1) Indentation¶

Tabs are 8 characters, and thus indentations are also 8 characters.There are heretic movements that try to make indentations 4 (or even 2!)characters deep, and that is akin to trying to define the value of PI tobe 3.

Rationale: The whole idea behind indentation is to clearly define wherea block of control starts and ends. Especially when you’ve been lookingat your screen for 20 straight hours, you’ll find it a lot easier to seehow the indentation works if you have large indentations.

Now, some people will claim that having 8-character indentations makesthe code move too far to the right, and makes it hard to read on a80-character terminal screen. The answer to that is that if you needmore than 3 levels of indentation, you’re screwed anyway, and should fixyour program.

In short, 8-char indents make things easier to read, and have the addedbenefit of warning you when you’re nesting your functions too deep.Heed that warning.

The preferred way to ease multiple indentation levels in a switch statement isto align theswitch and its subordinatecase labels in the same columninstead ofdouble-indenting thecase labels. E.g.:

switch(suffix){case'G':case'g':mem<<=30;break;case'M':case'm':mem<<=20;break;case'K':case'k':mem<<=10;fallthrough;default:break;}

Don’t put multiple statements on a single line unless you havesomething to hide:

if(condition)do_this;do_something_everytime;

Don’t use commas to avoid using braces:

if(condition)do_this(),do_that();

Always use braces for multiple statements:

if(condition){do_this();do_that();}

Don’t put multiple assignments on a single line either. Kernel coding styleis super simple. Avoid tricky expressions.

Outside of comments, documentation and except in Kconfig, spaces are neverused for indentation, and the above example is deliberately broken.

Get a decent editor and don’t leave whitespace at the end of lines.

2) Breaking long lines and strings¶

Coding style is all about readability and maintainability using commonlyavailable tools.

The preferred limit on the length of a single line is 80 columns.

Statements longer than 80 columns should be broken into sensible chunks,unless exceeding 80 columns significantly increases readability and doesnot hide information.

Descendants are always substantially shorter than the parent andare placed substantially to the right. A very commonly used styleis to align descendants to a function open parenthesis.

These same rules are applied to function headers with a long argument list.

However, never break user-visible strings such as printk messages becausethat breaks the ability to grep for them.

3) Placing Braces and Spaces¶

The other issue that always comes up in C styling is the placement ofbraces. Unlike the indent size, there are few technical reasons tochoose one placement strategy over the other, but the preferred way, asshown to us by the prophets Kernighan and Ritchie, is to put the openingbrace last on the line, and put the closing brace first, thusly:

if(xistrue){wedoy}

This applies to all non-function statement blocks (if, switch, for,while, do). E.g.:

switch(action){caseKOBJ_ADD:return"add";caseKOBJ_REMOVE:return"remove";caseKOBJ_CHANGE:return"change";default:returnNULL;}

However, there is one special case, namely functions: they have theopening brace at the beginning of the next line, thus:

intfunction(intx){bodyoffunction}

Heretic people all over the world have claimed that this inconsistencyis ... well ... inconsistent, but all right-thinking people know that(a) K&R areright and (b) K&R are right. Besides, functions arespecial anyway (you can’t nest them in C).

Note that the closing brace is empty on a line of its own,except inthe cases where it is followed by a continuation of the same statement,ie awhile in a do-statement or anelse in an if-statement, likethis:

do{bodyofdo-loop}while(condition);

and

if(x==y){..}elseif(x>y){...}else{....}

Rationale: K&R.

Also, note that this brace-placement also minimizes the number of empty(or almost empty) lines, without any loss of readability. Thus, as thesupply of new-lines on your screen is not a renewable resource (think25-line terminal screens here), you have more empty lines to putcomments on.

