3 How to Get Your Change Into the Linux Kernel or Care And Operation Of Your Linus Torvalds
4 =========================================================================================
6 For a person or company who wishes to submit a change to the Linux
7 kernel, the process can sometimes be daunting if you're not familiar
8 with "the system." This text is a collection of suggestions which
9 can greatly increase the chances of your change being accepted.
11 This document contains a large number of suggestions in a relatively terse
12 format. For detailed information on how the kernel development process
13 works, see Documentation/development-process. Also, read
14 Documentation/SubmitChecklist for a list of items to check before
15 submitting code. If you are submitting a driver, also read
16 Documentation/SubmittingDrivers; for device tree binding patches, read
17 Documentation/devicetree/bindings/submitting-patches.txt.
19 Many of these steps describe the default behavior of the ``git`` version
20 control system; if you use ``git`` to prepare your patches, you'll find much
21 of the mechanical work done for you, though you'll still need to prepare
22 and document a sensible set of patches. In general, use of ``git`` will make
23 your life as a kernel developer easier.
25 Creating and Sending your Change
26 ********************************
29 0) Obtain a current source tree
30 -------------------------------
32 If you do not have a repository with the current kernel source handy, use
33 ``git`` to obtain one. You'll want to start with the mainline repository,
34 which can be grabbed with::
36 git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
38 Note, however, that you may not want to develop against the mainline tree
39 directly. Most subsystem maintainers run their own trees and want to see
40 patches prepared against those trees. See the **T:** entry for the subsystem
41 in the MAINTAINERS file to find that tree, or simply ask the maintainer if
42 the tree is not listed there.
44 It is still possible to download kernel releases via tarballs (as described
45 in the next section), but that is the hard way to do kernel development.
50 If you must generate your patches by hand, use ``diff -up`` or ``diff -uprN``
51 to create patches. Git generates patches in this form by default; if
52 you're using ``git``, you can skip this section entirely.
54 All changes to the Linux kernel occur in the form of patches, as
55 generated by :manpage:`diff(1)`. When creating your patch, make sure to
56 create it in "unified diff" format, as supplied by the ``-u`` argument
57 to :manpage:`diff(1)`.
58 Also, please use the ``-p`` argument which shows which C function each
59 change is in - that makes the resultant ``diff`` a lot easier to read.
60 Patches should be based in the root kernel source directory,
61 not in any lower subdirectory.
63 To create a patch for a single file, it is often sufficient to do::
66 MYFILE= drivers/net/mydriver.c
69 cp $MYFILE $MYFILE.orig
70 vi $MYFILE # make your change
72 diff -up $SRCTREE/$MYFILE{.orig,} > /tmp/patch
74 To create a patch for multiple files, you should unpack a "vanilla",
75 or unmodified kernel source tree, and generate a ``diff`` against your
76 own source tree. For example::
80 tar xvfz linux-3.19.tar.gz
81 mv linux-3.19 linux-3.19-vanilla
82 diff -uprN -X linux-3.19-vanilla/Documentation/dontdiff \
83 linux-3.19-vanilla $MYSRC > /tmp/patch
85 ``dontdiff`` is a list of files which are generated by the kernel during
86 the build process, and should be ignored in any :manpage:`diff(1)`-generated
89 Make sure your patch does not include any extra files which do not
90 belong in a patch submission. Make sure to review your patch -after-
91 generating it with :manpage:`diff(1)`, to ensure accuracy.
93 If your changes produce a lot of deltas, you need to split them into
94 individual patches which modify things in logical stages; see
95 :ref:`split_changes`. This will facilitate review by other kernel developers,
96 very important if you want your patch accepted.
98 If you're using ``git``, ``git rebase -i`` can help you with this process. If
99 you're not using ``git``, ``quilt`` <http://savannah.nongnu.org/projects/quilt>
100 is another popular alternative.
