Based on kernel version 4.9. Page generated on 2016-12-21 14:34 EST.
1 HOWTO do Linux kernel development 2 ================================= 3 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programming, 8 but will help point you in the right direction for that. 9 10 If anything in this document becomes out of date, please send in patches 11 to the maintainer of this file, who is listed at the bottom of the 12 document. 13 14 15 Introduction 16 ------------ 17 18 So, you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. 24 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 32 - "The C Programming Language" by Kernighan and Ritchie [Prentice Hall] 33 - "Practical C Programming" by Steve Oualline [O'Reilly] 34 - "C: A Reference Manual" by Harbison and Steele [Prentice Hall] 35 36 The kernel is written using GNU C and the GNU toolchain. While it 37 adheres to the ISO C89 standard, it uses a number of extensions that are 38 not featured in the standard. The kernel is a freestanding C 39 environment, with no reliance on the standard C library, so some 40 portions of the C standard are not supported. Arbitrary long long 41 divisions and floating point are not allowed. It can sometimes be 42 difficult to understand the assumptions the kernel has on the toolchain 43 and the extensions that it uses, and unfortunately there is no 44 definitive reference for them. Please check the gcc info pages (`info 45 gcc`) for some information on them. 46 47 Please remember that you are trying to learn how to work with the 48 existing development community. It is a diverse group of people, with 49 high standards for coding, style and procedure. These standards have 50 been created over time based on what they have found to work best for 51 such a large and geographically dispersed team. Try to learn as much as 52 possible about these standards ahead of time, as they are well 53 documented; do not expect people to adapt to you or your company's way 54 of doing things. 55 56 57 Legal Issues 58 ------------ 59 60 The Linux kernel source code is released under the GPL. Please see the 61 file, COPYING, in the main directory of the source tree, for details on 62 the license. If you have further questions about the license, please 63 contact a lawyer, and do not ask on the Linux kernel mailing list. The 64 people on the mailing lists are not lawyers, and you should not rely on 65 their statements on legal matters. 66 67 For common questions and answers about the GPL, please see: 68 69 https://www.gnu.org/licenses/gpl-faq.html 70 71 72 Documentation 73 ------------- 74 75 The Linux kernel source tree has a large range of documents that are 76 invaluable for learning how to interact with the kernel community. When 77 new features are added to the kernel, it is recommended that new 78 documentation files are also added which explain how to use the feature. 79 When a kernel change causes the interface that the kernel exposes to 80 userspace to change, it is recommended that you send the information or 81 a patch to the manual pages explaining the change to the manual pages 82 maintainer at firstname.lastname@example.org, and CC the list 83 email@example.com. 84 85 Here is a list of files that are in the kernel source tree that are 86 required reading: 87 88 README 89 This file gives a short background on the Linux kernel and describes 90 what is necessary to do to configure and build the kernel. People 91 who are new to the kernel should start here. 92 93 :ref:`Documentation/Changes <changes>` 94 This file gives a list of the minimum levels of various software 95 packages that are necessary to build and run the kernel 96 successfully. 97 98 :ref:`Documentation/CodingStyle <codingstyle>` 99 This describes the Linux kernel coding style, and some of the 100 rationale behind it. All new code is expected to follow the 101 guidelines in this document. Most maintainers will only accept 102 patches if these rules are followed, and many people will only 103 review code if it is in the proper style. 104 105 :ref:`Documentation/SubmittingPatches <submittingpatches>` and :ref:`Documentation/SubmittingDrivers <submittingdrivers>` 106 These files describe in explicit detail how to successfully create 107 and send a patch, including (but not limited to): 108 109 - Email contents 110 - Email format 111 - Who to send it to 112 113 Following these rules will not guarantee success (as all patches are 114 subject to scrutiny for content and style), but not following them 115 will almost always prevent it. 116 117 Other excellent descriptions of how to create patches properly are: 118 119 "The Perfect Patch" 120 https://www.ozlabs.org/~akpm/stuff/tpp.txt 121 122 "Linux kernel patch submission format" 123 http://linux.yyz.us/patch-format.html 124 125 :ref:`Documentation/stable_api_nonsense.txt <stable_api_nonsense>` 126 This file describes the rationale behind the conscious decision to 127 not have a stable API within the kernel, including things like: 128 129 - Subsystem shim-layers (for compatibility?) 