1 .. _submittingpatches:
2
3 Submitting patches: the essential guide to getting your code into the kernel
4 ============================================================================
5
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.
10
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/process/development-process.rst. Also, read
14 Documentation/process/submit-checklist.rst
15 for a list of items to check before submitting code.
16 For device tree binding patches, read
17 Documentation/devicetree/bindings/submitting-patches.rst.
18
19 This documentation assumes that you're using ``git`` to prepare your patches.
20 If you're unfamiliar with ``git``, you would be well-advised to learn how to
21 use it, it will make your life as a kernel developer and in general much
22 easier.
23
24 Some subsystems and maintainer trees have additional information about
25 their workflow and expectations, see
26 :ref:`Documentation/process/maintainer-handbooks.rst <maintainer_handbooks_main>`.
27
28 Obtain a current source tree
29 ----------------------------
30
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::
34
35 git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
36
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.
42
43 .. _describe_changes:
44
45 Describe your changes
46 ---------------------
47
48 Describe your problem. Whether your patch is a one-line bug fix or
49 5000 lines of a new feature, there must be an underlying problem that
50 motivated you to do this work. Convince the reviewer that there is a
51 problem worth fixing and that it makes sense for them to read past the
52 first paragraph.
53
54 Describe user-visible impact. Straight up crashes and lockups are
55 pretty convincing, but not all bugs are that blatant. Even if the
56 problem was spotted during code review, describe the impact you think
57 it can have on users. Keep in mind that the majority of Linux
58 installations run kernels from secondary stable trees or
59 vendor/product-specific trees that cherry-pick only specific patches
60 from upstream, so include anything that could help route your change
61 downstream: provoking circumstances, excerpts from dmesg, crash
62 descriptions, performance regressions, latency spikes, lockups, etc.
63
64 Quantify optimizations and trade-offs. If you claim improvements in
65 performance, memory consumption, stack footprint, or binary size,
66 include numbers that back them up. But also describe non-obvious
67 costs. Optimizations usually aren't free but trade-offs between CPU,
68 memory, and readability; or, when it comes to heuristics, between
69 different workloads. Describe the expected downsides of your
70 optimization so that the reviewer can weigh costs against benefits.
71
72 Once the problem is established, describe what you are actually doing
73 about it in technical detail. It's important to describe the change
74 in plain English for the reviewer to verify that the code is behaving
75 as you intend it to.
76
77 The maintainer will thank you if you write your patch description in a
78 form which can be easily pulled into Linux's source code management
79 system, ``git``, as a "commit log". See :ref:`the_canonical_patch_format`.
80
81 Solve only one problem per patch. If your description starts to get
82 long, that's a sign that you probably need to split up your patch.
83 See :ref:`split_changes`.
84
85 When you submit or resubmit a patch or patch series, include the
86 complete patch description and justification for it. Don't just
87 say that this is version N of the patch (series). Don't expect the
88 subsystem maintainer to refer back to earlier patch versions or referenced
89 URLs to find the patch description and put that into the patch.
90 I.e., the patch (series) and its description should be self-contained.
91 This benefits both the maintainers and reviewers. Some reviewers
92 probably didn't even receive earlier versions of the patch.
93
94 Describe your changes in imperative mood, e.g. "make xyzzy do frotz"
95 instead of "[This patch] makes xyzzy do frotz" or "[I] changed xyzzy
96 to do frotz", as if you are giving orders to the codebase to change
97 its behaviour.
98
99 If you want to refer to a specific commit, don't just refer to the
100 SHA-1 ID of the commit. Please also include the oneline summary of
101 the commit, to make it easier for reviewers to know what it is about.
102 Example::
103
104 Commit e21d2170f36602ae2708 ("video: remove unnecessary
105 platform_set_drvdata()") removed the unnecessary
106 platform_set_drvdata(), but left the variable "dev" unused,
107 delete it.
108
109 You should also be sure to use at least the first twelve characters of the
110 SHA-1 ID. The kernel repository holds a *lot* of objects, making
111 collisions with shorter IDs a real possibility. Bear in mind that, even if
112 there is no collision with your six-character ID now, that condition may
113 change five years from now.
114
115 If related discussions or any other background information behind the change
116 can be found on the web, add 'Link:' tags pointing to it. If the patch is a
117 result of some earlier mailing list discussions or something documented on the
118 web, point to it.
119
120 When linking to mailing list archives, preferably use the lore.kernel.org
121 message archiver service. To create the link URL, use the contents of the
122 ``Message-ID`` header of the message without the surrounding angle brackets.
