blob: e439cd0d33752fd4231bcea896f39871aeff8dc3 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001
2 How to Get Your Change Into the Linux Kernel
3 or
4 Care And Operation Of Your Linus Torvalds
5
6
7
8For a person or company who wishes to submit a change to the Linux
9kernel, the process can sometimes be daunting if you're not familiar
10with "the system." This text is a collection of suggestions which
11can greatly increase the chances of your change being accepted.
12
Randy Dunlapbc7455f2006-07-30 03:03:45 -070013Read Documentation/SubmitChecklist for a list of items to check
14before submitting code. If you are submitting a driver, also read
15Documentation/SubmittingDrivers.
Linus Torvalds1da177e2005-04-16 15:20:36 -070016
17
18
19--------------------------------------------
20SECTION 1 - CREATING AND SENDING YOUR CHANGE
21--------------------------------------------
22
23
24
251) "diff -up"
26------------
27
28Use "diff -up" or "diff -uprN" to create patches.
29
30All changes to the Linux kernel occur in the form of patches, as
31generated by diff(1). When creating your patch, make sure to create it
32in "unified diff" format, as supplied by the '-u' argument to diff(1).
33Also, please use the '-p' argument which shows which C function each
34change is in - that makes the resultant diff a lot easier to read.
35Patches should be based in the root kernel source directory,
36not in any lower subdirectory.
37
38To create a patch for a single file, it is often sufficient to do:
39
Randy Dunlap84da7c02005-06-28 20:45:30 -070040 SRCTREE= linux-2.6
Linus Torvalds1da177e2005-04-16 15:20:36 -070041 MYFILE= drivers/net/mydriver.c
42
43 cd $SRCTREE
44 cp $MYFILE $MYFILE.orig
45 vi $MYFILE # make your change
46 cd ..
47 diff -up $SRCTREE/$MYFILE{.orig,} > /tmp/patch
48
49To create a patch for multiple files, you should unpack a "vanilla",
50or unmodified kernel source tree, and generate a diff against your
51own source tree. For example:
52
Randy Dunlap84da7c02005-06-28 20:45:30 -070053 MYSRC= /devel/linux-2.6
Linus Torvalds1da177e2005-04-16 15:20:36 -070054
Randy Dunlap84da7c02005-06-28 20:45:30 -070055 tar xvfz linux-2.6.12.tar.gz
56 mv linux-2.6.12 linux-2.6.12-vanilla
57 diff -uprN -X linux-2.6.12-vanilla/Documentation/dontdiff \
58 linux-2.6.12-vanilla $MYSRC > /tmp/patch
Linus Torvalds1da177e2005-04-16 15:20:36 -070059
60"dontdiff" is a list of files which are generated by the kernel during
61the build process, and should be ignored in any diff(1)-generated
Randy Dunlap84da7c02005-06-28 20:45:30 -070062patch. The "dontdiff" file is included in the kernel tree in
632.6.12 and later. For earlier kernel versions, you can get it
64from <http://www.xenotime.net/linux/doc/dontdiff>.
Linus Torvalds1da177e2005-04-16 15:20:36 -070065
66Make sure your patch does not include any extra files which do not
67belong in a patch submission. Make sure to review your patch -after-
68generated it with diff(1), to ensure accuracy.
69
70If your changes produce a lot of deltas, you may want to look into
71splitting them into individual patches which modify things in
Randy Dunlap84da7c02005-06-28 20:45:30 -070072logical stages. This will facilitate easier reviewing by other
Linus Torvalds1da177e2005-04-16 15:20:36 -070073kernel developers, very important if you want your patch accepted.
Randy Dunlap84da7c02005-06-28 20:45:30 -070074There are a number of scripts which can aid in this:
Linus Torvalds1da177e2005-04-16 15:20:36 -070075
76Quilt:
77http://savannah.nongnu.org/projects/quilt
78
Linus Torvalds1da177e2005-04-16 15:20:36 -070079Andrew Morton's patch scripts:
FD Cami2223c652008-10-15 22:02:00 -070080http://userweb.kernel.org/~akpm/stuff/patch-scripts.tar.gz
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -080081Instead of these scripts, quilt is the recommended patch management
82tool (see above).
Randy Dunlap84da7c02005-06-28 20:45:30 -070083
84
Linus Torvalds1da177e2005-04-16 15:20:36 -070085
862) Describe your changes.
