1 .. _process_howto: 2 3 HOWTO do Linux kernel development 4 ================================= 5 6 This is the be-all, end-all document on this topic. It contains 7 instructions on how to become a Linux kernel developer and how to learn 8 to work with the Linux kernel development community. It tries to not 9 contain anything related to the technical aspects of kernel programming, 10 but will help point you in the right direction for that. 11 12 If anything in this document becomes out of date, please send in patches 13 to the maintainer of this file, who is listed at the bottom of the 14 document. 15 16 17 Introduction 18 ------------ 19 20 So, you want to learn how to become a Linux kernel developer? Or you 21 have been told by your manager, "Go write a Linux driver for this 22 device." This document's goal is to teach you everything you need to 23 know to achieve this by describing the process you need to go through, 24 and hints on how to work with the community. It will also try to 25 explain some of the reasons why the community works like it does. 26 27 The kernel is written mostly in C, with some architecture-dependent 28 parts written in assembly. A good understanding of C is required for 29 kernel development. Assembly (any architecture) is not required unless 30 you plan to do low-level development for that architecture. Though they 31 are not a good substitute for a solid C education and/or years of 32 experience, the following books are good for, if anything, reference: 33 34 - "The C Programming Language" by Kernighan and Ritchie [Prentice Hall] 35 - "Practical C Programming" by Steve Oualline [O'Reilly] 36 - "C: A Reference Manual" by Harbison and Steele [Prentice Hall] 37 38 The kernel is written using GNU C and the GNU toolchain. While it 39 adheres to the ISO C11 standard, it uses a number of extensions that are 40 not featured in the standard. The kernel is a freestanding C 41 environment, with no reliance on the standard C library, so some 42 portions of the C standard are not supported. Arbitrary long long 43 divisions and floating point are not allowed. It can sometimes be 44 difficult to understand the assumptions the kernel has on the toolchain 45 and the extensions that it uses, and unfortunately there is no 46 definitive reference for them. Please check the gcc info pages (`info 47 gcc`) for some information on them. 48 49 Please remember that you are trying to learn how to work with the 50 existing development community. It is a diverse group of people, with 51 high standards for coding, style and procedure. These standards have 52 been created over time based on what they have found to work best for 53 such a large and geographically dispersed team. Try to learn as much as 54 possible about these standards ahead of time, as they are well 55 documented; do not expect people to adapt to you or your company's way 56 of doing things. 57 58 59 Legal Issues 60 ------------ 61 62 The Linux kernel source code is released under the GPL. Please see the file 63 COPYING in the main directory of the source tree. The Linux kernel licensing 64 rules and how to use `SPDX <https://spdx.org/>`_ identifiers in source code are 65 described in :ref:`Documentation/process/license-rules.rst <kernel_licensing>`. 66 If you have further questions about the license, please contact a lawyer, and do 67 not ask on the Linux kernel mailing list. The people on the mailing lists are 68 not lawyers, and you should not rely on their statements on legal matters. 69 70 For common questions and answers about the GPL, please see: 71 72 https://www.gnu.org/licenses/gpl-faq.html 73 74 75 Documentation 76 ------------- 77 78 The Linux kernel source tree has a large range of documents that are 79 invaluable for learning how to interact with the kernel community. When 80 new features are added to the kernel, it is recommended that new 81 documentation files are also added which explain how to use the feature. 82 When a kernel change causes the interface that the kernel exposes to 83 userspace to change, it is recommended that you send the information or 84 a patch to the manual pages explaining the change to the manual pages 85 maintainer at alx@kernel.org, and CC the list linux-api@vger.kernel.org. 86 87 Here is a list of files that are in the kernel source tree that are 88 required reading: 89 90 :ref:`Documentation/admin-guide/README.rst <readme>` 91 This file gives a short background on the Linux kernel and describes 92 what is necessary to do to configure and build the kernel. People 93 who are new to the kernel should start here. 94 95 :ref:`Documentation/process/changes.rst <changes>` 96 This file gives a list of the minimum levels of various software 97 packages that are necessary to build and run the kernel 98 successfully. 