8 The configuration database is a collection of configuration options
9 organized in a tree structure::
11 +- Code maturity level options
12 | +- Prompt for development and/or incomplete code/drivers
14 | +- Networking support
16 | +- BSD Process Accounting
18 +- Loadable module support
19 | +- Enable loadable module support
20 | +- Set version information on all module symbols
21 | +- Kernel module loader
24 Every entry has its own dependencies. These dependencies are used
25 to determine the visibility of an entry. Any child entry is only
26 visible if its parent entry is also visible.
31 Most entries define a config option; all other entries help to organize
32 them. A single configuration option is defined like this::
35 bool "Set version information on all module symbols"
38 Usually, modules have to be recompiled whenever you switch to a new
41 Every line starts with a key word and can be followed by multiple
42 arguments. "config" starts a new config entry. The following lines
43 define attributes for this config option. Attributes can be the type of
44 the config option, input prompt, dependencies, help text and default
45 values. A config option can be defined multiple times with the same
46 name, but every definition can have only a single input prompt and the
47 type must not conflict.
52 A menu entry can have a number of attributes. Not all of them are
53 applicable everywhere (see syntax).
55 - type definition: "bool"/"tristate"/"string"/"hex"/"int"
56 Every config option must have a type. There are only two basic types:
57 tristate and string; the other types are based on these two. The type
58 definition optionally accepts an input prompt, so these two examples
61 bool "Networking support"
66 prompt "Networking support"
68 - input prompt: "prompt" <prompt> ["if" <expr>]
69 Every menu entry can have at most one prompt, which is used to display
70 to the user. Optionally dependencies only for this prompt can be added
73 - default value: "default" <expr> ["if" <expr>]
74 A config option can have any number of default values. If multiple
75 default values are visible, only the first defined one is active.
76 Default values are not limited to the menu entry where they are
77 defined. This means the default can be defined somewhere else or be
78 overridden by an earlier definition.
79 The default value is only assigned to the config symbol if no other
80 value was set by the user (via the input prompt above). If an input
81 prompt is visible the default value is presented to the user and can
83 Optionally, dependencies only for this default value can be added with
86 The default value deliberately defaults to 'n' in order to avoid bloating the
87 build. With few exceptions, new config options should not change this. The
88 intent is for "make oldconfig" to add as little as possible to the config from
92 Things that merit "default y/m" include:
94 a) A new Kconfig option for something that used to always be built
95 should be "default y".
97 b) A new gatekeeping Kconfig option that hides/shows other Kconfig
98 options (but does not generate any code of its own), should be
99 "default y" so people will see those other options.
101 c) Sub-driver behavior or similar options for a driver that is
102 "default n". This allows you to provide sane defaults.
104 d) Hardware or infrastructure that everybody expects, such as CONFIG_NET
105 or CONFIG_BLOCK. These are rare exceptions.
107 - type definition + default value::
109 "def_bool"/"def_tristate" <expr> ["if" <expr>]
111 This is a shorthand notation for a type definition plus a value.
112 Optionally dependencies for this default value can be added with "if".
114 - dependencies: "depends on" <expr>
115 This defines a dependency for this menu entry. If multiple
116 dependencies are defined, they are connected with '&&'. Dependencies
117 are applied to all other options within this menu entry (which also
118 accept an "if" expression), so these two examples are equivalent::
129 - reverse dependencies: "select" <symbol> ["if" <expr>]
130 While normal dependencies reduce the upper limit of a symbol (see
131 below), reverse dependencies can be used to force a lower limit of
132 another symbol. The value of the current menu symbol is used as the
133 minimal value <symbol> can be set to. If <symbol> is selected multiple
134 times, the limit is set to the largest selection.
135 Reverse dependencies can only be used with boolean or tristate
139 select should be used with care. select will force
140 a symbol to a value without visiting the dependencies.
141 By abusing select you are able to select a symbol FOO even
142 if FOO depends on BAR that is not set.
143 In general use select only for non-visible symbols
144 (no prompts anywhere) and for symbols with no dependencies.
145 That will limit the usefulness but on the other hand avoid
146 the illegal configurations all over.