Do not unnecessarily use braces where a single statement will do.

if(condition)action();

and

if(condition)do_this();elsedo_that();

This does not apply if only one branch of a conditional statement is a singlestatement; in the latter case use braces in both branches:

if(condition){do_this();do_that();}else{otherwise();}

Also, use braces when a loop contains more than a single simple statement:

while(condition){if(test)do_something();}

if, switch, case, for, do, while

s=sizeof(structfile);

s=sizeof(structfile);

char*linux_banner;unsignedlonglongmemparse(char*ptr,char**retptr);char*match_strdup(substring_t*s);

=  +  -  <  >  *  /  %  |  &  ^  <=  >=  ==  !=  ?  :

&  *  +  -  ~  !  sizeof  typeof  alignof  __attribute__  defined

++  --

++  --

4) Naming¶

C is a Spartan language, and your naming conventions should follow suit.Unlike Modula-2 and Pascal programmers, C programmers do not use cutenames like ThisVariableIsATemporaryCounter. A C programmer would call thatvariabletmp, which is much easier to write, and not the least moredifficult to understand.

HOWEVER, while mixed-case names are frowned upon, descriptive names forglobal variables are a must. To call a global functionfoo is ashooting offense.

GLOBAL variables (to be used only if youreally need them) need tohave descriptive names, as do global functions. If you have a functionthat counts the number of active users, you should call thatcount_active_users() or similar, you shouldnot call itcntusr().

Encoding the type of a function into the name (so-called Hungariannotation) is asinine - the compiler knows the types anyway and can checkthose, and it only confuses the programmer.

LOCAL variable names should be short, and to the point. If you havesome random integer loop counter, it should probably be calledi.Calling itloop_counter is non-productive, if there is no chance of itbeing mis-understood. Similarly,tmp can be just about any type ofvariable that is used to hold a temporary value.

If you are afraid to mix up your local variable names, you have anotherproblem, which is called the function-growth-hormone-imbalance syndrome.See chapter 6 (Functions).

For symbol names and documentation, avoid introducing new usage of‘master / slave’ (or ‘slave’ independent of ‘master’) and ‘blacklist /whitelist’.

Recommended replacements for ‘master / slave’ are:

‘{primary,main} / {secondary,replica,subordinate}’‘{initiator,requester} / {target,responder}’‘{controller,host} / {device,worker,proxy}’‘leader / follower’‘director / performer’

Recommended replacements for ‘blacklist/whitelist’ are:

‘denylist / allowlist’‘blocklist / passlist’

Exceptions for introducing new usage is to maintain a userspace ABI/API,or when updating code for an existing (as of 2020) hardware or protocolspecification that mandates those terms. For new specificationstranslate specification usage of the terminology to the kernel codingstandard where possible.

5) Typedefs¶

Please don’t use things likevps_t.It’s amistake to use typedef for structures and pointers. When you see a

vps_ta;

in the source, what does it mean?In contrast, if it says

structvirtual_container*a;

you can actually tell whata is.

Lots of people think that typedefshelpreadability. Not so. They areuseful only for:

1. totally opaque objects (where thetypedefis actively used tohidewhat the object is).

   Example:pte_t etc. opaque objects that you can only access usingthe proper accessor functions.

   Note

   Opaqueness andaccessorfunctions are not good in themselves.The reason we have them for things like pte_t etc. is that therereally is absolutelyzero portably accessible information there.

2. Clear integer types, where the abstractionhelps avoid confusionwhether it isint orlong.

   u8/u16/u32 are perfectly fine typedefs, although they fit intocategory (d) better than here.

   Note

   Again - there needs to be areason for this. If something isunsignedlong, then there’s no reason to do

   typedef unsigned long myflags_t;

   but if there is a clear reason for why it under certain circumstancesmight be anunsignedint and under other configurations might beunsignedlong, then by all means go ahead and use a typedef.

3. when you use sparse to literally create anew type fortype-checking.

4. New types which are identical to standard C99 types, in certainexceptional circumstances.

   Although it would only take a short amount of time for the eyes andbrain to become accustomed to the standard types likeuint32_t,some people object to their use anyway.

