CodingStyle 26 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728
  1. Linux kernel coding style
  2. This is a short document describing the preferred coding style for the
  3. linux kernel. Coding style is very personal, and I won't _force_ my
  4. views on anybody, but this is what goes for anything that I have to be
  5. able to maintain, and I'd prefer it for most other things too. Please
  6. at least consider the points made here.
  7. First off, I'd suggest printing out a copy of the GNU coding standards,
  8. and NOT read it. Burn them, it's a great symbolic gesture.
  9. Anyway, here goes:
  10. Chapter 1: Indentation
  11. Tabs are 8 characters, and thus indentations are also 8 characters.
  12. There are heretic movements that try to make indentations 4 (or even 2!)
  13. characters deep, and that is akin to trying to define the value of PI to
  14. be 3.
  15. Rationale: The whole idea behind indentation is to clearly define where
  16. a block of control starts and ends. Especially when you've been looking
  17. at your screen for 20 straight hours, you'll find it a lot easier to see
  18. how the indentation works if you have large indentations.
  19. Now, some people will claim that having 8-character indentations makes
  20. the code move too far to the right, and makes it hard to read on a
  21. 80-character terminal screen. The answer to that is that if you need
  22. more than 3 levels of indentation, you're screwed anyway, and should fix
  23. your program.
  24. In short, 8-char indents make things easier to read, and have the added
  25. benefit of warning you when you're nesting your functions too deep.
  26. Heed that warning.
  27. The preferred way to ease multiple indentation levels in a switch statement is
  28. to align the "switch" and its subordinate "case" labels in the same column
  29. instead of "double-indenting" the "case" labels. E.g.:
  30. switch (suffix) {
  31. case 'G':
  32. case 'g':
  33. mem <<= 30;
  34. break;
  35. case 'M':
  36. case 'm':
  37. mem <<= 20;
  38. break;
  39. case 'K':
  40. case 'k':
  41. mem <<= 10;
  42. /* fall through */
  43. default:
  44. break;
  45. }
  46. Don't put multiple statements on a single line unless you have
  47. something to hide:
  48. if (condition) do_this;
  49. do_something_everytime;
  50. Don't put multiple assignments on a single line either. Kernel coding style
  51. is super simple. Avoid tricky expressions.
  52. Outside of comments, documentation and except in Kconfig, spaces are never
  53. used for indentation, and the above example is deliberately broken.
  54. Get a decent editor and don't leave whitespace at the end of lines.
  55. Chapter 2: Breaking long lines and strings
  56. Coding style is all about readability and maintainability using commonly
  57. available tools.
  58. The limit on the length of lines is 80 columns and this is a hard limit.
  59. Statements longer than 80 columns will be broken into sensible chunks.
  60. Descendants are always substantially shorter than the parent and are placed
  61. substantially to the right. The same applies to function headers with a long
  62. argument list. Long strings are as well broken into shorter strings.
  63. void fun(int a, int b, int c)
  64. {
  65. if (condition)
  66. printk(KERN_WARNING "Warning this is a long printk with "
  67. "3 parameters a: %u b: %u "
  68. "c: %u \n", a, b, c);
  69. else
  70. next_statement;
  71. }
  72. Chapter 3: Placing Braces and Spaces
  73. The other issue that always comes up in C styling is the placement of
  74. braces. Unlike the indent size, there are few technical reasons to
  75. choose one placement strategy over the other, but the preferred way, as
  76. shown to us by the prophets Kernighan and Ritchie, is to put the opening
  77. brace last on the line, and put the closing brace first, thusly:
  78. if (x is true) {
  79. we do y
  80. }
  81. This applies to all non-function statement blocks (if, switch, for,
  82. while, do). E.g.:
  83. switch (action) {
  84. case KOBJ_ADD:
  85. return "add";
  86. case KOBJ_REMOVE:
  87. return "remove";
  88. case KOBJ_CHANGE:
  89. return "change";
  90. default:
  91. return NULL;
  92. }
  93. However, there is one special case, namely functions: they have the
  94. opening brace at the beginning of the next line, thus:
  95. int function(int x)
  96. {
  97. body of function
  98. }
  99. Heretic people all over the world have claimed that this inconsistency
  100. is ... well ... inconsistent, but all right-thinking people know that
  101. (a) K&R are _right_ and (b) K&R are right. Besides, functions are
  102. special anyway (you can't nest them in C).