148 - weak reverse dependencies: "imply" <symbol> ["if" <expr>]
149 This is similar to "select" as it enforces a lower limit on another
150 symbol except that the "implied" symbol's value may still be set to n
151 from a direct dependency or with a visible prompt.
153 Given the following example::
163 The following values are possible:
165 === === ============= ==============
166 FOO BAR BAZ's default choice for BAZ
167 === === ============= ==============
172 === === ============= ==============
174 This is useful e.g. with multiple drivers that want to indicate their
175 ability to hook into a secondary subsystem while allowing the user to
176 configure that subsystem out without also having to unset these drivers.
178 - limiting menu display: "visible if" <expr>
179 This attribute is only applicable to menu blocks, if the condition is
180 false, the menu block is not displayed to the user (the symbols
181 contained there can still be selected by other symbols, though). It is
182 similar to a conditional "prompt" attribute for individual menu
183 entries. Default value of "visible" is true.
185 - numerical ranges: "range" <symbol> <symbol> ["if" <expr>]
186 This allows to limit the range of possible input values for int
187 and hex symbols. The user can only input a value which is larger than
188 or equal to the first symbol and smaller than or equal to the second
191 - help text: "help" or "---help---"
192 This defines a help text. The end of the help text is determined by
193 the indentation level, this means it ends at the first line which has
194 a smaller indentation than the first line of the help text.
195 "---help---" and "help" do not differ in behaviour, "---help---" is
196 used to help visually separate configuration logic from help within
197 the file as an aid to developers.
199 - misc options: "option" <symbol>[=<value>]
200 Various less common options can be defined via this option syntax,
201 which can modify the behaviour of the menu entry and its config
202 symbol. These options are currently possible:
205 This declares a list of default entries which can be used when
206 looking for the default configuration (which is used when the main
207 .config doesn't exists yet.)
210 This declares the symbol to be used as the MODULES symbol, which
211 enables the third modular state for all config symbols.
212 At most one symbol may have the "modules" option set.
215 This declares the symbol as one that should have the value y when
216 using "allnoconfig". Used for symbols that hide other symbols.
221 Dependencies define the visibility of a menu entry and can also reduce
222 the input range of tristate symbols. The tristate logic used in the
223 expressions uses one more state than normal boolean logic to express the
224 module state. Dependency expressions have the following syntax::
226 <expr> ::= <symbol> (1)
227 <symbol> '=' <symbol> (2)
228 <symbol> '!=' <symbol> (3)
229 <symbol1> '<' <symbol2> (4)
230 <symbol1> '>' <symbol2> (4)
231 <symbol1> '<=' <symbol2> (4)
232 <symbol1> '>=' <symbol2> (4)
235 <expr> '&&' <expr> (7)
236 <expr> '||' <expr> (8)
238 Expressions are listed in decreasing order of precedence.
240 (1) Convert the symbol into an expression. Boolean and tristate symbols
241 are simply converted into the respective expression values. All
242 other symbol types result in 'n'.
243 (2) If the values of both symbols are equal, it returns 'y',
245 (3) If the values of both symbols are equal, it returns 'n',
247 (4) If value of <symbol1> is respectively lower, greater, lower-or-equal,
248 or greater-or-equal than value of <symbol2>, it returns 'y',
250 (5) Returns the value of the expression. Used to override precedence.
251 (6) Returns the result of (2-/expr/).
252 (7) Returns the result of min(/expr/, /expr/).
253 (8) Returns the result of max(/expr/, /expr/).
255 An expression can have a value of 'n', 'm' or 'y' (or 0, 1, 2
256 respectively for calculations). A menu entry becomes visible when its
257 expression evaluates to 'm' or 'y'.
259 There are two types of symbols: constant and non-constant symbols.
260 Non-constant symbols are the most common ones and are defined with the
261 'config' statement. Non-constant symbols consist entirely of alphanumeric
262 characters or underscores.
263 Constant symbols are only part of expressions. Constant symbols are
264 always surrounded by single or double quotes. Within the quote, any
265 other character is allowed and the quotes can be escaped using '\'.
270 The position of a menu entry in the tree is determined in two ways. First
271 it can be specified explicitly::
273 menu "Network device support"
281 All entries within the "menu" ... "endmenu" block become a submenu of
282 "Network device support". All subentries inherit the dependencies from
283 the menu entry, e.g. this means the dependency "NET" is added to the
284 dependency list of the config option NETDEVICES.