   Therefore, the Linux-specificu8/u16/u32/u64 types and theirsigned equivalents which are identical to standard types arepermitted -- although they are not mandatory in new code of yourown.

   When editing existing code which already uses one or the other setof types, you should conform to the existing choices in that code.

5. Types safe for use in userspace.

   In certain structures which are visible to userspace, we cannotrequire C99 types and cannot use theu32 form above. Thus, weuse __u32 and similar types in all structures which are sharedwith userspace.

Maybe there are other cases too, but the rule should basically be to NEVEREVER use atypedefunless you can clearly match one of those rules.

In general, a pointer, or astructthat has elements that can reasonablybe directly accessed shouldnever be a typedef.

6) Functions¶

Functions should be short and sweet, and do just one thing. They shouldfit on one or two screenfuls of text (the ISO/ANSI screen size is 80x24,as we all know), and do one thing and do that well.

The maximum length of a function is inversely proportional to thecomplexity and indentation level of that function. So, if you have aconceptually simple function that is just one long (but simple)case-statement, where you have to do lots of small things for a lot ofdifferent cases, it’s OK to have a longer function.

However, if you have a complex function, and you suspect that aless-than-gifted first-year high-school student might not evenunderstand what the function is all about, you should adhere to themaximum limits all the more closely. Use helper functions withdescriptive names (you can ask the compiler to in-line them if you thinkit’s performance-critical, and it will probably do a better job of itthan you would have done).

Another measure of the function is the number of local variables. Theyshouldn’t exceed 5-10, or you’re doing something wrong. Re-think thefunction, and split it into smaller pieces. A human brain cangenerally easily keep track of about 7 different things, anything moreand it gets confused. You know you’re brilliant, but maybe you’d liketo understand what you did 2 weeks from now.

In source files, separate functions with one blank line. If the function isexported, theEXPORT macro for it should follow immediately after theclosing function brace line. E.g.:

intsystem_is_up(void){returnsystem_state==SYSTEM_RUNNING;}EXPORT_SYMBOL(system_is_up);

__init void * __must_check action(enum magic value, size_t size, u8 count,                                  char *fmt, ...) __printf(4, 5) __malloc;

static __always_inline __init __printf(4, 5) void * __must_check action(enum magic value,               size_t size, u8 count, char *fmt, ...) __malloc{       ...}

7) Centralized exiting of functions¶

Albeit deprecated by some people, the equivalent of the goto statement isused frequently by compilers in form of the unconditional jump instruction.

The goto statement comes in handy when a function exits from multiplelocations and some common work such as cleanup has to be done. If there is nocleanup needed then just return directly.

Choose label names which say what the goto does or why the goto exists. Anexample of a good name could beout_free_buffer: if the goto freesbuffer.Avoid using GW-BASIC names likeerr1: anderr2:, as you would have torenumber them if you ever add or remove exit paths, and they make correctnessdifficult to verify anyway.

The rationale for using gotos is:

• unconditional statements are easier to understand and follow

• nesting is reduced

• errors by not updating individual exit points when makingmodifications are prevented

• saves the compiler work to optimize redundant code away ;)

intfun(inta){intresult=0;char*buffer;buffer=kmalloc(SIZE,GFP_KERNEL);if(!buffer)return-ENOMEM;if(condition1){while(loop1){...}result=1;gotoout_free_buffer;}...out_free_buffer:kfree(buffer);returnresult;}

A common type of bug to be aware of isoneerrbugs which look like this:

err:kfree(foo->bar);kfree(foo);returnret;

The bug in this code is that on some exit pathsfoo is NULL. Normally thefix for this is to split it up into two error labelserr_free_bar: anderr_free_foo::

err_free_bar:kfree(foo->bar);err_free_foo:kfree(foo);returnret;

Ideally you should simulate errors to test all exit paths.

