56 break;
57 }
58
59
Linus Torvalds1da177e2005-04-16 15:20:36 -070060Don't put multiple statements on a single line unless you have
61something to hide:
62
63 if (condition) do_this;
64 do_something_everytime;
65
Randy Dunlapb3fc9942006-12-10 02:18:56 -080066Don't put multiple assignments on a single line either. Kernel coding style
67is super simple. Avoid tricky expressions.
68
Linus Torvalds1da177e2005-04-16 15:20:36 -070069Outside of comments, documentation and except in Kconfig, spaces are never
70used for indentation, and the above example is deliberately broken.
71
72Get a decent editor and don't leave whitespace at the end of lines.
73
74
75 Chapter 2: Breaking long lines and strings
76
77Coding style is all about readability and maintainability using commonly
78available tools.
79
Alan Coxdff49822007-10-16 23:27:33 -070080The limit on the length of lines is 80 columns and this is a strongly
81preferred limit.
Linus Torvalds1da177e2005-04-16 15:20:36 -070082
Josh Triplett6f76b6f2011-08-03 12:19:07 -070083Statements longer than 80 columns will be broken into sensible chunks, unless
84exceeding 80 columns significantly increases readability and does not hide
85information. Descendants are always substantially shorter than the parent and
86are placed substantially to the right. The same applies to function headers
87with a long argument list. However, never break user-visible strings such as
88printk messages, because that breaks the ability to grep for them.
Linus Torvalds1da177e2005-04-16 15:20:36 -070089
Linus Torvalds1da177e2005-04-16 15:20:36 -070090
Randy Dunlapb3fc9942006-12-10 02:18:56 -080091 Chapter 3: Placing Braces and Spaces
Linus Torvalds1da177e2005-04-16 15:20:36 -070092
93The other issue that always comes up in C styling is the placement of
94braces. Unlike the indent size, there are few technical reasons to
95choose one placement strategy over the other, but the preferred way, as
96shown to us by the prophets Kernighan and Ritchie, is to put the opening
97brace last on the line, and put the closing brace first, thusly:
98
99 if (x is true) {
100 we do y
101 }
102
Randy Dunlapb3fc9942006-12-10 02:18:56 -0800103This applies to all non-function statement blocks (if, switch, for,
104while, do). E.g.:
105
106 switch (action) {
107 case KOBJ_ADD:
108 return "add";
109 case KOBJ_REMOVE:
110 return "remove";
111 case KOBJ_CHANGE:
112 return "change";
113 default:
114 return NULL;
115 }
116
Linus Torvalds1da177e2005-04-16 15:20:36 -0700117However, there is one special case, namely functions: they have the
118opening brace at the beginning of the next line, thus:
119
120 int function(int x)
121 {
122 body of function
123 }
124
125Heretic people all over the world have claimed that this inconsistency
126is ... well ... inconsistent, but all right-thinking people know that
127(a) K&R are _right_ and (b) K&R are right. Besides, functions are
128special anyway (you can't nest them in C).
129
130Note that the closing brace is empty on a line of its own, _except_ in
131the cases where it is followed by a continuation of the same statement,
132ie a "while" in a do-statement or an "else" in an if-statement, like
133this:
134
135 do {
136 body of do-loop
137 } while (condition);
138
139and
140
141 if (x == y) {
142 ..
143 } else if (x > y) {
144 ...
145 } else {
146 ....
147 }
148
149Rationale: K&R.
150
151Also, note that this brace-placement also minimizes the number of empty
152(or almost empty) lines, without any loss of readability. Thus, as the
153supply of new-lines on your screen is not a renewable resource (think
15425-line terminal screens here), you have more empty lines to put
155comments on.
156
Oliver Neukume659ba42007-05-08 00:30:34 -0700157Do not unnecessarily use braces where a single statement will do.
158
159if (condition)
160 action();
161
Harry Wei38829dc2011-03-22 16:35:01 -0700162and
163
164if (condition)
165 do_this();
166else
167 do_that();
168
Antonio Ospiteb218ab02011-11-04 11:22:19 -0700169This does not apply if only one branch of a conditional statement is a single
170statement; in the latter case use braces in both branches:
Oliver Neukume659ba42007-05-08 00:30:34 -0700171
172if (condition) {
173 do_this();
174 do_that();
175} else {
176 otherwise();
177}
178
Randy Dunlapb3fc9942006-12-10 02:18:56 -0800179 3.1: Spaces
180
181Linux kernel style for use of spaces depends (mostly) on
182function-versus-keyword usage. Use a space after (most) keywords. The
183notable exceptions are sizeof, typeof, alignof, and __attribute__, which look
184somewhat like functions (and are usually used with parentheses in Linux,
185although they are not required in the language, as in: "sizeof info" after
186"struct fileinfo info;" is declared).