286 The other way to generate the menu structure is done by analyzing the
287 dependencies. If a menu entry somehow depends on the previous entry, it
288 can be made a submenu of it. First, the previous (parent) symbol must
289 be part of the dependency list and then one of these two conditions
292 - the child entry must become invisible, if the parent is set to 'n'
293 - the child entry must only be visible, if the parent is visible::
296 bool "Enable loadable module support"
299 bool "Set version information on all module symbols"
302 comment "module support disabled"
305 MODVERSIONS directly depends on MODULES, this means it's only visible if
306 MODULES is different from 'n'. The comment on the other hand is only
307 visible when MODULES is set to 'n'.
313 The configuration file describes a series of menu entries, where every
314 line starts with a keyword (except help texts). The following keywords
325 The first five also start the definition of a menu entry.
331 This defines a config symbol <symbol> and accepts any of above
332 attributes as options.
335 "menuconfig" <symbol>
338 This is similar to the simple config entry above, but it also gives a
339 hint to front ends, that all suboptions should be displayed as a
340 separate list of options. To make sure all the suboptions will really
341 show up under the menuconfig entry and not outside of it, every item
342 from the <config options> list must depend on the menuconfig symbol.
343 In practice, this is achieved by using one of the next two constructs::
359 In the following examples (3) and (4), C1 and C2 still have the M
360 dependency, but will not appear under menuconfig M anymore, because
361 of C0, which doesn't depend on M::
386 This defines a choice group and accepts any of the above attributes as
387 options. A choice can only be of type bool or tristate. If no type is
388 specified for a choice, its type will be determined by the type of
389 the first choice element in the group or remain unknown if none of the
390 choice elements have a type specified, as well.
392 While a boolean choice only allows a single config entry to be
393 selected, a tristate choice also allows any number of config entries
394 to be set to 'm'. This can be used if multiple drivers for a single
395 hardware exists and only a single driver can be compiled/loaded into
396 the kernel, but all drivers can be compiled as modules.
398 A choice accepts another option "optional", which allows to set the
399 choice to 'n' and no entry needs to be selected.
400 If no [symbol] is associated with a choice, then you can not have multiple
401 definitions of that choice. If a [symbol] is associated to the choice,
402 then you may define the same choice (i.e. with the same entries) in another
410 This defines a comment which is displayed to the user during the
411 configuration process and is also echoed to the output files. The only
412 possible options are dependencies.
421 This defines a menu block, see "Menu structure" above for more
422 information. The only possible options are dependencies and "visible"
431 This defines an if block. The dependency expression <expr> is appended
432 to all enclosed menu entries.
438 This reads the specified configuration file. This file is always parsed.
444 This sets the config program's title bar if the config program chooses
445 to use it. It should be placed at the top of the configuration, before any
448 '#' Kconfig source file comment:
450 An unquoted '#' character anywhere in a source file line indicates
451 the beginning of a source file comment. The remainder of that line
457 This is a collection of Kconfig tips, most of which aren't obvious at
458 first glance and most of which have become idioms in several Kconfig
461 Adding common features and make the usage configurable
462 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
463 It is a common idiom to implement a feature/functionality that are
464 relevant for some architectures but not all.
465 The recommended way to do so is to use a config variable named HAVE_*
466 that is defined in a common Kconfig file and selected by the relevant
468 An example is the generic IOMAP functionality.
470 We would in lib/Kconfig see::
472 # Generic IOMAP is used to ...
473 config HAVE_GENERIC_IOMAP
476 depends on HAVE_GENERIC_IOMAP && FOO
478 And in lib/Makefile we would see::
480 obj-$(CONFIG_GENERIC_IOMAP) += iomap.o
482 For each architecture using the generic IOMAP functionality we would see::
486 select HAVE_GENERIC_IOMAP
489 Note: we use the existing config option and avoid creating a new
490 config variable to select HAVE_GENERIC_IOMAP.
492 Note: the use of the internal config variable HAVE_GENERIC_IOMAP, it is
493 introduced to overcome the limitation of select which will force a
494 config option to 'y' no matter the dependencies.