8) Commenting¶

Comments are good, but there is also a danger of over-commenting. NEVERtry to explain HOW your code works in a comment: it’s much better towrite the code so that theworking is obvious, and it’s a waste oftime to explain badly written code.

Generally, you want your comments to tell WHAT your code does, not HOW.Also, try to avoid putting comments inside a function body: if thefunction is so complex that you need to separately comment parts of it,you should probably go back to chapter 6 for a while. You can makesmall comments to note or warn about something particularly clever (orugly), but try to avoid excess. Instead, put the comments at the headof the function, telling people what it does, and possibly WHY it doesit.

When commenting the kernel API functions, please use the kernel-doc format.See the files atDocumentation/doc-guide/ andscripts/kernel-doc for details. Note that the danger of over-commentingapplies to kernel-doc comments all the same. Do not add boilerplatekernel-doc which simply reiterates what’s obvious from the signatureof the function.

The preferred style for long (multi-line) comments is:

/* * This is the preferred style for multi-line * comments in the Linux kernel source code. * Please use it consistently. * * Description:  A column of asterisks on the left side, * with beginning and ending almost-blank lines. */

It’s also important to comment data, whether they are basic types or derivedtypes. To this end, use just one data declaration per line (no commas formultiple data declarations). This leaves you room for a small comment on eachitem, explaining its use.

9) You’ve made a mess of it¶

That’s OK, we all do. You’ve probably been told by your long-time Unixuser helper thatGNUemacs automatically formats the C sources foryou, and you’ve noticed that yes, it does do that, but the defaults ituses are less than desirable (in fact, they are worse than randomtyping - an infinite number of monkeys typing into GNU emacs would nevermake a good program).

So, you can either get rid of GNU emacs, or change it to use sanervalues. To do the latter, you can stick the following in your .emacs file:

(defunc-lineup-arglist-tabs-only(ignored)"Line up argument lists by tabs, not spaces"(let*((anchor(c-langelem-posc-syntactic-element))(column(c-langelem-2nd-posc-syntactic-element))(offset(-(1+column)anchor))(steps(flooroffsetc-basic-offset)))(*(maxsteps1)c-basic-offset)))(dir-locals-set-class-variables'linux-kernel'((c-mode.((c-basic-offset.8)(c-label-minimum-indentation.0)(c-offsets-alist.((arglist-close.c-lineup-arglist-tabs-only)(arglist-cont-nonempty.(c-lineup-gcc-asm-regc-lineup-arglist-tabs-only))(arglist-intro.+)(brace-list-intro.+)(c.c-lineup-C-comments)(case-label.0)(comment-intro.c-lineup-comment)(cpp-define-intro.+)(cpp-macro.-1000)(cpp-macro-cont.+)(defun-block-intro.+)(else-clause.0)(func-decl-cont.+)(inclass.+)(inher-cont.c-lineup-multi-inher)(knr-argdecl-intro.0)(label.-1000)(statement.0)(statement-block-intro.+)(statement-case-intro.+)(statement-cont.+)(substatement.+)))(indent-tabs-mode.t)(show-trailing-whitespace.t)))))(dir-locals-set-directory-class(expand-file-name"~/src/linux-trees")'linux-kernel)

This will make emacs go better with the kernel coding style for Cfiles below~/src/linux-trees.

But even if you fail in getting emacs to do sane formatting, noteverything is lost: useindent.

Now, again, GNU indent has the same brain-dead settings that GNU emacshas, which is why you need to give it a few command line options.However, that’s not too bad, because even the makers of GNU indentrecognize the authority of K&R (the GNU people aren’t evil, they arejust severely misguided in this matter), so you just give indent theoptions-kr-i8 (stands forK&R,8characterindents), or usescripts/Lindent, which indents in the latest style.

indent has a lot of options, and especially when it comes to commentre-formatting you may want to take a look at the man page. Butremember:indent is not a fix for bad programming.

