blob: 3c12fc1469403261002dd7fac1d25d583d819b1d [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
13If you are submitting a driver, also read Documentation/SubmittingDrivers.
14
15
16
17--------------------------------------------
18SECTION 1 - CREATING AND SENDING YOUR CHANGE
19--------------------------------------------
20
21
22
231) "diff -up"
24------------
25
26Use "diff -up" or "diff -uprN" to create patches.
27
28All changes to the Linux kernel occur in the form of patches, as
29generated by diff(1). When creating your patch, make sure to create it
30in "unified diff" format, as supplied by the '-u' argument to diff(1).
31Also, please use the '-p' argument which shows which C function each
32change is in - that makes the resultant diff a lot easier to read.
33Patches should be based in the root kernel source directory,
34not in any lower subdirectory.
35
36To create a patch for a single file, it is often sufficient to do:
37
Randy Dunlap84da7c02005-06-28 20:45:30 -070038 SRCTREE= linux-2.6
Linus Torvalds1da177e2005-04-16 15:20:36 -070039 MYFILE= drivers/net/mydriver.c
40
41 cd $SRCTREE
42 cp $MYFILE $MYFILE.orig
43 vi $MYFILE # make your change
44 cd ..
45 diff -up $SRCTREE/$MYFILE{.orig,} > /tmp/patch
46
47To create a patch for multiple files, you should unpack a "vanilla",
48or unmodified kernel source tree, and generate a diff against your
49own source tree. For example:
50
Randy Dunlap84da7c02005-06-28 20:45:30 -070051 MYSRC= /devel/linux-2.6
Linus Torvalds1da177e2005-04-16 15:20:36 -070052
Randy Dunlap84da7c02005-06-28 20:45:30 -070053 tar xvfz linux-2.6.12.tar.gz
54 mv linux-2.6.12 linux-2.6.12-vanilla
55 diff -uprN -X linux-2.6.12-vanilla/Documentation/dontdiff \
56 linux-2.6.12-vanilla $MYSRC > /tmp/patch
Linus Torvalds1da177e2005-04-16 15:20:36 -070057
58"dontdiff" is a list of files which are generated by the kernel during
59the build process, and should be ignored in any diff(1)-generated
Randy Dunlap84da7c02005-06-28 20:45:30 -070060patch. The "dontdiff" file is included in the kernel tree in
612.6.12 and later. For earlier kernel versions, you can get it
62from <http://www.xenotime.net/linux/doc/dontdiff>.
Linus Torvalds1da177e2005-04-16 15:20:36 -070063
64Make sure your patch does not include any extra files which do not
65belong in a patch submission. Make sure to review your patch -after-
66generated it with diff(1), to ensure accuracy.
67
68If your changes produce a lot of deltas, you may want to look into
69splitting them into individual patches which modify things in
Randy Dunlap84da7c02005-06-28 20:45:30 -070070logical stages. This will facilitate easier reviewing by other
Linus Torvalds1da177e2005-04-16 15:20:36 -070071kernel developers, very important if you want your patch accepted.
Randy Dunlap84da7c02005-06-28 20:45:30 -070072There are a number of scripts which can aid in this:
Linus Torvalds1da177e2005-04-16 15:20:36 -070073
74Quilt:
75http://savannah.nongnu.org/projects/quilt
76
77Randy Dunlap's patch scripts:
Randy Dunlap84da7c02005-06-28 20:45:30 -070078http://www.xenotime.net/linux/scripts/patching-scripts-002.tar.gz
Linus Torvalds1da177e2005-04-16 15:20:36 -070079
80Andrew Morton's patch scripts:
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -080081http://www.zip.com.au/~akpm/linux/patches/
82Instead of these scripts, quilt is the recommended patch management
83tool (see above).
Randy Dunlap84da7c02005-06-28 20:45:30 -070084
85
Linus Torvalds1da177e2005-04-16 15:20:36 -070086
872) Describe your changes.