87
88Describe the technical detail of the change(s) your patch includes.
89
90Be as specific as possible. The WORST descriptions possible include
91things like "update driver X", "bug fix for driver X", or "this patch
92includes updates for subsystem X. Please apply."
93
Theodore Ts'o2ae19ac2009-04-16 07:44:45 -040094The maintainer will thank you if you write your patch description in a
95form which can be easily pulled into Linux's source code management
96system, git, as a "commit log". See #15, below.
97
Linus Torvalds1da177e2005-04-16 15:20:36 -070098If your description starts to get long, that's a sign that you probably
99need to split up your patch. See #3, next.
100
Randy Dunlapd89b1942010-08-09 17:20:21 -0700101When you submit or resubmit a patch or patch series, include the
102complete patch description and justification for it. Don't just
103say that this is version N of the patch (series). Don't expect the
104patch merger to refer back to earlier patch versions or referenced
105URLs to find the patch description and put that into the patch.
106I.e., the patch (series) and its description should be self-contained.
107This benefits both the patch merger(s) and reviewers. Some reviewers
108probably didn't even receive earlier versions of the patch.
109
110If the patch fixes a logged bug entry, refer to that bug entry by
111number and URL.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700112
113
1143) Separate your changes.
115
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800116Separate _logical changes_ into a single patch file.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700117
118For example, if your changes include both bug fixes and performance
119enhancements for a single driver, separate those changes into two
120or more patches. If your changes include an API update, and a new
121driver which uses that new API, separate those into two patches.
122
123On the other hand, if you make a single change to numerous files,
124group those changes into a single patch. Thus a single logical change
125is contained within a single patch.
126
127If one patch depends on another patch in order for a change to be
128complete, that is OK. Simply note "this patch depends on patch X"
129in your patch description.
130
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800131If you cannot condense your patch set into a smaller set of patches,
132then only post say 15 or so at a time and wait for review and integration.
133
134
Linus Torvalds1da177e2005-04-16 15:20:36 -0700135
Andy Whitcroft0a920b52007-06-01 00:46:48 -07001364) Style check your changes.
137
138Check your patch for basic style violations, details of which can be
139found in Documentation/CodingStyle. Failure to do so simply wastes
Linus Nilssonf56d35e2007-07-21 17:49:06 +0200140the reviewers time and will get your patch rejected, probably
Andy Whitcroft0a920b52007-06-01 00:46:48 -0700141without even being read.
142
143At a minimum you should check your patches with the patch style
Andre Haupta570ab62007-09-11 15:23:47 -0700144checker prior to submission (scripts/checkpatch.pl). You should
Andy Whitcroft0a920b52007-06-01 00:46:48 -0700145be able to justify all violations that remain in your patch.
146
147
148
1495) Select e-mail destination.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700150
151Look through the MAINTAINERS file and the source code, and determine
152if your change applies to a specific subsystem of the kernel, with
153an assigned maintainer. If so, e-mail that person.
154
155If no maintainer is listed, or the maintainer does not respond, send
156your patch to the primary Linux kernel developer's mailing list,
157linux-kernel@vger.kernel.org. Most kernel developers monitor this
158e-mail list, and can comment on your changes.
159
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800160
161Do not send more than 15 patches at once to the vger mailing lists!!!
162
163
Linus Torvalds1da177e2005-04-16 15:20:36 -0700164Linus Torvalds is the final arbiter of all changes accepted into the
Linus Torvalds99ddcc72007-01-23 14:22:35 -0800165Linux kernel. His e-mail address is <torvalds@linux-foundation.org>.
166He gets a lot of e-mail, so typically you should do your best to -avoid-
167sending him e-mail.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700168
169Patches which are bug fixes, are "obvious" changes, or similarly
170require little discussion should be sent or CC'd to Linus. Patches
171which require discussion or do not have a clear advantage should
172usually be sent first to linux-kernel. Only after the patch is
173discussed should the patch then be submitted to Linus.
174
Linus Torvalds1da177e2005-04-16 15:20:36 -0700175
176
Andy Whitcroft0a920b52007-06-01 00:46:48 -07001776) Select your CC (e-mail carbon copy) list.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700178
179Unless you have a reason NOT to do so, CC linux-kernel@vger.kernel.org.
180
181Other kernel developers besides Linus need to be aware of your change,
182so that they may comment on it and offer code review and suggestions.