187
188So use a space after these keywords:
189 if, switch, case, for, do, while
190but not with sizeof, typeof, alignof, or __attribute__. E.g.,
191 s = sizeof(struct file);
192
193Do not add spaces around (inside) parenthesized expressions. This example is
194*bad*:
195
196 s = sizeof( struct file );
197
198When declaring pointer data or a function that returns a pointer type, the
199preferred use of '*' is adjacent to the data name or function name and not
200adjacent to the type name. Examples:
201
202 char *linux_banner;
203 unsigned long long memparse(char *ptr, char **retptr);
204 char *match_strdup(substring_t *s);
205
206Use one space around (on each side of) most binary and ternary operators,
207such as any of these:
208
209 = + - < > * / % | & ^ <= >= == != ? :
210
211but no space after unary operators:
212 & * + - ~ ! sizeof typeof alignof __attribute__ defined
213
214no space before the postfix increment & decrement unary operators:
215 ++ --
216
217no space after the prefix increment & decrement unary operators:
218 ++ --
219
220and no space around the '.' and "->" structure member operators.
221
Josh Tripletta923fd62007-07-15 23:41:37 -0700222Do not leave trailing whitespace at the ends of lines. Some editors with
223"smart" indentation will insert whitespace at the beginning of new lines as
224appropriate, so you can start typing the next line of code right away.
225However, some such editors do not remove the whitespace if you end up not
226putting a line of code there, such as if you leave a blank line. As a result,
227you end up with lines containing trailing whitespace.
228
229Git will warn you about patches that introduce trailing whitespace, and can
230optionally strip the trailing whitespace for you; however, if applying a series
231of patches, this may make later patches in the series fail by changing their
232context lines.
233
Linus Torvalds1da177e2005-04-16 15:20:36 -0700234
235 Chapter 4: Naming
236
237C is a Spartan language, and so should your naming be. Unlike Modula-2
238and Pascal programmers, C programmers do not use cute names like
239ThisVariableIsATemporaryCounter. A C programmer would call that
240variable "tmp", which is much easier to write, and not the least more
241difficult to understand.
242
243HOWEVER, while mixed-case names are frowned upon, descriptive names for
244global variables are a must. To call a global function "foo" is a
245shooting offense.
246
247GLOBAL variables (to be used only if you _really_ need them) need to
248have descriptive names, as do global functions. If you have a function
249that counts the number of active users, you should call that
250"count_active_users()" or similar, you should _not_ call it "cntusr()".
251
252Encoding the type of a function into the name (so-called Hungarian
253notation) is brain damaged - the compiler knows the types anyway and can
254check those, and it only confuses the programmer. No wonder MicroSoft
255makes buggy programs.
256
257LOCAL variable names should be short, and to the point. If you have
258some random integer loop counter, it should probably be called "i".
259Calling it "loop_counter" is non-productive, if there is no chance of it
260being mis-understood. Similarly, "tmp" can be just about any type of
261variable that is used to hold a temporary value.
262
263If you are afraid to mix up your local variable names, you have another
264problem, which is called the function-growth-hormone-imbalance syndrome.
Randy Dunlapb3fc9942006-12-10 02:18:56 -0800265See chapter 6 (Functions).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700266
267
Randy Dunlap226a6b82006-06-23 02:05:58 -0700268 Chapter 5: Typedefs
269
270Please don't use things like "vps_t".
271
272It's a _mistake_ to use typedef for structures and pointers. When you see a
273
274 vps_t a;
275
276in the source, what does it mean?
277
278In contrast, if it says
279
280 struct virtual_container *a;
281
282you can actually tell what "a" is.
283
284Lots of people think that typedefs "help readability". Not so. They are
285useful only for:
286
287 (a) totally opaque objects (where the typedef is actively used to _hide_
288 what the object is).
289
290 Example: "pte_t" etc. opaque objects that you can only access using
291 the proper accessor functions.
292
293 NOTE! Opaqueness and "accessor functions" are not good in themselves.
294 The reason we have them for things like pte_t etc. is that there
295 really is absolutely _zero_ portably accessible information there.
296
297 (b) Clear integer types, where the abstraction _helps_ avoid confusion
298 whether it is "int" or "long".
299
300 u8/u16/u32 are perfectly fine typedefs, although they fit into
301 category (d) better than here.
302
303 NOTE! Again - there needs to be a _reason_ for this. If something is
304 "unsigned long", then there's no reason to do
305
306 typedef unsigned long myflags_t;
307
308 but if there is a clear reason for why it under certain circumstances
309 might be an "unsigned int" and under other configurations might be
310 "unsigned long", then by all means go ahead and use a typedef.