102 .. _describe_changes:
104 2) Describe your changes
105 ------------------------
107 Describe your problem. Whether your patch is a one-line bug fix or
108 5000 lines of a new feature, there must be an underlying problem that
109 motivated you to do this work. Convince the reviewer that there is a
110 problem worth fixing and that it makes sense for them to read past the
113 Describe user-visible impact. Straight up crashes and lockups are
114 pretty convincing, but not all bugs are that blatant. Even if the
115 problem was spotted during code review, describe the impact you think
116 it can have on users. Keep in mind that the majority of Linux
117 installations run kernels from secondary stable trees or
118 vendor/product-specific trees that cherry-pick only specific patches
119 from upstream, so include anything that could help route your change
120 downstream: provoking circumstances, excerpts from dmesg, crash
121 descriptions, performance regressions, latency spikes, lockups, etc.
123 Quantify optimizations and trade-offs. If you claim improvements in
124 performance, memory consumption, stack footprint, or binary size,
125 include numbers that back them up. But also describe non-obvious
126 costs. Optimizations usually aren't free but trade-offs between CPU,
127 memory, and readability; or, when it comes to heuristics, between
128 different workloads. Describe the expected downsides of your
129 optimization so that the reviewer can weigh costs against benefits.
131 Once the problem is established, describe what you are actually doing
132 about it in technical detail. It's important to describe the change
133 in plain English for the reviewer to verify that the code is behaving
136 The maintainer will thank you if you write your patch description in a
137 form which can be easily pulled into Linux's source code management
138 system, ``git``, as a "commit log". See :ref:`explicit_in_reply_to`.
140 Solve only one problem per patch. If your description starts to get
141 long, that's a sign that you probably need to split up your patch.
142 See :ref:`split_changes`.
144 When you submit or resubmit a patch or patch series, include the
145 complete patch description and justification for it. Don't just
146 say that this is version N of the patch (series). Don't expect the
147 subsystem maintainer to refer back to earlier patch versions or referenced
148 URLs to find the patch description and put that into the patch.
149 I.e., the patch (series) and its description should be self-contained.
150 This benefits both the maintainers and reviewers. Some reviewers
151 probably didn't even receive earlier versions of the patch.
153 Describe your changes in imperative mood, e.g. "make xyzzy do frotz"
154 instead of "[This patch] makes xyzzy do frotz" or "[I] changed xyzzy
155 to do frotz", as if you are giving orders to the codebase to change
158 If the patch fixes a logged bug entry, refer to that bug entry by
159 number and URL. If the patch follows from a mailing list discussion,
160 give a URL to the mailing list archive; use the https://lkml.kernel.org/
161 redirector with a ``Message-Id``, to ensure that the links cannot become
164 However, try to make your explanation understandable without external
165 resources. In addition to giving a URL to a mailing list archive or
166 bug, summarize the relevant points of the discussion that led to the
169 If you want to refer to a specific commit, don't just refer to the
170 SHA-1 ID of the commit. Please also include the oneline summary of
171 the commit, to make it easier for reviewers to know what it is about.
174 Commit e21d2170f36602ae2708 ("video: remove unnecessary
175 platform_set_drvdata()") removed the unnecessary
176 platform_set_drvdata(), but left the variable "dev" unused,
179 You should also be sure to use at least the first twelve characters of the
180 SHA-1 ID. The kernel repository holds a *lot* of objects, making
181 collisions with shorter IDs a real possibility. Bear in mind that, even if
182 there is no collision with your six-character ID now, that condition may
183 change five years from now.
185 If your patch fixes a bug in a specific commit, e.g. you found an issue using
186 ``git bisect``, please use the 'Fixes:' tag with the first 12 characters of
187 the SHA-1 ID, and the one line summary. For example::
189 Fixes: e21d2170f366 ("video: remove unnecessary platform_set_drvdata()")
191 The following ``git config`` settings can be used to add a pretty format for
192 outputting the above style in the ``git log`` or ``git show`` commands::
197 fixes = Fixes: %h (\"%s\")
201 3) Separate your changes
202 ------------------------
204 Separate each **logical change** into a separate patch.
206 For example, if your changes include both bug fixes and performance
207 enhancements for a single driver, separate those changes into two
208 or more patches. If your changes include an API update, and a new
209 driver which uses that new API, separate those into two patches.