183linux-kernel is the primary Linux kernel developer mailing list.
184Other mailing lists are available for specific subsystems, such as
185USB, framebuffer devices, the VFS, the SCSI subsystem, etc. See the
186MAINTAINERS file for a mailing list that relates specifically to
187your change.
188
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800189Majordomo lists of VGER.KERNEL.ORG at:
190 <http://vger.kernel.org/vger-lists.html>
191
Paul Jackson1caf1f02005-07-31 22:34:48 -0700192If changes affect userland-kernel interfaces, please send
193the MAN-PAGES maintainer (as listed in the MAINTAINERS file)
194a man-pages patch, or at least a notification of the change,
195so that some information makes its way into the manual pages.
196
Michael Brunner8103b5c2009-08-04 00:41:11 +0200197Even if the maintainer did not respond in step #5, make sure to ALWAYS
Linus Torvalds1da177e2005-04-16 15:20:36 -0700198copy the maintainer when you change their code.
199
200For small patches you may want to CC the Trivial Patch Monkey
Markus Heidelberg82d27b22009-06-12 01:02:34 +0200201trivial@kernel.org which collects "trivial" patches. Have a look
202into the MAINTAINERS file for its current manager.
203Trivial patches must qualify for one of the following rules:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700204 Spelling fixes in documentation
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700205 Spelling fixes which could break grep(1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700206 Warning fixes (cluttering with useless warnings is bad)
207 Compilation fixes (only if they are actually correct)
208 Runtime fixes (only if they actually fix things)
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700209 Removing use of deprecated functions/macros (eg. check_region)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700210 Contact detail and documentation fixes
211 Non-portable code replaced by portable code (even in arch-specific,
212 since people copy, as long as it's trivial)
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700213 Any fix by the author/maintainer of the file (ie. patch monkey
Linus Torvalds1da177e2005-04-16 15:20:36 -0700214 in re-transmission mode)
Randy Dunlap84da7c02005-06-28 20:45:30 -0700215
Linus Torvalds1da177e2005-04-16 15:20:36 -0700216
217
Andy Whitcroft0a920b52007-06-01 00:46:48 -07002187) No MIME, no links, no compression, no attachments. Just plain text.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700219
220Linus and other kernel developers need to be able to read and comment
221on the changes you are submitting. It is important for a kernel
222developer to be able to "quote" your changes, using standard e-mail
223tools, so that they may comment on specific portions of your code.
224
225For this reason, all patches should be submitting e-mail "inline".
226WARNING: Be wary of your editor's word-wrap corrupting your patch,
227if you choose to cut-n-paste your patch.
228
229Do not attach the patch as a MIME attachment, compressed or not.
230Many popular e-mail applications will not always transmit a MIME
231attachment as plain text, making it impossible to comment on your
232code. A MIME attachment also takes Linus a bit more time to process,
233decreasing the likelihood of your MIME-attached change being accepted.
234
235Exception: If your mailer is mangling patches then someone may ask
236you to re-send them using MIME.
237
Michael Opdenacker097091c2008-02-03 18:06:58 +0200238See Documentation/email-clients.txt for hints about configuring
239your e-mail client so that it sends your patches untouched.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700240
Andy Whitcroft0a920b52007-06-01 00:46:48 -07002418) E-mail size.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700242
Andy Whitcroft0a920b52007-06-01 00:46:48 -0700243When sending patches to Linus, always follow step #7.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700244
245Large changes are not appropriate for mailing lists, and some
Randy Dunlap4932be72009-10-01 15:44:06 -0700246maintainers. If your patch, uncompressed, exceeds 300 kB in size,
Linus Torvalds1da177e2005-04-16 15:20:36 -0700247it is preferred that you store your patch on an Internet-accessible
248server, and provide instead a URL (link) pointing to your patch.
249
250
251
Andy Whitcroft0a920b52007-06-01 00:46:48 -07002529) Name your kernel version.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700253
254It is important to note, either in the subject line or in the patch
255description, the kernel version to which this patch applies.
256
257If the patch does not apply cleanly to the latest kernel version,
258Linus will not apply it.
259
260
261
Andy Whitcroft0a920b52007-06-01 00:46:48 -070026210) Don't get discouraged. Re-submit.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700263
264After you have submitted your change, be patient and wait. If Linus
265likes your change and applies it, it will appear in the next version
266of the kernel that he releases.
