1HOWTO do Linux kernel development 2--------------------------------- 3 4This is the be-all, end-all document on this topic. It contains 5instructions on how to become a Linux kernel developer and how to learn 6to work with the Linux kernel development community. It tries to not 7contain anything related to the technical aspects of kernel programming, 8but will help point you in the right direction for that. 9 10If anything in this document becomes out of date, please send in patches 11to the maintainer of this file, who is listed at the bottom of the 12document. 13 14 15Introduction 16------------ 17 18So, you want to learn how to become a Linux kernel developer? Or you 19have been told by your manager, "Go write a Linux driver for this 20device." This document's goal is to teach you everything you need to 21know to achieve this by describing the process you need to go through, 22and hints on how to work with the community. It will also try to 23explain some of the reasons why the community works like it does. 24 25The kernel is written mostly in C, with some architecture-dependent 26parts written in assembly. A good understanding of C is required for 27kernel development. Assembly (any architecture) is not required unless 28you plan to do low-level development for that architecture. Though they 29are not a good substitute for a solid C education and/or years of 30experience, 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] 34 35The kernel is written using GNU C and the GNU toolchain. While it 36adheres to the ISO C89 standard, it uses a number of extensions that are 37not featured in the standard. The kernel is a freestanding C 38environment, with no reliance on the standard C library, so some 39portions of the C standard are not supported. Arbitrary long long 40divisions and floating point are not allowed. It can sometimes be 41difficult to understand the assumptions the kernel has on the toolchain 42and the extensions that it uses, and unfortunately there is no 43definitive reference for them. Please check the gcc info pages (`info 44gcc`) for some information on them. 45 46Please remember that you are trying to learn how to work with the 47existing development community. It is a diverse group of people, with 48high standards for coding, style and procedure. These standards have 49been created over time based on what they have found to work best for 50such a large and geographically dispersed team. Try to learn as much as 51possible about these standards ahead of time, as they are well 52documented; do not expect people to adapt to you or your company's way 53of doing things. 54 55 56Legal Issues 57------------ 58 59The Linux kernel source code is released under the GPL. Please see the 60file, COPYING, in the main directory of the source tree, for details on 61the license. If you have further questions about the license, please 62contact a lawyer, and do not ask on the Linux kernel mailing list. The 63people on the mailing lists are not lawyers, and you should not rely on 64their statements on legal matters. 65 66For common questions and answers about the GPL, please see: 67 http://www.gnu.org/licenses/gpl-faq.html 68 69 70Documentation 71------------ 72 73The Linux kernel source tree has a large range of documents that are 74invaluable for learning how to interact with the kernel community. When 75new features are added to the kernel, it is recommended that new 76documentation files are also added which explain how to use the feature. 77When a kernel change causes the interface that the kernel exposes to 78userspace to change, it is recommended that you send the information or 79a patch to the manual pages explaining the change to the manual pages 80maintainer at mtk-manpages@gmx.net. 81 82Here is a list of files that are in the kernel source tree that are 83required reading: 84 README 85 This file gives a short background on the Linux kernel and describes 86 what is necessary to do to configure and build the kernel. People 87 who are new to the kernel should start here. 88 89 Documentation/Changes 90 This file gives a list of the minimum levels of various software 91 packages that are necessary to build and run the kernel 92 successfully. 93 94 Documentation/CodingStyle 95 This describes the Linux kernel coding style, and some of the 96 rationale behind it. All new code is expected to follow the 97 guidelines in this document. Most maintainers will only accept 98 patches if these rules are followed, and many people will only 99 review code if it is in the proper style. 