311
312 (c) when you use sparse to literally create a _new_ type for
313 type-checking.
314
315 (d) New types which are identical to standard C99 types, in certain
316 exceptional circumstances.
317
318 Although it would only take a short amount of time for the eyes and
319 brain to become accustomed to the standard types like 'uint32_t',
320 some people object to their use anyway.
321
322 Therefore, the Linux-specific 'u8/u16/u32/u64' types and their
323 signed equivalents which are identical to standard types are
324 permitted -- although they are not mandatory in new code of your
325 own.
326
327 When editing existing code which already uses one or the other set
328 of types, you should conform to the existing choices in that code.
329
330 (e) Types safe for use in userspace.
331
332 In certain structures which are visible to userspace, we cannot
333 require C99 types and cannot use the 'u32' form above. Thus, we
334 use __u32 and similar types in all structures which are shared
335 with userspace.
336
337Maybe there are other cases too, but the rule should basically be to NEVER
338EVER use a typedef unless you can clearly match one of those rules.
339
340In general, a pointer, or a struct that has elements that can reasonably
341be directly accessed should _never_ be a typedef.
342
343
344 Chapter 6: Functions
Linus Torvalds1da177e2005-04-16 15:20:36 -0700345
346Functions should be short and sweet, and do just one thing. They should
347fit on one or two screenfuls of text (the ISO/ANSI screen size is 80x24,
348as we all know), and do one thing and do that well.
349
350The maximum length of a function is inversely proportional to the
351complexity and indentation level of that function. So, if you have a
352conceptually simple function that is just one long (but simple)
353case-statement, where you have to do lots of small things for a lot of
354different cases, it's OK to have a longer function.
355
356However, if you have a complex function, and you suspect that a
357less-than-gifted first-year high-school student might not even
358understand what the function is all about, you should adhere to the
359maximum limits all the more closely. Use helper functions with
360descriptive names (you can ask the compiler to in-line them if you think
361it's performance-critical, and it will probably do a better job of it
362than you would have done).
363
364Another measure of the function is the number of local variables. They
365shouldn't exceed 5-10, or you're doing something wrong. Re-think the
366function, and split it into smaller pieces. A human brain can
367generally easily keep track of about 7 different things, anything more
368and it gets confused. You know you're brilliant, but maybe you'd like
369to understand what you did 2 weeks from now.
370
Randy Dunlapb3fc9942006-12-10 02:18:56 -0800371In source files, separate functions with one blank line. If the function is
372exported, the EXPORT* macro for it should follow immediately after the closing
373function brace line. E.g.:
374
375int system_is_up(void)
376{
377 return system_state == SYSTEM_RUNNING;
378}
379EXPORT_SYMBOL(system_is_up);
380
381In function prototypes, include parameter names with their data types.
382Although this is not required by the C language, it is preferred in Linux
383because it is a simple way to add valuable information for the reader.
384
Linus Torvalds1da177e2005-04-16 15:20:36 -0700385
Randy Dunlap226a6b82006-06-23 02:05:58 -0700386 Chapter 7: Centralized exiting of functions
Linus Torvalds1da177e2005-04-16 15:20:36 -0700387
388Albeit deprecated by some people, the equivalent of the goto statement is
389used frequently by compilers in form of the unconditional jump instruction.
390
391The goto statement comes in handy when a function exits from multiple
Dan Carpenterb57a0502013-07-03 15:08:08 -0700392locations and some common work such as cleanup has to be done. If there is no
393cleanup needed then just return directly.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700394
Dan Carpenterea040362014-12-02 11:59:50 +0300395Choose label names which say what the goto does or why the goto exists. An
396example of a good name could be "out_buffer:" if the goto frees "buffer". Avoid
397using GW-BASIC names like "err1:" and "err2:". Also don't name them after the
398goto location like "err_kmalloc_failed:"
399
400The rationale for using gotos is:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700401
402- unconditional statements are easier to understand and follow
403- nesting is reduced
404- errors by not updating individual exit points when making
405 modifications are prevented
406- saves the compiler work to optimize redundant code away ;)
407
Jesper Juhldc3d28d2006-01-09 20:53:51 -0800408int fun(int a)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700409{
410 int result = 0;
Dan Carpenterea040362014-12-02 11:59:50 +0300411 char *buffer;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700412
Dan Carpenterea040362014-12-02 11:59:50 +0300413 buffer = kmalloc(SIZE, GFP_KERNEL);
414 if (!buffer)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700415 return -ENOMEM;
416
417 if (condition1) {
418 while (loop1) {
419 ...
420 }
421 result = 1;
Dan Carpenterea040362014-12-02 11:59:50 +0300422 goto out_buffer;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700423 }
424 ...
