path: root/Documentation/SubmittingPatches (plain)
blob: 36f1dedc944ccf846afe5c58b8b3918c94a13977

1	.. _submittingpatches:
2
3	How to Get Your Change Into the Linux Kernel or Care And Operation Of Your Linus Torvalds
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 :ref:`Documentation/development-process <development_process_main>`.
14	Also, read :ref:`Documentation/SubmitChecklist <submitchecklist>`
15	for a list of items to check before
16	submitting code. If you are submitting a driver, also read
17	:ref:`Documentation/SubmittingDrivers <submittingdrivers>`;
18	for device tree binding patches, read
19	Documentation/devicetree/bindings/submitting-patches.txt.
20
21	Many of these steps describe the default behavior of the ``git`` version
22	control system; if you use ``git`` to prepare your patches, you'll find much
23	of the mechanical work done for you, though you'll still need to prepare
24	and document a sensible set of patches. In general, use of ``git`` will make
25	your life as a kernel developer easier.
26
27	Creating and Sending your Change
28	********************************
29
30
31	0) Obtain a current source tree
32	-------------------------------
33
34	If you do not have a repository with the current kernel source handy, use
35	``git`` to obtain one. You'll want to start with the mainline repository,
36	which can be grabbed with::
37
38	git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
39
40	Note, however, that you may not want to develop against the mainline tree
41	directly. Most subsystem maintainers run their own trees and want to see
42	patches prepared against those trees. See the **T:** entry for the subsystem
43	in the MAINTAINERS file to find that tree, or simply ask the maintainer if
44	the tree is not listed there.
45
46	It is still possible to download kernel releases via tarballs (as described
47	in the next section), but that is the hard way to do kernel development.
48
49	1) ``diff -up``
50	---------------
51
52	If you must generate your patches by hand, use ``diff -up`` or ``diff -uprN``
53	to create patches. Git generates patches in this form by default; if
54	you're using ``git``, you can skip this section entirely.
55
56	All changes to the Linux kernel occur in the form of patches, as
57	generated by :manpage:`diff(1)`. When creating your patch, make sure to
58	create it in "unified diff" format, as supplied by the ``-u`` argument
59	to :manpage:`diff(1)`.
60	Also, please use the ``-p`` argument which shows which C function each
61	change is in - that makes the resultant ``diff`` a lot easier to read.
62	Patches should be based in the root kernel source directory,
63	not in any lower subdirectory.
64
65	To create a patch for a single file, it is often sufficient to do::
66
67	SRCTREE= linux
68	MYFILE= drivers/net/mydriver.c
69
70	cd $SRCTREE
71	cp $MYFILE $MYFILE.orig
72	vi $MYFILE # make your change
73	cd ..
74	diff -up $SRCTREE/$MYFILE{.orig,} > /tmp/patch
75
76	To create a patch for multiple files, you should unpack a "vanilla",
77	or unmodified kernel source tree, and generate a ``diff`` against your
78	own source tree. For example::
79
80	MYSRC= /devel/linux
81
82	tar xvfz linux-3.19.tar.gz
83	mv linux-3.19 linux-3.19-vanilla
84	diff -uprN -X linux-3.19-vanilla/Documentation/dontdiff \
85	linux-3.19-vanilla $MYSRC > /tmp/patch
86
87	``dontdiff`` is a list of files which are generated by the kernel during
88	the build process, and should be ignored in any :manpage:`diff(1)`-generated
89	patch.
90
91	Make sure your patch does not include any extra files which do not
92	belong in a patch submission. Make sure to review your patch -after-
93	generating it with :manpage:`diff(1)`, to ensure accuracy.
94
95	If your changes produce a lot of deltas, you need to split them into
96	individual patches which modify things in logical stages; see
97	:ref:`split_changes`. This will facilitate review by other kernel developers,
98	very important if you want your patch accepted.
99
100	If you're using ``git``, ``git rebase -i`` can help you with this process. If
101	you're not using ``git``, ``quilt`` <http://savannah.nongnu.org/projects/quilt>
102	is another popular alternative.
103
104	.. _describe_changes:
105
106	2) Describe your changes
107	------------------------
108
109	Describe your problem. Whether your patch is a one-line bug fix or
110	5000 lines of a new feature, there must be an underlying problem that
111	motivated you to do this work. Convince the reviewer that there is a
112	problem worth fixing and that it makes sense for them to read past the
113	first paragraph.