100 101 Documentation/SubmittingPatches 102 Documentation/SubmittingDrivers 103 These files describe in explicit detail how to successfully create 104 and send a patch, including (but not limited to): 105 - Email contents 106 - Email format 107 - Who to send it to 108 Following these rules will not guarantee success (as all patches are 109 subject to scrutiny for content and style), but not following them 110 will almost always prevent it. 111 112 Other excellent descriptions of how to create patches properly are: 113 "The Perfect Patch" 114 http://www.zip.com.au/~akpm/linux/patches/stuff/tpp.txt 115 "Linux kernel patch submission format" 116 http://linux.yyz.us/patch-format.html 117 118 Documentation/stable_api_nonsense.txt 119 This file describes the rationale behind the conscious decision to 120 not have a stable API within the kernel, including things like: 121 - Subsystem shim-layers (for compatibility?) 122 - Driver portability between Operating Systems. 123 - Mitigating rapid change within the kernel source tree (or 124 preventing rapid change) 125 This document is crucial for understanding the Linux development 126 philosophy and is very important for people moving to Linux from 127 development on other Operating Systems. 128 129 Documentation/SecurityBugs 130 If you feel you have found a security problem in the Linux kernel, 131 please follow the steps in this document to help notify the kernel 132 developers, and help solve the issue. 133 134 Documentation/ManagementStyle 135 This document describes how Linux kernel maintainers operate and the 136 shared ethos behind their methodologies. This is important reading 137 for anyone new to kernel development (or anyone simply curious about 138 it), as it resolves a lot of common misconceptions and confusion 139 about the unique behavior of kernel maintainers. 140 141 Documentation/stable_kernel_rules.txt 142 This file describes the rules on how the stable kernel releases 143 happen, and what to do if you want to get a change into one of these 144 releases. 145 146 Documentation/kernel-docs.txt 147 A list of external documentation that pertains to kernel 148 development. Please consult this list if you do not find what you 149 are looking for within the in-kernel documentation. 150 151 Documentation/applying-patches.txt 152 A good introduction describing exactly what a patch is and how to 153 apply it to the different development branches of the kernel. 154 155The kernel also has a large number of documents that can be 156automatically generated from the source code itself. This includes a 157full description of the in-kernel API, and rules on how to handle 158locking properly. The documents will be created in the 159Documentation/DocBook/ directory and can be generated as PDF, 160Postscript, HTML, and man pages by running: 161 make pdfdocs 162 make psdocs 163 make htmldocs 164 make mandocs 165respectively from the main kernel source directory. 166 167 168Becoming A Kernel Developer 169--------------------------- 170 171If you do not know anything about Linux kernel development, you should 172look at the Linux KernelNewbies project: 173 http://kernelnewbies.org 174It consists of a helpful mailing list where you can ask almost any type 175of basic kernel development question (make sure to search the archives 176first, before asking something that has already been answered in the 177past.) It also has an IRC channel that you can use to ask questions in 178real-time, and a lot of helpful documentation that is useful for 179learning about Linux kernel development. 180 181The website has basic information about code organization, subsystems, 182and current projects (both in-tree and out-of-tree). It also describes 183some basic logistical information, like how to compile a kernel and 184apply a patch. 185 186If you do not know where you want to start, but you want to look for 187some task to start doing to join into the kernel development community, 188go to the Linux Kernel Janitor's project: 189 http://janitor.kernelnewbies.org/ 190It is a great place to start. It describes a list of relatively simple 191problems that need to be cleaned up and fixed within the Linux kernel 192source tree. Working with the developers in charge of this project, you 193will learn the basics of getting your patch into the Linux kernel tree, 194and possibly be pointed in the direction of what to go work on next, if 195you do not already have an idea. 196 197If you already have a chunk of code that you want to put into the kernel 198tree, but need some help getting it in the proper form, the 199kernel-mentors project was created to help you out with this. It is a 200mailing list, and can be found at: 201 http://selenic.com/mailman/listinfo/kernel-mentors 202 203Before making any actual modifications to the Linux kernel code, it is 204imperative to understand how the code in question works. For this 205purpose, nothing is better than reading through it directly (most tricky 206bits are commented well), perhaps even with the help of specialized 207tools. One such tool that is particularly recommended is the Linux 208Cross-Reference project, which is able to present source code in a 209self-referential, indexed webpage format. An excellent up-to-date 210repository of the kernel code may be found at: 211 http://sosdg.org/~coywolf/lxr/ 212 213 