267
268However, if your change doesn't appear in the next version of the
269kernel, there could be any number of reasons. It's YOUR job to
270narrow down those reasons, correct what was wrong, and submit your
271updated change.
272
273It is quite common for Linus to "drop" your patch without comment.
274That's the nature of the system. If he drops your patch, it could be
275due to
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700276* Your patch did not apply cleanly to the latest kernel version.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700277* Your patch was not sufficiently discussed on linux-kernel.
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700278* A style issue (see section 2).
279* An e-mail formatting issue (re-read this section).
280* A technical problem with your change.
281* He gets tons of e-mail, and yours got lost in the shuffle.
282* You are being annoying.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700283
284When in doubt, solicit comments on linux-kernel mailing list.
285
286
287
Andy Whitcroft0a920b52007-06-01 00:46:48 -070028811) Include PATCH in the subject
Linus Torvalds1da177e2005-04-16 15:20:36 -0700289
290Due to high e-mail traffic to Linus, and to linux-kernel, it is common
291convention to prefix your subject line with [PATCH]. This lets Linus
292and other kernel developers more easily distinguish patches from other
293e-mail discussions.
294
295
296
Andy Whitcroft0a920b52007-06-01 00:46:48 -070029712) Sign your work
Linus Torvalds1da177e2005-04-16 15:20:36 -0700298
299To improve tracking of who did what, especially with patches that can
300percolate to their final resting place in the kernel through several
301layers of maintainers, we've introduced a "sign-off" procedure on
302patches that are being emailed around.
303
304The sign-off is a simple line at the end of the explanation for the
305patch, which certifies that you wrote it or otherwise have the right to
306pass it on as a open-source patch. The rules are pretty simple: if you
307can certify the below:
308
Linus Torvaldscbd83da2005-06-13 17:51:55 -0700309 Developer's Certificate of Origin 1.1
Linus Torvalds1da177e2005-04-16 15:20:36 -0700310
311 By making a contribution to this project, I certify that:
312
313 (a) The contribution was created in whole or in part by me and I
314 have the right to submit it under the open source license
315 indicated in the file; or
316
317 (b) The contribution is based upon previous work that, to the best
318 of my knowledge, is covered under an appropriate open source
319 license and I have the right under that license to submit that
320 work with modifications, whether created in whole or in part
321 by me, under the same open source license (unless I am
322 permitted to submit under a different license), as indicated
323 in the file; or
324
325 (c) The contribution was provided directly to me by some other
326 person who certified (a), (b) or (c) and I have not modified
327 it.
328
Linus Torvaldscbd83da2005-06-13 17:51:55 -0700329 (d) I understand and agree that this project and the contribution
330 are public and that a record of the contribution (including all
331 personal information I submit with it, including my sign-off) is
332 maintained indefinitely and may be redistributed consistent with
333 this project or the open source license(s) involved.
334
Linus Torvalds1da177e2005-04-16 15:20:36 -0700335then you just add a line saying
336
Alexey Dobriyan9fd55592005-06-25 14:59:34 -0700337 Signed-off-by: Random J Developer <random@developer.example.org>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700338
Greg KHaf45f322006-09-12 20:35:52 -0700339using your real name (sorry, no pseudonyms or anonymous contributions.)
340
Linus Torvalds1da177e2005-04-16 15:20:36 -0700341Some people also put extra tags at the end. They'll just be ignored for
342now, but you can do this to mark internal company procedures or just
343point out some special detail about the sign-off.
344
Willy Tarreauadbd5882008-06-03 00:20:28 +0200345If you are a subsystem or branch maintainer, sometimes you need to slightly
346modify patches you receive in order to merge them, because the code is not
347exactly the same in your tree and the submitters'. If you stick strictly to
348rule (c), you should ask the submitter to rediff, but this is a totally
349counter-productive waste of time and energy. Rule (b) allows you to adjust
350the code, but then it is very impolite to change one submitter's code and
351make him endorse your bugs. To solve this problem, it is recommended that
352you add a line between the last Signed-off-by header and yours, indicating
353the nature of your changes. While there is nothing mandatory about this, it
354seems like prepending the description with your mail and/or name, all
355enclosed in square brackets, is noticeable enough to make it obvious that
356you are responsible for last-minute changes. Example :
357
358 Signed-off-by: Random J Developer <random@developer.example.org>
359 [lucky@maintainer.example.org: struct foo moved from foo.c to foo.h]
360 Signed-off-by: Lucky K Maintainer <lucky@maintainer.example.org>
361
362This practise is particularly helpful if you maintain a stable branch and
363want at the same time to credit the author, track changes, merge the fix,
364and protect the submitter from complaints. Note that under no circumstances
365can you change the author's identity (the From header), as it is the one
366which appears in the changelog.