Note that you can also use theclang-format tool to help you withthese rules, to quickly re-format parts of your code automatically,and to review full files in order to spot coding style mistakes,typos and possible improvements. It is also handy for sorting#includes,for aligning variables/macros, for reflowing text and other similar tasks.See the fileDocumentation/dev-tools/clang-format.rstfor more details.

Some basic editor settings, such as indentation and line endings, will beset automatically if you are using an editor that is compatible withEditorConfig. See the official EditorConfig website for more information:https://editorconfig.org/

10) Kconfig configuration files¶

For all of the Kconfig* configuration files throughout the source tree,the indentation is somewhat different. Lines under aconfig definitionare indented with one tab, while help text is indented an additional twospaces. Example:

config AUDIT      bool "Auditing support"      depends on NET      help        Enable auditing infrastructure that can be used with another        kernel subsystem, such as SELinux (which requires this for        logging of avc messages output).  Does not do system-call        auditing without CONFIG_AUDITSYSCALL.

Seriously dangerous features (such as write support for certainfilesystems) should advertise this prominently in their prompt string:

config ADFS_FS_RW      bool "ADFS write support (DANGEROUS)"      depends on ADFS_FS      ...

For full documentation on the configuration files, see the fileKconfig Language.

11) Data structures¶

Data structures that have visibility outside the single-threadedenvironment they are created and destroyed in should always havereference counts. In the kernel, garbage collection doesn’t exist (andoutside the kernel garbage collection is slow and inefficient), whichmeans that you absolutelyhave to reference count all your uses.

Reference counting means that you can avoid locking, and allows multipleusers to have access to the data structure in parallel - and not havingto worry about the structure suddenly going away from under them justbecause they slept or did something else for a while.

Note that locking isnot a replacement for reference counting.Locking is used to keep data structures coherent, while referencecounting is a memory management technique. Usually both are needed, andthey are not to be confused with each other.

Many data structures can indeed have two levels of reference counting,when there are users of differentclasses. The subclass count countsthe number of subclass users, and decrements the global count just oncewhen the subclass count goes to zero.

Examples of this kind ofmulti-level-reference-counting can be found inmemory management (structmm_struct: mm_users and mm_count), and infilesystem code (structsuper_block: s_count and s_active).

Remember: if another thread can find your data structure, and you don’thave a reference count on it, you almost certainly have a bug.

12) Macros, Enums and RTL¶

Names of macros defining constants and labels in enums are capitalized.

#define CONSTANT 0x12345

Enums are preferred when defining several related constants.

CAPITALIZED macro names are appreciated but macros resembling functionsmay be named in lower case.

Generally, inline functions are preferable to macros resembling functions.

Macros with multiple statements should be enclosed in a do - while block:

#define macrofun(a, b, c)                       \        do {                                    \                if (a == 5)                     \                        do_this(b, c);          \        } while (0)

Function-like macros with unused parameters should be replaced by staticinline functions to avoid the issue of unused variables:

staticinlinevoidfun(structfoo*foo){}

Due to historical practices, many files still employ the “cast to (void)”approach to evaluate parameters. However, this method is not advisable.Inline functions address the issue of “expression with side effectsevaluated more than once”, circumvent unused-variable problems, andare generally better documented than macros for some reason.

/* * Avoid doing this whenever possible and instead opt for static * inline functions */#define macrofun(foo) do { (void) (foo); } while (0)

Things to avoid when using macros:

1. macros that affect control flow:

#define FOO(x)                                  \        do {                                    \                if (blah(x) < 0)                \                        return -EBUGGERED;      \        } while (0)

is avery bad idea. It looks like a function call but exits thecallingfunction; don’t break the internal parsers of those who will read the code.

2. macros that depend on having a local variable with a magic name:

#define FOO(val) bar(index, val)

might look like a good thing, but it’s confusing as hell when one reads thecode and it’s prone to breakage from seemingly innocent changes.