88
89Describe the technical detail of the change(s) your patch includes.
90
91Be as specific as possible. The WORST descriptions possible include
92things like "update driver X", "bug fix for driver X", or "this patch
93includes updates for subsystem X. Please apply."
94
95If your description starts to get long, that's a sign that you probably
96need to split up your patch. See #3, next.
97
98
99
1003) Separate your changes.
101
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800102Separate _logical changes_ into a single patch file.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700103
104For example, if your changes include both bug fixes and performance
105enhancements for a single driver, separate those changes into two
106or more patches. If your changes include an API update, and a new
107driver which uses that new API, separate those into two patches.
108
109On the other hand, if you make a single change to numerous files,
110group those changes into a single patch. Thus a single logical change
111is contained within a single patch.
112
113If one patch depends on another patch in order for a change to be
114complete, that is OK. Simply note "this patch depends on patch X"
115in your patch description.
116
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800117If you cannot condense your patch set into a smaller set of patches,
118then only post say 15 or so at a time and wait for review and integration.
119
120
Linus Torvalds1da177e2005-04-16 15:20:36 -0700121
1224) Select e-mail destination.
123
124Look through the MAINTAINERS file and the source code, and determine
125if your change applies to a specific subsystem of the kernel, with
126an assigned maintainer. If so, e-mail that person.
127
128If no maintainer is listed, or the maintainer does not respond, send
129your patch to the primary Linux kernel developer's mailing list,
130linux-kernel@vger.kernel.org. Most kernel developers monitor this
131e-mail list, and can comment on your changes.
132
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800133
134Do not send more than 15 patches at once to the vger mailing lists!!!
135
136
Linus Torvalds1da177e2005-04-16 15:20:36 -0700137Linus Torvalds is the final arbiter of all changes accepted into the
138Linux kernel. His e-mail address is <torvalds@osdl.org>. He gets
139a lot of e-mail, so typically you should do your best to -avoid- sending
140him e-mail.
141
142Patches which are bug fixes, are "obvious" changes, or similarly
143require little discussion should be sent or CC'd to Linus. Patches
144which require discussion or do not have a clear advantage should
145usually be sent first to linux-kernel. Only after the patch is
146discussed should the patch then be submitted to Linus.
147
Linus Torvalds1da177e2005-04-16 15:20:36 -0700148
149
1505) Select your CC (e-mail carbon copy) list.
151
152Unless you have a reason NOT to do so, CC linux-kernel@vger.kernel.org.
153
154Other kernel developers besides Linus need to be aware of your change,
155so that they may comment on it and offer code review and suggestions.
156linux-kernel is the primary Linux kernel developer mailing list.
157Other mailing lists are available for specific subsystems, such as
158USB, framebuffer devices, the VFS, the SCSI subsystem, etc. See the
159MAINTAINERS file for a mailing list that relates specifically to
160your change.
161
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800162Majordomo lists of VGER.KERNEL.ORG at:
163 <http://vger.kernel.org/vger-lists.html>
164
Paul Jackson1caf1f02005-07-31 22:34:48 -0700165If changes affect userland-kernel interfaces, please send
166the MAN-PAGES maintainer (as listed in the MAINTAINERS file)
167a man-pages patch, or at least a notification of the change,
168so that some information makes its way into the manual pages.
169
Linus Torvalds1da177e2005-04-16 15:20:36 -0700170Even if the maintainer did not respond in step #4, make sure to ALWAYS
171copy the maintainer when you change their code.