114
115	Describe user-visible impact. Straight up crashes and lockups are
116	pretty convincing, but not all bugs are that blatant. Even if the
117	problem was spotted during code review, describe the impact you think
118	it can have on users. Keep in mind that the majority of Linux
119	installations run kernels from secondary stable trees or
120	vendor/product-specific trees that cherry-pick only specific patches
121	from upstream, so include anything that could help route your change
122	downstream: provoking circumstances, excerpts from dmesg, crash
123	descriptions, performance regressions, latency spikes, lockups, etc.
124
125	Quantify optimizations and trade-offs. If you claim improvements in
126	performance, memory consumption, stack footprint, or binary size,
127	include numbers that back them up. But also describe non-obvious
128	costs. Optimizations usually aren't free but trade-offs between CPU,
129	memory, and readability; or, when it comes to heuristics, between
130	different workloads. Describe the expected downsides of your
131	optimization so that the reviewer can weigh costs against benefits.
132
133	Once the problem is established, describe what you are actually doing
134	about it in technical detail. It's important to describe the change
135	in plain English for the reviewer to verify that the code is behaving
136	as you intend it to.
137
138	The maintainer will thank you if you write your patch description in a
139	form which can be easily pulled into Linux's source code management
140	system, ``git``, as a "commit log". See :ref:`explicit_in_reply_to`.
141
142	Solve only one problem per patch. If your description starts to get
143	long, that's a sign that you probably need to split up your patch.
144	See :ref:`split_changes`.
145
146	When you submit or resubmit a patch or patch series, include the
147	complete patch description and justification for it. Don't just
148	say that this is version N of the patch (series). Don't expect the
149	subsystem maintainer to refer back to earlier patch versions or referenced
150	URLs to find the patch description and put that into the patch.
151	I.e., the patch (series) and its description should be self-contained.
152	This benefits both the maintainers and reviewers. Some reviewers
153	probably didn't even receive earlier versions of the patch.
154
155	Describe your changes in imperative mood, e.g. "make xyzzy do frotz"
156	instead of "[This patch] makes xyzzy do frotz" or "[I] changed xyzzy
157	to do frotz", as if you are giving orders to the codebase to change
158	its behaviour.
159
160	If the patch fixes a logged bug entry, refer to that bug entry by
161	number and URL. If the patch follows from a mailing list discussion,
162	give a URL to the mailing list archive; use the https://lkml.kernel.org/
163	redirector with a ``Message-Id``, to ensure that the links cannot become
164	stale.
165
166	However, try to make your explanation understandable without external
167	resources. In addition to giving a URL to a mailing list archive or
168	bug, summarize the relevant points of the discussion that led to the
169	patch as submitted.
170
171	If you want to refer to a specific commit, don't just refer to the
172	SHA-1 ID of the commit. Please also include the oneline summary of
173	the commit, to make it easier for reviewers to know what it is about.
174	Example::
175
176	Commit e21d2170f36602ae2708 ("video: remove unnecessary
177	platform_set_drvdata()") removed the unnecessary
178	platform_set_drvdata(), but left the variable "dev" unused,
179	delete it.
180
181	You should also be sure to use at least the first twelve characters of the
182	SHA-1 ID. The kernel repository holds a *lot* of objects, making
183	collisions with shorter IDs a real possibility. Bear in mind that, even if
184	there is no collision with your six-character ID now, that condition may
185	change five years from now.
186
187	If your patch fixes a bug in a specific commit, e.g. you found an issue using
188	``git bisect``, please use the 'Fixes:' tag with the first 12 characters of
189	the SHA-1 ID, and the one line summary. For example::
190
191	Fixes: e21d2170f366 ("video: remove unnecessary platform_set_drvdata()")
192
193	The following ``git config`` settings can be used to add a pretty format for
194	outputting the above style in the ``git log`` or ``git show`` commands::
195
196	[core]
197	abbrev = 12
198	[pretty]
199	fixes = Fixes: %h (\"%s\")
200
201	.. _split_changes:
202
203	3) Separate your changes
204	------------------------
205
206	Separate each **logical change** into a separate patch.
207
208	For example, if your changes include both bug fixes and performance
209	enhancements for a single driver, separate those changes into two
210	or more patches. If your changes include an API update, and a new
211	driver which uses that new API, separate those into two patches.