3) macros with arguments that are used as l-values: FOO(x) = y; willbite you if somebody e.g. turns FOO into an inline function.

4) forgetting about precedence: macros defining constants using expressionsmust enclose the expression in parentheses. Beware of similar issues withmacros using parameters.

#define CONSTANT 0x4000#define CONSTEXP (CONSTANT | 3)

5) namespace collisions when defining local variables in macros resemblingfunctions:

#define FOO(x)                          \({                                      \        typeof(x) ret;                  \        ret = calc_ret(x);              \        (ret);                          \})

ret is a common name for a local variable - __foo_ret is less likelyto collide with an existing variable.

The cpp manual deals with macros exhaustively. The gcc internals manual alsocovers RTL which is used frequently with assembly language in the kernel.

13) Printing kernel messages¶

Kernel developers like to be seen as literate. Do mind the spellingof kernel messages to make a good impression. Do not use incorrectcontractions likedont; usedonot ordon't instead. Make themessages concise, clear, and unambiguous.

Kernel messages do not have to be terminated with a period.

Printing numbers in parentheses (%d) adds no value and should be avoided.

There are a number of driver model diagnostic macros in <linux/dev_printk.h>which you should use to make sure messages are matched to the right deviceand driver, and are tagged with the right level:dev_err(),dev_warn(),dev_info(), and so forth. For messages that aren’t associated with aparticular device, <linux/printk.h> definespr_notice(),pr_info(),pr_warn(),pr_err(), etc. When drivers are working properly they are quiet,so prefer to use dev_dbg/pr_debug unless something is wrong.

Coming up with good debugging messages can be quite a challenge; and onceyou have them, they can be a huge help for remote troubleshooting. Howeverdebug message printing is handled differently than printing other non-debugmessages. While the otherpr_XXX() functions print unconditionally,pr_debug() does not; it is compiled out by default, unless either DEBUG isdefined or CONFIG_DYNAMIC_DEBUG is set. That is true fordev_dbg() also,and a related convention uses VERBOSE_DEBUG to adddev_vdbg() messages tothe ones already enabled by DEBUG.

Many subsystems have Kconfig debug options to turn on -DDEBUG in thecorresponding Makefile; in other cases specific files #define DEBUG. Andwhen a debug message should be unconditionally printed, such as if it isalready inside a debug-related #ifdef section, printk(KERN_DEBUG ...) can beused.

14) Allocating memory¶

The kernel provides the following general purpose memory allocators:kmalloc(),kzalloc(),kmalloc_array(),kcalloc(),vmalloc(), andvzalloc(). Please refer to the API documentation for further informationabout them.Documentation/core-api/memory-allocation.rst

The preferred form for passing a size of astructis the following:

p=kmalloc(sizeof(*p),...);

The alternative form wherestructname is spelled out hurts readability andintroduces an opportunity for a bug when the pointer variable type is changedbut the corresponding sizeof that is passed to a memory allocator is not.

Casting the return value which is a void pointer is redundant. The conversionfrom void pointer to any other pointer type is guaranteed by the C programminglanguage.

The preferred form for allocating an array is the following:

p=kmalloc_array(n,sizeof(...),...);

The preferred form for allocating a zeroed array is the following:

p=kcalloc(n,sizeof(...),...);

Both forms check for overflow on the allocation size n * sizeof(...),and return NULL if that occurred.

These generic allocation functions all emit a stack dump on failure when usedwithout __GFP_NOWARN so there is no use in emitting an additional failuremessage when NULL is returned.

15) The inline disease¶

There appears to be a common misperception that gcc has a magic “make mefaster” speedup option calledinline. While the use of inlines can beappropriate (for example as a means of replacing macros, see Chapter 12), itvery often is not. Abundant use of the inline keyword leads to a much biggerkernel, which in turn slows the system as a whole down, due to a biggericache footprint for the CPU and simply because there is less memoryavailable for the pagecache. Just think about it; a pagecache miss causes adisk seek, which easily takes 5 milliseconds. There are a LOT of cpu cyclesthat can go into these 5 milliseconds.