214The development process 215----------------------- 216 217Linux kernel development process currently consists of a few different 218main kernel "branches" and lots of different subsystem-specific kernel 219branches. These different branches are: 220 - main 2.6.x kernel tree 221 - 2.6.x.y -stable kernel tree 222 - 2.6.x -git kernel patches 223 - 2.6.x -mm kernel patches 224 - subsystem specific kernel trees and patches 225 2262.6.x kernel tree 227----------------- 2282.6.x kernels are maintained by Linus Torvalds, and can be found on 229kernel.org in the pub/linux/kernel/v2.6/ directory. Its development 230process is as follows: 231 - As soon as a new kernel is released a two weeks window is open, 232 during this period of time maintainers can submit big diffs to 233 Linus, usually the patches that have already been included in the 234 -mm kernel for a few weeks. The preferred way to submit big changes 235 is using git (the kernel's source management tool, more information 236 can be found at http://git.or.cz/) but plain patches are also just 237 fine. 238 - After two weeks a -rc1 kernel is released it is now possible to push 239 only patches that do not include new features that could affect the 240 stability of the whole kernel. Please note that a whole new driver 241 (or filesystem) might be accepted after -rc1 because there is no 242 risk of causing regressions with such a change as long as the change 243 is self-contained and does not affect areas outside of the code that 244 is being added. git can be used to send patches to Linus after -rc1 245 is released, but the patches need to also be sent to a public 246 mailing list for review. 247 - A new -rc is released whenever Linus deems the current git tree to 248 be in a reasonably sane state adequate for testing. The goal is to 249 release a new -rc kernel every week. 250 - Process continues until the kernel is considered "ready", the 251 process should last around 6 weeks. 252 253It is worth mentioning what Andrew Morton wrote on the linux-kernel 254mailing list about kernel releases: 255 "Nobody knows when a kernel will be released, because it's 256 released according to perceived bug status, not according to a 257 preconceived timeline." 258 2592.6.x.y -stable kernel tree 260--------------------------- 261Kernels with 4 digit versions are -stable kernels. They contain 262relatively small and critical fixes for security problems or significant 263regressions discovered in a given 2.6.x kernel. 264 265This is the recommended branch for users who want the most recent stable 266kernel and are not interested in helping test development/experimental 267versions. 268 269If no 2.6.x.y kernel is available, then the highest numbered 2.6.x 270kernel is the current stable kernel. 271 2722.6.x.y are maintained by the "stable" team <stable@kernel.org>, and are 273released almost every other week. 274 275The file Documentation/stable_kernel_rules.txt in the kernel tree 276documents what kinds of changes are acceptable for the -stable tree, and 277how the release process works. 278 2792.6.x -git patches 280------------------ 281These are daily snapshots of Linus' kernel tree which are managed in a 282git repository (hence the name.) These patches are usually released 283daily and represent the current state of Linus' tree. They are more 284experimental than -rc kernels since they are generated automatically 285without even a cursory glance to see if they are sane. 286 2872.6.x -mm kernel patches 288------------------------ 289These are experimental kernel patches released by Andrew Morton. Andrew 290takes all of the different subsystem kernel trees and patches and mushes 291them together, along with a lot of patches that have been plucked from 292the linux-kernel mailing list. This tree serves as a proving ground for 293new features and patches. Once a patch has proved its worth in -mm for 294a while Andrew or the subsystem maintainer pushes it on to Linus for 295inclusion in mainline. 296 297It is heavily encouraged that all new patches get tested in the -mm tree 298before they are sent to Linus for inclusion in the main kernel tree. 299 300These kernels are not appropriate for use on systems that are supposed 301to be stable and they are more risky to run than any of the other 302branches. 303 304If you wish to help out with the kernel development process, please test 305and use these kernel releases and provide feedback to the linux-kernel 306mailing list if you have any problems, and if everything works properly. 307 308In addition to all the other experimental patches, these kernels usually 309also contain any changes in the mainline -git kernels available at the 310time of release. 311 312The -mm kernels are not released on a fixed schedule, but usually a few 313-mm kernels are released in between each -rc kernel (1 to 3 is common). 