2 How to Get Your Change Into the Linux Kernel or Care And Operation Of Your Linus Torvalds
3 =========================================================================================
5 For a person or company who wishes to submit a change to the Linux
6 kernel, the process can sometimes be daunting if you're not familiar
7 with "the system." This text is a collection of suggestions which
8 can greatly increase the chances of your change being accepted.
10 This document contains a large number of suggestions in a relatively terse
11 format. For detailed information on how the kernel development process
12 works, see Documentation/development-process. Also, read
13 Documentation/SubmitChecklist for a list of items to check before
14 submitting code. If you are submitting a driver, also read
15 Documentation/SubmittingDrivers; for device tree binding patches, read
16 Documentation/devicetree/bindings/submitting-patches.txt.
18 Many of these steps describe the default behavior of the git version
19 control system; if you use git to prepare your patches, you'll find much
20 of the mechanical work done for you, though you'll still need to prepare
21 and document a sensible set of patches. In general, use of git will make
22 your life as a kernel developer easier.
24 Creating and Sending your Change
25 ********************************
28 0) Obtain a current source tree
29 -------------------------------
31 If you do not have a repository with the current kernel source handy, use
32 git to obtain one. You'll want to start with the mainline repository,
33 which can be grabbed with::
35 git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
37 Note, however, that you may not want to develop against the mainline tree
38 directly. Most subsystem maintainers run their own trees and want to see
39 patches prepared against those trees. See the **T:** entry for the subsystem
40 in the MAINTAINERS file to find that tree, or simply ask the maintainer if
41 the tree is not listed there.
43 It is still possible to download kernel releases via tarballs (as described
44 in the next section), but that is the hard way to do kernel development.
49 If you must generate your patches by hand, use ``diff -up`` or ``diff -uprN``
50 to create patches. Git generates patches in this form by default; if
51 you're using git, you can skip this section entirely.
53 All changes to the Linux kernel occur in the form of patches, as
54 generated by diff(1). When creating your patch, make sure to create it
55 in "unified diff" format, as supplied by the ``-u`` argument to diff(1).
56 Also, please use the ``-p`` argument which shows which C function each
57 change is in - that makes the resultant diff a lot easier to read.
58 Patches should be based in the root kernel source directory,
59 not in any lower subdirectory.
61 To create a patch for a single file, it is often sufficient to do::
64 MYFILE= drivers/net/mydriver.c
67 cp $MYFILE $MYFILE.orig
68 vi $MYFILE # make your change
70 diff -up $SRCTREE/$MYFILE{.orig,} > /tmp/patch
72 To create a patch for multiple files, you should unpack a "vanilla",
73 or unmodified kernel source tree, and generate a diff against your
74 own source tree. For example::
78 tar xvfz linux-3.19.tar.gz
79 mv linux-3.19 linux-3.19-vanilla
80 diff -uprN -X linux-3.19-vanilla/Documentation/dontdiff \
81 linux-3.19-vanilla $MYSRC > /tmp/patch
83 ``dontdiff`` is a list of files which are generated by the kernel during
84 the build process, and should be ignored in any diff(1)-generated
87 Make sure your patch does not include any extra files which do not
88 belong in a patch submission. Make sure to review your patch -after-
89 generating it with diff(1), to ensure accuracy.
91 If your changes produce a lot of deltas, you need to split them into
92 individual patches which modify things in logical stages; see
93 :ref:`split_changes`. This will facilitate review by other kernel developers,
94 very important if you want your patch accepted.
96 If you're using git, ``git rebase -i`` can help you with this process. If
97 you're not using git, quilt <http://savannah.nongnu.org/projects/quilt>
98 is another popular alternative.