  103. Note that the closing brace is empty on a line of its own, _except_ in
  104. the cases where it is followed by a continuation of the same statement,
  105. ie a "while" in a do-statement or an "else" in an if-statement, like
  106. this:
  107. do {
  108. body of do-loop
  109. } while (condition);
  110. and
  111. if (x == y) {
  112. ..
  113. } else if (x > y) {
  114. ...
  115. } else {
  116. ....
  117. }
  118. Rationale: K&R.
  119. Also, note that this brace-placement also minimizes the number of empty
  120. (or almost empty) lines, without any loss of readability. Thus, as the
  121. supply of new-lines on your screen is not a renewable resource (think
  122. 25-line terminal screens here), you have more empty lines to put
  123. comments on.
  124. 3.1: Spaces
  125. Linux kernel style for use of spaces depends (mostly) on
  126. function-versus-keyword usage. Use a space after (most) keywords. The
  127. notable exceptions are sizeof, typeof, alignof, and __attribute__, which look
  128. somewhat like functions (and are usually used with parentheses in Linux,
  129. although they are not required in the language, as in: "sizeof info" after
  130. "struct fileinfo info;" is declared).
  131. So use a space after these keywords:
  132. if, switch, case, for, do, while
  133. but not with sizeof, typeof, alignof, or __attribute__. E.g.,
  134. s = sizeof(struct file);
  135. Do not add spaces around (inside) parenthesized expressions. This example is
  136. *bad*:
  137. s = sizeof( struct file );
  138. When declaring pointer data or a function that returns a pointer type, the
  139. preferred use of '*' is adjacent to the data name or function name and not
  140. adjacent to the type name. Examples:
  141. char *linux_banner;
  142. unsigned long long memparse(char *ptr, char **retptr);
  143. char *match_strdup(substring_t *s);
  144. Use one space around (on each side of) most binary and ternary operators,
  145. such as any of these:
  146. = + - < > * / % | & ^ <= >= == != ? :
  147. but no space after unary operators:
  148. & * + - ~ ! sizeof typeof alignof __attribute__ defined
  149. no space before the postfix increment & decrement unary operators:
  150. ++ --
  151. no space after the prefix increment & decrement unary operators:
  152. ++ --
  153. and no space around the '.' and "->" structure member operators.
  154. Chapter 4: Naming
  155. C is a Spartan language, and so should your naming be. Unlike Modula-2
  156. and Pascal programmers, C programmers do not use cute names like
  157. ThisVariableIsATemporaryCounter. A C programmer would call that
  158. variable "tmp", which is much easier to write, and not the least more
  159. difficult to understand.
  160. HOWEVER, while mixed-case names are frowned upon, descriptive names for
  161. global variables are a must. To call a global function "foo" is a
  162. shooting offense.
  163. GLOBAL variables (to be used only if you _really_ need them) need to
  164. have descriptive names, as do global functions. If you have a function
  165. that counts the number of active users, you should call that
  166. "count_active_users()" or similar, you should _not_ call it "cntusr()".
  167. Encoding the type of a function into the name (so-called Hungarian
  168. notation) is brain damaged - the compiler knows the types anyway and can
  169. check those, and it only confuses the programmer. No wonder MicroSoft
  170. makes buggy programs.
  171. LOCAL variable names should be short, and to the point. If you have
  172. some random integer loop counter, it should probably be called "i".
  173. Calling it "loop_counter" is non-productive, if there is no chance of it
  174. being mis-understood. Similarly, "tmp" can be just about any type of
  175. variable that is used to hold a temporary value.
  176. If you are afraid to mix up your local variable names, you have another
  177. problem, which is called the function-growth-hormone-imbalance syndrome.
  178. See chapter 6 (Functions).
  179. Chapter 5: Typedefs
  180. Please don't use things like "vps_t".
  181. It's a _mistake_ to use typedef for structures and pointers. When you see a
  182. vps_t a;
  183. in the source, what does it mean?
  184. In contrast, if it says
  185. struct virtual_container *a;
  186. you can actually tell what "a" is.
  187. Lots of people think that typedefs "help readability". Not so. They are
  188. useful only for:
  189. (a) totally opaque objects (where the typedef is actively used to _hide_
  190. what the object is).
  191. Example: "pte_t" etc. opaque objects that you can only access using
  192. the proper accessor functions.