211 On the other hand, if you make a single change to numerous files,
212 group those changes into a single patch. Thus a single logical change
213 is contained within a single patch.
215 The point to remember is that each patch should make an easily understood
216 change that can be verified by reviewers. Each patch should be justifiable
219 If one patch depends on another patch in order for a change to be
220 complete, that is OK. Simply note **"this patch depends on patch X"**
221 in your patch description.
223 When dividing your change into a series of patches, take special care to
224 ensure that the kernel builds and runs properly after each patch in the
225 series. Developers using ``git bisect`` to track down a problem can end up
226 splitting your patch series at any point; they will not thank you if you
227 introduce bugs in the middle.
229 If you cannot condense your patch set into a smaller set of patches,
230 then only post say 15 or so at a time and wait for review and integration.
234 4) Style-check your changes
235 ---------------------------
237 Check your patch for basic style violations, details of which can be
238 found in Documentation/CodingStyle. Failure to do so simply wastes
239 the reviewers time and will get your patch rejected, probably
240 without even being read.
242 One significant exception is when moving code from one file to
243 another -- in this case you should not modify the moved code at all in
244 the same patch which moves it. This clearly delineates the act of
245 moving the code and your changes. This greatly aids review of the
246 actual differences and allows tools to better track the history of
249 Check your patches with the patch style checker prior to submission
250 (scripts/checkpatch.pl). Note, though, that the style checker should be
251 viewed as a guide, not as a replacement for human judgment. If your code
252 looks better with a violation then its probably best left alone.
254 The checker reports at three levels:
255 - ERROR: things that are very likely to be wrong
256 - WARNING: things requiring careful review
257 - CHECK: things requiring thought
259 You should be able to justify all violations that remain in your
263 5) Select the recipients for your patch
264 ---------------------------------------
266 You should always copy the appropriate subsystem maintainer(s) on any patch
267 to code that they maintain; look through the MAINTAINERS file and the
268 source code revision history to see who those maintainers are. The
269 script scripts/get_maintainer.pl can be very useful at this step. If you
270 cannot find a maintainer for the subsystem you are working on, Andrew
271 Morton (akpm@linux-foundation.org) serves as a maintainer of last resort.
273 You should also normally choose at least one mailing list to receive a copy
274 of your patch set. linux-kernel@vger.kernel.org functions as a list of
275 last resort, but the volume on that list has caused a number of developers
276 to tune it out. Look in the MAINTAINERS file for a subsystem-specific
277 list; your patch will probably get more attention there. Please do not
278 spam unrelated lists, though.
280 Many kernel-related lists are hosted on vger.kernel.org; you can find a
281 list of them at http://vger.kernel.org/vger-lists.html. There are
282 kernel-related lists hosted elsewhere as well, though.
284 Do not send more than 15 patches at once to the vger mailing lists!!!
286 Linus Torvalds is the final arbiter of all changes accepted into the
287 Linux kernel. His e-mail address is <torvalds@linux-foundation.org>.
288 He gets a lot of e-mail, and, at this point, very few patches go through
289 Linus directly, so typically you should do your best to -avoid-
292 If you have a patch that fixes an exploitable security bug, send that patch
293 to security@kernel.org. For severe bugs, a short embargo may be considered
294 to allow distributors to get the patch out to users; in such cases,
295 obviously, the patch should not be sent to any public lists.
297 Patches that fix a severe bug in a released kernel should be directed
298 toward the stable maintainers by putting a line like this::
300 Cc: stable@vger.kernel.org
302 into the sign-off area of your patch (note, NOT an email recipient). You
303 should also read Documentation/stable_kernel_rules.txt in addition to this
306 Note, however, that some subsystem maintainers want to come to their own
307 conclusions on which patches should go to the stable trees. The networking
308 maintainer, in particular, would rather not see individual developers
309 adding lines like the above to their patches.
311 If changes affect userland-kernel interfaces, please send the MAN-PAGES
312 maintainer (as listed in the MAINTAINERS file) a man-pages patch, or at
313 least a notification of the change, so that some information makes its way
314 into the manual pages. User-space API changes should also be copied to
315 linux-api@vger.kernel.org.