100 .. _describe_changes:
102 2) Describe your changes
103 ------------------------
105 Describe your problem. Whether your patch is a one-line bug fix or
106 5000 lines of a new feature, there must be an underlying problem that
107 motivated you to do this work. Convince the reviewer that there is a
108 problem worth fixing and that it makes sense for them to read past the
111 Describe user-visible impact. Straight up crashes and lockups are
112 pretty convincing, but not all bugs are that blatant. Even if the
113 problem was spotted during code review, describe the impact you think
114 it can have on users. Keep in mind that the majority of Linux
115 installations run kernels from secondary stable trees or
116 vendor/product-specific trees that cherry-pick only specific patches
117 from upstream, so include anything that could help route your change
118 downstream: provoking circumstances, excerpts from dmesg, crash
119 descriptions, performance regressions, latency spikes, lockups, etc.
121 Quantify optimizations and trade-offs. If you claim improvements in
122 performance, memory consumption, stack footprint, or binary size,
123 include numbers that back them up. But also describe non-obvious
124 costs. Optimizations usually aren't free but trade-offs between CPU,
125 memory, and readability; or, when it comes to heuristics, between
126 different workloads. Describe the expected downsides of your
127 optimization so that the reviewer can weigh costs against benefits.
129 Once the problem is established, describe what you are actually doing
130 about it in technical detail. It's important to describe the change
131 in plain English for the reviewer to verify that the code is behaving
134 The maintainer will thank you if you write your patch description in a
135 form which can be easily pulled into Linux's source code management
136 system, git, as a "commit log". See :ref:`explicit_in_reply_to`.
138 Solve only one problem per patch. If your description starts to get
139 long, that's a sign that you probably need to split up your patch.
140 See :ref:`split_changes`.
142 When you submit or resubmit a patch or patch series, include the
143 complete patch description and justification for it. Don't just
144 say that this is version N of the patch (series). Don't expect the
145 subsystem maintainer to refer back to earlier patch versions or referenced
146 URLs to find the patch description and put that into the patch.
147 I.e., the patch (series) and its description should be self-contained.
148 This benefits both the maintainers and reviewers. Some reviewers
149 probably didn't even receive earlier versions of the patch.
151 Describe your changes in imperative mood, e.g. "make xyzzy do frotz"
152 instead of "[This patch] makes xyzzy do frotz" or "[I] changed xyzzy
153 to do frotz", as if you are giving orders to the codebase to change
156 If the patch fixes a logged bug entry, refer to that bug entry by
157 number and URL. If the patch follows from a mailing list discussion,
158 give a URL to the mailing list archive; use the https://lkml.kernel.org/
159 redirector with a Message-Id, to ensure that the links cannot become
162 However, try to make your explanation understandable without external
163 resources. In addition to giving a URL to a mailing list archive or
164 bug, summarize the relevant points of the discussion that led to the
167 If you want to refer to a specific commit, don't just refer to the
168 SHA-1 ID of the commit. Please also include the oneline summary of
169 the commit, to make it easier for reviewers to know what it is about.
172 Commit e21d2170f36602ae2708 ("video: remove unnecessary
173 platform_set_drvdata()") removed the unnecessary
174 platform_set_drvdata(), but left the variable "dev" unused,
177 You should also be sure to use at least the first twelve characters of the
178 SHA-1 ID. The kernel repository holds a *lot* of objects, making
179 collisions with shorter IDs a real possibility. Bear in mind that, even if
180 there is no collision with your six-character ID now, that condition may
181 change five years from now.
183 If your patch fixes a bug in a specific commit, e.g. you found an issue using
184 git-bisect, please use the 'Fixes:' tag with the first 12 characters of the
185 SHA-1 ID, and the one line summary. For example::
187 Fixes: e21d2170f366 ("video: remove unnecessary platform_set_drvdata()")
189 The following git-config settings can be used to add a pretty format for
190 outputting the above style in the git log or git show commands::
195 fixes = Fixes: %h (\"%s\")
199 3) Separate your changes
200 ------------------------
202 Separate each **logical change** into a separate patch.
204 For example, if your changes include both bug fixes and performance
205 enhancements for a single driver, separate those changes into two
206 or more patches. If your changes include an API update, and a new
207 driver which uses that new API, separate those into two patches.
209 On the other hand, if you make a single change to numerous files,
210 group those changes into a single patch. Thus a single logical change
211 is contained within a single patch.
