2 How to Get Your Change Into the Linux Kernel
4 Care And Operation Of Your Linus Torvalds
8 For a person or company who wishes to submit a change to the Linux
9 kernel, the process can sometimes be daunting if you're not familiar
10 with "the system." This text is a collection of suggestions which
11 can greatly increase the chances of your change being accepted.
13 If you are submitting a driver, also read Documentation/SubmittingDrivers.
17 --------------------------------------------
18 SECTION 1 - CREATING AND SENDING YOUR CHANGE
19 --------------------------------------------
26 Use "diff -up" or "diff -uprN" to create patches.
28 All changes to the Linux kernel occur in the form of patches, as
29 generated by diff(1). When creating your patch, make sure to create it
30 in "unified diff" format, as supplied by the '-u' argument to diff(1).
31 Also, please use the '-p' argument which shows which C function each
32 change is in - that makes the resultant diff a lot easier to read.
33 Patches should be based in the root kernel source directory,
34 not in any lower subdirectory.
36 To create a patch for a single file, it is often sufficient to do:
38 SRCTREE= /devel/linux-2.4
39 MYFILE= drivers/net/mydriver.c
42 cp $MYFILE $MYFILE.orig
43 vi $MYFILE # make your change
44 diff -up $MYFILE.orig $MYFILE > /tmp/patch
46 To create a patch for multiple files, you should unpack a "vanilla",
47 or unmodified kernel source tree, and generate a diff against your
48 own source tree. For example:
50 MYSRC= /devel/linux-2.4
52 tar xvfz linux-2.4.0-test11.tar.gz
53 mv linux linux-vanilla
54 wget http://www.moses.uklinux.net/patches/dontdiff
55 diff -uprN -X dontdiff linux-vanilla $MYSRC > /tmp/patch
58 "dontdiff" is a list of files which are generated by the kernel during
59 the build process, and should be ignored in any diff(1)-generated
60 patch. dontdiff is maintained by Tigran Aivazian <tigran@veritas.com>
62 Make sure your patch does not include any extra files which do not
63 belong in a patch submission. Make sure to review your patch -after-
64 generated it with diff(1), to ensure accuracy.
67 2) Describe your changes.
69 Describe the technical detail of the change(s) your patch includes.
71 Be as specific as possible. The WORST descriptions possible include
72 things like "update driver X", "bug fix for driver X", or "this patch
73 includes updates for subsystem X. Please apply."
75 If your description starts to get long, that's a sign that you probably
76 need to split up your patch. See #3, next.
80 3) Separate your changes.
82 Separate each logical change into its own patch.
84 For example, if your changes include both bug fixes and performance
85 enhancements for a single driver, separate those changes into two
86 or more patches. If your changes include an API update, and a new
87 driver which uses that new API, separate those into two patches.
89 On the other hand, if you make a single change to numerous files,
90 group those changes into a single patch. Thus a single logical change
91 is contained within a single patch.
93 If one patch depends on another patch in order for a change to be
94 complete, that is OK. Simply note "this patch depends on patch X"
95 in your patch description.
98 4) Select e-mail destination.
100 Look through the MAINTAINERS file and the source code, and determine
101 if your change applies to a specific subsystem of the kernel, with
102 an assigned maintainer. If so, e-mail that person.
104 If no maintainer is listed, or the maintainer does not respond, send
105 your patch to the primary Linux kernel developer's mailing list,
106 linux-kernel@vger.kernel.org. Most kernel developers monitor this
107 e-mail list, and can comment on your changes.
109 Linus Torvalds is the final arbiter of all changes accepted into the
110 Linux kernel. His e-mail address is <torvalds@osdl.org>. He gets
111 a lot of e-mail, so typically you should do your best to -avoid- sending
114 Patches which are bug fixes, are "obvious" changes, or similarly
115 require little discussion should be sent or CC'd to Linus. Patches
116 which require discussion or do not have a clear advantage should
117 usually be sent first to linux-kernel. Only after the patch is
118 discussed should the patch then be submitted to Linus.
122 5) Select your CC (e-mail carbon copy) list.
124 Unless you have a reason NOT to do so, CC linux-kernel@vger.kernel.org.
126 Other kernel developers besides Linus need to be aware of your change,
127 so that they may comment on it and offer code review and suggestions.
128 linux-kernel is the primary Linux kernel developer mailing list.
129 Other mailing lists are available for specific subsystems, such as
130 USB, framebuffer devices, the VFS, the SCSI subsystem, etc. See the
131 MAINTAINERS file for a mailing list that relates specifically to
134 Even if the maintainer did not respond in step #4, make sure to ALWAYS
135 copy the maintainer when you change their code.