  193. NOTE! Opaqueness and "accessor functions" are not good in themselves.
  194. The reason we have them for things like pte_t etc. is that there
  195. really is absolutely _zero_ portably accessible information there.
  196. (b) Clear integer types, where the abstraction _helps_ avoid confusion
  197. whether it is "int" or "long".
  198. u8/u16/u32 are perfectly fine typedefs, although they fit into
  199. category (d) better than here.
  200. NOTE! Again - there needs to be a _reason_ for this. If something is
  201. "unsigned long", then there's no reason to do
  202. typedef unsigned long myflags_t;
  203. but if there is a clear reason for why it under certain circumstances
  204. might be an "unsigned int" and under other configurations might be
  205. "unsigned long", then by all means go ahead and use a typedef.
  206. (c) when you use sparse to literally create a _new_ type for
  207. type-checking.
  208. (d) New types which are identical to standard C99 types, in certain
  209. exceptional circumstances.
  210. Although it would only take a short amount of time for the eyes and
  211. brain to become accustomed to the standard types like 'uint32_t',
  212. some people object to their use anyway.
  213. Therefore, the Linux-specific 'u8/u16/u32/u64' types and their
  214. signed equivalents which are identical to standard types are
  215. permitted -- although they are not mandatory in new code of your
  216. own.
  217. When editing existing code which already uses one or the other set
  218. of types, you should conform to the existing choices in that code.
  219. (e) Types safe for use in userspace.
  220. In certain structures which are visible to userspace, we cannot
  221. require C99 types and cannot use the 'u32' form above. Thus, we
  222. use __u32 and similar types in all structures which are shared
  223. with userspace.
  224. Maybe there are other cases too, but the rule should basically be to NEVER
  225. EVER use a typedef unless you can clearly match one of those rules.
  226. In general, a pointer, or a struct that has elements that can reasonably
  227. be directly accessed should _never_ be a typedef.
  228. Chapter 6: Functions
  229. Functions should be short and sweet, and do just one thing. They should
  230. fit on one or two screenfuls of text (the ISO/ANSI screen size is 80x24,
  231. as we all know), and do one thing and do that well.
  232. The maximum length of a function is inversely proportional to the
  233. complexity and indentation level of that function. So, if you have a
  234. conceptually simple function that is just one long (but simple)
  235. case-statement, where you have to do lots of small things for a lot of
  236. different cases, it's OK to have a longer function.
  237. However, if you have a complex function, and you suspect that a
  238. less-than-gifted first-year high-school student might not even
  239. understand what the function is all about, you should adhere to the
  240. maximum limits all the more closely. Use helper functions with
  241. descriptive names (you can ask the compiler to in-line them if you think
  242. it's performance-critical, and it will probably do a better job of it
  243. than you would have done).
  244. Another measure of the function is the number of local variables. They
  245. shouldn't exceed 5-10, or you're doing something wrong. Re-think the
  246. function, and split it into smaller pieces. A human brain can
  247. generally easily keep track of about 7 different things, anything more
  248. and it gets confused. You know you're brilliant, but maybe you'd like
  249. to understand what you did 2 weeks from now.
  250. In source files, separate functions with one blank line. If the function is
  251. exported, the EXPORT* macro for it should follow immediately after the closing
  252. function brace line. E.g.:
  253. int system_is_up(void)
  254. {
  255. return system_state == SYSTEM_RUNNING;
  256. }
  257. EXPORT_SYMBOL(system_is_up);
  258. In function prototypes, include parameter names with their data types.
  259. Although this is not required by the C language, it is preferred in Linux
  260. because it is a simple way to add valuable information for the reader.
  261. Chapter 7: Centralized exiting of functions
  262. Albeit deprecated by some people, the equivalent of the goto statement is
  263. used frequently by compilers in form of the unconditional jump instruction.
  264. The goto statement comes in handy when a function exits from multiple
  265. locations and some common work such as cleanup has to be done.
  266. The rationale is:
  267. - unconditional statements are easier to understand and follow
  268. - nesting is reduced
  269. - errors by not updating individual exit points when making
  270. modifications are prevented
  271. - saves the compiler work to optimize redundant code away ;)
  272. int fun(int a)
  273. {
  274. int result = 0;
  275. char *buffer = kmalloc(SIZE);
  276. if (buffer == NULL)
  277. return -ENOMEM;
  278. if (condition1) {
  279. while (loop1) {
  280. ...