Dan Carpenterea040362014-12-02 11:59:50 +0300425out_buffer:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700426 kfree(buffer);
427 return result;
428}
429
Dan Carpenterea040362014-12-02 11:59:50 +0300430A common type of bug to be aware of it "one err bugs" which look like this:
431
432err:
433 kfree(foo->bar);
434 kfree(foo);
435 return ret;
436
437The bug in this code is that on some exit paths "foo" is NULL. Normally the
438fix for this is to split it up into two error labels "err_bar:" and "err_foo:".
439
440
Randy Dunlap226a6b82006-06-23 02:05:58 -0700441 Chapter 8: Commenting
Linus Torvalds1da177e2005-04-16 15:20:36 -0700442
443Comments are good, but there is also a danger of over-commenting. NEVER
444try to explain HOW your code works in a comment: it's much better to
445write the code so that the _working_ is obvious, and it's a waste of
446time to explain badly written code.
447
448Generally, you want your comments to tell WHAT your code does, not HOW.
449Also, try to avoid putting comments inside a function body: if the
450function is so complex that you need to separately comment parts of it,
Randy Dunlapb3fc9942006-12-10 02:18:56 -0800451you should probably go back to chapter 6 for a while. You can make
Linus Torvalds1da177e2005-04-16 15:20:36 -0700452small comments to note or warn about something particularly clever (or
453ugly), but try to avoid excess. Instead, put the comments at the head
454of the function, telling people what it does, and possibly WHY it does
455it.
456
Randy Dunlapb3fc9942006-12-10 02:18:56 -0800457When commenting the kernel API functions, please use the kernel-doc format.
Pekka J Enberge776eba2005-09-10 00:26:44 -0700458See the files Documentation/kernel-doc-nano-HOWTO.txt and scripts/kernel-doc
459for details.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700460
Randy Dunlapb3fc9942006-12-10 02:18:56 -0800461Linux style for comments is the C89 "/* ... */" style.
462Don't use C99-style "// ..." comments.
463
464The preferred style for long (multi-line) comments is:
465
466 /*
467 * This is the preferred style for multi-line
468 * comments in the Linux kernel source code.
469 * Please use it consistently.
470 *
471 * Description: A column of asterisks on the left side,
472 * with beginning and ending almost-blank lines.
473 */
474
Joe Perchesc4ff1b52012-10-04 17:13:36 -0700475For files in net/ and drivers/net/ the preferred style for long (multi-line)
476comments is a little different.
477
478 /* The preferred comment style for files in net/ and drivers/net
479 * looks like this.
480 *
481 * It is nearly the same as the generally preferred comment style,
482 * but there is no initial almost-blank line.
483 */
484
Randy Dunlapb3fc9942006-12-10 02:18:56 -0800485It's also important to comment data, whether they are basic types or derived
486types. To this end, use just one data declaration per line (no commas for
487multiple data declarations). This leaves you room for a small comment on each
488item, explaining its use.
489
490
Randy Dunlap226a6b82006-06-23 02:05:58 -0700491 Chapter 9: You've made a mess of it
Linus Torvalds1da177e2005-04-16 15:20:36 -0700492
493That's OK, we all do. You've probably been told by your long-time Unix
494user helper that "GNU emacs" automatically formats the C sources for
495you, and you've noticed that yes, it does do that, but the defaults it
496uses are less than desirable (in fact, they are worse than random
497typing - an infinite number of monkeys typing into GNU emacs would never
498make a good program).
499
500So, you can either get rid of GNU emacs, or change it to use saner
501values. To do the latter, you can stick the following in your .emacs file:
502
Johannes Weinera7f371e2008-07-25 01:45:51 -0700503(defun c-lineup-arglist-tabs-only (ignored)
504 "Line up argument lists by tabs, not spaces"
505 (let* ((anchor (c-langelem-pos c-syntactic-element))
506 (column (c-langelem-2nd-pos c-syntactic-element))
507 (offset (- (1+ column) anchor))
508 (steps (floor offset c-basic-offset)))
509 (* (max steps 1)
510 c-basic-offset)))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700511
Teemu Likonen0acbc6c2009-01-29 16:28:16 -0800512(add-hook 'c-mode-common-hook
513 (lambda ()
514 ;; Add kernel style
515 (c-add-style
516 "linux-tabs-only"
517 '("linux" (c-offsets-alist
518 (arglist-cont-nonempty
519 c-lineup-gcc-asm-reg
520 c-lineup-arglist-tabs-only))))))
521
Johannes Weinera7f371e2008-07-25 01:45:51 -0700522(add-hook 'c-mode-hook
523 (lambda ()
524 (let ((filename (buffer-file-name)))
525 ;; Enable kernel mode for the appropriate files
526 (when (and filename
Dan Carpenter70221392009-01-29 16:28:28 -0800527 (string-match (expand-file-name "~/src/linux-trees")
528 filename))
Johannes Weinera7f371e2008-07-25 01:45:51 -0700529 (setq indent-tabs-mode t)
Alison Chaiken039d19a2015-01-25 19:26:01 -0800530 (setq show-trailing-whitespace t)
Teemu Likonen0acbc6c2009-01-29 16:28:16 -0800531 (c-set-style "linux-tabs-only")))))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700532
Johannes Weinera7f371e2008-07-25 01:45:51 -0700533This will make emacs go better with the kernel coding style for C
534files below ~/src/linux-trees.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700535
536But even if you fail in getting emacs to do sane formatting, not
537everything is lost: use "indent".