317 For small patches you may want to CC the Trivial Patch Monkey
318 trivial@kernel.org which collects "trivial" patches. Have a look
319 into the MAINTAINERS file for its current manager.
321 Trivial patches must qualify for one of the following rules:
323 - Spelling fixes in documentation
324 - Spelling fixes for errors which could break :manpage:`grep(1)`
325 - Warning fixes (cluttering with useless warnings is bad)
326 - Compilation fixes (only if they are actually correct)
327 - Runtime fixes (only if they actually fix things)
328 - Removing use of deprecated functions/macros
329 - Contact detail and documentation fixes
330 - Non-portable code replaced by portable code (even in arch-specific,
331 since people copy, as long as it's trivial)
332 - Any fix by the author/maintainer of the file (ie. patch monkey
333 in re-transmission mode)
337 6) No MIME, no links, no compression, no attachments. Just plain text
338 ----------------------------------------------------------------------
340 Linus and other kernel developers need to be able to read and comment
341 on the changes you are submitting. It is important for a kernel
342 developer to be able to "quote" your changes, using standard e-mail
343 tools, so that they may comment on specific portions of your code.
345 For this reason, all patches should be submitted by e-mail "inline".
349 Be wary of your editor's word-wrap corrupting your patch,
350 if you choose to cut-n-paste your patch.
352 Do not attach the patch as a MIME attachment, compressed or not.
353 Many popular e-mail applications will not always transmit a MIME
354 attachment as plain text, making it impossible to comment on your
355 code. A MIME attachment also takes Linus a bit more time to process,
356 decreasing the likelihood of your MIME-attached change being accepted.
358 Exception: If your mailer is mangling patches then someone may ask
359 you to re-send them using MIME.
361 See Documentation/email-clients.txt for hints about configuring
362 your e-mail client so that it sends your patches untouched.
367 Large changes are not appropriate for mailing lists, and some
368 maintainers. If your patch, uncompressed, exceeds 300 kB in size,
369 it is preferred that you store your patch on an Internet-accessible
370 server, and provide instead a URL (link) pointing to your patch. But note
371 that if your patch exceeds 300 kB, it almost certainly needs to be broken up
374 8) Respond to review comments
375 -----------------------------
377 Your patch will almost certainly get comments from reviewers on ways in
378 which the patch can be improved. You must respond to those comments;
379 ignoring reviewers is a good way to get ignored in return. Review comments
380 or questions that do not lead to a code change should almost certainly
381 bring about a comment or changelog entry so that the next reviewer better
382 understands what is going on.
384 Be sure to tell the reviewers what changes you are making and to thank them
385 for their time. Code review is a tiring and time-consuming process, and
386 reviewers sometimes get grumpy. Even in that case, though, respond
387 politely and address the problems they have pointed out.
390 9) Don't get discouraged - or impatient
391 ---------------------------------------
393 After you have submitted your change, be patient and wait. Reviewers are
394 busy people and may not get to your patch right away.
396 Once upon a time, patches used to disappear into the void without comment,
397 but the development process works more smoothly than that now. You should
398 receive comments within a week or so; if that does not happen, make sure
399 that you have sent your patches to the right place. Wait for a minimum of
400 one week before resubmitting or pinging reviewers - possibly longer during
401 busy times like merge windows.
404 10) Include PATCH in the subject
405 --------------------------------
407 Due to high e-mail traffic to Linus, and to linux-kernel, it is common
408 convention to prefix your subject line with [PATCH]. This lets Linus
409 and other kernel developers more easily distinguish patches from other
417 To improve tracking of who did what, especially with patches that can
418 percolate to their final resting place in the kernel through several
419 layers of maintainers, we've introduced a "sign-off" procedure on
420 patches that are being emailed around.