  281. }
  282. result = 1;
  283. goto out;
  284. }
  285. ...
  286. out:
  287. kfree(buffer);
  288. return result;
  289. }
  290. Chapter 8: Commenting
  291. Comments are good, but there is also a danger of over-commenting. NEVER
  292. try to explain HOW your code works in a comment: it's much better to
  293. write the code so that the _working_ is obvious, and it's a waste of
  294. time to explain badly written code.
  295. Generally, you want your comments to tell WHAT your code does, not HOW.
  296. Also, try to avoid putting comments inside a function body: if the
  297. function is so complex that you need to separately comment parts of it,
  298. you should probably go back to chapter 6 for a while. You can make
  299. small comments to note or warn about something particularly clever (or
  300. ugly), but try to avoid excess. Instead, put the comments at the head
  301. of the function, telling people what it does, and possibly WHY it does
  302. it.
  303. When commenting the kernel API functions, please use the kernel-doc format.
  304. See the files Documentation/kernel-doc-nano-HOWTO.txt and scripts/kernel-doc
  305. for details.
  306. Linux style for comments is the C89 "/* ... */" style.
  307. Don't use C99-style "// ..." comments.
  308. The preferred style for long (multi-line) comments is:
  309. /*
  310. * This is the preferred style for multi-line
  311. * comments in the Linux kernel source code.
  312. * Please use it consistently.
  313. *
  314. * Description: A column of asterisks on the left side,
  315. * with beginning and ending almost-blank lines.
  316. */
  317. It's also important to comment data, whether they are basic types or derived
  318. types. To this end, use just one data declaration per line (no commas for
  319. multiple data declarations). This leaves you room for a small comment on each
  320. item, explaining its use.
  321. Chapter 9: You've made a mess of it
  322. That's OK, we all do. You've probably been told by your long-time Unix
  323. user helper that "GNU emacs" automatically formats the C sources for
  324. you, and you've noticed that yes, it does do that, but the defaults it
  325. uses are less than desirable (in fact, they are worse than random
  326. typing - an infinite number of monkeys typing into GNU emacs would never
  327. make a good program).
  328. So, you can either get rid of GNU emacs, or change it to use saner
  329. values. To do the latter, you can stick the following in your .emacs file:
  330. (defun linux-c-mode ()
  331. "C mode with adjusted defaults for use with the Linux kernel."
  332. (interactive)
  333. (c-mode)
  334. (c-set-style "K&R")
  335. (setq tab-width 8)
  336. (setq indent-tabs-mode t)
  337. (setq c-basic-offset 8))
  338. This will define the M-x linux-c-mode command. When hacking on a
  339. module, if you put the string -*- linux-c -*- somewhere on the first
  340. two lines, this mode will be automatically invoked. Also, you may want
  341. to add
  342. (setq auto-mode-alist (cons '("/usr/src/linux.*/.*\\.[ch]$" . linux-c-mode)
  343. auto-mode-alist))
  344. to your .emacs file if you want to have linux-c-mode switched on
  345. automagically when you edit source files under /usr/src/linux.
  346. But even if you fail in getting emacs to do sane formatting, not
  347. everything is lost: use "indent".
  348. Now, again, GNU indent has the same brain-dead settings that GNU emacs
  349. has, which is why you need to give it a few command line options.
  350. However, that's not too bad, because even the makers of GNU indent
  351. recognize the authority of K&R (the GNU people aren't evil, they are
  352. just severely misguided in this matter), so you just give indent the
  353. options "-kr -i8" (stands for "K&R, 8 character indents"), or use
  354. "scripts/Lindent", which indents in the latest style.
  355. "indent" has a lot of options, and especially when it comes to comment
  356. re-formatting you may want to take a look at the man page. But
  357. remember: "indent" is not a fix for bad programming.
  358. Chapter 10: Configuration-files
  359. For configuration options (arch/xxx/Kconfig, and all the Kconfig files),
  360. somewhat different indentation is used.
  361. Help text is indented with 2 spaces.
  362. if CONFIG_EXPERIMENTAL
  363. tristate CONFIG_BOOM
  364. default n
  365. help
  366. Apply nitroglycerine inside the keyboard (DANGEROUS)
  367. bool CONFIG_CHEER
  368. depends on CONFIG_BOOM
  369. default y
  370. help
  371. Output nice messages when you explode
  372. endif
  373. Generally, CONFIG_EXPERIMENTAL should surround all options not considered
  374. stable. All options that are known to trash data (experimental write-
  375. support for file-systems, for instance) should be denoted (DANGEROUS), other
  376. experimental options should be denoted (EXPERIMENTAL).