538
539Now, again, GNU indent has the same brain-dead settings that GNU emacs
540has, which is why you need to give it a few command line options.
541However, that's not too bad, because even the makers of GNU indent
542recognize the authority of K&R (the GNU people aren't evil, they are
543just severely misguided in this matter), so you just give indent the
544options "-kr -i8" (stands for "K&R, 8 character indents"), or use
545"scripts/Lindent", which indents in the latest style.
546
547"indent" has a lot of options, and especially when it comes to comment
548re-formatting you may want to take a look at the man page. But
549remember: "indent" is not a fix for bad programming.
550
551
Robert P. J. Day6754bb42007-05-23 13:57:42 -0700552 Chapter 10: Kconfig configuration files
Linus Torvalds1da177e2005-04-16 15:20:36 -0700553
Robert P. J. Day6754bb42007-05-23 13:57:42 -0700554For all of the Kconfig* configuration files throughout the source tree,
555the indentation is somewhat different. Lines under a "config" definition
556are indented with one tab, while help text is indented an additional two
557spaces. Example:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700558
Robert P. J. Day6754bb42007-05-23 13:57:42 -0700559config AUDIT
560 bool "Auditing support"
561 depends on NET
Linus Torvalds1da177e2005-04-16 15:20:36 -0700562 help
Robert P. J. Day6754bb42007-05-23 13:57:42 -0700563 Enable auditing infrastructure that can be used with another
564 kernel subsystem, such as SELinux (which requires this for
565 logging of avc messages output). Does not do system-call
566 auditing without CONFIG_AUDITSYSCALL.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700567
Kees Cook0335cb42012-10-02 11:16:15 -0700568Seriously dangerous features (such as write support for certain
Robert P. J. Day6754bb42007-05-23 13:57:42 -0700569filesystems) should advertise this prominently in their prompt string:
570
571config ADFS_FS_RW
572 bool "ADFS write support (DANGEROUS)"
573 depends on ADFS_FS
574 ...
575
576For full documentation on the configuration files, see the file
577Documentation/kbuild/kconfig-language.txt.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700578
579
Randy Dunlap226a6b82006-06-23 02:05:58 -0700580 Chapter 11: Data structures
Linus Torvalds1da177e2005-04-16 15:20:36 -0700581
582Data structures that have visibility outside the single-threaded
583environment they are created and destroyed in should always have
584reference counts. In the kernel, garbage collection doesn't exist (and
585outside the kernel garbage collection is slow and inefficient), which
586means that you absolutely _have_ to reference count all your uses.
587
588Reference counting means that you can avoid locking, and allows multiple
589users to have access to the data structure in parallel - and not having
590to worry about the structure suddenly going away from under them just
591because they slept or did something else for a while.
592
593Note that locking is _not_ a replacement for reference counting.
594Locking is used to keep data structures coherent, while reference
595counting is a memory management technique. Usually both are needed, and
596they are not to be confused with each other.
597
598Many data structures can indeed have two levels of reference counting,
599when there are users of different "classes". The subclass count counts
600the number of subclass users, and decrements the global count just once
601when the subclass count goes to zero.
602
603Examples of this kind of "multi-level-reference-counting" can be found in
604memory management ("struct mm_struct": mm_users and mm_count), and in
605filesystem code ("struct super_block": s_count and s_active).
606
607Remember: if another thread can find your data structure, and you don't
608have a reference count on it, you almost certainly have a bug.
609
610
Randy Dunlap226a6b82006-06-23 02:05:58 -0700611 Chapter 12: Macros, Enums and RTL
Linus Torvalds1da177e2005-04-16 15:20:36 -0700612
613Names of macros defining constants and labels in enums are capitalized.
614
615#define CONSTANT 0x12345
616
617Enums are preferred when defining several related constants.
618
619CAPITALIZED macro names are appreciated but macros resembling functions
620may be named in lower case.