213 The point to remember is that each patch should make an easily understood
214 change that can be verified by reviewers. Each patch should be justifiable
217 If one patch depends on another patch in order for a change to be
218 complete, that is OK. Simply note **"this patch depends on patch X"**
219 in your patch description.
221 When dividing your change into a series of patches, take special care to
222 ensure that the kernel builds and runs properly after each patch in the
223 series. Developers using ``git bisect`` to track down a problem can end up
224 splitting your patch series at any point; they will not thank you if you
225 introduce bugs in the middle.
227 If you cannot condense your patch set into a smaller set of patches,
228 then only post say 15 or so at a time and wait for review and integration.
232 4) Style-check your changes
233 ---------------------------
235 Check your patch for basic style violations, details of which can be
236 found in Documentation/CodingStyle. Failure to do so simply wastes
237 the reviewers time and will get your patch rejected, probably
238 without even being read.
240 One significant exception is when moving code from one file to
241 another -- in this case you should not modify the moved code at all in
242 the same patch which moves it. This clearly delineates the act of
243 moving the code and your changes. This greatly aids review of the
244 actual differences and allows tools to better track the history of
247 Check your patches with the patch style checker prior to submission
248 (scripts/checkpatch.pl). Note, though, that the style checker should be
249 viewed as a guide, not as a replacement for human judgment. If your code
250 looks better with a violation then its probably best left alone.
252 The checker reports at three levels:
253 - ERROR: things that are very likely to be wrong
254 - WARNING: things requiring careful review
255 - CHECK: things requiring thought
257 You should be able to justify all violations that remain in your
261 5) Select the recipients for your patch
262 ---------------------------------------
264 You should always copy the appropriate subsystem maintainer(s) on any patch
265 to code that they maintain; look through the MAINTAINERS file and the
266 source code revision history to see who those maintainers are. The
267 script scripts/get_maintainer.pl can be very useful at this step. If you
268 cannot find a maintainer for the subsystem you are working on, Andrew
269 Morton (akpm@linux-foundation.org) serves as a maintainer of last resort.
271 You should also normally choose at least one mailing list to receive a copy
272 of your patch set. linux-kernel@vger.kernel.org functions as a list of
273 last resort, but the volume on that list has caused a number of developers
274 to tune it out. Look in the MAINTAINERS file for a subsystem-specific
275 list; your patch will probably get more attention there. Please do not
276 spam unrelated lists, though.
278 Many kernel-related lists are hosted on vger.kernel.org; you can find a
279 list of them at http://vger.kernel.org/vger-lists.html. There are
280 kernel-related lists hosted elsewhere as well, though.
282 Do not send more than 15 patches at once to the vger mailing lists!!!
284 Linus Torvalds is the final arbiter of all changes accepted into the
285 Linux kernel. His e-mail address is <torvalds@linux-foundation.org>.
286 He gets a lot of e-mail, and, at this point, very few patches go through
287 Linus directly, so typically you should do your best to -avoid-
290 If you have a patch that fixes an exploitable security bug, send that patch
291 to security@kernel.org. For severe bugs, a short embargo may be considered
292 to allow distributors to get the patch out to users; in such cases,
293 obviously, the patch should not be sent to any public lists.
295 Patches that fix a severe bug in a released kernel should be directed
296 toward the stable maintainers by putting a line like this::
298 Cc: stable@vger.kernel.org
300 into the sign-off area of your patch (note, NOT an email recipient). You
301 should also read Documentation/stable_kernel_rules.txt in addition to this
304 Note, however, that some subsystem maintainers want to come to their own
305 conclusions on which patches should go to the stable trees. The networking
306 maintainer, in particular, would rather not see individual developers
307 adding lines like the above to their patches.
309 If changes affect userland-kernel interfaces, please send the MAN-PAGES
310 maintainer (as listed in the MAINTAINERS file) a man-pages patch, or at
311 least a notification of the change, so that some information makes its way
312 into the manual pages. User-space API changes should also be copied to
313 linux-api@vger.kernel.org.