314 315Subsystem Specific kernel trees and patches 316------------------------------------------- 317A number of the different kernel subsystem developers expose their 318development trees so that others can see what is happening in the 319different areas of the kernel. These trees are pulled into the -mm 320kernel releases as described above. 321 322Here is a list of some of the different kernel trees available: 323 git trees: 324 - Kbuild development tree, Sam Ravnborg <sam@ravnborg.org> 325 git.kernel.org:/pub/scm/linux/kernel/git/sam/kbuild.git 326 327 - ACPI development tree, Len Brown <len.brown@intel.com> 328 git.kernel.org:/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6.git 329 330 - Block development tree, Jens Axboe <axboe@suse.de> 331 git.kernel.org:/pub/scm/linux/kernel/git/axboe/linux-2.6-block.git 332 333 - DRM development tree, Dave Airlie <airlied@linux.ie> 334 git.kernel.org:/pub/scm/linux/kernel/git/airlied/drm-2.6.git 335 336 - ia64 development tree, Tony Luck <tony.luck@intel.com> 337 git.kernel.org:/pub/scm/linux/kernel/git/aegl/linux-2.6.git 338 339 - infiniband, Roland Dreier <rolandd@cisco.com> 340 git.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband.git 341 342 - libata, Jeff Garzik <jgarzik@pobox.com> 343 git.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev.git 344 345 - network drivers, Jeff Garzik <jgarzik@pobox.com> 346 git.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6.git 347 348 - pcmcia, Dominik Brodowski <linux@dominikbrodowski.net> 349 git.kernel.org:/pub/scm/linux/kernel/git/brodo/pcmcia-2.6.git 350 351 - SCSI, James Bottomley <James.Bottomley@SteelEye.com> 352 git.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6.git 353 354 quilt trees: 355 - USB, PCI, Driver Core, and I2C, Greg Kroah-Hartman <gregkh@suse.de> 356 kernel.org/pub/linux/kernel/people/gregkh/gregkh-2.6/ 357 - x86-64, partly i386, Andi Kleen <ak@suse.de> 358 ftp.firstfloor.org:/pub/ak/x86_64/quilt/ 359 360 Other kernel trees can be found listed at http://git.kernel.org/ and in 361 the MAINTAINERS file. 362 363Bug Reporting 364------------- 365 366bugzilla.kernel.org is where the Linux kernel developers track kernel 367bugs. Users are encouraged to report all bugs that they find in this 368tool. For details on how to use the kernel bugzilla, please see: 369 http://test.kernel.org/bugzilla/faq.html 370 371The file REPORTING-BUGS in the main kernel source directory has a good 372template for how to report a possible kernel bug, and details what kind 373of information is needed by the kernel developers to help track down the 374problem. 375 376 377Managing bug reports 378-------------------- 379 380One of the best ways to put into practice your hacking skills is by fixing 381bugs reported by other people. Not only you will help to make the kernel 382more stable, you'll learn to fix real world problems and you will improve 383your skills, and other developers will be aware of your presence. Fixing 384bugs is one of the best ways to earn merit amongst the developers, because 385not many people like wasting time fixing other people's bugs. 386 387To work in the already reported bug reports, go to http://bugzilla.kernel.org. 388If you want to be advised of the future bug reports, you can subscribe to the 389bugme-new mailing list (only new bug reports are mailed here) or to the 390bugme-janitor mailing list (every change in the bugzilla is mailed here) 391 392 http://lists.osdl.org/mailman/listinfo/bugme-new 393 http://lists.osdl.org/mailman/listinfo/bugme-janitors 394 395 396 397Mailing lists 398------------- 399 400As some of the above documents describe, the majority of the core kernel 401developers participate on the Linux Kernel Mailing list. Details on how 402to subscribe and unsubscribe from the list can be found at: 403 http://vger.kernel.org/vger-lists.html#linux-kernel 404There are archives of the mailing list on the web in many different 405places. Use a search engine to find these archives. For example: 406 http://dir.gmane.org/gmane.linux.kernel 407It is highly recommended that you search the archives about the topic 408you want to bring up, before you post it to the list. A lot of things 409already discussed in detail are only recorded at the mailing list 410archives. 411 412Most of the individual kernel subsystems also have their own separate 413mailing list where they do their development efforts. See the 414MAINTAINERS file for a list of what these lists are for the different 415groups. 416 417Many of the lists are hosted on kernel.org. Information on them can be 418found at: 419 http://vger.kernel.org/vger-lists.html 420 421Please remember to follow good behavioral habits when using the lists. 