From 119dcf6bd689a20951a2460f1e6e43f66c3368cb Mon Sep 17 00:00:00 2001 From: Natalia Ogoreltseva Date: Mon, 14 Dec 2020 14:51:31 +0300 Subject: [PATCH 1/8] gh-1031 refactor C Style Guide --- doc/dev_guide/c_style_guide.rst | 1111 ++++++++++++++------------- doc/dev_guide/c_style_guide_new.rst | 1046 +++++++++++++++++++++++++ doc/dev_guide/c_style_guide_old.rst | 1037 +++++++++++++++++++++++++ 3 files changed, 2643 insertions(+), 551 deletions(-) create mode 100644 doc/dev_guide/c_style_guide_new.rst create mode 100644 doc/dev_guide/c_style_guide_old.rst diff --git a/doc/dev_guide/c_style_guide.rst b/doc/dev_guide/c_style_guide.rst index e8aebf7141..4657470d05 100644 --- a/doc/dev_guide/c_style_guide.rst +++ b/doc/dev_guide/c_style_guide.rst @@ -2,124 +2,36 @@ C Style Guide ------------------------------------------------------------------------------- -The project's coding style is based on a version of the Linux kernel coding style. - -The latest version of the Linux style can be found at: -http://www.kernel.org/doc/Documentation/CodingStyle - -Since it is open for changes, the version of style that we follow, -one from 2007-July-13, will be also copied later in this document. - -There are a few additional guidelines, either unique -to Tarantool or deviating from the Kernel guidelines. - -A. Chapters 10 "Kconfig configuration files", 11 "Data structures", - 13 "Printing kernel messages", 14 "Allocating memory" and 17 - "Don't re-invent the kernel macros" do not apply, since they are - specific to Linux kernel programming environment. - -B. The rest of Linux Kernel Coding Style is amended as follows: - -=========================================================== - General guidelines -=========================================================== - We use Git for revision control. The latest development is happening in the -default branch (currently ``master``). -Our git repository is hosted on github, and can be checked out with -``git clone git://github.com/tarantool/tarantool.git`` (anonymous read-only access). +default branch (currently ``master``). Our git repository is hosted on GitHub, +and can be checked out with ``git clone git://github.com/tarantool/tarantool.git`` +(anonymous read-only access). If you have any questions about Tarantool internals, please post them on the -developer discussion list, https://groups.google.com/forum/#!forum/tarantool. However, -please be warned: Launchpad silently deletes posts from non-subscribed members, -thus please be sure to have subscribed to the list prior to posting. Additionally, -some engineers are always present on #tarantool channel on irc.freenode.net. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Commenting style -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Use Doxygen comment format, Javadoc flavor, i.e. `@tag` rather than `\\tag`. -The main tags in use are @param, @retval, @return, @see, @note and @todo. - -Every function, except perhaps a very short and obvious one, should have a -comment. A sample function comment may look like below: - -.. code-block:: c - - /** Write all data to a descriptor. - * - * This function is equivalent to 'write', except it would ensure - * that all data is written to the file unless a non-ignorable - * error occurs. - * - * @retval 0 Success - * - * @retval 1 An error occurred (not EINTR) - * / - static int - write_all(int fd, void \*data, size_t len); - -Public structures and important structure members should be commented as well. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Header files -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Use header guards. Put the header guard in the first line in the header, -before the copyright or declarations. Use all-uppercase name for the header -guard. Derive the header guard name from the file name, and append _INCLUDED -to get a macro name. For example, core/log_io.h -> CORE_LOG_IO_H_INCLUDED. In -``.c`` (implementation) file, include the respective declaration header before all -other headers, to ensure that the header is self- sufficient. Header "header.h" -is self-sufficient if the following compiles without errors: - -.. code-block:: c - - #include "header.h" - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Allocating memory -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Prefer the supplied slab (salloc) and pool (palloc) allocators to malloc()/free() -for any performance-intensive or large memory allocations. Repetitive use of -malloc()/free() can lead to memory fragmentation and should therefore be avoided. - -Always free all allocated memory, even allocated at start-up. We aim at being -valgrind leak-check clean, and in most cases it's just as easy to free() the -allocated memory as it is to write a valgrind suppression. Freeing all allocated -memory is also dynamic-load friendly: assuming a plug-in can be dynamically loaded -and unloaded multiple times, reload should not lead to a memory leak. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Function naming -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Our convention is to use: +`developer discussion list `_ +or on `StackOverflow `_. +Additionally, some engineers are always present on #tarantool channel on +irc.freenode.net. -* ``new``/``delete`` for functions which - allocate + initialize and destroy + deallocate an object, -* ``create``/``destroy`` for functions which initialize/destroy an object - but do not handle memory management, -* ``init``/``free`` for functions which initialize/destroy libraries and subsystems. +**General guidelines** -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Other -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +The project's coding style is based on the `Linux kernel coding style +`_. -Select GNU C99 extensions are acceptable. It's OK to mix declarations and statements, -use true and false. +However, we have some additional guidelines, either unique to Tarantool or +deviating from the Kernel guidelines. Below we cite the Linux kernel +coding style noting Tarantool's style features. -The not-so-current list of all GCC C extensions can be found at: -http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html +We don't cite chapters 10 "Kconfig configuration files", 11 "Data structures", +13 "Printing kernel messages", and 17 "Don't re-invent the kernel macros" since +they are specific to Linux kernel programming environment. -=========================================================== - Linux kernel coding style -=========================================================== +================================================================================ + Linux kernel coding style +================================================================================ This is a short document describing the preferred coding style for the -linux kernel. Coding style is very personal, and I won't _force_ my +linux kernel. Coding style is very personal, and I won't **force** my views on anybody, but this is what goes for anything that I have to be able to maintain, and I'd prefer it for most other things too. Please at least consider the points made here. @@ -130,13 +42,13 @@ and NOT read it. Burn them, it's a great symbolic gesture. Anyway, here goes: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 1: Indentation +Chapter 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 to -be 3. +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 to be 3. Rationale: The whole idea behind indentation is to clearly define where a block of control starts and ends. Especially when you've been looking @@ -154,105 +66,96 @@ benefit 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 is -to align the "switch" and its subordinate "case" labels in the same column -instead of "double-indenting" the "case" labels. e.g.: +to align the ``switch`` and its subordinate ``case`` labels in the same column +instead of ``double-indenting`` the ``case`` labels. E.g.: .. code-block:: c switch (suffix) { case 'G': case 'g': - mem <<= 30; - break; + mem <<= 30; + break; case 'M': case 'm': - mem <<= 20; - break; + mem <<= 20; + break; case 'K': case 'k': - mem <<= 10; - /* fall through */ + mem <<= 10; + /* fall through */ default: - break; + break; } - Don't put multiple statements on a single line unless you have something to hide: -.. code-block:: none +.. code-block:: c - if (condition) do_this; - do_something_everytime; + if (condition) do_this; + do_something_everytime; Don't put multiple assignments on a single line either. Kernel coding style is super simple. Avoid tricky expressions. -Outside of comments, documentation and except in Kconfig, spaces are never +Outside of comments and documentation, spaces are never used for indentation, and the above example is deliberately broken. Get a decent editor and don't leave whitespace at the end of lines. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 2: Breaking long lines and strings -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 2: Breaking long lines and strings +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Coding style is all about readability and maintainability using commonly available tools. -The limit on the length of lines is 80 columns, and this is a strongly -preferred limit. As for comments, the same limit of 80 columns is applied. +The limit on the length of lines is 80 columns and this is a strongly +preferred limit. -Statements longer than 80 columns will be broken into sensible chunks. -Descendants are always substantially shorter than the parent and are placed -substantially to the right. The same applies to function headers with a long -argument list. Long strings are as well broken into shorter strings. The -only exception to this is where exceeding 80 columns significantly increases -readability and does not hide information. +.. admonition:: Tarantool Style + :class: FACT -.. code-block:: c + As for comments, the same limit of 80 columns is applied. - void fun(int a, int b, int c) - { - if (condition) - printk(KERN_WARNING "Warning this is a long printk with " - "3 parameters a: %u b: %u " - "c: %u \n", a, b, c); - else - next_statement; - } +Statements longer than 80 columns will be broken into sensible chunks, unless +exceeding 80 columns significantly increases readability and does not hide +information. Descendants are always substantially shorter than the parent and +are placed substantially to the right. The same applies to function headers +with a long argument list. However, never break user-visible strings such as +printk messages, because that breaks the ability to grep for them. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 3: Placing Braces and Spaces -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 3: Placing Braces and Spaces +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The other issue that always comes up in C styling is the placement of -braces. Unlike the indent size, there are few technical reasons to +braces. Unlike the indent size, there are few technical reasons to choose one placement strategy over the other, but the preferred way, as shown to us by the prophets Kernighan and Ritchie, is to put the opening -brace last on the line, and put the closing brace first, thusly: +brace last on the line, and put the closing brace first, thus: -.. code-block:: none +.. code-block:: c if (x is true) { - we do y + we do y } This applies to all non-function statement blocks (if, switch, for, -while, do). e.g.: +while, do). E.g.: .. code-block:: c switch (action) { case KOBJ_ADD: - return "add"; + return "add"; case KOBJ_REMOVE: - return "remove"; + return "remove"; case KOBJ_CHANGE: - return "change"; + return "change"; default: - return NULL; + return NULL; } However, there is one special case, namely functions: they have the @@ -260,25 +163,26 @@ opening brace at the beginning of the next line, thus: .. code-block:: c - int function(int x) + int + function(int x) { - body of function; + body of function } Heretic people all over the world have claimed that this inconsistency is ... well ... inconsistent, but all right-thinking people know that -(a) K&R are _right_ and (b) K&R are right. Besides, functions are +(a) K&R are **right** and (b) K&R are right. Besides, functions are special anyway (you can't nest them in C). -Note that the closing brace is empty on a line of its own, _except_ in +Note that the closing brace is empty on a line of its own, **except** in the cases where it is followed by a continuation of the same statement, -ie a "while" in a do-statement or an "else" in an if-statement, like +ie a ``while`` in a do-statement or an ``else`` in an if-statement, like this: .. code-block:: c do { - body of do-loop; + body of do-loop } while (condition); and @@ -286,11 +190,11 @@ and .. code-block:: c if (x == y) { - .. + .. } else if (x > y) { - ... + ... } else { - .... + .... } Rationale: K&R. @@ -306,79 +210,92 @@ Do not unnecessarily use braces where a single statement will do. .. code-block:: c if (condition) - action(); + action(); + and + +.. code-block:: none + + if (condition) + do_this(); + else + do_that(); -This does not apply if one branch of a conditional statement is a single -statement. Use braces in both branches. +This does not apply if only one branch of a conditional statement is a single +statement; in the latter case use braces in both branches: .. code-block:: c if (condition) { - do_this(); - do_that(); + do_this(); + do_that(); } else { - otherwise(); + otherwise(); } -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 3.1: Spaces -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +********************** +Chapter 3.1: Spaces +********************** -Linux kernel style for use of spaces depends (mostly) on -function-versus-keyword usage. Use a space after (most) keywords. The +Tarantool style for use of spaces depends (mostly) on +function-versus-keyword usage. Use a space after (most) keywords. The notable exceptions are sizeof, typeof, alignof, and __attribute__, which look somewhat like functions (and are usually used with parentheses in Linux, -although they are not required in the language, as in: "sizeof info" after -"struct fileinfo info;" is declared). +although they are not required in the language, as in: ``sizeof info`` after +``struct fileinfo info;`` is declared). + +So use a space after these keywords: + +.. code-block:: c + + if, switch, case, for, do, while -So use a space after these keywords: if, switch, case, for, do, while but not with sizeof, typeof, alignof, or __attribute__. E.g., .. code-block:: c - s = sizeof(struct file); + s = sizeof(struct file); Do not add spaces around (inside) parenthesized expressions. This example is **bad**: .. code-block:: c - s = sizeof( struct file ); + s = sizeof( struct file ); When declaring pointer data or a function that returns a pointer type, the -preferred use of '*' is adjacent to the data name or function name and not -adjacent to the type name. Examples: +preferred use of ``*`` is adjacent to the data name or function name and not +adjacent to the type name. Examples: .. code-block:: c - char *linux_banner; - unsigned long long memparse(char *ptr, char **retptr); - char *match_strdup(substring_t *s); + char *linux_banner; + unsigned long long memparse(char *ptr, char **retptr); + char *match_strdup(substring_t *s); Use one space around (on each side of) most binary and ternary operators, -such as any of these: +such as any of these:: = + - < > * / % | & ^ <= >= == != ? : -but no space after unary operators: +but no space after unary operators:: & * + - ~ ! sizeof typeof alignof __attribute__ defined -no space before the postfix increment & decrement unary operators: +no space before the postfix increment & decrement unary operators:: ++ -- -no space after the prefix increment & decrement unary operators: +no space after the prefix increment & decrement unary operators:: ++ -- -and no space around the '.' and "->" structure member operators. +and no space around the ``.`` and ``->`` structure member operators. -Do not leave trailing whitespace at the ends of lines. Some editors with -"smart" indentation will insert whitespace at the beginning of new lines as +Do not leave trailing whitespace at the ends of lines. Some editors with +``smart`` indentation will insert whitespace at the beginning of new lines as appropriate, so you can start typing the next line of code right away. However, some such editors do not remove the whitespace if you end up not -putting a line of code there, such as if you leave a blank line. As a result, +putting a line of code there, such as if you leave a blank line. As a result, you end up with lines containing trailing whitespace. Git will warn you about patches that introduce trailing whitespace, and can @@ -386,118 +303,132 @@ optionally strip the trailing whitespace for you; however, if applying a series of patches, this may make later patches in the series fail by changing their context lines. +.. admonition:: Tarantool Style + :class: FACT -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 4: Naming -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Do not split a cast operator from its argument with a whitespace, + e.g. ``(ssize_t)inj->iparam``. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 4: Naming +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -C is a Spartan language, and so should your naming be. Unlike Modula-2 +C is a Spartan language, and so should your naming be. Unlike Modula-2 and Pascal programmers, C programmers do not use cute names like -ThisVariableIsATemporaryCounter. A C programmer would call that -variable "tmp", which is much easier to write, and not the least more +ThisVariableIsATemporaryCounter. A C programmer would call that +variable ``tmp``, which is much easier to write, and not the least more difficult to understand. HOWEVER, while mixed-case names are frowned upon, descriptive names for -global variables are a must. To call a global function "foo" is a +global variables are a must. To call a global function ``foo`` is a shooting offense. -GLOBAL variables (to be used only if you _really_ need them) need to +GLOBAL variables (to be used only if you **really** need them) need to have descriptive names, as do global functions. If you have a function that counts the number of active users, you should call that -"count_active_users()" or similar, you should _not_ call it "cntusr()". +``count_active_users()`` or similar, you should **not** call it ``cntusr()``. Encoding the type of a function into the name (so-called Hungarian notation) is brain damaged - the compiler knows the types anyway and can -check those, and it only confuses the programmer. No wonder MicroSoft +check those, and it only confuses the programmer. No wonder MicroSoft makes buggy programs. LOCAL variable names should be short, and to the point. If you have -some random integer loop counter, it should probably be called "i". -Calling it "loop_counter" is non-productive, if there is no chance of it -being mis-understood. Similarly, "tmp" can be just about any type of +some random integer loop counter, it should probably be called ``i``. +Calling it ``loop_counter`` is non-productive, if there is no chance of it +being mis-understood. Similarly, ``tmp`` can be just about any type of variable that is used to hold a temporary value. If you are afraid to mix up your local variable names, you have another problem, which is called the function-growth-hormone-imbalance syndrome. See chapter 6 (Functions). +.. admonition:: Tarantool Style + :class: FACT -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 5: Typedefs -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + For function naming we have a convention is to use: -Please don't use things like "vps_t". + * ``new``/``delete`` for functions which + allocate + initialize and destroy + deallocate an object, + * ``create``/``destroy`` for functions which initialize/destroy an object + but do not handle memory management, + * ``init``/``free`` for functions which initialize/destroy libraries and subsystems. -It's a _mistake_ to use typedef for structures and pointers. When you see a +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 5: Typedefs +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Please don't use things like ``vps_t``. +It's a **mistake** to use typedef for structures and pointers. When you see a .. code-block:: c - vps_t a; + vps_t a; in the source, what does it mean? - In contrast, if it says .. code-block:: c - struct virtual_container *a; + struct virtual_container *a; -you can actually tell what "a" is. +you can actually tell what ``a`` is. -Lots of people think that typedefs "help readability". Not so. They are +Lots of people think that typedefs ``help readability``. Not so. They are useful only for: -(a) totally opaque objects (where the typedef is actively used to _hide_ +#. totally opaque objects (where the typedef is actively used to **hide** what the object is). - Example: "pte_t" etc. opaque objects that you can only access using - the proper accessor functions. + Example: ``pte_t`` etc. opaque objects that you can only access using + the proper accessor functions. - NOTE! Opaqueness and "accessor functions" are not good in themselves. - The reason we have them for things like pte_t etc. is that there - really is absolutely _zero_ portably accessible information there. + .. note:: -(b) Clear integer types, where the abstraction _helps_ avoid confusion - whether it is "int" or "long". + Opaqueness and ``accessor functions`` are not good in themselves. + The reason we have them for things like pte_t etc. is that there + really is absolutely **zero** portably accessible information there. - u8/u16/u32 are perfectly fine typedefs, although they fit into - category (d) better than here. +#. Clear integer types, where the abstraction **helps** avoid confusion + whether it is ``int`` or ``long``. - NOTE! Again - there needs to be a _reason_ for this. If something is - "unsigned long", then there's no reason to do + u8/u16/u32 are perfectly fine typedefs, although they fit into + point 4 better than here. - .. code-block:: c + .. note:: - typedef unsigned long myflags_t; + Again - there needs to be a **reason** for this. If something is + ``unsigned long``, then there's no reason to do + typedef unsigned long myflags_t; - but if there is a clear reason for why it under certain circumstances - might be an "unsigned int" and under other configurations might be - "unsigned long", then by all means go ahead and use a typedef. + but if there is a clear reason for why it under certain circumstances + might be an ``unsigned int`` and under other configurations might be + ``unsigned long``, then by all means go ahead and use a typedef. -(c) when you use sparse to literally create a _new_ type for - type-checking. +#. when you use sparse to literally create a **new** type for + type-checking. -(d) New types which are identical to standard C99 types, in certain - exceptional circumstances. +#. New types which are identical to standard C99 types, in certain + exceptional circumstances. - Although it would only take a short amount of time for the eyes and - brain to become accustomed to the standard types like 'uint32_t', - some people object to their use anyway. + Although it would only take a short amount of time for the eyes and + brain to become accustomed to the standard types like ``uint32_t``, + some people object to their use anyway. - Therefore, the Linux-specific 'u8/u16/u32/u64' types and their - signed equivalents which are identical to standard types are - permitted -- although they are not mandatory in new code of your - own. + Therefore, the Linux-specific ``u8/u16/u32/u64`` types and their + signed equivalents which are identical to standard types are + permitted -- although they are not mandatory in new code of your + own. - When editing existing code which already uses one or the other set - of types, you should conform to the existing choices in that code. + When editing existing code which already uses one or the other set + of types, you should conform to the existing choices in that code. -(e) Types safe for use in userspace. +#. Types safe for use in userspace. - In certain structures which are visible to userspace, we cannot - require C99 types and cannot use the 'u32' form above. Thus, we - use __u32 and similar types in all structures which are shared - with userspace. + In certain structures which are visible to userspace, we cannot + require C99 types and cannot use the ``u32`` form above. Thus, we + use __u32 and similar types in all structures which are shared + with userspace. Maybe there are other cases too, but the rule should basically be to NEVER EVER use a typedef unless you can clearly match one of those rules. @@ -505,12 +436,11 @@ EVER use a typedef unless you can clearly match one of those rules. In general, a pointer, or a struct that has elements that can reasonably be directly accessed should **never** be a typedef. +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 6: Functions +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 6: Functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Functions should be short and sweet, and do just one thing. They should +Functions should be short and sweet, and do just one thing. They should fit 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. @@ -523,45 +453,59 @@ different cases, it's OK to have a longer function. However, if you have a complex function, and you suspect that a less-than-gifted first-year high-school student might not even understand what the function is all about, you should adhere to the -maximum limits all the more closely. Use helper functions with +maximum limits all the more closely. Use helper functions with descriptive names (you can ask the compiler to in-line them if you think it's performance-critical, and it will probably do a better job of it than you would have done). -Another measure of the function is the number of local variables. They -shouldn't exceed 5-10, or you're doing something wrong. Re-think the -function, and split it into smaller pieces. A human brain can +Another measure of the function is the number of local variables. They +shouldn't exceed 5-10, or you're doing something wrong. Re-think the +function, and split it into smaller pieces. A human brain can generally easily keep track of about 7 different things, anything more -and it gets confu/sed. You know you're brilliant, but maybe you'd like +and it gets confused. You know you're brilliant, but maybe you'd like to understand what you did 2 weeks from now. -In source files, separate functions with one blank line. If the function is -exported, the EXPORT* macro for it should follow immediately after the closing -function brace line. E.g.: +In source files, separate functions with one blank line. If the function is +exported, the **EXPORT** macro for it should follow immediately after the +closing function brace line. E.g.: .. code-block:: c - int system_is_up(void) - { - return system_state == SYSTEM_RUNNING; - } - EXPORT_SYMBOL(system_is_up); + int + system_is_up(void) + { + return system_state == SYSTEM_RUNNING; + } + EXPORT_SYMBOL(system_is_up); In function prototypes, include parameter names with their data types. Although this is not required by the C language, it is preferred in Linux because it is a simple way to add valuable information for the reader. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 7: Centralized exiting of functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +.. admonition:: Tarantool Style + :class: FACT + + Note that in Tarantool, we place the function return type on the + line before the name and signature. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 7: Centralized exiting of functions +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Albeit deprecated by some people, the equivalent of the goto statement is used frequently by compilers in form of the unconditional jump instruction. The goto statement comes in handy when a function exits from multiple -locations and some common work such as cleanup has to be done. +locations and some common work such as cleanup has to be done. If there is no +cleanup needed then just return directly. -The rationale is: +Choose label names which say what the goto does or why the goto exists. An +example of a good name could be ``out_free_buffer:`` if the goto frees ``buffer``. +Avoid using GW-BASIC names like ``err1:`` and ``err2:``, as you would have to +renumber them if you ever add or remove exit paths, and they make correctness +difficult to verify anyway. + +The rationale for using gotos is: - unconditional statements are easier to understand and follow - nesting is reduced @@ -571,219 +515,208 @@ The rationale is: .. code-block:: c - int fun(int a) - { - int result = 0; - char *buffer = kmalloc(SIZE); - - if (buffer == NULL) - return -ENOMEM; - - if (condition1) { - while (loop1) { - ... - } - result = 1; - goto out; - } + int + fun(int a) + { + int result = 0; + char *buffer; + + buffer = kmalloc(SIZE, GFP_KERNEL); + if (!buffer) + return -ENOMEM; + + if (condition1) { + while (loop1) { ... - out: - kfree(buffer); - return result; + } + result = 1; + goto out_free_buffer; } + ... + out_free_buffer: + kfree(buffer); + return result; + } -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 8: Commenting -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +A common type of bug to be aware of is ``one err bugs`` which look like this: + +.. code-block:: c + + err: + kfree(foo->bar); + kfree(foo); + return ret; + +The bug in this code is that on some exit paths ``foo`` is NULL. Normally the +fix for this is to split it up into two error labels ``err_free_bar:`` and +``err_free_foo:``: + +.. code-block:: c -Comments are good, but there is also a danger of over-commenting. NEVER + err_free_bar: + kfree(foo->bar); + err_free_foo: + kfree(foo); + return ret; + +Ideally you should simulate errors to test all exit paths. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 8: Commenting +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Comments are good, but there is also a danger of over-commenting. NEVER try to explain HOW your code works in a comment: it's much better to -write the code so that the _working_ is obvious, and it's a waste of +write the code so that the **working** is obvious, and it's a waste of time to explain badly written code. -с + Generally, you want your comments to tell WHAT your code does, not HOW. Also, try to avoid putting comments inside a function body: if the function 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 make +you should probably go back to chapter 6 for a while. You can make small comments to note or warn about something particularly clever (or -ugly), but try to avoid excess. Instead, put the comments at the head +ugly), but try to avoid excess. Instead, put the comments at the head of the function, telling people what it does, and possibly WHY it does it. -When commenting the kernel API functions, please use the kernel-doc format. -See the files Documentation/kernel-doc-nano-HOWTO.txt and scripts/kernel-doc -for details. +.. admonition:: Tarantool Style + :class: FACT -Linux style for comments is the C89 :code:`/* ... */`` style. -Don't use C99-style :code:`// ...` comments. + When commenting the Tarantool C API functions, please use Doxygen comment format, + Javadoc flavor, i.e. `@tag` rather than `\\tag`. + The main tags in use are ``@param``, ``@retval``, ``@return``, ``@see``, + ``@note`` and ``@todo``. -The preferred style for long (multi-line) comments is: + Every function, except perhaps a very short and obvious one, should have a + comment. A sample function comment may look like below: -.. code-block:: c + .. code-block:: c - /* - * 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. - */ + /** Write all data to a descriptor. + * + * This function is equivalent to 'write', except it would ensure + * that all data is written to the file unless a non-ignorable + * error occurs. + * + * @retval 0 Success + * + * @retval 1 An error occurred (not EINTR) + * / + static int + write_all(int fd, void \*data, size_t len); + + It's also important to comment data, whether they are basic types or derived + types. To this end, use just one data declaration per line (no commas for + multiple data declarations). This leaves you room for a small comment on each + item, explaining its use. + + Public structures and important structure members should be commented as well. + + In C comments out of functions and inside of functions should be different in + how they are started. Everything else is wrong. Below are correct examples. + /** comes for documentation comments, /* for local not documented comments. + However the difference is vague already, so the rule is simple: + out of function = /\**, inside = /\*. -It's also important to comment data, whether they are basic types or derived -types. To this end, use just one data declaration per line (no commas for -multiple data declarations). This leaves you room for a small comment on each -item, explaining its use. + .. code-block:: c -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 9: You've made a mess of it -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + /** + * Out of function comment, option 1. + */ + + /** Out of function comment, option 2. */ + + int + function() + { + /* Comment inside function, option 1. */ + + /* + * Comment inside function, option 2. + */ + } + + If a function has declaration and implementation separated, the function comment + should be for the declaration. Usually in the header file. Don't duplicate the + comment. + + A comment and the function signature should be synchronized. Double-check if the + parameter names are the same as used in the comment, and mean the same. + Especially when you change one of them - ensure you changed the other. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 9: You've made a mess of it +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ That's OK, we all do. You've probably been told by your long-time Unix -user helper that "GNU emacs" automatically formats the C sources for +user helper that ``GNU emacs`` automatically formats the C sources for you, and you've noticed that yes, it does do that, but the defaults it uses are less than desirable (in fact, they are worse than random typing - an infinite number of monkeys typing into GNU emacs would never make a good program). So, you can either get rid of GNU emacs, or change it to use saner -values. To do the latter, you can stick the following in your .emacs file: +values. To do the latter, you can stick the following in your .emacs file: -.. code-block:: lisp +.. code-block:: none - (defun c-lineup-arglist-tabs-only (ignored) + (defun c-lineup-arglist-tabs-only (ignored) "Line up argument lists by tabs, not spaces" (let* ((anchor (c-langelem-pos c-syntactic-element)) - (column (c-langelem-2nd-pos c-syntactic-element)) - (offset (- (1+ column) anchor)) - (steps (floor offset c-basic-offset))) - (* (max steps 1) - c-basic-offset))) + (column (c-langelem-2nd-pos c-syntactic-element)) + (offset (- (1+ column) anchor)) + (steps (floor offset c-basic-offset))) + (* (max steps 1) + c-basic-offset))) - (add-hook 'c-mode-common-hook + (add-hook 'c-mode-common-hook (lambda () - ;; Add kernel style - (c-add-style - "linux-tabs-only" - '("linux" (c-offsets-alist - (arglist-cont-nonempty - c-lineup-gcc-asm-reg - c-lineup-arglist-tabs-only)))))) - - (add-hook 'c-mode-hook + ;; Add kernel style + (c-add-style + "linux-tabs-only" + '("linux" (c-offsets-alist + (arglist-cont-nonempty + c-lineup-gcc-asm-reg + c-lineup-arglist-tabs-only)))))) + + (add-hook 'c-mode-hook (lambda () - (let ((filename (buffer-file-name))) + (let ((filename (buffer-file-name))) ;; Enable kernel mode for the appropriate files (when (and filename - (string-match (expand-file-name "~/src/linux-trees") - filename)) - (setq indent-tabs-mode t) - (c-set-style "linux-tabs-only"))))) + (string-match (expand-file-name "~/src/linux-trees") + filename)) + (setq indent-tabs-mode t) + (setq show-trailing-whitespace t) + (c-set-style "linux-tabs-only"))))) This will make emacs go better with the kernel coding style for C -files below ~/src/linux-trees. +files below ``~/src/linux-trees``. But even if you fail in getting emacs to do sane formatting, not -everything is lost: use "indent". +everything is lost: use ``indent``. Now, again, GNU indent has the same brain-dead settings that GNU emacs has, 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 indent recognize the authority of K&R (the GNU people aren't evil, they are just severely misguided in this matter), so you just give indent the -options "-kr -i8" (stands for "K&R, 8 character indents"), or use -"scripts/Lindent", which indents in the latest style. +options ``-kr -i8`` (stands for ``K&R, 8 character indents``), or use +``scripts/Lindent``, which indents in the latest style. -"indent" has a lot of options, and especially when it comes to comment +``indent`` has a lot of options, and especially when it comes to comment re-formatting you may want to take a look at the man page. But -remember: "indent" is not a fix for bad programming. +remember: ``indent`` is not a fix for bad programming. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 10: Kconfig configuration files -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -For all of the Kconfig* configuration files throughout the source tree, -the indentation is somewhat different. Lines under a "config" definition -are indented with one tab, while help text is indented an additional two -spaces. Example: - -.. code-block:: kconfig - - 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. - -Features that might still be considered unstable should be defined as -dependent on "EXPERIMENTAL": - -.. code-block:: kconfig - - config SLUB - depends on EXPERIMENTAL && !ARCH_USES_SLAB_PAGE_STRUCT - bool "SLUB (Unqueued Allocator)" - ... - -while seriously dangerous features (such as write support for certain -filesystems) should advertise this prominently in their prompt string: - -.. code-block:: kconfig - - config ADFS_FS_RW - bool "ADFS write support (DANGEROUS)" - depends on ADFS_FS - ... - -For full documentation on the configuration files, see the file -Documentation/kbuild/kconfig-language.txt. - - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 11: Data structures -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Data structures that have visibility outside the single-threaded -environment they are created and destroyed in should always have -reference counts. In the kernel, garbage collection doesn't exist (and -outside the kernel garbage collection is slow and inefficient), which -means that you absolutely _have_ to reference count all your uses. - -Reference counting means that you can avoid locking, and allows multiple -users to have access to the data structure in parallel - and not having -to worry about the structure suddenly going away from under them just -because they slept or did something else for a while. - -Note that locking is _not_ a replacement for reference counting. -Locking is used to keep data structures coherent, while reference -counting is a memory management technique. Usually both are needed, and -they are not to be confused with each other. - -Many data structures can indeed have two levels of reference counting, -when there are users of different "classes". The subclass count counts -the number of subclass users, and decrements the global count just once -when the subclass count goes to zero. - -Examples of this kind of "multi-level-reference-counting" can be found in -memory management ("struct mm_struct": mm_users and mm_count), and in -filesystem code ("struct super_block": s_count and s_active). - -Remember: if another thread can find your data structure, and you don't -have a reference count on it, you almost certainly have a bug. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 12: Macros, Enums and RTL -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 10: Macros, Enums and RTL +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Names of macros defining constants and labels in enums are capitalized. .. code-block:: c - #define CONSTANT 0x12345 + #define CONSTANT 0x12345 Enums are preferred when defining several related constants. @@ -796,106 +729,89 @@ Macros with multiple statements should be enclosed in a do - while block: .. code-block:: c - #define macrofun(a, b, c) \ - do { \ - if (a == 5) \ - do_this(b, c); \ - } while (0) + #define macrofun(a, b, c) \ + do { \ + if (a == 5) \ + do_this(b, c); \ + } while (0) Things to avoid when using macros: -1. macros that affect control flow: +1) macros that affect control flow: .. code-block:: c - #define FOO(x) \ - do { \ - if (blah(x) < 0) \ - return -EBUGGERED; \ - } while(0) + #define FOO(x) \ + do { \ + if (blah(x) < 0) \ + return -EBUGGERED; \ + } while (0) - is a _very_ bad idea. It looks like a function call but exits the "calling" + is a **very** bad idea. It looks like a function call but exits the ``calling`` function; 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: +2) macros that depend on having a local variable with a magic name: .. code-block:: c - #define FOO(val) bar(index, val) + #define FOO(val) bar(index, val) might look like a good thing, but it's confusing as hell when one reads the code and it's prone to breakage from seemingly innocent changes. -3. macros with arguments that are used as l-values: FOO(x) = y; will +3) macros with arguments that are used as l-values: FOO(x) = y; will bite you if somebody e.g. turns FOO into an inline function. -4. forgetting about precedence: macros defining constants using expressions +4) forgetting about precedence: macros defining constants using expressions must enclose the expression in parentheses. Beware of similar issues with macros using parameters. .. code-block:: c - #define CONSTANT 0x4000 - #define CONSTEXP (CONSTANT | 3) - - The cpp manual deals with macros exhaustively. The gcc internals manual also - covers RTL which is used frequently with assembly language in the kernel. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 13: Printing kernel messages -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Kernel developers like to be seen as literate. Do mind the spelling -of kernel messages to make a good impression. Do not use crippled -words like "dont"; use "do not" or "don't" instead. Make the messages -concise, clear, and unambiguous. + #define CONSTANT 0x4000 + #define CONSTEXP (CONSTANT | 3) -Kernel messages do not have to be terminated with a period. +5) namespace collisions when defining local variables in macros resembling + functions: -Printing numbers in parentheses (%d) adds no value and should be avoided. - -There are a number of driver model diagnostic macros in -which you should use to make sure messages are matched to the right device -and driver, and are tagged with the right level: dev_err(), dev_warn(), -dev_info(), and so forth. For messages that aren't associated with a -particular device, defines pr_debug() and pr_info(). + .. code-block:: c -Coming up with good debugging messages can be quite a challenge; and once -you have them, they can be a huge help for remote troubleshooting. Such -messages should be compiled out when the DEBUG symbol is not defined (that -is, by default they are not included). When you use dev_dbg() or pr_debug(), -that's automatic. Many subsystems have Kconfig options to turn on -DDEBUG. -A related convention uses VERBOSE_DEBUG to add dev_vdbg() messages to the -ones already enabled by DEBUG. + #define FOO(x) \ + ({ \ + typeof(x) ret; \ + ret = calc_ret(x); \ + (ret); \ + }) -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 14: Allocating memory -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + ret is a common name for a local variable - __foo_ret is less likely + to collide with an existing variable. -The kernel provides the following general purpose memory allocators: -kmalloc(), kzalloc(), kcalloc(), and vmalloc(). Please refer to the API -documentation for further information about them. - -The preferred form for passing a size of a struct is the following: + The cpp manual deals with macros exhaustively. The gcc internals manual also + covers RTL which is used frequently with assembly language in the kernel. -.. code-block:: c +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 11: Allocating memory +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - p = kmalloc(sizeof(*p), ...); +.. admonition:: Tarantool Style + :class: FACT -The alternative form where struct name is spelled out hurts readability and -introduces an opportunity for a bug when the pointer variable type is changed -but the corresponding sizeof that is passed to a memory allocator is not. + Prefer the supplied slab (salloc) and pool (palloc) allocators to malloc()/free() + for any performance-intensive or large memory allocations. Repetitive use of + malloc()/free() can lead to memory fragmentation and should therefore be avoided. -Casting the return value which is a void pointer is redundant. The conversion -from void pointer to any other pointer type is guaranteed by the C programming -language. + Always free all allocated memory, even allocated at start-up. We aim at being + valgrind leak-check clean, and in most cases it's just as easy to free() the + allocated memory as it is to write a valgrind suppression. Freeing all allocated + memory is also dynamic-load friendly: assuming a plug-in can be dynamically + loaded and unloaded multiple times, reload should not lead to a memory leak. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 15: The inline disease -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 12: The inline disease +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There appears to be a common misperception that gcc has a magic "make me -faster" speedup option called "inline". While the use of inlines can be +faster" speedup option called ``inline``. While the use of inlines can be appropriate (for example as a means of replacing macros, see Chapter 12), it very often is not. Abundant use of the inline keyword leads to a much bigger kernel, which in turn slows the system as a whole down, due to a bigger @@ -918,14 +834,14 @@ help, and the maintenance issue of removing the inline when a second user appears outweighs the potential value of the hint that tells gcc to do something it would have done anyway. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 16: Function return values and names -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 13: Function return values and names +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Functions can return values of many different kinds, and one of the most common is a value indicating whether the function succeeded or failed. Such a value can be represented as an error-code integer -(-Exxx = failure, 0 = success) or a "succeeded" boolean (0 = failure, +(-Exxx = failure, 0 = success) or a ``succeeded`` boolean (0 = failure, non-zero = success). Mixing up these two sorts of representations is a fertile source of @@ -934,14 +850,12 @@ between integers and booleans then the compiler would find these mistakes for us... but it doesn't. To help prevent such bugs, always follow this convention: -:: + If the name of a function is an action or an imperative command, + the function should return an error-code integer. If the name + is a predicate, the function should return a "succeeded" boolean. - If the name of a function is an action or an imperative command, - the function should return an error-code integer. If the name - is a predicate, the function should return a "succeeded" boolean. - -For example, "add work" is a command, and the add_work() function returns 0 -for success or -EBUSY for failure. In the same way, "PCI device present" is +For example, ``add work`` is a command, and the add_work() function returns 0 +for success or -EBUSY for failure. In the same way, ``PCI device present`` is a predicate, and the pci_dev_present() function returns 1 if it succeeds in finding a matching device or 0 if it doesn't. @@ -955,63 +869,157 @@ this rule. Generally they indicate failure by returning some out-of-range result. Typical examples would be functions that return pointers; they use NULL or the ERR_PTR mechanism to report failure. +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 14: Editor modelines and other cruft +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 17: Don't re-invent the kernel macros -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Some editors can interpret configuration information embedded in source files, +indicated with special markers. For example, emacs interprets lines marked +like this: + +.. code-block:: c + + -*- mode: c -*- -The header file include/linux/kernel.h contains a number of macros that -you should use, rather than explicitly coding some variant of them yourself. -For example, if you need to calculate the length of an array, take advantage -of the macro +Or like this: .. code-block:: c - #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) + /* + Local Variables: + compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" + End: + */ -Similarly, if you need to calculate the size of some structure member, use +Vim interprets markers that look like this: .. code-block:: c - #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f)) + /* vim:set sw=8 noet */ -There are also min() and max() macros that do strict type checking if you -need them. Feel free to peruse that header file to see what else is already -defined that you shouldn't reproduce in your code. +Do not include any of these in source files. People have their own personal +editor configurations, and your source files should not override them. This +includes markers for indentation and mode configuration. People may use their +own custom mode, or may have some other magic method for making indentation +work correctly. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 18: Editor modelines and other cruft -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 15: Inline assembly +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Some editors can interpret configuration information embedded in source files, -indicated with special markers. For example, emacs interprets lines marked -like this: +In architecture-specific code, you may need to use inline assembly to interface +with 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 can +and should poke hardware from C when possible. -.. code-block:: none +Consider writing simple helper functions that wrap common bits of inline +assembly, rather than repeatedly writing them with slight variations. Remember +that inline assembly can use C parameters. - -*- mode: c -*- +Large, non-trivial assembly functions should go in .S files, with corresponding +C prototypes defined in C header files. The C prototypes for assembly +functions should use ``asmlinkage``. -Or like this: +You may need to mark your asm statement as volatile, to prevent GCC from +removing it if GCC doesn't notice any side effects. You don't always need to +do so, though, and doing so unnecessarily can limit optimization. -.. code-block:: none +When writing a single inline assembly statement containing multiple +instructions, put each instruction on a separate line in a separate quoted +string, and end each string except the last with \n\t to properly indent the +next instruction in the assembly output: - /* - Local Variables: - compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" - End: - */ +.. code-block:: c -Vim interprets markers that look like this: + asm ("magic %reg1, #42\n\t" + "more_magic %reg2, %reg3" + : /* outputs */ : /* inputs */ : /* clobbers */); -.. code-block:: none +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 16: Conditional Compilation +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - /* vim:set sw=8 noet */ +Wherever possible, don't use preprocessor conditionals (#if, #ifdef) in .c +files; 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 .c +files, providing no-op stub versions in the #else case, and then call those +functions unconditionally from .c files. The compiler will avoid generating +any code for the stub calls, producing identical results, but the logic will +remain easy to follow. -Do not include any of these in source files. People have their own personal -editor configurations, and your source files should not override them. This -includes markers for indentation and mode configuration. People may use their -own custom mode, or may have some other magic method for making indentation -work correctly. +Prefer to compile out entire functions, rather than portions of functions or +portions of expressions. Rather than putting an ifdef in an expression, factor +out part or all of the expression into a separate helper function and apply the +conditional to that function. + +If you have a function or variable which may potentially go unused in a +particular configuration, and the compiler would warn about its definition +going unused, mark the definition as __maybe_unused rather than wrapping it in +a preprocessor conditional. (However, if a function or variable *always* goes +unused, delete it.) + +Within code, where possible, use the IS_ENABLED macro to convert a Kconfig +symbol into a C boolean expression, and use it in a normal C conditional: + +.. code-block:: c + + if (IS_ENABLED(CONFIG_SOMETHING)) { + ... + } + +The compiler will constant-fold the conditional away, and include or exclude +the block of code just as with an #ifdef, so this will not add any runtime +overhead. +However, this approach still allows the C compiler to see the code +inside the block, and check it for correctness (syntax, types, symbol +references, etc). Thus, you still have to use an #ifdef if the code inside the +block 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 conditional +expression used. For instance: + +.. code-block:: c + + #ifdef CONFIG_SOMETHING + ... + #endif /* CONFIG_SOMETHING */ + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 17: Header files +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +.. admonition:: Tarantool Style + :class: FACT + + Use header guards. Put the header guard in the first line in the header, + before the copyright or declarations. Use all-uppercase name for the header + guard. Derive the header guard name from the file name, and append _INCLUDED + to get a macro name. For example, core/log_io.h -> CORE_LOG_IO_H_INCLUDED. In + ``.c`` (implementation) file, include the respective declaration header before + all other headers, to ensure that the header is self- sufficient. Header + "header.h" is self-sufficient if the following compiles without errors: + + .. code-block:: c + + #include "header.h" + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 18: Other +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +.. admonition:: Tarantool Style + :class: FACT + + * We don't apply ``!`` operator to non-boolean values. It means, to check + if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, + you use ``!= NULL``. The same for ``==``. + + * Select GNU C99 extensions are acceptable. It's OK to mix declarations and + statements, use true and false. + + * The not-so-current list of all GCC C extensions can be found at: + http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Appendix I: References @@ -1035,3 +1043,4 @@ work correctly. * `Kernel CodingStyle, by greg@kroah.com at OLS 2002 `_ + diff --git a/doc/dev_guide/c_style_guide_new.rst b/doc/dev_guide/c_style_guide_new.rst new file mode 100644 index 0000000000..4657470d05 --- /dev/null +++ b/doc/dev_guide/c_style_guide_new.rst @@ -0,0 +1,1046 @@ +------------------------------------------------------------------------------- + C Style Guide +------------------------------------------------------------------------------- + +We use Git for revision control. The latest development is happening in the +default branch (currently ``master``). Our git repository is hosted on GitHub, +and can be checked out with ``git clone git://github.com/tarantool/tarantool.git`` +(anonymous read-only access). + +If you have any questions about Tarantool internals, please post them on the +`developer discussion list `_ +or on `StackOverflow `_. +Additionally, some engineers are always present on #tarantool channel on +irc.freenode.net. + +**General guidelines** + +The project's coding style is based on the `Linux kernel coding style +`_. + +However, we have some additional guidelines, either unique to Tarantool or +deviating from the Kernel guidelines. Below we cite the Linux kernel +coding style noting Tarantool's style features. + +We don't cite chapters 10 "Kconfig configuration files", 11 "Data structures", +13 "Printing kernel messages", and 17 "Don't re-invent the kernel macros" since +they are specific to Linux kernel programming environment. + +================================================================================ + Linux kernel coding style +================================================================================ + +This is a short document describing the preferred coding style for the +linux kernel. Coding style is very personal, and I won't **force** my +views on anybody, but this is what goes for anything that I have to be +able to maintain, and I'd prefer it for most other things too. Please +at least consider the points made here. + +First off, I'd suggest printing out a copy of the GNU coding standards, +and NOT read it. Burn them, it's a great symbolic gesture. + +Anyway, here goes: + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 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 to be 3. + +Rationale: The whole idea behind indentation is to clearly define where +a block of control starts and ends. Especially when you've been looking +at your screen for 20 straight hours, you'll find it a lot easier to see +how the indentation works if you have large indentations. + +Now, some people will claim that having 8-character indentations makes +the code move too far to the right, and makes it hard to read on a +80-character terminal screen. The answer to that is that if you need +more than 3 levels of indentation, you're screwed anyway, and should fix +your program. + +In short, 8-char indents make things easier to read, and have the added +benefit 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 is +to align the ``switch`` and its subordinate ``case`` labels in the same column +instead of ``double-indenting`` the ``case`` labels. E.g.: + +.. code-block:: c + + switch (suffix) { + case 'G': + case 'g': + mem <<= 30; + break; + case 'M': + case 'm': + mem <<= 20; + break; + case 'K': + case 'k': + mem <<= 10; + /* fall through */ + default: + break; + } + +Don't put multiple statements on a single line unless you have +something to hide: + +.. code-block:: c + + if (condition) do_this; + do_something_everytime; + +Don't put multiple assignments on a single line either. Kernel coding style +is super simple. Avoid tricky expressions. + +Outside of comments and documentation, spaces are never +used for indentation, and the above example is deliberately broken. + +Get a decent editor and don't leave whitespace at the end of lines. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 2: Breaking long lines and strings +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Coding style is all about readability and maintainability using commonly +available tools. + +The limit on the length of lines is 80 columns and this is a strongly +preferred limit. + +.. admonition:: Tarantool Style + :class: FACT + + As for comments, the same limit of 80 columns is applied. + +Statements longer than 80 columns will be broken into sensible chunks, unless +exceeding 80 columns significantly increases readability and does not hide +information. Descendants are always substantially shorter than the parent and +are placed substantially to the right. The same applies to function headers +with a long argument list. However, never break user-visible strings such as +printk messages, because that breaks the ability to grep for them. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 3: Placing Braces and Spaces +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +The other issue that always comes up in C styling is the placement of +braces. Unlike the indent size, there are few technical reasons to +choose one placement strategy over the other, but the preferred way, as +shown to us by the prophets Kernighan and Ritchie, is to put the opening +brace last on the line, and put the closing brace first, thus: + +.. code-block:: c + + if (x is true) { + we do y + } + +This applies to all non-function statement blocks (if, switch, for, +while, do). E.g.: + +.. code-block:: c + + switch (action) { + case KOBJ_ADD: + return "add"; + case KOBJ_REMOVE: + return "remove"; + case KOBJ_CHANGE: + return "change"; + default: + return NULL; + } + +However, there is one special case, namely functions: they have the +opening brace at the beginning of the next line, thus: + +.. code-block:: c + + int + function(int x) + { + body of function + } + +Heretic people all over the world have claimed that this inconsistency +is ... well ... inconsistent, but all right-thinking people know that +(a) K&R are **right** and (b) K&R are right. Besides, functions are +special anyway (you can't nest them in C). + +Note that the closing brace is empty on a line of its own, **except** in +the cases where it is followed by a continuation of the same statement, +ie a ``while`` in a do-statement or an ``else`` in an if-statement, like +this: + +.. code-block:: c + + do { + body of do-loop + } while (condition); + +and + +.. code-block:: c + + if (x == y) { + .. + } else if (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 the +supply of new-lines on your screen is not a renewable resource (think +25-line terminal screens here), you have more empty lines to put +comments on. + +Do not unnecessarily use braces where a single statement will do. + +.. code-block:: c + + if (condition) + action(); + and + +.. code-block:: none + + if (condition) + do_this(); + else + do_that(); + +This does not apply if only one branch of a conditional statement is a single +statement; in the latter case use braces in both branches: + +.. code-block:: c + + if (condition) { + do_this(); + do_that(); + } else { + otherwise(); + } + +********************** +Chapter 3.1: Spaces +********************** + +Tarantool style for use of spaces depends (mostly) on +function-versus-keyword usage. Use a space after (most) keywords. The +notable exceptions are sizeof, typeof, alignof, and __attribute__, which look +somewhat like functions (and are usually used with parentheses in Linux, +although they are not required in the language, as in: ``sizeof info`` after +``struct fileinfo info;`` is declared). + +So use a space after these keywords: + +.. code-block:: c + + if, switch, case, for, do, while + +but not with sizeof, typeof, alignof, or __attribute__. E.g., + +.. code-block:: c + + s = sizeof(struct file); + +Do not add spaces around (inside) parenthesized expressions. This example is +**bad**: + +.. code-block:: c + + s = sizeof( struct file ); + +When declaring pointer data or a function that returns a pointer type, the +preferred use of ``*`` is adjacent to the data name or function name and not +adjacent to the type name. Examples: + +.. code-block:: c + + char *linux_banner; + unsigned long long memparse(char *ptr, char **retptr); + char *match_strdup(substring_t *s); + +Use one space around (on each side of) most binary and ternary operators, +such as any of these:: + + = + - < > * / % | & ^ <= >= == != ? : + +but no space after unary operators:: + + & * + - ~ ! sizeof typeof alignof __attribute__ defined + +no space before the postfix increment & decrement unary operators:: + + ++ -- + +no space after the prefix increment & decrement unary operators:: + + ++ -- + +and no space around the ``.`` and ``->`` structure member operators. + +Do not leave trailing whitespace at the ends of lines. Some editors with +``smart`` indentation will insert whitespace at the beginning of new lines as +appropriate, so you can start typing the next line of code right away. +However, some such editors do not remove the whitespace if you end up not +putting a line of code there, such as if you leave a blank line. As a result, +you end up with lines containing trailing whitespace. + +Git will warn you about patches that introduce trailing whitespace, and can +optionally strip the trailing whitespace for you; however, if applying a series +of patches, this may make later patches in the series fail by changing their +context lines. + +.. admonition:: Tarantool Style + :class: FACT + + Do not split a cast operator from its argument with a whitespace, + e.g. ``(ssize_t)inj->iparam``. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 4: Naming +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +C is a Spartan language, and so should your naming be. Unlike Modula-2 +and Pascal programmers, C programmers do not use cute names like +ThisVariableIsATemporaryCounter. A C programmer would call that +variable ``tmp``, which is much easier to write, and not the least more +difficult to understand. + +HOWEVER, while mixed-case names are frowned upon, descriptive names for +global variables are a must. To call a global function ``foo`` is a +shooting offense. + +GLOBAL variables (to be used only if you **really** need them) need to +have descriptive names, as do global functions. If you have a function +that counts the number of active users, you should call that +``count_active_users()`` or similar, you should **not** call it ``cntusr()``. + +Encoding the type of a function into the name (so-called Hungarian +notation) is brain damaged - the compiler knows the types anyway and can +check those, and it only confuses the programmer. No wonder MicroSoft +makes buggy programs. + +LOCAL variable names should be short, and to the point. If you have +some random integer loop counter, it should probably be called ``i``. +Calling it ``loop_counter`` is non-productive, if there is no chance of it +being mis-understood. Similarly, ``tmp`` can be just about any type of +variable that is used to hold a temporary value. + +If you are afraid to mix up your local variable names, you have another +problem, which is called the function-growth-hormone-imbalance syndrome. +See chapter 6 (Functions). + +.. admonition:: Tarantool Style + :class: FACT + + For function naming we have a convention is to use: + + * ``new``/``delete`` for functions which + allocate + initialize and destroy + deallocate an object, + * ``create``/``destroy`` for functions which initialize/destroy an object + but do not handle memory management, + * ``init``/``free`` for functions which initialize/destroy libraries and subsystems. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 5: Typedefs +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Please don't use things like ``vps_t``. +It's a **mistake** to use typedef for structures and pointers. When you see a + +.. code-block:: c + + vps_t a; + +in the source, what does it mean? +In contrast, if it says + +.. code-block:: c + + struct virtual_container *a; + +you can actually tell what ``a`` is. + +Lots of people think that typedefs ``help readability``. Not so. They are +useful only for: + +#. totally opaque objects (where the typedef is actively used to **hide** + what the object is). + + Example: ``pte_t`` etc. opaque objects that you can only access using + the proper accessor functions. + + .. note:: + + Opaqueness and ``accessor functions`` are not good in themselves. + The reason we have them for things like pte_t etc. is that there + really is absolutely **zero** portably accessible information there. + +#. Clear integer types, where the abstraction **helps** avoid confusion + whether it is ``int`` or ``long``. + + u8/u16/u32 are perfectly fine typedefs, although they fit into + point 4 better than here. + + .. note:: + + Again - there needs to be a **reason** for this. If something is + ``unsigned long``, then there's no reason to do + typedef unsigned long myflags_t; + + but if there is a clear reason for why it under certain circumstances + might be an ``unsigned int`` and under other configurations might be + ``unsigned long``, then by all means go ahead and use a typedef. + +#. when you use sparse to literally create a **new** type for + type-checking. + +#. New types which are identical to standard C99 types, in certain + exceptional circumstances. + + Although it would only take a short amount of time for the eyes and + brain to become accustomed to the standard types like ``uint32_t``, + some people object to their use anyway. + + Therefore, the Linux-specific ``u8/u16/u32/u64`` types and their + signed equivalents which are identical to standard types are + permitted -- although they are not mandatory in new code of your + own. + + When editing existing code which already uses one or the other set + of types, you should conform to the existing choices in that code. + +#. Types safe for use in userspace. + + In certain structures which are visible to userspace, we cannot + require C99 types and cannot use the ``u32`` form above. Thus, we + use __u32 and similar types in all structures which are shared + with userspace. + +Maybe there are other cases too, but the rule should basically be to NEVER +EVER use a typedef unless you can clearly match one of those rules. + +In general, a pointer, or a struct that has elements that can reasonably +be directly accessed should **never** be a typedef. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 6: Functions +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Functions should be short and sweet, and do just one thing. They should +fit 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 the +complexity and indentation level of that function. So, if you have a +conceptually simple function that is just one long (but simple) +case-statement, where you have to do lots of small things for a lot of +different cases, it's OK to have a longer function. + +However, if you have a complex function, and you suspect that a +less-than-gifted first-year high-school student might not even +understand what the function is all about, you should adhere to the +maximum limits all the more closely. Use helper functions with +descriptive names (you can ask the compiler to in-line them if you think +it's performance-critical, and it will probably do a better job of it +than you would have done). + +Another measure of the function is the number of local variables. They +shouldn't exceed 5-10, or you're doing something wrong. Re-think the +function, and split it into smaller pieces. A human brain can +generally easily keep track of about 7 different things, anything more +and it gets confused. You know you're brilliant, but maybe you'd like +to understand what you did 2 weeks from now. + +In source files, separate functions with one blank line. If the function is +exported, the **EXPORT** macro for it should follow immediately after the +closing function brace line. E.g.: + +.. code-block:: c + + int + system_is_up(void) + { + return system_state == SYSTEM_RUNNING; + } + EXPORT_SYMBOL(system_is_up); + +In function prototypes, include parameter names with their data types. +Although this is not required by the C language, it is preferred in Linux +because it is a simple way to add valuable information for the reader. + +.. admonition:: Tarantool Style + :class: FACT + + Note that in Tarantool, we place the function return type on the + line before the name and signature. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 7: Centralized exiting of functions +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Albeit deprecated by some people, the equivalent of the goto statement is +used frequently by compilers in form of the unconditional jump instruction. + +The goto statement comes in handy when a function exits from multiple +locations and some common work such as cleanup has to be done. If there is no +cleanup needed then just return directly. + +Choose label names which say what the goto does or why the goto exists. An +example of a good name could be ``out_free_buffer:`` if the goto frees ``buffer``. +Avoid using GW-BASIC names like ``err1:`` and ``err2:``, as you would have to +renumber them if you ever add or remove exit paths, and they make correctness +difficult 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 making + modifications are prevented +- saves the compiler work to optimize redundant code away ;) + +.. code-block:: c + + int + fun(int a) + { + int result = 0; + char *buffer; + + buffer = kmalloc(SIZE, GFP_KERNEL); + if (!buffer) + return -ENOMEM; + + if (condition1) { + while (loop1) { + ... + } + result = 1; + goto out_free_buffer; + } + ... + out_free_buffer: + kfree(buffer); + return result; + } + +A common type of bug to be aware of is ``one err bugs`` which look like this: + +.. code-block:: c + + err: + kfree(foo->bar); + kfree(foo); + return ret; + +The bug in this code is that on some exit paths ``foo`` is NULL. Normally the +fix for this is to split it up into two error labels ``err_free_bar:`` and +``err_free_foo:``: + +.. code-block:: c + + err_free_bar: + kfree(foo->bar); + err_free_foo: + kfree(foo); + return ret; + +Ideally you should simulate errors to test all exit paths. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 8: Commenting +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Comments are good, but there is also a danger of over-commenting. NEVER +try to explain HOW your code works in a comment: it's much better to +write the code so that the **working** is obvious, and it's a waste of +time 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 the +function 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 make +small comments to note or warn about something particularly clever (or +ugly), but try to avoid excess. Instead, put the comments at the head +of the function, telling people what it does, and possibly WHY it does +it. + +.. admonition:: Tarantool Style + :class: FACT + + When commenting the Tarantool C API functions, please use Doxygen comment format, + Javadoc flavor, i.e. `@tag` rather than `\\tag`. + The main tags in use are ``@param``, ``@retval``, ``@return``, ``@see``, + ``@note`` and ``@todo``. + + Every function, except perhaps a very short and obvious one, should have a + comment. A sample function comment may look like below: + + .. code-block:: c + + /** Write all data to a descriptor. + * + * This function is equivalent to 'write', except it would ensure + * that all data is written to the file unless a non-ignorable + * error occurs. + * + * @retval 0 Success + * + * @retval 1 An error occurred (not EINTR) + * / + static int + write_all(int fd, void \*data, size_t len); + + It's also important to comment data, whether they are basic types or derived + types. To this end, use just one data declaration per line (no commas for + multiple data declarations). This leaves you room for a small comment on each + item, explaining its use. + + Public structures and important structure members should be commented as well. + + In C comments out of functions and inside of functions should be different in + how they are started. Everything else is wrong. Below are correct examples. + /** comes for documentation comments, /* for local not documented comments. + However the difference is vague already, so the rule is simple: + out of function = /\**, inside = /\*. + + .. code-block:: c + + /** + * Out of function comment, option 1. + */ + + /** Out of function comment, option 2. */ + + int + function() + { + /* Comment inside function, option 1. */ + + /* + * Comment inside function, option 2. + */ + } + + If a function has declaration and implementation separated, the function comment + should be for the declaration. Usually in the header file. Don't duplicate the + comment. + + A comment and the function signature should be synchronized. Double-check if the + parameter names are the same as used in the comment, and mean the same. + Especially when you change one of them - ensure you changed the other. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 9: You've made a mess of it +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +That's OK, we all do. You've probably been told by your long-time Unix +user helper that ``GNU emacs`` automatically formats the C sources for +you, and you've noticed that yes, it does do that, but the defaults it +uses are less than desirable (in fact, they are worse than random +typing - an infinite number of monkeys typing into GNU emacs would never +make a good program). + +So, you can either get rid of GNU emacs, or change it to use saner +values. To do the latter, you can stick the following in your .emacs file: + +.. code-block:: none + + (defun c-lineup-arglist-tabs-only (ignored) + "Line up argument lists by tabs, not spaces" + (let* ((anchor (c-langelem-pos c-syntactic-element)) + (column (c-langelem-2nd-pos c-syntactic-element)) + (offset (- (1+ column) anchor)) + (steps (floor offset c-basic-offset))) + (* (max steps 1) + c-basic-offset))) + + (add-hook 'c-mode-common-hook + (lambda () + ;; Add kernel style + (c-add-style + "linux-tabs-only" + '("linux" (c-offsets-alist + (arglist-cont-nonempty + c-lineup-gcc-asm-reg + c-lineup-arglist-tabs-only)))))) + + (add-hook 'c-mode-hook + (lambda () + (let ((filename (buffer-file-name))) + ;; Enable kernel mode for the appropriate files + (when (and filename + (string-match (expand-file-name "~/src/linux-trees") + filename)) + (setq indent-tabs-mode t) + (setq show-trailing-whitespace t) + (c-set-style "linux-tabs-only"))))) + +This will make emacs go better with the kernel coding style for C +files below ``~/src/linux-trees``. + +But even if you fail in getting emacs to do sane formatting, not +everything is lost: use ``indent``. + +Now, again, GNU indent has the same brain-dead settings that GNU emacs +has, 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 indent +recognize the authority of K&R (the GNU people aren't evil, they are +just severely misguided in this matter), so you just give indent the +options ``-kr -i8`` (stands for ``K&R, 8 character indents``), or use +``scripts/Lindent``, which indents in the latest style. + +``indent`` has a lot of options, and especially when it comes to comment +re-formatting you may want to take a look at the man page. But +remember: ``indent`` is not a fix for bad programming. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 10: Macros, Enums and RTL +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Names of macros defining constants and labels in enums are capitalized. + +.. code-block:: c + + #define CONSTANT 0x12345 + +Enums are preferred when defining several related constants. + +CAPITALIZED macro names are appreciated but macros resembling functions +may 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: + +.. code-block:: c + + #define macrofun(a, b, c) \ + do { \ + if (a == 5) \ + do_this(b, c); \ + } while (0) + +Things to avoid when using macros: + +1) macros that affect control flow: + + .. code-block:: c + + #define FOO(x) \ + do { \ + if (blah(x) < 0) \ + return -EBUGGERED; \ + } while (0) + + is a **very** bad idea. It looks like a function call but exits the ``calling`` + function; 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: + + .. code-block:: c + + #define FOO(val) bar(index, val) + + might look like a good thing, but it's confusing as hell when one reads the + code and it's prone to breakage from seemingly innocent changes. + +3) macros with arguments that are used as l-values: FOO(x) = y; will + bite you if somebody e.g. turns FOO into an inline function. + +4) forgetting about precedence: macros defining constants using expressions + must enclose the expression in parentheses. Beware of similar issues with + macros using parameters. + + .. code-block:: c + + #define CONSTANT 0x4000 + #define CONSTEXP (CONSTANT | 3) + +5) namespace collisions when defining local variables in macros resembling + functions: + + .. code-block:: c + + #define FOO(x) \ + ({ \ + typeof(x) ret; \ + ret = calc_ret(x); \ + (ret); \ + }) + + ret is a common name for a local variable - __foo_ret is less likely + to collide with an existing variable. + + The cpp manual deals with macros exhaustively. The gcc internals manual also + covers RTL which is used frequently with assembly language in the kernel. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 11: Allocating memory +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +.. admonition:: Tarantool Style + :class: FACT + + Prefer the supplied slab (salloc) and pool (palloc) allocators to malloc()/free() + for any performance-intensive or large memory allocations. Repetitive use of + malloc()/free() can lead to memory fragmentation and should therefore be avoided. + + Always free all allocated memory, even allocated at start-up. We aim at being + valgrind leak-check clean, and in most cases it's just as easy to free() the + allocated memory as it is to write a valgrind suppression. Freeing all allocated + memory is also dynamic-load friendly: assuming a plug-in can be dynamically + loaded and unloaded multiple times, reload should not lead to a memory leak. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 12: The inline disease +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +There appears to be a common misperception that gcc has a magic "make me +faster" speedup option called ``inline``. While the use of inlines can be +appropriate (for example as a means of replacing macros, see Chapter 12), it +very often is not. Abundant use of the inline keyword leads to a much bigger +kernel, which in turn slows the system as a whole down, due to a bigger +icache footprint for the CPU and simply because there is less memory +available for the pagecache. Just think about it; a pagecache miss causes a +disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles +that can go into these 5 milliseconds. + +A reasonable rule of thumb is to not put inline at functions that have more +than 3 lines of code in them. An exception to this rule are the cases where +a parameter is known to be a compiletime constant, and as a result of this +constantness you *know* the compiler will be able to optimize most of your +function away at compile time. For a good example of this later case, see +the kmalloc() inline function. + +Often people argue that adding inline to functions that are static and used +only once is always a win since there is no space tradeoff. While this is +technically correct, gcc is capable of inlining these automatically without +help, and the maintenance issue of removing the inline when a second user +appears outweighs the potential value of the hint that tells gcc to do +something it would have done anyway. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 13: Function return values and names +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Functions can return values of many different kinds, and one of the +most common is a value indicating whether the function succeeded or +failed. Such a value can be represented as an error-code integer +(-Exxx = failure, 0 = success) or a ``succeeded`` boolean (0 = failure, +non-zero = success). + +Mixing up these two sorts of representations is a fertile source of +difficult-to-find bugs. If the C language included a strong distinction +between integers and booleans then the compiler would find these mistakes +for us... but it doesn't. To help prevent such bugs, always follow this +convention: + + If the name of a function is an action or an imperative command, + the function should return an error-code integer. If the name + is a predicate, the function should return a "succeeded" boolean. + +For example, ``add work`` is a command, and the add_work() function returns 0 +for success or -EBUSY for failure. In the same way, ``PCI device present`` is +a predicate, and the pci_dev_present() function returns 1 if it succeeds in +finding a matching device or 0 if it doesn't. + +All EXPORTed functions must respect this convention, and so should all +public functions. Private (static) functions need not, but it is +recommended that they do. + +Functions whose return value is the actual result of a computation, rather +than an indication of whether the computation succeeded, are not subject to +this rule. Generally they indicate failure by returning some out-of-range +result. Typical examples would be functions that return pointers; they use +NULL or the ERR_PTR mechanism to report failure. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 14: Editor modelines and other cruft +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Some editors can interpret configuration information embedded in source files, +indicated with special markers. For example, emacs interprets lines marked +like this: + +.. code-block:: c + + -*- mode: c -*- + +Or like this: + +.. code-block:: c + + /* + Local Variables: + compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" + End: + */ + +Vim interprets markers that look like this: + +.. code-block:: c + + /* vim:set sw=8 noet */ + +Do not include any of these in source files. People have their own personal +editor configurations, and your source files should not override them. This +includes markers for indentation and mode configuration. People may use their +own custom mode, or may have some other magic method for making indentation +work correctly. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 15: Inline assembly +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +In architecture-specific code, you may need to use inline assembly to interface +with 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 can +and should poke hardware from C when possible. + +Consider writing simple helper functions that wrap common bits of inline +assembly, rather than repeatedly writing them with slight variations. Remember +that inline assembly can use C parameters. + +Large, non-trivial assembly functions should go in .S files, with corresponding +C prototypes defined in C header files. The C prototypes for assembly +functions should use ``asmlinkage``. + +You may need to mark your asm statement as volatile, to prevent GCC from +removing it if GCC doesn't notice any side effects. You don't always need to +do so, though, and doing so unnecessarily can limit optimization. + +When writing a single inline assembly statement containing multiple +instructions, put each instruction on a separate line in a separate quoted +string, and end each string except the last with \n\t to properly indent the +next instruction in the assembly output: + +.. code-block:: c + + asm ("magic %reg1, #42\n\t" + "more_magic %reg2, %reg3" + : /* outputs */ : /* inputs */ : /* clobbers */); + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 16: Conditional Compilation +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Wherever possible, don't use preprocessor conditionals (#if, #ifdef) in .c +files; 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 .c +files, providing no-op stub versions in the #else case, and then call those +functions unconditionally from .c files. The compiler will avoid generating +any code for the stub calls, producing identical results, but the logic will +remain easy to follow. + +Prefer to compile out entire functions, rather than portions of functions or +portions of expressions. Rather than putting an ifdef in an expression, factor +out part or all of the expression into a separate helper function and apply the +conditional to that function. + +If you have a function or variable which may potentially go unused in a +particular configuration, and the compiler would warn about its definition +going unused, mark the definition as __maybe_unused rather than wrapping it in +a preprocessor conditional. (However, if a function or variable *always* goes +unused, delete it.) + +Within code, where possible, use the IS_ENABLED macro to convert a Kconfig +symbol into a C boolean expression, and use it in a normal C conditional: + +.. code-block:: c + + if (IS_ENABLED(CONFIG_SOMETHING)) { + ... + } + +The compiler will constant-fold the conditional away, and include or exclude +the block of code just as with an #ifdef, so this will not add any runtime +overhead. +However, this approach still allows the C compiler to see the code +inside the block, and check it for correctness (syntax, types, symbol +references, etc). Thus, you still have to use an #ifdef if the code inside the +block 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 conditional +expression used. For instance: + +.. code-block:: c + + #ifdef CONFIG_SOMETHING + ... + #endif /* CONFIG_SOMETHING */ + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 17: Header files +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +.. admonition:: Tarantool Style + :class: FACT + + Use header guards. Put the header guard in the first line in the header, + before the copyright or declarations. Use all-uppercase name for the header + guard. Derive the header guard name from the file name, and append _INCLUDED + to get a macro name. For example, core/log_io.h -> CORE_LOG_IO_H_INCLUDED. In + ``.c`` (implementation) file, include the respective declaration header before + all other headers, to ensure that the header is self- sufficient. Header + "header.h" is self-sufficient if the following compiles without errors: + + .. code-block:: c + + #include "header.h" + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 18: Other +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +.. admonition:: Tarantool Style + :class: FACT + + * We don't apply ``!`` operator to non-boolean values. It means, to check + if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, + you use ``!= NULL``. The same for ``==``. + + * Select GNU C99 extensions are acceptable. It's OK to mix declarations and + statements, use true and false. + + * The not-so-current list of all GCC C extensions can be found at: + http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Appendix I: References +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +* `The C Programming Language, Second Edition `_ + by Brian W. Kernighan and Dennis M. Ritchie. |br| + Prentice Hall, Inc., 1988. |br| + ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback). + +* `The Practice of Programming `_ + by Brian W. Kernighan and Rob Pike. |br| + Addison-Wesley, Inc., 1999. |br| + ISBN 0-201-61586-X. + +* `GNU manuals `_ - where in compliance with K&R and this text - for **cpp**, **gcc**, + **gcc internals** and **indent** + +* `WG14 International standardization workgroup for the programming + language C `_ + +* `Kernel CodingStyle, by greg@kroah.com at OLS 2002 + `_ + diff --git a/doc/dev_guide/c_style_guide_old.rst b/doc/dev_guide/c_style_guide_old.rst new file mode 100644 index 0000000000..e8aebf7141 --- /dev/null +++ b/doc/dev_guide/c_style_guide_old.rst @@ -0,0 +1,1037 @@ +------------------------------------------------------------------------------- + C Style Guide +------------------------------------------------------------------------------- + +The project's coding style is based on a version of the Linux kernel coding style. + +The latest version of the Linux style can be found at: +http://www.kernel.org/doc/Documentation/CodingStyle + +Since it is open for changes, the version of style that we follow, +one from 2007-July-13, will be also copied later in this document. + +There are a few additional guidelines, either unique +to Tarantool or deviating from the Kernel guidelines. + +A. Chapters 10 "Kconfig configuration files", 11 "Data structures", + 13 "Printing kernel messages", 14 "Allocating memory" and 17 + "Don't re-invent the kernel macros" do not apply, since they are + specific to Linux kernel programming environment. + +B. The rest of Linux Kernel Coding Style is amended as follows: + +=========================================================== + General guidelines +=========================================================== + +We use Git for revision control. The latest development is happening in the +default branch (currently ``master``). +Our git repository is hosted on github, and can be checked out with +``git clone git://github.com/tarantool/tarantool.git`` (anonymous read-only access). + +If you have any questions about Tarantool internals, please post them on the +developer discussion list, https://groups.google.com/forum/#!forum/tarantool. However, +please be warned: Launchpad silently deletes posts from non-subscribed members, +thus please be sure to have subscribed to the list prior to posting. Additionally, +some engineers are always present on #tarantool channel on irc.freenode.net. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Commenting style +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Use Doxygen comment format, Javadoc flavor, i.e. `@tag` rather than `\\tag`. +The main tags in use are @param, @retval, @return, @see, @note and @todo. + +Every function, except perhaps a very short and obvious one, should have a +comment. A sample function comment may look like below: + +.. code-block:: c + + /** Write all data to a descriptor. + * + * This function is equivalent to 'write', except it would ensure + * that all data is written to the file unless a non-ignorable + * error occurs. + * + * @retval 0 Success + * + * @retval 1 An error occurred (not EINTR) + * / + static int + write_all(int fd, void \*data, size_t len); + +Public structures and important structure members should be commented as well. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Header files +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Use header guards. Put the header guard in the first line in the header, +before the copyright or declarations. Use all-uppercase name for the header +guard. Derive the header guard name from the file name, and append _INCLUDED +to get a macro name. For example, core/log_io.h -> CORE_LOG_IO_H_INCLUDED. In +``.c`` (implementation) file, include the respective declaration header before all +other headers, to ensure that the header is self- sufficient. Header "header.h" +is self-sufficient if the following compiles without errors: + +.. code-block:: c + + #include "header.h" + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Allocating memory +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Prefer the supplied slab (salloc) and pool (palloc) allocators to malloc()/free() +for any performance-intensive or large memory allocations. Repetitive use of +malloc()/free() can lead to memory fragmentation and should therefore be avoided. + +Always free all allocated memory, even allocated at start-up. We aim at being +valgrind leak-check clean, and in most cases it's just as easy to free() the +allocated memory as it is to write a valgrind suppression. Freeing all allocated +memory is also dynamic-load friendly: assuming a plug-in can be dynamically loaded +and unloaded multiple times, reload should not lead to a memory leak. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Function naming +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Our convention is to use: + +* ``new``/``delete`` for functions which + allocate + initialize and destroy + deallocate an object, +* ``create``/``destroy`` for functions which initialize/destroy an object + but do not handle memory management, +* ``init``/``free`` for functions which initialize/destroy libraries and subsystems. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Other +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Select GNU C99 extensions are acceptable. It's OK to mix declarations and statements, +use true and false. + +The not-so-current list of all GCC C extensions can be found at: +http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html + +=========================================================== + Linux kernel coding style +=========================================================== + +This is a short document describing the preferred coding style for the +linux kernel. Coding style is very personal, and I won't _force_ my +views on anybody, but this is what goes for anything that I have to be +able to maintain, and I'd prefer it for most other things too. Please +at least consider the points made here. + +First off, I'd suggest printing out a copy of the GNU coding standards, +and NOT read it. Burn them, it's a great symbolic gesture. + +Anyway, here goes: + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 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 to +be 3. + +Rationale: The whole idea behind indentation is to clearly define where +a block of control starts and ends. Especially when you've been looking +at your screen for 20 straight hours, you'll find it a lot easier to see +how the indentation works if you have large indentations. + +Now, some people will claim that having 8-character indentations makes +the code move too far to the right, and makes it hard to read on a +80-character terminal screen. The answer to that is that if you need +more than 3 levels of indentation, you're screwed anyway, and should fix +your program. + +In short, 8-char indents make things easier to read, and have the added +benefit 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 is +to align the "switch" and its subordinate "case" labels in the same column +instead of "double-indenting" the "case" labels. e.g.: + +.. code-block:: c + + switch (suffix) { + case 'G': + case 'g': + mem <<= 30; + break; + case 'M': + case 'm': + mem <<= 20; + break; + case 'K': + case 'k': + mem <<= 10; + /* fall through */ + default: + break; + } + + +Don't put multiple statements on a single line unless you have +something to hide: + +.. code-block:: none + + if (condition) do_this; + do_something_everytime; + +Don't put multiple assignments on a single line either. Kernel coding style +is super simple. Avoid tricky expressions. + +Outside of comments, documentation and except in Kconfig, spaces are never +used for indentation, and the above example is deliberately broken. + +Get a decent editor and don't leave whitespace at the end of lines. + + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 2: Breaking long lines and strings +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Coding style is all about readability and maintainability using commonly +available tools. + +The limit on the length of lines is 80 columns, and this is a strongly +preferred limit. As for comments, the same limit of 80 columns is applied. + +Statements longer than 80 columns will be broken into sensible chunks. +Descendants are always substantially shorter than the parent and are placed +substantially to the right. The same applies to function headers with a long +argument list. Long strings are as well broken into shorter strings. The +only exception to this is where exceeding 80 columns significantly increases +readability and does not hide information. + +.. code-block:: c + + void fun(int a, int b, int c) + { + if (condition) + printk(KERN_WARNING "Warning this is a long printk with " + "3 parameters a: %u b: %u " + "c: %u \n", a, b, c); + else + next_statement; + } + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 3: Placing Braces and Spaces +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +The other issue that always comes up in C styling is the placement of +braces. Unlike the indent size, there are few technical reasons to +choose one placement strategy over the other, but the preferred way, as +shown to us by the prophets Kernighan and Ritchie, is to put the opening +brace last on the line, and put the closing brace first, thusly: + +.. code-block:: none + + if (x is true) { + we do y + } + +This applies to all non-function statement blocks (if, switch, for, +while, do). e.g.: + +.. code-block:: c + + switch (action) { + case KOBJ_ADD: + return "add"; + case KOBJ_REMOVE: + return "remove"; + case KOBJ_CHANGE: + return "change"; + default: + return NULL; + } + +However, there is one special case, namely functions: they have the +opening brace at the beginning of the next line, thus: + +.. code-block:: c + + int function(int x) + { + body of function; + } + +Heretic people all over the world have claimed that this inconsistency +is ... well ... inconsistent, but all right-thinking people know that +(a) K&R are _right_ and (b) K&R are right. Besides, functions are +special anyway (you can't nest them in C). + +Note that the closing brace is empty on a line of its own, _except_ in +the cases where it is followed by a continuation of the same statement, +ie a "while" in a do-statement or an "else" in an if-statement, like +this: + +.. code-block:: c + + do { + body of do-loop; + } while (condition); + +and + +.. code-block:: c + + if (x == y) { + .. + } else if (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 the +supply of new-lines on your screen is not a renewable resource (think +25-line terminal screens here), you have more empty lines to put +comments on. + +Do not unnecessarily use braces where a single statement will do. + +.. code-block:: c + + if (condition) + action(); + +This does not apply if one branch of a conditional statement is a single +statement. Use braces in both branches. + +.. code-block:: c + + if (condition) { + do_this(); + do_that(); + } else { + otherwise(); + } + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 3.1: Spaces +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Linux kernel style for use of spaces depends (mostly) on +function-versus-keyword usage. Use a space after (most) keywords. The +notable exceptions are sizeof, typeof, alignof, and __attribute__, which look +somewhat like functions (and are usually used with parentheses in Linux, +although they are not required in the language, as in: "sizeof info" after +"struct fileinfo info;" is declared). + +So use a space after these keywords: if, switch, case, for, do, while +but not with sizeof, typeof, alignof, or __attribute__. E.g., + +.. code-block:: c + + s = sizeof(struct file); + +Do not add spaces around (inside) parenthesized expressions. This example is +**bad**: + +.. code-block:: c + + s = sizeof( struct file ); + +When declaring pointer data or a function that returns a pointer type, the +preferred use of '*' is adjacent to the data name or function name and not +adjacent to the type name. Examples: + +.. code-block:: c + + char *linux_banner; + unsigned long long memparse(char *ptr, char **retptr); + char *match_strdup(substring_t *s); + +Use one space around (on each side of) most binary and ternary operators, +such as any of these: + + = + - < > * / % | & ^ <= >= == != ? : + +but no space after unary operators: + + & * + - ~ ! sizeof typeof alignof __attribute__ defined + +no space before the postfix increment & decrement unary operators: + + ++ -- + +no space after the prefix increment & decrement unary operators: + + ++ -- + +and no space around the '.' and "->" structure member operators. + +Do not leave trailing whitespace at the ends of lines. Some editors with +"smart" indentation will insert whitespace at the beginning of new lines as +appropriate, so you can start typing the next line of code right away. +However, some such editors do not remove the whitespace if you end up not +putting a line of code there, such as if you leave a blank line. As a result, +you end up with lines containing trailing whitespace. + +Git will warn you about patches that introduce trailing whitespace, and can +optionally strip the trailing whitespace for you; however, if applying a series +of patches, this may make later patches in the series fail by changing their +context lines. + + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 4: Naming +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +C is a Spartan language, and so should your naming be. Unlike Modula-2 +and Pascal programmers, C programmers do not use cute names like +ThisVariableIsATemporaryCounter. A C programmer would call that +variable "tmp", which is much easier to write, and not the least more +difficult to understand. + +HOWEVER, while mixed-case names are frowned upon, descriptive names for +global variables are a must. To call a global function "foo" is a +shooting offense. + +GLOBAL variables (to be used only if you _really_ need them) need to +have descriptive names, as do global functions. If you have a function +that counts the number of active users, you should call that +"count_active_users()" or similar, you should _not_ call it "cntusr()". + +Encoding the type of a function into the name (so-called Hungarian +notation) is brain damaged - the compiler knows the types anyway and can +check those, and it only confuses the programmer. No wonder MicroSoft +makes buggy programs. + +LOCAL variable names should be short, and to the point. If you have +some random integer loop counter, it should probably be called "i". +Calling it "loop_counter" is non-productive, if there is no chance of it +being mis-understood. Similarly, "tmp" can be just about any type of +variable that is used to hold a temporary value. + +If you are afraid to mix up your local variable names, you have another +problem, which is called the function-growth-hormone-imbalance syndrome. +See chapter 6 (Functions). + + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 5: Typedefs +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Please don't use things like "vps_t". + +It's a _mistake_ to use typedef for structures and pointers. When you see a + +.. code-block:: c + + vps_t a; + +in the source, what does it mean? + +In contrast, if it says + +.. code-block:: c + + struct virtual_container *a; + +you can actually tell what "a" is. + +Lots of people think that typedefs "help readability". Not so. They are +useful only for: + +(a) totally opaque objects (where the typedef is actively used to _hide_ + what the object is). + + Example: "pte_t" etc. opaque objects that you can only access using + the proper accessor functions. + + NOTE! Opaqueness and "accessor functions" are not good in themselves. + The reason we have them for things like pte_t etc. is that there + really is absolutely _zero_ portably accessible information there. + +(b) Clear integer types, where the abstraction _helps_ avoid confusion + whether it is "int" or "long". + + u8/u16/u32 are perfectly fine typedefs, although they fit into + category (d) better than here. + + NOTE! Again - there needs to be a _reason_ for this. If something is + "unsigned long", then there's no reason to do + + .. code-block:: c + + typedef unsigned long myflags_t; + + but if there is a clear reason for why it under certain circumstances + might be an "unsigned int" and under other configurations might be + "unsigned long", then by all means go ahead and use a typedef. + +(c) when you use sparse to literally create a _new_ type for + type-checking. + +(d) New types which are identical to standard C99 types, in certain + exceptional circumstances. + + Although it would only take a short amount of time for the eyes and + brain to become accustomed to the standard types like 'uint32_t', + some people object to their use anyway. + + Therefore, the Linux-specific 'u8/u16/u32/u64' types and their + signed equivalents which are identical to standard types are + permitted -- although they are not mandatory in new code of your + own. + + When editing existing code which already uses one or the other set + of types, you should conform to the existing choices in that code. + +(e) Types safe for use in userspace. + + In certain structures which are visible to userspace, we cannot + require C99 types and cannot use the 'u32' form above. Thus, we + use __u32 and similar types in all structures which are shared + with userspace. + +Maybe there are other cases too, but the rule should basically be to NEVER +EVER use a typedef unless you can clearly match one of those rules. + +In general, a pointer, or a struct that has elements that can reasonably +be directly accessed should **never** be a typedef. + + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 6: Functions +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Functions should be short and sweet, and do just one thing. They should +fit 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 the +complexity and indentation level of that function. So, if you have a +conceptually simple function that is just one long (but simple) +case-statement, where you have to do lots of small things for a lot of +different cases, it's OK to have a longer function. + +However, if you have a complex function, and you suspect that a +less-than-gifted first-year high-school student might not even +understand what the function is all about, you should adhere to the +maximum limits all the more closely. Use helper functions with +descriptive names (you can ask the compiler to in-line them if you think +it's performance-critical, and it will probably do a better job of it +than you would have done). + +Another measure of the function is the number of local variables. They +shouldn't exceed 5-10, or you're doing something wrong. Re-think the +function, and split it into smaller pieces. A human brain can +generally easily keep track of about 7 different things, anything more +and it gets confu/sed. You know you're brilliant, but maybe you'd like +to understand what you did 2 weeks from now. + +In source files, separate functions with one blank line. If the function is +exported, the EXPORT* macro for it should follow immediately after the closing +function brace line. E.g.: + +.. code-block:: c + + int system_is_up(void) + { + return system_state == SYSTEM_RUNNING; + } + EXPORT_SYMBOL(system_is_up); + +In function prototypes, include parameter names with their data types. +Although this is not required by the C language, it is preferred in Linux +because it is a simple way to add valuable information for the reader. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 7: Centralized exiting of functions +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Albeit deprecated by some people, the equivalent of the goto statement is +used frequently by compilers in form of the unconditional jump instruction. + +The goto statement comes in handy when a function exits from multiple +locations and some common work such as cleanup has to be done. + +The rationale is: + +- unconditional statements are easier to understand and follow +- nesting is reduced +- errors by not updating individual exit points when making + modifications are prevented +- saves the compiler work to optimize redundant code away ;) + +.. code-block:: c + + int fun(int a) + { + int result = 0; + char *buffer = kmalloc(SIZE); + + if (buffer == NULL) + return -ENOMEM; + + if (condition1) { + while (loop1) { + ... + } + result = 1; + goto out; + } + ... + out: + kfree(buffer); + return result; + } + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 8: Commenting +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Comments are good, but there is also a danger of over-commenting. NEVER +try to explain HOW your code works in a comment: it's much better to +write the code so that the _working_ is obvious, and it's a waste of +time to explain badly written code. +с +Generally, you want your comments to tell WHAT your code does, not HOW. +Also, try to avoid putting comments inside a function body: if the +function 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 make +small comments to note or warn about something particularly clever (or +ugly), but try to avoid excess. Instead, put the comments at the head +of the function, telling people what it does, and possibly WHY it does +it. + +When commenting the kernel API functions, please use the kernel-doc format. +See the files Documentation/kernel-doc-nano-HOWTO.txt and scripts/kernel-doc +for details. + +Linux style for comments is the C89 :code:`/* ... */`` style. +Don't use C99-style :code:`// ...` comments. + +The preferred style for long (multi-line) comments is: + +.. code-block:: c + + /* + * 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 derived +types. To this end, use just one data declaration per line (no commas for +multiple data declarations). This leaves you room for a small comment on each +item, explaining its use. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 9: You've made a mess of it +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +That's OK, we all do. You've probably been told by your long-time Unix +user helper that "GNU emacs" automatically formats the C sources for +you, and you've noticed that yes, it does do that, but the defaults it +uses are less than desirable (in fact, they are worse than random +typing - an infinite number of monkeys typing into GNU emacs would never +make a good program). + +So, you can either get rid of GNU emacs, or change it to use saner +values. To do the latter, you can stick the following in your .emacs file: + +.. code-block:: lisp + + (defun c-lineup-arglist-tabs-only (ignored) + "Line up argument lists by tabs, not spaces" + (let* ((anchor (c-langelem-pos c-syntactic-element)) + (column (c-langelem-2nd-pos c-syntactic-element)) + (offset (- (1+ column) anchor)) + (steps (floor offset c-basic-offset))) + (* (max steps 1) + c-basic-offset))) + + (add-hook 'c-mode-common-hook + (lambda () + ;; Add kernel style + (c-add-style + "linux-tabs-only" + '("linux" (c-offsets-alist + (arglist-cont-nonempty + c-lineup-gcc-asm-reg + c-lineup-arglist-tabs-only)))))) + + (add-hook 'c-mode-hook + (lambda () + (let ((filename (buffer-file-name))) + ;; Enable kernel mode for the appropriate files + (when (and filename + (string-match (expand-file-name "~/src/linux-trees") + filename)) + (setq indent-tabs-mode t) + (c-set-style "linux-tabs-only"))))) + +This will make emacs go better with the kernel coding style for C +files below ~/src/linux-trees. + +But even if you fail in getting emacs to do sane formatting, not +everything is lost: use "indent". + +Now, again, GNU indent has the same brain-dead settings that GNU emacs +has, 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 indent +recognize the authority of K&R (the GNU people aren't evil, they are +just severely misguided in this matter), so you just give indent the +options "-kr -i8" (stands for "K&R, 8 character indents"), or use +"scripts/Lindent", which indents in the latest style. + +"indent" has a lot of options, and especially when it comes to comment +re-formatting you may want to take a look at the man page. But +remember: "indent" is not a fix for bad programming. + + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 10: Kconfig configuration files +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +For all of the Kconfig* configuration files throughout the source tree, +the indentation is somewhat different. Lines under a "config" definition +are indented with one tab, while help text is indented an additional two +spaces. Example: + +.. code-block:: kconfig + + 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. + +Features that might still be considered unstable should be defined as +dependent on "EXPERIMENTAL": + +.. code-block:: kconfig + + config SLUB + depends on EXPERIMENTAL && !ARCH_USES_SLAB_PAGE_STRUCT + bool "SLUB (Unqueued Allocator)" + ... + +while seriously dangerous features (such as write support for certain +filesystems) should advertise this prominently in their prompt string: + +.. code-block:: kconfig + + config ADFS_FS_RW + bool "ADFS write support (DANGEROUS)" + depends on ADFS_FS + ... + +For full documentation on the configuration files, see the file +Documentation/kbuild/kconfig-language.txt. + + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 11: Data structures +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Data structures that have visibility outside the single-threaded +environment they are created and destroyed in should always have +reference counts. In the kernel, garbage collection doesn't exist (and +outside the kernel garbage collection is slow and inefficient), which +means that you absolutely _have_ to reference count all your uses. + +Reference counting means that you can avoid locking, and allows multiple +users to have access to the data structure in parallel - and not having +to worry about the structure suddenly going away from under them just +because they slept or did something else for a while. + +Note that locking is _not_ a replacement for reference counting. +Locking is used to keep data structures coherent, while reference +counting is a memory management technique. Usually both are needed, and +they are not to be confused with each other. + +Many data structures can indeed have two levels of reference counting, +when there are users of different "classes". The subclass count counts +the number of subclass users, and decrements the global count just once +when the subclass count goes to zero. + +Examples of this kind of "multi-level-reference-counting" can be found in +memory management ("struct mm_struct": mm_users and mm_count), and in +filesystem code ("struct super_block": s_count and s_active). + +Remember: if another thread can find your data structure, and you don't +have a reference count on it, you almost certainly have a bug. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 12: Macros, Enums and RTL +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Names of macros defining constants and labels in enums are capitalized. + +.. code-block:: c + + #define CONSTANT 0x12345 + +Enums are preferred when defining several related constants. + +CAPITALIZED macro names are appreciated but macros resembling functions +may 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: + +.. code-block:: c + + #define macrofun(a, b, c) \ + do { \ + if (a == 5) \ + do_this(b, c); \ + } while (0) + +Things to avoid when using macros: + +1. macros that affect control flow: + + .. code-block:: c + + #define FOO(x) \ + do { \ + if (blah(x) < 0) \ + return -EBUGGERED; \ + } while(0) + + is a _very_ bad idea. It looks like a function call but exits the "calling" + function; 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: + + .. code-block:: c + + #define FOO(val) bar(index, val) + + might look like a good thing, but it's confusing as hell when one reads the + code and it's prone to breakage from seemingly innocent changes. + +3. macros with arguments that are used as l-values: FOO(x) = y; will + bite you if somebody e.g. turns FOO into an inline function. + +4. forgetting about precedence: macros defining constants using expressions + must enclose the expression in parentheses. Beware of similar issues with + macros using parameters. + + .. code-block:: c + + #define CONSTANT 0x4000 + #define CONSTEXP (CONSTANT | 3) + + The cpp manual deals with macros exhaustively. The gcc internals manual also + covers RTL which is used frequently with assembly language in the kernel. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 13: Printing kernel messages +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Kernel developers like to be seen as literate. Do mind the spelling +of kernel messages to make a good impression. Do not use crippled +words like "dont"; use "do not" or "don't" instead. Make the messages +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 +which you should use to make sure messages are matched to the right device +and driver, and are tagged with the right level: dev_err(), dev_warn(), +dev_info(), and so forth. For messages that aren't associated with a +particular device, defines pr_debug() and pr_info(). + +Coming up with good debugging messages can be quite a challenge; and once +you have them, they can be a huge help for remote troubleshooting. Such +messages should be compiled out when the DEBUG symbol is not defined (that +is, by default they are not included). When you use dev_dbg() or pr_debug(), +that's automatic. Many subsystems have Kconfig options to turn on -DDEBUG. +A related convention uses VERBOSE_DEBUG to add dev_vdbg() messages to the +ones already enabled by DEBUG. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 14: Allocating memory +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +The kernel provides the following general purpose memory allocators: +kmalloc(), kzalloc(), kcalloc(), and vmalloc(). Please refer to the API +documentation for further information about them. + +The preferred form for passing a size of a struct is the following: + +.. code-block:: c + + p = kmalloc(sizeof(*p), ...); + +The alternative form where struct name is spelled out hurts readability and +introduces an opportunity for a bug when the pointer variable type is changed +but the corresponding sizeof that is passed to a memory allocator is not. + +Casting the return value which is a void pointer is redundant. The conversion +from void pointer to any other pointer type is guaranteed by the C programming +language. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 15: The inline disease +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +There appears to be a common misperception that gcc has a magic "make me +faster" speedup option called "inline". While the use of inlines can be +appropriate (for example as a means of replacing macros, see Chapter 12), it +very often is not. Abundant use of the inline keyword leads to a much bigger +kernel, which in turn slows the system as a whole down, due to a bigger +icache footprint for the CPU and simply because there is less memory +available for the pagecache. Just think about it; a pagecache miss causes a +disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles +that can go into these 5 milliseconds. + +A reasonable rule of thumb is to not put inline at functions that have more +than 3 lines of code in them. An exception to this rule are the cases where +a parameter is known to be a compiletime constant, and as a result of this +constantness you *know* the compiler will be able to optimize most of your +function away at compile time. For a good example of this later case, see +the kmalloc() inline function. + +Often people argue that adding inline to functions that are static and used +only once is always a win since there is no space tradeoff. While this is +technically correct, gcc is capable of inlining these automatically without +help, and the maintenance issue of removing the inline when a second user +appears outweighs the potential value of the hint that tells gcc to do +something it would have done anyway. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 16: Function return values and names +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Functions can return values of many different kinds, and one of the +most common is a value indicating whether the function succeeded or +failed. Such a value can be represented as an error-code integer +(-Exxx = failure, 0 = success) or a "succeeded" boolean (0 = failure, +non-zero = success). + +Mixing up these two sorts of representations is a fertile source of +difficult-to-find bugs. If the C language included a strong distinction +between integers and booleans then the compiler would find these mistakes +for us... but it doesn't. To help prevent such bugs, always follow this +convention: + +:: + + If the name of a function is an action or an imperative command, + the function should return an error-code integer. If the name + is a predicate, the function should return a "succeeded" boolean. + +For example, "add work" is a command, and the add_work() function returns 0 +for success or -EBUSY for failure. In the same way, "PCI device present" is +a predicate, and the pci_dev_present() function returns 1 if it succeeds in +finding a matching device or 0 if it doesn't. + +All EXPORTed functions must respect this convention, and so should all +public functions. Private (static) functions need not, but it is +recommended that they do. + +Functions whose return value is the actual result of a computation, rather +than an indication of whether the computation succeeded, are not subject to +this rule. Generally they indicate failure by returning some out-of-range +result. Typical examples would be functions that return pointers; they use +NULL or the ERR_PTR mechanism to report failure. + + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 17: Don't re-invent the kernel macros +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +The header file include/linux/kernel.h contains a number of macros that +you should use, rather than explicitly coding some variant of them yourself. +For example, if you need to calculate the length of an array, take advantage +of the macro + +.. code-block:: c + + #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) + +Similarly, if you need to calculate the size of some structure member, use + +.. code-block:: c + + #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f)) + +There are also min() and max() macros that do strict type checking if you +need them. Feel free to peruse that header file to see what else is already +defined that you shouldn't reproduce in your code. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Chapter 18: Editor modelines and other cruft +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Some editors can interpret configuration information embedded in source files, +indicated with special markers. For example, emacs interprets lines marked +like this: + +.. code-block:: none + + -*- mode: c -*- + +Or like this: + +.. code-block:: none + + /* + Local Variables: + compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" + End: + */ + +Vim interprets markers that look like this: + +.. code-block:: none + + /* vim:set sw=8 noet */ + +Do not include any of these in source files. People have their own personal +editor configurations, and your source files should not override them. This +includes markers for indentation and mode configuration. People may use their +own custom mode, or may have some other magic method for making indentation +work correctly. + +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + Appendix I: References +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +* `The C Programming Language, Second Edition `_ + by Brian W. Kernighan and Dennis M. Ritchie. |br| + Prentice Hall, Inc., 1988. |br| + ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback). + +* `The Practice of Programming `_ + by Brian W. Kernighan and Rob Pike. |br| + Addison-Wesley, Inc., 1999. |br| + ISBN 0-201-61586-X. + +* `GNU manuals `_ - where in compliance with K&R and this text - for **cpp**, **gcc**, + **gcc internals** and **indent** + +* `WG14 International standardization workgroup for the programming + language C `_ + +* `Kernel CodingStyle, by greg@kroah.com at OLS 2002 + `_ From 57cf0143f5c09078f55f15b915711898664c3b31 Mon Sep 17 00:00:00 2001 From: Natalia Ogoreltseva Date: Mon, 15 Mar 2021 14:31:12 +0300 Subject: [PATCH 2/8] fix indentation according to tarantool/doc guidelines --- doc/dev_guide/c_style_guide_new.rst | 642 ++++++++++++++-------------- 1 file changed, 321 insertions(+), 321 deletions(-) diff --git a/doc/dev_guide/c_style_guide_new.rst b/doc/dev_guide/c_style_guide_new.rst index 4657470d05..d87e260aa4 100644 --- a/doc/dev_guide/c_style_guide_new.rst +++ b/doc/dev_guide/c_style_guide_new.rst @@ -1,6 +1,6 @@ -------------------------------------------------------------------------------- +================================================================================ C Style Guide -------------------------------------------------------------------------------- +================================================================================ We use Git for revision control. The latest development is happening in the default branch (currently ``master``). Our git repository is hosted on GitHub, @@ -26,24 +26,24 @@ We don't cite chapters 10 "Kconfig configuration files", 11 "Data structures", 13 "Printing kernel messages", and 17 "Don't re-invent the kernel macros" since they are specific to Linux kernel programming environment. -================================================================================ +-------------------------------------------------------------------------------- Linux kernel coding style -================================================================================ +-------------------------------------------------------------------------------- This is a short document describing the preferred coding style for the linux kernel. Coding style is very personal, and I won't **force** my views on anybody, but this is what goes for anything that I have to be -able to maintain, and I'd prefer it for most other things too. Please +able to maintain, and I'd prefer it for most other things too. Please at least consider the points made here. First off, I'd suggest printing out a copy of the GNU coding standards, -and NOT read it. Burn them, it's a great symbolic gesture. +and NOT read it. Burn them, it's a great symbolic gesture. Anyway, here goes: -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 1: Indentation -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Tabs are 8 characters, and thus indentations are also 8 characters. There are heretic movements that try to make indentations @@ -51,13 +51,13 @@ also 8 characters. There are heretic movements that try to make indentations value of PI to be 3. Rationale: The whole idea behind indentation is to clearly define where -a block of control starts and ends. Especially when you've been looking +a block of control starts and ends. Especially when you've been looking at your screen for 20 straight hours, you'll find it a lot easier to see how the indentation works if you have large indentations. Now, some people will claim that having 8-character indentations makes the code move too far to the right, and makes it hard to read on a -80-character terminal screen. The answer to that is that if you need +80-character terminal screen. The answer to that is that if you need more than 3 levels of indentation, you're screwed anyway, and should fix your program. @@ -69,7 +69,7 @@ The preferred way to ease multiple indentation levels in a switch statement is to align the ``switch`` and its subordinate ``case`` labels in the same column instead of ``double-indenting`` the ``case`` labels. E.g.: -.. code-block:: c +.. code-block:: c switch (suffix) { case 'G': @@ -91,10 +91,10 @@ instead of ``double-indenting`` the ``case`` labels. E.g.: Don't put multiple statements on a single line unless you have something to hide: -.. code-block:: c +.. code-block:: c - if (condition) do_this; - do_something_everytime; + if (condition) do_this; + do_something_everytime; Don't put multiple assignments on a single line either. Kernel coding style is super simple. Avoid tricky expressions. @@ -104,9 +104,9 @@ used for indentation, and the above example is deliberately broken. Get a decent editor and don't leave whitespace at the end of lines. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 2: Breaking long lines and strings -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Coding style is all about readability and maintainability using commonly available tools. @@ -114,7 +114,7 @@ available tools. The limit on the length of lines is 80 columns and this is a strongly preferred limit. -.. admonition:: Tarantool Style +.. admonition:: Tarantool Style :class: FACT As for comments, the same limit of 80 columns is applied. @@ -126,9 +126,9 @@ are placed substantially to the right. The same applies to function headers with a long argument list. However, never break user-visible strings such as printk messages, because that breaks the ability to grep for them. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 3: Placing Braces and Spaces -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The other issue that always comes up in C styling is the placement of braces. Unlike the indent size, there are few technical reasons to @@ -136,16 +136,16 @@ choose one placement strategy over the other, but the preferred way, as shown to us by the prophets Kernighan and Ritchie, is to put the opening brace last on the line, and put the closing brace first, thus: -.. code-block:: c +.. code-block:: c if (x is true) { we do y } This applies to all non-function statement blocks (if, switch, for, -while, do). E.g.: +while, do). E.g.: -.. code-block:: c +.. code-block:: c switch (action) { case KOBJ_ADD: @@ -161,7 +161,7 @@ while, do). E.g.: However, there is one special case, namely functions: they have the opening brace at the beginning of the next line, thus: -.. code-block:: c +.. code-block:: c int function(int x) @@ -170,8 +170,8 @@ opening brace at the beginning of the next line, thus: } Heretic people all over the world have claimed that this inconsistency -is ... well ... inconsistent, but all right-thinking people know that -(a) K&R are **right** and (b) K&R are right. Besides, functions are +is ... well ... inconsistent, but all right-thinking people know that +(a) K&R are **right** and (b) K&R are right. Besides, functions are special anyway (you can't nest them in C). Note that the closing brace is empty on a line of its own, **except** in @@ -179,7 +179,7 @@ the cases where it is followed by a continuation of the same statement, ie a ``while`` in a do-statement or an ``else`` in an if-statement, like this: -.. code-block:: c +.. code-block:: c do { body of do-loop @@ -187,7 +187,7 @@ this: and -.. code-block:: c +.. code-block:: c if (x == y) { .. @@ -200,20 +200,20 @@ and 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 the +(or almost empty) lines, without any loss of readability. Thus, as the supply of new-lines on your screen is not a renewable resource (think 25-line terminal screens here), you have more empty lines to put comments on. Do not unnecessarily use braces where a single statement will do. -.. code-block:: c +.. code-block:: c if (condition) action(); and -.. code-block:: none +.. code-block:: none if (condition) do_this(); @@ -223,7 +223,7 @@ Do not unnecessarily use braces where a single statement will do. This does not apply if only one branch of a conditional statement is a single statement; in the latter case use braces in both branches: -.. code-block:: c +.. code-block:: c if (condition) { do_this(); @@ -232,9 +232,9 @@ statement; in the latter case use braces in both branches: otherwise(); } -********************** +******************************************************************************** Chapter 3.1: Spaces -********************** +******************************************************************************** Tarantool style for use of spaces depends (mostly) on function-versus-keyword usage. Use a space after (most) keywords. The @@ -245,32 +245,32 @@ although they are not required in the language, as in: ``sizeof info`` after So use a space after these keywords: -.. code-block:: c +.. code-block:: c - if, switch, case, for, do, while + if, switch, case, for, do, while -but not with sizeof, typeof, alignof, or __attribute__. E.g., +but not with sizeof, typeof, alignof, or __attribute__. E.g., -.. code-block:: c +.. code-block:: c - s = sizeof(struct file); + s = sizeof(struct file); Do not add spaces around (inside) parenthesized expressions. This example is **bad**: -.. code-block:: c +.. code-block:: c - s = sizeof( struct file ); + s = sizeof( struct file ); When declaring pointer data or a function that returns a pointer type, the preferred use of ``*`` is adjacent to the data name or function name and not adjacent to the type name. Examples: -.. code-block:: c +.. code-block:: c - char *linux_banner; - unsigned long long memparse(char *ptr, char **retptr); - char *match_strdup(substring_t *s); + char *linux_banner; + unsigned long long memparse(char *ptr, char **retptr); + char *match_strdup(substring_t *s); Use one space around (on each side of) most binary and ternary operators, such as any of these:: @@ -303,15 +303,15 @@ optionally strip the trailing whitespace for you; however, if applying a series of patches, this may make later patches in the series fail by changing their context lines. -.. admonition:: Tarantool Style +.. admonition:: Tarantool Style :class: FACT Do not split a cast operator from its argument with a whitespace, e.g. ``(ssize_t)inj->iparam``. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 4: Naming -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ C is a Spartan language, and so should your naming be. Unlike Modula-2 and Pascal programmers, C programmers do not use cute names like @@ -320,11 +320,11 @@ variable ``tmp``, which is much easier to write, and not the least more difficult to understand. HOWEVER, while mixed-case names are frowned upon, descriptive names for -global variables are a must. To call a global function ``foo`` is a +global variables are a must. To call a global function ``foo`` is a shooting offense. GLOBAL variables (to be used only if you **really** need them) need to -have descriptive names, as do global functions. If you have a function +have descriptive names, as do global functions. If you have a function that counts the number of active users, you should call that ``count_active_users()`` or similar, you should **not** call it ``cntusr()``. @@ -333,102 +333,102 @@ notation) is brain damaged - the compiler knows the types anyway and can check those, and it only confuses the programmer. No wonder MicroSoft makes buggy programs. -LOCAL variable names should be short, and to the point. If you have +LOCAL variable names should be short, and to the point. If you have some random integer loop counter, it should probably be called ``i``. Calling it ``loop_counter`` is non-productive, if there is no chance of it -being mis-understood. Similarly, ``tmp`` can be just about any type of +being mis-understood. Similarly, ``tmp`` can be just about any type of variable that is used to hold a temporary value. If you are afraid to mix up your local variable names, you have another problem, which is called the function-growth-hormone-imbalance syndrome. See chapter 6 (Functions). -.. admonition:: Tarantool Style +.. admonition:: Tarantool Style :class: FACT For function naming we have a convention is to use: - * ``new``/``delete`` for functions which - allocate + initialize and destroy + deallocate an object, - * ``create``/``destroy`` for functions which initialize/destroy an object - but do not handle memory management, - * ``init``/``free`` for functions which initialize/destroy libraries and subsystems. + * ``new``/``delete`` for functions which + allocate + initialize and destroy + deallocate an object, + * ``create``/``destroy`` for functions which initialize/destroy an object + but do not handle memory management, + * ``init``/``free`` for functions which initialize/destroy libraries and subsystems. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 5: Typedefs -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Please don't use things like ``vps_t``. It's a **mistake** to use typedef for structures and pointers. When you see a -.. code-block:: c +.. code-block:: c - vps_t a; + vps_t a; in the source, what does it mean? In contrast, if it says -.. code-block:: c +.. code-block:: c - struct virtual_container *a; + struct virtual_container *a; you can actually tell what ``a`` is. Lots of people think that typedefs ``help readability``. Not so. They are useful only for: -#. totally opaque objects (where the typedef is actively used to **hide** +#. totally opaque objects (where the typedef is actively used to **hide** what the object is). - Example: ``pte_t`` etc. opaque objects that you can only access using - the proper accessor functions. + Example: ``pte_t`` etc. opaque objects that you can only access using + the proper accessor functions. - .. note:: + .. note:: - Opaqueness and ``accessor functions`` are not good in themselves. - The reason we have them for things like pte_t etc. is that there - really is absolutely **zero** portably accessible information there. + Opaqueness and ``accessor functions`` are not good in themselves. + The reason we have them for things like pte_t etc. is that there + really is absolutely **zero** portably accessible information there. -#. Clear integer types, where the abstraction **helps** avoid confusion - whether it is ``int`` or ``long``. +#. Clear integer types, where the abstraction **helps** avoid confusion + whether it is ``int`` or ``long``. - u8/u16/u32 are perfectly fine typedefs, although they fit into - point 4 better than here. + u8/u16/u32 are perfectly fine typedefs, although they fit into + point 4 better than here. - .. note:: + .. note:: - Again - there needs to be a **reason** for this. If something is - ``unsigned long``, then there's no reason to do - typedef unsigned long myflags_t; + Again - there needs to be a **reason** for this. If something is + ``unsigned long``, then there's no reason to do + typedef unsigned long myflags_t; - but if there is a clear reason for why it under certain circumstances - might be an ``unsigned int`` and under other configurations might be - ``unsigned long``, then by all means go ahead and use a typedef. + but if there is a clear reason for why it under certain circumstances + might be an ``unsigned int`` and under other configurations might be + ``unsigned long``, then by all means go ahead and use a typedef. -#. when you use sparse to literally create a **new** type for - type-checking. +#. when you use sparse to literally create a **new** type for + type-checking. -#. New types which are identical to standard C99 types, in certain - exceptional circumstances. +#. New types which are identical to standard C99 types, in certain + exceptional circumstances. - Although it would only take a short amount of time for the eyes and - brain to become accustomed to the standard types like ``uint32_t``, - some people object to their use anyway. + Although it would only take a short amount of time for the eyes and + brain to become accustomed to the standard types like ``uint32_t``, + some people object to their use anyway. - Therefore, the Linux-specific ``u8/u16/u32/u64`` types and their - signed equivalents which are identical to standard types are - permitted -- although they are not mandatory in new code of your - own. + Therefore, the Linux-specific ``u8/u16/u32/u64`` types and their + signed equivalents which are identical to standard types are + permitted -- although they are not mandatory in new code of your + own. - When editing existing code which already uses one or the other set - of types, you should conform to the existing choices in that code. + When editing existing code which already uses one or the other set + of types, you should conform to the existing choices in that code. -#. Types safe for use in userspace. +#. Types safe for use in userspace. - In certain structures which are visible to userspace, we cannot - require C99 types and cannot use the ``u32`` form above. Thus, we - use __u32 and similar types in all structures which are shared - with userspace. + In certain structures which are visible to userspace, we cannot + require C99 types and cannot use the ``u32`` form above. Thus, we + use __u32 and similar types in all structures which are shared + with userspace. Maybe there are other cases too, but the rule should basically be to NEVER EVER use a typedef unless you can clearly match one of those rules. @@ -436,16 +436,16 @@ EVER use a typedef unless you can clearly match one of those rules. In general, a pointer, or a struct that has elements that can reasonably be directly accessed should **never** be a typedef. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 6: Functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Functions should be short and sweet, and do just one thing. They should fit 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 the -complexity and indentation level of that function. So, if you have a +complexity and indentation level of that function. So, if you have a conceptually simple function that is just one long (but simple) case-statement, where you have to do lots of small things for a lot of different cases, it's OK to have a longer function. @@ -467,39 +467,39 @@ to understand what you did 2 weeks from now. In source files, separate functions with one blank line. If the function is exported, the **EXPORT** macro for it should follow immediately after the -closing function brace line. E.g.: +closing function brace line. E.g.: -.. code-block:: c +.. code-block:: c - int - system_is_up(void) - { - return system_state == SYSTEM_RUNNING; - } - EXPORT_SYMBOL(system_is_up); + int + system_is_up(void) + { + return system_state == SYSTEM_RUNNING; + } + EXPORT_SYMBOL(system_is_up); In function prototypes, include parameter names with their data types. Although this is not required by the C language, it is preferred in Linux because it is a simple way to add valuable information for the reader. -.. admonition:: Tarantool Style +.. admonition:: Tarantool Style :class: FACT Note that in Tarantool, we place the function return type on the line before the name and signature. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 7: Centralized exiting of functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Albeit deprecated by some people, the equivalent of the goto statement is used frequently by compilers in form of the unconditional jump instruction. The goto statement comes in handy when a function exits from multiple -locations and some common work such as cleanup has to be done. If there is no +locations and some common work such as cleanup has to be done. If there is no cleanup needed then just return directly. -Choose label names which say what the goto does or why the goto exists. An +Choose label names which say what the goto does or why the goto exists. An example of a good name could be ``out_free_buffer:`` if the goto frees ``buffer``. Avoid using GW-BASIC names like ``err1:`` and ``err2:``, as you would have to renumber them if you ever add or remove exit paths, and they make correctness @@ -513,59 +513,59 @@ The rationale for using gotos is: modifications are prevented - saves the compiler work to optimize redundant code away ;) -.. code-block:: c +.. code-block:: c - int - fun(int a) - { - int result = 0; - char *buffer; + int + fun(int a) + { + int result = 0; + char *buffer; - buffer = kmalloc(SIZE, GFP_KERNEL); - if (!buffer) - return -ENOMEM; + buffer = kmalloc(SIZE, GFP_KERNEL); + if (!buffer) + return -ENOMEM; - if (condition1) { - while (loop1) { - ... + if (condition1) { + while (loop1) { + ... + } + result = 1; + goto out_free_buffer; } - result = 1; - goto out_free_buffer; + ... + out_free_buffer: + kfree(buffer); + return result; } - ... - out_free_buffer: - kfree(buffer); - return result; - } A common type of bug to be aware of is ``one err bugs`` which look like this: -.. code-block:: c +.. code-block:: c - err: - kfree(foo->bar); - kfree(foo); - return ret; + err: + kfree(foo->bar); + kfree(foo); + return ret; -The bug in this code is that on some exit paths ``foo`` is NULL. Normally the +The bug in this code is that on some exit paths ``foo`` is NULL. Normally the fix for this is to split it up into two error labels ``err_free_bar:`` and ``err_free_foo:``: -.. code-block:: c +.. code-block:: c - err_free_bar: - kfree(foo->bar); - err_free_foo: - kfree(foo); - return ret; + err_free_bar: + kfree(foo->bar); + err_free_foo: + kfree(foo); + return ret; Ideally you should simulate errors to test all exit paths. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 8: Commenting -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Comments are good, but there is also a danger of over-commenting. NEVER +Comments are good, but there is also a danger of over-commenting. NEVER try to explain HOW your code works in a comment: it's much better to write the code so that the **working** is obvious, and it's a waste of time to explain badly written code. @@ -579,7 +579,7 @@ ugly), but try to avoid excess. Instead, put the comments at the head of the function, telling people what it does, and possibly WHY it does it. -.. admonition:: Tarantool Style +.. admonition:: Tarantool Style :class: FACT When commenting the Tarantool C API functions, please use Doxygen comment format, @@ -590,7 +590,7 @@ it. Every function, except perhaps a very short and obvious one, should have a comment. A sample function comment may look like below: - .. code-block:: c + .. code-block:: c /** Write all data to a descriptor. * @@ -618,7 +618,7 @@ it. However the difference is vague already, so the rule is simple: out of function = /\**, inside = /\*. - .. code-block:: c + .. code-block:: c /** * Out of function comment, option 1. @@ -644,11 +644,11 @@ it. parameter names are the same as used in the comment, and mean the same. Especially when you change one of them - ensure you changed the other. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 9: You've made a mess of it -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -That's OK, we all do. You've probably been told by your long-time Unix +That's OK, we all do. You've probably been told by your long-time Unix user helper that ``GNU emacs`` automatically formats the C sources for you, and you've noticed that yes, it does do that, but the defaults it uses are less than desirable (in fact, they are worse than random @@ -658,38 +658,38 @@ make a good program). So, you can either get rid of GNU emacs, or change it to use saner values. To do the latter, you can stick the following in your .emacs file: -.. code-block:: none - - (defun c-lineup-arglist-tabs-only (ignored) - "Line up argument lists by tabs, not spaces" - (let* ((anchor (c-langelem-pos c-syntactic-element)) - (column (c-langelem-2nd-pos c-syntactic-element)) - (offset (- (1+ column) anchor)) - (steps (floor offset c-basic-offset))) - (* (max steps 1) - c-basic-offset))) - - (add-hook 'c-mode-common-hook - (lambda () - ;; Add kernel style - (c-add-style - "linux-tabs-only" - '("linux" (c-offsets-alist - (arglist-cont-nonempty - c-lineup-gcc-asm-reg - c-lineup-arglist-tabs-only)))))) - - (add-hook 'c-mode-hook - (lambda () - (let ((filename (buffer-file-name))) - ;; Enable kernel mode for the appropriate files - (when (and filename - (string-match (expand-file-name "~/src/linux-trees") - filename)) - (setq indent-tabs-mode t) - (setq show-trailing-whitespace t) - (c-set-style "linux-tabs-only"))))) - +.. code-block:: none + + (defun c-lineup-arglist-tabs-only (ignored) + "Line up argument lists by tabs, not spaces" + (let* ((anchor (c-langelem-pos c-syntactic-element)) + (column (c-langelem-2nd-pos c-syntactic-element)) + (offset (- (1+ column) anchor)) + (steps (floor offset c-basic-offset))) + (* (max steps 1) + c-basic-offset))) + + (add-hook 'c-mode-common-hook + (lambda () + ;; Add kernel style + (c-add-style + "linux-tabs-only" + '("linux" (c-offsets-alist + (arglist-cont-nonempty + c-lineup-gcc-asm-reg + c-lineup-arglist-tabs-only)))))) + + (add-hook 'c-mode-hook + (lambda () + (let ((filename (buffer-file-name))) + ;; Enable kernel mode for the appropriate files + (when (and filename + (string-match (expand-file-name "~/src/linux-trees") + filename)) + (setq indent-tabs-mode t) + (setq show-trailing-whitespace t) + (c-set-style "linux-tabs-only"))))) + This will make emacs go better with the kernel coding style for C files below ``~/src/linux-trees``. @@ -705,18 +705,18 @@ options ``-kr -i8`` (stands for ``K&R, 8 character indents``), or use ``scripts/Lindent``, which indents in the latest style. ``indent`` has a lot of options, and especially when it comes to comment -re-formatting you may want to take a look at the man page. But +re-formatting you may want to take a look at the man page. But remember: ``indent`` is not a fix for bad programming. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 10: Macros, Enums and RTL -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Names of macros defining constants and labels in enums are capitalized. -.. code-block:: c +.. code-block:: c - #define CONSTANT 0x12345 + #define CONSTANT 0x12345 Enums are preferred when defining several related constants. @@ -727,73 +727,73 @@ Generally, inline functions are preferable to macros resembling functions. Macros with multiple statements should be enclosed in a do - while block: -.. code-block:: c +.. code-block:: c - #define macrofun(a, b, c) \ - do { \ - if (a == 5) \ - do_this(b, c); \ - } while (0) + #define macrofun(a, b, c) \ + do { \ + if (a == 5) \ + do_this(b, c); \ + } while (0) Things to avoid when using macros: -1) macros that affect control flow: +1) macros that affect control flow: - .. code-block:: c + .. code-block:: c - #define FOO(x) \ - do { \ - if (blah(x) < 0) \ - return -EBUGGERED; \ - } while (0) + #define FOO(x) \ + do { \ + if (blah(x) < 0) \ + return -EBUGGERED; \ + } while (0) - is a **very** bad idea. It looks like a function call but exits the ``calling`` - function; don't break the internal parsers of those who will read the code. + is a **very** bad idea. It looks like a function call but exits the ``calling`` + function; 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: +2) macros that depend on having a local variable with a magic name: - .. code-block:: c + .. code-block:: c - #define FOO(val) bar(index, val) + #define FOO(val) bar(index, val) - might look like a good thing, but it's confusing as hell when one reads the - code and it's prone to breakage from seemingly innocent changes. + might look like a good thing, but it's confusing as hell when one reads the + code and it's prone to breakage from seemingly innocent changes. -3) macros with arguments that are used as l-values: FOO(x) = y; will - bite you if somebody e.g. turns FOO into an inline function. +3) macros with arguments that are used as l-values: FOO(x) = y; will + bite you if somebody e.g. turns FOO into an inline function. -4) forgetting about precedence: macros defining constants using expressions - must enclose the expression in parentheses. Beware of similar issues with - macros using parameters. +4) forgetting about precedence: macros defining constants using expressions + must enclose the expression in parentheses. Beware of similar issues with + macros using parameters. - .. code-block:: c + .. code-block:: c - #define CONSTANT 0x4000 - #define CONSTEXP (CONSTANT | 3) + #define CONSTANT 0x4000 + #define CONSTEXP (CONSTANT | 3) -5) namespace collisions when defining local variables in macros resembling - functions: +5) namespace collisions when defining local variables in macros resembling + functions: - .. code-block:: c + .. code-block:: c - #define FOO(x) \ - ({ \ - typeof(x) ret; \ - ret = calc_ret(x); \ - (ret); \ - }) + #define FOO(x) \ + ({ \ + typeof(x) ret; \ + ret = calc_ret(x); \ + (ret); \ + }) - ret is a common name for a local variable - __foo_ret is less likely - to collide with an existing variable. + ret is a common name for a local variable - __foo_ret is less likely + to collide with an existing variable. - The cpp manual deals with macros exhaustively. The gcc internals manual also - covers RTL which is used frequently with assembly language in the kernel. + The cpp manual deals with macros exhaustively. The gcc internals manual also + covers RTL which is used frequently with assembly language in the kernel. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 11: Allocating memory -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -.. admonition:: Tarantool Style +.. admonition:: Tarantool Style :class: FACT Prefer the supplied slab (salloc) and pool (palloc) allocators to malloc()/free() @@ -806,9 +806,9 @@ Chapter 11: Allocating memory memory is also dynamic-load friendly: assuming a plug-in can be dynamically loaded and unloaded multiple times, reload should not lead to a memory leak. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 12: The inline disease -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There appears to be a common misperception that gcc has a magic "make me faster" speedup option called ``inline``. While the use of inlines can be @@ -834,94 +834,94 @@ help, and the maintenance issue of removing the inline when a second user appears outweighs the potential value of the hint that tells gcc to do something it would have done anyway. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 13: Function return values and names -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Functions can return values of many different kinds, and one of the most common is a value indicating whether the function succeeded or -failed. Such a value can be represented as an error-code integer +failed. Such a value can be represented as an error-code integer (-Exxx = failure, 0 = success) or a ``succeeded`` boolean (0 = failure, non-zero = success). Mixing up these two sorts of representations is a fertile source of -difficult-to-find bugs. If the C language included a strong distinction +difficult-to-find bugs. If the C language included a strong distinction between integers and booleans then the compiler would find these mistakes -for us... but it doesn't. To help prevent such bugs, always follow this +for us... but it doesn't. To help prevent such bugs, always follow this convention: If the name of a function is an action or an imperative command, - the function should return an error-code integer. If the name + the function should return an error-code integer. If the name is a predicate, the function should return a "succeeded" boolean. For example, ``add work`` is a command, and the add_work() function returns 0 -for success or -EBUSY for failure. In the same way, ``PCI device present`` is +for success or -EBUSY for failure. In the same way, ``PCI device present`` is a predicate, and the pci_dev_present() function returns 1 if it succeeds in finding a matching device or 0 if it doesn't. All EXPORTed functions must respect this convention, and so should all -public functions. Private (static) functions need not, but it is +public functions. Private (static) functions need not, but it is recommended that they do. Functions whose return value is the actual result of a computation, rather than an indication of whether the computation succeeded, are not subject to -this rule. Generally they indicate failure by returning some out-of-range -result. Typical examples would be functions that return pointers; they use +this rule. Generally they indicate failure by returning some out-of-range +result. Typical examples would be functions that return pointers; they use NULL or the ERR_PTR mechanism to report failure. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 14: Editor modelines and other cruft -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Some editors can interpret configuration information embedded in source files, -indicated with special markers. For example, emacs interprets lines marked +indicated with special markers. For example, emacs interprets lines marked like this: -.. code-block:: c +.. code-block:: c - -*- mode: c -*- + -*- mode: c -*- Or like this: -.. code-block:: c +.. code-block:: c - /* - Local Variables: - compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" - End: - */ + /* + Local Variables: + compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" + End: + */ Vim interprets markers that look like this: -.. code-block:: c +.. code-block:: c - /* vim:set sw=8 noet */ + /* vim:set sw=8 noet */ -Do not include any of these in source files. People have their own personal -editor configurations, and your source files should not override them. This -includes markers for indentation and mode configuration. People may use their +Do not include any of these in source files. People have their own personal +editor configurations, and your source files should not override them. This +includes markers for indentation and mode configuration. People may use their own custom mode, or may have some other magic method for making indentation work correctly. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 15: Inline assembly -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In architecture-specific code, you may need to use inline assembly to interface -with 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 can +with 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 can and should poke hardware from C when possible. Consider writing simple helper functions that wrap common bits of inline -assembly, rather than repeatedly writing them with slight variations. Remember +assembly, rather than repeatedly writing them with slight variations. Remember that inline assembly can use C parameters. Large, non-trivial assembly functions should go in .S files, with corresponding -C prototypes defined in C header files. The C prototypes for assembly +C prototypes defined in C header files. The C prototypes for assembly functions should use ``asmlinkage``. You may need to mark your asm statement as volatile, to prevent GCC from -removing it if GCC doesn't notice any side effects. You don't always need to +removing it if GCC doesn't notice any side effects. You don't always need to do so, though, and doing so unnecessarily can limit optimization. When writing a single inline assembly statement containing multiple @@ -929,67 +929,67 @@ instructions, put each instruction on a separate line in a separate quoted string, and end each string except the last with \n\t to properly indent the next instruction in the assembly output: -.. code-block:: c +.. code-block:: c - asm ("magic %reg1, #42\n\t" - "more_magic %reg2, %reg3" - : /* outputs */ : /* inputs */ : /* clobbers */); + asm ("magic %reg1, #42\n\t" + "more_magic %reg2, %reg3" + : /* outputs */ : /* inputs */ : /* clobbers */); -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 16: Conditional Compilation -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Wherever possible, don't use preprocessor conditionals (#if, #ifdef) in .c -files; doing so makes code harder to read and logic harder to follow. Instead, +files; 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 .c files, providing no-op stub versions in the #else case, and then call those -functions unconditionally from .c files. The compiler will avoid generating +functions unconditionally from .c files. The compiler will avoid generating any code for the stub calls, producing identical results, but the logic will remain easy to follow. Prefer to compile out entire functions, rather than portions of functions or -portions of expressions. Rather than putting an ifdef in an expression, factor +portions of expressions. Rather than putting an ifdef in an expression, factor out part or all of the expression into a separate helper function and apply the conditional to that function. If you have a function or variable which may potentially go unused in a particular configuration, and the compiler would warn about its definition going unused, mark the definition as __maybe_unused rather than wrapping it in -a preprocessor conditional. (However, if a function or variable *always* goes +a preprocessor conditional. (However, if a function or variable *always* goes unused, delete it.) Within code, where possible, use the IS_ENABLED macro to convert a Kconfig symbol into a C boolean expression, and use it in a normal C conditional: -.. code-block:: c +.. code-block:: c - if (IS_ENABLED(CONFIG_SOMETHING)) { - ... - } + if (IS_ENABLED(CONFIG_SOMETHING)) { + ... + } The compiler will constant-fold the conditional away, and include or exclude the block of code just as with an #ifdef, so this will not add any runtime overhead. However, this approach still allows the C compiler to see the code inside the block, and check it for correctness (syntax, types, symbol -references, etc). Thus, you still have to use an #ifdef if the code inside the +references, etc). Thus, you still have to use an #ifdef if the code inside the block 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 conditional -expression used. For instance: +expression used. For instance: -.. code-block:: c +.. code-block:: c - #ifdef CONFIG_SOMETHING - ... - #endif /* CONFIG_SOMETHING */ + #ifdef CONFIG_SOMETHING + ... + #endif /* CONFIG_SOMETHING */ -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 17: Header files -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -.. admonition:: Tarantool Style +.. admonition:: Tarantool Style :class: FACT Use header guards. Put the header guard in the first line in the header, @@ -1000,47 +1000,47 @@ Chapter 17: Header files all other headers, to ensure that the header is self- sufficient. Header "header.h" is self-sufficient if the following compiles without errors: - .. code-block:: c + .. code-block:: c #include "header.h" -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 18: Other -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -.. admonition:: Tarantool Style +.. admonition:: Tarantool Style :class: FACT - * We don't apply ``!