stable-kernel-rules.rst 8.8 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234
  1. .. _stable_kernel_rules:
  2. Everything you ever wanted to know about Linux -stable releases
  3. ===============================================================
  4. Rules on what kind of patches are accepted, and which ones are not, into the
  5. "-stable" tree:
  6. - It or an equivalent fix must already exist in Linux mainline (upstream).
  7. - It must be obviously correct and tested.
  8. - It cannot be bigger than 100 lines, with context.
  9. - It must follow the
  10. :ref:`Documentation/process/submitting-patches.rst <submittingpatches>`
  11. rules.
  12. - It must either fix a real bug that bothers people or just add a device ID.
  13. To elaborate on the former:
  14. - It fixes a problem like an oops, a hang, data corruption, a real security
  15. issue, a hardware quirk, a build error (but not for things marked
  16. CONFIG_BROKEN), or some "oh, that's not good" issue.
  17. - Serious issues as reported by a user of a distribution kernel may also
  18. be considered if they fix a notable performance or interactivity issue.
  19. As these fixes are not as obvious and have a higher risk of a subtle
  20. regression they should only be submitted by a distribution kernel
  21. maintainer and include an addendum linking to a bugzilla entry if it
  22. exists and additional information on the user-visible impact.
  23. - No "This could be a problem..." type of things like a "theoretical race
  24. condition", unless an explanation of how the bug can be exploited is also
  25. provided.
  26. - No "trivial" fixes without benefit for users (spelling changes, whitespace
  27. cleanups, etc).
  28. Procedure for submitting patches to the -stable tree
  29. ----------------------------------------------------
  30. .. note::
  31. Security patches should not be handled (solely) by the -stable review
  32. process but should follow the procedures in
  33. :ref:`Documentation/process/security-bugs.rst <securitybugs>`.
  34. There are three options to submit a change to -stable trees:
  35. 1. Add a 'stable tag' to the description of a patch you then submit for
  36. mainline inclusion.
  37. 2. Ask the stable team to pick up a patch already mainlined.
  38. 3. Submit a patch to the stable team that is equivalent to a change already
  39. mainlined.
  40. The sections below describe each of the options in more detail.
  41. :ref:`option_1` is **strongly** preferred, it is the easiest and most common.
  42. :ref:`option_2` is mainly meant for changes where backporting was not considered
  43. at the time of submission. :ref:`option_3` is an alternative to the two earlier
  44. options for cases where a mainlined patch needs adjustments to apply in older
  45. series (for example due to API changes).
  46. When using option 2 or 3 you can ask for your change to be included in specific
  47. stable series. When doing so, ensure the fix or an equivalent is applicable,
  48. submitted, or already present in all newer stable trees still supported. This is
  49. meant to prevent regressions that users might later encounter on updating, if
  50. e.g. a fix merged for 5.19-rc1 would be backported to 5.10.y, but not to 5.15.y.
  51. .. _option_1:
  52. Option 1
  53. ********
  54. To have a patch you submit for mainline inclusion later automatically picked up
  55. for stable trees, add this tag in the sign-off area::
  56. Cc: stable@vger.kernel.org
  57. Use ``Cc: stable@kernel.org`` instead when fixing unpublished vulnerabilities:
  58. it reduces the chance of accidentally exposing the fix to the public by way of
  59. 'git send-email', as mails sent to that address are not delivered anywhere.
  60. Once the patch is mainlined it will be applied to the stable tree without
  61. anything else needing to be done by the author or subsystem maintainer.
  62. To send additional instructions to the stable team, use a shell-style inline
  63. comment to pass arbitrary or predefined notes:
  64. * Specify any additional patch prerequisites for cherry picking::
  65. Cc: <stable@vger.kernel.org> # 3.3.x: a1f84a3: sched: Check for idle
  66. Cc: <stable@vger.kernel.org> # 3.3.x: 1b9508f: sched: Rate-limit newidle
  67. Cc: <stable@vger.kernel.org> # 3.3.x: fd21073: sched: Fix affinity logic
  68. Cc: <stable@vger.kernel.org> # 3.3.x
  69. Signed-off-by: Ingo Molnar <mingo@elte.hu>
  70. The tag sequence has the meaning of::
  71. git cherry-pick a1f84a3
  72. git cherry-pick 1b9508f
  73. git cherry-pick fd21073
  74. git cherry-pick <this commit>
  75. Note that for a patch series, you do not have to list as prerequisites the
  76. patches present in the series itself. For example, if you have the following
  77. patch series::
  78. patch1
  79. patch2
  80. where patch2 depends on patch1, you do not have to list patch1 as
  81. prerequisite of patch2 if you have already marked patch1 for stable
  82. inclusion.