212
213	On the other hand, if you make a single change to numerous files,
214	group those changes into a single patch. Thus a single logical change
215	is contained within a single patch.
216
217	The point to remember is that each patch should make an easily understood
218	change that can be verified by reviewers. Each patch should be justifiable
219	on its own merits.
220
221	If one patch depends on another patch in order for a change to be
222	complete, that is OK. Simply note **"this patch depends on patch X"**
223	in your patch description.
224
225	When dividing your change into a series of patches, take special care to
226	ensure that the kernel builds and runs properly after each patch in the
227	series. Developers using ``git bisect`` to track down a problem can end up
228	splitting your patch series at any point; they will not thank you if you
229	introduce bugs in the middle.
230
231	If you cannot condense your patch set into a smaller set of patches,
232	then only post say 15 or so at a time and wait for review and integration.
233
234
235
236	4) Style-check your changes
237	---------------------------
238
239	Check your patch for basic style violations, details of which can be
240	found in
241	:ref:`Documentation/CodingStyle <codingstyle>`.
242	Failure to do so simply wastes
243	the reviewers time and will get your patch rejected, probably
244	without even being read.
245
246	One significant exception is when moving code from one file to
247	another -- in this case you should not modify the moved code at all in
248	the same patch which moves it. This clearly delineates the act of
249	moving the code and your changes. This greatly aids review of the
250	actual differences and allows tools to better track the history of
251	the code itself.
252
253	Check your patches with the patch style checker prior to submission
254	(scripts/checkpatch.pl). Note, though, that the style checker should be
255	viewed as a guide, not as a replacement for human judgment. If your code
256	looks better with a violation then its probably best left alone.
257
258	The checker reports at three levels:
259	- ERROR: things that are very likely to be wrong
260	- WARNING: things requiring careful review
261	- CHECK: things requiring thought
262
263	You should be able to justify all violations that remain in your
264	patch.
265
266
267	5) Select the recipients for your patch
268	---------------------------------------
269
270	You should always copy the appropriate subsystem maintainer(s) on any patch
271	to code that they maintain; look through the MAINTAINERS file and the
272	source code revision history to see who those maintainers are. The
273	script scripts/get_maintainer.pl can be very useful at this step. If you
274	cannot find a maintainer for the subsystem you are working on, Andrew
275	Morton (akpm@linux-foundation.org) serves as a maintainer of last resort.
276
277	You should also normally choose at least one mailing list to receive a copy
278	of your patch set. linux-kernel@vger.kernel.org functions as a list of
279	last resort, but the volume on that list has caused a number of developers
280	to tune it out. Look in the MAINTAINERS file for a subsystem-specific
281	list; your patch will probably get more attention there. Please do not
282	spam unrelated lists, though.
283
284	Many kernel-related lists are hosted on vger.kernel.org; you can find a
285	list of them at http://vger.kernel.org/vger-lists.html. There are
286	kernel-related lists hosted elsewhere as well, though.
287
288	Do not send more than 15 patches at once to the vger mailing lists!!!
289
290	Linus Torvalds is the final arbiter of all changes accepted into the
291	Linux kernel. His e-mail address is <torvalds@linux-foundation.org>.
292	He gets a lot of e-mail, and, at this point, very few patches go through
293	Linus directly, so typically you should do your best to -avoid-
294	sending him e-mail.
295
296	If you have a patch that fixes an exploitable security bug, send that patch
297	to security@kernel.org. For severe bugs, a short embargo may be considered
298	to allow distributors to get the patch out to users; in such cases,
299	obviously, the patch should not be sent to any public lists.
300
301	Patches that fix a severe bug in a released kernel should be directed
302	toward the stable maintainers by putting a line like this::
303
304	Cc: stable@vger.kernel.org
305
306	into the sign-off area of your patch (note, NOT an email recipient). You
307	should also read
308	:ref:`Documentation/stable_kernel_rules.txt <stable_kernel_rules>`
309	in addition to this file.
310
311	Note, however, that some subsystem maintainers want to come to their own
312	conclusions on which patches should go to the stable trees. The networking
313	maintainer, in particular, would rather not see individual developers
314	adding lines like the above to their patches.