A reasonable rule of thumb is to not put inline at functions that have morethan 3 lines of code in them. An exception to this rule are the cases wherea parameter is known to be a compile time constant, and as a result of thisconstantness youknow the compiler will be able to optimize most of yourfunction away at compile time. For a good example of this later case, seethekmalloc() inline function.

Often people argue that adding inline to functions that are static and usedonly once is always a win since there is no space tradeoff. While this istechnically correct, gcc is capable of inlining these automatically withouthelp, and the maintenance issue of removing the inline when a second userappears outweighs the potential value of the hint that tells gcc to dosomething it would have done anyway.

16) Function return values and names¶

Functions can return values of many different kinds, and one of themost common is a value indicating whether the function succeeded orfailed. Such a value can be represented as an error-code integer(-Exxx = failure, 0 = success) or asucceeded boolean (0 = failure,non-zero = success).

Mixing up these two sorts of representations is a fertile source ofdifficult-to-find bugs. If the C language included a strong distinctionbetween integers and booleans then the compiler would find these mistakesfor us... but it doesn’t. To help prevent such bugs, always follow thisconvention:

If the name of a function is an action or an imperative command,the function should return an error-code integer.  If the nameis a predicate, the function should return a "succeeded" boolean.

For example,addwork is a command, and theadd_work() function returns 0for success or -EBUSY for failure. In the same way,PCIdevicepresent isa predicate, and thepci_dev_present() function returns 1 if it succeeds infinding a matching device or 0 if it doesn’t.

All EXPORTed functions must respect this convention, and so should allpublic functions. Private (static) functions need not, but it isrecommended that they do.

Functions whose return value is the actual result of a computation, ratherthan an indication of whether the computation succeeded, are not subject tothis rule. Generally they indicate failure by returning some out-of-rangeresult. Typical examples would be functions that return pointers; they useNULL or the ERR_PTR mechanism to report failure.

17) Using bool¶

The Linux kernel bool type is an alias for the C99 _Bool type. bool values canonly evaluate to 0 or 1, and implicit or explicit conversion to boolautomatically converts the value to true or false. When using bool types the!! construction is not needed, which eliminates a class of bugs.

When working with bool values the true and false definitions should be usedinstead of 1 and 0.

bool function return types and stack variables are always fine to use wheneverappropriate. Use of bool is encouraged to improve readability and is often abetter option than ‘int’ for storing boolean values.

Do not use bool if cache line layout or size of the value matters, as its sizeand alignment varies based on the compiled architecture. Structures that areoptimized for alignment and size should not use bool.

If a structure has many true/false values, consider consolidating them into abitfield with 1 bit members, or using an appropriate fixed width type, such asu8.

Similarly for function arguments, many true/false values can be consolidatedinto a single bitwise ‘flags’ argument and ‘flags’ can often be a morereadable alternative if the call-sites have naked true/false constants.

Otherwise limited use of bool in structures and arguments can improvereadability.

18) Don’t re-invent the kernel macros¶

The header file include/linux/kernel.h contains a number of macros thatyou should use, rather than explicitly coding some variant of them yourself.For example, if you need to calculate the length of an array, take advantageof the macro

#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))

Similarly, if you need to calculate the size of some structure member, use

#define sizeof_field(t, f) (sizeof(((t*)0)->f))

There are alsomin() andmax() macros that do strict type checking if youneed them. Feel free to peruse that header file to see what else is alreadydefined that you shouldn’t reproduce in your code.

