3 Linux kernel coding style
4 =========================
6 This is a short document describing the preferred coding style for the
7 linux kernel. Coding style is very personal, and I won't **force** my
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.
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.
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
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.
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
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.
41 The preferred way to ease multiple indentation levels in a switch statement is
42 to align the ``switch`` and its subordinate ``case`` labels in the same column
43 instead of ``double-indenting`` the ``case`` labels. E.g.:
64 Don't put multiple statements on a single line unless you have
69 if (condition) do_this;
70 do_something_everytime;
72 Don't put multiple assignments on a single line either. Kernel coding style
73 is super simple. Avoid tricky expressions.
75 Outside of comments, documentation and except in Kconfig, spaces are never
76 used for indentation, and the above example is deliberately broken.
78 Get a decent editor and don't leave whitespace at the end of lines.
81 2) Breaking long lines and strings
82 ----------------------------------
84 Coding style is all about readability and maintainability using commonly
87 The limit on the length of lines is 80 columns and this is a strongly
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.
98 3) Placing Braces and Spaces
99 ----------------------------
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:
113 This applies to all non-function statement blocks (if, switch, for,
129 However, there is one special case, namely functions: they have the
130 opening brace at the beginning of the next line, thus:
139 Heretic people all over the world have claimed that this inconsistency
140 is ... well ... inconsistent, but all right-thinking people know that
141 (a) K&R are **right** and (b) K&R are right. Besides, functions are
142 special anyway (you can't nest them in C).
144 Note that the closing brace is empty on a line of its own, **except** in
145 the cases where it is followed by a continuation of the same statement,
146 ie a ``while`` in a do-statement or an ``else`` in an if-statement, like
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
175 Do not unnecessarily use braces where a single statement will do.
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:
203 Also, use braces when a loop contains more than a single simple statement:
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,
219 although they are not required in the language, as in: ``sizeof info`` after
220 ``struct fileinfo info;`` is declared).
222 So use a space after these keywords::
224 if, switch, case, for, do, while
226 but not with sizeof, typeof, alignof, or __attribute__. E.g.,
231 s = sizeof(struct file);
233 Do not add spaces around (inside) parenthesized expressions. This example is
239 s = sizeof( struct file );
241 When declaring pointer data or a function that returns a pointer type, the
242 preferred use of ``*`` is adjacent to the data name or function name and not
243 adjacent to the type name. Examples:
249 unsigned long long memparse(char *ptr, char **retptr);
250 char *match_strdup(substring_t *s);
252 Use one space around (on each side of) most binary and ternary operators,
253 such as any of these::
255 = + - < > * / % | & ^ <= >= == != ? :
257 but no space after unary operators::
259 & * + - ~ ! sizeof typeof alignof __attribute__ defined
261 no space before the postfix increment & decrement unary operators::
265 no space after the prefix increment & decrement unary operators::
269 and no space around the ``.`` and ``->`` structure member operators.
271 Do not leave trailing whitespace at the ends of lines. Some editors with
272 ``smart`` indentation will insert whitespace at the beginning of new lines as
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.
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
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
290 variable ``tmp``, which is much easier to write, and not the least more
291 difficult to understand.
293 HOWEVER, while mixed-case names are frowned upon, descriptive names for
294 global variables are a must. To call a global function ``foo`` is a
297 GLOBAL variables (to be used only if you **really** need them) need to
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
300 ``count_active_users()`` or similar, you should **not** call it ``cntusr()``.
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.
307 LOCAL variable names should be short, and to the point. If you have
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
311 variable that is used to hold a temporary value.
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.
315 See chapter 6 (Functions).
321 Please don't use things like ``vps_t``.
322 It's a **mistake** to use typedef for structures and pointers. When you see a
329 in the source, what does it mean?
330 In contrast, if it says
334 struct virtual_container *a;
336 you can actually tell what ``a`` is.
338 Lots of people think that typedefs ``help readability``. Not so. They are
341 (a) totally opaque objects (where the typedef is actively used to **hide**
344 Example: ``pte_t`` etc. opaque objects that you can only access using
345 the proper accessor functions.
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.
353 (b) Clear integer types, where the abstraction **helps** avoid confusion
354 whether it is ``int`` or ``long``.
