| xj | b04a402 | 2021-11-25 15:01:52 +0800 | [diff] [blame] | 1 | # Cumulative Kconfig recursive issue | 
|  | 2 | # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | 
|  | 3 | # | 
|  | 4 | # Test with: | 
|  | 5 | # | 
|  | 6 | # make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.recursion-issue-02 allnoconfig | 
|  | 7 | # | 
|  | 8 | # The recursive limitations with Kconfig has some non intuitive implications on | 
|  | 9 | # kconfig sematics which are documented here. One known practical implication | 
|  | 10 | # of the recursive limitation is that drivers cannot negate features from other | 
|  | 11 | # drivers if they share a common core requirement and use disjoint semantics to | 
|  | 12 | # annotate those requirements, ie, some drivers use "depends on" while others | 
|  | 13 | # use "select". For instance it means if a driver A and driver B share the same | 
|  | 14 | # core requirement, and one uses "select" while the other uses "depends on" to | 
|  | 15 | # annotate this, all features that driver A selects cannot now be negated by | 
|  | 16 | # driver B. | 
|  | 17 | # | 
|  | 18 | # A perhaps not so obvious implication of this is that, if semantics on these | 
|  | 19 | # core requirements are not carefully synced, as drivers evolve features | 
|  | 20 | # they select or depend on end up becoming shared requirements which cannot be | 
|  | 21 | # negated by other drivers. | 
|  | 22 | # | 
|  | 23 | # The example provided in Documentation/kbuild/Kconfig.recursion-issue-02 | 
|  | 24 | # describes a simple driver core layout of example features a kernel might | 
|  | 25 | # have. Let's assume we have some CORE functionality, then the kernel has a | 
|  | 26 | # series of bells and whistles it desires to implement, its not so advanced so | 
|  | 27 | # it only supports bells at this time: CORE_BELL_A and CORE_BELL_B. If | 
|  | 28 | # CORE_BELL_A has some advanced feature CORE_BELL_A_ADVANCED which selects | 
|  | 29 | # CORE_BELL_A then CORE_BELL_A ends up becoming a common BELL feature which | 
|  | 30 | # other bells in the system cannot negate. The reason for this issue is | 
|  | 31 | # due to the disjoint use of semantics on expressing each bell's relationship | 
|  | 32 | # with CORE, one uses "depends on" while the other uses "select". Another | 
|  | 33 | # more important reason is that kconfig does not check for dependencies listed | 
|  | 34 | # under 'select' for a symbol, when such symbols are selected kconfig them | 
|  | 35 | # as mandatory required symbols. For more details on the heavy handed nature | 
|  | 36 | # of select refer to Documentation/kbuild/Kconfig.select-break | 
|  | 37 | # | 
|  | 38 | # To fix this the "depends on CORE" must be changed to "select CORE", or the | 
|  | 39 | # "select CORE" must be changed to "depends on CORE". | 
|  | 40 | # | 
|  | 41 | # For an example real world scenario issue refer to the attempt to remove | 
|  | 42 | # "select FW_LOADER" [0], in the end the simple alternative solution to this | 
|  | 43 | # problem consisted on matching semantics with newly introduced features. | 
|  | 44 | # | 
|  | 45 | # [0] http://lkml.kernel.org/r/1432241149-8762-1-git-send-email-mcgrof@do-not-panic.com | 
|  | 46 |  | 
|  | 47 | mainmenu "Simple example to demo cumulative kconfig recursive dependency implication" | 
|  | 48 |  | 
|  | 49 | config CORE | 
|  | 50 | tristate | 
|  | 51 |  | 
|  | 52 | config CORE_BELL_A | 
|  | 53 | tristate | 
|  | 54 | depends on CORE | 
|  | 55 |  | 
|  | 56 | config CORE_BELL_A_ADVANCED | 
|  | 57 | tristate | 
|  | 58 | select CORE_BELL_A | 
|  | 59 |  | 
|  | 60 | config CORE_BELL_B | 
|  | 61 | tristate | 
|  | 62 | depends on !CORE_BELL_A | 
|  | 63 | select CORE |