130 - Driver portability between Operating Systems. 131 - Mitigating rapid change within the kernel source tree (or 132 preventing rapid change) 133 134 This document is crucial for understanding the Linux development 135 philosophy and is very important for people moving to Linux from 136 development on other Operating Systems. 137 138 :ref:`Documentation/SecurityBugs <securitybugs>` 139 If you feel you have found a security problem in the Linux kernel, 140 please follow the steps in this document to help notify the kernel 141 developers, and help solve the issue. 142 143 :ref:`Documentation/ManagementStyle <managementstyle>` 144 This document describes how Linux kernel maintainers operate and the 145 shared ethos behind their methodologies. This is important reading 146 for anyone new to kernel development (or anyone simply curious about 147 it), as it resolves a lot of common misconceptions and confusion 148 about the unique behavior of kernel maintainers. 149 150 :ref:`Documentation/stable_kernel_rules.txt <stable_kernel_rules>` 151 This file describes the rules on how the stable kernel releases 152 happen, and what to do if you want to get a change into one of these 153 releases. 154 155 :ref:`Documentation/kernel-docs.txt <kernel_docs>` 156 A list of external documentation that pertains to kernel 157 development. Please consult this list if you do not find what you 158 are looking for within the in-kernel documentation. 159 160 :ref:`Documentation/applying-patches.txt <applying_patches>` 161 A good introduction describing exactly what a patch is and how to 162 apply it to the different development branches of the kernel. 163 164 The kernel also has a large number of documents that can be 165 automatically generated from the source code itself or from 166 ReStructuredText markups (ReST), like this one. This includes a 167 full description of the in-kernel API, and rules on how to handle 168 locking properly. 169 170 All such documents can be generated as PDF or HTML by running:: 171 172 make pdfdocs 173 make htmldocs 174 175 respectively from the main kernel source directory. 176 177 The documents that uses ReST markup will be generated at Documentation/output. 178 They can also be generated on LaTeX and ePub formats with:: 179 180 make latexdocs 181 make epubdocs 182 183 Currently, there are some documents written on DocBook that are in 184 the process of conversion to ReST. Such documents will be created in the 185 Documentation/DocBook/ directory and can be generated also as 186 Postscript or man pages by running:: 187 188 make psdocs 189 make mandocs 190 191 Becoming A Kernel Developer 192 --------------------------- 193 194 If you do not know anything about Linux kernel development, you should 195 look at the Linux KernelNewbies project: 196 197 https://kernelnewbies.org 198 199 It consists of a helpful mailing list where you can ask almost any type 200 of basic kernel development question (make sure to search the archives 201 first, before asking something that has already been answered in the 202 past.) It also has an IRC channel that you can use to ask questions in 203 real-time, and a lot of helpful documentation that is useful for 204 learning about Linux kernel development. 205 206 The website has basic information about code organization, subsystems, 207 and current projects (both in-tree and out-of-tree). It also describes 208 some basic logistical information, like how to compile a kernel and 209 apply a patch. 210 211 If you do not know where you want to start, but you want to look for 212 some task to start doing to join into the kernel development community, 213 go to the Linux Kernel Janitor's project: 214 215 https://kernelnewbies.org/KernelJanitors 216 217 It is a great place to start. It describes a list of relatively simple 218 problems that need to be cleaned up and fixed within the Linux kernel 219 source tree. Working with the developers in charge of this project, you 220 will learn the basics of getting your patch into the Linux kernel tree, 221 and possibly be pointed in the direction of what to go work on next, if 222 you do not already have an idea. 223 224 If you already have a chunk of code that you want to put into the kernel 225 tree, but need some help getting it in the proper form, the 226 kernel-mentors project was created to help you out with this. It is a 227 mailing list, and can be found at: 228 229 https://selenic.com/mailman/listinfo/kernel-mentors 230 231 Before making any actual modifications to the Linux kernel code, it is 232 imperative to understand how the code in question works. For this 233 purpose, nothing is better than reading through it directly (most tricky 234 bits are commented well), perhaps even with the help of specialized 235 tools. One such tool that is particularly recommended is the Linux 236 Cross-Reference project, which is able to present source code in a 237 self-referential, indexed webpage format. An excellent up-to-date 238 repository of the