367
368Special note to back-porters: It seems to be a common and useful practise
369to insert an indication of the origin of a patch at the top of the commit
370message (just after the subject line) to facilitate tracking. For instance,
371here's what we see in 2.6-stable :
372
373 Date: Tue May 13 19:10:30 2008 +0000
374
375 SCSI: libiscsi regression in 2.6.25: fix nop timer handling
376
377 commit 4cf1043593db6a337f10e006c23c69e5fc93e722 upstream
378
379And here's what appears in 2.4 :
380
381 Date: Tue May 13 22:12:27 2008 +0200
382
383 wireless, airo: waitbusy() won't delay
384
385 [backport of 2.6 commit b7acbdfbd1f277c1eb23f344f899cfa4cd0bf36a]
386
387Whatever the format, this information provides a valuable help to people
388tracking your trees, and to people trying to trouble-shoot bugs in your
389tree.
390
Linus Torvalds1da177e2005-04-16 15:20:36 -0700391
Jonathan Corbetef402032008-03-28 11:22:38 -060039213) When to use Acked-by: and Cc:
Andy Whitcroft0a920b52007-06-01 00:46:48 -0700393
Andrew Morton0f44cd22007-06-08 13:46:45 -0700394The Signed-off-by: tag indicates that the signer was involved in the
395development of the patch, or that he/she was in the patch's delivery path.
396
397If a person was not directly involved in the preparation or handling of a
398patch but wishes to signify and record their approval of it then they can
399arrange to have an Acked-by: line added to the patch's changelog.
400
401Acked-by: is often used by the maintainer of the affected code when that
402maintainer neither contributed to nor forwarded the patch.
403
404Acked-by: is not as formal as Signed-off-by:. It is a record that the acker
405has at least reviewed the patch and has indicated acceptance. Hence patch
406mergers will sometimes manually convert an acker's "yep, looks good to me"
407into an Acked-by:.
408
409Acked-by: does not necessarily indicate acknowledgement of the entire patch.
410For example, if a patch affects multiple subsystems and has an Acked-by: from
411one subsystem maintainer then this usually indicates acknowledgement of just
412the part which affects that maintainer's code. Judgement should be used here.
Jonathan Corbetef402032008-03-28 11:22:38 -0600413When in doubt people should refer to the original discussion in the mailing
Andrew Morton0f44cd22007-06-08 13:46:45 -0700414list archives.
415
Jonathan Corbetef402032008-03-28 11:22:38 -0600416If a person has had the opportunity to comment on a patch, but has not
417provided such comments, you may optionally add a "Cc:" tag to the patch.
418This is the only tag which might be added without an explicit action by the
419person it names. This tag documents that potentially interested parties
420have been included in the discussion
Andrew Morton0f44cd22007-06-08 13:46:45 -0700421
Jonathan Corbetef402032008-03-28 11:22:38 -0600422
Jonathan Corbetbbb0a4242009-01-16 09:49:50 -070042314) Using Reported-by:, Tested-by: and Reviewed-by:
424
425If this patch fixes a problem reported by somebody else, consider adding a
426Reported-by: tag to credit the reporter for their contribution. Please
427note that this tag should not be added without the reporter's permission,
428especially if the problem was not reported in a public forum. That said,
429if we diligently credit our bug reporters, they will, hopefully, be
430inspired to help us again in the future.
Jonathan Corbetef402032008-03-28 11:22:38 -0600431
432A Tested-by: tag indicates that the patch has been successfully tested (in
433some environment) by the person named. This tag informs maintainers that
434some testing has been performed, provides a means to locate testers for
435future patches, and ensures credit for the testers.
436
437Reviewed-by:, instead, indicates that the patch has been reviewed and found
438acceptable according to the Reviewer's Statement:
439
440 Reviewer's statement of oversight
441
442 By offering my Reviewed-by: tag, I state that:
443
444 (a) I have carried out a technical review of this patch to
445 evaluate its appropriateness and readiness for inclusion into
446 the mainline kernel.