621
622Generally, inline functions are preferable to macros resembling functions.
623
624Macros with multiple statements should be enclosed in a do - while block:
625
626#define macrofun(a, b, c) \
627 do { \
628 if (a == 5) \
629 do_this(b, c); \
630 } while (0)
631
632Things to avoid when using macros:
633
6341) macros that affect control flow:
635
636#define FOO(x) \
637 do { \
638 if (blah(x) < 0) \
639 return -EBUGGERED; \
640 } while(0)
641
642is a _very_ bad idea. It looks like a function call but exits the "calling"
643function; don't break the internal parsers of those who will read the code.
644
6452) macros that depend on having a local variable with a magic name:
646
647#define FOO(val) bar(index, val)
648
649might look like a good thing, but it's confusing as hell when one reads the
650code and it's prone to breakage from seemingly innocent changes.
651
6523) macros with arguments that are used as l-values: FOO(x) = y; will
653bite you if somebody e.g. turns FOO into an inline function.
654
6554) forgetting about precedence: macros defining constants using expressions
656must enclose the expression in parentheses. Beware of similar issues with
657macros using parameters.
658
659#define CONSTANT 0x4000
660#define CONSTEXP (CONSTANT | 3)
661
662The cpp manual deals with macros exhaustively. The gcc internals manual also
663covers RTL which is used frequently with assembly language in the kernel.
664
665
Randy Dunlap226a6b82006-06-23 02:05:58 -0700666 Chapter 13: Printing kernel messages
Linus Torvalds1da177e2005-04-16 15:20:36 -0700667
668Kernel developers like to be seen as literate. Do mind the spelling
669of kernel messages to make a good impression. Do not use crippled
David Brownell6b094482007-07-13 16:32:09 -0700670words like "dont"; use "do not" or "don't" instead. Make the messages
671concise, clear, and unambiguous.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700672
673Kernel messages do not have to be terminated with a period.
674
675Printing numbers in parentheses (%d) adds no value and should be avoided.
676
David Brownell6b094482007-07-13 16:32:09 -0700677There are a number of driver model diagnostic macros in <linux/device.h>
678which you should use to make sure messages are matched to the right device
679and driver, and are tagged with the right level: dev_err(), dev_warn(),
680dev_info(), and so forth. For messages that aren't associated with a
Dan Streetman6e099f52014-06-04 16:11:44 -0700681particular device, <linux/printk.h> defines pr_notice(), pr_info(),
682pr_warn(), pr_err(), etc.
David Brownell6b094482007-07-13 16:32:09 -0700683
684Coming up with good debugging messages can be quite a challenge; and once
Dan Streetman6e099f52014-06-04 16:11:44 -0700685you have them, they can be a huge help for remote troubleshooting. However
686debug message printing is handled differently than printing other non-debug
687messages. While the other pr_XXX() functions print unconditionally,
688pr_debug() does not; it is compiled out by default, unless either DEBUG is
689defined or CONFIG_DYNAMIC_DEBUG is set. That is true for dev_dbg() also,
690and a related convention uses VERBOSE_DEBUG to add dev_vdbg() messages to
691the ones already enabled by DEBUG.
692
693Many subsystems have Kconfig debug options to turn on -DDEBUG in the
694corresponding Makefile; in other cases specific files #define DEBUG. And
695when a debug message should be unconditionally printed, such as if it is
Raymond L. Rivera7c18fd72014-07-24 02:39:44 -0700696already inside a debug-related #ifdef section, printk(KERN_DEBUG ...) can be
Dan Streetman6e099f52014-06-04 16:11:44 -0700697used.
David Brownell6b094482007-07-13 16:32:09 -0700698
Linus Torvalds1da177e2005-04-16 15:20:36 -0700699
Randy Dunlap226a6b82006-06-23 02:05:58 -0700700 Chapter 14: Allocating memory
Pekka J Enbergaf4e5a22005-09-16 19:28:11 -0700701
702The kernel provides the following general purpose memory allocators:
Xi Wang158372942012-05-31 16:26:04 -0700703kmalloc(), kzalloc(), kmalloc_array(), kcalloc(), vmalloc(), and
704vzalloc(). Please refer to the API documentation for further information
705about them.
Pekka J Enbergaf4e5a22005-09-16 19:28:11 -0700706
707The preferred form for passing a size of a struct is the following:
708
709 p = kmalloc(sizeof(*p), ...);
710
711The alternative form where struct name is spelled out hurts readability and
712introduces an opportunity for a bug when the pointer variable type is changed
713but the corresponding sizeof that is passed to a memory allocator is not.
714
715Casting the return value which is a void pointer is redundant. The conversion
716from void pointer to any other pointer type is guaranteed by the C programming
717language.