kernel code may be found at: 239 240 http://lxr.free-electrons.com/ 241 242 243 The development process 244 ----------------------- 245 246 Linux kernel development process currently consists of a few different 247 main kernel "branches" and lots of different subsystem-specific kernel 248 branches. These different branches are: 249 250 - main 4.x kernel tree 251 - 4.x.y -stable kernel tree 252 - 4.x -git kernel patches 253 - subsystem specific kernel trees and patches 254 - the 4.x -next kernel tree for integration tests 255 256 4.x kernel tree 257 ----------------- 258 4.x kernels are maintained by Linus Torvalds, and can be found on 259 https://kernel.org in the pub/linux/kernel/v4.x/ directory. Its development 260 process is as follows: 261 262 - As soon as a new kernel is released a two weeks window is open, 263 during this period of time maintainers can submit big diffs to 264 Linus, usually the patches that have already been included in the 265 -next kernel for a few weeks. The preferred way to submit big changes 266 is using git (the kernel's source management tool, more information 267 can be found at https://git-scm.com/) but plain patches are also just 268 fine. 269 - After two weeks a -rc1 kernel is released it is now possible to push 270 only patches that do not include new features that could affect the 271 stability of the whole kernel. Please note that a whole new driver 272 (or filesystem) might be accepted after -rc1 because there is no 273 risk of causing regressions with such a change as long as the change 274 is self-contained and does not affect areas outside of the code that 275 is being added. git can be used to send patches to Linus after -rc1 276 is released, but the patches need to also be sent to a public 277 mailing list for review. 278 - A new -rc is released whenever Linus deems the current git tree to 279 be in a reasonably sane state adequate for testing. The goal is to 280 release a new -rc kernel every week. 281 - Process continues until the kernel is considered "ready", the 282 process should last around 6 weeks. 283 284 It is worth mentioning what Andrew Morton wrote on the linux-kernel 285 mailing list about kernel releases: 286 287 *"Nobody knows when a kernel will be released, because it's 288 released according to perceived bug status, not according to a 289 preconceived timeline."* 290 291 4.x.y -stable kernel tree 292 ------------------------- 293 Kernels with 3-part versions are -stable kernels. They contain 294 relatively small and critical fixes for security problems or significant 295 regressions discovered in a given 4.x kernel. 296 297 This is the recommended branch for users who want the most recent stable 298 kernel and are not interested in helping test development/experimental 299 versions. 300 301 If no 4.x.y kernel is available, then the highest numbered 4.x 302 kernel is the current stable kernel. 303 304 4.x.y are maintained by the "stable" team <firstname.lastname@example.org>, and 305 are released as needs dictate. The normal release period is approximately 306 two weeks, but it can be longer if there are no pressing problems. A 307 security-related problem, instead, can cause a release to happen almost 308 instantly. 309 310 The file Documentation/stable_kernel_rules.txt in the kernel tree 311 documents what kinds of changes are acceptable for the -stable tree, and 312 how the release process works. 313 314 4.x -git patches 315 ---------------- 316 These are daily snapshots of Linus' kernel tree which are managed in a 317 git repository (hence the name.) These patches are usually released 318 daily and represent the current state of Linus' tree. They are more 319 experimental than -rc kernels since they are generated automatically 320 without even a cursory glance to see if they are sane. 321 322 Subsystem Specific kernel trees and patches 323 ------------------------------------------- 324 The maintainers of the various kernel subsystems --- and also many 325 kernel subsystem developers --- expose their current state of 326 development in source repositories. That way, others can see what is 327 happening in the different areas of the kernel. In areas where 328 development is rapid, a developer may be asked to base his submissions 329 onto such a subsystem kernel tree so that conflicts between the 330 submission and other already ongoing work are avoided. 331 332 Most of these repositories are git trees, but there are also other SCMs 333 in use, or patch queues being published as quilt series. Addresses of 334 these subsystem repositories are listed in the MAINTAINERS file. Many 335 of them can be browsed at https://git.kernel.org/. 336 337 Before a proposed patch is committed to such a subsystem tree, it is 338 subject to review which primarily happens on mailing lists (see the 339 respective section below). For several kernel subsystems, this review 340 process is tracked with the tool patchwork. Patchwork offers a web 341 interface which shows patch postings, any comments on a patch or 342 revisions to it, and maintainers can mark patches as under review, 343 accepted, or rejected. Most of these patchwork sites are listed at 344 https://patchwork.kernel.org/. 