137 For small patches you may want to CC the Trivial Patch Monkey
138 trivial@rustcorp.com.au set up by Rusty Russell; which collects "trivial"
139 patches. Trivial patches must qualify for one of the following rules:
140 Spelling fixes in documentation
141 Spelling fixes which could break grep(1).
142 Warning fixes (cluttering with useless warnings is bad)
143 Compilation fixes (only if they are actually correct)
144 Runtime fixes (only if they actually fix things)
145 Removing use of deprecated functions/macros (eg. check_region).
146 Contact detail and documentation fixes
147 Non-portable code replaced by portable code (even in arch-specific,
148 since people copy, as long as it's trivial)
149 Any fix by the author/maintainer of the file. (ie. patch monkey
150 in re-transmission mode)
154 6) No MIME, no links, no compression, no attachments. Just plain text.
156 Linus and other kernel developers need to be able to read and comment
157 on the changes you are submitting. It is important for a kernel
158 developer to be able to "quote" your changes, using standard e-mail
159 tools, so that they may comment on specific portions of your code.
161 For this reason, all patches should be submitting e-mail "inline".
162 WARNING: Be wary of your editor's word-wrap corrupting your patch,
163 if you choose to cut-n-paste your patch.
165 Do not attach the patch as a MIME attachment, compressed or not.
166 Many popular e-mail applications will not always transmit a MIME
167 attachment as plain text, making it impossible to comment on your
168 code. A MIME attachment also takes Linus a bit more time to process,
169 decreasing the likelihood of your MIME-attached change being accepted.
171 Exception: If your mailer is mangling patches then someone may ask
172 you to re-send them using MIME.
178 When sending patches to Linus, always follow step #6.
180 Large changes are not appropriate for mailing lists, and some
181 maintainers. If your patch, uncompressed, exceeds 40 kB in size,
182 it is preferred that you store your patch on an Internet-accessible
183 server, and provide instead a URL (link) pointing to your patch.
187 8) Name your kernel version.
189 It is important to note, either in the subject line or in the patch
190 description, the kernel version to which this patch applies.
192 If the patch does not apply cleanly to the latest kernel version,
193 Linus will not apply it.
197 9) Don't get discouraged. Re-submit.
199 After you have submitted your change, be patient and wait. If Linus
200 likes your change and applies it, it will appear in the next version
201 of the kernel that he releases.
203 However, if your change doesn't appear in the next version of the
204 kernel, there could be any number of reasons. It's YOUR job to
205 narrow down those reasons, correct what was wrong, and submit your
208 It is quite common for Linus to "drop" your patch without comment.
209 That's the nature of the system. If he drops your patch, it could be
211 * Your patch did not apply cleanly to the latest kernel version
212 * Your patch was not sufficiently discussed on linux-kernel.
213 * A style issue (see section 2),
214 * An e-mail formatting issue (re-read this section)
215 * A technical problem with your change
216 * He gets tons of e-mail, and yours got lost in the shuffle
217 * You are being annoying (See Figure 1)
219 When in doubt, solicit comments on linux-kernel mailing list.
223 10) Include PATCH in the subject
225 Due to high e-mail traffic to Linus, and to linux-kernel, it is common
226 convention to prefix your subject line with [PATCH]. This lets Linus
227 and other kernel developers more easily distinguish patches from other
232 -----------------------------------
233 SECTION 2 - HINTS, TIPS, AND TRICKS
234 -----------------------------------
236 This section lists many of the common "rules" associated with code
237 submitted to the kernel. There are always exceptions... but you must
238 have a really good reason for doing so. You could probably call this
239 section Linus Computer Science 101.
243 1) Read Documentation/CodingStyle
245 Nuff said. If your code deviates too much from this, it is likely
246 to be rejected without further review, and without comment.
252 Code cluttered with ifdefs is difficult to read and maintain. Don't do
253 it. Instead, put your ifdefs in a header, and conditionally define
254 'static inline' functions, or macros, which are used in the code.
255 Let the compiler optimize away the "no-op" case.
257 Simple example, of poor code:
259 dev = alloc_etherdev (sizeof(struct funky_private));
262 #ifdef CONFIG_NET_FUNKINESS
269 #ifndef CONFIG_NET_FUNKINESS
270 static inline void init_funky_net (struct net_device *d) {}
274 dev = alloc_etherdev (sizeof(struct funky_private));
281 3) 'static inline' is better than a macro
283 Static inline functions are greatly preferred over macros.
284 They provide type safety, have no length limitations, no formatting
285 limitations, and under gcc they are as cheap as macros.
287 Macros should only be used for cases where a static inline is clearly
288 suboptimal [there a few, isolated cases of this in fast paths],
289 or where it is impossible to use a static inline function [such as
292 'static inline' is preferred over 'static __inline__', 'extern inline',
293 and 'extern __inline__'.
297 4) Don't over-design.
299 Don't try to anticipate nebulous future cases which may or may not
300 be useful: "Make it as simple as you can, and no simpler"