315 For small patches you may want to CC the Trivial Patch Monkey
316 trivial@kernel.org which collects "trivial" patches. Have a look
317 into the MAINTAINERS file for its current manager.
319 Trivial patches must qualify for one of the following rules:
321 Spelling fixes in documentation
322 Spelling fixes for errors which could break grep(1)
323 Warning fixes (cluttering with useless warnings is bad)
324 Compilation fixes (only if they are actually correct)
325 Runtime fixes (only if they actually fix things)
326 Removing use of deprecated functions/macros
327 Contact detail and documentation fixes
328 Non-portable code replaced by portable code (even in arch-specific,
329 since people copy, as long as it's trivial)
330 Any fix by the author/maintainer of the file (ie. patch monkey
331 in re-transmission mode)
335 6) No MIME, no links, no compression, no attachments. Just plain text
336 ----------------------------------------------------------------------
338 Linus and other kernel developers need to be able to read and comment
339 on the changes you are submitting. It is important for a kernel
340 developer to be able to "quote" your changes, using standard e-mail
341 tools, so that they may comment on specific portions of your code.
343 For this reason, all patches should be submitted by e-mail "inline".
344 WARNING: Be wary of your editor's word-wrap corrupting your patch,
345 if you choose to cut-n-paste your patch.
347 Do not attach the patch as a MIME attachment, compressed or not.
348 Many popular e-mail applications will not always transmit a MIME
349 attachment as plain text, making it impossible to comment on your
350 code. A MIME attachment also takes Linus a bit more time to process,
351 decreasing the likelihood of your MIME-attached change being accepted.
353 Exception: If your mailer is mangling patches then someone may ask
354 you to re-send them using MIME.
356 See Documentation/email-clients.txt for hints about configuring
357 your e-mail client so that it sends your patches untouched.
362 Large changes are not appropriate for mailing lists, and some
363 maintainers. If your patch, uncompressed, exceeds 300 kB in size,
364 it is preferred that you store your patch on an Internet-accessible
365 server, and provide instead a URL (link) pointing to your patch. But note
366 that if your patch exceeds 300 kB, it almost certainly needs to be broken up
369 8) Respond to review comments
370 -----------------------------
372 Your patch will almost certainly get comments from reviewers on ways in
373 which the patch can be improved. You must respond to those comments;
374 ignoring reviewers is a good way to get ignored in return. Review comments
375 or questions that do not lead to a code change should almost certainly
376 bring about a comment or changelog entry so that the next reviewer better
377 understands what is going on.
379 Be sure to tell the reviewers what changes you are making and to thank them
380 for their time. Code review is a tiring and time-consuming process, and
381 reviewers sometimes get grumpy. Even in that case, though, respond
382 politely and address the problems they have pointed out.
385 9) Don't get discouraged - or impatient
386 ---------------------------------------
388 After you have submitted your change, be patient and wait. Reviewers are
389 busy people and may not get to your patch right away.
391 Once upon a time, patches used to disappear into the void without comment,
392 but the development process works more smoothly than that now. You should
393 receive comments within a week or so; if that does not happen, make sure
394 that you have sent your patches to the right place. Wait for a minimum of
395 one week before resubmitting or pinging reviewers - possibly longer during
396 busy times like merge windows.
399 10) Include PATCH in the subject
400 --------------------------------
402 Due to high e-mail traffic to Linus, and to linux-kernel, it is common
403 convention to prefix your subject line with [PATCH]. This lets Linus
404 and other kernel developers more easily distinguish patches from other
412 To improve tracking of who did what, especially with patches that can
413 percolate to their final resting place in the kernel through several
414 layers of maintainers, we've introduced a "sign-off" procedure on
415 patches that are being emailed around.