315
316	If changes affect userland-kernel interfaces, please send the MAN-PAGES
317	maintainer (as listed in the MAINTAINERS file) a man-pages patch, or at
318	least a notification of the change, so that some information makes its way
319	into the manual pages. User-space API changes should also be copied to
320	linux-api@vger.kernel.org.
321
322	For small patches you may want to CC the Trivial Patch Monkey
323	trivial@kernel.org which collects "trivial" patches. Have a look
324	into the MAINTAINERS file for its current manager.
325
326	Trivial patches must qualify for one of the following rules:
327
328	- Spelling fixes in documentation
329	- Spelling fixes for errors which could break :manpage:`grep(1)`
330	- Warning fixes (cluttering with useless warnings is bad)
331	- Compilation fixes (only if they are actually correct)
332	- Runtime fixes (only if they actually fix things)
333	- Removing use of deprecated functions/macros
334	- Contact detail and documentation fixes
335	- Non-portable code replaced by portable code (even in arch-specific,
336	since people copy, as long as it's trivial)
337	- Any fix by the author/maintainer of the file (ie. patch monkey
338	in re-transmission mode)
339
340
341
342	6) No MIME, no links, no compression, no attachments. Just plain text
343	----------------------------------------------------------------------
344
345	Linus and other kernel developers need to be able to read and comment
346	on the changes you are submitting. It is important for a kernel
347	developer to be able to "quote" your changes, using standard e-mail
348	tools, so that they may comment on specific portions of your code.
349
350	For this reason, all patches should be submitted by e-mail "inline".
351
352	.. warning::
353
354	Be wary of your editor's word-wrap corrupting your patch,
355	if you choose to cut-n-paste your patch.
356
357	Do not attach the patch as a MIME attachment, compressed or not.
358	Many popular e-mail applications will not always transmit a MIME
359	attachment as plain text, making it impossible to comment on your
360	code. A MIME attachment also takes Linus a bit more time to process,
361	decreasing the likelihood of your MIME-attached change being accepted.
362
363	Exception: If your mailer is mangling patches then someone may ask
364	you to re-send them using MIME.
365
366	See :ref:`Documentation/email-clients.txt <email_clients>`
367	for hints about configuring your e-mail client so that it sends your patches
368	untouched.
369
370	7) E-mail size
371	--------------
372
373	Large changes are not appropriate for mailing lists, and some
374	maintainers. If your patch, uncompressed, exceeds 300 kB in size,
375	it is preferred that you store your patch on an Internet-accessible
376	server, and provide instead a URL (link) pointing to your patch. But note
377	that if your patch exceeds 300 kB, it almost certainly needs to be broken up
378	anyway.
379
380	8) Respond to review comments
381	-----------------------------
382
383	Your patch will almost certainly get comments from reviewers on ways in
384	which the patch can be improved. You must respond to those comments;
385	ignoring reviewers is a good way to get ignored in return. Review comments
386	or questions that do not lead to a code change should almost certainly
387	bring about a comment or changelog entry so that the next reviewer better
388	understands what is going on.
389
390	Be sure to tell the reviewers what changes you are making and to thank them
391	for their time. Code review is a tiring and time-consuming process, and
392	reviewers sometimes get grumpy. Even in that case, though, respond
393	politely and address the problems they have pointed out.
394
395
396	9) Don't get discouraged - or impatient
397	---------------------------------------
398
399	After you have submitted your change, be patient and wait. Reviewers are
400	busy people and may not get to your patch right away.
401
402	Once upon a time, patches used to disappear into the void without comment,
403	but the development process works more smoothly than that now. You should
404	receive comments within a week or so; if that does not happen, make sure
405	that you have sent your patches to the right place. Wait for a minimum of
406	one week before resubmitting or pinging reviewers - possibly longer during
407	busy times like merge windows.
408
409
410	10) Include PATCH in the subject
411	--------------------------------
412
413	Due to high e-mail traffic to Linus, and to linux-kernel, it is common
414	convention to prefix your subject line with [PATCH]. This lets Linus
415	and other kernel developers more easily distinguish patches from other
416	e-mail discussions.
417
418
419
420	11) Sign your work
421	------------------
422
423	To improve tracking of who did what, especially with patches that can
424	percolate to their final resting place in the kernel through several
425	layers of maintainers, we've introduced a "sign-off" procedure on
426	patches that are being emailed around.