123 For example::
124
125 Link: https://lore.kernel.org/30th.anniversary.repost@klaava.Helsinki.FI
126
127 Please check the link to make sure that it is actually working and points
128 to the relevant message.
129
130 However, try to make your explanation understandable without external
131 resources. In addition to giving a URL to a mailing list archive or bug,
132 summarize the relevant points of the discussion that led to the
133 patch as submitted.
134
135 In case your patch fixes a bug, use the 'Closes:' tag with a URL referencing
136 the report in the mailing list archives or a public bug tracker. For example::
137
138 Closes: https://example.com/issues/1234
139
140 Some bug trackers have the ability to close issues automatically when a
141 commit with such a tag is applied. Some bots monitoring mailing lists can
142 also track such tags and take certain actions. Private bug trackers and
143 invalid URLs are forbidden.
144
145 If your patch fixes a bug in a specific commit, e.g. you found an issue using
146 ``git bisect``, please use the 'Fixes:' tag with the first 12 characters of
147 the SHA-1 ID, and the one line summary. Do not split the tag across multiple
148 lines, tags are exempt from the "wrap at 75 columns" rule in order to simplify
149 parsing scripts. For example::
150
151 Fixes: 54a4f0239f2e ("KVM: MMU: make kvm_mmu_zap_page() return the number of pages it actually freed")
152
153 The following ``git config`` settings can be used to add a pretty format for
154 outputting the above style in the ``git log`` or ``git show`` commands::
155
156 [core]
157 abbrev = 12
158 [pretty]
159 fixes = Fixes: %h (\"%s\")
160
161 An example call::
162
163 $ git log -1 --pretty=fixes 54a4f0239f2e
164 Fixes: 54a4f0239f2e ("KVM: MMU: make kvm_mmu_zap_page() return the number of pages it actually freed")
165
166 .. _split_changes:
167
168 Separate your changes
169 ---------------------
170
171 Separate each **logical change** into a separate patch.
172
173 For example, if your changes include both bug fixes and performance
174 enhancements for a single driver, separate those changes into two
175 or more patches. If your changes include an API update, and a new
176 driver which uses that new API, separate those into two patches.
177
178 On the other hand, if you make a single change to numerous files,
179 group those changes into a single patch. Thus a single logical change
180 is contained within a single patch.
181
182 The point to remember is that each patch should make an easily understood
183 change that can be verified by reviewers. Each patch should be justifiable
184 on its own merits.
185
186 If one patch depends on another patch in order for a change to be
187 complete, that is OK. Simply note **"this patch depends on patch X"**
188 in your patch description.
189
190 When dividing your change into a series of patches, take special care to
191 ensure that the kernel builds and runs properly after each patch in the
192 series. Developers using ``git bisect`` to track down a problem can end up
193 splitting your patch series at any point; they will not thank you if you
194 introduce bugs in the middle.
195
196 If you cannot condense your patch set into a smaller set of patches,
197 then only post say 15 or so at a time and wait for review and integration.
198
199
200
201 Style-check your changes
202 ------------------------
203
204 Check your patch for basic style violations, details of which can be
205 found in Documentation/process/coding-style.rst.
206 Failure to do so simply wastes
207 the reviewers time and will get your patch rejected, probably
208 without even being read.
209
210 One significant exception is when moving code from one file to
211 another -- in this case you should not modify the moved code at all in
212 the same patch which moves it. This clearly delineates the act of
213 moving the code and your changes. This greatly aids review of the
214 actual differences and allows tools to better track the history of
215 the code itself.
216
217 Check your patches with the patch style checker prior to submission
218 (scripts/checkpatch.pl). Note, though, that the style checker should be
219 viewed as a guide, not as a replacement for human judgment. If your code
220 looks better with a violation then its probably best left alone.
221
222 The checker reports at three levels:
223 - ERROR: things that are very likely to be wrong
224 - WARNING: things requiring careful review
225 - CHECK: things requiring thought
226
227 You should be able to justify all violations that remain in your
228 patch.
229
230
231 Select the recipients for your patch
232 ------------------------------------
233
234 You should always copy the appropriate subsystem maintainer(s) and list(s) on
235 any patch to code that they maintain; look through the MAINTAINERS file and the
236 source code revision history to see who those maintainers are. The script
237 scripts/get_maintainer.pl can be very useful at this step (pass paths to your
238 patches as arguments to scripts/get_maintainer.pl). If you cannot find a
239 maintainer for the subsystem you are working on, Andrew Morton
240 (akpm@linux-foundation.org) serves as a maintainer of last resort.