422 The sign-off is a simple line at the end of the explanation for the
423 patch, which certifies that you wrote it or otherwise have the right to
424 pass it on as an open-source patch. The rules are pretty simple: if you
425 can certify the below:
427 Developer's Certificate of Origin 1.1
428 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
430 By making a contribution to this project, I certify that:
432 (a) The contribution was created in whole or in part by me and I
433 have the right to submit it under the open source license
434 indicated in the file; or
436 (b) The contribution is based upon previous work that, to the best
437 of my knowledge, is covered under an appropriate open source
438 license and I have the right under that license to submit that
439 work with modifications, whether created in whole or in part
440 by me, under the same open source license (unless I am
441 permitted to submit under a different license), as indicated
444 (c) The contribution was provided directly to me by some other
445 person who certified (a), (b) or (c) and I have not modified
448 (d) I understand and agree that this project and the contribution
449 are public and that a record of the contribution (including all
450 personal information I submit with it, including my sign-off) is
451 maintained indefinitely and may be redistributed consistent with
452 this project or the open source license(s) involved.
454 then you just add a line saying::
456 Signed-off-by: Random J Developer <random@developer.example.org>
458 using your real name (sorry, no pseudonyms or anonymous contributions.)
460 Some people also put extra tags at the end. They'll just be ignored for
461 now, but you can do this to mark internal company procedures or just
462 point out some special detail about the sign-off.
464 If you are a subsystem or branch maintainer, sometimes you need to slightly
465 modify patches you receive in order to merge them, because the code is not
466 exactly the same in your tree and the submitters'. If you stick strictly to
467 rule (c), you should ask the submitter to rediff, but this is a totally
468 counter-productive waste of time and energy. Rule (b) allows you to adjust
469 the code, but then it is very impolite to change one submitter's code and
470 make him endorse your bugs. To solve this problem, it is recommended that
471 you add a line between the last Signed-off-by header and yours, indicating
472 the nature of your changes. While there is nothing mandatory about this, it
473 seems like prepending the description with your mail and/or name, all
474 enclosed in square brackets, is noticeable enough to make it obvious that
475 you are responsible for last-minute changes. Example::
477 Signed-off-by: Random J Developer <random@developer.example.org>
478 [lucky@maintainer.example.org: struct foo moved from foo.c to foo.h]
479 Signed-off-by: Lucky K Maintainer <lucky@maintainer.example.org>
481 This practice is particularly helpful if you maintain a stable branch and
482 want at the same time to credit the author, track changes, merge the fix,
483 and protect the submitter from complaints. Note that under no circumstances
484 can you change the author's identity (the From header), as it is the one
485 which appears in the changelog.
487 Special note to back-porters: It seems to be a common and useful practice
488 to insert an indication of the origin of a patch at the top of the commit
489 message (just after the subject line) to facilitate tracking. For instance,
490 here's what we see in a 3.x-stable release::
492 Date: Tue Oct 7 07:26:38 2014 -0400
494 libata: Un-break ATA blacklist
496 commit 1c40279960bcd7d52dbdf1d466b20d24b99176c8 upstream.
498 And here's what might appear in an older kernel once a patch is backported::
500 Date: Tue May 13 22:12:27 2008 +0200
502 wireless, airo: waitbusy() won't delay
504 [backport of 2.6 commit b7acbdfbd1f277c1eb23f344f899cfa4cd0bf36a]
506 Whatever the format, this information provides a valuable help to people
507 tracking your trees, and to people trying to troubleshoot bugs in your
511 12) When to use Acked-by: and Cc:
512 ---------------------------------
514 The Signed-off-by: tag indicates that the signer was involved in the
515 development of the patch, or that he/she was in the patch's delivery path.
517 If a person was not directly involved in the preparation or handling of a
518 patch but wishes to signify and record their approval of it then they can
519 ask to have an Acked-by: line added to the patch's changelog.
521 Acked-by: is often used by the maintainer of the affected code when that
522 maintainer neither contributed to nor forwarded the patch.
524 Acked-by: is not as formal as Signed-off-by:. It is a record that the acker
525 has at least reviewed the patch and has indicated acceptance. Hence patch
526 mergers will sometimes manually convert an acker's "yep, looks good to me"
527 into an Acked-by: (but note that it is usually better to ask for an
530 Acked-by: does not necessarily indicate acknowledgement of the entire patch.
531 For example, if a patch affects multiple subsystems and has an Acked-by: from
532 one subsystem maintainer then this usually indicates acknowledgement of just
533 the part which affects that maintainer's code. Judgement should be used here.