427
428	The sign-off is a simple line at the end of the explanation for the
429	patch, which certifies that you wrote it or otherwise have the right to
430	pass it on as an open-source patch. The rules are pretty simple: if you
431	can certify the below:
432
433	Developer's Certificate of Origin 1.1
434	^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
435
436	By making a contribution to this project, I certify that:
437
438	(a) The contribution was created in whole or in part by me and I
439	have the right to submit it under the open source license
440	indicated in the file; or
441
442	(b) The contribution is based upon previous work that, to the best
443	of my knowledge, is covered under an appropriate open source
444	license and I have the right under that license to submit that
445	work with modifications, whether created in whole or in part
446	by me, under the same open source license (unless I am
447	permitted to submit under a different license), as indicated
448	in the file; or
449
450	(c) The contribution was provided directly to me by some other
451	person who certified (a), (b) or (c) and I have not modified
452	it.
453
454	(d) I understand and agree that this project and the contribution
455	are public and that a record of the contribution (including all
456	personal information I submit with it, including my sign-off) is
457	maintained indefinitely and may be redistributed consistent with
458	this project or the open source license(s) involved.
459
460	then you just add a line saying::
461
462	Signed-off-by: Random J Developer <random@developer.example.org>
463
464	using your real name (sorry, no pseudonyms or anonymous contributions.)
465
466	Some people also put extra tags at the end. They'll just be ignored for
467	now, but you can do this to mark internal company procedures or just
468	point out some special detail about the sign-off.
469
470	If you are a subsystem or branch maintainer, sometimes you need to slightly
471	modify patches you receive in order to merge them, because the code is not
472	exactly the same in your tree and the submitters'. If you stick strictly to
473	rule (c), you should ask the submitter to rediff, but this is a totally
474	counter-productive waste of time and energy. Rule (b) allows you to adjust
475	the code, but then it is very impolite to change one submitter's code and
476	make him endorse your bugs. To solve this problem, it is recommended that
477	you add a line between the last Signed-off-by header and yours, indicating
478	the nature of your changes. While there is nothing mandatory about this, it
479	seems like prepending the description with your mail and/or name, all
480	enclosed in square brackets, is noticeable enough to make it obvious that
481	you are responsible for last-minute changes. Example::
482
483	Signed-off-by: Random J Developer <random@developer.example.org>
484	[lucky@maintainer.example.org: struct foo moved from foo.c to foo.h]
485	Signed-off-by: Lucky K Maintainer <lucky@maintainer.example.org>
486
487	This practice is particularly helpful if you maintain a stable branch and
488	want at the same time to credit the author, track changes, merge the fix,
489	and protect the submitter from complaints. Note that under no circumstances
490	can you change the author's identity (the From header), as it is the one
491	which appears in the changelog.
492
493	Special note to back-porters: It seems to be a common and useful practice
494	to insert an indication of the origin of a patch at the top of the commit
495	message (just after the subject line) to facilitate tracking. For instance,
496	here's what we see in a 3.x-stable release::
497
498	Date: Tue Oct 7 07:26:38 2014 -0400
499
500	libata: Un-break ATA blacklist
501
502	commit 1c40279960bcd7d52dbdf1d466b20d24b99176c8 upstream.
503
504	And here's what might appear in an older kernel once a patch is backported::
505
506	Date: Tue May 13 22:12:27 2008 +0200
507
508	wireless, airo: waitbusy() won't delay
509
510	[backport of 2.6 commit b7acbdfbd1f277c1eb23f344f899cfa4cd0bf36a]
511
512	Whatever the format, this information provides a valuable help to people
513	tracking your trees, and to people trying to troubleshoot bugs in your
514	tree.
515
516
517	12) When to use Acked-by: and Cc:
518	---------------------------------
519
520	The Signed-off-by: tag indicates that the signer was involved in the
521	development of the patch, or that he/she was in the patch's delivery path.
522
523	If a person was not directly involved in the preparation or handling of a
524	patch but wishes to signify and record their approval of it then they can
525	ask to have an Acked-by: line added to the patch's changelog.
526
527	Acked-by: is often used by the maintainer of the affected code when that
528	maintainer neither contributed to nor forwarded the patch.
529
530	Acked-by: is not as formal as Signed-off-by:. It is a record that the acker
531	has at least reviewed the patch and has indicated acceptance. Hence patch
532	mergers will sometimes manually convert an acker's "yep, looks good to me"
533	into an Acked-by: (but note that it is usually better to ask for an
534	explicit ack).