356 u8/u16/u32 are perfectly fine typedefs, although they fit into
357 category (d) better than here.
361 Again - there needs to be a **reason** for this. If something is
362 ``unsigned long``, then there's no reason to do
364 typedef unsigned long myflags_t;
366 but if there is a clear reason for why it under certain circumstances
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.
370 (c) when you use sparse to literally create a **new** type for
373 (d) New types which are identical to standard C99 types, in certain
374 exceptional circumstances.
376 Although it would only take a short amount of time for the eyes and
377 brain to become accustomed to the standard types like ``uint32_t``,
378 some people object to their use anyway.
380 Therefore, the Linux-specific ``u8/u16/u32/u64`` types and their
381 signed equivalents which are identical to standard types are
382 permitted -- although they are not mandatory in new code of your
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.
388 (e) Types safe for use in userspace.
390 In certain structures which are visible to userspace, we cannot
391 require C99 types and cannot use the ``u32`` form above. Thus, we
392 use __u32 and similar types in all structures which are shared
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.
398 In general, a pointer, or a struct that has elements that can reasonably
399 be directly accessed should **never** be a typedef.
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.
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.
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).
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.
430 In source files, separate functions with one blank line. If the function is
431 exported, the **EXPORT** macro for it should follow immediately after the
432 closing function brace line. E.g.:
436 int system_is_up(void)
438 return system_state == SYSTEM_RUNNING;
440 EXPORT_SYMBOL(system_is_up);
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.
446 Do not use the `extern' keyword with function prototypes as this makes
447 lines longer and isn't strictly necessary.
450 7) Centralized exiting of functions
451 -----------------------------------
453 Albeit deprecated by some people, the equivalent of the goto statement is
454 used frequently by compilers in form of the unconditional jump instruction.
456 The goto statement comes in handy when a function exits from multiple
457 locations and some common work such as cleanup has to be done. If there is no
458 cleanup needed then just return directly.
460 Choose label names which say what the goto does or why the goto exists. An
461 example of a good name could be ``out_free_buffer:`` if the goto frees ``buffer``.
462 Avoid using GW-BASIC names like ``err1:`` and ``err2:``, as you would have to
463 renumber them if you ever add or remove exit paths, and they make correctness
464 difficult to verify anyway.
466 The rationale for using gotos is:
468 - unconditional statements are easier to understand and follow
470 - errors by not updating individual exit points when making
471 modifications are prevented
472 - saves the compiler work to optimize redundant code away ;)
481 buffer = kmalloc(SIZE, GFP_KERNEL);
490 goto out_free_buffer;
498 A common type of bug to be aware of is ``one err bugs`` which look like this:
507 The bug in this code is that on some exit paths ``foo`` is NULL. Normally the
508 fix for this is to split it up into two error labels ``err_free_bar:`` and
519 Ideally you should simulate errors to test all exit paths.
525 Comments are good, but there is also a danger of over-commenting. NEVER
526 try to explain HOW your code works in a comment: it's much better to
527 write the code so that the **working** is obvious, and it's a waste of
528 time to explain badly written code.
530 Generally, you want your comments to tell WHAT your code does, not HOW.
531 Also, try to avoid putting comments inside a function body: if the
532 function is so complex that you need to separately comment parts of it,
533 you should probably go back to chapter 6 for a while. You can make
534 small comments to note or warn about something particularly clever (or
535 ugly), but try to avoid excess. Instead, put the comments at the head
536 of the function, telling people what it does, and possibly WHY it does
539 When commenting the kernel API functions, please use the kernel-doc format.
540 See the files at :ref:`Documentation/doc-guide/ <doc_guide>` and
541 ``scripts/kernel-doc`` for details.
543 The preferred style for long (multi-line) comments is:
548 * This is the preferred style for multi-line
549 * comments in the Linux kernel source code.
550 * Please use it consistently.
552 * Description: A column of asterisks on the left side,
553 * with beginning and ending almost-blank lines.
556 For files in net/ and drivers/net/ the preferred style for long (multi-line)
557 comments is a little different.