241
242 linux-kernel@vger.kernel.org should be used by default for all patches, but the
243 volume on that list has caused a number of developers to tune it out. Please
244 do not spam unrelated lists and unrelated people, though.
245
246 Many kernel-related lists are hosted at kernel.org; you can find a list
247 of them at https://subspace.kernel.org. There are kernel-related lists
248 hosted elsewhere as well, though.
249
250 Linus Torvalds is the final arbiter of all changes accepted into the
251 Linux kernel. His e-mail address is <torvalds@linux-foundation.org>.
252 He gets a lot of e-mail, and, at this point, very few patches go through
253 Linus directly, so typically you should do your best to -avoid-
254 sending him e-mail.
255
256 If you have a patch that fixes an exploitable security bug, send that patch
257 to security@kernel.org. For severe bugs, a short embargo may be considered
258 to allow distributors to get the patch out to users; in such cases,
259 obviously, the patch should not be sent to any public lists. See also
260 Documentation/process/security-bugs.rst.
261
262 Patches that fix a severe bug in a released kernel should be directed
263 toward the stable maintainers by putting a line like this::
264
265 Cc: stable@vger.kernel.org
266
267 into the sign-off area of your patch (note, NOT an email recipient). You
268 should also read Documentation/process/stable-kernel-rules.rst
269 in addition to this document.
270
271 If changes affect userland-kernel interfaces, please send the MAN-PAGES
272 maintainer (as listed in the MAINTAINERS file) a man-pages patch, or at
273 least a notification of the change, so that some information makes its way
274 into the manual pages. User-space API changes should also be copied to
275 linux-api@vger.kernel.org.
276
277
278 No MIME, no links, no compression, no attachments. Just plain text
279 -------------------------------------------------------------------
280
281 Linus and other kernel developers need to be able to read and comment
282 on the changes you are submitting. It is important for a kernel
283 developer to be able to "quote" your changes, using standard e-mail
284 tools, so that they may comment on specific portions of your code.
285
286 For this reason, all patches should be submitted by e-mail "inline". The
287 easiest way to do this is with ``git send-email``, which is strongly
288 recommended. An interactive tutorial for ``git send-email`` is available at
289 https://git-send-email.io.
290
291 If you choose not to use ``git send-email``:
292
293 .. warning::
294
295 Be wary of your editor's word-wrap corrupting your patch,
296 if you choose to cut-n-paste your patch.
297
298 Do not attach the patch as a MIME attachment, compressed or not.
299 Many popular e-mail applications will not always transmit a MIME
300 attachment as plain text, making it impossible to comment on your
301 code. A MIME attachment also takes Linus a bit more time to process,
302 decreasing the likelihood of your MIME-attached change being accepted.
303
304 Exception: If your mailer is mangling patches then someone may ask
305 you to re-send them using MIME.
306
307 See Documentation/process/email-clients.rst for hints about configuring
308 your e-mail client so that it sends your patches untouched.
309
310 Respond to review comments
311 --------------------------
312
313 Your patch will almost certainly get comments from reviewers on ways in
314 which the patch can be improved, in the form of a reply to your email. You must
315 respond to those comments; ignoring reviewers is a good way to get ignored in
316 return. You can simply reply to their emails to answer their comments. Review
317 comments or questions that do not lead to a code change should almost certainly
318 bring about a comment or changelog entry so that the next reviewer better
319 understands what is going on.
320
321 Be sure to tell the reviewers what changes you are making and to thank them
322 for their time. Code review is a tiring and time-consuming process, and
323 reviewers sometimes get grumpy. Even in that case, though, respond
324 politely and address the problems they have pointed out. When sending a next
325 version, add a ``patch changelog`` to the cover letter or to individual patches
326 explaining difference against previous submission (see
327 :ref:`the_canonical_patch_format`).
328 Notify people that commented on your patch about new versions by adding them to
329 the patches CC list.
330
331 See Documentation/process/email-clients.rst for recommendations on email
332 clients and mailing list etiquette.