447
448 (b) Any problems, concerns, or questions relating to the patch
449 have been communicated back to the submitter. I am satisfied
450 with the submitter's response to my comments.
451
452 (c) While there may be things that could be improved with this
453 submission, I believe that it is, at this time, (1) a
454 worthwhile modification to the kernel, and (2) free of known
455 issues which would argue against its inclusion.
456
457 (d) While I have reviewed the patch and believe it to be sound, I
458 do not (unless explicitly stated elsewhere) make any
459 warranties or guarantees that it will achieve its stated
460 purpose or function properly in any given situation.
461
462A Reviewed-by tag is a statement of opinion that the patch is an
463appropriate modification of the kernel without any remaining serious
464technical issues. Any interested reviewer (who has done the work) can
465offer a Reviewed-by tag for a patch. This tag serves to give credit to
466reviewers and to inform maintainers of the degree of review which has been
467done on the patch. Reviewed-by: tags, when supplied by reviewers known to
468understand the subject area and to perform thorough reviews, will normally
Pavel Machek5801da12009-06-04 16:26:50 +0200469increase the likelihood of your patch getting into the kernel.
Jonathan Corbetef402032008-03-28 11:22:38 -0600470
471
47215) The canonical patch format
Randy Dunlap84da7c02005-06-28 20:45:30 -0700473
Paul Jackson75f84262005-10-02 18:01:42 -0700474The canonical patch subject line is:
475
Paul Jacksond6b9acc2005-10-03 00:29:10 -0700476 Subject: [PATCH 001/123] subsystem: summary phrase
Paul Jackson75f84262005-10-02 18:01:42 -0700477
478The canonical patch message body contains the following:
479
480 - A "from" line specifying the patch author.
481
482 - An empty line.
483
484 - The body of the explanation, which will be copied to the
485 permanent changelog to describe this patch.
486
487 - The "Signed-off-by:" lines, described above, which will
488 also go in the changelog.
489
490 - A marker line containing simply "---".
491
492 - Any additional comments not suitable for the changelog.
493
494 - The actual patch (diff output).
495
496The Subject line format makes it very easy to sort the emails
497alphabetically by subject line - pretty much any email reader will
498support that - since because the sequence number is zero-padded,
499the numerical and alphabetic sort is the same.
500
Paul Jacksond6b9acc2005-10-03 00:29:10 -0700501The "subsystem" in the email's Subject should identify which
502area or subsystem of the kernel is being patched.
503
504The "summary phrase" in the email's Subject should concisely
505describe the patch which that email contains. The "summary
506phrase" should not be a filename. Do not use the same "summary
Randy Dunlap66effdc2007-05-09 02:33:42 -0700507phrase" for every patch in a whole patch series (where a "patch
508series" is an ordered sequence of multiple, related patches).
Paul Jacksond6b9acc2005-10-03 00:29:10 -0700509
Theodore Ts'o2ae19ac2009-04-16 07:44:45 -0400510Bear in mind that the "summary phrase" of your email becomes a
511globally-unique identifier for that patch. It propagates all the way
512into the git changelog. The "summary phrase" may later be used in
513developer discussions which refer to the patch. People will want to
514google for the "summary phrase" to read discussion regarding that
515patch. It will also be the only thing that people may quickly see
516when, two or three months later, they are going through perhaps
517thousands of patches using tools such as "gitk" or "git log
518--oneline".
519
520For these reasons, the "summary" must be no more than 70-75
521characters, and it must describe both what the patch changes, as well
522as why the patch might be necessary. It is challenging to be both
523succinct and descriptive, but that is what a well-written summary
524should do.
525
526The "summary phrase" may be prefixed by tags enclosed in square
527brackets: "Subject: [PATCH tag] <summary phrase>". The tags are not
528considered part of the summary phrase, but describe how the patch
529should be treated. Common tags might include a version descriptor if
530the multiple versions of the patch have been sent out in response to
531comments (i.e., "v1, v2, v3"), or "RFC" to indicate a request for
532comments. If there are four patches in a patch series the individual
533patches may be numbered like this: 1/4, 2/4, 3/4, 4/4. This assures
534that developers understand the order in which the patches should be
535applied and that they have reviewed or applied all of the patches in
536the patch series.