495 The dependencies are moved to the symbol GENERIC_IOMAP and we avoid the
496 situation where select forces a symbol equals to 'y'.
498 Adding features that need compiler support
499 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
501 There are several features that need compiler support. The recommended way
502 to describe the dependency on the compiler feature is to use "depends on"
503 followed by a test macro::
505 config STACKPROTECTOR
506 bool "Stack Protector buffer overflow detection"
507 depends on $(cc-option,-fstack-protector)
510 If you need to expose a compiler capability to makefiles and/or C source files,
511 `CC_HAS_` is the recommended prefix for the config option::
513 config CC_HAS_STACKPROTECTOR_NONE
514 def_bool $(cc-option,-fno-stack-protector)
518 To restrict a component build to module-only, qualify its config symbol
519 with "depends on m". E.g.::
524 limits FOO to module (=m) or disabled (=n).
526 Kconfig recursive dependency limitations
527 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
529 If you've hit the Kconfig error: "recursive dependency detected" you've run
530 into a recursive dependency issue with Kconfig, a recursive dependency can be
531 summarized as a circular dependency. The kconfig tools need to ensure that
532 Kconfig files comply with specified configuration requirements. In order to do
533 that kconfig must determine the values that are possible for all Kconfig
534 symbols, this is currently not possible if there is a circular relation
535 between two or more Kconfig symbols. For more details refer to the "Simple
536 Kconfig recursive issue" subsection below. Kconfig does not do recursive
537 dependency resolution; this has a few implications for Kconfig file writers.
538 We'll first explain why this issues exists and then provide an example
539 technical limitation which this brings upon Kconfig developers. Eager
540 developers wishing to try to address this limitation should read the next
543 Simple Kconfig recursive issue
544 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
546 Read: Documentation/kbuild/Kconfig.recursion-issue-01
550 make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.recursion-issue-01 allnoconfig
552 Cumulative Kconfig recursive issue
553 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
555 Read: Documentation/kbuild/Kconfig.recursion-issue-02
559 make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.recursion-issue-02 allnoconfig
561 Practical solutions to kconfig recursive issue
562 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
564 Developers who run into the recursive Kconfig issue have two options
565 at their disposal. We document them below and also provide a list of
566 historical issues resolved through these different solutions.
568 a) Remove any superfluous "select FOO" or "depends on FOO"
569 b) Match dependency semantics:
571 b1) Swap all "select FOO" to "depends on FOO" or,
573 b2) Swap all "depends on FOO" to "select FOO"
575 The resolution to a) can be tested with the sample Kconfig file
576 Documentation/kbuild/Kconfig.recursion-issue-01 through the removal
577 of the "select CORE" from CORE_BELL_A_ADVANCED as that is implicit already
578 since CORE_BELL_A depends on CORE. At times it may not be possible to remove
579 some dependency criteria, for such cases you can work with solution b).
581 The two different resolutions for b) can be tested in the sample Kconfig file
582 Documentation/kbuild/Kconfig.recursion-issue-02.
584 Below is a list of examples of prior fixes for these types of recursive issues;
585 all errors appear to involve one or more select's and one or more "depends on".
587 ============ ===================================
589 ============ ===================================
590 06b718c01208 select A -> depends on A
591 c22eacfe82f9 depends on A -> depends on B
592 6a91e854442c select A -> depends on A
593 118c565a8f2e select A -> select B
594 f004e5594705 select A -> depends on A
595 c7861f37b4c6 depends on A -> (null)
596 80c69915e5fb select A -> (null) (1)
597 c2218e26c0d0 select A -> depends on A (1)
598 d6ae99d04e1c select A -> depends on A
599 95ca19cf8cbf select A -> depends on A
600 8f057d7bca54 depends on A -> (null)
601 8f057d7bca54 depends on A -> select A
602 a0701f04846e select A -> depends on A
603 0c8b92f7f259 depends on A -> (null)
604 e4e9e0540928 select A -> depends on A (2)
605 7453ea886e87 depends on A > (null) (1)
606 7b1fff7e4fdf select A -> depends on A
607 86c747d2a4f0 select A -> depends on A
608 d9f9ab51e55e select A -> depends on A
609 0c51a4d8abd6 depends on A -> select A (3)
610 e98062ed6dc4 select A -> depends on A (3)
611 91e5d284a7f1 select A -> (null)
612 ============ ===================================
614 (1) Partial (or no) quote of error.