417 The sign-off is a simple line at the end of the explanation for the
418 patch, which certifies that you wrote it or otherwise have the right to
419 pass it on as an open-source patch. The rules are pretty simple: if you
420 can certify the below:
422 Developer's Certificate of Origin 1.1
423 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
425 By making a contribution to this project, I certify that:
427 (a) The contribution was created in whole or in part by me and I
428 have the right to submit it under the open source license
429 indicated in the file; or
431 (b) The contribution is based upon previous work that, to the best
432 of my knowledge, is covered under an appropriate open source
433 license and I have the right under that license to submit that
434 work with modifications, whether created in whole or in part
435 by me, under the same open source license (unless I am
436 permitted to submit under a different license), as indicated
439 (c) The contribution was provided directly to me by some other
440 person who certified (a), (b) or (c) and I have not modified
443 (d) I understand and agree that this project and the contribution
444 are public and that a record of the contribution (including all
445 personal information I submit with it, including my sign-off) is
446 maintained indefinitely and may be redistributed consistent with
447 this project or the open source license(s) involved.
449 then you just add a line saying::
451 Signed-off-by: Random J Developer <random@developer.example.org>
453 using your real name (sorry, no pseudonyms or anonymous contributions.)
455 Some people also put extra tags at the end. They'll just be ignored for
456 now, but you can do this to mark internal company procedures or just
457 point out some special detail about the sign-off.
459 If you are a subsystem or branch maintainer, sometimes you need to slightly
460 modify patches you receive in order to merge them, because the code is not
461 exactly the same in your tree and the submitters'. If you stick strictly to
462 rule (c), you should ask the submitter to rediff, but this is a totally
463 counter-productive waste of time and energy. Rule (b) allows you to adjust
464 the code, but then it is very impolite to change one submitter's code and
465 make him endorse your bugs. To solve this problem, it is recommended that
466 you add a line between the last Signed-off-by header and yours, indicating
467 the nature of your changes. While there is nothing mandatory about this, it
468 seems like prepending the description with your mail and/or name, all
469 enclosed in square brackets, is noticeable enough to make it obvious that
470 you are responsible for last-minute changes. Example::
472 Signed-off-by: Random J Developer <random@developer.example.org>
473 [lucky@maintainer.example.org: struct foo moved from foo.c to foo.h]
474 Signed-off-by: Lucky K Maintainer <lucky@maintainer.example.org>
476 This practice is particularly helpful if you maintain a stable branch and
477 want at the same time to credit the author, track changes, merge the fix,
478 and protect the submitter from complaints. Note that under no circumstances
479 can you change the author's identity (the From header), as it is the one
480 which appears in the changelog.
482 Special note to back-porters: It seems to be a common and useful practice
483 to insert an indication of the origin of a patch at the top of the commit
484 message (just after the subject line) to facilitate tracking. For instance,
485 here's what we see in a 3.x-stable release::
487 Date: Tue Oct 7 07:26:38 2014 -0400
489 libata: Un-break ATA blacklist
491 commit 1c40279960bcd7d52dbdf1d466b20d24b99176c8 upstream.
493 And here's what might appear in an older kernel once a patch is backported::
495 Date: Tue May 13 22:12:27 2008 +0200
497 wireless, airo: waitbusy() won't delay
499 [backport of 2.6 commit b7acbdfbd1f277c1eb23f344f899cfa4cd0bf36a]
501 Whatever the format, this information provides a valuable help to people
502 tracking your trees, and to people trying to troubleshoot bugs in your
506 12) When to use Acked-by: and Cc:
507 ---------------------------------
509 The Signed-off-by: tag indicates that the signer was involved in the
510 development of the patch, or that he/she was in the patch's delivery path.
512 If a person was not directly involved in the preparation or handling of a
513 patch but wishes to signify and record their approval of it then they can
514 ask to have an Acked-by: line added to the patch's changelog.
516 Acked-by: is often used by the maintainer of the affected code when that
517 maintainer neither contributed to nor forwarded the patch.
519 Acked-by: is not as formal as Signed-off-by:. It is a record that the acker
520 has at least reviewed the patch and has indicated acceptance. Hence patch
521 mergers will sometimes manually convert an acker's "yep, looks good to me"
522 into an Acked-by: (but note that it is usually better to ask for an
525 Acked-by: does not necessarily indicate acknowledgement of the entire patch.
526 For example, if a patch affects multiple subsystems and has an Acked-by: from
527 one subsystem maintainer then this usually indicates acknowledgement of just
528 the part which affects that maintainer's code. Judgement should be used here.