561 /* The preferred comment style for files in net/ and drivers/net
564 * It is nearly the same as the generally preferred comment style,
565 * but there is no initial almost-blank line.
568 It's also important to comment data, whether they are basic types or derived
569 types. To this end, use just one data declaration per line (no commas for
570 multiple data declarations). This leaves you room for a small comment on each
571 item, explaining its use.
574 9) You've made a mess of it
575 ---------------------------
577 That's OK, we all do. You've probably been told by your long-time Unix
578 user helper that ``GNU emacs`` automatically formats the C sources for
579 you, and you've noticed that yes, it does do that, but the defaults it
580 uses are less than desirable (in fact, they are worse than random
581 typing - an infinite number of monkeys typing into GNU emacs would never
582 make a good program).
584 So, you can either get rid of GNU emacs, or change it to use saner
585 values. To do the latter, you can stick the following in your .emacs file:
589 (defun c-lineup-arglist-tabs-only (ignored)
590 "Line up argument lists by tabs, not spaces"
591 (let* ((anchor (c-langelem-pos c-syntactic-element))
592 (column (c-langelem-2nd-pos c-syntactic-element))
593 (offset (- (1+ column) anchor))
594 (steps (floor offset c-basic-offset)))
598 (add-hook 'c-mode-common-hook
603 '("linux" (c-offsets-alist
604 (arglist-cont-nonempty
606 c-lineup-arglist-tabs-only))))))
608 (add-hook 'c-mode-hook
610 (let ((filename (buffer-file-name)))
611 ;; Enable kernel mode for the appropriate files
613 (string-match (expand-file-name "~/src/linux-trees")
615 (setq indent-tabs-mode t)
616 (setq show-trailing-whitespace t)
617 (c-set-style "linux-tabs-only")))))
619 This will make emacs go better with the kernel coding style for C
620 files below ``~/src/linux-trees``.
622 But even if you fail in getting emacs to do sane formatting, not
623 everything is lost: use ``indent``.
625 Now, again, GNU indent has the same brain-dead settings that GNU emacs
626 has, which is why you need to give it a few command line options.
627 However, that's not too bad, because even the makers of GNU indent
628 recognize the authority of K&R (the GNU people aren't evil, they are
629 just severely misguided in this matter), so you just give indent the
630 options ``-kr -i8`` (stands for ``K&R, 8 character indents``), or use
631 ``scripts/Lindent``, which indents in the latest style.
633 ``indent`` has a lot of options, and especially when it comes to comment
634 re-formatting you may want to take a look at the man page. But
635 remember: ``indent`` is not a fix for bad programming.
637 Note that you can also use the ``clang-format`` tool to help you with
638 these rules, to quickly re-format parts of your code automatically,
639 and to review full files in order to spot coding style mistakes,
640 typos and possible improvements. It is also handy for sorting ``#includes``,
641 for aligning variables/macros, for reflowing text and other similar tasks.
642 See the file :ref:`Documentation/process/clang-format.rst <clangformat>`
646 10) Kconfig configuration files
647 -------------------------------
649 For all of the Kconfig* configuration files throughout the source tree,
650 the indentation is somewhat different. Lines under a ``config`` definition
651 are indented with one tab, while help text is indented an additional two
655 bool "Auditing support"
658 Enable auditing infrastructure that can be used with another
659 kernel subsystem, such as SELinux (which requires this for
660 logging of avc messages output). Does not do system-call
661 auditing without CONFIG_AUDITSYSCALL.
663 Seriously dangerous features (such as write support for certain
664 filesystems) should advertise this prominently in their prompt string::
667 bool "ADFS write support (DANGEROUS)"
671 For full documentation on the configuration files, see the file
672 Documentation/kbuild/kconfig-language.txt.
678 Data structures that have visibility outside the single-threaded
679 environment they are created and destroyed in should always have
680 reference counts. In the kernel, garbage collection doesn't exist (and
681 outside the kernel garbage collection is slow and inefficient), which
682 means that you absolutely **have** to reference count all your uses.
684 Reference counting means that you can avoid locking, and allows multiple
685 users to have access to the data structure in parallel - and not having
686 to worry about the structure suddenly going away from under them just
687 because they slept or did something else for a while.