  377. Chapter 11: Data structures
  378. Data structures that have visibility outside the single-threaded
  379. environment they are created and destroyed in should always have
  380. reference counts. In the kernel, garbage collection doesn't exist (and
  381. outside the kernel garbage collection is slow and inefficient), which
  382. means that you absolutely _have_ to reference count all your uses.
  383. Reference counting means that you can avoid locking, and allows multiple
  384. users to have access to the data structure in parallel - and not having
  385. to worry about the structure suddenly going away from under them just
  386. because they slept or did something else for a while.
  387. Note that locking is _not_ a replacement for reference counting.
  388. Locking is used to keep data structures coherent, while reference
  389. counting is a memory management technique. Usually both are needed, and
  390. they are not to be confused with each other.
  391. Many data structures can indeed have two levels of reference counting,
  392. when there are users of different "classes". The subclass count counts
  393. the number of subclass users, and decrements the global count just once
  394. when the subclass count goes to zero.
  395. Examples of this kind of "multi-level-reference-counting" can be found in
  396. memory management ("struct mm_struct": mm_users and mm_count), and in
  397. filesystem code ("struct super_block": s_count and s_active).
  398. Remember: if another thread can find your data structure, and you don't
  399. have a reference count on it, you almost certainly have a bug.
  400. Chapter 12: Macros, Enums and RTL
  401. Names of macros defining constants and labels in enums are capitalized.
  402. #define CONSTANT 0x12345
  403. Enums are preferred when defining several related constants.
  404. CAPITALIZED macro names are appreciated but macros resembling functions
  405. may be named in lower case.
  406. Generally, inline functions are preferable to macros resembling functions.
  407. Macros with multiple statements should be enclosed in a do - while block:
  408. #define macrofun(a, b, c) \
  409. do { \
  410. if (a == 5) \
  411. do_this(b, c); \
  412. } while (0)
  413. Things to avoid when using macros:
  414. 1) macros that affect control flow:
  415. #define FOO(x) \
  416. do { \
  417. if (blah(x) < 0) \
  418. return -EBUGGERED; \
  419. } while(0)
  420. is a _very_ bad idea. It looks like a function call but exits the "calling"
  421. function; don't break the internal parsers of those who will read the code.
  422. 2) macros that depend on having a local variable with a magic name:
  423. #define FOO(val) bar(index, val)
  424. might look like a good thing, but it's confusing as hell when one reads the
  425. code and it's prone to breakage from seemingly innocent changes.
  426. 3) macros with arguments that are used as l-values: FOO(x) = y; will
  427. bite you if somebody e.g. turns FOO into an inline function.
  428. 4) forgetting about precedence: macros defining constants using expressions
  429. must enclose the expression in parentheses. Beware of similar issues with
  430. macros using parameters.
  431. #define CONSTANT 0x4000
  432. #define CONSTEXP (CONSTANT | 3)
  433. The cpp manual deals with macros exhaustively. The gcc internals manual also
  434. covers RTL which is used frequently with assembly language in the kernel.
  435. Chapter 13: Printing kernel messages
  436. Kernel developers like to be seen as literate. Do mind the spelling
  437. of kernel messages to make a good impression. Do not use crippled
  438. words like "dont" and use "do not" or "don't" instead.
  439. Kernel messages do not have to be terminated with a period.
  440. Printing numbers in parentheses (%d) adds no value and should be avoided.
  441. Chapter 14: Allocating memory
  442. The kernel provides the following general purpose memory allocators:
  443. kmalloc(), kzalloc(), kcalloc(), and vmalloc(). Please refer to the API
  444. documentation for further information about them.
  445. The preferred form for passing a size of a struct is the following:
  446. p = kmalloc(sizeof(*p), ...);
  447. The alternative form where struct name is spelled out hurts readability and
  448. introduces an opportunity for a bug when the pointer variable type is changed
  449. but the corresponding sizeof that is passed to a memory allocator is not.
  450. Casting the return value which is a void pointer is redundant. The conversion
  451. from void pointer to any other pointer type is guaranteed by the C programming
  452. language.