529 When in doubt people should refer to the original discussion in the mailing
532 If a person has had the opportunity to comment on a patch, but has not
533 provided such comments, you may optionally add a ``Cc:`` tag to the patch.
534 This is the only tag which might be added without an explicit action by the
535 person it names - but it should indicate that this person was copied on the
536 patch. This tag documents that potentially interested parties
537 have been included in the discussion.
540 13) Using Reported-by:, Tested-by:, Reviewed-by:, Suggested-by: and Fixes:
541 --------------------------------------------------------------------------
543 The Reported-by tag gives credit to people who find bugs and report them and it
544 hopefully inspires them to help us again in the future. Please note that if
545 the bug was reported in private, then ask for permission first before using the
548 A Tested-by: tag indicates that the patch has been successfully tested (in
549 some environment) by the person named. This tag informs maintainers that
550 some testing has been performed, provides a means to locate testers for
551 future patches, and ensures credit for the testers.
553 Reviewed-by:, instead, indicates that the patch has been reviewed and found
554 acceptable according to the Reviewer's Statement:
556 Reviewer's statement of oversight
557 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
559 By offering my Reviewed-by: tag, I state that:
561 (a) I have carried out a technical review of this patch to
562 evaluate its appropriateness and readiness for inclusion into
565 (b) Any problems, concerns, or questions relating to the patch
566 have been communicated back to the submitter. I am satisfied
567 with the submitter's response to my comments.
569 (c) While there may be things that could be improved with this
570 submission, I believe that it is, at this time, (1) a
571 worthwhile modification to the kernel, and (2) free of known
572 issues which would argue against its inclusion.
574 (d) While I have reviewed the patch and believe it to be sound, I
575 do not (unless explicitly stated elsewhere) make any
576 warranties or guarantees that it will achieve its stated
577 purpose or function properly in any given situation.
579 A Reviewed-by tag is a statement of opinion that the patch is an
580 appropriate modification of the kernel without any remaining serious
581 technical issues. Any interested reviewer (who has done the work) can
582 offer a Reviewed-by tag for a patch. This tag serves to give credit to
583 reviewers and to inform maintainers of the degree of review which has been
584 done on the patch. Reviewed-by: tags, when supplied by reviewers known to
585 understand the subject area and to perform thorough reviews, will normally
586 increase the likelihood of your patch getting into the kernel.
588 A Suggested-by: tag indicates that the patch idea is suggested by the person
589 named and ensures credit to the person for the idea. Please note that this
590 tag should not be added without the reporter's permission, especially if the
591 idea was not posted in a public forum. That said, if we diligently credit our
592 idea reporters, they will, hopefully, be inspired to help us again in the
595 A Fixes: tag indicates that the patch fixes an issue in a previous commit. It
596 is used to make it easy to determine where a bug originated, which can help
597 review a bug fix. This tag also assists the stable kernel team in determining
598 which stable kernel versions should receive your fix. This is the preferred
599 method for indicating a bug fixed by the patch. See :ref:`describe_changes`
603 14) The canonical patch format
604 ------------------------------
606 This section describes how the patch itself should be formatted. Note
607 that, if you have your patches stored in a git repository, proper patch
608 formatting can be had with ``git format-patch``. The tools cannot create
609 the necessary text, though, so read the instructions below anyway.
611 The canonical patch subject line is::
613 Subject: [PATCH 001/123] subsystem: summary phrase
615 The canonical patch message body contains the following:
617 - A ``from`` line specifying the patch author (only needed if the person
618 sending the patch is not the author).
622 - The body of the explanation, line wrapped at 75 columns, which will
623 be copied to the permanent changelog to describe this patch.
625 - The ``Signed-off-by:`` lines, described above, which will
626 also go in the changelog.
628 - A marker line containing simply ``---``.