422Though a bit cheesy, the following URL has some simple guidelines for 423interacting with the list (or any list): 424 http://www.albion.com/netiquette/ 425 426If multiple people respond to your mail, the CC: list of recipients may 427get pretty large. Don't remove anybody from the CC: list without a good 428reason, or don't reply only to the list address. Get used to receiving the 429mail twice, one from the sender and the one from the list, and don't try 430to tune that by adding fancy mail-headers, people will not like it. 431 432Remember to keep the context and the attribution of your replies intact, 433keep the "John Kernelhacker wrote ...:" lines at the top of your reply, and 434add your statements between the individual quoted sections instead of 435writing at the top of the mail. 436 437If you add patches to your mail, make sure they are plain readable text 438as stated in Documentation/SubmittingPatches. Kernel developers don't 439want to deal with attachments or compressed patches; they may want 440to comment on individual lines of your patch, which works only that way. 441Make sure you use a mail program that does not mangle spaces and tab 442characters. A good first test is to send the mail to yourself and try 443to apply your own patch by yourself. If that doesn't work, get your 444mail program fixed or change it until it works. 445 446Above all, please remember to show respect to other subscribers. 447 448 449Working with the community 450-------------------------- 451 452The goal of the kernel community is to provide the best possible kernel 453there is. When you submit a patch for acceptance, it will be reviewed 454on its technical merits and those alone. So, what should you be 455expecting? 456 - criticism 457 - comments 458 - requests for change 459 - requests for justification 460 - silence 461 462Remember, this is part of getting your patch into the kernel. You have 463to be able to take criticism and comments about your patches, evaluate 464them at a technical level and either rework your patches or provide 465clear and concise reasoning as to why those changes should not be made. 466If there are no responses to your posting, wait a few days and try 467again, sometimes things get lost in the huge volume. 468 469What should you not do? 470 - expect your patch to be accepted without question 471 - become defensive 472 - ignore comments 473 - resubmit the patch without making any of the requested changes 474 475In a community that is looking for the best technical solution possible, 476there will always be differing opinions on how beneficial a patch is. 477You have to be cooperative, and willing to adapt your idea to fit within 478the kernel. Or at least be willing to prove your idea is worth it. 479Remember, being wrong is acceptable as long as you are willing to work 480toward a solution that is right. 481 482It is normal that the answers to your first patch might simply be a list 483of a dozen things you should correct. This does _not_ imply that your 484patch will not be accepted, and it is _not_ meant against you 485personally. Simply correct all issues raised against your patch and 486resend it. 487 488 489Differences between the kernel community and corporate structures 490----------------------------------------------------------------- 491 492The kernel community works differently than most traditional corporate 493development environments. Here are a list of things that you can try to 494do to try to avoid problems: 495 Good things to say regarding your proposed changes: 496 - "This solves multiple problems." 497 - "This deletes 2000 lines of code." 498 - "Here is a patch that explains what I am trying to describe." 499 - "I tested it on 5 different architectures..." 500 - "Here is a series of small patches that..." 501 - "This increases performance on typical machines..." 502 503 Bad things you should avoid saying: 504 - "We did it this way in AIX/ptx/Solaris, so therefore it must be 505 good..." 506 - "I've being doing this for 20 years, so..." 507 - "This is required for my company to make money" 508 - "This is for our Enterprise product line." 509 - "Here is my 1000 page design document that describes my idea" 510 - "I've been working on this for 6 months..." 511 - "Here's a 5000 line patch that..." 512 - "I rewrote all of the current mess, and here it is..." 513 - "I have a deadline, and this patch needs to be applied now." 514 515Another way the kernel community is different than most traditional 516software engineering work environments is the faceless nature of 517interaction. One benefit of using email and irc as the primary forms of 518communication is the lack of discrimination based on gender or race. 519The Linux kernel work environment is accepting of women and minorities 520because all you are is an email address. The international aspect also 521helps to level the playing field because you can't guess gender based on 522a person's name. A man may be named Andrea and a woman may be named Pat. 523Most women who have worked in the Linux kernel and have expressed an 524opinion have had positive experiences. 