99 100 :ref:`Documentation/process/coding-style.rst <codingstyle>` 101 This describes the Linux kernel coding style, and some of the 102 rationale behind it. All new code is expected to follow the 103 guidelines in this document. Most maintainers will only accept 104 patches if these rules are followed, and many people will only 105 review code if it is in the proper style. 106 107 :ref:`Documentation/process/submitting-patches.rst <submittingpatches>` 108 This file describes in explicit detail how to successfully create 109 and send a patch, including (but not limited to): 110 111 - Email contents 112 - Email format 113 - Who to send it to 114 115 Following these rules will not guarantee success (as all patches are 116 subject to scrutiny for content and style), but not following them 117 will almost always prevent it. 118 119 Other excellent descriptions of how to create patches properly are: 120 121 "The Perfect Patch" 122 https://www.ozlabs.org/~akpm/stuff/tpp.txt 123 124 "Linux kernel patch submission format" 125 https://web.archive.org/web/20180829112450/http://linux.yyz.us/patch-format.html 126 127 :ref:`Documentation/process/stable-api-nonsense.rst <stable_api_nonsense>` 128 This file describes the rationale behind the conscious decision to 129 not have a stable API within the kernel, including things like: 130 131 - Subsystem shim-layers (for compatibility?) 132 - Driver portability between Operating Systems. 133 - Mitigating rapid change within the kernel source tree (or 134 preventing rapid change) 135 136 This document is crucial for understanding the Linux development 137 philosophy and is very important for people moving to Linux from 138 development on other Operating Systems. 139 140 :ref:`Documentation/process/security-bugs.rst <securitybugs>` 141 If you feel you have found a security problem in the Linux kernel, 142 please follow the steps in this document to help notify the kernel 143 developers, and help solve the issue. 144 145 :ref:`Documentation/process/management-style.rst <managementstyle>` 146 This document describes how Linux kernel maintainers operate and the 147 shared ethos behind their methodologies. This is important reading 148 for anyone new to kernel development (or anyone simply curious about 149 it), as it resolves a lot of common misconceptions and confusion 150 about the unique behavior of kernel maintainers. 151 152 :ref:`Documentation/process/stable-kernel-rules.rst <stable_kernel_rules>` 153 This file describes the rules on how the stable kernel releases 154 happen, and what to do if you want to get a change into one of these 155 releases. 156 157 :ref:`Documentation/process/kernel-docs.rst <kernel_docs>` 158 A list of external documentation that pertains to kernel 159 development. Please consult this list if you do not find what you 160 are looking for within the in-kernel documentation. 161 162 :ref:`Documentation/process/applying-patches.rst <applying_patches>` 163 A good introduction describing exactly what a patch is and how to 164 apply it to the different development branches of the kernel. 165 166 The kernel also has a large number of documents that can be 167 automatically generated from the source code itself or from 168 ReStructuredText markups (ReST), like this one. This includes a 169 full description of the in-kernel API, and rules on how to handle 170 locking properly. 171 172 All such documents can be generated as PDF or HTML by running:: 173 174 make pdfdocs 175 make htmldocs 176 177 respectively from the main kernel source directory. 178 179 The documents that uses ReST markup will be generated at Documentation/output. 180 They can also be generated on LaTeX and ePub formats with:: 181 182 make latexdocs 183 make epubdocs 184 185 Becoming A Kernel Developer 186 --------------------------- 187 188 If you do not know anything about Linux kernel development, you should 189 look at the Linux KernelNewbies project: 190 191 https://kernelnewbies.org 192 193 It consists of a helpful mailing list where you can ask almost any type 194 of basic kernel development question (make sure to search the archives 195 first, before asking something that has already been answered in the 196 past.) It also has an IRC channel that you can use to ask questions in 197 real-time, and a lot of helpful documentation that is useful for 198 learning about Linux kernel development. 199 200 The website has basic information about code organization, subsystems, 201 and current projects (both in-tree and out-of-tree). It also describes 202 some basic logistical information, like how to compile a kernel and 203 apply a patch. 