630 - Any additional comments not suitable for the changelog.
632 - The actual patch (diff output).
634 The Subject line format makes it very easy to sort the emails
635 alphabetically by subject line - pretty much any email reader will
636 support that - since because the sequence number is zero-padded,
637 the numerical and alphabetic sort is the same.
639 The ``subsystem`` in the email's Subject should identify which
640 area or subsystem of the kernel is being patched.
642 The ``summary phrase`` in the email's Subject should concisely
643 describe the patch which that email contains. The ``summary
644 phrase`` should not be a filename. Do not use the same ``summary
645 phrase`` for every patch in a whole patch series (where a ``patch
646 series`` is an ordered sequence of multiple, related patches).
648 Bear in mind that the ``summary phrase`` of your email becomes a
649 globally-unique identifier for that patch. It propagates all the way
650 into the git changelog. The ``summary phrase`` may later be used in
651 developer discussions which refer to the patch. People will want to
652 google for the ``summary phrase`` to read discussion regarding that
653 patch. It will also be the only thing that people may quickly see
654 when, two or three months later, they are going through perhaps
655 thousands of patches using tools such as ``gitk`` or "git log
658 For these reasons, the ``summary`` must be no more than 70-75
659 characters, and it must describe both what the patch changes, as well
660 as why the patch might be necessary. It is challenging to be both
661 succinct and descriptive, but that is what a well-written summary
664 The ``summary phrase`` may be prefixed by tags enclosed in square
665 brackets: "Subject: [PATCH <tag>...] <summary phrase>". The tags are
666 not considered part of the summary phrase, but describe how the patch
667 should be treated. Common tags might include a version descriptor if
668 the multiple versions of the patch have been sent out in response to
669 comments (i.e., "v1, v2, v3"), or "RFC" to indicate a request for
670 comments. If there are four patches in a patch series the individual
671 patches may be numbered like this: 1/4, 2/4, 3/4, 4/4. This assures
672 that developers understand the order in which the patches should be
673 applied and that they have reviewed or applied all of the patches in
676 A couple of example Subjects::
678 Subject: [PATCH 2/5] ext2: improve scalability of bitmap searching
679 Subject: [PATCH v2 01/27] x86: fix eflags tracking
681 The ``from`` line must be the very first line in the message body,
684 From: Original Author <author@example.com>
686 The ``from`` line specifies who will be credited as the author of the
687 patch in the permanent changelog. If the ``from`` line is missing,
688 then the ``From:`` line from the email header will be used to determine
689 the patch author in the changelog.
691 The explanation body will be committed to the permanent source
692 changelog, so should make sense to a competent reader who has long
693 since forgotten the immediate details of the discussion that might
694 have led to this patch. Including symptoms of the failure which the
695 patch addresses (kernel log messages, oops messages, etc.) is
696 especially useful for people who might be searching the commit logs
697 looking for the applicable patch. If a patch fixes a compile failure,
698 it may not be necessary to include _all_ of the compile failures; just
699 enough that it is likely that someone searching for the patch can find
700 it. As in the ``summary phrase``, it is important to be both succinct as
703 The ``---`` marker line serves the essential purpose of marking for patch
704 handling tools where the changelog message ends.
706 One good use for the additional comments after the ``---`` marker is for
707 a diffstat, to show what files have changed, and the number of
708 inserted and deleted lines per file. A diffstat is especially useful
709 on bigger patches. Other comments relevant only to the moment or the
710 maintainer, not suitable for the permanent changelog, should also go
711 here. A good example of such comments might be ``patch changelogs``
712 which describe what has changed between the v1 and v2 version of the
715 If you are going to include a diffstat after the ``---`` marker, please
716 use diffstat options ``-p 1 -w 70`` so that filenames are listed from
717 the top of the kernel source tree and don't use too much horizontal
718 space (easily fit in 80 columns, maybe with some indentation). (git
719 generates appropriate diffstats by default.)
