1
HOWTO do Linux kernel development
2
---------------------------------
4
This is the be-all, end-all document on this topic. It contains
5
instructions on how to become a Linux kernel developer and how to learn
6
to work with the Linux kernel development community. It tries to not
7
contain anything related to the technical aspects of kernel programming,
8
but will help point you in the right direction for that.
10
If anything in this document becomes out of date, please send in patches
11
to the maintainer of this file, who is listed at the bottom of the
18
So, you want to learn how to become a Linux kernel developer? Or you
19
have been told by your manager, "Go write a Linux driver for this
20
device." This document's goal is to teach you everything you need to
21
know to achieve this by describing the process you need to go through,
22
and hints on how to work with the community. It will also try to
23
explain some of the reasons why the community works like it does.
25
The kernel is written mostly in C, with some architecture-dependent
26
parts written in assembly. A good understanding of C is required for
27
kernel development. Assembly (any architecture) is not required unless
28
you plan to do low-level development for that architecture. Though they
29
are not a good substitute for a solid C education and/or years of
30
experience, the following books are good for, if anything, reference:
31
- "The C Programming Language" by Kernighan and Ritchie [Prentice Hall]
32
- "Practical C Programming" by Steve Oualline [O'Reilly]
33
- "C: A Reference Manual" by Harbison and Steele [Prentice Hall]
35
The kernel is written using GNU C and the GNU toolchain. While it
36
adheres to the ISO C89 standard, it uses a number of extensions that are
37
not featured in the standard. The kernel is a freestanding C
38
environment, with no reliance on the standard C library, so some
39
portions of the C standard are not supported. Arbitrary long long
40
divisions and floating point are not allowed. It can sometimes be
41
difficult to understand the assumptions the kernel has on the toolchain
42
and the extensions that it uses, and unfortunately there is no
43
definitive reference for them. Please check the gcc info pages (`info
44
gcc`) for some information on them.
46
Please remember that you are trying to learn how to work with the
47
existing development community. It is a diverse group of people, with
48
high standards for coding, style and procedure. These standards have
49
been created over time based on what they have found to work best for
50
such a large and geographically dispersed team. Try to learn as much as
51
possible about these standards ahead of time, as they are well
52
documented; do not expect people to adapt to you or your company's way
59
The Linux kernel source code is released under the GPL. Please see the
60
file, COPYING, in the main directory of the source tree, for details on
61
the license. If you have further questions about the license, please
62
contact a lawyer, and do not ask on the Linux kernel mailing list. The
63
people on the mailing lists are not lawyers, and you should not rely on
64
their statements on legal matters.
66
For common questions and answers about the GPL, please see:
67
http://www.gnu.org/licenses/gpl-faq.html
73
The Linux kernel source tree has a large range of documents that are
74
invaluable for learning how to interact with the kernel community. When
75
new features are added to the kernel, it is recommended that new
76
documentation files are also added which explain how to use the feature.
77
When a kernel change causes the interface that the kernel exposes to
78
userspace to change, it is recommended that you send the information or
79
a patch to the manual pages explaining the change to the manual pages
80
maintainer at mtk.manpages@gmail.com, and CC the list
81
linux-api@vger.kernel.org.
83
Here is a list of files that are in the kernel source tree that are
86
This file gives a short background on the Linux kernel and describes
87
what is necessary to do to configure and build the kernel. People
88
who are new to the kernel should start here.
91
This file gives a list of the minimum levels of various software
92
packages that are necessary to build and run the kernel
95
Documentation/CodingStyle
96
This describes the Linux kernel coding style, and some of the
97
rationale behind it. All new code is expected to follow the
98
guidelines in this document. Most maintainers will only accept
99
patches if these rules are followed, and many people will only
100
review code if it is in the proper style.