204 205 If you do not know where you want to start, but you want to look for 206 some task to start doing to join into the kernel development community, 207 go to the Linux Kernel Janitor's project: 208 209 https://kernelnewbies.org/KernelJanitors 210 211 It is a great place to start. It describes a list of relatively simple 212 problems that need to be cleaned up and fixed within the Linux kernel 213 source tree. Working with the developers in charge of this project, you 214 will learn the basics of getting your patch into the Linux kernel tree, 215 and possibly be pointed in the direction of what to go work on next, if 216 you do not already have an idea. 217 218 Before making any actual modifications to the Linux kernel code, it is 219 imperative to understand how the code in question works. For this 220 purpose, nothing is better than reading through it directly (most tricky 221 bits are commented well), perhaps even with the help of specialized 222 tools. One such tool that is particularly recommended is the Linux 223 Cross-Reference project, which is able to present source code in a 224 self-referential, indexed webpage format. An excellent up-to-date 225 repository of the kernel code may be found at: 226 227 https://elixir.bootlin.com/ 228 229 230 The development process 231 ----------------------- 232 233 Linux kernel development process currently consists of a few different 234 main kernel "branches" and lots of different subsystem-specific kernel 235 branches. These different branches are: 236 237 - Linus's mainline tree 238 - Various stable trees with multiple major numbers 239 - Subsystem-specific trees 240 - linux-next integration testing tree 241 242 Mainline tree 243 ~~~~~~~~~~~~~ 244 245 The mainline tree is maintained by Linus Torvalds, and can be found at 246 https://kernel.org or in the repo. Its development process is as follows: 247 248 - As soon as a new kernel is released a two week window is open, 249 during this period of time maintainers can submit big diffs to 250 Linus, usually the patches that have already been included in the 251 linux-next for a few weeks. The preferred way to submit big changes 252 is using git (the kernel's source management tool, more information 253 can be found at https://git-scm.com/) but plain patches are also just 254 fine. 255 - After two weeks a -rc1 kernel is released and the focus is on making the 256 new kernel as rock solid as possible. Most of the patches at this point 257 should fix a regression. Bugs that have always existed are not 258 regressions, so only push these kinds of fixes if they are important. 259 Please note that a whole new driver (or filesystem) might be accepted 260 after -rc1 because there is no risk of causing regressions with such a 261 change as long as the change is self-contained and does not affect areas 262 outside of the code that is being added. git can be used to send 263 patches to Linus after -rc1 is released, but the patches need to also be 264 sent to a public mailing list for review. 265 - A new -rc is released whenever Linus deems the current git tree to 266 be in a reasonably sane state adequate for testing. The goal is to 267 release a new -rc kernel every week. 268 - Process continues until the kernel is considered "ready", the 269 process should last around 6 weeks. 270 271 It is worth mentioning what Andrew Morton wrote on the linux-kernel 272 mailing list about kernel releases: 273 274 *"Nobody knows when a kernel will be released, because it's 275 released according to perceived bug status, not according to a 276 preconceived timeline."* 277 278 Various stable trees with multiple major numbers 279 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 280 281 Kernels with 3-part versions are -stable kernels. They contain 282 relatively small and critical fixes for security problems or significant 283 regressions discovered in a given major mainline release. Each release 284 in a major stable series increments the third part of the version 285 number, keeping the first two parts the same. 286 287 This is the recommended branch for users who want the most recent stable 288 kernel and are not interested in helping test development/experimental 289 versions. 290 291 Stable trees are maintained by the "stable" team <stable@vger.kernel.org>, and 292 are released as needs dictate. The normal release period is approximately 293 two weeks, but it can be longer if there are no pressing problems. A 294 security-related problem, instead, can cause a release to happen almost 295 instantly. 296 297 The file :ref:`Documentation/process/stable-kernel-rules.rst <stable_kernel_rules>` 298 in the kernel tree documents what kinds of changes are acceptable for 299 the -stable tree, and how the release process works. 300 301 Subsystem-specific trees 302 ~~~~~~~~~~~~~~~~~~~~~~~~ 303 304 The maintainers of the various kernel subsystems --- and also many 305 kernel subsystem developers --- expose their current state of 306 development in source repositories. That way, others can see what is 307 happening in the different areas of the kernel. In areas where 308 development is rapid, a developer may be asked to base his submissions 309 onto such a subsystem kernel tree so that conflicts between the 310 submission and other already ongoing work are avoided. 