333
334 .. _interleaved_replies:
335
336 Use trimmed interleaved replies in email discussions
337 ----------------------------------------------------
338 Top-posting is strongly discouraged in Linux kernel development
339 discussions. Interleaved (or "inline") replies make conversations much
340 easier to follow. For more details see:
341 https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
342
343 As is frequently quoted on the mailing list::
344
345 A: http://en.wikipedia.org/wiki/Top_post
346 Q: Were do I find info about this thing called top-posting?
347 A: Because it messes up the order in which people normally read text.
348 Q: Why is top-posting such a bad thing?
349 A: Top-posting.
350 Q: What is the most annoying thing in e-mail?
351
352 Similarly, please trim all unneeded quotations that aren't relevant
353 to your reply. This makes responses easier to find, and saves time and
354 space. For more details see: http://daringfireball.net/2007/07/on_top ::
355
356 A: No.
357 Q: Should I include quotations after my reply?
358
359 .. _resend_reminders:
360
361 Don't get discouraged - or impatient
362 ------------------------------------
363
364 After you have submitted your change, be patient and wait. Reviewers are
365 busy people and may not get to your patch right away.
366
367 Once upon a time, patches used to disappear into the void without comment,
368 but the development process works more smoothly than that now. You should
369 receive comments within a few weeks (typically 2-3); if that does not
370 happen, make sure that you have sent your patches to the right place.
371 Wait for a minimum of one week before resubmitting or pinging reviewers
372 - possibly longer during busy times like merge windows.
373
374 It's also ok to resend the patch or the patch series after a couple of
375 weeks with the word "RESEND" added to the subject line::
376
377 [PATCH Vx RESEND] sub/sys: Condensed patch summary
378
379 Don't add "RESEND" when you are submitting a modified version of your
380 patch or patch series - "RESEND" only applies to resubmission of a
381 patch or patch series which have not been modified in any way from the
382 previous submission.
383
384
385 Include PATCH in the subject
386 -----------------------------
387
388 Due to high e-mail traffic to Linus, and to linux-kernel, it is common
389 convention to prefix your subject line with [PATCH]. This lets Linus
390 and other kernel developers more easily distinguish patches from other
391 e-mail discussions.
392
393 ``git send-email`` will do this for you automatically.
394
395
396 Sign your work - the Developer's Certificate of Origin
397 ------------------------------------------------------
398
399 To improve tracking of who did what, especially with patches that can
400 percolate to their final resting place in the kernel through several
401 layers of maintainers, we've introduced a "sign-off" procedure on
402 patches that are being emailed around.
403
404 The sign-off is a simple line at the end of the explanation for the
405 patch, which certifies that you wrote it or otherwise have the right to
406 pass it on as an open-source patch. The rules are pretty simple: if you
407 can certify the below:
408
409 Developer's Certificate of Origin 1.1
410 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
411
412 By making a contribution to this project, I certify that:
413
414 (a) The contribution was created in whole or in part by me and I
415 have the right to submit it under the open source license
416 indicated in the file; or
417
418 (b) The contribution is based upon previous work that, to the best
419 of my knowledge, is covered under an appropriate open source
420 license and I have the right under that license to submit that
421 work with modifications, whether created in whole or in part
422 by me, under the same open source license (unless I am
423 permitted to submit under a different license), as indicated
424 in the file; or
425
426 (c) The contribution was provided directly to me by some other
427 person who certified (a), (b) or (c) and I have not modified
428 it.
429
430 (d) I understand and agree that this project and the contribution
431 are public and that a record of the contribution (including all
432 personal information I submit with it, including my sign-off) is
433 maintained indefinitely and may be redistributed consistent with
434 this project or the open source license(s) involved.
435
436 then you just add a line saying::
437
438 Signed-off-by: Random J Developer <random@developer.example.org>
439
440 using a known identity (sorry, no anonymous contributions.)
441 This will be done for you automatically if you use ``git commit -s``.
442 Reverts should also include "Signed-off-by". ``git revert -s`` does that
443 for you.
444
445 Some people also put extra tags at the end. They'll just be ignored for
446 now, but you can do this to mark internal company procedures or just
447 point out some special detail about the sign-off.
448
449 Any further SoBs (Signed-off-by:'s) following the author's SoB are from
450 people handling and transporting the patch, but were not involved in its
451 development. SoB chains should reflect the **real** route a patch took
452 as it was propagated to the maintainers and ultimately to Linus, with
453 the first SoB entry signalling primary authorship of a single author.
454
455
456 When to use Acked-by:, Cc:, and Co-developed-by:
457 ------------------------------------------------
458
459 The Signed-off-by: tag indicates that the signer was involved in the
460 development of the patch, or that he/she was in the patch's delivery path.
