1.Intro.rst 14 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268
  1. .. _development_process_intro:
  2. Introduction
  3. ============
  4. Executive summary
  5. -----------------
  6. The rest of this section covers the scope of the kernel development process
  7. and the kinds of frustrations that developers and their employers can
  8. encounter there. There are a great many reasons why kernel code should be
  9. merged into the official ("mainline") kernel, including automatic
  10. availability to users, community support in many forms, and the ability to
  11. influence the direction of kernel development. Code contributed to the
  12. Linux kernel must be made available under a GPL-compatible license.
  13. :ref:`development_process` introduces the development process, the kernel
  14. release cycle, and the mechanics of the merge window. The various phases in
  15. the patch development, review, and merging cycle are covered. There is some
  16. discussion of tools and mailing lists. Developers wanting to get started
  17. with kernel development are encouraged to track down and fix bugs as an
  18. initial exercise.
  19. :ref:`development_early_stage` covers early-stage project planning, with an
  20. emphasis on involving the development community as soon as possible.
  21. :ref:`development_coding` is about the coding process; several pitfalls which
  22. have been encountered by other developers are discussed. Some requirements for
  23. patches are covered, and there is an introduction to some of the tools
  24. which can help to ensure that kernel patches are correct.
  25. :ref:`development_posting` talks about the process of posting patches for
  26. review. To be taken seriously by the development community, patches must be
  27. properly formatted and described, and they must be sent to the right place.
  28. Following the advice in this section should help to ensure the best
  29. possible reception for your work.
  30. :ref:`development_followthrough` covers what happens after posting patches; the
  31. job is far from done at that point. Working with reviewers is a crucial part
  32. of the development process; this section offers a number of tips on how to
  33. avoid problems at this important stage. Developers are cautioned against
  34. assuming that the job is done when a patch is merged into the mainline.
  35. :ref:`development_advancedtopics` introduces a couple of "advanced" topics:
  36. managing patches with git and reviewing patches posted by others.
  37. :ref:`development_conclusion` concludes the document with pointers to sources
  38. for more information on kernel development.
  39. What this document is about
  40. ---------------------------
  41. The Linux kernel, at over 8 million lines of code and well over 1000
  42. contributors to each release, is one of the largest and most active free
  43. software projects in existence. Since its humble beginning in 1991, this
  44. kernel has evolved into a best-of-breed operating system component which
  45. runs on pocket-sized digital music players, desktop PCs, the largest
  46. supercomputers in existence, and all types of systems in between. It is a
  47. robust, efficient, and scalable solution for almost any situation.
  48. With the growth of Linux has come an increase in the number of developers
  49. (and companies) wishing to participate in its development. Hardware
  50. vendors want to ensure that Linux supports their products well, making
  51. those products attractive to Linux users. Embedded systems vendors, who
  52. use Linux as a component in an integrated product, want Linux to be as
  53. capable and well-suited to the task at hand as possible. Distributors and
  54. other software vendors who base their products on Linux have a clear
  55. interest in the capabilities, performance, and reliability of the Linux
  56. kernel. And end users, too, will often wish to change Linux to make it
  57. better suit their needs.
  58. One of the most compelling features of Linux is that it is accessible to
  59. these developers; anybody with the requisite skills can improve Linux and
  60. influence the direction of its development. Proprietary products cannot
  61. offer this kind of openness, which is a characteristic of the free software
  62. process. But, if anything, the kernel is even more open than most other
  63. free software projects. A typical three-month kernel development cycle can
  64. involve over 1000 developers working for more than 100 different companies
  65. (or for no company at all).
  66. Working with the kernel development community is not especially hard. But,
  67. that notwithstanding, many potential contributors have experienced
  68. difficulties when trying to do kernel work. The kernel community has
  69. evolved its own distinct ways of operating which allow it to function
  70. smoothly (and produce a high-quality product) in an environment where
  71. thousands of lines of code are being changed every day. So it is not
  72. surprising that Linux kernel development process differs greatly from
  73. proprietary development methods.