535
536	Acked-by: does not necessarily indicate acknowledgement of the entire patch.
537	For example, if a patch affects multiple subsystems and has an Acked-by: from
538	one subsystem maintainer then this usually indicates acknowledgement of just
539	the part which affects that maintainer's code. Judgement should be used here.
540	When in doubt people should refer to the original discussion in the mailing
541	list archives.
542
543	If a person has had the opportunity to comment on a patch, but has not
544	provided such comments, you may optionally add a ``Cc:`` tag to the patch.
545	This is the only tag which might be added without an explicit action by the
546	person it names - but it should indicate that this person was copied on the
547	patch. This tag documents that potentially interested parties
548	have been included in the discussion.
549
550
551	13) Using Reported-by:, Tested-by:, Reviewed-by:, Suggested-by: and Fixes:
552	--------------------------------------------------------------------------
553
554	The Reported-by tag gives credit to people who find bugs and report them and it
555	hopefully inspires them to help us again in the future. Please note that if
556	the bug was reported in private, then ask for permission first before using the
557	Reported-by tag.
558
559	A Tested-by: tag indicates that the patch has been successfully tested (in
560	some environment) by the person named. This tag informs maintainers that
561	some testing has been performed, provides a means to locate testers for
562	future patches, and ensures credit for the testers.
563
564	Reviewed-by:, instead, indicates that the patch has been reviewed and found
565	acceptable according to the Reviewer's Statement:
566
567	Reviewer's statement of oversight
568	^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
569
570	By offering my Reviewed-by: tag, I state that:
571
572	(a) I have carried out a technical review of this patch to
573	evaluate its appropriateness and readiness for inclusion into
574	the mainline kernel.
575
576	(b) Any problems, concerns, or questions relating to the patch
577	have been communicated back to the submitter. I am satisfied
578	with the submitter's response to my comments.
579
580	(c) While there may be things that could be improved with this
581	submission, I believe that it is, at this time, (1) a
582	worthwhile modification to the kernel, and (2) free of known
583	issues which would argue against its inclusion.
584
585	(d) While I have reviewed the patch and believe it to be sound, I
586	do not (unless explicitly stated elsewhere) make any
587	warranties or guarantees that it will achieve its stated
588	purpose or function properly in any given situation.
589
590	A Reviewed-by tag is a statement of opinion that the patch is an
591	appropriate modification of the kernel without any remaining serious
592	technical issues. Any interested reviewer (who has done the work) can
593	offer a Reviewed-by tag for a patch. This tag serves to give credit to
594	reviewers and to inform maintainers of the degree of review which has been
595	done on the patch. Reviewed-by: tags, when supplied by reviewers known to
596	understand the subject area and to perform thorough reviews, will normally
597	increase the likelihood of your patch getting into the kernel.
598
599	A Suggested-by: tag indicates that the patch idea is suggested by the person
600	named and ensures credit to the person for the idea. Please note that this
601	tag should not be added without the reporter's permission, especially if the
602	idea was not posted in a public forum. That said, if we diligently credit our
603	idea reporters, they will, hopefully, be inspired to help us again in the
604	future.
605
606	A Fixes: tag indicates that the patch fixes an issue in a previous commit. It
607	is used to make it easy to determine where a bug originated, which can help
608	review a bug fix. This tag also assists the stable kernel team in determining
609	which stable kernel versions should receive your fix. This is the preferred
610	method for indicating a bug fixed by the patch. See :ref:`describe_changes`
611	for more details.
612
613
614	14) The canonical patch format
615	------------------------------
616
617	This section describes how the patch itself should be formatted. Note
618	that, if you have your patches stored in a ``git`` repository, proper patch
619	formatting can be had with ``git format-patch``. The tools cannot create
620	the necessary text, though, so read the instructions below anyway.
621
622	The canonical patch subject line is::
623
624	Subject: [PATCH 001/123] subsystem: summary phrase
625
626	The canonical patch message body contains the following:
627
628	- A ``from`` line specifying the patch author (only needed if the person
629	sending the patch is not the author).
630
631	- An empty line.
632
633	- The body of the explanation, line wrapped at 75 columns, which will
634	be copied to the permanent changelog to describe this patch.
635
636	- The ``Signed-off-by:`` lines, described above, which will
637	also go in the changelog.