311 312 Most of these repositories are git trees, but there are also other SCMs 313 in use, or patch queues being published as quilt series. Addresses of 314 these subsystem repositories are listed in the MAINTAINERS file. Many 315 of them can be browsed at https://git.kernel.org/. 316 317 Before a proposed patch is committed to such a subsystem tree, it is 318 subject to review which primarily happens on mailing lists (see the 319 respective section below). For several kernel subsystems, this review 320 process is tracked with the tool patchwork. Patchwork offers a web 321 interface which shows patch postings, any comments on a patch or 322 revisions to it, and maintainers can mark patches as under review, 323 accepted, or rejected. Most of these patchwork sites are listed at 324 https://patchwork.kernel.org/. 325 326 linux-next integration testing tree 327 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 328 329 Before updates from subsystem trees are merged into the mainline tree, 330 they need to be integration-tested. For this purpose, a special 331 testing repository exists into which virtually all subsystem trees are 332 pulled on an almost daily basis: 333 334 https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git 335 336 This way, the linux-next gives a summary outlook onto what will be 337 expected to go into the mainline kernel at the next merge period. 338 Adventurous testers are very welcome to runtime-test the linux-next. 339 340 341 Bug Reporting 342 ------------- 343 344 The file 'Documentation/admin-guide/reporting-issues.rst' in the main kernel 345 source directory describes how to report a possible kernel bug, and details 346 what kind of information is needed by the kernel developers to help track 347 down the problem. 348 349 350 Managing bug reports 351 -------------------- 352 353 One of the best ways to put into practice your hacking skills is by fixing 354 bugs reported by other people. Not only will you help to make the kernel 355 more stable, but you'll also learn to fix real-world problems and you will 356 improve your skills, and other developers will be aware of your presence. 357 Fixing bugs is one of the best ways to get merits among other developers, 358 because not many people like wasting time fixing other people's bugs. 359 360 To work on already reported bug reports, find a subsystem you are interested in. 361 Check the MAINTAINERS file where bugs for that subsystem get reported to; often 362 it will be a mailing list, rarely a bugtracker. Search the archives of said 363 place for recent reports and help where you see fit. You may also want to check 364 https://bugzilla.kernel.org for bug reports; only a handful of kernel subsystems 365 use it actively for reporting or tracking, nevertheless bugs for the whole 366 kernel get filed there. 367 368 369 Mailing lists 370 ------------- 371 372 As some of the above documents describe, the majority of the core kernel 373 developers participate on the Linux Kernel Mailing list. Details on how 374 to subscribe and unsubscribe from the list can be found at: 375 376 https://subspace.kernel.org/subscribing.html 377 378 There are archives of the mailing list on the web in many different 379 places. Use a search engine to find these archives. For example: 380 381 https://lore.kernel.org/linux-kernel/ 382 383 It is highly recommended that you search the archives about the topic 384 you want to bring up, before you post it to the list. A lot of things 385 already discussed in detail are only recorded at the mailing list 386 archives. 387 388 Most of the individual kernel subsystems also have their own separate 389 mailing list where they do their development efforts. See the 390 MAINTAINERS file for a list of what these lists are for the different 391 groups. 392 393 Many of the lists are hosted on kernel.org. Information on them can be 394 found at: 395 396 https://subspace.kernel.org 397 398 Please remember to follow good behavioral habits when using the lists. 399 Though a bit cheesy, the following URL has some simple guidelines for 400 interacting with the list (or any list): 401 402 https://subspace.kernel.org/etiquette.html 403 404 If multiple people respond to your mail, the CC: list of recipients may 405 get pretty large. Don't remove anybody from the CC: list without a good 406 reason, or don't reply only to the list address. Get used to receiving the 407 mail twice, one from the sender and the one from the list, and don't try 408 to tune that by adding fancy mail-headers, people will not like it. 409 410 Remember to keep the context and the attribution of your replies intact, 411 keep the "John Kernelhacker wrote ...:" lines at the top of your reply, and 412 add your statements between the individual quoted sections instead of 413 writing at the top of the mail. 414 415 If you add patches to your mail, make sure they are plain readable text 416 as stated in :ref:`Documentation/process/submitting-patches.rst <submittingpatches>`. 