172
173For small patches you may want to CC the Trivial Patch Monkey
Kees Cookf62870d2006-01-03 13:33:31 +0100174trivial@kernel.org managed by Adrian Bunk; which collects "trivial"
Linus Torvalds1da177e2005-04-16 15:20:36 -0700175patches. Trivial patches must qualify for one of the following rules:
176 Spelling fixes in documentation
177 Spelling fixes which could break grep(1).
178 Warning fixes (cluttering with useless warnings is bad)
179 Compilation fixes (only if they are actually correct)
180 Runtime fixes (only if they actually fix things)
181 Removing use of deprecated functions/macros (eg. check_region).
182 Contact detail and documentation fixes
183 Non-portable code replaced by portable code (even in arch-specific,
184 since people copy, as long as it's trivial)
185 Any fix by the author/maintainer of the file. (ie. patch monkey
186 in re-transmission mode)
Kees Cookf62870d2006-01-03 13:33:31 +0100187URL: <http://www.kernel.org/pub/linux/kernel/people/bunk/trivial/>
Randy Dunlap84da7c02005-06-28 20:45:30 -0700188
Linus Torvalds1da177e2005-04-16 15:20:36 -0700189
190
191
1926) No MIME, no links, no compression, no attachments. Just plain text.
193
194Linus and other kernel developers need to be able to read and comment
195on the changes you are submitting. It is important for a kernel
196developer to be able to "quote" your changes, using standard e-mail
197tools, so that they may comment on specific portions of your code.
198
199For this reason, all patches should be submitting e-mail "inline".
200WARNING: Be wary of your editor's word-wrap corrupting your patch,
201if you choose to cut-n-paste your patch.
202
203Do not attach the patch as a MIME attachment, compressed or not.
204Many popular e-mail applications will not always transmit a MIME
205attachment as plain text, making it impossible to comment on your
206code. A MIME attachment also takes Linus a bit more time to process,
207decreasing the likelihood of your MIME-attached change being accepted.
208
209Exception: If your mailer is mangling patches then someone may ask
210you to re-send them using MIME.
211
212
213
2147) E-mail size.
215
216When sending patches to Linus, always follow step #6.
217
218Large changes are not appropriate for mailing lists, and some
219maintainers. If your patch, uncompressed, exceeds 40 kB in size,
220it is preferred that you store your patch on an Internet-accessible
221server, and provide instead a URL (link) pointing to your patch.
222
223
224
2258) Name your kernel version.
226
227It is important to note, either in the subject line or in the patch
228description, the kernel version to which this patch applies.
229
230If the patch does not apply cleanly to the latest kernel version,
231Linus will not apply it.
232
233
234
2359) Don't get discouraged. Re-submit.
236
237After you have submitted your change, be patient and wait. If Linus
238likes your change and applies it, it will appear in the next version
239of the kernel that he releases.
240
241However, if your change doesn't appear in the next version of the
242kernel, there could be any number of reasons. It's YOUR job to
243narrow down those reasons, correct what was wrong, and submit your
244updated change.
245
246It is quite common for Linus to "drop" your patch without comment.
247That's the nature of the system. If he drops your patch, it could be
248due to
249* Your patch did not apply cleanly to the latest kernel version
250* Your patch was not sufficiently discussed on linux-kernel.
251* A style issue (see section 2),
252* An e-mail formatting issue (re-read this section)
253* A technical problem with your change
254* He gets tons of e-mail, and yours got lost in the shuffle
255* You are being annoying (See Figure 1)
256
257When in doubt, solicit comments on linux-kernel mailing list.
258
259
260
26110) Include PATCH in the subject
262
263Due to high e-mail traffic to Linus, and to linux-kernel, it is common
264convention to prefix your subject line with [PATCH]. This lets Linus
265and other kernel developers more easily distinguish patches from other
266e-mail discussions.
267
268
269
27011) Sign your work
271
272To improve tracking of who did what, especially with patches that can
273percolate to their final resting place in the kernel through several
274layers of maintainers, we've introduced a "sign-off" procedure on
275patches that are being emailed around.