  74. The kernel's development process may come across as strange and
  75. intimidating to new developers, but there are good reasons and solid
  76. experience behind it. A developer who does not understand the kernel
  77. community's ways (or, worse, who tries to flout or circumvent them) will
  78. have a frustrating experience in store. The development community, while
  79. being helpful to those who are trying to learn, has little time for those
  80. who will not listen or who do not care about the development process.
  81. It is hoped that those who read this document will be able to avoid that
  82. frustrating experience. There is a lot of material here, but the effort
  83. involved in reading it will be repaid in short order. The development
  84. community is always in need of developers who will help to make the kernel
  85. better; the following text should help you - or those who work for you -
  86. join our community.
  87. Credits
  88. -------
  89. This document was written by Jonathan Corbet, corbet@lwn.net. It has been
  90. improved by comments from Johannes Berg, James Berry, Alex Chiang, Roland
  91. Dreier, Randy Dunlap, Jake Edge, Jiri Kosina, Matt Mackall, Arthur Marsh,
  92. Amanda McPherson, Andrew Morton, Andrew Price, Tsugikazu Shibata, and
  93. Jochen Voß.
  94. This work was supported by the Linux Foundation; thanks especially to
  95. Amanda McPherson, who saw the value of this effort and made it all happen.
  96. The importance of getting code into the mainline
  97. ------------------------------------------------
  98. Some companies and developers occasionally wonder why they should bother
  99. learning how to work with the kernel community and get their code into the
  100. mainline kernel (the "mainline" being the kernel maintained by Linus
  101. Torvalds and used as a base by Linux distributors). In the short term,
  102. contributing code can look like an avoidable expense; it seems easier to
  103. just keep the code separate and support users directly. The truth of the
  104. matter is that keeping code separate ("out of tree") is a false economy.
  105. As a way of illustrating the costs of out-of-tree code, here are a few
  106. relevant aspects of the kernel development process; most of these will be
  107. discussed in greater detail later in this document. Consider:
  108. - Code which has been merged into the mainline kernel is available to all
  109. Linux users. It will automatically be present on all distributions which
  110. enable it. There is no need for driver disks, downloads, or the hassles
  111. of supporting multiple versions of multiple distributions; it all just
  112. works, for the developer and for the user. Incorporation into the
  113. mainline solves a large number of distribution and support problems.
  114. - While kernel developers strive to maintain a stable interface to user
  115. space, the internal kernel API is in constant flux. The lack of a stable
  116. internal interface is a deliberate design decision; it allows fundamental
  117. improvements to be made at any time and results in higher-quality code.
  118. But one result of that policy is that any out-of-tree code requires
  119. constant upkeep if it is to work with new kernels. Maintaining
  120. out-of-tree code requires significant amounts of work just to keep that
  121. code working.
  122. Code which is in the mainline, instead, does not require this work as the
  123. result of a simple rule requiring any developer who makes an API change
  124. to also fix any code that breaks as the result of that change. So code
  125. which has been merged into the mainline has significantly lower
  126. maintenance costs.
  127. - Beyond that, code which is in the kernel will often be improved by other
  128. developers. Surprising results can come from empowering your user
  129. community and customers to improve your product.
  130. - Kernel code is subjected to review, both before and after merging into
  131. the mainline. No matter how strong the original developer's skills are,
  132. this review process invariably finds ways in which the code can be
  133. improved. Often review finds severe bugs and security problems. This is
  134. especially true for code which has been developed in a closed
  135. environment; such code benefits strongly from review by outside
  136. developers. Out-of-tree code is lower-quality code.
  137. - Participation in the development process is your way to influence the
  138. direction of kernel development. Users who complain from the sidelines
  139. are heard, but active developers have a stronger voice - and the ability
  140. to implement changes which make the kernel work better for their needs.
  141. - When code is maintained separately, the possibility that a third party
  142. will contribute a different implementation of a similar feature always
  143. exists. Should that happen, getting your code merged will become much
  144. harder - to the point of impossibility. Then you will be faced with the
  145. unpleasant alternatives of either (1) maintaining a nonstandard feature
  146. out of tree indefinitely, or (2) abandoning your code and migrating your
  147. users over to the in-tree version.