417 Kernel developers don't want to deal with 418 attachments or compressed patches; they may want to comment on 419 individual lines of your patch, which works only that way. Make sure you 420 use a mail program that does not mangle spaces and tab characters. A 421 good first test is to send the mail to yourself and try to apply your 422 own patch by yourself. If that doesn't work, get your mail program fixed 423 or change it until it works. 424 425 Above all, please remember to show respect to other subscribers. 426 427 428 Working with the community 429 -------------------------- 430 431 The goal of the kernel community is to provide the best possible kernel 432 there is. When you submit a patch for acceptance, it will be reviewed 433 on its technical merits and those alone. So, what should you be 434 expecting? 435 436 - criticism 437 - comments 438 - requests for change 439 - requests for justification 440 - silence 441 442 Remember, this is part of getting your patch into the kernel. You have 443 to be able to take criticism and comments about your patches, evaluate 444 them at a technical level and either rework your patches or provide 445 clear and concise reasoning as to why those changes should not be made. 446 If there are no responses to your posting, wait a few days and try 447 again, sometimes things get lost in the huge volume. 448 449 What should you not do? 450 451 - expect your patch to be accepted without question 452 - become defensive 453 - ignore comments 454 - resubmit the patch without making any of the requested changes 455 456 In a community that is looking for the best technical solution possible, 457 there will always be differing opinions on how beneficial a patch is. 458 You have to be cooperative, and willing to adapt your idea to fit within 459 the kernel. Or at least be willing to prove your idea is worth it. 460 Remember, being wrong is acceptable as long as you are willing to work 461 toward a solution that is right. 462 463 It is normal that the answers to your first patch might simply be a list 464 of a dozen things you should correct. This does **not** imply that your 465 patch will not be accepted, and it is **not** meant against you 466 personally. Simply correct all issues raised against your patch and 467 resend it. 468 469 470 Differences between the kernel community and corporate structures 471 ----------------------------------------------------------------- 472 473 The kernel community works differently than most traditional corporate 474 development environments. Here are a list of things that you can try to 475 do to avoid problems: 476 477 Good things to say regarding your proposed changes: 478 479 - "This solves multiple problems." 480 - "This deletes 2000 lines of code." 481 - "Here is a patch that explains what I am trying to describe." 482 - "I tested it on 5 different architectures..." 483 - "Here is a series of small patches that..." 484 - "This increases performance on typical machines..." 485 486 Bad things you should avoid saying: 487 488 - "We did it this way in AIX/ptx/Solaris, so therefore it must be 489 good..." 490 - "I've being doing this for 20 years, so..." 491 - "This is required for my company to make money" 492 - "This is for our Enterprise product line." 493 - "Here is my 1000 page design document that describes my idea" 494 - "I've been working on this for 6 months..." 495 - "Here's a 5000 line patch that..." 496 - "I rewrote all of the current mess, and here it is..." 497 - "I have a deadline, and this patch needs to be applied now." 498 499 Another way the kernel community is different than most traditional 500 software engineering work environments is the faceless nature of 501 interaction. One benefit of using email and irc as the primary forms of 502 communication is the lack of discrimination based on gender or race. 503 The Linux kernel work environment is accepting of women and minorities 504 because all you are is an email address. The international aspect also 505 helps to level the playing field because you can't guess gender based on 506 a person's name. A man may be named Andrea and a woman may be named Pat. 507 Most women who have worked in the Linux kernel and have expressed an 508 opinion have had positive experiences. 