718
Xi Wang158372942012-05-31 16:26:04 -0700719The preferred form for allocating an array is the following:
720
721 p = kmalloc_array(n, sizeof(...), ...);
722
723The preferred form for allocating a zeroed array is the following:
724
725 p = kcalloc(n, sizeof(...), ...);
726
727Both forms check for overflow on the allocation size n * sizeof(...),
728and return NULL if that occurred.
729
Pekka J Enbergaf4e5a22005-09-16 19:28:11 -0700730
Randy Dunlap226a6b82006-06-23 02:05:58 -0700731 Chapter 15: The inline disease
Arjan van de Vena771f2b2006-01-08 01:05:04 -0800732
733There appears to be a common misperception that gcc has a magic "make me
734faster" speedup option called "inline". While the use of inlines can be
Jesper Juhl53ab97a2007-05-08 00:31:06 -0700735appropriate (for example as a means of replacing macros, see Chapter 12), it
Arjan van de Vena771f2b2006-01-08 01:05:04 -0800736very often is not. Abundant use of the inline keyword leads to a much bigger
737kernel, which in turn slows the system as a whole down, due to a bigger
738icache footprint for the CPU and simply because there is less memory
739available for the pagecache. Just think about it; a pagecache miss causes a
Martin Olsson19af5cd2009-04-23 11:37:37 +0200740disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles
741that can go into these 5 milliseconds.
Arjan van de Vena771f2b2006-01-08 01:05:04 -0800742
743A reasonable rule of thumb is to not put inline at functions that have more
744than 3 lines of code in them. An exception to this rule are the cases where
745a parameter is known to be a compiletime constant, and as a result of this
746constantness you *know* the compiler will be able to optimize most of your
747function away at compile time. For a good example of this later case, see
748the kmalloc() inline function.
749
750Often people argue that adding inline to functions that are static and used
751only once is always a win since there is no space tradeoff. While this is
752technically correct, gcc is capable of inlining these automatically without
753help, and the maintenance issue of removing the inline when a second user
754appears outweighs the potential value of the hint that tells gcc to do
755something it would have done anyway.
756
757
Alan Sternc16a02d2006-09-29 02:01:21 -0700758 Chapter 16: Function return values and names
759
760Functions can return values of many different kinds, and one of the
761most common is a value indicating whether the function succeeded or
762failed. Such a value can be represented as an error-code integer
763(-Exxx = failure, 0 = success) or a "succeeded" boolean (0 = failure,
764non-zero = success).
765
766Mixing up these two sorts of representations is a fertile source of
767difficult-to-find bugs. If the C language included a strong distinction
768between integers and booleans then the compiler would find these mistakes
769for us... but it doesn't. To help prevent such bugs, always follow this
770convention:
771
772 If the name of a function is an action or an imperative command,
773 the function should return an error-code integer. If the name
774 is a predicate, the function should return a "succeeded" boolean.
775
776For example, "add work" is a command, and the add_work() function returns 0
777for success or -EBUSY for failure. In the same way, "PCI device present" is
778a predicate, and the pci_dev_present() function returns 1 if it succeeds in
779finding a matching device or 0 if it doesn't.
780
781All EXPORTed functions must respect this convention, and so should all
782public functions. Private (static) functions need not, but it is
783recommended that they do.
784
785Functions whose return value is the actual result of a computation, rather
786than an indication of whether the computation succeeded, are not subject to
787this rule. Generally they indicate failure by returning some out-of-range
788result. Typical examples would be functions that return pointers; they use
789NULL or the ERR_PTR mechanism to report failure.
790
791
Robert P. J. Day58637ec2006-12-22 01:09:11 -0800792 Chapter 17: Don't re-invent the kernel macros
793
794The header file include/linux/kernel.h contains a number of macros that
795you should use, rather than explicitly coding some variant of them yourself.
796For example, if you need to calculate the length of an array, take advantage
797of the macro
798
799 #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))
800
801Similarly, if you need to calculate the size of some structure member, use
802
803 #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f))
804
805There are also min() and max() macros that do strict type checking if you
806need them. Feel free to peruse that header file to see what else is already
807defined that you shouldn't reproduce in your code.
808
809
Josh Triplett4e7bd662007-07-15 23:41:37 -0700810 Chapter 18: Editor modelines and other cruft
811
812Some editors can interpret configuration information embedded in source files,
813indicated with special markers. For example, emacs interprets lines marked
814like this:
815
816-*- mode: c -*-
817
818Or like this:
819
820/*
821Local Variables:
822compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c"
823End:
824*/
825
826Vim interprets markers that look like this:
827
828/* vim:set sw=8 noet */
829
830Do not include any of these in source files. People have their own personal
831editor configurations, and your source files should not override them. This
832includes markers for indentation and mode configuration. People may use their
833own custom mode, or may have some other magic method for making indentation
834work correctly.