`` operator to non-boolean values. It means, to check - if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, - you use ``!= NULL``. The same for ``==``. + * We don't apply ``!`` operator to non-boolean values. It means, to check + if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, + you use ``!= NULL``. The same for ``==``. - * Select GNU C99 extensions are acceptable. It's OK to mix declarations and - statements, use true and false. + * Select GNU C99 extensions are acceptable. It's OK to mix declarations and + statements, use true and false. - * The not-so-current list of all GCC C extensions can be found at: - http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html + * The not-so-current list of all GCC C extensions can be found at: + http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Appendix I: References -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -* `The C Programming Language, Second Edition `_ - by Brian W. Kernighan and Dennis M. Ritchie. |br| - Prentice Hall, Inc., 1988. |br| - ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback). +* `The C Programming Language, Second Edition `_ + by 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 Programming `_ - by Brian W. Kernighan and Rob Pike. |br| - Addison-Wesley, Inc., 1999. |br| - ISBN 0-201-61586-X. +* `The Practice of Programming `_ + by 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** +* `GNU manuals `_ - where in compliance with K&R + and this text - for **cpp**, **gcc**, **gcc internals** and **indent** -* `WG14 International standardization workgroup for the programming - language C `_ +* `WG14 International standardization workgroup for the programming + language C `_ -* `Kernel CodingStyle, by greg@kroah.com at OLS 2002 - `_ +* `Kernel CodingStyle, by greg@kroah.com at OLS 2002 + `_ From bda860444533e061860fe42be657249b29cd4d52 Mon Sep 17 00:00:00 2001 From: Natalia Ogoreltseva Date: Mon, 15 Mar 2021 14:57:26 +0300 Subject: [PATCH 3/8] eliminate errors from review that look simple --- doc/dev_guide/c_style_guide_new.rst | 81 +++++++++-------------------- 1 file changed, 24 insertions(+), 57 deletions(-) diff --git a/doc/dev_guide/c_style_guide_new.rst b/doc/dev_guide/c_style_guide_new.rst index d87e260aa4..830d65e05b 100644 --- a/doc/dev_guide/c_style_guide_new.rst +++ b/doc/dev_guide/c_style_guide_new.rst @@ -7,39 +7,26 @@ default branch (currently ``master``). Our git repository is hosted on GitHub, and can be checked out with ``git clone git://github.com/tarantool/tarantool.git`` (anonymous read-only access). -If you have any questions about Tarantool internals, please post them on the -`developer discussion list `_ -or on `StackOverflow `_. -Additionally, some engineers are always present on #tarantool channel on -irc.freenode.net. +If you have any questions about Tarantool internals, please post them on +`StackOverflow `_ or +ask Tarantool developers directly in `telegram `_. **General guidelines** -The project's coding style is based on the `Linux kernel coding style +The project's coding style is inspired by the `Linux kernel coding style `_. However, we have some additional guidelines, either unique to Tarantool or -deviating from the Kernel guidelines. Below we cite the Linux kernel +deviating from the Kernel guidelines. Below we rewrite the Linux kernel coding style noting Tarantool's style features. -We don't cite chapters 10 "Kconfig configuration files", 11 "Data structures", -13 "Printing kernel messages", and 17 "Don't re-invent the kernel macros" since -they are specific to Linux kernel programming environment. - -------------------------------------------------------------------------------- Linux kernel coding style -------------------------------------------------------------------------------- This is a short document describing the preferred coding style for the -linux kernel. Coding style is very personal, and I won't **force** my -views on anybody, but this is what goes for anything that I have to be -able to maintain, and I'd prefer it for most other things too. Please -at least consider the points made here. - -First off, I'd suggest printing out a copy of the GNU coding standards, -and NOT read it. Burn them, it's a great symbolic gesture. - -Anyway, here goes: +Tarantool developers and contributors. We insist on following these rules +in order to make our code consistent and understandable to any developer. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 1: Indentation @@ -123,8 +110,7 @@ Statements longer than 80 columns will be broken into sensible chunks, unless exceeding 80 columns significantly increases readability and does not hide information. Descendants are always substantially shorter than the parent and are placed substantially to the right. The same applies to function headers -with a long argument list. However, never break user-visible strings such as -printk messages, because that breaks the ability to grep for them. +with a long argument list. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 3: Placing Braces and Spaces @@ -211,7 +197,8 @@ Do not unnecessarily use braces where a single statement will do. if (condition) action(); - and + +and .. code-block:: none @@ -236,7 +223,7 @@ statement; in the latter case use braces in both branches: Chapter 3.1: Spaces ******************************************************************************** -Tarantool style for use of spaces depends (mostly) on +Like Linux kernel, Tarantool style for use of spaces depends (mostly) on function-versus-keyword usage. Use a space after (most) keywords. The notable exceptions are sizeof, typeof, alignof, and __attribute__, which look somewhat like functions (and are usually used with parentheses in Linux, @@ -291,6 +278,12 @@ no space after the prefix increment & decrement unary operators:: and no space around the ``.`` and ``->`` structure member operators. +.. admonition:: Tarantool Style + :class: FACT + + Do not split a cast operator from its argument with a whitespace, + e.g. ``(ssize_t)inj->iparam``. + Do not leave trailing whitespace at the ends of lines. Some editors with ``smart`` indentation will insert whitespace at the beginning of new lines as appropriate, so you can start typing the next line of code right away. @@ -303,12 +296,6 @@ optionally strip the trailing whitespace for you; however, if applying a series of patches, this may make later patches in the series fail by changing their context lines. -.. admonition:: Tarantool Style - :class: FACT - - Do not split a cast operator from its argument with a whitespace, - e.g. ``(ssize_t)inj->iparam``. - ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 4: Naming ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ @@ -423,13 +410,6 @@ useful only for: When editing existing code which already uses one or the other set of types, you should conform to the existing choices in that code. -#. Types safe for use in userspace. - - In certain structures which are visible to userspace, we cannot - require C99 types and cannot use the ``u32`` form above. Thus, we - use __u32 and similar types in all structures which are shared - with userspace. - Maybe there are other cases too, but the rule should basically be to NEVER EVER use a typedef unless you can clearly match one of those rules. @@ -465,19 +445,6 @@ generally easily keep track of about 7 different things, anything more and it gets confused. You know you're brilliant, but maybe you'd like to understand what you did 2 weeks from now. -In source files, separate functions with one blank line. If the function is -exported, the **EXPORT** macro for it should follow immediately after the -closing function brace line. E.g.: - -.. code-block:: c - - int - system_is_up(void) - { - return system_state == SYSTEM_RUNNING; - } - EXPORT_SYMBOL(system_is_up); - In function prototypes, include parameter names with their data types. Although this is not required by the C language, it is preferred in Linux because it is a simple way to add valuable information for the reader. @@ -592,23 +559,23 @@ it. .. code-block:: c - /** Write all data to a descriptor. + /** + * Write all data to a descriptor. * * This function is equivalent to 'write', except it would ensure * that all data is written to the file unless a non-ignorable * error occurs. * * @retval 0 Success - * * @retval 1 An error occurred (not EINTR) - * / + */ static int write_all(int fd, void \*data, size_t len); - It's also important to comment data, whether they are basic types or derived - types. To this end, use just one data declaration per line (no commas for - multiple data declarations). This leaves you room for a small comment on each - item, explaining its use. + It's also important to comment data types, whether they are basic types or + derived ones. To this end, use just one data declaration per line (no commas + for multiple data declarations). This leaves you room for a small comment on + each item, explaining its use. Public structures and important structure members should be commented as well. From d7e7391754795cfcc59ea9084fb1e51e22212216 Mon Sep 17 00:00:00 2001 From: Natalia Ogoreltseva Date: Tue, 16 Mar 2021 12:24:16 +0300 Subject: [PATCH 4/8] add code formatting where needed --- doc/dev_guide/c_style_guide_new.rst | 34 ++++++++++++++--------------- 1 file changed, 17 insertions(+), 17 deletions(-) diff --git a/doc/dev_guide/c_style_guide_new.rst b/doc/dev_guide/c_style_guide_new.rst index 830d65e05b..3a3ec0cf44 100644 --- a/doc/dev_guide/c_style_guide_new.rst +++ b/doc/dev_guide/c_style_guide_new.rst @@ -225,8 +225,8 @@ Chapter 3.1: Spaces Like Linux kernel, Tarantool style for use of spaces depends (mostly) on function-versus-keyword usage. Use a space after (most) keywords. The -notable exceptions are sizeof, typeof, alignof, and __attribute__, which look -somewhat like functions (and are usually used with parentheses in Linux, +notable exceptions are ``sizeof``, ``typeof``, ``alignof``, and ``__attribute__``, +which look somewhat like functions (and are usually used with parentheses in Linux, although they are not required in the language, as in: ``sizeof info`` after ``struct fileinfo info;`` is declared). @@ -236,7 +236,7 @@ So use a space after these keywords: if, switch, case, for, do, while -but not with sizeof, typeof, alignof, or __attribute__. E.g., +but not with ``sizeof``, ``typeof``, ``alignof``, or ``__attribute__``. E.g., .. code-block:: c @@ -581,9 +581,9 @@ it. In C comments out of functions and inside of functions should be different in how they are started. Everything else is wrong. Below are correct examples. - /** comes for documentation comments, /* for local not documented comments. + ``/**`` comes for documentation comments, ``/*`` for local not documented comments. However the difference is vague already, so the rule is simple: - out of function = /\**, inside = /\*. + out of function use ``/\**``, inside use ``/\*``. .. code-block:: c @@ -726,7 +726,7 @@ Things to avoid when using macros: might look like a good thing, but it's confusing as hell when one reads the code and it's prone to breakage from seemingly innocent changes. -3) macros with arguments that are used as l-values: FOO(x) = y; will +3) macros with arguments that are used as l-values: ``FOO(x) = y;`` will bite you if somebody e.g. turns FOO into an inline function. 4) forgetting about precedence: macros defining constants using expressions @@ -750,7 +750,7 @@ Things to avoid when using macros: (ret); \ }) - ret is a common name for a local variable - __foo_ret is less likely + ret is a common name for a local variable - ``__foo_ret`` is less likely to collide with an existing variable. The cpp manual deals with macros exhaustively. The gcc internals manual also @@ -768,7 +768,7 @@ Chapter 11: Allocating memory malloc()/free() can lead to memory fragmentation and should therefore be avoided. Always free all allocated memory, even allocated at start-up. We aim at being - valgrind leak-check clean, and in most cases it's just as easy to free() the + valgrind leak-check clean, and in most cases it's just as easy to ``free()`` the allocated memory as it is to write a valgrind suppression. Freeing all allocated memory is also dynamic-load friendly: assuming a plug-in can be dynamically loaded and unloaded multiple times, reload should not lead to a memory leak. @@ -893,7 +893,7 @@ do so, though, and doing so unnecessarily can limit optimization. When writing a single inline assembly statement containing multiple instructions, put each instruction on a separate line in a separate quoted -string, and end each string except the last with \n\t to properly indent the +string, and end each string except the last with ``\n\t`` to properly indent the next instruction in the assembly output: .. code-block:: c @@ -906,8 +906,8 @@ next instruction in the assembly output: Chapter 16: Conditional Compilation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Wherever possible, don't use preprocessor conditionals (#if, #ifdef) in .c -files; doing so makes code harder to read and logic harder to follow. Instead, +Wherever possible, don't use preprocessor conditionals (``#if``, ``#ifdef``) in +.c files; 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 .c files, providing no-op stub versions in the #else case, and then call those functions unconditionally from .c files. The compiler will avoid generating @@ -915,9 +915,9 @@ any code for the stub calls, producing identical results, but the logic will remain easy to follow. Prefer to compile out entire functions, rather than portions of functions or -portions of expressions. Rather than putting an ifdef in an expression, factor -out part or all of the expression into a separate helper function and apply the -conditional to that function. +portions of expressions. Rather than putting an ``#ifdef`` in an expression, +factor out part or all of the expression into a separate helper function and +apply the conditional to that function. If you have a function or variable which may potentially go unused in a particular configuration, and the compiler would warn about its definition @@ -939,10 +939,10 @@ the block of code just as with an #ifdef, so this will not add any runtime overhead. However, this approach still allows the C compiler to see the code inside the block, and check it for correctness (syntax, types, symbol -references, etc). Thus, you still have to use an #ifdef if the code inside the -block references symbols that will not exist if the condition is not met. +references, etc). Thus, you still have to use an ``#ifdef`` if the code inside +the block 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), +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 conditional expression used. For instance: From 77a7b8c3ba86beee23ca06980e104b00cf8692c5 Mon Sep 17 00:00:00 2001 From: Natalia Ogoreltseva Date: Tue, 16 Mar 2021 13:16:02 +0300 Subject: [PATCH 5/8] update guide --- doc/dev_guide/c_style_guide_new.rst | 339 ++++++++++------------------ 1 file changed, 123 insertions(+), 216 deletions(-) diff --git a/doc/dev_guide/c_style_guide_new.rst b/doc/dev_guide/c_style_guide_new.rst index 3a3ec0cf44..3d89ac4f9c 100644 --- a/doc/dev_guide/c_style_guide_new.rst +++ b/doc/dev_guide/c_style_guide_new.rst @@ -18,14 +18,14 @@ The project's coding style is inspired by the `Linux kernel coding style However, we have some additional guidelines, either unique to Tarantool or deviating from the Kernel guidelines. Below we rewrite the Linux kernel -coding style noting Tarantool's style features. +coding style according to the Tarantool's style features. -------------------------------------------------------------------------------- - Linux kernel coding style + Tarantool coding style -------------------------------------------------------------------------------- This is a short document describing the preferred coding style for the -Tarantool developers and contributors. We insist on following these rules +Tarantool developers and contributors. We **insist** on following these rules in order to make our code consistent and understandable to any developer. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ @@ -99,12 +99,7 @@ Coding style is all about readability and maintainability using commonly available tools. The limit on the length of lines is 80 columns and this is a strongly -preferred limit. - -.. admonition:: Tarantool Style - :class: FACT - - As for comments, the same limit of 80 columns is applied. +preferred limit. As for comments, the same limit of 80 columns is applied. Statements longer than 80 columns will be broken into sensible chunks, unless exceeding 80 columns significantly increases readability and does not hide @@ -200,7 +195,7 @@ Do not unnecessarily use braces where a single statement will do. and -.. code-block:: none +.. code-block:: c if (condition) do_this(); @@ -278,11 +273,8 @@ no space after the prefix increment & decrement unary operators:: and no space around the ``.`` and ``->`` structure member operators. -.. admonition:: Tarantool Style - :class: FACT - - Do not split a cast operator from its argument with a whitespace, - e.g. ``(ssize_t)inj->iparam``. +Do not split a cast operator from its argument with a whitespace, +e.g. ``(ssize_t)inj->iparam``. Do not leave trailing whitespace at the ends of lines. Some editors with ``smart`` indentation will insert whitespace at the beginning of new lines as @@ -330,16 +322,13 @@ If you are afraid to mix up your local variable names, you have another problem, which is called the function-growth-hormone-imbalance syndrome. See chapter 6 (Functions). -.. admonition:: Tarantool Style - :class: FACT +For function naming we have a convention is to use: - For function naming we have a convention is to use: - - * ``new``/``delete`` for functions which - allocate + initialize and destroy + deallocate an object, - * ``create``/``destroy`` for functions which initialize/destroy an object - but do not handle memory management, - * ``init``/``free`` for functions which initialize/destroy libraries and subsystems. +* ``new``/``delete`` for functions which + allocate + initialize and destroy + deallocate an object, +* ``create``/``destroy`` for functions which initialize/destroy an object + but do not handle memory management, +* ``init``/``free`` for functions which initialize/destroy libraries and subsystems. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 5: Typedefs @@ -449,11 +438,7 @@ In function prototypes, include parameter names with their data types. Although this is not required by the C language, it is preferred in Linux because it is a simple way to add valuable information for the reader. -.. admonition:: Tarantool Style - :class: FACT - - Note that in Tarantool, we place the function return type on the - line before the name and signature. +Note that we place the function return type on the line before the name and signature. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 7: Centralized exiting of functions @@ -546,137 +531,70 @@ ugly), but try to avoid excess. Instead, put the comments at the head of the function, telling people what it does, and possibly WHY it does it. -.. admonition:: Tarantool Style - :class: FACT +When commenting the Tarantool C API functions, please use Doxygen comment format, +Javadoc flavor, i.e. `@tag` rather than `\\tag`. +The main tags in use are ``@param``, ``@retval``, ``@return``, ``@see``, +``@note`` and ``@todo``. - When commenting the Tarantool C API functions, please use Doxygen comment format, - Javadoc flavor, i.e. `@tag` rather than `\\tag`. - The main tags in use are ``@param``, ``@retval``, ``@return``, ``@see``, - ``@note`` and ``@todo``. +Every function, except perhaps a very short and obvious one, should have a +comment. A sample function comment may look like below: - Every function, except perhaps a very short and obvious one, should have a - comment. A sample function comment may look like below: - - .. code-block:: c +.. code-block:: c - /** - * Write all data to a descriptor. - * - * This function is equivalent to 'write', except it would ensure - * that all data is written to the file unless a non-ignorable - * error occurs. - * - * @retval 0 Success - * @retval 1 An error occurred (not EINTR) - */ - static int - write_all(int fd, void \*data, size_t len); + /** + * Write all data to a descriptor. + * + * This function is equivalent to 'write', except it would ensure + * that all data is written to the file unless a non-ignorable + * error occurs. + * + * @retval 0 Success + * @retval 1 An error occurred (not EINTR) + */ + static int + write_all(int fd, void \*data, size_t len); + +It's also important to comment data types, whether they are basic types or +derived ones. To this end, use just one data declaration per line (no commas +for multiple data declarations). This leaves you room for a small comment on +each item, explaining its use. + +Public structures and important structure members should be commented as well. + +In C comments out of functions and inside of functions should be different in +how they are started. Everything else is wrong. Below are correct examples. +``/**`` comes for documentation comments, ``/*`` for local not documented comments. +However the difference is vague already, so the rule is simple: +out of function use ``/**``, inside use ``/*``. - It's also important to comment data types, whether they are basic types or - derived ones. To this end, use just one data declaration per line (no commas - for multiple data declarations). This leaves you room for a small comment on - each item, explaining its use. +.. code-block:: c - Public structures and important structure members should be commented as well. + /** + * Out of function comment, option 1. + */ - In C comments out of functions and inside of functions should be different in - how they are started. Everything else is wrong. Below are correct examples. - ``/**`` comes for documentation comments, ``/*`` for local not documented comments. - However the difference is vague already, so the rule is simple: - out of function use ``/\**``, inside use ``/\*``. + /** Out of function comment, option 2. */ - .. code-block:: c + int + function() + { + /* Comment inside function, option 1. */ - /** - * Out of function comment, option 1. + /* + * Comment inside function, option 2. */ + } - /** Out of function comment, option 2. */ - - int - function() - { - /* Comment inside function, option 1. */ +If a function has declaration and implementation separated, the function comment +should be for the declaration. Usually in the header file. Don't duplicate the +comment. - /* - * Comment inside function, option 2. - */ - } +A comment and the function signature should be synchronized. Double-check if the +parameter names are the same as used in the comment, and mean the same. +Especially when you change one of them - ensure you changed the other. - If a function has declaration and implementation separated, the function comment - should be for the declaration. Usually in the header file. Don't duplicate the - comment. - - A comment and the function signature should be synchronized. Double-check if the - parameter names are the same as used in the comment, and mean the same. - Especially when you change one of them - ensure you changed the other. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 9: You've made a mess of it -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -That's OK, we all do. You've probably been told by your long-time Unix -user helper that ``GNU emacs`` automatically formats the C sources for -you, and you've noticed that yes, it does do that, but the defaults it -uses are less than desirable (in fact, they are worse than random -typing - an infinite number of monkeys typing into GNU emacs would never -make a good program). - -So, you can either get rid of GNU emacs, or change it to use saner -values. To do the latter, you can stick the following in your .emacs file: - -.. code-block:: none - - (defun c-lineup-arglist-tabs-only (ignored) - "Line up argument lists by tabs, not spaces" - (let* ((anchor (c-langelem-pos c-syntactic-element)) - (column (c-langelem-2nd-pos c-syntactic-element)) - (offset (- (1+ column) anchor)) - (steps (floor offset c-basic-offset))) - (* (max steps 1) - c-basic-offset))) - - (add-hook 'c-mode-common-hook - (lambda () - ;; Add kernel style - (c-add-style - "linux-tabs-only" - '("linux" (c-offsets-alist - (arglist-cont-nonempty - c-lineup-gcc-asm-reg - c-lineup-arglist-tabs-only)))))) - - (add-hook 'c-mode-hook - (lambda () - (let ((filename (buffer-file-name))) - ;; Enable kernel mode for the appropriate files - (when (and filename - (string-match (expand-file-name "~/src/linux-trees") - filename)) - (setq indent-tabs-mode t) - (setq show-trailing-whitespace t) - (c-set-style "linux-tabs-only"))))) - -This will make emacs go better with the kernel coding style for C -files below ``~/src/linux-trees``. - -But even if you fail in getting emacs to do sane formatting, not -everything is lost: use ``indent``. - -Now, again, GNU indent has the same brain-dead settings that GNU emacs -has, 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 indent -recognize the authority of K&R (the GNU people aren't evil, they are -just severely misguided in this matter), so you just give indent the -options ``-kr -i8`` (stands for ``K&R, 8 character indents``), or use -``scripts/Lindent``, which indents in the latest style. - -``indent`` has a lot of options, and especially when it comes to comment -re-formatting you may want to take a look at the man page. But -remember: ``indent`` is not a fix for bad programming. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 10: Macros, Enums and RTL +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 9: Macros, Enums and RTL ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Names of macros defining constants and labels in enums are capitalized. @@ -757,32 +675,29 @@ Things to avoid when using macros: covers RTL which is used frequently with assembly language in the kernel. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 11: Allocating memory +Chapter 10: Allocating memory ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -.. admonition:: Tarantool Style - :class: FACT +Prefer specialized allocators like ``region``, ``mempool``, ``smalloc`` to +``malloc()/free()`` for any performance-intensive or large memory allocations. +Repetitive use of ``malloc()``/``free()`` can lead to memory fragmentation +and should therefore be avoided. - Prefer the supplied slab (salloc) and pool (palloc) allocators to malloc()/free() - for any performance-intensive or large memory allocations. Repetitive use of - malloc()/free() can lead to memory fragmentation and should therefore be avoided. - - Always free all allocated memory, even allocated at start-up. We aim at being - valgrind leak-check clean, and in most cases it's just as easy to ``free()`` the - allocated memory as it is to write a valgrind suppression. Freeing all allocated - memory is also dynamic-load friendly: assuming a plug-in can be dynamically - loaded and unloaded multiple times, reload should not lead to a memory leak. +Always free all allocated memory, even allocated at start-up. We aim at being +valgrind leak-check clean, and in most cases it's just as easy to ``free()`` the +allocated memory as it is to write a valgrind suppression. Freeing all allocated +memory is also dynamic-load friendly: assuming a plug-in can be dynamically +loaded and unloaded multiple times, reload should not lead to a memory leak. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 12: The inline disease +Chapter 11: The inline disease ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There appears to be a common misperception that gcc has a magic "make me faster" speedup option called ``inline``. While the use of inlines can be -appropriate (for example as a means of replacing macros, see Chapter 12), it -very often is not. Abundant use of the inline keyword leads to a much bigger -kernel, which in turn slows the system as a whole down, due to a bigger -icache footprint for the CPU and simply because there is less memory +appropriate, it very often is not. Abundant use of the inline keyword leads to +a much bigger kernel, which in turn slows the system as a whole down, due to a +bigger icache footprint for the CPU and simply because there is less memory available for the pagecache. Just think about it; a pagecache miss causes a disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles that can go into these 5 milliseconds. @@ -802,33 +717,16 @@ appears outweighs the potential value of the hint that tells gcc to do something it would have done anyway. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 13: Function return values and names +Chapter 12: Function return values and names ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Functions can return values of many different kinds, and one of the most common is a value indicating whether the function succeeded or -failed. Such a value can be represented as an error-code integer -(-Exxx = failure, 0 = success) or a ``succeeded`` boolean (0 = failure, -non-zero = success). - -Mixing up these two sorts of representations is a fertile source of -difficult-to-find bugs. If the C language included a strong distinction -between integers and booleans then the compiler would find these mistakes -for us... but it doesn't. To help prevent such bugs, always follow this -convention: +failed. - If the name of a function is an action or an imperative command, - the function should return an error-code integer. If the name - is a predicate, the function should return a "succeeded" boolean. - -For example, ``add work`` is a command, and the add_work() function returns 0 -for success or -EBUSY for failure. In the same way, ``PCI device present`` is -a predicate, and the pci_dev_present() function returns 1 if it succeeds in -finding a matching device or 0 if it doesn't. - -All EXPORTed functions must respect this convention, and so should all -public functions. Private (static) functions need not, but it is -recommended that they do. +In 99.99999% of all cases in Tarantool we return 0 on success, non-zero on error +(-1 usually). Errors are saved into a diagnostics area which is global per fiber. +We never return error codes as a result of a function. Functions whose return value is the actual result of a computation, rather than an indication of whether the computation succeeded, are not subject to @@ -837,7 +735,7 @@ result. Typical examples would be functions that return pointers; they use NULL or the ERR_PTR mechanism to report failure. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 14: Editor modelines and other cruft +Chapter 13: Editor modelines and other cruft ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Some editors can interpret configuration information embedded in source files, @@ -871,7 +769,7 @@ own custom mode, or may have some other magic method for making indentation work correctly. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 15: Inline assembly +Chapter 14: Inline assembly ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In architecture-specific code, you may need to use inline assembly to interface @@ -903,7 +801,7 @@ next instruction in the assembly output: : /* outputs */ : /* inputs */ : /* clobbers */); ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 16: Conditional Compilation +Chapter 15: Conditional Compilation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Wherever possible, don't use preprocessor conditionals (``#if``, ``#ifdef``) in @@ -921,9 +819,8 @@ apply the conditional to that function. If you have a function or variable which may potentially go unused in a particular configuration, and the compiler would warn about its definition -going unused, mark the definition as __maybe_unused rather than wrapping it in -a preprocessor conditional. (However, if a function or variable *always* goes -unused, delete it.) +going unused, do not compile it and use #if for this. +(However, if a function or variable *always* goes unused, delete it.) Within code, where possible, use the IS_ENABLED macro to convert a Kconfig symbol into a C boolean expression, and use it in a normal C conditional: @@ -953,40 +850,50 @@ expression used. For instance: #endif /* CONFIG_SOMETHING */ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 17: Header files +Chapter 16: Header files ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -.. admonition:: Tarantool Style - :class: FACT +Use ``#pragma once`` in the headers. As the header guards we refer to this +construction: - Use header guards. Put the header guard in the first line in the header, - before the copyright or declarations. Use all-uppercase name for the header - guard. Derive the header guard name from the file name, and append _INCLUDED - to get a macro name. For example, core/log_io.h -> CORE_LOG_IO_H_INCLUDED. In - ``.c`` (implementation) file, include the respective declaration header before - all other headers, to ensure that the header is self- sufficient. Header - "header.h" is self-sufficient if the following compiles without errors: +.. code-block:: c - .. code-block:: c + #ifndef THE_HEADER_IS_INCLUDED + #define THE_HEADER_IS_INCLUDED - #include "header.h" + // ... the header code ... + + #endif // THE_HEADER_IS_INCLUDED + +It works fine, but the guard name ``THE_HEADER_IS_INCLUDED`` tends to +become outdated when the file is moved or renamed. This is especially +painful with multiple files having the same name in the project, but +different path. For instance, we have 3 ``error.h`` files, which means for +each of them we need to invent a new header guard name, and not forget to +update them if the files are moved or renamed. + +For that reason we use ``#pragma once`` in all the new code, which shortens +the header file down to this: + +.. code-block:: c + + #pragma once + + // ... header code ... ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 18: Other +Chapter 17: Other ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -.. admonition:: Tarantool Style - :class: FACT - - * We don't apply ``!`` operator to non-boolean values. It means, to check - if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, - you use ``!= NULL``. The same for ``==``. +* We don't apply ``!`` operator to non-boolean values. It means, to check + if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, + you use ``!= NULL``. The same for ``==``. - * Select GNU C99 extensions are acceptable. It's OK to mix declarations and - statements, use true and false. +* Select GNU C99 extensions are acceptable. It's OK to mix declarations and + statements, use true and false. - * The not-so-current list of all GCC C extensions can be found at: - http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html +* The not-so-current list of all GCC C extensions can be found at: + http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Appendix I: References From 27d69edcadf9724b422cbdafbedc2dfa12cda0f2 Mon Sep 17 00:00:00 2001 From: Natalia Ogoreltseva Date: Tue, 30 Mar 2021 15:32:50 +0300 Subject: [PATCH 6/8] apply changes after review --- doc/dev_guide/c_style_guide_new.rst | 88 +++++------------------------ 1 file changed, 14 insertions(+), 74 deletions(-) diff --git a/doc/dev_guide/c_style_guide_new.rst b/doc/dev_guide/c_style_guide_new.rst index 3d89ac4f9c..b24fd17799 100644 --- a/doc/dev_guide/c_style_guide_new.rst +++ b/doc/dev_guide/c_style_guide_new.rst @@ -32,8 +32,8 @@ in order to make our code consistent and understandable to any developer. Chapter 1: Indentation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Tabs are 8 characters, and thus indentations are -also 8 characters. There are heretic movements that try to make indentations +Tabs are 8 characters (8-width tabs, not 8 whitespaces), 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 to be 3. @@ -83,8 +83,7 @@ something to hide: if (condition) do_this; do_something_everytime; -Don't put multiple assignments on a single line either. Kernel coding style -is super simple. Avoid tricky expressions. +Don't put multiple assignments on a single line either. Avoid tricky expressions. Outside of comments and documentation, spaces are never used for indentation, and the above example is deliberately broken. @@ -353,7 +352,7 @@ you can actually tell what ``a`` is. Lots of people think that typedefs ``help readability``. Not so. They are useful only for: -#. totally opaque objects (where the typedef is actively used to **hide** +#. Totally opaque objects (where the typedef is actively used to **hide** what the object is). Example: ``pte_t`` etc. opaque objects that you can only access using @@ -381,7 +380,7 @@ useful only for: might be an ``unsigned int`` and under other configurations might be ``unsigned long``, then by all means go ahead and use a typedef. -#. when you use sparse to literally create a **new** type for +#. When you use sparse to literally create a **new** type for type-checking. #. New types which are identical to standard C99 types, in certain @@ -391,11 +390,6 @@ useful only for: brain to become accustomed to the standard types like ``uint32_t``, some people object to their use anyway. - Therefore, the Linux-specific ``u8/u16/u32/u64`` types and their - signed equivalents which are identical to standard types are - permitted -- although they are not mandatory in new code of your - own. - When editing existing code which already uses one or the other set of types, you should conform to the existing choices in that code. @@ -435,7 +429,7 @@ and it gets confused. You know you're brilliant, but maybe you'd like to understand what you did 2 weeks from now. In function prototypes, include parameter names with their data types. -Although this is not required by the C language, it is preferred in Linux +Although this is not required by the C language, it is preferred in Tarantool because it is a simple way to add valuable information for the reader. Note that we place the function return type on the line before the name and signature. @@ -549,10 +543,10 @@ comment. A sample function comment may look like below: * error occurs. * * @retval 0 Success - * @retval 1 An error occurred (not EINTR) + * @retval 1 An error occurred (not EINTR) */ static int - write_all(int fd, void \*data, size_t len); + write_all(int fd, void *data, size_t len); It's also important to comment data types, whether they are basic types or derived ones. To this end, use just one data declaration per line (no commas @@ -671,9 +665,6 @@ Things to avoid when using macros: ret is a common name for a local variable - ``__foo_ret`` is less likely to collide with an existing variable. - The cpp manual deals with macros exhaustively. The gcc internals manual also - covers RTL which is used frequently with assembly language in the kernel. - ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 10: Allocating memory ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ @@ -706,8 +697,7 @@ A reasonable rule of thumb is to not put inline at functions that have more than 3 lines of code in them. An exception to this rule are the cases where a parameter is known to be a compiletime constant, and as a result of this constantness you *know* the compiler will be able to optimize most of your -function away at compile time. For a good example of this later case, see -the kmalloc() inline function. +function away at compile time. Often people argue that adding inline to functions that are static and used only once is always a win since there is no space tradeoff. While this is @@ -732,7 +722,7 @@ Functions whose return value is the actual result of a computation, rather than an indication of whether the computation succeeded, are not subject to this rule. Generally they indicate failure by returning some out-of-range result. Typical examples would be functions that return pointers; they use -NULL or the ERR_PTR mechanism to report failure. +NULL or the mechanism to report failure. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 13: Editor modelines and other cruft @@ -769,39 +759,7 @@ own custom mode, or may have some other magic method for making indentation work correctly. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 14: Inline assembly -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -In architecture-specific code, you may need to use inline assembly to interface -with 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 can -and should poke hardware from C when possible. - -Consider writing simple helper functions that wrap common bits of inline -assembly, rather than repeatedly writing them with slight variations. Remember -that inline assembly can use C parameters. - -Large, non-trivial assembly functions should go in .S files, with corresponding -C prototypes defined in C header files. The C prototypes for assembly -functions should use ``asmlinkage``. - -You may need to mark your asm statement as volatile, to prevent GCC from -removing it if GCC doesn't notice any side effects. You don't always need to -do so, though, and doing so unnecessarily can limit optimization. - -When writing a single inline assembly statement containing multiple -instructions, put each instruction on a separate line in a separate quoted -string, and end each string except the last with ``\n\t`` to properly indent the -next instruction in the assembly output: - -.. code-block:: c - - asm ("magic %reg1, #42\n\t" - "more_magic %reg2, %reg3" - : /* outputs */ : /* inputs */ : /* clobbers */); - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 15: Conditional Compilation +Chapter 14: Conditional Compilation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Wherever possible, don't use preprocessor conditionals (``#if``, ``#ifdef``) in @@ -815,29 +773,11 @@ remain easy to follow. Prefer to compile out entire functions, rather than portions of functions or portions of expressions. Rather than putting an ``#ifdef`` in an expression, factor out part or all of the expression into a separate helper function and -apply the conditional to that function. +apply the condition to that function. If you have a function or variable which may potentially go unused in a particular configuration, and the compiler would warn about its definition going unused, do not compile it and use #if for this. -(However, if a function or variable *always* goes unused, delete it.) - -Within code, where possible, use the IS_ENABLED macro to convert a Kconfig -symbol into a C boolean expression, and use it in a normal C conditional: - -.. code-block:: c - - if (IS_ENABLED(CONFIG_SOMETHING)) { - ... - } - -The compiler will constant-fold the conditional away, and include or exclude -the block of code just as with an #ifdef, so this will not add any runtime -overhead. -However, this approach still allows the C compiler to see the code -inside the block, and check it for correctness (syntax, types, symbol -references, etc). Thus, you still have to use an ``#ifdef`` if the code inside -the block 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 conditional @@ -850,7 +790,7 @@ expression used. For instance: #endif /* CONFIG_SOMETHING */ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 16: Header files +Chapter 15: Header files ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Use ``#pragma once`` in the headers. As the header guards we refer to this @@ -882,7 +822,7 @@ the header file down to this: // ... header code ... ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 17: Other +Chapter 16: Other ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ * We don't apply ``!`` operator to non-boolean values. It means, to check From 56139a46dadbd6250616f0b1eeabf07cb16d2595 Mon Sep 17 00:00:00 2001 From: Natalia Ogoreltseva Date: Tue, 30 Mar 2021 17:48:54 +0300 Subject: [PATCH 7/8] fix headings formatting according to documentation guidelines --- doc/dev_guide/c_style_guide_new.rst | 67 +++++++++++------------------ 1 file changed, 24 insertions(+), 43 deletions(-) diff --git a/doc/dev_guide/c_style_guide_new.rst b/doc/dev_guide/c_style_guide_new.rst index b24fd17799..6c8a26e95a 100644 --- a/doc/dev_guide/c_style_guide_new.rst +++ b/doc/dev_guide/c_style_guide_new.rst @@ -1,6 +1,6 @@ -================================================================================ - C Style Guide -================================================================================ + +C Style Guide +============= We use Git for revision control. The latest development is happening in the default branch (currently ``master``). Our git repository is hosted on GitHub, @@ -20,17 +20,15 @@ However, we have some additional guidelines, either unique to Tarantool or deviating from the Kernel guidelines. Below we rewrite the Linux kernel coding style according to the Tarantool's style features. --------------------------------------------------------------------------------- - Tarantool coding style --------------------------------------------------------------------------------- +Tarantool coding style +---------------------- This is a short document describing the preferred coding style for the Tarantool developers and contributors. We **insist** on following these rules in order to make our code consistent and understandable to any developer. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 1: Indentation -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~ Tabs are 8 characters (8-width tabs, not 8 whitespaces), and thus indentations are also 8 characters. There are heretic movements that try to make indentations @@ -90,9 +88,8 @@ used for indentation, and the above example is deliberately broken. Get a decent editor and don't leave whitespace at the end of lines. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 2: Breaking long lines and strings -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Coding style is all about readability and maintainability using commonly available tools. @@ -106,9 +103,8 @@ information. Descendants are always substantially shorter than the parent and are placed substantially to the right. The same applies to function headers with a long argument list. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 3: Placing Braces and Spaces -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The other issue that always comes up in C styling is the placement of braces. Unlike the indent size, there are few technical reasons to @@ -213,9 +209,8 @@ statement; in the latter case use braces in both branches: otherwise(); } -******************************************************************************** Chapter 3.1: Spaces -******************************************************************************** +^^^^^^^^^^^^^^^^^^^ Like Linux kernel, Tarantool style for use of spaces depends (mostly) on function-versus-keyword usage. Use a space after (most) keywords. The @@ -287,9 +282,8 @@ optionally strip the trailing whitespace for you; however, if applying a series of patches, this may make later patches in the series fail by changing their context lines. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 4: Naming -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~ C is a Spartan language, and so should your naming be. Unlike Modula-2 and Pascal programmers, C programmers do not use cute names like @@ -329,9 +323,8 @@ For function naming we have a convention is to use: but do not handle memory management, * ``init``/``free`` for functions which initialize/destroy libraries and subsystems. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 5: Typedefs -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~ Please don't use things like ``vps_t``. It's a **mistake** to use typedef for structures and pointers. When you see a @@ -399,9 +392,8 @@ EVER use a typedef unless you can clearly match one of those rules. In general, a pointer, or a struct that has elements that can reasonably be directly accessed should **never** be a typedef. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 6: Functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~ Functions should be short and sweet, and do just one thing. They should fit on one or two screenfuls of text (the ISO/ANSI screen size is 80x24, @@ -434,9 +426,8 @@ because it is a simple way to add valuable information for the reader. Note that we place the function return type on the line before the name and signature. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 7: Centralized exiting of functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Albeit deprecated by some people, the equivalent of the goto statement is used frequently by compilers in form of the unconditional jump instruction. @@ -507,9 +498,8 @@ fix for this is to split it up into two error labels ``err_free_bar:`` and Ideally you should simulate errors to test all exit paths. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 8: Commenting -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~ Comments are good, but there is also a danger of over-commenting. NEVER try to explain HOW your code works in a comment: it's much better to @@ -587,9 +577,8 @@ A comment and the function signature should be synchronized. Double-check if the parameter names are the same as used in the comment, and mean the same. Especially when you change one of them - ensure you changed the other. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 9: Macros, Enums and RTL -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Names of macros defining constants and labels in enums are capitalized. @@ -665,9 +654,8 @@ Things to avoid when using macros: ret is a common name for a local variable - ``__foo_ret`` is less likely to collide with an existing variable. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 10: Allocating memory -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Prefer specialized allocators like ``region``, ``mempool``, ``smalloc`` to ``malloc()/free()`` for any performance-intensive or large memory allocations. @@ -680,9 +668,8 @@ allocated memory as it is to write a valgrind suppression. Freeing all allocated memory is also dynamic-load friendly: assuming a plug-in can be dynamically loaded and unloaded multiple times, reload should not lead to a memory leak. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 11: The inline disease -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There appears to be a common misperception that gcc has a magic "make me faster" speedup option called ``inline``. While the use of inlines can be @@ -706,9 +693,8 @@ help, and the maintenance issue of removing the inline when a second user appears outweighs the potential value of the hint that tells gcc to do something it would have done anyway. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 12: Function return values and names -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Functions can return values of many different kinds, and one of the most common is a value indicating whether the function succeeded or @@ -724,9 +710,8 @@ this rule. Generally they indicate failure by returning some out-of-range result. Typical examples would be functions that return pointers; they use NULL or the mechanism to report failure. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 13: Editor modelines and other cruft -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Some editors can interpret configuration information embedded in source files, indicated with special markers. For example, emacs interprets lines marked @@ -758,9 +743,8 @@ includes markers for indentation and mode configuration. People may use their own custom mode, or may have some other magic method for making indentation work correctly. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 14: Conditional Compilation -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Wherever possible, don't use preprocessor conditionals (``#if``, ``#ifdef``) in .c files; doing so makes code harder to read and logic harder to follow. Instead, @@ -789,9 +773,8 @@ expression used. For instance: ... #endif /* CONFIG_SOMETHING */ -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 15: Header files -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~ Use ``#pragma once`` in the headers. As the header guards we refer to this construction: @@ -821,9 +804,8 @@ the header file down to this: // ... header code ... -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 16: Other -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~ * We don't apply ``!`` operator to non-boolean values. It means, to check if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, @@ -835,9 +817,8 @@ Chapter 16: Other * The not-so-current list of all GCC C extensions can be found at: http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Appendix I: References -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Appendix I: References +~~~~~~~~~~~~~~~~~~~~~~ * `The C Programming Language, Second Edition `_ by Brian W. Kernighan and Dennis M. Ritchie. From c0cfa1219e7cb58ec60f743dea8273378206fcdb Mon Sep 17 00:00:00 2001 From: Natalia Ogoreltseva Date: Wed, 7 Apr 2021 11:28:04 +0300 Subject: [PATCH 8/8] remove extra files --- doc/dev_guide/c_style_guide.rst | 895 +++++++++-------------- doc/dev_guide/c_style_guide_new.rst | 841 ---------------------- doc/dev_guide/c_style_guide_old.rst | 1037 --------------------------- 3 files changed, 345 insertions(+), 2428 deletions(-) delete mode 100644 doc/dev_guide/c_style_guide_new.rst delete mode 100644 doc/dev_guide/c_style_guide_old.rst diff --git a/doc/dev_guide/c_style_guide.rst b/doc/dev_guide/c_style_guide.rst index 4657470d05..2efd655bef 100644 --- a/doc/dev_guide/c_style_guide.rst +++ b/doc/dev_guide/c_style_guide.rst @@ -1,63 +1,48 @@ -------------------------------------------------------------------------------- - C Style Guide -------------------------------------------------------------------------------- + +C Style Guide +============= We use Git for revision control. The latest development is happening in the default branch (currently ``master``). Our git repository is hosted on GitHub, and can be checked out with ``git clone git://github.com/tarantool/tarantool.git`` (anonymous read-only access). -If you have any questions about Tarantool internals, please post them on the -`developer discussion list `_ -or on `StackOverflow `_. -Additionally, some engineers are always present on #tarantool channel on -irc.freenode.net. +If you have any questions about Tarantool internals, please post them on +`StackOverflow `_ or +ask Tarantool developers directly in `telegram `_. **General guidelines** -The project's coding style is based on the `Linux kernel coding style +The project's coding style is inspired by the `Linux kernel coding style `_. However, we have some additional guidelines, either unique to Tarantool or -deviating from the Kernel guidelines. Below we cite the Linux kernel -coding style noting Tarantool's style features. - -We don't cite chapters 10 "Kconfig configuration files", 11 "Data structures", -13 "Printing kernel messages", and 17 "Don't re-invent the kernel macros" since -they are specific to Linux kernel programming environment. +deviating from the Kernel guidelines. Below we rewrite the Linux kernel +coding style according to the Tarantool's style features. -================================================================================ - Linux kernel coding style -================================================================================ +Tarantool coding style +---------------------- This is a short document describing the preferred coding style for the -linux kernel. Coding style is very personal, and I won't **force** my -views on anybody, but this is what goes for anything that I have to be -able to maintain, and I'd prefer it for most other things too. Please -at least consider the points made here. - -First off, I'd suggest printing out a copy of the GNU coding standards, -and NOT read it. Burn them, it's a great symbolic gesture. - -Anyway, here goes: +Tarantool developers and contributors. We **insist** on following these rules +in order to make our code consistent and understandable to any developer. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 1: Indentation -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~ -Tabs are 8 characters, and thus indentations are -also 8 characters. There are heretic movements that try to make indentations +Tabs are 8 characters (8-width tabs, not 8 whitespaces), 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 to be 3. Rationale: The whole idea behind indentation is to clearly define where -a block of control starts and ends. Especially when you've been looking +a block of control starts and ends. Especially when you've been looking at your screen for 20 straight hours, you'll find it a lot easier to see how the indentation works if you have large indentations. Now, some people will claim that having 8-character indentations makes the code move too far to the right, and makes it hard to read on a -80-character terminal screen. The answer to that is that if you need +80-character terminal screen. The answer to that is that if you need more than 3 levels of indentation, you're screwed anyway, and should fix your program. @@ -69,7 +54,7 @@ The preferred way to ease multiple indentation levels in a switch statement is to align the ``switch`` and its subordinate ``case`` labels in the same column instead of ``double-indenting`` the ``case`` labels. E.g.: -.. code-block:: c +.. code-block:: c switch (suffix) { case 'G': @@ -91,44 +76,35 @@ instead of ``double-indenting`` the ``case`` labels. E.g.: Don't put multiple statements on a single line unless you have something to hide: -.. code-block:: c +.. code-block:: c - if (condition) do_this; - do_something_everytime; + if (condition) do_this; + do_something_everytime; -Don't put multiple assignments on a single line either. Kernel coding style -is super simple. Avoid tricky expressions. +Don't put multiple assignments on a single line either. Avoid tricky expressions. Outside of comments and documentation, spaces are never used for indentation, and the above example is deliberately broken. Get a decent editor and don't leave whitespace at the end of lines. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 2: Breaking long lines and strings -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Coding style is all about readability and maintainability using commonly available tools. The limit on the length of lines is 80 columns and this is a strongly -preferred limit. - -.. admonition:: Tarantool Style - :class: FACT - - As for comments, the same limit of 80 columns is applied. +preferred limit. As for comments, the same limit of 80 columns is applied. Statements longer than 80 columns will be broken into sensible chunks, unless exceeding 80 columns significantly increases readability and does not hide information. Descendants are always substantially shorter than the parent and are placed substantially to the right. The same applies to function headers -with a long argument list. However, never break user-visible strings such as -printk messages, because that breaks the ability to grep for them. +with a long argument list. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 3: Placing Braces and Spaces -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The other issue that always comes up in C styling is the placement of braces. Unlike the indent size, there are few technical reasons to @@ -136,16 +112,16 @@ choose one placement strategy over the other, but the preferred way, as shown to us by the prophets Kernighan and Ritchie, is to put the opening brace last on the line, and put the closing brace first, thus: -.. code-block:: c +.. code-block:: c if (x is true) { we do y } This applies to all non-function statement blocks (if, switch, for, -while, do). E.g.: +while, do). E.g.: -.. code-block:: c +.. code-block:: c switch (action) { case KOBJ_ADD: @@ -161,7 +137,7 @@ while, do). E.g.: However, there is one special case, namely functions: they have the opening brace at the beginning of the next line, thus: -.. code-block:: c +.. code-block:: c int function(int x) @@ -170,8 +146,8 @@ opening brace at the beginning of the next line, thus: } Heretic people all over the world have claimed that this inconsistency -is ... well ... inconsistent, but all right-thinking people know that -(a) K&R are **right** and (b) K&R are right. Besides, functions are +is ... well ... inconsistent, but all right-thinking people know that +(a) K&R are **right** and (b) K&R are right. Besides, functions are special anyway (you can't nest them in C). Note that the closing brace is empty on a line of its own, **except** in @@ -179,7 +155,7 @@ the cases where it is followed by a continuation of the same statement, ie a ``while`` in a do-statement or an ``else`` in an if-statement, like this: -.. code-block:: c +.. code-block:: c do { body of do-loop @@ -187,7 +163,7 @@ this: and -.. code-block:: c +.. code-block:: c if (x == y) { .. @@ -200,20 +176,21 @@ and 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 the +(or almost empty) lines, without any loss of readability. Thus, as the supply of new-lines on your screen is not a renewable resource (think 25-line terminal screens here), you have more empty lines to put comments on. Do not unnecessarily use braces where a single statement will do. -.. code-block:: c +.. code-block:: c if (condition) action(); - and -.. code-block:: none +and + +.. code-block:: c if (condition) do_this(); @@ -223,7 +200,7 @@ Do not unnecessarily use braces where a single statement will do. This does not apply if only one branch of a conditional statement is a single statement; in the latter case use braces in both branches: -.. code-block:: c +.. code-block:: c if (condition) { do_this(); @@ -232,45 +209,44 @@ statement; in the latter case use braces in both branches: otherwise(); } -********************** Chapter 3.1: Spaces -********************** +^^^^^^^^^^^^^^^^^^^ Tarantool style for use of spaces depends (mostly) on function-versus-keyword usage. Use a space after (most) keywords. The -notable exceptions are sizeof, typeof, alignof, and __attribute__, which look -somewhat like functions (and are usually used with parentheses in Linux, +notable exceptions are ``sizeof``, ``typeof``, ``alignof``, and ``__attribute__``, +which look somewhat like functions (and are usually used with parentheses, although they are not required in the language, as in: ``sizeof info`` after ``struct fileinfo info;`` is declared). So use a space after these keywords: -.. code-block:: c +.. code-block:: c - if, switch, case, for, do, while + if, switch, case, for, do, while -but not with sizeof, typeof, alignof, or __attribute__. E.g., +but not with ``sizeof``, ``typeof``, ``alignof``, or ``__attribute__``. E.g., -.. code-block:: c +.. code-block:: c - s = sizeof(struct file); + s = sizeof(struct file); Do not add spaces around (inside) parenthesized expressions. This example is **bad**: -.. code-block:: c +.. code-block:: c - s = sizeof( struct file ); + s = sizeof( struct file ); When declaring pointer data or a function that returns a pointer type, the preferred use of ``*`` is adjacent to the data name or function name and not adjacent to the type name. Examples: -.. code-block:: c +.. code-block:: c - char *linux_banner; - unsigned long long memparse(char *ptr, char **retptr); - char *match_strdup(substring_t *s); + char *linux_banner; + unsigned long long memparse(char *ptr, char **retptr); + char *match_strdup(substring_t *s); Use one space around (on each side of) most binary and ternary operators, such as any of these:: @@ -291,6 +267,9 @@ no space after the prefix increment & decrement unary operators:: and no space around the ``.`` and ``->`` structure member operators. +Do not split a cast operator from its argument with a whitespace, +e.g. ``(ssize_t)inj->iparam``. + Do not leave trailing whitespace at the ends of lines. Some editors with ``smart`` indentation will insert whitespace at the beginning of new lines as appropriate, so you can start typing the next line of code right away. @@ -303,15 +282,8 @@ optionally strip the trailing whitespace for you; however, if applying a series of patches, this may make later patches in the series fail by changing their context lines. -.. admonition:: Tarantool Style - :class: FACT - - Do not split a cast operator from its argument with a whitespace, - e.g. ``(ssize_t)inj->iparam``. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 4: Naming -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~ C is a Spartan language, and so should your naming be. Unlike Modula-2 and Pascal programmers, C programmers do not use cute names like @@ -320,11 +292,11 @@ variable ``tmp``, which is much easier to write, and not the least more difficult to understand. HOWEVER, while mixed-case names are frowned upon, descriptive names for -global variables are a must. To call a global function ``foo`` is a +global variables are a must. To call a global function ``foo`` is a shooting offense. GLOBAL variables (to be used only if you **really** need them) need to -have descriptive names, as do global functions. If you have a function +have descriptive names, as do global functions. If you have a function that counts the number of active users, you should call that ``count_active_users()`` or similar, you should **not** call it ``cntusr()``. @@ -333,102 +305,86 @@ notation) is brain damaged - the compiler knows the types anyway and can check those, and it only confuses the programmer. No wonder MicroSoft makes buggy programs. -LOCAL variable names should be short, and to the point. If you have +LOCAL variable names should be short, and to the point. If you have some random integer loop counter, it should probably be called ``i``. Calling it ``loop_counter`` is non-productive, if there is no chance of it -being mis-understood. Similarly, ``tmp`` can be just about any type of +being mis-understood. Similarly, ``tmp`` can be just about any type of variable that is used to hold a temporary value. If you are afraid to mix up your local variable names, you have another problem, which is called the function-growth-hormone-imbalance syndrome. See chapter 6 (Functions). -.. admonition:: Tarantool Style - :class: FACT +For function naming we have a convention is to use: - For function naming we have a convention is to use: +* ``new``/``delete`` for functions which + allocate + initialize and destroy + deallocate an object, +* ``create``/``destroy`` for functions which initialize/destroy an object + but do not handle memory management, +* ``init``/``free`` for functions which initialize/destroy libraries and subsystems. - * ``new``/``delete`` for functions which - allocate + initialize and destroy + deallocate an object, - * ``create``/``destroy`` for functions which initialize/destroy an object - but do not handle memory management, - * ``init``/``free`` for functions which initialize/destroy libraries and subsystems. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 5: Typedefs -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~ Please don't use things like ``vps_t``. It's a **mistake** to use typedef for structures and pointers. When you see a -.. code-block:: c +.. code-block:: c - vps_t a; + vps_t a; in the source, what does it mean? In contrast, if it says -.. code-block:: c +.. code-block:: c - struct virtual_container *a; + struct virtual_container *a; you can actually tell what ``a`` is. Lots of people think that typedefs ``help readability``. Not so. They are useful only for: -#. totally opaque objects (where the typedef is actively used to **hide** +#. Totally opaque objects (where the typedef is actively used to **hide** what the object is). - Example: ``pte_t`` etc. opaque objects that you can only access using - the proper accessor functions. - - .. note:: - - Opaqueness and ``accessor functions`` are not good in themselves. - The reason we have them for things like pte_t etc. is that there - really is absolutely **zero** portably accessible information there. + Example: ``pte_t`` etc. opaque objects that you can only access using + the proper accessor functions. -#. Clear integer types, where the abstraction **helps** avoid confusion - whether it is ``int`` or ``long``. + .. note:: - u8/u16/u32 are perfectly fine typedefs, although they fit into - point 4 better than here. + Opaqueness and ``accessor functions`` are not good in themselves. + The reason we have them for things like pte_t etc. is that there + really is absolutely **zero** portably accessible information there. - .. note:: +#. Clear integer types, where the abstraction **helps** avoid confusion + whether it is ``int`` or ``long``. - Again - there needs to be a **reason** for this. If something is - ``unsigned long``, then there's no reason to do - typedef unsigned long myflags_t; + u8/u16/u32 are perfectly fine typedefs, although they fit into + point 4 better than here. - but if there is a clear reason for why it under certain circumstances - might be an ``unsigned int`` and under other configurations might be - ``unsigned long``, then by all means go ahead and use a typedef. + .. note:: -#. when you use sparse to literally create a **new** type for - type-checking. + Again - there needs to be a **reason** for this. If something is + ``unsigned long``, then there's no reason to do + typedef unsigned long myflags_t; -#. New types which are identical to standard C99 types, in certain - exceptional circumstances. + but if there is a clear reason for why it under certain circumstances + might be an ``unsigned int`` and under other configurations might be + ``unsigned long``, then by all means go ahead and use a typedef. - Although it would only take a short amount of time for the eyes and - brain to become accustomed to the standard types like ``uint32_t``, - some people object to their use anyway. +#. When you use sparse to literally create a **new** type for + type-checking. - Therefore, the Linux-specific ``u8/u16/u32/u64`` types and their - signed equivalents which are identical to standard types are - permitted -- although they are not mandatory in new code of your - own. +#. New types which are identical to standard C99 types, in certain + exceptional circumstances. - When editing existing code which already uses one or the other set - of types, you should conform to the existing choices in that code. + Although it would only take a short amount of time for the eyes and + brain to become accustomed to the standard types like ``uint32_t``, + some people object to their use anyway. -#. Types safe for use in userspace. - - In certain structures which are visible to userspace, we cannot - require C99 types and cannot use the ``u32`` form above. Thus, we - use __u32 and similar types in all structures which are shared - with userspace. + When editing existing code which already uses one or the other set + of types, you should conform to the existing choices in that code. Maybe there are other cases too, but the rule should basically be to NEVER EVER use a typedef unless you can clearly match one of those rules. @@ -436,16 +392,15 @@ EVER use a typedef unless you can clearly match one of those rules. In general, a pointer, or a struct that has elements that can reasonably be directly accessed should **never** be a typedef. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 6: Functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~ Functions should be short and sweet, and do just one thing. They should fit 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 the -complexity and indentation level of that function. So, if you have a +complexity and indentation level of that function. So, if you have a conceptually simple function that is just one long (but simple) case-statement, where you have to do lots of small things for a lot of different cases, it's OK to have a longer function. @@ -465,41 +420,23 @@ generally easily keep track of about 7 different things, anything more and it gets confused. You know you're brilliant, but maybe you'd like to understand what you did 2 weeks from now. -In source files, separate functions with one blank line. If the function is -exported, the **EXPORT** macro for it should follow immediately after the -closing function brace line. E.g.: - -.. code-block:: c - - int - system_is_up(void) - { - return system_state == SYSTEM_RUNNING; - } - EXPORT_SYMBOL(system_is_up); - In function prototypes, include parameter names with their data types. -Although this is not required by the C language, it is preferred in Linux +Although this is not required by the C language, it is preferred in Tarantool because it is a simple way to add valuable information for the reader. -.. admonition:: Tarantool Style - :class: FACT +Note that we place the function return type on the line before the name and signature. - Note that in Tarantool, we place the function return type on the - line before the name and signature. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 7: Centralized exiting of functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Albeit deprecated by some people, the equivalent of the goto statement is used frequently by compilers in form of the unconditional jump instruction. The goto statement comes in handy when a function exits from multiple -locations and some common work such as cleanup has to be done. If there is no +locations and some common work such as cleanup has to be done. If there is no cleanup needed then just return directly. -Choose label names which say what the goto does or why the goto exists. An +Choose label names which say what the goto does or why the goto exists. An example of a good name could be ``out_free_buffer:`` if the goto frees ``buffer``. Avoid using GW-BASIC names like ``err1:`` and ``err2:``, as you would have to renumber them if you ever add or remove exit paths, and they make correctness @@ -513,59 +450,58 @@ The rationale for using gotos is: modifications are prevented - saves the compiler work to optimize redundant code away ;) -.. code-block:: c +.. code-block:: c - int - fun(int a) - { - int result = 0; - char *buffer; + int + fun(int a) + { + int result = 0; + char *buffer; - buffer = kmalloc(SIZE, GFP_KERNEL); - if (!buffer) - return -ENOMEM; + buffer = kmalloc(SIZE, GFP_KERNEL); + if (!buffer) + return -ENOMEM; - if (condition1) { - while (loop1) { - ... + if (condition1) { + while (loop1) { + ... + } + result = 1; + goto out_free_buffer; } - result = 1; - goto out_free_buffer; + ... + out_free_buffer: + kfree(buffer); + return result; } - ... - out_free_buffer: - kfree(buffer); - return result; - } A common type of bug to be aware of is ``one err bugs`` which look like this: -.. code-block:: c +.. code-block:: c - err: - kfree(foo->bar); - kfree(foo); - return ret; + err: + kfree(foo->bar); + kfree(foo); + return ret; -The bug in this code is that on some exit paths ``foo`` is NULL. Normally the +The bug in this code is that on some exit paths ``foo`` is NULL. Normally the fix for this is to split it up into two error labels ``err_free_bar:`` and ``err_free_foo:``: -.. code-block:: c +.. code-block:: c - err_free_bar: - kfree(foo->bar); - err_free_foo: - kfree(foo); - return ret; + err_free_bar: + kfree(foo->bar); + err_free_foo: + kfree(foo); + return ret; Ideally you should simulate errors to test all exit paths. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chapter 8: Commenting -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~ -Comments are good, but there is also a danger of over-commenting. NEVER +Comments are good, but there is also a danger of over-commenting. NEVER try to explain HOW your code works in a comment: it's much better to write the code so that the **working** is obvious, and it's a waste of time to explain badly written code. @@ -579,144 +515,76 @@ ugly), but try to avoid excess. Instead, put the comments at the head of the function, telling people what it does, and possibly WHY it does it. -.. admonition:: Tarantool Style - :class: FACT - - When commenting the Tarantool C API functions, please use Doxygen comment format, - Javadoc flavor, i.e. `@tag` rather than `\\tag`. - The main tags in use are ``@param``, ``@retval``, ``@return``, ``@see``, - ``@note`` and ``@todo``. - - Every function, except perhaps a very short and obvious one, should have a - comment. A sample function comment may look like below: - - .. code-block:: c - - /** Write all data to a descriptor. - * - * This function is equivalent to 'write', except it would ensure - * that all data is written to the file unless a non-ignorable - * error occurs. - * - * @retval 0 Success - * - * @retval 1 An error occurred (not EINTR) - * / - static int - write_all(int fd, void \*data, size_t len); - - It's also important to comment data, whether they are basic types or derived - types. To this end, use just one data declaration per line (no commas for - multiple data declarations). This leaves you room for a small comment on each - item, explaining its use. - - Public structures and important structure members should be commented as well. - - In C comments out of functions and inside of functions should be different in - how they are started. Everything else is wrong. Below are correct examples. - /** comes for documentation comments, /* for local not documented comments. - However the difference is vague already, so the rule is simple: - out of function = /\**, inside = /\*. - - .. code-block:: c - - /** - * Out of function comment, option 1. - */ +When commenting the Tarantool C API functions, please use Doxygen comment format, +Javadoc flavor, i.e. `@tag` rather than `\\tag`. +The main tags in use are ``@param``, ``@retval``, ``@return``, ``@see``, +``@note`` and ``@todo``. - /** Out of function comment, option 2. */ +Every function, except perhaps a very short and obvious one, should have a +comment. A sample function comment may look like below: - int - function() - { - /* Comment inside function, option 1. */ +.. code-block:: c - /* - * Comment inside function, option 2. - */ - } + /** + * Write all data to a descriptor. + * + * This function is equivalent to 'write', except it would ensure + * that all data is written to the file unless a non-ignorable + * error occurs. + * + * @retval 0 Success + * @retval 1 An error occurred (not EINTR) + */ + static int + write_all(int fd, void *data, size_t len); + +It's also important to comment data types, whether they are basic types or +derived ones. To this end, use just one data declaration per line (no commas +for multiple data declarations). This leaves you room for a small comment on +each item, explaining its use. + +Public structures and important structure members should be commented as well. - If a function has declaration and implementation separated, the function comment - should be for the declaration. Usually in the header file. Don't duplicate the - comment. - - A comment and the function signature should be synchronized. Double-check if the - parameter names are the same as used in the comment, and mean the same. - Especially when you change one of them - ensure you changed the other. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 9: You've made a mess of it -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -That's OK, we all do. You've probably been told by your long-time Unix -user helper that ``GNU emacs`` automatically formats the C sources for -you, and you've noticed that yes, it does do that, but the defaults it -uses are less than desirable (in fact, they are worse than random -typing - an infinite number of monkeys typing into GNU emacs would never -make a good program). - -So, you can either get rid of GNU emacs, or change it to use saner -values. To do the latter, you can stick the following in your .emacs file: - -.. code-block:: none - - (defun c-lineup-arglist-tabs-only (ignored) - "Line up argument lists by tabs, not spaces" - (let* ((anchor (c-langelem-pos c-syntactic-element)) - (column (c-langelem-2nd-pos c-syntactic-element)) - (offset (- (1+ column) anchor)) - (steps (floor offset c-basic-offset))) - (* (max steps 1) - c-basic-offset))) - - (add-hook 'c-mode-common-hook - (lambda () - ;; Add kernel style - (c-add-style - "linux-tabs-only" - '("linux" (c-offsets-alist - (arglist-cont-nonempty - c-lineup-gcc-asm-reg - c-lineup-arglist-tabs-only)))))) - - (add-hook 'c-mode-hook - (lambda () - (let ((filename (buffer-file-name))) - ;; Enable kernel mode for the appropriate files - (when (and filename - (string-match (expand-file-name "~/src/linux-trees") - filename)) - (setq indent-tabs-mode t) - (setq show-trailing-whitespace t) - (c-set-style "linux-tabs-only"))))) - -This will make emacs go better with the kernel coding style for C -files below ``~/src/linux-trees``. - -But even if you fail in getting emacs to do sane formatting, not -everything is lost: use ``indent``. - -Now, again, GNU indent has the same brain-dead settings that GNU emacs -has, 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 indent -recognize the authority of K&R (the GNU people aren't evil, they are -just severely misguided in this matter), so you just give indent the -options ``-kr -i8`` (stands for ``K&R, 8 character indents``), or use -``scripts/Lindent``, which indents in the latest style. - -``indent`` has a lot of options, and especially when it comes to comment -re-formatting you may want to take a look at the man page. But -remember: ``indent`` is not a fix for bad programming. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 10: Macros, Enums and RTL -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +In C comments out of functions and inside of functions should be different in +how they are started. Everything else is wrong. Below are correct examples. +``/**`` comes for documentation comments, ``/*`` for local not documented comments. +However the difference is vague already, so the rule is simple: +out of function use ``/**``, inside use ``/*``. + +.. code-block:: c + + /** + * Out of function comment, option 1. + */ + + /** Out of function comment, option 2. */ + + int + function() + { + /* Comment inside function, option 1. */ + + /* + * Comment inside function, option 2. + */ + } + +If a function has declaration and implementation separated, the function comment +should be for the declaration. Usually in the header file. Don't duplicate the +comment. + +A comment and the function signature should be synchronized. Double-check if the +parameter names are the same as used in the comment, and mean the same. +Especially when you change one of them - ensure you changed the other. + +Chapter 9: Macros, Enums and RTL +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Names of macros defining constants and labels in enums are capitalized. -.. code-block:: c +.. code-block:: c - #define CONSTANT 0x12345 + #define CONSTANT 0x12345 Enums are preferred when defining several related constants. @@ -727,95 +595,87 @@ Generally, inline functions are preferable to macros resembling functions. Macros with multiple statements should be enclosed in a do - while block: -.. code-block:: c +.. code-block:: c - #define macrofun(a, b, c) \ - do { \ - if (a == 5) \ - do_this(b, c); \ - } while (0) + #define macrofun(a, b, c) \ + do { \ + if (a == 5) \ + do_this(b, c); \ + } while (0) Things to avoid when using macros: -1) macros that affect control flow: - - .. code-block:: c - - #define FOO(x) \ - do { \ - if (blah(x) < 0) \ - return -EBUGGERED; \ - } while (0) +1) macros that affect control flow: - is a **very** bad idea. It looks like a function call but exits the ``calling`` - function; don't break the internal parsers of those who will read the code. + .. code-block:: c -2) macros that depend on having a local variable with a magic name: + #define FOO(x) \ + do { \ + if (blah(x) < 0) \ + return -EBUGGERED; \ + } while (0) - .. code-block:: c + is a **very** bad idea. It looks like a function call but exits the ``calling`` + function; don't break the internal parsers of those who will read the code. - #define FOO(val) bar(index, val) +2) macros that depend on having a local variable with a magic name: - might look like a good thing, but it's confusing as hell when one reads the - code and it's prone to breakage from seemingly innocent changes. + .. code-block:: c -3) macros with arguments that are used as l-values: FOO(x) = y; will - bite you if somebody e.g. turns FOO into an inline function. + #define FOO(val) bar(index, val) -4) forgetting about precedence: macros defining constants using expressions - must enclose the expression in parentheses. Beware of similar issues with - macros using parameters. + might look like a good thing, but it's confusing as hell when one reads the + code and it's prone to breakage from seemingly innocent changes. - .. code-block:: c +3) macros with arguments that are used as l-values: ``FOO(x) = y;`` will + bite you if somebody e.g. turns FOO into an inline function. - #define CONSTANT 0x4000 - #define CONSTEXP (CONSTANT | 3) +4) forgetting about precedence: macros defining constants using expressions + must enclose the expression in parentheses. Beware of similar issues with + macros using parameters. -5) namespace collisions when defining local variables in macros resembling - functions: + .. code-block:: c - .. code-block:: c + #define CONSTANT 0x4000 + #define CONSTEXP (CONSTANT | 3) - #define FOO(x) \ - ({ \ - typeof(x) ret; \ - ret = calc_ret(x); \ - (ret); \ - }) +5) namespace collisions when defining local variables in macros resembling + functions: - ret is a common name for a local variable - __foo_ret is less likely - to collide with an existing variable. + .. code-block:: c - The cpp manual deals with macros exhaustively. The gcc internals manual also - covers RTL which is used frequently with assembly language in the kernel. + #define FOO(x) \ + ({ \ + typeof(x) ret; \ + ret = calc_ret(x); \ + (ret); \ + }) -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 11: Allocating memory -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + ret is a common name for a local variable - ``__foo_ret`` is less likely + to collide with an existing variable. -.. admonition:: Tarantool Style - :class: FACT +Chapter 10: Allocating memory +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Prefer the supplied slab (salloc) and pool (palloc) allocators to malloc()/free() - for any performance-intensive or large memory allocations. Repetitive use of - malloc()/free() can lead to memory fragmentation and should therefore be avoided. +Prefer specialized allocators like ``region``, ``mempool``, ``smalloc`` to +``malloc()/free()`` for any performance-intensive or large memory allocations. +Repetitive use of ``malloc()``/``free()`` can lead to memory fragmentation +and should therefore be avoided. - Always free all allocated memory, even allocated at start-up. We aim at being - valgrind leak-check clean, and in most cases it's just as easy to free() the - allocated memory as it is to write a valgrind suppression. Freeing all allocated - memory is also dynamic-load friendly: assuming a plug-in can be dynamically - loaded and unloaded multiple times, reload should not lead to a memory leak. +Always free all allocated memory, even allocated at start-up. We aim at being +valgrind leak-check clean, and in most cases it's just as easy to ``free()`` the +allocated memory as it is to write a valgrind suppression. Freeing all allocated +memory is also dynamic-load friendly: assuming a plug-in can be dynamically +loaded and unloaded multiple times, reload should not lead to a memory leak. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 12: The inline disease -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 11: The inline disease +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There appears to be a common misperception that gcc has a magic "make me faster" speedup option called ``inline``. While the use of inlines can be -appropriate (for example as a means of replacing macros, see Chapter 12), it -very often is not. Abundant use of the inline keyword leads to a much bigger -kernel, which in turn slows the system as a whole down, due to a bigger -icache footprint for the CPU and simply because there is less memory +appropriate, it very often is not. Abundant use of the inline keyword leads to +a much bigger kernel, which in turn slows the system as a whole down, due to a +bigger icache footprint for the CPU and simply because there is less memory available for the pagecache. Just think about it; a pagecache miss causes a disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles that can go into these 5 milliseconds. @@ -824,8 +684,7 @@ A reasonable rule of thumb is to not put inline at functions that have more than 3 lines of code in them. An exception to this rule are the cases where a parameter is known to be a compiletime constant, and as a result of this constantness you *know* the compiler will be able to optimize most of your -function away at compile time. For a good example of this later case, see -the kmalloc() inline function. +function away at compile time. Often people argue that adding inline to functions that are static and used only once is always a win since there is no space tradeoff. While this is @@ -834,213 +693,149 @@ help, and the maintenance issue of removing the inline when a second user appears outweighs the potential value of the hint that tells gcc to do something it would have done anyway. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 13: Function return values and names -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 12: Function return values and names +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Functions can return values of many different kinds, and one of the most common is a value indicating whether the function succeeded or -failed. Such a value can be represented as an error-code integer -(-Exxx = failure, 0 = success) or a ``succeeded`` boolean (0 = failure, -non-zero = success). +failed. -Mixing up these two sorts of representations is a fertile source of -difficult-to-find bugs. If the C language included a strong distinction -between integers and booleans then the compiler would find these mistakes -for us... but it doesn't. To help prevent such bugs, always follow this -convention: - - If the name of a function is an action or an imperative command, - the function should return an error-code integer. If the name - is a predicate, the function should return a "succeeded" boolean. - -For example, ``add work`` is a command, and the add_work() function returns 0 -for success or -EBUSY for failure. In the same way, ``PCI device present`` is -a predicate, and the pci_dev_present() function returns 1 if it succeeds in -finding a matching device or 0 if it doesn't. - -All EXPORTed functions must respect this convention, and so should all -public functions. Private (static) functions need not, but it is -recommended that they do. +In 99.99999% of all cases in Tarantool we return 0 on success, non-zero on error +(-1 usually). Errors are saved into a diagnostics area which is global per fiber. +We never return error codes as a result of a function. Functions whose return value is the actual result of a computation, rather than an indication of whether the computation succeeded, are not subject to -this rule. Generally they indicate failure by returning some out-of-range -result. Typical examples would be functions that return pointers; they use -NULL or the ERR_PTR mechanism to report failure. +this rule. Generally they indicate failure by returning some out-of-range +result. Typical examples would be functions that return pointers; they use +NULL or the mechanism to report failure. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 14: Editor modelines and other cruft -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Chapter 13: Editor modelines and other cruft +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Some editors can interpret configuration information embedded in source files, -indicated with special markers. For example, emacs interprets lines marked +indicated with special markers. For example, emacs interprets lines marked like this: -.. code-block:: c +.. code-block:: c - -*- mode: c -*- + -*- mode: c -*- Or like this: -.. code-block:: c +.. code-block:: c - /* - Local Variables: - compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" - End: - */ + /* + Local Variables: + compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" + End: + */ Vim interprets markers that look like this: -.. code-block:: c +.. code-block:: c - /* vim:set sw=8 noet */ + /* vim:set sw=8 noet */ -Do not include any of these in source files. People have their own personal -editor configurations, and your source files should not override them. This -includes markers for indentation and mode configuration. People may use their +Do not include any of these in source files. People have their own personal +editor configurations, and your source files should not override them. This +includes markers for indentation and mode configuration. People may use their own custom mode, or may have some other magic method for making indentation work correctly. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 15: Inline assembly -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -In architecture-specific code, you may need to use inline assembly to interface -with 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 can -and should poke hardware from C when possible. - -Consider writing simple helper functions that wrap common bits of inline -assembly, rather than repeatedly writing them with slight variations. Remember -that inline assembly can use C parameters. - -Large, non-trivial assembly functions should go in .S files, with corresponding -C prototypes defined in C header files. The C prototypes for assembly -functions should use ``asmlinkage``. - -You may need to mark your asm statement as volatile, to prevent GCC from -removing it if GCC doesn't notice any side effects. You don't always need to -do so, though, and doing so unnecessarily can limit optimization. +Chapter 14: Conditional Compilation +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -When writing a single inline assembly statement containing multiple -instructions, put each instruction on a separate line in a separate quoted -string, and end each string except the last with \n\t to properly indent the -next instruction in the assembly output: - -.. code-block:: c - - asm ("magic %reg1, #42\n\t" - "more_magic %reg2, %reg3" - : /* outputs */ : /* inputs */ : /* clobbers */); - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 16: Conditional Compilation -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Wherever possible, don't use preprocessor conditionals (#if, #ifdef) in .c -files; doing so makes code harder to read and logic harder to follow. Instead, +Wherever possible, don't use preprocessor conditionals (``#if``, ``#ifdef``) in +.c files; 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 .c files, providing no-op stub versions in the #else case, and then call those -functions unconditionally from .c files. The compiler will avoid generating +functions unconditionally from .c files. The compiler will avoid generating any code for the stub calls, producing identical results, but the logic will remain easy to follow. Prefer to compile out entire functions, rather than portions of functions or -portions of expressions. Rather than putting an ifdef in an expression, factor -out part or all of the expression into a separate helper function and apply the -conditional to that function. +portions of expressions. Rather than putting an ``#ifdef`` in an expression, +factor out part or all of the expression into a separate helper function and +apply the condition to that function. If you have a function or variable which may potentially go unused in a particular configuration, and the compiler would warn about its definition -going unused, mark the definition as __maybe_unused rather than wrapping it in -a preprocessor conditional. (However, if a function or variable *always* goes -unused, delete it.) +going unused, do not compile it and use #if for this. -Within code, where possible, use the IS_ENABLED macro to convert a Kconfig -symbol into a C boolean expression, and use it in a normal C conditional: +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 conditional +expression used. For instance: -.. code-block:: c +.. code-block:: c - if (IS_ENABLED(CONFIG_SOMETHING)) { + #ifdef CONFIG_SOMETHING ... - } + #endif /* CONFIG_SOMETHING */ -The compiler will constant-fold the conditional away, and include or exclude -the block of code just as with an #ifdef, so this will not add any runtime -overhead. -However, this approach still allows the C compiler to see the code -inside the block, and check it for correctness (syntax, types, symbol -references, etc). Thus, you still have to use an #ifdef if the code inside the -block references symbols that will not exist if the condition is not met. +Chapter 15: Header files +~~~~~~~~~~~~~~~~~~~~~~~~ -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 conditional -expression used. For instance: +Use ``#pragma once`` in the headers. As the header guards we refer to this +construction: + +.. code-block:: c -.. code-block:: c + #ifndef THE_HEADER_IS_INCLUDED + #define THE_HEADER_IS_INCLUDED - #ifdef CONFIG_SOMETHING - ... - #endif /* CONFIG_SOMETHING */ + // ... the header code ... -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 17: Header files -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + #endif // THE_HEADER_IS_INCLUDED -.. admonition:: Tarantool Style - :class: FACT +It works fine, but the guard name ``THE_HEADER_IS_INCLUDED`` tends to +become outdated when the file is moved or renamed. This is especially +painful with multiple files having the same name in the project, but +different path. For instance, we have 3 ``error.h`` files, which means for +each of them we need to invent a new header guard name, and not forget to +update them if the files are moved or renamed. - Use header guards. Put the header guard in the first line in the header, - before the copyright or declarations. Use all-uppercase name for the header - guard. Derive the header guard name from the file name, and append _INCLUDED - to get a macro name. For example, core/log_io.h -> CORE_LOG_IO_H_INCLUDED. In - ``.c`` (implementation) file, include the respective declaration header before - all other headers, to ensure that the header is self- sufficient. Header - "header.h" is self-sufficient if the following compiles without errors: +For that reason we use ``#pragma once`` in all the new code, which shortens +the header file down to this: - .. code-block:: c +.. code-block:: c - #include "header.h" + #pragma once -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Chapter 18: Other -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + // ... header code ... -.. admonition:: Tarantool Style - :class: FACT +Chapter 16: Other +~~~~~~~~~~~~~~~~~ - * We don't apply ``!`` operator to non-boolean values. It means, to check - if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, - you use ``!= NULL``. The same for ``==``. +* We don't apply ``!`` operator to non-boolean values. It means, to check + if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, + you use ``!= NULL``. The same for ``==``. - * Select GNU C99 extensions are acceptable. It's OK to mix declarations and - statements, use true and false. +* Select GNU C99 extensions are acceptable. It's OK to mix declarations and + statements, use true and false. - * The not-so-current list of all GCC C extensions can be found at: - http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html +* The not-so-current list of all GCC C extensions can be found at: + http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Appendix I: References -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Appendix I: References +~~~~~~~~~~~~~~~~~~~~~~ -* `The C Programming Language, Second Edition `_ - by Brian W. Kernighan and Dennis M. Ritchie. |br| - Prentice Hall, Inc., 1988. |br| - ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback). +* `The C Programming Language, Second Edition `_ + by 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 Programming `_ - by Brian W. Kernighan and Rob Pike. |br| - Addison-Wesley, Inc., 1999. |br| - ISBN 0-201-61586-X. +* `The Practice of Programming `_ + by 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** +* `GNU manuals `_ - where in compliance with K&R + and this text - for **cpp**, **gcc**, **gcc internals** and **indent** -* `WG14 International standardization workgroup for the programming - language C `_ +* `WG14 International standardization workgroup for the programming + language C `_ -* `Kernel CodingStyle, by greg@kroah.com at OLS 2002 - `_ +* `Kernel CodingStyle, by greg@kroah.com at OLS 2002 + `_ diff --git a/doc/dev_guide/c_style_guide_new.rst b/doc/dev_guide/c_style_guide_new.rst deleted file mode 100644 index 6c8a26e95a..0000000000 --- a/doc/dev_guide/c_style_guide_new.rst +++ /dev/null @@ -1,841 +0,0 @@ - -C Style Guide -============= - -We use Git for revision control. The latest development is happening in the -default branch (currently ``master``). Our git repository is hosted on GitHub, -and can be checked out with ``git clone git://github.com/tarantool/tarantool.git`` -(anonymous read-only access). - -If you have any questions about Tarantool internals, please post them on -`StackOverflow `_ or -ask Tarantool developers directly in `telegram `_. - -**General guidelines** - -The project's coding style is inspired by the `Linux kernel coding style -`_. - -However, we have some additional guidelines, either unique to Tarantool or -deviating from the Kernel guidelines. Below we rewrite the Linux kernel -coding style according to the Tarantool's style features. - -Tarantool coding style ----------------------- - -This is a short document describing the preferred coding style for the -Tarantool developers and contributors. We **insist** on following these rules -in order to make our code consistent and understandable to any developer. - -Chapter 1: Indentation -~~~~~~~~~~~~~~~~~~~~~~ - -Tabs are 8 characters (8-width tabs, not 8 whitespaces), 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 to be 3. - -Rationale: The whole idea behind indentation is to clearly define where -a block of control starts and ends. Especially when you've been looking -at your screen for 20 straight hours, you'll find it a lot easier to see -how the indentation works if you have large indentations. - -Now, some people will claim that having 8-character indentations makes -the code move too far to the right, and makes it hard to read on a -80-character terminal screen. The answer to that is that if you need -more than 3 levels of indentation, you're screwed anyway, and should fix -your program. - -In short, 8-char indents make things easier to read, and have the added -benefit 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 is -to align the ``switch`` and its subordinate ``case`` labels in the same column -instead of ``double-indenting`` the ``case`` labels. E.g.: - -.. code-block:: c - - switch (suffix) { - case 'G': - case 'g': - mem <<= 30; - break; - case 'M': - case 'm': - mem <<= 20; - break; - case 'K': - case 'k': - mem <<= 10; - /* fall through */ - default: - break; - } - -Don't put multiple statements on a single line unless you have -something to hide: - -.. code-block:: c - - if (condition) do_this; - do_something_everytime; - -Don't put multiple assignments on a single line either. Avoid tricky expressions. - -Outside of comments and documentation, spaces are never -used for indentation, and the above example is deliberately broken. - -Get a decent editor and don't leave whitespace at the end of lines. - -Chapter 2: Breaking long lines and strings -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Coding style is all about readability and maintainability using commonly -available tools. - -The limit on the length of lines is 80 columns and this is a strongly -preferred limit. As for comments, the same limit of 80 columns is applied. - -Statements longer than 80 columns will be broken into sensible chunks, unless -exceeding 80 columns significantly increases readability and does not hide -information. Descendants are always substantially shorter than the parent and -are placed substantially to the right. The same applies to function headers -with a long argument list. - -Chapter 3: Placing Braces and Spaces -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -The other issue that always comes up in C styling is the placement of -braces. Unlike the indent size, there are few technical reasons to -choose one placement strategy over the other, but the preferred way, as -shown to us by the prophets Kernighan and Ritchie, is to put the opening -brace last on the line, and put the closing brace first, thus: - -.. code-block:: c - - if (x is true) { - we do y - } - -This applies to all non-function statement blocks (if, switch, for, -while, do). E.g.: - -.. code-block:: c - - switch (action) { - case KOBJ_ADD: - return "add"; - case KOBJ_REMOVE: - return "remove"; - case KOBJ_CHANGE: - return "change"; - default: - return NULL; - } - -However, there is one special case, namely functions: they have the -opening brace at the beginning of the next line, thus: - -.. code-block:: c - - int - function(int x) - { - body of function - } - -Heretic people all over the world have claimed that this inconsistency -is ... well ... inconsistent, but all right-thinking people know that -(a) K&R are **right** and (b) K&R are right. Besides, functions are -special anyway (you can't nest them in C). - -Note that the closing brace is empty on a line of its own, **except** in -the cases where it is followed by a continuation of the same statement, -ie a ``while`` in a do-statement or an ``else`` in an if-statement, like -this: - -.. code-block:: c - - do { - body of do-loop - } while (condition); - -and - -.. code-block:: c - - if (x == y) { - .. - } else if (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 the -supply of new-lines on your screen is not a renewable resource (think -25-line terminal screens here), you have more empty lines to put -comments on. - -Do not unnecessarily use braces where a single statement will do. - -.. code-block:: c - - if (condition) - action(); - -and - -.. code-block:: c - - if (condition) - do_this(); - else - do_that(); - -This does not apply if only one branch of a conditional statement is a single -statement; in the latter case use braces in both branches: - -.. code-block:: c - - if (condition) { - do_this(); - do_that(); - } else { - otherwise(); - } - -Chapter 3.1: Spaces -^^^^^^^^^^^^^^^^^^^ - -Like Linux kernel, Tarantool style for use of spaces depends (mostly) on -function-versus-keyword usage. Use a space after (most) keywords. The -notable exceptions are ``sizeof``, ``typeof``, ``alignof``, and ``__attribute__``, -which look somewhat like functions (and are usually used with parentheses in Linux, -although they are not required in the language, as in: ``sizeof info`` after -``struct fileinfo info;`` is declared). - -So use a space after these keywords: - -.. code-block:: c - - if, switch, case, for, do, while - -but not with ``sizeof``, ``typeof``, ``alignof``, or ``__attribute__``. E.g., - -.. code-block:: c - - s = sizeof(struct file); - -Do not add spaces around (inside) parenthesized expressions. This example is -**bad**: - -.. code-block:: c - - s = sizeof( struct file ); - -When declaring pointer data or a function that returns a pointer type, the -preferred use of ``*`` is adjacent to the data name or function name and not -adjacent to the type name. Examples: - -.. code-block:: c - - char *linux_banner; - unsigned long long memparse(char *ptr, char **retptr); - char *match_strdup(substring_t *s); - -Use one space around (on each side of) most binary and ternary operators, -such as any of these:: - - = + - < > * / % | & ^ <= >= == != ? : - -but no space after unary operators:: - - & * + - ~ ! sizeof typeof alignof __attribute__ defined - -no space before the postfix increment & decrement unary operators:: - - ++ -- - -no space after the prefix increment & decrement unary operators:: - - ++ -- - -and no space around the ``.`` and ``->`` structure member operators. - -Do not split a cast operator from its argument with a whitespace, -e.g. ``(ssize_t)inj->iparam``. - -Do not leave trailing whitespace at the ends of lines. Some editors with -``smart`` indentation will insert whitespace at the beginning of new lines as -appropriate, so you can start typing the next line of code right away. -However, some such editors do not remove the whitespace if you end up not -putting a line of code there, such as if you leave a blank line. As a result, -you end up with lines containing trailing whitespace. - -Git will warn you about patches that introduce trailing whitespace, and can -optionally strip the trailing whitespace for you; however, if applying a series -of patches, this may make later patches in the series fail by changing their -context lines. - -Chapter 4: Naming -~~~~~~~~~~~~~~~~~ - -C is a Spartan language, and so should your naming be. Unlike Modula-2 -and Pascal programmers, C programmers do not use cute names like -ThisVariableIsATemporaryCounter. A C programmer would call that -variable ``tmp``, which is much easier to write, and not the least more -difficult to understand. - -HOWEVER, while mixed-case names are frowned upon, descriptive names for -global variables are a must. To call a global function ``foo`` is a -shooting offense. - -GLOBAL variables (to be used only if you **really** need them) need to -have descriptive names, as do global functions. If you have a function -that counts the number of active users, you should call that -``count_active_users()`` or similar, you should **not** call it ``cntusr()``. - -Encoding the type of a function into the name (so-called Hungarian -notation) is brain damaged - the compiler knows the types anyway and can -check those, and it only confuses the programmer. No wonder MicroSoft -makes buggy programs. - -LOCAL variable names should be short, and to the point. If you have -some random integer loop counter, it should probably be called ``i``. -Calling it ``loop_counter`` is non-productive, if there is no chance of it -being mis-understood. Similarly, ``tmp`` can be just about any type of -variable that is used to hold a temporary value. - -If you are afraid to mix up your local variable names, you have another -problem, which is called the function-growth-hormone-imbalance syndrome. -See chapter 6 (Functions). - -For function naming we have a convention is to use: - -* ``new``/``delete`` for functions which - allocate + initialize and destroy + deallocate an object, -* ``create``/``destroy`` for functions which initialize/destroy an object - but do not handle memory management, -* ``init``/``free`` for functions which initialize/destroy libraries and subsystems. - -Chapter 5: Typedefs -~~~~~~~~~~~~~~~~~~~ - -Please don't use things like ``vps_t``. -It's a **mistake** to use typedef for structures and pointers. When you see a - -.. code-block:: c - - vps_t a; - -in the source, what does it mean? -In contrast, if it says - -.. code-block:: c - - struct virtual_container *a; - -you can actually tell what ``a`` is. - -Lots of people think that typedefs ``help readability``. Not so. They are -useful only for: - -#. Totally opaque objects (where the typedef is actively used to **hide** - what the object is). - - Example: ``pte_t`` etc. opaque objects that you can only access using - the proper accessor functions. - - .. note:: - - Opaqueness and ``accessor functions`` are not good in themselves. - The reason we have them for things like pte_t etc. is that there - really is absolutely **zero** portably accessible information there. - -#. Clear integer types, where the abstraction **helps** avoid confusion - whether it is ``int`` or ``long``. - - u8/u16/u32 are perfectly fine typedefs, although they fit into - point 4 better than here. - - .. note:: - - Again - there needs to be a **reason** for this. If something is - ``unsigned long``, then there's no reason to do - typedef unsigned long myflags_t; - - but if there is a clear reason for why it under certain circumstances - might be an ``unsigned int`` and under other configurations might be - ``unsigned long``, then by all means go ahead and use a typedef. - -#. When you use sparse to literally create a **new** type for - type-checking. - -#. New types which are identical to standard C99 types, in certain - exceptional circumstances. - - Although it would only take a short amount of time for the eyes and - brain to become accustomed to the standard types like ``uint32_t``, - some people object to their use anyway. - - When editing existing code which already uses one or the other set - of types, you should conform to the existing choices in that code. - -Maybe there are other cases too, but the rule should basically be to NEVER -EVER use a typedef unless you can clearly match one of those rules. - -In general, a pointer, or a struct that has elements that can reasonably -be directly accessed should **never** be a typedef. - -Chapter 6: Functions -~~~~~~~~~~~~~~~~~~~~ - -Functions should be short and sweet, and do just one thing. They should -fit 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 the -complexity and indentation level of that function. So, if you have a -conceptually simple function that is just one long (but simple) -case-statement, where you have to do lots of small things for a lot of -different cases, it's OK to have a longer function. - -However, if you have a complex function, and you suspect that a -less-than-gifted first-year high-school student might not even -understand what the function is all about, you should adhere to the -maximum limits all the more closely. Use helper functions with -descriptive names (you can ask the compiler to in-line them if you think -it's performance-critical, and it will probably do a better job of it -than you would have done). - -Another measure of the function is the number of local variables. They -shouldn't exceed 5-10, or you're doing something wrong. Re-think the -function, and split it into smaller pieces. A human brain can -generally easily keep track of about 7 different things, anything more -and it gets confused. You know you're brilliant, but maybe you'd like -to understand what you did 2 weeks from now. - -In function prototypes, include parameter names with their data types. -Although this is not required by the C language, it is preferred in Tarantool -because it is a simple way to add valuable information for the reader. - -Note that we place the function return type on the line before the name and signature. - -Chapter 7: Centralized exiting of functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Albeit deprecated by some people, the equivalent of the goto statement is -used frequently by compilers in form of the unconditional jump instruction. - -The goto statement comes in handy when a function exits from multiple -locations and some common work such as cleanup has to be done. If there is no -cleanup needed then just return directly. - -Choose label names which say what the goto does or why the goto exists. An -example of a good name could be ``out_free_buffer:`` if the goto frees ``buffer``. -Avoid using GW-BASIC names like ``err1:`` and ``err2:``, as you would have to -renumber them if you ever add or remove exit paths, and they make correctness -difficult 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 making - modifications are prevented -- saves the compiler work to optimize redundant code away ;) - -.. code-block:: c - - int - fun(int a) - { - int result = 0; - char *buffer; - - buffer = kmalloc(SIZE, GFP_KERNEL); - if (!buffer) - return -ENOMEM; - - if (condition1) { - while (loop1) { - ... - } - result = 1; - goto out_free_buffer; - } - ... - out_free_buffer: - kfree(buffer); - return result; - } - -A common type of bug to be aware of is ``one err bugs`` which look like this: - -.. code-block:: c - - err: - kfree(foo->bar); - kfree(foo); - return ret; - -The bug in this code is that on some exit paths ``foo`` is NULL. Normally the -fix for this is to split it up into two error labels ``err_free_bar:`` and -``err_free_foo:``: - -.. code-block:: c - - err_free_bar: - kfree(foo->bar); - err_free_foo: - kfree(foo); - return ret; - -Ideally you should simulate errors to test all exit paths. - -Chapter 8: Commenting -~~~~~~~~~~~~~~~~~~~~~ - -Comments are good, but there is also a danger of over-commenting. NEVER -try to explain HOW your code works in a comment: it's much better to -write the code so that the **working** is obvious, and it's a waste of -time 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 the -function 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 make -small comments to note or warn about something particularly clever (or -ugly), but try to avoid excess. Instead, put the comments at the head -of the function, telling people what it does, and possibly WHY it does -it. - -When commenting the Tarantool C API functions, please use Doxygen comment format, -Javadoc flavor, i.e. `@tag` rather than `\\tag`. -The main tags in use are ``@param``, ``@retval``, ``@return``, ``@see``, -``@note`` and ``@todo``. - -Every function, except perhaps a very short and obvious one, should have a -comment. A sample function comment may look like below: - -.. code-block:: c - - /** - * Write all data to a descriptor. - * - * This function is equivalent to 'write', except it would ensure - * that all data is written to the file unless a non-ignorable - * error occurs. - * - * @retval 0 Success - * @retval 1 An error occurred (not EINTR) - */ - static int - write_all(int fd, void *data, size_t len); - -It's also important to comment data types, whether they are basic types or -derived ones. To this end, use just one data declaration per line (no commas -for multiple data declarations). This leaves you room for a small comment on -each item, explaining its use. - -Public structures and important structure members should be commented as well. - -In C comments out of functions and inside of functions should be different in -how they are started. Everything else is wrong. Below are correct examples. -``/**`` comes for documentation comments, ``/*`` for local not documented comments. -However the difference is vague already, so the rule is simple: -out of function use ``/**``, inside use ``/*``. - -.. code-block:: c - - /** - * Out of function comment, option 1. - */ - - /** Out of function comment, option 2. */ - - int - function() - { - /* Comment inside function, option 1. */ - - /* - * Comment inside function, option 2. - */ - } - -If a function has declaration and implementation separated, the function comment -should be for the declaration. Usually in the header file. Don't duplicate the -comment. - -A comment and the function signature should be synchronized. Double-check if the -parameter names are the same as used in the comment, and mean the same. -Especially when you change one of them - ensure you changed the other. - -Chapter 9: Macros, Enums and RTL -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Names of macros defining constants and labels in enums are capitalized. - -.. code-block:: c - - #define CONSTANT 0x12345 - -Enums are preferred when defining several related constants. - -CAPITALIZED macro names are appreciated but macros resembling functions -may 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: - -.. code-block:: c - - #define macrofun(a, b, c) \ - do { \ - if (a == 5) \ - do_this(b, c); \ - } while (0) - -Things to avoid when using macros: - -1) macros that affect control flow: - - .. code-block:: c - - #define FOO(x) \ - do { \ - if (blah(x) < 0) \ - return -EBUGGERED; \ - } while (0) - - is a **very** bad idea. It looks like a function call but exits the ``calling`` - function; 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: - - .. code-block:: c - - #define FOO(val) bar(index, val) - - might look like a good thing, but it's confusing as hell when one reads the - code and it's prone to breakage from seemingly innocent changes. - -3) macros with arguments that are used as l-values: ``FOO(x) = y;`` will - bite you if somebody e.g. turns FOO into an inline function. - -4) forgetting about precedence: macros defining constants using expressions - must enclose the expression in parentheses. Beware of similar issues with - macros using parameters. - - .. code-block:: c - - #define CONSTANT 0x4000 - #define CONSTEXP (CONSTANT | 3) - -5) namespace collisions when defining local variables in macros resembling - functions: - - .. code-block:: c - - #define FOO(x) \ - ({ \ - typeof(x) ret; \ - ret = calc_ret(x); \ - (ret); \ - }) - - ret is a common name for a local variable - ``__foo_ret`` is less likely - to collide with an existing variable. - -Chapter 10: Allocating memory -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Prefer specialized allocators like ``region``, ``mempool``, ``smalloc`` to -``malloc()/free()`` for any performance-intensive or large memory allocations. -Repetitive use of ``malloc()``/``free()`` can lead to memory fragmentation -and should therefore be avoided. - -Always free all allocated memory, even allocated at start-up. We aim at being -valgrind leak-check clean, and in most cases it's just as easy to ``free()`` the -allocated memory as it is to write a valgrind suppression. Freeing all allocated -memory is also dynamic-load friendly: assuming a plug-in can be dynamically -loaded and unloaded multiple times, reload should not lead to a memory leak. - -Chapter 11: The inline disease -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -There appears to be a common misperception that gcc has a magic "make me -faster" speedup option called ``inline``. While the use of inlines can be -appropriate, it very often is not. Abundant use of the inline keyword leads to -a much bigger kernel, which in turn slows the system as a whole down, due to a -bigger icache footprint for the CPU and simply because there is less memory -available for the pagecache. Just think about it; a pagecache miss causes a -disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles -that can go into these 5 milliseconds. - -A reasonable rule of thumb is to not put inline at functions that have more -than 3 lines of code in them. An exception to this rule are the cases where -a parameter is known to be a compiletime constant, and as a result of this -constantness you *know* the compiler will be able to optimize most of your -function away at compile time. - -Often people argue that adding inline to functions that are static and used -only once is always a win since there is no space tradeoff. While this is -technically correct, gcc is capable of inlining these automatically without -help, and the maintenance issue of removing the inline when a second user -appears outweighs the potential value of the hint that tells gcc to do -something it would have done anyway. - -Chapter 12: Function return values and names -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Functions can return values of many different kinds, and one of the -most common is a value indicating whether the function succeeded or -failed. - -In 99.99999% of all cases in Tarantool we return 0 on success, non-zero on error -(-1 usually). Errors are saved into a diagnostics area which is global per fiber. -We never return error codes as a result of a function. - -Functions whose return value is the actual result of a computation, rather -than an indication of whether the computation succeeded, are not subject to -this rule. Generally they indicate failure by returning some out-of-range -result. Typical examples would be functions that return pointers; they use -NULL or the mechanism to report failure. - -Chapter 13: Editor modelines and other cruft -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Some editors can interpret configuration information embedded in source files, -indicated with special markers. For example, emacs interprets lines marked -like this: - -.. code-block:: c - - -*- mode: c -*- - -Or like this: - -.. code-block:: c - - /* - Local Variables: - compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" - End: - */ - -Vim interprets markers that look like this: - -.. code-block:: c - - /* vim:set sw=8 noet */ - -Do not include any of these in source files. People have their own personal -editor configurations, and your source files should not override them. This -includes markers for indentation and mode configuration. People may use their -own custom mode, or may have some other magic method for making indentation -work correctly. - -Chapter 14: Conditional Compilation -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Wherever possible, don't use preprocessor conditionals (``#if``, ``#ifdef``) in -.c files; 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 .c -files, providing no-op stub versions in the #else case, and then call those -functions unconditionally from .c files. The compiler will avoid generating -any code for the stub calls, producing identical results, but the logic will -remain easy to follow. - -Prefer to compile out entire functions, rather than portions of functions or -portions of expressions. Rather than putting an ``#ifdef`` in an expression, -factor out part or all of the expression into a separate helper function and -apply the condition to that function. - -If you have a function or variable which may potentially go unused in a -particular configuration, and the compiler would warn about its definition -going unused, do not compile it and use #if for this. - -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 conditional -expression used. For instance: - -.. code-block:: c - - #ifdef CONFIG_SOMETHING - ... - #endif /* CONFIG_SOMETHING */ - -Chapter 15: Header files -~~~~~~~~~~~~~~~~~~~~~~~~ - -Use ``#pragma once`` in the headers. As the header guards we refer to this -construction: - -.. code-block:: c - - #ifndef THE_HEADER_IS_INCLUDED - #define THE_HEADER_IS_INCLUDED - - // ... the header code ... - - #endif // THE_HEADER_IS_INCLUDED - -It works fine, but the guard name ``THE_HEADER_IS_INCLUDED`` tends to -become outdated when the file is moved or renamed. This is especially -painful with multiple files having the same name in the project, but -different path. For instance, we have 3 ``error.h`` files, which means for -each of them we need to invent a new header guard name, and not forget to -update them if the files are moved or renamed. - -For that reason we use ``#pragma once`` in all the new code, which shortens -the header file down to this: - -.. code-block:: c - - #pragma once - - // ... header code ... - -Chapter 16: Other -~~~~~~~~~~~~~~~~~ - -* We don't apply ``!