509 510 The language barrier can cause problems for some people who are not 511 comfortable with English. A good grasp of the language can be needed in 512 order to get ideas across properly on mailing lists, so it is 513 recommended that you check your emails to make sure they make sense in 514 English before sending them. 515 516 517 Break up your changes 518 --------------------- 519 520 The Linux kernel community does not gladly accept large chunks of code 521 dropped on it all at once. The changes need to be properly introduced, 522 discussed, and broken up into tiny, individual portions. This is almost 523 the exact opposite of what companies are used to doing. Your proposal 524 should also be introduced very early in the development process, so that 525 you can receive feedback on what you are doing. It also lets the 526 community feel that you are working with them, and not simply using them 527 as a dumping ground for your feature. However, don't send 50 emails at 528 one time to a mailing list, your patch series should be smaller than 529 that almost all of the time. 530 531 The reasons for breaking things up are the following: 532 533 1) Small patches increase the likelihood that your patches will be 534 applied, since they don't take much time or effort to verify for 535 correctness. A 5 line patch can be applied by a maintainer with 536 barely a second glance. However, a 500 line patch may take hours to 537 review for correctness (the time it takes is exponentially 538 proportional to the size of the patch, or something). 539 540 Small patches also make it very easy to debug when something goes 541 wrong. It's much easier to back out patches one by one than it is 542 to dissect a very large patch after it's been applied (and broken 543 something). 544 545 2) It's important not only to send small patches, but also to rewrite 546 and simplify (or simply re-order) patches before submitting them. 547 548 Here is an analogy from kernel developer Al Viro: 549 550 *"Think of a teacher grading homework from a math student. The 551 teacher does not want to see the student's trials and errors 552 before they came up with the solution. They want to see the 553 cleanest, most elegant answer. A good student knows this, and 554 would never submit her intermediate work before the final 555 solution.* 556 557 *The same is true of kernel development. The maintainers and 558 reviewers do not want to see the thought process behind the 559 solution to the problem one is solving. They want to see a 560 simple and elegant solution."* 561 562 It may be challenging to keep the balance between presenting an elegant 563 solution and working together with the community and discussing your 564 unfinished work. Therefore it is good to get early in the process to 565 get feedback to improve your work, but also keep your changes in small 566 chunks that they may get already accepted, even when your whole task is 567 not ready for inclusion now. 568 569 Also realize that it is not acceptable to send patches for inclusion 570 that are unfinished and will be "fixed up later." 571 572 573 Justify your change 574 ------------------- 575 576 Along with breaking up your patches, it is very important for you to let 577 the Linux community know why they should add this change. New features 578 must be justified as being needed and useful. 579 580 581 Document your change 582 -------------------- 583 584 When sending in your patches, pay special attention to what you say in 585 the text in your email. This information will become the ChangeLog 586 information for the patch, and will be preserved for everyone to see for 587 all time. It should describe the patch completely, containing: 588 589 - why the change is necessary 590 - the overall design approach in the patch 591 - implementation details 592 - testing results 593 594 For more details on what this should all look like, please see the 595 ChangeLog section of the document: 596 597 "The Perfect Patch" 598 https://www.ozlabs.org/~akpm/stuff/tpp.txt 599 600 601 All of these things are sometimes very hard to do. It can take years to 602 perfect these practices (if at all). It's a continuous process of 603 improvement that requires a lot of patience and determination. But 604 don't give up, it's possible. Many have done it before, and each had to 605 start exactly where you are now. 606 607 608 609 610 ---------- 611 612 Thanks to Paolo Ciarrocchi who allowed the "Development Process" 613 (https://lwn.net/Articles/94386/) section 614 to be based on text he had written, and to Randy Dunlap and Gerrit 615 Huizenga for some of the list of things you should and should not say. 616 Also thanks to Pat Mochel, Hanna Linder, Randy Dunlap, Kay Sievers, 617 Vojtech Pavlik, Jan Kara, Josh Boyer, Kees Cook, Andrew Morton, Andi 618 Kleen, Vadim Lobanov, Jesper Juhl, Adrian Bunk, Keri Harris, Frans Pop, 619 David A. Wheeler, Junio Hamano, Michael Kerrisk, and Alex Shepard for 620 their review, comments, and contributions. Without their help, this 621 document would not have been possible. 622 623 624 625 Maintainer: Greg Kroah-Hartman <greg@kroah.com>
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.