102
Documentation/SubmittingPatches
103
Documentation/SubmittingDrivers
104
These files describe in explicit detail how to successfully create
105
and send a patch, including (but not limited to):
109
Following these rules will not guarantee success (as all patches are
110
subject to scrutiny for content and style), but not following them
111
will almost always prevent it.
113
Other excellent descriptions of how to create patches properly are:
115
http://userweb.kernel.org/~akpm/stuff/tpp.txt
116
"Linux kernel patch submission format"
117
http://linux.yyz.us/patch-format.html
119
Documentation/stable_api_nonsense.txt
120
This file describes the rationale behind the conscious decision to
121
not have a stable API within the kernel, including things like:
122
- Subsystem shim-layers (for compatibility?)
123
- Driver portability between Operating Systems.
124
- Mitigating rapid change within the kernel source tree (or
125
preventing rapid change)
126
This document is crucial for understanding the Linux development
127
philosophy and is very important for people moving to Linux from
128
development on other Operating Systems.
130
Documentation/SecurityBugs
131
If you feel you have found a security problem in the Linux kernel,
132
please follow the steps in this document to help notify the kernel
133
developers, and help solve the issue.
135
Documentation/ManagementStyle
136
This document describes how Linux kernel maintainers operate and the
137
shared ethos behind their methodologies. This is important reading
138
for anyone new to kernel development (or anyone simply curious about
139
it), as it resolves a lot of common misconceptions and confusion
140
about the unique behavior of kernel maintainers.
142
Documentation/stable_kernel_rules.txt
143
This file describes the rules on how the stable kernel releases
144
happen, and what to do if you want to get a change into one of these
147
Documentation/kernel-docs.txt
148
A list of external documentation that pertains to kernel
149
development. Please consult this list if you do not find what you
150
are looking for within the in-kernel documentation.
152
Documentation/applying-patches.txt
153
A good introduction describing exactly what a patch is and how to
154
apply it to the different development branches of the kernel.
156
The kernel also has a large number of documents that can be
157
automatically generated from the source code itself. This includes a
158
full description of the in-kernel API, and rules on how to handle
159
locking properly. The documents will be created in the
160
Documentation/DocBook/ directory and can be generated as PDF,
161
Postscript, HTML, and man pages by running:
166
respectively from the main kernel source directory.
169
Becoming A Kernel Developer
170
---------------------------
172
If you do not know anything about Linux kernel development, you should
173
look at the Linux KernelNewbies project:
174
http://kernelnewbies.org
175
It consists of a helpful mailing list where you can ask almost any type
176
of basic kernel development question (make sure to search the archives
177
first, before asking something that has already been answered in the
178
past.) It also has an IRC channel that you can use to ask questions in
179
real-time, and a lot of helpful documentation that is useful for
180
learning about Linux kernel development.
182
The website has basic information about code organization, subsystems,
183
and current projects (both in-tree and out-of-tree). It also describes
184
some basic logistical information, like how to compile a kernel and
187
If you do not know where you want to start, but you want to look for
188
some task to start doing to join into the kernel development community,
189
go to the Linux Kernel Janitor's project:
190
http://kernelnewbies.org/KernelJanitors
191
It is a great place to start. It describes a list of relatively simple
192
problems that need to be cleaned up and fixed within the Linux kernel
193
source tree. Working with the developers in charge of this project, you
194
will learn the basics of getting your patch into the Linux kernel tree,
195
and possibly be pointed in the direction of what to go work on next, if
196
you do not already have an idea.