276
277The sign-off is a simple line at the end of the explanation for the
278patch, which certifies that you wrote it or otherwise have the right to
279pass it on as a open-source patch. The rules are pretty simple: if you
280can certify the below:
281
Linus Torvaldscbd83da2005-06-13 17:51:55 -0700282 Developer's Certificate of Origin 1.1
Linus Torvalds1da177e2005-04-16 15:20:36 -0700283
284 By making a contribution to this project, I certify that:
285
286 (a) The contribution was created in whole or in part by me and I
287 have the right to submit it under the open source license
288 indicated in the file; or
289
290 (b) The contribution is based upon previous work that, to the best
291 of my knowledge, is covered under an appropriate open source
292 license and I have the right under that license to submit that
293 work with modifications, whether created in whole or in part
294 by me, under the same open source license (unless I am
295 permitted to submit under a different license), as indicated
296 in the file; or
297
298 (c) The contribution was provided directly to me by some other
299 person who certified (a), (b) or (c) and I have not modified
300 it.
301
Linus Torvaldscbd83da2005-06-13 17:51:55 -0700302 (d) I understand and agree that this project and the contribution
303 are public and that a record of the contribution (including all
304 personal information I submit with it, including my sign-off) is
305 maintained indefinitely and may be redistributed consistent with
306 this project or the open source license(s) involved.
307
Linus Torvalds1da177e2005-04-16 15:20:36 -0700308then you just add a line saying
309
Alexey Dobriyan9fd55592005-06-25 14:59:34 -0700310 Signed-off-by: Random J Developer <random@developer.example.org>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700311
312Some people also put extra tags at the end. They'll just be ignored for
313now, but you can do this to mark internal company procedures or just
314point out some special detail about the sign-off.
315
316
Paul Jackson75f84262005-10-02 18:01:42 -070031712) The canonical patch format
Randy Dunlap84da7c02005-06-28 20:45:30 -0700318
Paul Jackson75f84262005-10-02 18:01:42 -0700319The canonical patch subject line is:
320
Paul Jacksond6b9acc2005-10-03 00:29:10 -0700321 Subject: [PATCH 001/123] subsystem: summary phrase
Paul Jackson75f84262005-10-02 18:01:42 -0700322
323The canonical patch message body contains the following:
324
325 - A "from" line specifying the patch author.
326
327 - An empty line.
328
329 - The body of the explanation, which will be copied to the
330 permanent changelog to describe this patch.
331
332 - The "Signed-off-by:" lines, described above, which will
333 also go in the changelog.
334
335 - A marker line containing simply "---".
336
337 - Any additional comments not suitable for the changelog.
338
339 - The actual patch (diff output).
340
341The Subject line format makes it very easy to sort the emails
342alphabetically by subject line - pretty much any email reader will
343support that - since because the sequence number is zero-padded,
344the numerical and alphabetic sort is the same.
345
Paul Jacksond6b9acc2005-10-03 00:29:10 -0700346The "subsystem" in the email's Subject should identify which
347area or subsystem of the kernel is being patched.
348
349The "summary phrase" in the email's Subject should concisely
350describe the patch which that email contains. The "summary
351phrase" should not be a filename. Do not use the same "summary
352phrase" for every patch in a whole patch series.
353
354Bear in mind that the "summary phrase" of your email becomes
355a globally-unique identifier for that patch. It propagates
356all the way into the git changelog. The "summary phrase" may
357later be used in developer discussions which refer to the patch.
358People will want to google for the "summary phrase" to read
359discussion regarding that patch.
360
361A couple of example Subjects:
362
363 Subject: [patch 2/5] ext2: improve scalability of bitmap searching
364 Subject: [PATCHv2 001/207] x86: fix eflags tracking
Paul Jackson75f84262005-10-02 18:01:42 -0700365
366The "from" line must be the very first line in the message body,
367and has the form:
368
369 From: Original Author <author@example.com>
370
371The "from" line specifies who will be credited as the author of the
372patch in the permanent changelog. If the "from" line is missing,
373then the "From:" line from the email header will be used to determine
374the patch author in the changelog.
375
376The explanation body will be committed to the permanent source
377changelog, so should make sense to a competent reader who has long
378since forgotten the immediate details of the discussion that might
379have led to this patch.
380
381The "---" marker line serves the essential purpose of marking for patch
382handling tools where the changelog message ends.