`` operator to non-boolean values. It means, to check - if an integer is not 0, you use ``!= 0``. To check if a pointer is not NULL, - you use ``!= NULL``. The same for ``==``. - -* Select GNU C99 extensions are acceptable. It's OK to mix declarations and - statements, use true and false. - -* The not-so-current list of all GCC C extensions can be found at: - http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html - -Appendix I: References -~~~~~~~~~~~~~~~~~~~~~~ - -* `The C Programming Language, Second Edition `_ - by 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 Programming `_ - by 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** - -* `WG14 International standardization workgroup for the programming - language C `_ - -* `Kernel CodingStyle, by greg@kroah.com at OLS 2002 - `_ - diff --git a/doc/dev_guide/c_style_guide_old.rst b/doc/dev_guide/c_style_guide_old.rst deleted file mode 100644 index e8aebf7141..0000000000 --- a/doc/dev_guide/c_style_guide_old.rst +++ /dev/null @@ -1,1037 +0,0 @@ -------------------------------------------------------------------------------- - C Style Guide -------------------------------------------------------------------------------- - -The project's coding style is based on a version of the Linux kernel coding style. - -The latest version of the Linux style can be found at: -http://www.kernel.org/doc/Documentation/CodingStyle - -Since it is open for changes, the version of style that we follow, -one from 2007-July-13, will be also copied later in this document. - -There are a few additional guidelines, either unique -to Tarantool or deviating from the Kernel guidelines. - -A. Chapters 10 "Kconfig configuration files", 11 "Data structures", - 13 "Printing kernel messages", 14 "Allocating memory" and 17 - "Don't re-invent the kernel macros" do not apply, since they are - specific to Linux kernel programming environment. - -B. The rest of Linux Kernel Coding Style is amended as follows: - -=========================================================== - General guidelines -=========================================================== - -We use Git for revision control. The latest development is happening in the -default branch (currently ``master``). -Our git repository is hosted on github, and can be checked out with -``git clone git://github.com/tarantool/tarantool.git`` (anonymous read-only access). - -If you have any questions about Tarantool internals, please post them on the -developer discussion list, https://groups.google.com/forum/#!forum/tarantool. However, -please be warned: Launchpad silently deletes posts from non-subscribed members, -thus please be sure to have subscribed to the list prior to posting. Additionally, -some engineers are always present on #tarantool channel on irc.freenode.net. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Commenting style -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Use Doxygen comment format, Javadoc flavor, i.e. `@tag` rather than `\\tag`. -The main tags in use are @param, @retval, @return, @see, @note and @todo. - -Every function, except perhaps a very short and obvious one, should have a -comment. A sample function comment may look like below: - -.. code-block:: c - - /** Write all data to a descriptor. - * - * This function is equivalent to 'write', except it would ensure - * that all data is written to the file unless a non-ignorable - * error occurs. - * - * @retval 0 Success - * - * @retval 1 An error occurred (not EINTR) - * / - static int - write_all(int fd, void \*data, size_t len); - -Public structures and important structure members should be commented as well. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Header files -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Use header guards. Put the header guard in the first line in the header, -before the copyright or declarations. Use all-uppercase name for the header -guard. Derive the header guard name from the file name, and append _INCLUDED -to get a macro name. For example, core/log_io.h -> CORE_LOG_IO_H_INCLUDED. In -``.c`` (implementation) file, include the respective declaration header before all -other headers, to ensure that the header is self- sufficient. Header "header.h" -is self-sufficient if the following compiles without errors: - -.. code-block:: c - - #include "header.h" - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Allocating memory -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Prefer the supplied slab (salloc) and pool (palloc) allocators to malloc()/free() -for any performance-intensive or large memory allocations. Repetitive use of -malloc()/free() can lead to memory fragmentation and should therefore be avoided. - -Always free all allocated memory, even allocated at start-up. We aim at being -valgrind leak-check clean, and in most cases it's just as easy to free() the -allocated memory as it is to write a valgrind suppression. Freeing all allocated -memory is also dynamic-load friendly: assuming a plug-in can be dynamically loaded -and unloaded multiple times, reload should not lead to a memory leak. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Function naming -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Our convention is to use: - -* ``new``/``delete`` for functions which - allocate + initialize and destroy + deallocate an object, -* ``create``/``destroy`` for functions which initialize/destroy an object - but do not handle memory management, -* ``init``/``free`` for functions which initialize/destroy libraries and subsystems. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Other -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Select GNU C99 extensions are acceptable. It's OK to mix declarations and statements, -use true and false. - -The not-so-current list of all GCC C extensions can be found at: -http://gcc.gnu.org/onlinedocs/gcc-4.3.5/gcc/C-Extensions.html - -=========================================================== - Linux kernel coding style -=========================================================== - -This is a short document describing the preferred coding style for the -linux kernel. Coding style is very personal, and I won't _force_ my -views on anybody, but this is what goes for anything that I have to be -able to maintain, and I'd prefer it for most other things too. Please -at least consider the points made here. - -First off, I'd suggest printing out a copy of the GNU coding standards, -and NOT read it. Burn them, it's a great symbolic gesture. - -Anyway, here goes: - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 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 to -be 3. - -Rationale: The whole idea behind indentation is to clearly define where -a block of control starts and ends. Especially when you've been looking -at your screen for 20 straight hours, you'll find it a lot easier to see -how the indentation works if you have large indentations. - -Now, some people will claim that having 8-character indentations makes -the code move too far to the right, and makes it hard to read on a -80-character terminal screen. The answer to that is that if you need -more than 3 levels of indentation, you're screwed anyway, and should fix -your program. - -In short, 8-char indents make things easier to read, and have the added -benefit 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 is -to align the "switch" and its subordinate "case" labels in the same column -instead of "double-indenting" the "case" labels. e.g.: - -.. code-block:: c - - switch (suffix) { - case 'G': - case 'g': - mem <<= 30; - break; - case 'M': - case 'm': - mem <<= 20; - break; - case 'K': - case 'k': - mem <<= 10; - /* fall through */ - default: - break; - } - - -Don't put multiple statements on a single line unless you have -something to hide: - -.. code-block:: none - - if (condition) do_this; - do_something_everytime; - -Don't put multiple assignments on a single line either. Kernel coding style -is super simple. Avoid tricky expressions. - -Outside of comments, documentation and except in Kconfig, spaces are never -used for indentation, and the above example is deliberately broken. - -Get a decent editor and don't leave whitespace at the end of lines. - - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 2: Breaking long lines and strings -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Coding style is all about readability and maintainability using commonly -available tools. - -The limit on the length of lines is 80 columns, and this is a strongly -preferred limit. As for comments, the same limit of 80 columns is applied. - -Statements longer than 80 columns will be broken into sensible chunks. -Descendants are always substantially shorter than the parent and are placed -substantially to the right. The same applies to function headers with a long -argument list. Long strings are as well broken into shorter strings. The -only exception to this is where exceeding 80 columns significantly increases -readability and does not hide information. - -.. code-block:: c - - void fun(int a, int b, int c) - { - if (condition) - printk(KERN_WARNING "Warning this is a long printk with " - "3 parameters a: %u b: %u " - "c: %u \n", a, b, c); - else - next_statement; - } - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 3: Placing Braces and Spaces -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -The other issue that always comes up in C styling is the placement of -braces. Unlike the indent size, there are few technical reasons to -choose one placement strategy over the other, but the preferred way, as -shown to us by the prophets Kernighan and Ritchie, is to put the opening -brace last on the line, and put the closing brace first, thusly: - -.. code-block:: none - - if (x is true) { - we do y - } - -This applies to all non-function statement blocks (if, switch, for, -while, do). e.g.: - -.. code-block:: c - - switch (action) { - case KOBJ_ADD: - return "add"; - case KOBJ_REMOVE: - return "remove"; - case KOBJ_CHANGE: - return "change"; - default: - return NULL; - } - -However, there is one special case, namely functions: they have the -opening brace at the beginning of the next line, thus: - -.. code-block:: c - - int function(int x) - { - body of function; - } - -Heretic people all over the world have claimed that this inconsistency -is ... well ... inconsistent, but all right-thinking people know that -(a) K&R are _right_ and (b) K&R are right. Besides, functions are -special anyway (you can't nest them in C). - -Note that the closing brace is empty on a line of its own, _except_ in -the cases where it is followed by a continuation of the same statement, -ie a "while" in a do-statement or an "else" in an if-statement, like -this: - -.. code-block:: c - - do { - body of do-loop; - } while (condition); - -and - -.. code-block:: c - - if (x == y) { - .. - } else if (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 the -supply of new-lines on your screen is not a renewable resource (think -25-line terminal screens here), you have more empty lines to put -comments on. - -Do not unnecessarily use braces where a single statement will do. - -.. code-block:: c - - if (condition) - action(); - -This does not apply if one branch of a conditional statement is a single -statement. Use braces in both branches. - -.. code-block:: c - - if (condition) { - do_this(); - do_that(); - } else { - otherwise(); - } - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 3.1: Spaces -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Linux kernel style for use of spaces depends (mostly) on -function-versus-keyword usage. Use a space after (most) keywords. The -notable exceptions are sizeof, typeof, alignof, and __attribute__, which look -somewhat like functions (and are usually used with parentheses in Linux, -although they are not required in the language, as in: "sizeof info" after -"struct fileinfo info;" is declared). - -So use a space after these keywords: if, switch, case, for, do, while -but not with sizeof, typeof, alignof, or __attribute__. E.g., - -.. code-block:: c - - s = sizeof(struct file); - -Do not add spaces around (inside) parenthesized expressions. This example is -**bad**: - -.. code-block:: c - - s = sizeof( struct file ); - -When declaring pointer data or a function that returns a pointer type, the -preferred use of '*' is adjacent to the data name or function name and not -adjacent to the type name. Examples: - -.. code-block:: c - - char *linux_banner; - unsigned long long memparse(char *ptr, char **retptr); - char *match_strdup(substring_t *s); - -Use one space around (on each side of) most binary and ternary operators, -such as any of these: - - = + - < > * / % | & ^ <= >= == != ? : - -but no space after unary operators: - - & * + - ~ ! sizeof typeof alignof __attribute__ defined - -no space before the postfix increment & decrement unary operators: - - ++ -- - -no space after the prefix increment & decrement unary operators: - - ++ -- - -and no space around the '.' and "->" structure member operators. - -Do not leave trailing whitespace at the ends of lines. Some editors with -"smart" indentation will insert whitespace at the beginning of new lines as -appropriate, so you can start typing the next line of code right away. -However, some such editors do not remove the whitespace if you end up not -putting a line of code there, such as if you leave a blank line. As a result, -you end up with lines containing trailing whitespace. - -Git will warn you about patches that introduce trailing whitespace, and can -optionally strip the trailing whitespace for you; however, if applying a series -of patches, this may make later patches in the series fail by changing their -context lines. - - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 4: Naming -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -C is a Spartan language, and so should your naming be. Unlike Modula-2 -and Pascal programmers, C programmers do not use cute names like -ThisVariableIsATemporaryCounter. A C programmer would call that -variable "tmp", which is much easier to write, and not the least more -difficult to understand. - -HOWEVER, while mixed-case names are frowned upon, descriptive names for -global variables are a must. To call a global function "foo" is a -shooting offense. - -GLOBAL variables (to be used only if you _really_ need them) need to -have descriptive names, as do global functions. If you have a function -that counts the number of active users, you should call that -"count_active_users()" or similar, you should _not_ call it "cntusr()". - -Encoding the type of a function into the name (so-called Hungarian -notation) is brain damaged - the compiler knows the types anyway and can -check those, and it only confuses the programmer. No wonder MicroSoft -makes buggy programs. - -LOCAL variable names should be short, and to the point. If you have -some random integer loop counter, it should probably be called "i". -Calling it "loop_counter" is non-productive, if there is no chance of it -being mis-understood. Similarly, "tmp" can be just about any type of -variable that is used to hold a temporary value. - -If you are afraid to mix up your local variable names, you have another -problem, which is called the function-growth-hormone-imbalance syndrome. -See chapter 6 (Functions). - - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 5: Typedefs -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Please don't use things like "vps_t". - -It's a _mistake_ to use typedef for structures and pointers. When you see a - -.. code-block:: c - - vps_t a; - -in the source, what does it mean? - -In contrast, if it says - -.. code-block:: c - - struct virtual_container *a; - -you can actually tell what "a" is. - -Lots of people think that typedefs "help readability". Not so. They are -useful only for: - -(a) totally opaque objects (where the typedef is actively used to _hide_ - what the object is). - - Example: "pte_t" etc. opaque objects that you can only access using - the proper accessor functions. - - NOTE! Opaqueness and "accessor functions" are not good in themselves. - The reason we have them for things like pte_t etc. is that there - really is absolutely _zero_ portably accessible information there. - -(b) Clear integer types, where the abstraction _helps_ avoid confusion - whether it is "int" or "long". - - u8/u16/u32 are perfectly fine typedefs, although they fit into - category (d) better than here. - - NOTE! Again - there needs to be a _reason_ for this. If something is - "unsigned long", then there's no reason to do - - .. code-block:: c - - typedef unsigned long myflags_t; - - but if there is a clear reason for why it under certain circumstances - might be an "unsigned int" and under other configurations might be - "unsigned long", then by all means go ahead and use a typedef. - -(c) when you use sparse to literally create a _new_ type for - type-checking. - -(d) New types which are identical to standard C99 types, in certain - exceptional circumstances. - - Although it would only take a short amount of time for the eyes and - brain to become accustomed to the standard types like 'uint32_t', - some people object to their use anyway. - - Therefore, the Linux-specific 'u8/u16/u32/u64' types and their - signed equivalents which are identical to standard types are - permitted -- although they are not mandatory in new code of your - own. - - When editing existing code which already uses one or the other set - of types, you should conform to the existing choices in that code. - -(e) Types safe for use in userspace. - - In certain structures which are visible to userspace, we cannot - require C99 types and cannot use the 'u32' form above. Thus, we - use __u32 and similar types in all structures which are shared - with userspace. - -Maybe there are other cases too, but the rule should basically be to NEVER -EVER use a typedef unless you can clearly match one of those rules. - -In general, a pointer, or a struct that has elements that can reasonably -be directly accessed should **never** be a typedef. - - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 6: Functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Functions should be short and sweet, and do just one thing. They should -fit 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 the -complexity and indentation level of that function. So, if you have a -conceptually simple function that is just one long (but simple) -case-statement, where you have to do lots of small things for a lot of -different cases, it's OK to have a longer function. - -However, if you have a complex function, and you suspect that a -less-than-gifted first-year high-school student might not even -understand what the function is all about, you should adhere to the -maximum limits all the more closely. Use helper functions with -descriptive names (you can ask the compiler to in-line them if you think -it's performance-critical, and it will probably do a better job of it -than you would have done). - -Another measure of the function is the number of local variables. They -shouldn't exceed 5-10, or you're doing something wrong. Re-think the -function, and split it into smaller pieces. A human brain can -generally easily keep track of about 7 different things, anything more -and it gets confu/sed. You know you're brilliant, but maybe you'd like -to understand what you did 2 weeks from now. - -In source files, separate functions with one blank line. If the function is -exported, the EXPORT* macro for it should follow immediately after the closing -function brace line. E.g.: - -.. code-block:: c - - int system_is_up(void) - { - return system_state == SYSTEM_RUNNING; - } - EXPORT_SYMBOL(system_is_up); - -In function prototypes, include parameter names with their data types. -Although this is not required by the C language, it is preferred in Linux -because it is a simple way to add valuable information for the reader. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 7: Centralized exiting of functions -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Albeit deprecated by some people, the equivalent of the goto statement is -used frequently by compilers in form of the unconditional jump instruction. - -The goto statement comes in handy when a function exits from multiple -locations and some common work such as cleanup has to be done. - -The rationale is: - -- unconditional statements are easier to understand and follow -- nesting is reduced -- errors by not updating individual exit points when making - modifications are prevented -- saves the compiler work to optimize redundant code away ;) - -.. code-block:: c - - int fun(int a) - { - int result = 0; - char *buffer = kmalloc(SIZE); - - if (buffer == NULL) - return -ENOMEM; - - if (condition1) { - while (loop1) { - ... - } - result = 1; - goto out; - } - ... - out: - kfree(buffer); - return result; - } - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 8: Commenting -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Comments are good, but there is also a danger of over-commenting. NEVER -try to explain HOW your code works in a comment: it's much better to -write the code so that the _working_ is obvious, and it's a waste of -time to explain badly written code. -с -Generally, you want your comments to tell WHAT your code does, not HOW. -Also, try to avoid putting comments inside a function body: if the -function 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 make -small comments to note or warn about something particularly clever (or -ugly), but try to avoid excess. Instead, put the comments at the head -of the function, telling people what it does, and possibly WHY it does -it. - -When commenting the kernel API functions, please use the kernel-doc format. -See the files Documentation/kernel-doc-nano-HOWTO.txt and scripts/kernel-doc -for details. - -Linux style for comments is the C89 :code:`/* ... */`` style. -Don't use C99-style :code:`// ...` comments. - -The preferred style for long (multi-line) comments is: - -.. code-block:: c - - /* - * 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 derived -types. To this end, use just one data declaration per line (no commas for -multiple data declarations). This leaves you room for a small comment on each -item, explaining its use. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 9: You've made a mess of it -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -That's OK, we all do. You've probably been told by your long-time Unix -user helper that "GNU emacs" automatically formats the C sources for -you, and you've noticed that yes, it does do that, but the defaults it -uses are less than desirable (in fact, they are worse than random -typing - an infinite number of monkeys typing into GNU emacs would never -make a good program). - -So, you can either get rid of GNU emacs, or change it to use saner -values. To do the latter, you can stick the following in your .emacs file: - -.. code-block:: lisp - - (defun c-lineup-arglist-tabs-only (ignored) - "Line up argument lists by tabs, not spaces" - (let* ((anchor (c-langelem-pos c-syntactic-element)) - (column (c-langelem-2nd-pos c-syntactic-element)) - (offset (- (1+ column) anchor)) - (steps (floor offset c-basic-offset))) - (* (max steps 1) - c-basic-offset))) - - (add-hook 'c-mode-common-hook - (lambda () - ;; Add kernel style - (c-add-style - "linux-tabs-only" - '("linux" (c-offsets-alist - (arglist-cont-nonempty - c-lineup-gcc-asm-reg - c-lineup-arglist-tabs-only)))))) - - (add-hook 'c-mode-hook - (lambda () - (let ((filename (buffer-file-name))) - ;; Enable kernel mode for the appropriate files - (when (and filename - (string-match (expand-file-name "~/src/linux-trees") - filename)) - (setq indent-tabs-mode t) - (c-set-style "linux-tabs-only"))))) - -This will make emacs go better with the kernel coding style for C -files below ~/src/linux-trees. - -But even if you fail in getting emacs to do sane formatting, not -everything is lost: use "indent". - -Now, again, GNU indent has the same brain-dead settings that GNU emacs -has, 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 indent -recognize the authority of K&R (the GNU people aren't evil, they are -just severely misguided in this matter), so you just give indent the -options "-kr -i8" (stands for "K&R, 8 character indents"), or use -"scripts/Lindent", which indents in the latest style. - -"indent" has a lot of options, and especially when it comes to comment -re-formatting you may want to take a look at the man page. But -remember: "indent" is not a fix for bad programming. - - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 10: Kconfig configuration files -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -For all of the Kconfig* configuration files throughout the source tree, -the indentation is somewhat different. Lines under a "config" definition -are indented with one tab, while help text is indented an additional two -spaces. Example: - -.. code-block:: kconfig - - 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. - -Features that might still be considered unstable should be defined as -dependent on "EXPERIMENTAL": - -.. code-block:: kconfig - - config SLUB - depends on EXPERIMENTAL && !ARCH_USES_SLAB_PAGE_STRUCT - bool "SLUB (Unqueued Allocator)" - ... - -while seriously dangerous features (such as write support for certain -filesystems) should advertise this prominently in their prompt string: - -.. code-block:: kconfig - - config ADFS_FS_RW - bool "ADFS write support (DANGEROUS)" - depends on ADFS_FS - ... - -For full documentation on the configuration files, see the file -Documentation/kbuild/kconfig-language.txt. - - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 11: Data structures -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Data structures that have visibility outside the single-threaded -environment they are created and destroyed in should always have -reference counts. In the kernel, garbage collection doesn't exist (and -outside the kernel garbage collection is slow and inefficient), which -means that you absolutely _have_ to reference count all your uses. - -Reference counting means that you can avoid locking, and allows multiple -users to have access to the data structure in parallel - and not having -to worry about the structure suddenly going away from under them just -because they slept or did something else for a while. - -Note that locking is _not_ a replacement for reference counting. -Locking is used to keep data structures coherent, while reference -counting is a memory management technique. Usually both are needed, and -they are not to be confused with each other. - -Many data structures can indeed have two levels of reference counting, -when there are users of different "classes". The subclass count counts -the number of subclass users, and decrements the global count just once -when the subclass count goes to zero. - -Examples of this kind of "multi-level-reference-counting" can be found in -memory management ("struct mm_struct": mm_users and mm_count), and in -filesystem code ("struct super_block": s_count and s_active). - -Remember: if another thread can find your data structure, and you don't -have a reference count on it, you almost certainly have a bug. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 12: Macros, Enums and RTL -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Names of macros defining constants and labels in enums are capitalized. - -.. code-block:: c - - #define CONSTANT 0x12345 - -Enums are preferred when defining several related constants. - -CAPITALIZED macro names are appreciated but macros resembling functions -may 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: - -.. code-block:: c - - #define macrofun(a, b, c) \ - do { \ - if (a == 5) \ - do_this(b, c); \ - } while (0) - -Things to avoid when using macros: - -1. macros that affect control flow: - - .. code-block:: c - - #define FOO(x) \ - do { \ - if (blah(x) < 0) \ - return -EBUGGERED; \ - } while(0) - - is a _very_ bad idea. It looks like a function call but exits the "calling" - function; 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: - - .. code-block:: c - - #define FOO(val) bar(index, val) - - might look like a good thing, but it's confusing as hell when one reads the - code and it's prone to breakage from seemingly innocent changes. - -3. macros with arguments that are used as l-values: FOO(x) = y; will - bite you if somebody e.g. turns FOO into an inline function. - -4. forgetting about precedence: macros defining constants using expressions - must enclose the expression in parentheses. Beware of similar issues with - macros using parameters. - - .. code-block:: c - - #define CONSTANT 0x4000 - #define CONSTEXP (CONSTANT | 3) - - The cpp manual deals with macros exhaustively. The gcc internals manual also - covers RTL which is used frequently with assembly language in the kernel. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 13: Printing kernel messages -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Kernel developers like to be seen as literate. Do mind the spelling -of kernel messages to make a good impression. Do not use crippled -words like "dont"; use "do not" or "don't" instead. Make the messages -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 -which you should use to make sure messages are matched to the right device -and driver, and are tagged with the right level: dev_err(), dev_warn(), -dev_info(), and so forth. For messages that aren't associated with a -particular device, defines pr_debug() and pr_info(). - -Coming up with good debugging messages can be quite a challenge; and once -you have them, they can be a huge help for remote troubleshooting. Such -messages should be compiled out when the DEBUG symbol is not defined (that -is, by default they are not included). When you use dev_dbg() or pr_debug(), -that's automatic. Many subsystems have Kconfig options to turn on -DDEBUG. -A related convention uses VERBOSE_DEBUG to add dev_vdbg() messages to the -ones already enabled by DEBUG. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 14: Allocating memory -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -The kernel provides the following general purpose memory allocators: -kmalloc(), kzalloc(), kcalloc(), and vmalloc(). Please refer to the API -documentation for further information about them. - -The preferred form for passing a size of a struct is the following: - -.. code-block:: c - - p = kmalloc(sizeof(*p), ...); - -The alternative form where struct name is spelled out hurts readability and -introduces an opportunity for a bug when the pointer variable type is changed -but the corresponding sizeof that is passed to a memory allocator is not. - -Casting the return value which is a void pointer is redundant. The conversion -from void pointer to any other pointer type is guaranteed by the C programming -language. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 15: The inline disease -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -There appears to be a common misperception that gcc has a magic "make me -faster" speedup option called "inline". While the use of inlines can be -appropriate (for example as a means of replacing macros, see Chapter 12), it -very often is not. Abundant use of the inline keyword leads to a much bigger -kernel, which in turn slows the system as a whole down, due to a bigger -icache footprint for the CPU and simply because there is less memory -available for the pagecache. Just think about it; a pagecache miss causes a -disk seek, which easily takes 5 milliseconds. There are a LOT of cpu cycles -that can go into these 5 milliseconds. - -A reasonable rule of thumb is to not put inline at functions that have more -than 3 lines of code in them. An exception to this rule are the cases where -a parameter is known to be a compiletime constant, and as a result of this -constantness you *know* the compiler will be able to optimize most of your -function away at compile time. For a good example of this later case, see -the kmalloc() inline function. - -Often people argue that adding inline to functions that are static and used -only once is always a win since there is no space tradeoff. While this is -technically correct, gcc is capable of inlining these automatically without -help, and the maintenance issue of removing the inline when a second user -appears outweighs the potential value of the hint that tells gcc to do -something it would have done anyway. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 16: Function return values and names -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Functions can return values of many different kinds, and one of the -most common is a value indicating whether the function succeeded or -failed. Such a value can be represented as an error-code integer -(-Exxx = failure, 0 = success) or a "succeeded" boolean (0 = failure, -non-zero = success). - -Mixing up these two sorts of representations is a fertile source of -difficult-to-find bugs. If the C language included a strong distinction -between integers and booleans then the compiler would find these mistakes -for us... but it doesn't. To help prevent such bugs, always follow this -convention: - -:: - - If the name of a function is an action or an imperative command, - the function should return an error-code integer. If the name - is a predicate, the function should return a "succeeded" boolean. - -For example, "add work" is a command, and the add_work() function returns 0 -for success or -EBUSY for failure. In the same way, "PCI device present" is -a predicate, and the pci_dev_present() function returns 1 if it succeeds in -finding a matching device or 0 if it doesn't. - -All EXPORTed functions must respect this convention, and so should all -public functions. Private (static) functions need not, but it is -recommended that they do. - -Functions whose return value is the actual result of a computation, rather -than an indication of whether the computation succeeded, are not subject to -this rule. Generally they indicate failure by returning some out-of-range -result. Typical examples would be functions that return pointers; they use -NULL or the ERR_PTR mechanism to report failure. - - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 17: Don't re-invent the kernel macros -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -The header file include/linux/kernel.h contains a number of macros that -you should use, rather than explicitly coding some variant of them yourself. -For example, if you need to calculate the length of an array, take advantage -of the macro - -.. code-block:: c - - #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) - -Similarly, if you need to calculate the size of some structure member, use - -.. code-block:: c - - #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f)) - -There are also min() and max() macros that do strict type checking if you -need them. Feel free to peruse that header file to see what else is already -defined that you shouldn't reproduce in your code. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Chapter 18: Editor modelines and other cruft -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -Some editors can interpret configuration information embedded in source files, -indicated with special markers. For example, emacs interprets lines marked -like this: - -.. code-block:: none - - -*- mode: c -*- - -Or like this: - -.. code-block:: none - - /* - Local Variables: - compile-command: "gcc -DMAGIC_DEBUG_FLAG foo.c" - End: - */ - -Vim interprets markers that look like this: - -.. code-block:: none - - /* vim:set sw=8 noet */ - -Do not include any of these in source files. People have their own personal -editor configurations, and your source files should not override them. This -includes markers for indentation and mode configuration. People may use their -own custom mode, or may have some other magic method for making indentation -work correctly. - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Appendix I: References -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -* `The C Programming Language, Second Edition `_ - by Brian W. Kernighan and Dennis M. Ritchie. |br| - Prentice Hall, Inc., 1988. |br| - ISBN 0-13-110362-8 (paperback), 0-13-110370-9 (hardback). - -* `The Practice of Programming `_ - by Brian W. Kernighan and Rob Pike. |br| - Addison-Wesley, Inc., 1999. |br| - ISBN 0-201-61586-X. - -* `GNU manuals `_ - where in compliance with K&R and this text - for **cpp**, **gcc**, - **gcc internals** and **indent** - -* `WG14 International standardization workgroup for the programming - language C `_ - -* `Kernel CodingStyle, by greg@kroah.com at OLS 2002 - `_