198
If you already have a chunk of code that you want to put into the kernel
199
tree, but need some help getting it in the proper form, the
200
kernel-mentors project was created to help you out with this. It is a
201
mailing list, and can be found at:
202
http://selenic.com/mailman/listinfo/kernel-mentors
204
Before making any actual modifications to the Linux kernel code, it is
205
imperative to understand how the code in question works. For this
206
purpose, nothing is better than reading through it directly (most tricky
207
bits are commented well), perhaps even with the help of specialized
208
tools. One such tool that is particularly recommended is the Linux
209
Cross-Reference project, which is able to present source code in a
210
self-referential, indexed webpage format. An excellent up-to-date
211
repository of the kernel code may be found at:
212
http://lxr.linux.no/+trees
215
The development process
216
-----------------------
218
Linux kernel development process currently consists of a few different
219
main kernel "branches" and lots of different subsystem-specific kernel
220
branches. These different branches are:
221
- main 2.6.x kernel tree
222
- 2.6.x.y -stable kernel tree
223
- 2.6.x -git kernel patches
224
- subsystem specific kernel trees and patches
225
- the 2.6.x -next kernel tree for integration tests
229
2.6.x kernels are maintained by Linus Torvalds, and can be found on
230
kernel.org in the pub/linux/kernel/v2.6/ directory. Its development
231
process is as follows:
232
- As soon as a new kernel is released a two weeks window is open,
233
during this period of time maintainers can submit big diffs to
234
Linus, usually the patches that have already been included in the
235
-next kernel for a few weeks. The preferred way to submit big changes
236
is using git (the kernel's source management tool, more information
237
can be found at http://git-scm.com/) but plain patches are also just
239
- After two weeks a -rc1 kernel is released it is now possible to push
240
only patches that do not include new features that could affect the
241
stability of the whole kernel. Please note that a whole new driver
242
(or filesystem) might be accepted after -rc1 because there is no
243
risk of causing regressions with such a change as long as the change
244
is self-contained and does not affect areas outside of the code that
245
is being added. git can be used to send patches to Linus after -rc1
246
is released, but the patches need to also be sent to a public
247
mailing list for review.
248
- A new -rc is released whenever Linus deems the current git tree to
249
be in a reasonably sane state adequate for testing. The goal is to
250
release a new -rc kernel every week.
251
- Process continues until the kernel is considered "ready", the
252
process should last around 6 weeks.
253
- Known regressions in each release are periodically posted to the
254
linux-kernel mailing list. The goal is to reduce the length of
255
that list to zero before declaring the kernel to be "ready," but, in
256
the real world, a small number of regressions often remain at
259
It is worth mentioning what Andrew Morton wrote on the linux-kernel
260
mailing list about kernel releases:
261
"Nobody knows when a kernel will be released, because it's
262
released according to perceived bug status, not according to a
263
preconceived timeline."
265
2.6.x.y -stable kernel tree
266
---------------------------
267
Kernels with 4-part versions are -stable kernels. They contain
268
relatively small and critical fixes for security problems or significant
269
regressions discovered in a given 2.6.x kernel.
271
This is the recommended branch for users who want the most recent stable
272
kernel and are not interested in helping test development/experimental
275
If no 2.6.x.y kernel is available, then the highest numbered 2.6.x
276
kernel is the current stable kernel.
278
2.6.x.y are maintained by the "stable" team <stable@kernel.org>, and are
279
released as needs dictate. The normal release period is approximately
280
two weeks, but it can be longer if there are no pressing problems. A
281
security-related problem, instead, can cause a release to happen almost
284
The file Documentation/stable_kernel_rules.txt in the kernel tree
285
documents what kinds of changes are acceptable for the -stable tree, and
286
how the release process works.
290
These are daily snapshots of Linus' kernel tree which are managed in a
291
git repository (hence the name.) These patches are usually released
292
daily and represent the current state of Linus' tree. They are more
293
experimental than -rc kernels since they are generated automatically
294
without even a cursory glance to see if they are sane.
296
Subsystem Specific kernel trees and patches
297
-------------------------------------------
298
The maintainers of the various kernel subsystems --- and also many
299
kernel subsystem developers --- expose their current state of
300
development in source repositories. That way, others can see what is
301
happening in the different areas of the kernel. In areas where
302
development is rapid, a developer may be asked to base his submissions
303
onto such a subsystem kernel tree so that conflicts between the
304
submission and other already ongoing work are avoided.
