blob: 958182f6aa1f104a22659df31ac3a97ed8bf15cd [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:
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -080080http://www.zip.com.au/~akpm/linux/patches/
81Instead 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
94If your description starts to get long, that's a sign that you probably
95need to split up your patch. See #3, next.
96
97
98
993) Separate your changes.
100
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800101Separate _logical changes_ into a single patch file.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700102
103For example, if your changes include both bug fixes and performance
104enhancements for a single driver, separate those changes into two
105or more patches. If your changes include an API update, and a new
106driver which uses that new API, separate those into two patches.
107
108On the other hand, if you make a single change to numerous files,
109group those changes into a single patch. Thus a single logical change
110is contained within a single patch.
111
112If one patch depends on another patch in order for a change to be
113complete, that is OK. Simply note "this patch depends on patch X"
114in your patch description.
115
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800116If you cannot condense your patch set into a smaller set of patches,
117then only post say 15 or so at a time and wait for review and integration.
118
119
Linus Torvalds1da177e2005-04-16 15:20:36 -0700120
1214) Select e-mail destination.
122
123Look through the MAINTAINERS file and the source code, and determine
124if your change applies to a specific subsystem of the kernel, with
125an assigned maintainer. If so, e-mail that person.
126
127If no maintainer is listed, or the maintainer does not respond, send
128your patch to the primary Linux kernel developer's mailing list,
129linux-kernel@vger.kernel.org. Most kernel developers monitor this
130e-mail list, and can comment on your changes.
131
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800132
133Do not send more than 15 patches at once to the vger mailing lists!!!
134
135
Linus Torvalds1da177e2005-04-16 15:20:36 -0700136Linus Torvalds is the final arbiter of all changes accepted into the
137Linux kernel. His e-mail address is <torvalds@osdl.org>. He gets
138a lot of e-mail, so typically you should do your best to -avoid- sending
139him e-mail.
140
141Patches which are bug fixes, are "obvious" changes, or similarly
142require little discussion should be sent or CC'd to Linus. Patches
143which require discussion or do not have a clear advantage should
144usually be sent first to linux-kernel. Only after the patch is
145discussed should the patch then be submitted to Linus.
146
Linus Torvalds1da177e2005-04-16 15:20:36 -0700147
148
1495) Select your CC (e-mail carbon copy) list.
150
151Unless you have a reason NOT to do so, CC linux-kernel@vger.kernel.org.
152
153Other kernel developers besides Linus need to be aware of your change,
154so that they may comment on it and offer code review and suggestions.
155linux-kernel is the primary Linux kernel developer mailing list.
156Other mailing lists are available for specific subsystems, such as
157USB, framebuffer devices, the VFS, the SCSI subsystem, etc. See the
158MAINTAINERS file for a mailing list that relates specifically to
159your change.
160
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800161Majordomo lists of VGER.KERNEL.ORG at:
162 <http://vger.kernel.org/vger-lists.html>
163
Paul Jackson1caf1f02005-07-31 22:34:48 -0700164If changes affect userland-kernel interfaces, please send
165the MAN-PAGES maintainer (as listed in the MAINTAINERS file)
166a man-pages patch, or at least a notification of the change,
167so that some information makes its way into the manual pages.
168
Linus Torvalds1da177e2005-04-16 15:20:36 -0700169Even if the maintainer did not respond in step #4, make sure to ALWAYS
170copy the maintainer when you change their code.
171
172For small patches you may want to CC the Trivial Patch Monkey
Kees Cookf62870d2006-01-03 13:33:31 +0100173trivial@kernel.org managed by Adrian Bunk; which collects "trivial"
Linus Torvalds1da177e2005-04-16 15:20:36 -0700174patches. Trivial patches must qualify for one of the following rules:
175 Spelling fixes in documentation
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700176 Spelling fixes which could break grep(1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700177 Warning fixes (cluttering with useless warnings is bad)
178 Compilation fixes (only if they are actually correct)
179 Runtime fixes (only if they actually fix things)
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700180 Removing use of deprecated functions/macros (eg. check_region)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700181 Contact detail and documentation fixes
182 Non-portable code replaced by portable code (even in arch-specific,
183 since people copy, as long as it's trivial)
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700184 Any fix by the author/maintainer of the file (ie. patch monkey
Linus Torvalds1da177e2005-04-16 15:20:36 -0700185 in re-transmission mode)
Kees Cookf62870d2006-01-03 13:33:31 +0100186URL: <http://www.kernel.org/pub/linux/kernel/people/bunk/trivial/>
Randy Dunlap84da7c02005-06-28 20:45:30 -0700187
Linus Torvalds1da177e2005-04-16 15:20:36 -0700188
189
190
1916) No MIME, no links, no compression, no attachments. Just plain text.