534 When in doubt people should refer to the original discussion in the mailing
537 If a person has had the opportunity to comment on a patch, but has not
538 provided such comments, you may optionally add a ``Cc:`` tag to the patch.
539 This is the only tag which might be added without an explicit action by the
540 person it names - but it should indicate that this person was copied on the
541 patch. This tag documents that potentially interested parties
542 have been included in the discussion.
545 13) Using Reported-by:, Tested-by:, Reviewed-by:, Suggested-by: and Fixes:
546 --------------------------------------------------------------------------
548 The Reported-by tag gives credit to people who find bugs and report them and it
549 hopefully inspires them to help us again in the future. Please note that if
550 the bug was reported in private, then ask for permission first before using the
553 A Tested-by: tag indicates that the patch has been successfully tested (in
554 some environment) by the person named. This tag informs maintainers that
555 some testing has been performed, provides a means to locate testers for
556 future patches, and ensures credit for the testers.
558 Reviewed-by:, instead, indicates that the patch has been reviewed and found
559 acceptable according to the Reviewer's Statement:
561 Reviewer's statement of oversight
562 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
564 By offering my Reviewed-by: tag, I state that:
566 (a) I have carried out a technical review of this patch to
567 evaluate its appropriateness and readiness for inclusion into
570 (b) Any problems, concerns, or questions relating to the patch
571 have been communicated back to the submitter. I am satisfied
572 with the submitter's response to my comments.
574 (c) While there may be things that could be improved with this
575 submission, I believe that it is, at this time, (1) a
576 worthwhile modification to the kernel, and (2) free of known
577 issues which would argue against its inclusion.
579 (d) While I have reviewed the patch and believe it to be sound, I
580 do not (unless explicitly stated elsewhere) make any
581 warranties or guarantees that it will achieve its stated
582 purpose or function properly in any given situation.
584 A Reviewed-by tag is a statement of opinion that the patch is an
585 appropriate modification of the kernel without any remaining serious
586 technical issues. Any interested reviewer (who has done the work) can
587 offer a Reviewed-by tag for a patch. This tag serves to give credit to
588 reviewers and to inform maintainers of the degree of review which has been
589 done on the patch. Reviewed-by: tags, when supplied by reviewers known to
590 understand the subject area and to perform thorough reviews, will normally
591 increase the likelihood of your patch getting into the kernel.
593 A Suggested-by: tag indicates that the patch idea is suggested by the person
594 named and ensures credit to the person for the idea. Please note that this
595 tag should not be added without the reporter's permission, especially if the
596 idea was not posted in a public forum. That said, if we diligently credit our
597 idea reporters, they will, hopefully, be inspired to help us again in the
600 A Fixes: tag indicates that the patch fixes an issue in a previous commit. It
601 is used to make it easy to determine where a bug originated, which can help
602 review a bug fix. This tag also assists the stable kernel team in determining
603 which stable kernel versions should receive your fix. This is the preferred
604 method for indicating a bug fixed by the patch. See :ref:`describe_changes`
608 14) The canonical patch format
609 ------------------------------
611 This section describes how the patch itself should be formatted. Note
612 that, if you have your patches stored in a ``git`` repository, proper patch
613 formatting can be had with ``git format-patch``. The tools cannot create
614 the necessary text, though, so read the instructions below anyway.
616 The canonical patch subject line is::
618 Subject: [PATCH 001/123] subsystem: summary phrase
620 The canonical patch message body contains the following:
622 - A ``from`` line specifying the patch author (only needed if the person
623 sending the patch is not the author).
627 - The body of the explanation, line wrapped at 75 columns, which will
628 be copied to the permanent changelog to describe this patch.
630 - The ``Signed-off-by:`` lines, described above, which will
631 also go in the changelog.
633 - A marker line containing simply ``---``.