638
639	- A marker line containing simply ``---``.
640
641	- Any additional comments not suitable for the changelog.
642
643	- The actual patch (``diff`` output).
644
645	The Subject line format makes it very easy to sort the emails
646	alphabetically by subject line - pretty much any email reader will
647	support that - since because the sequence number is zero-padded,
648	the numerical and alphabetic sort is the same.
649
650	The ``subsystem`` in the email's Subject should identify which
651	area or subsystem of the kernel is being patched.
652
653	The ``summary phrase`` in the email's Subject should concisely
654	describe the patch which that email contains. The ``summary
655	phrase`` should not be a filename. Do not use the same ``summary
656	phrase`` for every patch in a whole patch series (where a ``patch
657	series`` is an ordered sequence of multiple, related patches).
658
659	Bear in mind that the ``summary phrase`` of your email becomes a
660	globally-unique identifier for that patch. It propagates all the way
661	into the ``git`` changelog. The ``summary phrase`` may later be used in
662	developer discussions which refer to the patch. People will want to
663	google for the ``summary phrase`` to read discussion regarding that
664	patch. It will also be the only thing that people may quickly see
665	when, two or three months later, they are going through perhaps
666	thousands of patches using tools such as ``gitk`` or ``git log
667	--oneline``.
668
669	For these reasons, the ``summary`` must be no more than 70-75
670	characters, and it must describe both what the patch changes, as well
671	as why the patch might be necessary. It is challenging to be both
672	succinct and descriptive, but that is what a well-written summary
673	should do.
674
675	The ``summary phrase`` may be prefixed by tags enclosed in square
676	brackets: "Subject: [PATCH <tag>...] <summary phrase>". The tags are
677	not considered part of the summary phrase, but describe how the patch
678	should be treated. Common tags might include a version descriptor if
679	the multiple versions of the patch have been sent out in response to
680	comments (i.e., "v1, v2, v3"), or "RFC" to indicate a request for
681	comments. If there are four patches in a patch series the individual
682	patches may be numbered like this: 1/4, 2/4, 3/4, 4/4. This assures
683	that developers understand the order in which the patches should be
684	applied and that they have reviewed or applied all of the patches in
685	the patch series.
686
687	A couple of example Subjects::
688
689	Subject: [PATCH 2/5] ext2: improve scalability of bitmap searching
690	Subject: [PATCH v2 01/27] x86: fix eflags tracking
691
692	The ``from`` line must be the very first line in the message body,
693	and has the form:
694
695	From: Original Author <author@example.com>
696
697	The ``from`` line specifies who will be credited as the author of the
698	patch in the permanent changelog. If the ``from`` line is missing,
699	then the ``From:`` line from the email header will be used to determine
700	the patch author in the changelog.
701
702	The explanation body will be committed to the permanent source
703	changelog, so should make sense to a competent reader who has long
704	since forgotten the immediate details of the discussion that might
705	have led to this patch. Including symptoms of the failure which the
706	patch addresses (kernel log messages, oops messages, etc.) is
707	especially useful for people who might be searching the commit logs
708	looking for the applicable patch. If a patch fixes a compile failure,
709	it may not be necessary to include _all_ of the compile failures; just
710	enough that it is likely that someone searching for the patch can find
711	it. As in the ``summary phrase``, it is important to be both succinct as
712	well as descriptive.
713
714	The ``---`` marker line serves the essential purpose of marking for patch
715	handling tools where the changelog message ends.
716
717	One good use for the additional comments after the ``---`` marker is for
718	a ``diffstat``, to show what files have changed, and the number of
719	inserted and deleted lines per file. A ``diffstat`` is especially useful
720	on bigger patches. Other comments relevant only to the moment or the
721	maintainer, not suitable for the permanent changelog, should also go
722	here. A good example of such comments might be ``patch changelogs``
723	which describe what has changed between the v1 and v2 version of the
724	patch.
725
726	If you are going to include a ``diffstat`` after the ``---`` marker, please
727	use ``diffstat`` options ``-p 1 -w 70`` so that filenames are listed from
728	the top of the kernel source tree and don't use too much horizontal
729	space (easily fit in 80 columns, maybe with some indentation). (``git``
730	generates appropriate diffstats by default.)
731
732	See more details on the proper patch format in the following
733	references.