689 Note that locking is **not** a replacement for reference counting.
690 Locking is used to keep data structures coherent, while reference
691 counting is a memory management technique. Usually both are needed, and
692 they are not to be confused with each other.
694 Many data structures can indeed have two levels of reference counting,
695 when there are users of different ``classes``. The subclass count counts
696 the number of subclass users, and decrements the global count just once
697 when the subclass count goes to zero.
699 Examples of this kind of ``multi-level-reference-counting`` can be found in
700 memory management (``struct mm_struct``: mm_users and mm_count), and in
701 filesystem code (``struct super_block``: s_count and s_active).
703 Remember: if another thread can find your data structure, and you don't
704 have a reference count on it, you almost certainly have a bug.
707 12) Macros, Enums and RTL
708 -------------------------
710 Names of macros defining constants and labels in enums are capitalized.
714 #define CONSTANT 0x12345
716 Enums are preferred when defining several related constants.
718 CAPITALIZED macro names are appreciated but macros resembling functions
719 may be named in lower case.
721 Generally, inline functions are preferable to macros resembling functions.
723 Macros with multiple statements should be enclosed in a do - while block:
727 #define macrofun(a, b, c) \
733 Things to avoid when using macros:
735 1) macros that affect control flow:
745 is a **very** bad idea. It looks like a function call but exits the ``calling``
746 function; don't break the internal parsers of those who will read the code.
748 2) macros that depend on having a local variable with a magic name:
752 #define FOO(val) bar(index, val)
754 might look like a good thing, but it's confusing as hell when one reads the
755 code and it's prone to breakage from seemingly innocent changes.
757 3) macros with arguments that are used as l-values: FOO(x) = y; will
758 bite you if somebody e.g. turns FOO into an inline function.
760 4) forgetting about precedence: macros defining constants using expressions
761 must enclose the expression in parentheses. Beware of similar issues with
762 macros using parameters.
766 #define CONSTANT 0x4000
767 #define CONSTEXP (CONSTANT | 3)
769 5) namespace collisions when defining local variables in macros resembling
781 ret is a common name for a local variable - __foo_ret is less likely
782 to collide with an existing variable.
784 The cpp manual deals with macros exhaustively. The gcc internals manual also
785 covers RTL which is used frequently with assembly language in the kernel.
788 13) Printing kernel messages
789 ----------------------------
791 Kernel developers like to be seen as literate. Do mind the spelling
792 of kernel messages to make a good impression. Do not use crippled
793 words like ``dont``; use ``do not`` or ``don't`` instead. Make the messages
794 concise, clear, and unambiguous.
796 Kernel messages do not have to be terminated with a period.
798 Printing numbers in parentheses (%d) adds no value and should be avoided.
800 There are a number of driver model diagnostic macros in <linux/device.h>
801 which you should use to make sure messages are matched to the right device
802 and driver, and are tagged with the right level: dev_err(), dev_warn(),
803 dev_info(), and so forth. For messages that aren't associated with a
804 particular device, <linux/printk.h> defines pr_notice(), pr_info(),
805 pr_warn(), pr_err(), etc.
807 Coming up with good debugging messages can be quite a challenge; and once
808 you have them, they can be a huge help for remote troubleshooting. However
809 debug message printing is handled differently than printing other non-debug
810 messages. While the other pr_XXX() functions print unconditionally,
811 pr_debug() does not; it is compiled out by default, unless either DEBUG is
812 defined or CONFIG_DYNAMIC_DEBUG is set. That is true for dev_dbg() also,
813 and a related convention uses VERBOSE_DEBUG to add dev_vdbg() messages to
814 the ones already enabled by DEBUG.
816 Many subsystems have Kconfig debug options to turn on -DDEBUG in the
817 corresponding Makefile; in other cases specific files #define DEBUG. And
818 when a debug message should be unconditionally printed, such as if it is
819 already inside a debug-related #ifdef section, printk(KERN_DEBUG ...) can be
823 14) Allocating memory
824 ---------------------
826 The kernel provides the following general purpose memory allocators:
827 kmalloc(), kzalloc(), kmalloc_array(), kcalloc(), vmalloc(), and
828 vzalloc(). Please refer to the API documentation for further information
831 The preferred form for passing a size of a struct is the following:
835 p = kmalloc(sizeof(*p), ...);
837 The alternative form where struct name is spelled out hurts readability and
838 introduces an opportunity for a bug when the pointer variable type is changed
839 but the corresponding sizeof that is passed to a memory allocator is not.