635 - Any additional comments not suitable for the changelog.
637 - The actual patch (``diff`` output).
639 The Subject line format makes it very easy to sort the emails
640 alphabetically by subject line - pretty much any email reader will
641 support that - since because the sequence number is zero-padded,
642 the numerical and alphabetic sort is the same.
644 The ``subsystem`` in the email's Subject should identify which
645 area or subsystem of the kernel is being patched.
647 The ``summary phrase`` in the email's Subject should concisely
648 describe the patch which that email contains. The ``summary
649 phrase`` should not be a filename. Do not use the same ``summary
650 phrase`` for every patch in a whole patch series (where a ``patch
651 series`` is an ordered sequence of multiple, related patches).
653 Bear in mind that the ``summary phrase`` of your email becomes a
654 globally-unique identifier for that patch. It propagates all the way
655 into the ``git`` changelog. The ``summary phrase`` may later be used in
656 developer discussions which refer to the patch. People will want to
657 google for the ``summary phrase`` to read discussion regarding that
658 patch. It will also be the only thing that people may quickly see
659 when, two or three months later, they are going through perhaps
660 thousands of patches using tools such as ``gitk`` or ``git log
663 For these reasons, the ``summary`` must be no more than 70-75
664 characters, and it must describe both what the patch changes, as well
665 as why the patch might be necessary. It is challenging to be both
666 succinct and descriptive, but that is what a well-written summary
669 The ``summary phrase`` may be prefixed by tags enclosed in square
670 brackets: "Subject: [PATCH <tag>...] <summary phrase>". The tags are
671 not considered part of the summary phrase, but describe how the patch
672 should be treated. Common tags might include a version descriptor if
673 the multiple versions of the patch have been sent out in response to
674 comments (i.e., "v1, v2, v3"), or "RFC" to indicate a request for
675 comments. If there are four patches in a patch series the individual
676 patches may be numbered like this: 1/4, 2/4, 3/4, 4/4. This assures
677 that developers understand the order in which the patches should be
678 applied and that they have reviewed or applied all of the patches in
681 A couple of example Subjects::
683 Subject: [PATCH 2/5] ext2: improve scalability of bitmap searching
684 Subject: [PATCH v2 01/27] x86: fix eflags tracking
686 The ``from`` line must be the very first line in the message body,
689 From: Original Author <author@example.com>
691 The ``from`` line specifies who will be credited as the author of the
692 patch in the permanent changelog. If the ``from`` line is missing,
693 then the ``From:`` line from the email header will be used to determine
694 the patch author in the changelog.
696 The explanation body will be committed to the permanent source
697 changelog, so should make sense to a competent reader who has long
698 since forgotten the immediate details of the discussion that might
699 have led to this patch. Including symptoms of the failure which the
700 patch addresses (kernel log messages, oops messages, etc.) is
701 especially useful for people who might be searching the commit logs
702 looking for the applicable patch. If a patch fixes a compile failure,
703 it may not be necessary to include _all_ of the compile failures; just
704 enough that it is likely that someone searching for the patch can find
705 it. As in the ``summary phrase``, it is important to be both succinct as
708 The ``---`` marker line serves the essential purpose of marking for patch
709 handling tools where the changelog message ends.
711 One good use for the additional comments after the ``---`` marker is for
712 a ``diffstat``, to show what files have changed, and the number of
713 inserted and deleted lines per file. A ``diffstat`` is especially useful
714 on bigger patches. Other comments relevant only to the moment or the
715 maintainer, not suitable for the permanent changelog, should also go
716 here. A good example of such comments might be ``patch changelogs``
717 which describe what has changed between the v1 and v2 version of the
720 If you are going to include a ``diffstat`` after the ``---`` marker, please
721 use ``diffstat`` options ``-p 1 -w 70`` so that filenames are listed from
722 the top of the kernel source tree and don't use too much horizontal
723 space (easily fit in 80 columns, maybe with some indentation). (``git``
724 generates appropriate diffstats by default.)