461
462 If a person was not directly involved in the preparation or handling of a
463 patch but wishes to signify and record their approval of it then they can
464 ask to have an Acked-by: line added to the patch's changelog.
465
466 Acked-by: is often used by the maintainer of the affected code when that
467 maintainer neither contributed to nor forwarded the patch.
468
469 Acked-by: is not as formal as Signed-off-by:. It is a record that the acker
470 has at least reviewed the patch and has indicated acceptance. Hence patch
471 mergers will sometimes manually convert an acker's "yep, looks good to me"
472 into an Acked-by: (but note that it is usually better to ask for an
473 explicit ack).
474
475 Acked-by: does not necessarily indicate acknowledgement of the entire patch.
476 For example, if a patch affects multiple subsystems and has an Acked-by: from
477 one subsystem maintainer then this usually indicates acknowledgement of just
478 the part which affects that maintainer's code. Judgement should be used here.
479 When in doubt people should refer to the original discussion in the mailing
480 list archives.
481
482 If a person has had the opportunity to comment on a patch, but has not
483 provided such comments, you may optionally add a ``Cc:`` tag to the patch.
484 This is the only tag which might be added without an explicit action by the
485 person it names - but it should indicate that this person was copied on the
486 patch. This tag documents that potentially interested parties
487 have been included in the discussion.
488
489 Co-developed-by: states that the patch was co-created by multiple developers;
490 it is used to give attribution to co-authors (in addition to the author
491 attributed by the From: tag) when several people work on a single patch. Since
492 Co-developed-by: denotes authorship, every Co-developed-by: must be immediately
493 followed by a Signed-off-by: of the associated co-author. Standard sign-off
494 procedure applies, i.e. the ordering of Signed-off-by: tags should reflect the
495 chronological history of the patch insofar as possible, regardless of whether
496 the author is attributed via From: or Co-developed-by:. Notably, the last
497 Signed-off-by: must always be that of the developer submitting the patch.
498
499 Note, the From: tag is optional when the From: author is also the person (and
500 email) listed in the From: line of the email header.
501
502 Example of a patch submitted by the From: author::
503
504 <changelog>
505
506 Co-developed-by: First Co-Author <first@coauthor.example.org>
507 Signed-off-by: First Co-Author <first@coauthor.example.org>
508 Co-developed-by: Second Co-Author <second@coauthor.example.org>
509 Signed-off-by: Second Co-Author <second@coauthor.example.org>
510 Signed-off-by: From Author <from@author.example.org>
511
512 Example of a patch submitted by a Co-developed-by: author::
513
514 From: From Author <from@author.example.org>
515
516 <changelog>
517
518 Co-developed-by: Random Co-Author <random@coauthor.example.org>
519 Signed-off-by: Random Co-Author <random@coauthor.example.org>
520 Signed-off-by: From Author <from@author.example.org>
521 Co-developed-by: Submitting Co-Author <sub@coauthor.example.org>
522 Signed-off-by: Submitting Co-Author <sub@coauthor.example.org>
523
524
525 Using Reported-by:, Tested-by:, Reviewed-by:, Suggested-by: and Fixes:
526 ----------------------------------------------------------------------
527
528 The Reported-by tag gives credit to people who find bugs and report them and it
529 hopefully inspires them to help us again in the future. The tag is intended for
530 bugs; please do not use it to credit feature requests. The tag should be
531 followed by a Closes: tag pointing to the report, unless the report is not
532 available on the web. The Link: tag can be used instead of Closes: if the patch
533 fixes a part of the issue(s) being reported. Please note that if the bug was
534 reported in private, then ask for permission first before using the Reported-by
535 tag.
536
537 A Tested-by: tag indicates that the patch has been successfully tested (in
538 some environment) by the person named. This tag informs maintainers that
539 some testing has been performed, provides a means to locate testers for
540 future patches, and ensures credit for the testers.
541
542 Reviewed-by:, instead, indicates that the patch has been reviewed and found
543 acceptable according to the Reviewer's Statement:
544
545 Reviewer's statement of oversight
546 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
547
548 By offering my Reviewed-by: tag, I state that:
549
550 (a) I have carried out a technical review of this patch to
551 evaluate its appropriateness and readiness for inclusion into
552 the mainline kernel.
553
554 (b) Any problems, concerns, or questions relating to the patch
555 have been communicated back to the submitter. I am satisfied
556 with the submitter's response to my comments.
557
558 (c) While there may be things that could be improved with this
559 submission, I believe that it is, at this time, (1) a
560 worthwhile modification to the kernel, and (2) free of known
561 issues which would argue against its inclusion.