Paul Jacksond6b9acc2005-10-03 00:29:10 -0700537
538A couple of example Subjects:
539
540 Subject: [patch 2/5] ext2: improve scalability of bitmap searching
541 Subject: [PATCHv2 001/207] x86: fix eflags tracking
Paul Jackson75f84262005-10-02 18:01:42 -0700542
543The "from" line must be the very first line in the message body,
544and has the form:
545
546 From: Original Author <author@example.com>
547
548The "from" line specifies who will be credited as the author of the
549patch in the permanent changelog. If the "from" line is missing,
550then the "From:" line from the email header will be used to determine
551the patch author in the changelog.
552
553The explanation body will be committed to the permanent source
554changelog, so should make sense to a competent reader who has long
555since forgotten the immediate details of the discussion that might
Theodore Ts'o2ae19ac2009-04-16 07:44:45 -0400556have led to this patch. Including symptoms of the failure which the
557patch addresses (kernel log messages, oops messages, etc.) is
558especially useful for people who might be searching the commit logs
559looking for the applicable patch. If a patch fixes a compile failure,
560it may not be necessary to include _all_ of the compile failures; just
561enough that it is likely that someone searching for the patch can find
562it. As in the "summary phrase", it is important to be both succinct as
563well as descriptive.
Paul Jackson75f84262005-10-02 18:01:42 -0700564
565The "---" marker line serves the essential purpose of marking for patch
566handling tools where the changelog message ends.
567
568One good use for the additional comments after the "---" marker is for
Theodore Ts'o2ae19ac2009-04-16 07:44:45 -0400569a diffstat, to show what files have changed, and the number of
570inserted and deleted lines per file. A diffstat is especially useful
571on bigger patches. Other comments relevant only to the moment or the
572maintainer, not suitable for the permanent changelog, should also go
573here. A good example of such comments might be "patch changelogs"
574which describe what has changed between the v1 and v2 version of the
575patch.
576
577If you are going to include a diffstat after the "---" marker, please
578use diffstat options "-p 1 -w 70" so that filenames are listed from
579the top of the kernel source tree and don't use too much horizontal
580space (easily fit in 80 columns, maybe with some indentation).
Paul Jackson75f84262005-10-02 18:01:42 -0700581
582See more details on the proper patch format in the following
583references.
584
585
Randy Dunlap14863612008-07-27 20:44:24 -070058616) Sending "git pull" requests (from Linus emails)
Randy Dunlap84da7c02005-06-28 20:45:30 -0700587
Randy Dunlap14863612008-07-27 20:44:24 -0700588Please write the git repo address and branch name alone on the same line
589so that I can't even by mistake pull from the wrong branch, and so
590that a triple-click just selects the whole thing.
591
592So the proper format is something along the lines of:
593
594 "Please pull from
595
596 git://jdelvare.pck.nerim.net/jdelvare-2.6 i2c-for-linus
597
598 to get these changes:"
599
600so that I don't have to hunt-and-peck for the address and inevitably
601get it wrong (actually, I've only gotten it wrong a few times, and
602checking against the diffstat tells me when I get it wrong, but I'm
603just a lot more comfortable when I don't have to "look for" the right
604thing to pull, and double-check that I have the right branch-name).
605
606
607Please use "git diff -M --stat --summary" to generate the diffstat:
608the -M enables rename detection, and the summary enables a summary of
609new/deleted or renamed files.
610
611With rename detection, the statistics are rather different [...]
612because git will notice that a fair number of the changes are renames.
Randy Dunlap84da7c02005-06-28 20:45:30 -0700613
Linus Torvalds1da177e2005-04-16 15:20:36 -0700614-----------------------------------
615SECTION 2 - HINTS, TIPS, AND TRICKS
616-----------------------------------
617
618This section lists many of the common "rules" associated with code
619submitted to the kernel. There are always exceptions... but you must
620have a really good reason for doing so. You could probably call this
621section Linus Computer Science 101.
622
623
624
6251) Read Documentation/CodingStyle
626
627Nuff said. If your code deviates too much from this, it is likely
628to be rejected without further review, and without comment.
629
Keiichi Kii5ab3bd52007-10-26 15:51:44 +0900630One significant exception is when moving code from one file to
631another -- in this case you should not modify the moved code at all in
Andy Whitcroftde7d4f02007-07-15 23:37:22 -0700632the same patch which moves it. This clearly delineates the act of
633moving the code and your changes. This greatly aids review of the
634actual differences and allows tools to better track the history of
635the code itself.