  453. Chapter 15: The inline disease
  454. There appears to be a common misperception that gcc has a magic "make me
  455. faster" speedup option called "inline". While the use of inlines can be
  456. appropriate (for example as a means of replacing macros, see Chapter 11), it
  457. very often is not. Abundant use of the inline keyword leads to a much bigger
  458. kernel, which in turn slows the system as a whole down, due to a bigger
  459. icache footprint for the CPU and simply because there is less memory
  460. available for the pagecache. Just think about it; a pagecache miss causes a
  461. disk seek, which easily takes 5 miliseconds. There are a LOT of cpu cycles
  462. that can go into these 5 miliseconds.
  463. A reasonable rule of thumb is to not put inline at functions that have more
  464. than 3 lines of code in them. An exception to this rule are the cases where
  465. a parameter is known to be a compiletime constant, and as a result of this
  466. constantness you *know* the compiler will be able to optimize most of your
  467. function away at compile time. For a good example of this later case, see
  468. the kmalloc() inline function.
  469. Often people argue that adding inline to functions that are static and used
  470. only once is always a win since there is no space tradeoff. While this is
  471. technically correct, gcc is capable of inlining these automatically without
  472. help, and the maintenance issue of removing the inline when a second user
  473. appears outweighs the potential value of the hint that tells gcc to do
  474. something it would have done anyway.
  475. Chapter 16: Function return values and names
  476. Functions can return values of many different kinds, and one of the
  477. most common is a value indicating whether the function succeeded or
  478. failed. Such a value can be represented as an error-code integer
  479. (-Exxx = failure, 0 = success) or a "succeeded" boolean (0 = failure,
  480. non-zero = success).
  481. Mixing up these two sorts of representations is a fertile source of
  482. difficult-to-find bugs. If the C language included a strong distinction
  483. between integers and booleans then the compiler would find these mistakes
  484. for us... but it doesn't. To help prevent such bugs, always follow this
  485. convention:
  486. If the name of a function is an action or an imperative command,
  487. the function should return an error-code integer. If the name
  488. is a predicate, the function should return a "succeeded" boolean.
  489. For example, "add work" is a command, and the add_work() function returns 0
  490. for success or -EBUSY for failure. In the same way, "PCI device present" is
  491. a predicate, and the pci_dev_present() function returns 1 if it succeeds in
  492. finding a matching device or 0 if it doesn't.
  493. All EXPORTed functions must respect this convention, and so should all
  494. public functions. Private (static) functions need not, but it is
  495. recommended that they do.
  496. Functions whose return value is the actual result of a computation, rather
  497. than an indication of whether the computation succeeded, are not subject to
  498. this rule. Generally they indicate failure by returning some out-of-range
  499. result. Typical examples would be functions that return pointers; they use
  500. NULL or the ERR_PTR mechanism to report failure.
  501. Chapter 17: Don't re-invent the kernel macros
  502. The header file include/linux/kernel.h contains a number of macros that
  503. you should use, rather than explicitly coding some variant of them yourself.
  504. For example, if you need to calculate the length of an array, take advantage
  505. of the macro
  506. #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))
  507. Similarly, if you need to calculate the size of some structure member, use
  508. #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f))
  509. There are also min() and max() macros that do strict type checking if you
  510. need them. Feel free to peruse that header file to see what else is already
  511. defined that you shouldn't reproduce in your code.
  512. Appendix I: References
  513. The C Programming Language, Second Edition
  514. by Brian W. Kernighan and Dennis M. Ritchie.
  515. Prentice Hall, Inc., 1988.
  516. ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback).
  517. URL: http://cm.bell-labs.com/cm/cs/cbook/
  518. The Practice of Programming
  519. by Brian W. Kernighan and Rob Pike.
  520. Addison-Wesley, Inc., 1999.
  521. ISBN 0-201-61586-X.
  522. URL: http://cm.bell-labs.com/cm/cs/tpop/
  523. GNU manuals - where in compliance with K&R and this text - for cpp, gcc,
  524. gcc internals and indent, all available from http://www.gnu.org/manual/
  525. WG14 is the international standardization working group for the programming
  526. language C, URL: http://www.open-std.org/JTC1/SC22/WG14/
  527. Kernel CodingStyle, by greg@kroah.com at OLS 2002:
  528. http://www.kroah.com/linux/talks/ols_2002_kernel_codingstyle_talk/html/
  529. --
  530. Last updated on 2006-December-06.