615 (2) That seems to be the gist of that fix.
621 Work on kconfig is welcomed on both areas of clarifying semantics and on
622 evaluating the use of a full SAT solver for it. A full SAT solver can be
623 desirable to enable more complex dependency mappings and / or queries,
624 for instance on possible use case for a SAT solver could be that of handling
625 the current known recursive dependency issues. It is not known if this would
626 address such issues but such evaluation is desirable. If support for a full SAT
627 solver proves too complex or that it cannot address recursive dependency issues
628 Kconfig should have at least clear and well defined semantics which also
629 addresses and documents limitations or requirements such as the ones dealing
630 with recursive dependencies.
632 Further work on both of these areas is welcomed on Kconfig. We elaborate
633 on both of these in the next two subsections.
638 The use of Kconfig is broad, Linux is now only one of Kconfig's users:
639 one study has completed a broad analysis of Kconfig use in 12 projects [0]_.
640 Despite its widespread use, and although this document does a reasonable job
641 in documenting basic Kconfig syntax a more precise definition of Kconfig
642 semantics is welcomed. One project deduced Kconfig semantics through
643 the use of the xconfig configurator [1]_. Work should be done to confirm if
644 the deduced semantics matches our intended Kconfig design goals.
646 Having well defined semantics can be useful for tools for practical
647 evaluation of depenencies, for instance one such use known case was work to
648 express in boolean abstraction of the inferred semantics of Kconfig to
649 translate Kconfig logic into boolean formulas and run a SAT solver on this to
650 find dead code / features (always inactive), 114 dead features were found in
651 Linux using this methodology [1]_ (Section 8: Threats to validity).
653 Confirming this could prove useful as Kconfig stands as one of the the leading
654 industrial variability modeling languages [1]_ [2]_. Its study would help
655 evaluate practical uses of such languages, their use was only theoretical
656 and real world requirements were not well understood. As it stands though
657 only reverse engineering techniques have been used to deduce semantics from
658 variability modeling languages such as Kconfig [3]_.
660 .. [0] http://www.eng.uwaterloo.ca/~shshe/kconfig_semantics.pdf
661 .. [1] http://gsd.uwaterloo.ca/sites/default/files/vm-2013-berger.pdf
662 .. [2] http://gsd.uwaterloo.ca/sites/default/files/ase241-berger_0.pdf
663 .. [3] http://gsd.uwaterloo.ca/sites/default/files/icse2011.pdf
665 Full SAT solver for Kconfig
666 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
668 Although SAT solvers [4]_ haven't yet been used by Kconfig directly, as noted
669 in the previous subsection, work has been done however to express in boolean
670 abstraction the inferred semantics of Kconfig to translate Kconfig logic into
671 boolean formulas and run a SAT solver on it [5]_. Another known related project
672 is CADOS [6]_ (former VAMOS [7]_) and the tools, mainly undertaker [8]_, which
673 has been introduced first with [9]_. The basic concept of undertaker is to
674 exract variability models from Kconfig, and put them together with a
675 propositional formula extracted from CPP #ifdefs and build-rules into a SAT
676 solver in order to find dead code, dead files, and dead symbols. If using a SAT
677 solver is desirable on Kconfig one approach would be to evaluate repurposing
678 such efforts somehow on Kconfig. There is enough interest from mentors of
679 existing projects to not only help advise how to integrate this work upstream
680 but also help maintain it long term. Interested developers should visit:
682 http://kernelnewbies.org/KernelProjects/kconfig-sat
684 .. [4] http://www.cs.cornell.edu/~sabhar/chapters/SATSolvers-KR-Handbook.pdf
685 .. [5] http://gsd.uwaterloo.ca/sites/default/files/vm-2013-berger.pdf
686 .. [6] https://cados.cs.fau.de
687 .. [7] https://vamos.cs.fau.de
688 .. [8] https://undertaker.cs.fau.de
689 .. [9] https://www4.cs.fau.de/Publications/2011/tartler_11_eurosys.pdf