835
836
Josh Triplett9a7c48b2012-03-30 13:37:10 -0700837 Chapter 19: Inline assembly
838
839In architecture-specific code, you may need to use inline assembly to interface
840with CPU or platform functionality. Don't hesitate to do so when necessary.
841However, don't use inline assembly gratuitously when C can do the job. You can
842and should poke hardware from C when possible.
843
844Consider writing simple helper functions that wrap common bits of inline
845assembly, rather than repeatedly writing them with slight variations. Remember
846that inline assembly can use C parameters.
847
848Large, non-trivial assembly functions should go in .S files, with corresponding
849C prototypes defined in C header files. The C prototypes for assembly
850functions should use "asmlinkage".
851
852You may need to mark your asm statement as volatile, to prevent GCC from
853removing it if GCC doesn't notice any side effects. You don't always need to
854do so, though, and doing so unnecessarily can limit optimization.
855
856When writing a single inline assembly statement containing multiple
857instructions, put each instruction on a separate line in a separate quoted
858string, and end each string except the last with \n\t to properly indent the
859next instruction in the assembly output:
860
861 asm ("magic %reg1, #42\n\t"
862 "more_magic %reg2, %reg3"
863 : /* outputs */ : /* inputs */ : /* clobbers */);
864
865
Josh Triplett21228a12014-10-29 11:15:17 -0700866 Chapter 20: Conditional Compilation
867
868Wherever possible, don't use preprocessor conditionals (#if, #ifdef) in .c
869files; doing so makes code harder to read and logic harder to follow. Instead,
870use such conditionals in a header file defining functions for use in those .c
871files, providing no-op stub versions in the #else case, and then call those
872functions unconditionally from .c files. The compiler will avoid generating
873any code for the stub calls, producing identical results, but the logic will
874remain easy to follow.
875
876Prefer to compile out entire functions, rather than portions of functions or
877portions of expressions. Rather than putting an ifdef in an expression, factor
878out part or all of the expression into a separate helper function and apply the
879conditional to that function.
880
881If you have a function or variable which may potentially go unused in a
882particular configuration, and the compiler would warn about its definition
883going unused, mark the definition as __maybe_unused rather than wrapping it in
884a preprocessor conditional. (However, if a function or variable *always* goes
885unused, delete it.)
886
887Within code, where possible, use the IS_ENABLED macro to convert a Kconfig
888symbol into a C boolean expression, and use it in a normal C conditional:
889
890 if (IS_ENABLED(CONFIG_SOMETHING)) {
891 ...
892 }
893
894The compiler will constant-fold the conditional away, and include or exclude
895the block of code just as with an #ifdef, so this will not add any runtime
896overhead. However, this approach still allows the C compiler to see the code
897inside the block, and check it for correctness (syntax, types, symbol
898references, etc). Thus, you still have to use an #ifdef if the code inside the
899block references symbols that will not exist if the condition is not met.
900
901At the end of any non-trivial #if or #ifdef block (more than a few lines),
902place a comment after the #endif on the same line, noting the conditional
903expression used. For instance:
904
905#ifdef CONFIG_SOMETHING
906...
907#endif /* CONFIG_SOMETHING */
908
Arjan van de Vena771f2b2006-01-08 01:05:04 -0800909
Randy Dunlap226a6b82006-06-23 02:05:58 -0700910 Appendix I: References
Linus Torvalds1da177e2005-04-16 15:20:36 -0700911
912The C Programming Language, Second Edition
913by Brian W. Kernighan and Dennis M. Ritchie.
914Prentice Hall, Inc., 1988.
915ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback).
916URL: http://cm.bell-labs.com/cm/cs/cbook/
917
918The Practice of Programming
919by Brian W. Kernighan and Rob Pike.
920Addison-Wesley, Inc., 1999.
921ISBN 0-201-61586-X.
922URL: http://cm.bell-labs.com/cm/cs/tpop/
923
924GNU manuals - where in compliance with K&R and this text - for cpp, gcc,
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800925gcc internals and indent, all available from http://www.gnu.org/manual/
Linus Torvalds1da177e2005-04-16 15:20:36 -0700926
927WG14 is the international standardization working group for the programming
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800928language C, URL: http://www.open-std.org/JTC1/SC22/WG14/
929
930Kernel CodingStyle, by greg@kroah.com at OLS 2002:
931http://www.kroah.com/linux/talks/ols_2002_kernel_codingstyle_talk/html/
Linus Torvalds1da177e2005-04-16 15:20:36 -0700932