841 Casting the return value which is a void pointer is redundant. The conversion
842 from void pointer to any other pointer type is guaranteed by the C programming
845 The preferred form for allocating an array is the following:
849 p = kmalloc_array(n, sizeof(...), ...);
851 The preferred form for allocating a zeroed array is the following:
855 p = kcalloc(n, sizeof(...), ...);
857 Both forms check for overflow on the allocation size n * sizeof(...),
858 and return NULL if that occurred.
861 15) The inline disease
862 ----------------------
864 There appears to be a common misperception that gcc has a magic "make me
865 faster" speedup option called ``inline``. While the use of inlines can be
866 appropriate (for example as a means of replacing macros, see Chapter 12), it
867 very often is not. Abundant use of the inline keyword leads to a much bigger
868 kernel, which in turn slows the system as a whole down, due to a bigger
869 icache footprint for the CPU and simply because there is less memory
870 available for the pagecache. Just think about it; a pagecache miss causes a
871 disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles
872 that can go into these 5 milliseconds.
874 A reasonable rule of thumb is to not put inline at functions that have more
875 than 3 lines of code in them. An exception to this rule are the cases where
876 a parameter is known to be a compiletime constant, and as a result of this
877 constantness you *know* the compiler will be able to optimize most of your
878 function away at compile time. For a good example of this later case, see
879 the kmalloc() inline function.
881 Often people argue that adding inline to functions that are static and used
882 only once is always a win since there is no space tradeoff. While this is
883 technically correct, gcc is capable of inlining these automatically without
884 help, and the maintenance issue of removing the inline when a second user
885 appears outweighs the potential value of the hint that tells gcc to do
886 something it would have done anyway.
889 16) Function return values and names
890 ------------------------------------
892 Functions can return values of many different kinds, and one of the
893 most common is a value indicating whether the function succeeded or
894 failed. Such a value can be represented as an error-code integer
895 (-Exxx = failure, 0 = success) or a ``succeeded`` boolean (0 = failure,
898 Mixing up these two sorts of representations is a fertile source of
899 difficult-to-find bugs. If the C language included a strong distinction
900 between integers and booleans then the compiler would find these mistakes
901 for us... but it doesn't. To help prevent such bugs, always follow this
904 If the name of a function is an action or an imperative command,
905 the function should return an error-code integer. If the name
906 is a predicate, the function should return a "succeeded" boolean.
908 For example, ``add work`` is a command, and the add_work() function returns 0
909 for success or -EBUSY for failure. In the same way, ``PCI device present`` is
910 a predicate, and the pci_dev_present() function returns 1 if it succeeds in
911 finding a matching device or 0 if it doesn't.
913 All EXPORTed functions must respect this convention, and so should all
914 public functions. Private (static) functions need not, but it is
915 recommended that they do.
917 Functions whose return value is the actual result of a computation, rather
918 than an indication of whether the computation succeeded, are not subject to
919 this rule. Generally they indicate failure by returning some out-of-range
920 result. Typical examples would be functions that return pointers; they use
921 NULL or the ERR_PTR mechanism to report failure.
924 17) Don't re-invent the kernel macros
925 -------------------------------------
927 The header file include/linux/kernel.h contains a number of macros that
928 you should use, rather than explicitly coding some variant of them yourself.
929 For example, if you need to calculate the length of an array, take advantage
934 #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))
936 Similarly, if you need to calculate the size of some structure member, use
940 #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f))
942 There are also min() and max() macros that do strict type checking if you
943 need them. Feel free to peruse that header file to see what else is already
944 defined that you shouldn't reproduce in your code.