19) Editor modelines and other cruft¶

Some editors can interpret configuration information embedded in source files,indicated with special markers. For example, emacs interprets lines markedlike this:

-*-mode:c-*-

Or like this:

/*Local Variables:compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c"End:*/

Vim interprets markers that look like this:

/* vim:set sw=8 noet */

Do not include any of these in source files. People have their own personaleditor configurations, and your source files should not override them. Thisincludes markers for indentation and mode configuration. People may use theirown custom mode, or may have some other magic method for making indentationwork correctly.

20) Inline assembly¶

In architecture-specific code, you may need to use inline assembly to interfacewith CPU or platform functionality. Don’t hesitate to do so when necessary.However, don’t use inline assembly gratuitously when C can do the job. You canand should poke hardware from C when possible.

Consider writing simple helper functions that wrap common bits of inlineassembly, rather than repeatedly writing them with slight variations. Rememberthat inline assembly can use C parameters.

Large, non-trivial assembly functions should go in .S files, with correspondingC prototypes defined in C header files. The C prototypes for assemblyfunctions should useasmlinkage.

You may need to mark your asm statement as volatile, to prevent GCC fromremoving it if GCC doesn’t notice any side effects. You don’t always need todo so, though, and doing so unnecessarily can limit optimization.

When writing a single inline assembly statement containing multipleinstructions, put each instruction on a separate line in a separate quotedstring, and end each string except the last with\n\t to properly indentthe next instruction in the assembly output:

asm("magic %reg1, #42\n\t""more_magic %reg2, %reg3":/* outputs */:/* inputs */:/* clobbers */);

21) Conditional Compilation¶

Wherever possible, don’t use preprocessor conditionals (#if, #ifdef) in .cfiles; doing so makes code harder to read and logic harder to follow. Instead,use such conditionals in a header file defining functions for use in those .cfiles, providing no-op stub versions in the #else case, and then call thosefunctions unconditionally from .c files. The compiler will avoid generatingany code for the stub calls, producing identical results, but the logic willremain easy to follow.

Prefer to compile out entire functions, rather than portions of functions orportions of expressions. Rather than putting an ifdef in an expression, factorout part or all of the expression into a separate helper function and apply theconditional to that function.

If you have a function or variable which may potentially go unused in aparticular configuration, and the compiler would warn about its definitiongoing unused, mark the definition as __maybe_unused rather than wrapping it ina preprocessor conditional. (However, if a function or variablealways goesunused, delete it.)

Within code, where possible, use the IS_ENABLED macro to convert a Kconfigsymbol into a C boolean expression, and use it in a normal C conditional:

if(IS_ENABLED(CONFIG_SOMETHING)){...}

The compiler will constant-fold the conditional away, and include or excludethe block of code just as with an #ifdef, so this will not add any runtimeoverhead. However, this approach still allows the C compiler to see the codeinside the block, and check it for correctness (syntax, types, symbolreferences, etc). Thus, you still have to use an #ifdef if the code inside theblock references symbols that will not exist if the condition is not met.

At the end of any non-trivial #if or #ifdef block (more than a few lines),place a comment after the #endif on the same line, noting the conditionalexpression used. For instance:

#ifdef CONFIG_SOMETHING...#endif/* CONFIG_SOMETHING */

22) Do not crash the kernel¶

In general, the decision to crash the kernel belongs to the user, ratherthan to the kernel developer.

Appendix I) References¶

The C Programming Language, Second Editionby Brian W. Kernighan and Dennis M. Ritchie.Prentice Hall, Inc., 1988.ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback).

The Practice of Programmingby Brian W. Kernighan and Rob Pike.Addison-Wesley, Inc., 1999.ISBN 0-201-61586-X.

GNU manuals - where in compliance with K&R and this text - for cpp, gcc,gcc internals and indent, all available fromhttps://www.gnu.org/manual/

WG14 is the international standardization working group for the programminglanguage C, URL:http://www.open-std.org/JTC1/SC22/WG14/

Kernel CodingStyle, bygreg@kroah.com at OLS 2002:http://www.kroah.com/linux/talks/ols_2002_kernel_codingstyle_talk/html/