  148. - Contribution of code is the fundamental action which makes the whole
  149. process work. By contributing your code you can add new functionality to
  150. the kernel and provide capabilities and examples which are of use to
  151. other kernel developers. If you have developed code for Linux (or are
  152. thinking about doing so), you clearly have an interest in the continued
  153. success of this platform; contributing code is one of the best ways to
  154. help ensure that success.
  155. All of the reasoning above applies to any out-of-tree kernel code,
  156. including code which is distributed in proprietary, binary-only form.
  157. There are, however, additional factors which should be taken into account
  158. before considering any sort of binary-only kernel code distribution. These
  159. include:
  160. - The legal issues around the distribution of proprietary kernel modules
  161. are cloudy at best; quite a few kernel copyright holders believe that
  162. most binary-only modules are derived products of the kernel and that, as
  163. a result, their distribution is a violation of the GNU General Public
  164. license (about which more will be said below). Your author is not a
  165. lawyer, and nothing in this document can possibly be considered to be
  166. legal advice. The true legal status of closed-source modules can only be
  167. determined by the courts. But the uncertainty which haunts those modules
  168. is there regardless.
  169. - Binary modules greatly increase the difficulty of debugging kernel
  170. problems, to the point that most kernel developers will not even try. So
  171. the distribution of binary-only modules will make it harder for your
  172. users to get support from the community.
  173. - Support is also harder for distributors of binary-only modules, who must
  174. provide a version of the module for every distribution and every kernel
  175. version they wish to support. Dozens of builds of a single module can
  176. be required to provide reasonably comprehensive coverage, and your users
  177. will have to upgrade your module separately every time they upgrade their
  178. kernel.
  179. - Everything that was said above about code review applies doubly to
  180. closed-source code. Since this code is not available at all, it cannot
  181. have been reviewed by the community and will, beyond doubt, have serious
  182. problems.
  183. Makers of embedded systems, in particular, may be tempted to disregard much
  184. of what has been said in this section in the belief that they are shipping
  185. a self-contained product which uses a frozen kernel version and requires no
  186. more development after its release. This argument misses the value of
  187. widespread code review and the value of allowing your users to add
  188. capabilities to your product. But these products, too, have a limited
  189. commercial life, after which a new version must be released. At that
  190. point, vendors whose code is in the mainline and well maintained will be
  191. much better positioned to get the new product ready for market quickly.
  192. Licensing
  193. ---------
  194. Code is contributed to the Linux kernel under a number of licenses, but all
  195. code must be compatible with version 2 of the GNU General Public License
  196. (GPLv2), which is the license covering the kernel distribution as a whole.
  197. In practice, that means that all code contributions are covered either by
  198. GPLv2 (with, optionally, language allowing distribution under later
  199. versions of the GPL) or the three-clause BSD license. Any contributions
  200. which are not covered by a compatible license will not be accepted into the
  201. kernel.
  202. Copyright assignments are not required (or requested) for code contributed
  203. to the kernel. All code merged into the mainline kernel retains its
  204. original ownership; as a result, the kernel now has thousands of owners.
  205. One implication of this ownership structure is that any attempt to change
  206. the licensing of the kernel is doomed to almost certain failure. There are
  207. few practical scenarios where the agreement of all copyright holders could
  208. be obtained (or their code removed from the kernel). So, in particular,
  209. there is no prospect of a migration to version 3 of the GPL in the
  210. foreseeable future.
  211. It is imperative that all code contributed to the kernel be legitimately
  212. free software. For that reason, code from anonymous (or pseudonymous)
  213. contributors will not be accepted. All contributors are required to "sign
  214. off" on their code, stating that the code can be distributed with the
  215. kernel under the GPL. Code which has not been licensed as free software by
  216. its owner, or which risks creating copyright-related problems for the
  217. kernel (such as code which derives from reverse-engineering efforts lacking
  218. proper safeguards) cannot be contributed.
  219. Questions about copyright-related issues are common on Linux development
  220. mailing lists. Such questions will normally receive no shortage of
  221. answers, but one should bear in mind that the people answering those
  222. questions are not lawyers and cannot provide legal advice. If you have
  223. legal questions relating to Linux source code, there is no substitute for
  224. talking with a lawyer who understands this field. Relying on answers
  225. obtained on technical mailing lists is a risky affair.