562
563 (d) While I have reviewed the patch and believe it to be sound, I
564 do not (unless explicitly stated elsewhere) make any
565 warranties or guarantees that it will achieve its stated
566 purpose or function properly in any given situation.
567
568 A Reviewed-by tag is a statement of opinion that the patch is an
569 appropriate modification of the kernel without any remaining serious
570 technical issues. Any interested reviewer (who has done the work) can
571 offer a Reviewed-by tag for a patch. This tag serves to give credit to
572 reviewers and to inform maintainers of the degree of review which has been
573 done on the patch. Reviewed-by: tags, when supplied by reviewers known to
574 understand the subject area and to perform thorough reviews, will normally
575 increase the likelihood of your patch getting into the kernel.
576
577 Both Tested-by and Reviewed-by tags, once received on mailing list from tester
578 or reviewer, should be added by author to the applicable patches when sending
579 next versions. However if the patch has changed substantially in following
580 version, these tags might not be applicable anymore and thus should be removed.
581 Usually removal of someone's Tested-by or Reviewed-by tags should be mentioned
582 in the patch changelog (after the '---' separator).
583
584 A Suggested-by: tag indicates that the patch idea is suggested by the person
585 named and ensures credit to the person for the idea. Please note that this
586 tag should not be added without the reporter's permission, especially if the
587 idea was not posted in a public forum. That said, if we diligently credit our
588 idea reporters, they will, hopefully, be inspired to help us again in the
589 future.
590
591 A Fixes: tag indicates that the patch fixes an issue in a previous commit. It
592 is used to make it easy to determine where a bug originated, which can help
593 review a bug fix. This tag also assists the stable kernel team in determining
594 which stable kernel versions should receive your fix. This is the preferred
595 method for indicating a bug fixed by the patch. See :ref:`describe_changes`
596 for more details.
597
598 Note: Attaching a Fixes: tag does not subvert the stable kernel rules
599 process nor the requirement to Cc: stable@vger.kernel.org on all stable
600 patch candidates. For more information, please read
601 Documentation/process/stable-kernel-rules.rst.
602
603 .. _the_canonical_patch_format:
604
605 The canonical patch format
606 --------------------------
607
608 This section describes how the patch itself should be formatted. Note
609 that, if you have your patches stored in a ``git`` repository, proper patch
610 formatting can be had with ``git format-patch``. The tools cannot create
611 the necessary text, though, so read the instructions below anyway.
612
613 The canonical patch subject line is::
614
615 Subject: [PATCH 001/123] subsystem: summary phrase
616
617 The canonical patch message body contains the following:
618
619 - A ``from`` line specifying the patch author, followed by an empty
620 line (only needed if the person sending the patch is not the author).
621
622 - The body of the explanation, line wrapped at 75 columns, which will
623 be copied to the permanent changelog to describe this patch.
624
625 - An empty line.
626
627 - The ``Signed-off-by:`` lines, described above, which will
628 also go in the changelog.
629
630 - A marker line containing simply ``---``.
631
632 - Any additional comments not suitable for the changelog.
633
634 - The actual patch (``diff`` output).
635
636 The Subject line format makes it very easy to sort the emails
637 alphabetically by subject line - pretty much any email reader will
638 support that - since because the sequence number is zero-padded,
639 the numerical and alphabetic sort is the same.
640
641 The ``subsystem`` in the email's Subject should identify which
642 area or subsystem of the kernel is being patched.
643
644 The ``summary phrase`` in the email's Subject should concisely
645 describe the patch which that email contains. The ``summary
646 phrase`` should not be a filename. Do not use the same ``summary
647 phrase`` for every patch in a whole patch series (where a ``patch
648 series`` is an ordered sequence of multiple, related patches).
649
650 Bear in mind that the ``summary phrase`` of your email becomes a
651 globally-unique identifier for that patch. It propagates all the way
652 into the ``git`` changelog. The ``summary phrase`` may later be used in
653 developer discussions which refer to the patch. People will want to
654 google for the ``summary phrase`` to read discussion regarding that
655 patch. It will also be the only thing that people may quickly see
656 when, two or three months later, they are going through perhaps
657 thousands of patches using tools such as ``gitk`` or ``git log
658 --oneline``.
659
660 For these reasons, the ``summary`` must be no more than 70-75
661 characters, and it must describe both what the patch changes, as well
662 as why the patch might be necessary. It is challenging to be both
663 succinct and descriptive, but that is what a well-written summary
664 should do.