525 526The language barrier can cause problems for some people who are not 527comfortable with English. A good grasp of the language can be needed in 528order to get ideas across properly on mailing lists, so it is 529recommended that you check your emails to make sure they make sense in 530English before sending them. 531 532 533Break up your changes 534--------------------- 535 536The Linux kernel community does not gladly accept large chunks of code 537dropped on it all at once. The changes need to be properly introduced, 538discussed, and broken up into tiny, individual portions. This is almost 539the exact opposite of what companies are used to doing. Your proposal 540should also be introduced very early in the development process, so that 541you can receive feedback on what you are doing. It also lets the 542community feel that you are working with them, and not simply using them 543as a dumping ground for your feature. However, don't send 50 emails at 544one time to a mailing list, your patch series should be smaller than 545that almost all of the time. 546 547The reasons for breaking things up are the following: 548 5491) Small patches increase the likelihood that your patches will be 550 applied, since they don't take much time or effort to verify for 551 correctness. A 5 line patch can be applied by a maintainer with 552 barely a second glance. However, a 500 line patch may take hours to 553 review for correctness (the time it takes is exponentially 554 proportional to the size of the patch, or something). 555 556 Small patches also make it very easy to debug when something goes 557 wrong. It's much easier to back out patches one by one than it is 558 to dissect a very large patch after it's been applied (and broken 559 something). 560 5612) It's important not only to send small patches, but also to rewrite 562 and simplify (or simply re-order) patches before submitting them. 563 564Here is an analogy from kernel developer Al Viro: 565 "Think of a teacher grading homework from a math student. The 566 teacher does not want to see the student's trials and errors 567 before they came up with the solution. They want to see the 568 cleanest, most elegant answer. A good student knows this, and 569 would never submit her intermediate work before the final 570 solution." 571 572 The same is true of kernel development. The maintainers and 573 reviewers do not want to see the thought process behind the 574 solution to the problem one is solving. They want to see a 575 simple and elegant solution." 576 577It may be challenging to keep the balance between presenting an elegant 578solution and working together with the community and discussing your 579unfinished work. Therefore it is good to get early in the process to 580get feedback to improve your work, but also keep your changes in small 581chunks that they may get already accepted, even when your whole task is 582not ready for inclusion now. 583 584Also realize that it is not acceptable to send patches for inclusion 585that are unfinished and will be "fixed up later." 586 587 588Justify your change 589------------------- 590 591Along with breaking up your patches, it is very important for you to let 592the Linux community know why they should add this change. New features 593must be justified as being needed and useful. 594 595 596Document your change 597-------------------- 598 599When sending in your patches, pay special attention to what you say in 600the text in your email. This information will become the ChangeLog 601information for the patch, and will be preserved for everyone to see for 602all time. It should describe the patch completely, containing: 603 - why the change is necessary 604 - the overall design approach in the patch 605 - implementation details 606 - testing results 607 608For more details on what this should all look like, please see the 609ChangeLog section of the document: 610 "The Perfect Patch" 611 http://www.zip.com.au/~akpm/linux/patches/stuff/tpp.txt 612 613 614 615 616All of these things are sometimes very hard to do. It can take years to 617perfect these practices (if at all). It's a continuous process of 618improvement that requires a lot of patience and determination. But 619don't give up, it's possible. Many have done it before, and each had to 620start exactly where you are now. 621 622 623 624 625---------- 626Thanks to Paolo Ciarrocchi who allowed the "Development Process" 627(http://linux.tar.bz/articles/2.6-development_process) section 628to be based on text he had written, and to Randy Dunlap and Gerrit 629Huizenga for some of the list of things you should and should not say. 630Also thanks to Pat Mochel, Hanna Linder, Randy Dunlap, Kay Sievers, 631Vojtech Pavlik, Jan Kara, Josh Boyer, Kees Cook, Andrew Morton, Andi 632Kleen, Vadim Lobanov, Jesper Juhl, Adrian Bunk, Keri Harris, Frans Pop, 633David A. Wheeler, Junio Hamano, Michael Kerrisk, and Alex Shepard for 634their review, comments, and contributions. Without their help, this 635document would not have been possible. 636 637 638 639Maintainer: Greg Kroah-Hartman <greg@kroah.com> 640