306
Most of these repositories are git trees, but there are also other SCMs
307
in use, or patch queues being published as quilt series. Addresses of
308
these subsystem repositories are listed in the MAINTAINERS file. Many
309
of them can be browsed at http://git.kernel.org/.
311
Before a proposed patch is committed to such a subsystem tree, it is
312
subject to review which primarily happens on mailing lists (see the
313
respective section below). For several kernel subsystems, this review
314
process is tracked with the tool patchwork. Patchwork offers a web
315
interface which shows patch postings, any comments on a patch or
316
revisions to it, and maintainers can mark patches as under review,
317
accepted, or rejected. Most of these patchwork sites are listed at
318
http://patchwork.kernel.org/.
320
2.6.x -next kernel tree for integration tests
321
---------------------------------------------
322
Before updates from subsystem trees are merged into the mainline 2.6.x
323
tree, they need to be integration-tested. For this purpose, a special
324
testing repository exists into which virtually all subsystem trees are
325
pulled on an almost daily basis:
326
http://git.kernel.org/?p=linux/kernel/git/sfr/linux-next.git
327
http://linux.f-seidel.de/linux-next/pmwiki/
329
This way, the -next kernel gives a summary outlook onto what will be
330
expected to go into the mainline kernel at the next merge period.
331
Adventurous testers are very welcome to runtime-test the -next kernel.
337
bugzilla.kernel.org is where the Linux kernel developers track kernel
338
bugs. Users are encouraged to report all bugs that they find in this
339
tool. For details on how to use the kernel bugzilla, please see:
340
http://bugzilla.kernel.org/page.cgi?id=faq.html
342
The file REPORTING-BUGS in the main kernel source directory has a good
343
template for how to report a possible kernel bug, and details what kind
344
of information is needed by the kernel developers to help track down the
351
One of the best ways to put into practice your hacking skills is by fixing
352
bugs reported by other people. Not only you will help to make the kernel
353
more stable, you'll learn to fix real world problems and you will improve
354
your skills, and other developers will be aware of your presence. Fixing
355
bugs is one of the best ways to get merits among other developers, because
356
not many people like wasting time fixing other people's bugs.
358
To work in the already reported bug reports, go to http://bugzilla.kernel.org.
359
If you want to be advised of the future bug reports, you can subscribe to the
360
bugme-new mailing list (only new bug reports are mailed here) or to the
361
bugme-janitor mailing list (every change in the bugzilla is mailed here)
363
http://lists.linux-foundation.org/mailman/listinfo/bugme-new
364
http://lists.linux-foundation.org/mailman/listinfo/bugme-janitors
371
As some of the above documents describe, the majority of the core kernel
372
developers participate on the Linux Kernel Mailing list. Details on how
373
to subscribe and unsubscribe from the list can be found at:
374
http://vger.kernel.org/vger-lists.html#linux-kernel
375
There are archives of the mailing list on the web in many different
376
places. Use a search engine to find these archives. For example:
377
http://dir.gmane.org/gmane.linux.kernel
378
It is highly recommended that you search the archives about the topic
379
you want to bring up, before you post it to the list. A lot of things
380
already discussed in detail are only recorded at the mailing list
383
Most of the individual kernel subsystems also have their own separate
384
mailing list where they do their development efforts. See the
385
MAINTAINERS file for a list of what these lists are for the different
388
Many of the lists are hosted on kernel.org. Information on them can be
390
http://vger.kernel.org/vger-lists.html
392
Please remember to follow good behavioral habits when using the lists.
393
Though a bit cheesy, the following URL has some simple guidelines for
394
interacting with the list (or any list):
395
http://www.albion.com/netiquette/
397
If multiple people respond to your mail, the CC: list of recipients may
398
get pretty large. Don't remove anybody from the CC: list without a good
399
reason, or don't reply only to the list address. Get used to receiving the
400
mail twice, one from the sender and the one from the list, and don't try
401
to tune that by adding fancy mail-headers, people will not like it.