192
193Linus and other kernel developers need to be able to read and comment
194on the changes you are submitting. It is important for a kernel
195developer to be able to "quote" your changes, using standard e-mail
196tools, so that they may comment on specific portions of your code.
197
198For this reason, all patches should be submitting e-mail "inline".
199WARNING: Be wary of your editor's word-wrap corrupting your patch,
200if you choose to cut-n-paste your patch.
201
202Do not attach the patch as a MIME attachment, compressed or not.
203Many popular e-mail applications will not always transmit a MIME
204attachment as plain text, making it impossible to comment on your
205code. A MIME attachment also takes Linus a bit more time to process,
206decreasing the likelihood of your MIME-attached change being accepted.
207
208Exception: If your mailer is mangling patches then someone may ask
209you to re-send them using MIME.
210
211
212
2137) E-mail size.
214
215When sending patches to Linus, always follow step #6.
216
217Large changes are not appropriate for mailing lists, and some
218maintainers. If your patch, uncompressed, exceeds 40 kB in size,
219it is preferred that you store your patch on an Internet-accessible
220server, and provide instead a URL (link) pointing to your patch.
221
222
223
2248) Name your kernel version.
225
226It is important to note, either in the subject line or in the patch
227description, the kernel version to which this patch applies.
228
229If the patch does not apply cleanly to the latest kernel version,
230Linus will not apply it.
231
232
233
2349) Don't get discouraged. Re-submit.
235
236After you have submitted your change, be patient and wait. If Linus
237likes your change and applies it, it will appear in the next version
238of the kernel that he releases.
239
240However, if your change doesn't appear in the next version of the
241kernel, there could be any number of reasons. It's YOUR job to
242narrow down those reasons, correct what was wrong, and submit your
243updated change.
244
245It is quite common for Linus to "drop" your patch without comment.
246That's the nature of the system. If he drops your patch, it could be
247due to
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700248* Your patch did not apply cleanly to the latest kernel version.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700249* Your patch was not sufficiently discussed on linux-kernel.
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700250* A style issue (see section 2).
251* An e-mail formatting issue (re-read this section).
252* A technical problem with your change.
253* He gets tons of e-mail, and yours got lost in the shuffle.
254* You are being annoying.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700255
256When in doubt, solicit comments on linux-kernel mailing list.
257
258
259
26010) Include PATCH in the subject
261
262Due to high e-mail traffic to Linus, and to linux-kernel, it is common
263convention to prefix your subject line with [PATCH]. This lets Linus
264and other kernel developers more easily distinguish patches from other
265e-mail discussions.
266
267
268
26911) Sign your work
270
271To improve tracking of who did what, especially with patches that can
272percolate to their final resting place in the kernel through several
273layers of maintainers, we've introduced a "sign-off" procedure on
274patches that are being emailed around.
275
276The sign-off is a simple line at the end of the explanation for the
277patch, which certifies that you wrote it or otherwise have the right to
278pass it on as a open-source patch. The rules are pretty simple: if you
279can certify the below:
280
Linus Torvaldscbd83da2005-06-13 17:51:55 -0700281 Developer's Certificate of Origin 1.1
Linus Torvalds1da177e2005-04-16 15:20:36 -0700282
283 By making a contribution to this project, I certify that:
284
285 (a) The contribution was created in whole or in part by me and I
286 have the right to submit it under the open source license
287 indicated in the file; or
288
289 (b) The contribution is based upon previous work that, to the best
290 of my knowledge, is covered under an appropriate open source
291 license and I have the right under that license to submit that
292 work with modifications, whether created in whole or in part
293 by me, under the same open source license (unless I am
294 permitted to submit under a different license), as indicated
295 in the file; or
296
297 (c) The contribution was provided directly to me by some other
298 person who certified (a), (b) or (c) and I have not modified
299 it.