734
735	.. _explicit_in_reply_to:
736
737	15) Explicit In-Reply-To headers
738	--------------------------------
739
740	It can be helpful to manually add In-Reply-To: headers to a patch
741	(e.g., when using ``git send-email``) to associate the patch with
742	previous relevant discussion, e.g. to link a bug fix to the email with
743	the bug report. However, for a multi-patch series, it is generally
744	best to avoid using In-Reply-To: to link to older versions of the
745	series. This way multiple versions of the patch don't become an
746	unmanageable forest of references in email clients. If a link is
747	helpful, you can use the https://lkml.kernel.org/ redirector (e.g., in
748	the cover email text) to link to an earlier version of the patch series.
749
750
751	16) Sending ``git pull`` requests
752	---------------------------------
753
754	If you have a series of patches, it may be most convenient to have the
755	maintainer pull them directly into the subsystem repository with a
756	``git pull`` operation. Note, however, that pulling patches from a developer
757	requires a higher degree of trust than taking patches from a mailing list.
758	As a result, many subsystem maintainers are reluctant to take pull
759	requests, especially from new, unknown developers. If in doubt you can use
760	the pull request as the cover letter for a normal posting of the patch
761	series, giving the maintainer the option of using either.
762
763	A pull request should have [GIT] or [PULL] in the subject line. The
764	request itself should include the repository name and the branch of
765	interest on a single line; it should look something like::
766
767	Please pull from
768
769	git://jdelvare.pck.nerim.net/jdelvare-2.6 i2c-for-linus
770
771	to get these changes:
772
773	A pull request should also include an overall message saying what will be
774	included in the request, a ``git shortlog`` listing of the patches
775	themselves, and a ``diffstat`` showing the overall effect of the patch series.
776	The easiest way to get all this information together is, of course, to let
777	``git`` do it for you with the ``git request-pull`` command.
778
779	Some maintainers (including Linus) want to see pull requests from signed
780	commits; that increases their confidence that the request actually came
781	from you. Linus, in particular, will not pull from public hosting sites
782	like GitHub in the absence of a signed tag.
783
784	The first step toward creating such tags is to make a GNUPG key and get it
785	signed by one or more core kernel developers. This step can be hard for
786	new developers, but there is no way around it. Attending conferences can
787	be a good way to find developers who can sign your key.
788
789	Once you have prepared a patch series in ``git`` that you wish to have somebody
790	pull, create a signed tag with ``git tag -s``. This will create a new tag
791	identifying the last commit in the series and containing a signature
792	created with your private key. You will also have the opportunity to add a
793	changelog-style message to the tag; this is an ideal place to describe the
794	effects of the pull request as a whole.
795
796	If the tree the maintainer will be pulling from is not the repository you
797	are working from, don't forget to push the signed tag explicitly to the
798	public tree.
799
800	When generating your pull request, use the signed tag as the target. A
801	command like this will do the trick::
802
803	git request-pull master git://my.public.tree/linux.git my-signed-tag
804
805
806	REFERENCES
807	**********
808
809	Andrew Morton, "The perfect patch" (tpp).
810	<http://www.ozlabs.org/~akpm/stuff/tpp.txt>
811
812	Jeff Garzik, "Linux kernel patch submission format".
813	<http://linux.yyz.us/patch-format.html>
814
815	Greg Kroah-Hartman, "How to piss off a kernel subsystem maintainer".
816	<http://www.kroah.com/log/linux/maintainer.html>
817
818	<http://www.kroah.com/log/linux/maintainer-02.html>
819
820	<http://www.kroah.com/log/linux/maintainer-03.html>
821
822	<http://www.kroah.com/log/linux/maintainer-04.html>
823
824	<http://www.kroah.com/log/linux/maintainer-05.html>
825
826	<http://www.kroah.com/log/linux/maintainer-06.html>
827
828	NO!!!! No more huge patch bombs to linux-kernel@vger.kernel.org people!
829	<https://lkml.org/lkml/2005/7/11/336>
830
831	Kernel Documentation/CodingStyle:
832	:ref:`Documentation/CodingStyle <codingstyle>`
833
834	Linus Torvalds's mail on the canonical patch format:
835	<http://lkml.org/lkml/2005/4/7/183>
836
837	Andi Kleen, "On submitting kernel patches"
838	Some strategies to get difficult or controversial changes in.
839
840	http://halobates.de/on-submitting-patches.pdf
841
842