947 18) Editor modelines and other cruft
948 ------------------------------------
950 Some editors can interpret configuration information embedded in source files,
951 indicated with special markers. For example, emacs interprets lines marked
964 compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c"
968 Vim interprets markers that look like this:
972 /* vim:set sw=8 noet */
974 Do not include any of these in source files. People have their own personal
975 editor configurations, and your source files should not override them. This
976 includes markers for indentation and mode configuration. People may use their
977 own custom mode, or may have some other magic method for making indentation
984 In architecture-specific code, you may need to use inline assembly to interface
985 with CPU or platform functionality. Don't hesitate to do so when necessary.
986 However, don't use inline assembly gratuitously when C can do the job. You can
987 and should poke hardware from C when possible.
989 Consider writing simple helper functions that wrap common bits of inline
990 assembly, rather than repeatedly writing them with slight variations. Remember
991 that inline assembly can use C parameters.
993 Large, non-trivial assembly functions should go in .S files, with corresponding
994 C prototypes defined in C header files. The C prototypes for assembly
995 functions should use ``asmlinkage``.
997 You may need to mark your asm statement as volatile, to prevent GCC from
998 removing it if GCC doesn't notice any side effects. You don't always need to
999 do so, though, and doing so unnecessarily can limit optimization.
1001 When writing a single inline assembly statement containing multiple
1002 instructions, put each instruction on a separate line in a separate quoted
1003 string, and end each string except the last with ``\n\t`` to properly indent
1004 the next instruction in the assembly output:
1008 asm ("magic %reg1, #42\n\t"
1009 "more_magic %reg2, %reg3"
1010 : /* outputs */ : /* inputs */ : /* clobbers */);
1013 20) Conditional Compilation
1014 ---------------------------
1016 Wherever possible, don't use preprocessor conditionals (#if, #ifdef) in .c
1017 files; doing so makes code harder to read and logic harder to follow. Instead,
1018 use such conditionals in a header file defining functions for use in those .c
1019 files, providing no-op stub versions in the #else case, and then call those
1020 functions unconditionally from .c files. The compiler will avoid generating
1021 any code for the stub calls, producing identical results, but the logic will
1022 remain easy to follow.
1024 Prefer to compile out entire functions, rather than portions of functions or
1025 portions of expressions. Rather than putting an ifdef in an expression, factor
1026 out part or all of the expression into a separate helper function and apply the
1027 conditional to that function.
1029 If you have a function or variable which may potentially go unused in a
1030 particular configuration, and the compiler would warn about its definition
1031 going unused, mark the definition as __maybe_unused rather than wrapping it in
1032 a preprocessor conditional. (However, if a function or variable *always* goes
1035 Within code, where possible, use the IS_ENABLED macro to convert a Kconfig
1036 symbol into a C boolean expression, and use it in a normal C conditional:
1040 if (IS_ENABLED(CONFIG_SOMETHING)) {
1044 The compiler will constant-fold the conditional away, and include or exclude
1045 the block of code just as with an #ifdef, so this will not add any runtime
1046 overhead. However, this approach still allows the C compiler to see the code
1047 inside the block, and check it for correctness (syntax, types, symbol
1048 references, etc). Thus, you still have to use an #ifdef if the code inside the
1049 block references symbols that will not exist if the condition is not met.
1051 At the end of any non-trivial #if or #ifdef block (more than a few lines),
1052 place a comment after the #endif on the same line, noting the conditional
1053 expression used. For instance:
1057 #ifdef CONFIG_SOMETHING
1059 #endif /* CONFIG_SOMETHING */
1062 Appendix I) References
1063 ----------------------
1065 The C Programming Language, Second Edition
1066 by Brian W. Kernighan and Dennis M. Ritchie.
1067 Prentice Hall, Inc., 1988.
1068 ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback).
1070 The Practice of Programming
1071 by Brian W. Kernighan and Rob Pike.
1072 Addison-Wesley, Inc., 1999.
1075 GNU manuals - where in compliance with K&R and this text - for cpp, gcc,
1076 gcc internals and indent, all available from http://www.gnu.org/manual/
1078 WG14 is the international standardization working group for the programming
1079 language C, URL: http://www.open-std.org/JTC1/SC22/WG14/
1081 Kernel :ref:`process/coding-style.rst <codingstyle>`, by greg@kroah.com at OLS 2002:
1082 http://www.kroah.com/linux/talks/ols_2002_kernel_codingstyle_talk/html/