403
Remember to keep the context and the attribution of your replies intact,
404
keep the "John Kernelhacker wrote ...:" lines at the top of your reply, and
405
add your statements between the individual quoted sections instead of
406
writing at the top of the mail.
408
If you add patches to your mail, make sure they are plain readable text
409
as stated in Documentation/SubmittingPatches. Kernel developers don't
410
want to deal with attachments or compressed patches; they may want
411
to comment on individual lines of your patch, which works only that way.
412
Make sure you use a mail program that does not mangle spaces and tab
413
characters. A good first test is to send the mail to yourself and try
414
to apply your own patch by yourself. If that doesn't work, get your
415
mail program fixed or change it until it works.
417
Above all, please remember to show respect to other subscribers.
420
Working with the community
421
--------------------------
423
The goal of the kernel community is to provide the best possible kernel
424
there is. When you submit a patch for acceptance, it will be reviewed
425
on its technical merits and those alone. So, what should you be
429
- requests for change
430
- requests for justification
433
Remember, this is part of getting your patch into the kernel. You have
434
to be able to take criticism and comments about your patches, evaluate
435
them at a technical level and either rework your patches or provide
436
clear and concise reasoning as to why those changes should not be made.
437
If there are no responses to your posting, wait a few days and try
438
again, sometimes things get lost in the huge volume.
440
What should you not do?
441
- expect your patch to be accepted without question
444
- resubmit the patch without making any of the requested changes
446
In a community that is looking for the best technical solution possible,
447
there will always be differing opinions on how beneficial a patch is.
448
You have to be cooperative, and willing to adapt your idea to fit within
449
the kernel. Or at least be willing to prove your idea is worth it.
450
Remember, being wrong is acceptable as long as you are willing to work
451
toward a solution that is right.
453
It is normal that the answers to your first patch might simply be a list
454
of a dozen things you should correct. This does _not_ imply that your
455
patch will not be accepted, and it is _not_ meant against you
456
personally. Simply correct all issues raised against your patch and
460
Differences between the kernel community and corporate structures
461
-----------------------------------------------------------------
463
The kernel community works differently than most traditional corporate
464
development environments. Here are a list of things that you can try to
465
do to try to avoid problems:
466
Good things to say regarding your proposed changes:
467
- "This solves multiple problems."
468
- "This deletes 2000 lines of code."
469
- "Here is a patch that explains what I am trying to describe."
470
- "I tested it on 5 different architectures..."
471
- "Here is a series of small patches that..."
472
- "This increases performance on typical machines..."
474
Bad things you should avoid saying:
475
- "We did it this way in AIX/ptx/Solaris, so therefore it must be
477
- "I've being doing this for 20 years, so..."
478
- "This is required for my company to make money"
479
- "This is for our Enterprise product line."
480
- "Here is my 1000 page design document that describes my idea"
481
- "I've been working on this for 6 months..."
482
- "Here's a 5000 line patch that..."
483
- "I rewrote all of the current mess, and here it is..."
484
- "I have a deadline, and this patch needs to be applied now."
486
Another way the kernel community is different than most traditional
487
software engineering work environments is the faceless nature of
488
interaction. One benefit of using email and irc as the primary forms of
489
communication is the lack of discrimination based on gender or race.
490
The Linux kernel work environment is accepting of women and minorities
491
because all you are is an email address. The international aspect also
492
helps to level the playing field because you can't guess gender based on
493
a person's name. A man may be named Andrea and a woman may be named Pat.
494
Most women who have worked in the Linux kernel and have expressed an
495
opinion have had positive experiences.
497
The language barrier can cause problems for some people who are not
498
comfortable with English. A good grasp of the language can be needed in
499
order to get ideas across properly on mailing lists, so it is
500
recommended that you check your emails to make sure they make sense in
501
English before sending them.