636
Andy Whitcroft0a920b52007-06-01 00:46:48 -0700637Check your patches with the patch style checker prior to submission
Andy Whitcroftde7d4f02007-07-15 23:37:22 -0700638(scripts/checkpatch.pl). The style checker should be viewed as
639a guide not as the final word. If your code looks better with
640a violation then its probably best left alone.
641
642The checker reports at three levels:
643 - ERROR: things that are very likely to be wrong
644 - WARNING: things requiring careful review
645 - CHECK: things requiring thought
646
647You should be able to justify all violations that remain in your
648patch.
Andy Whitcroft0a920b52007-06-01 00:46:48 -0700649
Linus Torvalds1da177e2005-04-16 15:20:36 -0700650
651
6522) #ifdefs are ugly
653
654Code cluttered with ifdefs is difficult to read and maintain. Don't do
655it. Instead, put your ifdefs in a header, and conditionally define
656'static inline' functions, or macros, which are used in the code.
657Let the compiler optimize away the "no-op" case.
658
659Simple example, of poor code:
660
661 dev = alloc_etherdev (sizeof(struct funky_private));
662 if (!dev)
663 return -ENODEV;
664 #ifdef CONFIG_NET_FUNKINESS
665 init_funky_net(dev);
666 #endif
667
668Cleaned-up example:
669
670(in header)
671 #ifndef CONFIG_NET_FUNKINESS
672 static inline void init_funky_net (struct net_device *d) {}
673 #endif
674
675(in the code itself)
676 dev = alloc_etherdev (sizeof(struct funky_private));
677 if (!dev)
678 return -ENODEV;
679 init_funky_net(dev);
680
681
682
6833) 'static inline' is better than a macro
684
685Static inline functions are greatly preferred over macros.
686They provide type safety, have no length limitations, no formatting
687limitations, and under gcc they are as cheap as macros.
688
689Macros should only be used for cases where a static inline is clearly
Jim Meyeringf2b2ea62008-04-02 13:04:46 -0700690suboptimal [there are a few, isolated cases of this in fast paths],
Linus Torvalds1da177e2005-04-16 15:20:36 -0700691or where it is impossible to use a static inline function [such as
692string-izing].
693
694'static inline' is preferred over 'static __inline__', 'extern inline',
695and 'extern __inline__'.
696
697
698
6994) Don't over-design.
700
701Don't try to anticipate nebulous future cases which may or may not
Randy Dunlap84da7c02005-06-28 20:45:30 -0700702be useful: "Make it as simple as you can, and no simpler."
Linus Torvalds1da177e2005-04-16 15:20:36 -0700703
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800704
705
706----------------------
707SECTION 3 - REFERENCES
708----------------------
709
710Andrew Morton, "The perfect patch" (tpp).
FD Cami2223c652008-10-15 22:02:00 -0700711 <http://userweb.kernel.org/~akpm/stuff/tpp.txt>
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800712
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700713Jeff Garzik, "Linux kernel patch submission format".
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800714 <http://linux.yyz.us/patch-format.html>
715
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700716Greg Kroah-Hartman, "How to piss off a kernel subsystem maintainer".
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800717 <http://www.kroah.com/log/2005/03/31/>
718 <http://www.kroah.com/log/2005/07/08/>
719 <http://www.kroah.com/log/2005/10/19/>
Xose Vazquez Pereze1b114e2006-01-14 12:57:41 +0100720 <http://www.kroah.com/log/2006/01/11/>
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800721
Randy Dunlapbc7455f2006-07-30 03:03:45 -0700722NO!!!! No more huge patch bombs to linux-kernel@vger.kernel.org people!
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800723 <http://marc.theaimsgroup.com/?l=linux-kernel&m=112112749912944&w=2>
724
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700725Kernel Documentation/CodingStyle:
Qi Yong4db29c12007-06-12 13:06:49 +0800726 <http://users.sosdg.org/~qiyong/lxr/source/Documentation/CodingStyle>
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800727
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700728Linus Torvalds's mail on the canonical patch format:
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800729 <http://lkml.org/lkml/2005/4/7/183>
Andi Kleen95367272008-10-15 22:02:02 -0700730
731Andi Kleen, "On submitting kernel patches"
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300732 Some strategies to get difficult or controversial changes in.
Andi Kleen95367272008-10-15 22:02:02 -0700733 http://halobates.de/on-submitting-patches.pdf
734
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800735--