721 See more details on the proper patch format in the following
724 .. _explicit_in_reply_to:
726 15) Explicit In-Reply-To headers
727 --------------------------------
729 It can be helpful to manually add In-Reply-To: headers to a patch
730 (e.g., when using ``git send-email``) to associate the patch with
731 previous relevant discussion, e.g. to link a bug fix to the email with
732 the bug report. However, for a multi-patch series, it is generally
733 best to avoid using In-Reply-To: to link to older versions of the
734 series. This way multiple versions of the patch don't become an
735 unmanageable forest of references in email clients. If a link is
736 helpful, you can use the https://lkml.kernel.org/ redirector (e.g., in
737 the cover email text) to link to an earlier version of the patch series.
740 16) Sending ``git pull`` requests
741 ---------------------------------
743 If you have a series of patches, it may be most convenient to have the
744 maintainer pull them directly into the subsystem repository with a
745 ``git pull`` operation. Note, however, that pulling patches from a developer
746 requires a higher degree of trust than taking patches from a mailing list.
747 As a result, many subsystem maintainers are reluctant to take pull
748 requests, especially from new, unknown developers. If in doubt you can use
749 the pull request as the cover letter for a normal posting of the patch
750 series, giving the maintainer the option of using either.
752 A pull request should have [GIT] or [PULL] in the subject line. The
753 request itself should include the repository name and the branch of
754 interest on a single line; it should look something like::
758 git://jdelvare.pck.nerim.net/jdelvare-2.6 i2c-for-linus
760 to get these changes:
762 A pull request should also include an overall message saying what will be
763 included in the request, a ``git shortlog`` listing of the patches
764 themselves, and a diffstat showing the overall effect of the patch series.
765 The easiest way to get all this information together is, of course, to let
766 git do it for you with the ``git request-pull`` command.
768 Some maintainers (including Linus) want to see pull requests from signed
769 commits; that increases their confidence that the request actually came
770 from you. Linus, in particular, will not pull from public hosting sites
771 like GitHub in the absence of a signed tag.
773 The first step toward creating such tags is to make a GNUPG key and get it
774 signed by one or more core kernel developers. This step can be hard for
775 new developers, but there is no way around it. Attending conferences can
776 be a good way to find developers who can sign your key.
778 Once you have prepared a patch series in git that you wish to have somebody
779 pull, create a signed tag with ``git tag -s``. This will create a new tag
780 identifying the last commit in the series and containing a signature
781 created with your private key. You will also have the opportunity to add a
782 changelog-style message to the tag; this is an ideal place to describe the
783 effects of the pull request as a whole.
785 If the tree the maintainer will be pulling from is not the repository you
786 are working from, don't forget to push the signed tag explicitly to the
789 When generating your pull request, use the signed tag as the target. A
790 command like this will do the trick::
792 git request-pull master git://my.public.tree/linux.git my-signed-tag
798 Andrew Morton, "The perfect patch" (tpp).
799 <http://www.ozlabs.org/~akpm/stuff/tpp.txt>
801 Jeff Garzik, "Linux kernel patch submission format".
802 <http://linux.yyz.us/patch-format.html>
804 Greg Kroah-Hartman, "How to piss off a kernel subsystem maintainer".
805 <http://www.kroah.com/log/linux/maintainer.html>
806 <http://www.kroah.com/log/linux/maintainer-02.html>
807 <http://www.kroah.com/log/linux/maintainer-03.html>
808 <http://www.kroah.com/log/linux/maintainer-04.html>
809 <http://www.kroah.com/log/linux/maintainer-05.html>
810 <http://www.kroah.com/log/linux/maintainer-06.html>
812 NO!!!! No more huge patch bombs to linux-kernel@vger.kernel.org people!
813 <https://lkml.org/lkml/2005/7/11/336>
815 Kernel Documentation/CodingStyle:
816 <Documentation/CodingStyle>
818 Linus Torvalds's mail on the canonical patch format:
819 <http://lkml.org/lkml/2005/4/7/183>
821 Andi Kleen, "On submitting kernel patches"
822 Some strategies to get difficult or controversial changes in.
823 http://halobates.de/on-submitting-patches.pdf