504
Break up your changes
505
---------------------
507
The Linux kernel community does not gladly accept large chunks of code
508
dropped on it all at once. The changes need to be properly introduced,
509
discussed, and broken up into tiny, individual portions. This is almost
510
the exact opposite of what companies are used to doing. Your proposal
511
should also be introduced very early in the development process, so that
512
you can receive feedback on what you are doing. It also lets the
513
community feel that you are working with them, and not simply using them
514
as a dumping ground for your feature. However, don't send 50 emails at
515
one time to a mailing list, your patch series should be smaller than
516
that almost all of the time.
518
The reasons for breaking things up are the following:
520
1) Small patches increase the likelihood that your patches will be
521
applied, since they don't take much time or effort to verify for
522
correctness. A 5 line patch can be applied by a maintainer with
523
barely a second glance. However, a 500 line patch may take hours to
524
review for correctness (the time it takes is exponentially
525
proportional to the size of the patch, or something).
527
Small patches also make it very easy to debug when something goes
528
wrong. It's much easier to back out patches one by one than it is
529
to dissect a very large patch after it's been applied (and broken
532
2) It's important not only to send small patches, but also to rewrite
533
and simplify (or simply re-order) patches before submitting them.
535
Here is an analogy from kernel developer Al Viro:
536
"Think of a teacher grading homework from a math student. The
537
teacher does not want to see the student's trials and errors
538
before they came up with the solution. They want to see the
539
cleanest, most elegant answer. A good student knows this, and
540
would never submit her intermediate work before the final
543
The same is true of kernel development. The maintainers and
544
reviewers do not want to see the thought process behind the
545
solution to the problem one is solving. They want to see a
546
simple and elegant solution."
548
It may be challenging to keep the balance between presenting an elegant
549
solution and working together with the community and discussing your
550
unfinished work. Therefore it is good to get early in the process to
551
get feedback to improve your work, but also keep your changes in small
552
chunks that they may get already accepted, even when your whole task is
553
not ready for inclusion now.
555
Also realize that it is not acceptable to send patches for inclusion
556
that are unfinished and will be "fixed up later."
562
Along with breaking up your patches, it is very important for you to let
563
the Linux community know why they should add this change. New features
564
must be justified as being needed and useful.
570
When sending in your patches, pay special attention to what you say in
571
the text in your email. This information will become the ChangeLog
572
information for the patch, and will be preserved for everyone to see for
573
all time. It should describe the patch completely, containing:
574
- why the change is necessary
575
- the overall design approach in the patch
576
- implementation details
579
For more details on what this should all look like, please see the
580
ChangeLog section of the document:
582
http://userweb.kernel.org/~akpm/stuff/tpp.txt
587
All of these things are sometimes very hard to do. It can take years to
588
perfect these practices (if at all). It's a continuous process of
589
improvement that requires a lot of patience and determination. But
590
don't give up, it's possible. Many have done it before, and each had to
591
start exactly where you are now.
597
Thanks to Paolo Ciarrocchi who allowed the "Development Process"
598
(http://lwn.net/Articles/94386/) section
599
to be based on text he had written, and to Randy Dunlap and Gerrit
600
Huizenga for some of the list of things you should and should not say.
601
Also thanks to Pat Mochel, Hanna Linder, Randy Dunlap, Kay Sievers,
602
Vojtech Pavlik, Jan Kara, Josh Boyer, Kees Cook, Andrew Morton, Andi
603
Kleen, Vadim Lobanov, Jesper Juhl, Adrian Bunk, Keri Harris, Frans Pop,
604
David A. Wheeler, Junio Hamano, Michael Kerrisk, and Alex Shepard for
605
their review, comments, and contributions. Without their help, this
606
document would not have been possible.
610
Maintainer: Greg Kroah-Hartman <greg@kroah.com>