665
666 The ``summary phrase`` may be prefixed by tags enclosed in square
667 brackets: "Subject: [PATCH <tag>...] <summary phrase>". The tags are
668 not considered part of the summary phrase, but describe how the patch
669 should be treated. Common tags might include a version descriptor if
670 the multiple versions of the patch have been sent out in response to
671 comments (i.e., "v1, v2, v3"), or "RFC" to indicate a request for
672 comments.
673
674 If there are four patches in a patch series the individual patches may
675 be numbered like this: 1/4, 2/4, 3/4, 4/4. This assures that developers
676 understand the order in which the patches should be applied and that
677 they have reviewed or applied all of the patches in the patch series.
678
679 Here are some good example Subjects::
680
681 Subject: [PATCH 2/5] ext2: improve scalability of bitmap searching
682 Subject: [PATCH v2 01/27] x86: fix eflags tracking
683 Subject: [PATCH v2] sub/sys: Condensed patch summary
684 Subject: [PATCH v2 M/N] sub/sys: Condensed patch summary
685
686 The ``from`` line must be the very first line in the message body,
687 and has the form:
688
689 From: Patch Author <author@example.com>
690
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.
695
696 The explanation body will be committed to the permanent source
697 changelog, so should make sense to a competent reader who has long since
698 forgotten the immediate details of the discussion that might have led to
699 this patch. Including symptoms of the failure which the patch addresses
700 (kernel log messages, oops messages, etc.) are especially useful for
701 people who might be searching the commit logs looking for the applicable
702 patch. The text should be written in such detail so that when read
703 weeks, months or even years later, it can give the reader the needed
704 details to grasp the reasoning for **why** the patch was created.
705
706 If a patch fixes a compile failure, it may not be necessary to include
707 _all_ of the compile failures; just enough that it is likely that
708 someone searching for the patch can find it. As in the ``summary
709 phrase``, it is important to be both succinct as well as descriptive.
710
711 The ``---`` marker line serves the essential purpose of marking for
712 patch handling tools where the changelog message ends.
713
714 One good use for the additional comments after the ``---`` marker is
715 for a ``diffstat``, to show what files have changed, and the number of
716 inserted and deleted lines per file. A ``diffstat`` is especially useful
717 on bigger patches. If you are going to include a ``diffstat`` after the
718 ``---`` marker, please use ``diffstat`` options ``-p 1 -w 70`` so that
719 filenames are listed from the top of the kernel source tree and don't
720 use too much horizontal space (easily fit in 80 columns, maybe with some
721 indentation). (``git`` generates appropriate diffstats by default.)
722
723 Other comments relevant only to the moment or the maintainer, not
724 suitable for the permanent changelog, should also go here. A good
725 example of such comments might be ``patch changelogs`` which describe
726 what has changed between the v1 and v2 version of the patch.
727
728 Please put this information **after** the ``---`` line which separates
729 the changelog from the rest of the patch. The version information is
730 not part of the changelog which gets committed to the git tree. It is
731 additional information for the reviewers. If it's placed above the
732 commit tags, it needs manual interaction to remove it. If it is below
733 the separator line, it gets automatically stripped off when applying the
734 patch::
735
736 <commit message>
737 ...
738 Signed-off-by: Author <author@mail>
739 ---
740 V2 -> V3: Removed redundant helper function
741 V1 -> V2: Cleaned up coding style and addressed review comments
742
743 path/to/file | 5+++--
744 ...
745
746 See more details on the proper patch format in the following
747 references.
748
749 .. _backtraces:
750
751 Backtraces in commit messages
752 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
753
754 Backtraces help document the call chain leading to a problem. However,
755 not all backtraces are helpful. For example, early boot call chains are
756 unique and obvious. Copying the full dmesg output verbatim, however,
757 adds distracting information like timestamps, module lists, register and
758 stack dumps.