300
Linus Torvaldscbd83da2005-06-13 17:51:55 -0700301 (d) I understand and agree that this project and the contribution
302 are public and that a record of the contribution (including all
303 personal information I submit with it, including my sign-off) is
304 maintained indefinitely and may be redistributed consistent with
305 this project or the open source license(s) involved.
306
Linus Torvalds1da177e2005-04-16 15:20:36 -0700307then you just add a line saying
308
Alexey Dobriyan9fd55592005-06-25 14:59:34 -0700309 Signed-off-by: Random J Developer <random@developer.example.org>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700310
Greg KHaf45f322006-09-12 20:35:52 -0700311using your real name (sorry, no pseudonyms or anonymous contributions.)
312
Linus Torvalds1da177e2005-04-16 15:20:36 -0700313Some people also put extra tags at the end. They'll just be ignored for
314now, but you can do this to mark internal company procedures or just
315point out some special detail about the sign-off.
316
317
Paul Jackson75f84262005-10-02 18:01:42 -070031812) The canonical patch format
Randy Dunlap84da7c02005-06-28 20:45:30 -0700319
Paul Jackson75f84262005-10-02 18:01:42 -0700320The canonical patch subject line is:
321
Paul Jacksond6b9acc2005-10-03 00:29:10 -0700322 Subject: [PATCH 001/123] subsystem: summary phrase
Paul Jackson75f84262005-10-02 18:01:42 -0700323
324The canonical patch message body contains the following:
325
326 - A "from" line specifying the patch author.
327
328 - An empty line.
329
330 - The body of the explanation, which will be copied to the
331 permanent changelog to describe this patch.
332
333 - The "Signed-off-by:" lines, described above, which will
334 also go in the changelog.
335
336 - A marker line containing simply "---".
337
338 - Any additional comments not suitable for the changelog.
339
340 - The actual patch (diff output).
341
342The Subject line format makes it very easy to sort the emails
343alphabetically by subject line - pretty much any email reader will
344support that - since because the sequence number is zero-padded,
345the numerical and alphabetic sort is the same.
346
Paul Jacksond6b9acc2005-10-03 00:29:10 -0700347The "subsystem" in the email's Subject should identify which
348area or subsystem of the kernel is being patched.
349
350The "summary phrase" in the email's Subject should concisely
351describe the patch which that email contains. The "summary
352phrase" should not be a filename. Do not use the same "summary
353phrase" for every patch in a whole patch series.
354
355Bear in mind that the "summary phrase" of your email becomes
356a globally-unique identifier for that patch. It propagates
357all the way into the git changelog. The "summary phrase" may
358later be used in developer discussions which refer to the patch.
359People will want to google for the "summary phrase" to read
360discussion regarding that patch.
361
362A couple of example Subjects:
363
364 Subject: [patch 2/5] ext2: improve scalability of bitmap searching
365 Subject: [PATCHv2 001/207] x86: fix eflags tracking
Paul Jackson75f84262005-10-02 18:01:42 -0700366
367The "from" line must be the very first line in the message body,
368and has the form:
369
370 From: Original Author <author@example.com>
371
372The "from" line specifies who will be credited as the author of the
373patch in the permanent changelog. If the "from" line is missing,
374then the "From:" line from the email header will be used to determine
375the patch author in the changelog.
376
377The explanation body will be committed to the permanent source
378changelog, so should make sense to a competent reader who has long
379since forgotten the immediate details of the discussion that might
380have led to this patch.
381
382The "---" marker line serves the essential purpose of marking for patch
383handling tools where the changelog message ends.
384
385One good use for the additional comments after the "---" marker is for
386a diffstat, to show what files have changed, and the number of inserted
387and deleted lines per file. A diffstat is especially useful on bigger
388patches. Other comments relevant only to the moment or the maintainer,
389not suitable for the permanent changelog, should also go here.