  83. * Point out kernel version prerequisites::
  84. Cc: <stable@vger.kernel.org> # 3.3.x
  85. The tag has the meaning of::
  86. git cherry-pick <this commit>
  87. For each "-stable" tree starting with the specified version.
  88. Note, such tagging is unnecessary if the stable team can derive the
  89. appropriate versions from Fixes: tags.
  90. * Delay pick up of patches::
  91. Cc: <stable@vger.kernel.org> # after -rc3
  92. * Point out known problems::
  93. Cc: <stable@vger.kernel.org> # see patch description, needs adjustments for <= 6.3
  94. There furthermore is a variant of the stable tag you can use to make the stable
  95. team's backporting tools (e.g AUTOSEL or scripts that look for commits
  96. containing a 'Fixes:' tag) ignore a change::
  97. Cc: <stable+noautosel@kernel.org> # reason goes here, and must be present
  98. .. _option_2:
  99. Option 2
  100. ********
  101. If the patch already has been merged to mainline, send an email to
  102. stable@vger.kernel.org containing the subject of the patch, the commit ID,
  103. why you think it should be applied, and what kernel versions you wish it to
  104. be applied to.
  105. .. _option_3:
  106. Option 3
  107. ********
  108. Send the patch, after verifying that it follows the above rules, to
  109. stable@vger.kernel.org and mention the kernel versions you wish it to be applied
  110. to. When doing so, you must note the upstream commit ID in the changelog of your
  111. submission with a separate line above the commit text, like this::
  112. commit <sha1> upstream.
  113. Or alternatively::
  114. [ Upstream commit <sha1> ]
  115. If the submitted patch deviates from the original upstream patch (for example
  116. because it had to be adjusted for the older API), this must be very clearly
  117. documented and justified in the patch description.
  118. Following the submission
  119. ------------------------
  120. The sender will receive an ACK when the patch has been accepted into the
  121. queue, or a NAK if the patch is rejected. This response might take a few
  122. days, according to the schedules of the stable team members.
  123. If accepted, the patch will be added to the -stable queue, for review by other
  124. developers and by the relevant subsystem maintainer.
  125. Review cycle
  126. ------------
  127. - When the -stable maintainers decide for a review cycle, the patches will be
  128. sent to the review committee, and the maintainer of the affected area of
  129. the patch (unless the submitter is the maintainer of the area) and CC: to
  130. the linux-kernel mailing list.
  131. - The review committee has 48 hours in which to ACK or NAK the patch.
  132. - If the patch is rejected by a member of the committee, or linux-kernel
  133. members object to the patch, bringing up issues that the maintainers and
  134. members did not realize, the patch will be dropped from the queue.
  135. - The ACKed patches will be posted again as part of release candidate (-rc)
  136. to be tested by developers and testers.
  137. - Usually only one -rc release is made, however if there are any outstanding
  138. issues, some patches may be modified or dropped or additional patches may
  139. be queued. Additional -rc releases are then released and tested until no
  140. issues are found.
  141. - Responding to the -rc releases can be done on the mailing list by sending
  142. a "Tested-by:" email with any testing information desired. The "Tested-by:"
  143. tags will be collected and added to the release commit.
  144. - At the end of the review cycle, the new -stable release will be released
  145. containing all the queued and tested patches.
  146. - Security patches will be accepted into the -stable tree directly from the
  147. security kernel team, and not go through the normal review cycle.
  148. Contact the kernel security team for more details on this procedure.
  149. Trees
  150. -----
  151. - The queues of patches, for both completed versions and in progress
  152. versions can be found at:
  153. https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git
  154. - The finalized and tagged releases of all stable kernels can be found
  155. in separate branches per version at:
  156. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git
  157. - The release candidate of all stable kernel versions can be found at:
  158. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git/
  159. .. warning::
  160. The -stable-rc tree is a snapshot in time of the stable-queue tree and
  161. will change frequently, hence will be rebased often. It should only be
  162. used for testing purposes (e.g. to be consumed by CI systems).
  163. Review committee
  164. ----------------
  165. - This is made up of a number of kernel developers who have volunteered for
  166. this task, and a few that haven't.