383
384One good use for the additional comments after the "---" marker is for
385a diffstat, to show what files have changed, and the number of inserted
386and deleted lines per file. A diffstat is especially useful on bigger
387patches. Other comments relevant only to the moment or the maintainer,
388not suitable for the permanent changelog, should also go here.
389
390See more details on the proper patch format in the following
391references.
392
393
Randy Dunlap84da7c02005-06-28 20:45:30 -0700394
395
Linus Torvalds1da177e2005-04-16 15:20:36 -0700396-----------------------------------
397SECTION 2 - HINTS, TIPS, AND TRICKS
398-----------------------------------
399
400This section lists many of the common "rules" associated with code
401submitted to the kernel. There are always exceptions... but you must
402have a really good reason for doing so. You could probably call this
403section Linus Computer Science 101.
404
405
406
4071) Read Documentation/CodingStyle
408
409Nuff said. If your code deviates too much from this, it is likely
410to be rejected without further review, and without comment.
411
412
413
4142) #ifdefs are ugly
415
416Code cluttered with ifdefs is difficult to read and maintain. Don't do
417it. Instead, put your ifdefs in a header, and conditionally define
418'static inline' functions, or macros, which are used in the code.
419Let the compiler optimize away the "no-op" case.
420
421Simple example, of poor code:
422
423 dev = alloc_etherdev (sizeof(struct funky_private));
424 if (!dev)
425 return -ENODEV;
426 #ifdef CONFIG_NET_FUNKINESS
427 init_funky_net(dev);
428 #endif
429
430Cleaned-up example:
431
432(in header)
433 #ifndef CONFIG_NET_FUNKINESS
434 static inline void init_funky_net (struct net_device *d) {}
435 #endif
436
437(in the code itself)
438 dev = alloc_etherdev (sizeof(struct funky_private));
439 if (!dev)
440 return -ENODEV;
441 init_funky_net(dev);
442
443
444
4453) 'static inline' is better than a macro
446
447Static inline functions are greatly preferred over macros.
448They provide type safety, have no length limitations, no formatting
449limitations, and under gcc they are as cheap as macros.
450
451Macros should only be used for cases where a static inline is clearly
452suboptimal [there a few, isolated cases of this in fast paths],
453or where it is impossible to use a static inline function [such as
454string-izing].
455
456'static inline' is preferred over 'static __inline__', 'extern inline',
457and 'extern __inline__'.
458
459
460
4614) Don't over-design.
462
463Don't try to anticipate nebulous future cases which may or may not
Randy Dunlap84da7c02005-06-28 20:45:30 -0700464be useful: "Make it as simple as you can, and no simpler."
Linus Torvalds1da177e2005-04-16 15:20:36 -0700465
Xose Vazquez Perez5b0ed2c2006-01-08 01:02:49 -0800466
467
468----------------------
469SECTION 3 - REFERENCES
470----------------------
471
472Andrew Morton, "The perfect patch" (tpp).
473 <http://www.zip.com.au/~akpm/linux/patches/stuff/tpp.txt>
474
475Jeff Garzik, "Linux kernel patch submission format."
476 <http://linux.yyz.us/patch-format.html>
477
478Greg Kroah, "How to piss off a kernel subsystem maintainer".
479 <http://www.kroah.com/log/2005/03/31/>
480 <http://www.kroah.com/log/2005/07/08/>
481 <http://www.kroah.com/log/2005/10/19/>
482
483NO!!!! No more huge patch bombs to linux-kernel@vger.kernel.org people!.
484 <http://marc.theaimsgroup.com/?l=linux-kernel&m=112112749912944&w=2>
485
486Kernel Documentation/CodingStyle
487 <http://sosdg.org/~coywolf/lxr/source/Documentation/CodingStyle>
488
489Linus Torvald's mail on the canonical patch format:
490 <http://lkml.org/lkml/2005/4/7/183>
491--
492Last updated on 17 Nov 2005.