Randy Dunlap58591e82006-01-08 01:04:52 -0800390Use diffstat options "-p 1 -w 70" so that filenames are listed from the
391top of the kernel source tree and don't use too much horizontal space
392(easily fit in 80 columns, maybe with some indentation).
Paul Jackson75f84262005-10-02 18:01:42 -0700393
394See more details on the proper patch format in the following
395references.
396
397
Randy Dunlap84da7c02005-06-28 20:45:30 -0700398
399
Linus Torvalds1da177e2005-04-16 15:20:36 -0700400-----------------------------------
401SECTION 2 - HINTS, TIPS, AND TRICKS
402-----------------------------------
403
404This section lists many of the common "rules" associated with code
405submitted to the kernel. There are always exceptions... but you must
406have a really good reason for doing so. You could probably call this
407section Linus Computer Science 101.
408
409
410
4111) Read Documentation/CodingStyle
412
413Nuff said. If your code deviates too much from this, it is likely
414to be rejected without further review, and without comment.
415
416
417
4182) #ifdefs are ugly
419
420Code cluttered with ifdefs is difficult to read and maintain. Don't do
421it. Instead, put your ifdefs in a header, and conditionally define
422'static inline' functions, or macros, which are used in the code.
423Let the compiler optimize away the "no-op" case.
424
425Simple example, of poor code:
426
427 dev = alloc_etherdev (sizeof(struct funky_private));
428 if (!dev)
429 return -ENODEV;
430 #ifdef CONFIG_NET_FUNKINESS
431 init_funky_net(dev);
432 #endif
433
434Cleaned-up example:
435
436(in header)
437 #ifndef CONFIG_NET_FUNKINESS
438 static inline void init_funky_net (struct net_device *d) {}
439 #endif
440
441(in the code itself)
442 dev = alloc_etherdev (sizeof(struct funky_private));
443 if (!dev)
444 return -ENODEV;
445 init_funky_net(dev);
446
447
448
4493) 'static inline' is better than a macro
450
451Static inline functions are greatly preferred over macros.
452They provide type safety, have no length limitations, no formatting
453limitations, and under gcc they are as cheap as macros.
454
455Macros should only be used for cases where a static inline is clearly
456suboptimal [there a few, isolated cases of this in fast paths],
457or where it is impossible to use a static inline function [such as
458string-izing].
459
460'static inline' is preferred over 'static __inline__', 'extern inline',
461and 'extern __inline__'.
462
463
464
4654) Don't over-design.
466
467Don't try to anticipate nebulous future cases which may or may not
Randy Dunlap84da7c02005-06-28 20:45:30 -0700468be useful: "Make it as simple as you can, and no simpler."
Linus Torvalds1da177e2005-04-16 15:20:36 -0700469
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800470
471
472----------------------
473SECTION 3 - REFERENCES
474----------------------
475
476Andrew Morton, "The perfect patch" (tpp).
477 <http://www.zip.com.au/~akpm/linux/patches/stuff/tpp.txt>
478
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700479Jeff Garzik, "Linux kernel patch submission format".
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800480 <http://linux.yyz.us/patch-format.html>
481
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700482Greg Kroah-Hartman, "How to piss off a kernel subsystem maintainer".
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800483 <http://www.kroah.com/log/2005/03/31/>
484 <http://www.kroah.com/log/2005/07/08/>
485 <http://www.kroah.com/log/2005/10/19/>
Xose Vazquez Pereze1b114e2006-01-14 12:57:41 +0100486 <http://www.kroah.com/log/2006/01/11/>
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800487
Randy Dunlapbc7455f2006-07-30 03:03:45 -0700488NO!!!! No more huge patch bombs to linux-kernel@vger.kernel.org people!
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800489 <http://marc.theaimsgroup.com/?l=linux-kernel&m=112112749912944&w=2>
490
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700491Kernel Documentation/CodingStyle:
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800492 <http://sosdg.org/~coywolf/lxr/source/Documentation/CodingStyle>
493
Pavel Machek8e9cb8f2006-09-29 02:01:29 -0700494Linus Torvalds's mail on the canonical patch format:
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800495 <http://lkml.org/lkml/2005/4/7/183>
496--