759
760 Therefore, the most useful backtraces should distill the relevant
761 information from the dump, which makes it easier to focus on the real
762 issue. Here is an example of a well-trimmed backtrace::
763
764 unchecked MSR access error: WRMSR to 0xd51 (tried to write 0x0000000000000064)
765 at rIP: 0xffffffffae059994 (native_write_msr+0x4/0x20)
766 Call Trace:
767 mba_wrmsr
768 update_domains
769 rdtgroup_mkdir
770
771 .. _explicit_in_reply_to:
772
773 Explicit In-Reply-To headers
774 ----------------------------
775
776 It can be helpful to manually add In-Reply-To: headers to a patch
777 (e.g., when using ``git send-email``) to associate the patch with
778 previous relevant discussion, e.g. to link a bug fix to the email with
779 the bug report. However, for a multi-patch series, it is generally
780 best to avoid using In-Reply-To: to link to older versions of the
781 series. This way multiple versions of the patch don't become an
782 unmanageable forest of references in email clients. If a link is
783 helpful, you can use the https://lore.kernel.org/ redirector (e.g., in
784 the cover email text) to link to an earlier version of the patch series.
785
786
787 Providing base tree information
788 -------------------------------
789
790 When other developers receive your patches and start the review process,
791 it is absolutely necessary for them to know what is the base
792 commit/branch your work applies on, considering the sheer amount of
793 maintainer trees present nowadays. Note again the **T:** entry in the
794 MAINTAINERS file explained above.
795
796 This is even more important for automated CI processes that attempt to
797 run a series of tests in order to establish the quality of your
798 submission before the maintainer starts the review.
799
800 If you are using ``git format-patch`` to generate your patches, you can
801 automatically include the base tree information in your submission by
802 using the ``--base`` flag. The easiest and most convenient way to use
803 this option is with topical branches::
804
805 $ git checkout -t -b my-topical-branch master
806 Branch 'my-topical-branch' set up to track local branch 'master'.
807 Switched to a new branch 'my-topical-branch'
808
809 [perform your edits and commits]
810
811 $ git format-patch --base=auto --cover-letter -o outgoing/ master
812 outgoing/0000-cover-letter.patch
813 outgoing/0001-First-Commit.patch
814 outgoing/...
815
816 When you open ``outgoing/0000-cover-letter.patch`` for editing, you will
817 notice that it will have the ``base-commit:`` trailer at the very
818 bottom, which provides the reviewer and the CI tools enough information
819 to properly perform ``git am`` without worrying about conflicts::
820
821 $ git checkout -b patch-review [base-commit-id]
822 Switched to a new branch 'patch-review'
823 $ git am patches.mbox
824 Applying: First Commit
825 Applying: ...
826
827 Please see ``man git-format-patch`` for more information about this
828 option.
829
830 .. note::
831
832 The ``--base`` feature was introduced in git version 2.9.0.
833
834 If you are not using git to format your patches, you can still include
835 the same ``base-commit`` trailer to indicate the commit hash of the tree
836 on which your work is based. You should add it either in the cover
837 letter or in the first patch of the series and it should be placed
838 either below the ``---`` line or at the very bottom of all other
839 content, right before your email signature.
840
841 Make sure that base commit is in an official maintainer/mainline tree
842 and not in some internal, accessible only to you tree - otherwise it
843 would be worthless.
844
845 Tooling
846 -------
847
848 Many of the technical aspects of this process can be automated using
849 b4, documented at <https://b4.docs.kernel.org/en/latest/>. This can
850 help with things like tracking dependencies, running checkpatch and
851 with formatting and sending mails.
852
853 References
854 ----------
855
856 Andrew Morton, "The perfect patch" (tpp).
857 <https://www.ozlabs.org/~akpm/stuff/tpp.txt>
858
859 Jeff Garzik, "Linux kernel patch submission format".
860 <https://web.archive.org/web/20180829112450/http://linux.yyz.us/patch-format.html>
861
862 Greg Kroah-Hartman, "How to piss off a kernel subsystem maintainer".
863 <http://www.kroah.com/log/linux/maintainer.html>
864
865 <http://www.kroah.com/log/linux/maintainer-02.html>
866
867 <http://www.kroah.com/log/linux/maintainer-03.html>
868
869 <http://www.kroah.com/log/linux/maintainer-04.html>
870
871 <http://www.kroah.com/log/linux/maintainer-05.html>
872
873 <http://www.kroah.com/log/linux/maintainer-06.html>
874
875 Kernel Documentation/process/coding-style.rst
876
877 Linus Torvalds's mail on the canonical patch format:
878 <https://lore.kernel.org/r/Pine.LNX.4.58.0504071023190.28951@ppc970.osdl.org">https://lore.kernel.org/r/Pine.LNX.4.58.0504071023190.28951@ppc970.osdl.org>
879
880 Andi Kleen, "On submitting kernel patches"
881 Some strategies to get difficult or controversial changes in.
882
883 http://halobates.de/on-submitting-patches.pdf
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.