726 See more details on the proper patch format in the following
729 .. _explicit_in_reply_to:
731 15) Explicit In-Reply-To headers
732 --------------------------------
734 It can be helpful to manually add In-Reply-To: headers to a patch
735 (e.g., when using ``git send-email``) to associate the patch with
736 previous relevant discussion, e.g. to link a bug fix to the email with
737 the bug report. However, for a multi-patch series, it is generally
738 best to avoid using In-Reply-To: to link to older versions of the
739 series. This way multiple versions of the patch don't become an
740 unmanageable forest of references in email clients. If a link is
741 helpful, you can use the https://lkml.kernel.org/ redirector (e.g., in
742 the cover email text) to link to an earlier version of the patch series.
745 16) Sending ``git pull`` requests
746 ---------------------------------
748 If you have a series of patches, it may be most convenient to have the
749 maintainer pull them directly into the subsystem repository with a
750 ``git pull`` operation. Note, however, that pulling patches from a developer
751 requires a higher degree of trust than taking patches from a mailing list.
752 As a result, many subsystem maintainers are reluctant to take pull
753 requests, especially from new, unknown developers. If in doubt you can use
754 the pull request as the cover letter for a normal posting of the patch
755 series, giving the maintainer the option of using either.
757 A pull request should have [GIT] or [PULL] in the subject line. The
758 request itself should include the repository name and the branch of
759 interest on a single line; it should look something like::
763 git://jdelvare.pck.nerim.net/jdelvare-2.6 i2c-for-linus
765 to get these changes:
767 A pull request should also include an overall message saying what will be
768 included in the request, a ``git shortlog`` listing of the patches
769 themselves, and a ``diffstat`` showing the overall effect of the patch series.
770 The easiest way to get all this information together is, of course, to let
771 ``git`` do it for you with the ``git request-pull`` command.
773 Some maintainers (including Linus) want to see pull requests from signed
774 commits; that increases their confidence that the request actually came
775 from you. Linus, in particular, will not pull from public hosting sites
776 like GitHub in the absence of a signed tag.
778 The first step toward creating such tags is to make a GNUPG key and get it
779 signed by one or more core kernel developers. This step can be hard for
780 new developers, but there is no way around it. Attending conferences can
781 be a good way to find developers who can sign your key.
783 Once you have prepared a patch series in ``git`` that you wish to have somebody
784 pull, create a signed tag with ``git tag -s``. This will create a new tag
785 identifying the last commit in the series and containing a signature
786 created with your private key. You will also have the opportunity to add a
787 changelog-style message to the tag; this is an ideal place to describe the
788 effects of the pull request as a whole.
790 If the tree the maintainer will be pulling from is not the repository you
791 are working from, don't forget to push the signed tag explicitly to the
794 When generating your pull request, use the signed tag as the target. A
795 command like this will do the trick::
797 git request-pull master git://my.public.tree/linux.git my-signed-tag
803 Andrew Morton, "The perfect patch" (tpp).
804 <http://www.ozlabs.org/~akpm/stuff/tpp.txt>
806 Jeff Garzik, "Linux kernel patch submission format".
807 <http://linux.yyz.us/patch-format.html>
809 Greg Kroah-Hartman, "How to piss off a kernel subsystem maintainer".
810 <http://www.kroah.com/log/linux/maintainer.html>
812 <http://www.kroah.com/log/linux/maintainer-02.html>
814 <http://www.kroah.com/log/linux/maintainer-03.html>
816 <http://www.kroah.com/log/linux/maintainer-04.html>
818 <http://www.kroah.com/log/linux/maintainer-05.html>
820 <http://www.kroah.com/log/linux/maintainer-06.html>
822 NO!!!! No more huge patch bombs to linux-kernel@vger.kernel.org people!
823 <https://lkml.org/lkml/2005/7/11/336>
825 Kernel Documentation/CodingStyle:
826 <Documentation/CodingStyle>
828 Linus Torvalds's mail on the canonical patch format:
829 <http://lkml.org/lkml/2005/4/7/183>
831 Andi Kleen, "On submitting kernel patches"
832 Some strategies to get difficult or controversial changes in.
834 http://halobates.de/on-submitting-patches.pdf