345 346 4.x -next kernel tree for integration tests 347 ------------------------------------------- 348 Before updates from subsystem trees are merged into the mainline 4.x 349 tree, they need to be integration-tested. For this purpose, a special 350 testing repository exists into which virtually all subsystem trees are 351 pulled on an almost daily basis: 352 353 https://git.kernel.org/?p=linux/kernel/git/next/linux-next.git 354 355 This way, the -next kernel gives a summary outlook onto what will be 356 expected to go into the mainline kernel at the next merge period. 357 Adventurous testers are very welcome to runtime-test the -next kernel. 358 359 360 Bug Reporting 361 ------------- 362 363 https://bugzilla.kernel.org is where the Linux kernel developers track kernel 364 bugs. Users are encouraged to report all bugs that they find in this 365 tool. For details on how to use the kernel bugzilla, please see: 366 367 https://bugzilla.kernel.org/page.cgi?id=faq.html 368 369 The file REPORTING-BUGS in the main kernel source directory has a good 370 template for how to report a possible kernel bug, and details what kind 371 of information is needed by the kernel developers to help track down the 372 problem. 373 374 375 Managing bug reports 376 -------------------- 377 378 One of the best ways to put into practice your hacking skills is by fixing 379 bugs reported by other people. Not only you will help to make the kernel 380 more stable, you'll learn to fix real world problems and you will improve 381 your skills, and other developers will be aware of your presence. Fixing 382 bugs is one of the best ways to get merits among other developers, because 383 not many people like wasting time fixing other people's bugs. 384 385 To work in the already reported bug reports, go to https://bugzilla.kernel.org. 386 If you want to be advised of the future bug reports, you can subscribe to the 387 bugme-new mailing list (only new bug reports are mailed here) or to the 388 bugme-janitor mailing list (every change in the bugzilla is mailed here) 389 390 https://lists.linux-foundation.org/mailman/listinfo/bugme-new 391 392 https://lists.linux-foundation.org/mailman/listinfo/bugme-janitors 393 394 395 396 Mailing lists 397 ------------- 398 399 As some of the above documents describe, the majority of the core kernel 400 developers participate on the Linux Kernel Mailing list. Details on how 401 to subscribe and unsubscribe from the list can be found at: 402 403 http://vger.kernel.org/vger-lists.html#linux-kernel 404 405 There are archives of the mailing list on the web in many different 406 places. Use a search engine to find these archives. For example: 407 408 http://dir.gmane.org/gmane.linux.kernel 409 410 It is highly recommended that you search the archives about the topic 411 you want to bring up, before you post it to the list. A lot of things 412 already discussed in detail are only recorded at the mailing list 413 archives. 414 415 Most of the individual kernel subsystems also have their own separate 416 mailing list where they do their development efforts. See the 417 MAINTAINERS file for a list of what these lists are for the different 418 groups. 419 420 Many of the lists are hosted on kernel.org. Information on them can be 421 found at: 422 423 http://vger.kernel.org/vger-lists.html 424 425 Please remember to follow good behavioral habits when using the lists. 426 Though a bit cheesy, the following URL has some simple guidelines for 427 interacting with the list (or any list): 428 429 http://www.albion.com/netiquette/ 430 431 If multiple people respond to your mail, the CC: list of recipients may 432 get pretty large. Don't remove anybody from the CC: list without a good 433 reason, or don't reply only to the list address. Get used to receiving the 434 mail twice, one from the sender and the one from the list, and don't try 435 to tune that by adding fancy mail-headers, people will not like it. 436 437 Remember to keep the context and the attribution of your replies intact, 438 keep the "John Kernelhacker wrote ...:" lines at the top of your reply, and 439 add your statements between the individual quoted sections instead of 440 writing at the top of the mail. 441 442 If you add patches to your mail, make sure they are plain readable text 443 as stated in Documentation/SubmittingPatches. 444 Kernel developers don't want to deal with 445 attachments or compressed patches; they may want to comment on 446 individual lines of your patch, which works only that way. Make sure you 447 use a mail program that does not mangle spaces and tab characters. A 448 good first test is to send the mail to yourself and try to apply your 449 own patch by yourself. If that doesn't work, get your mail program fixed 450 or change it until it works. 451 452 Above all, please remember to show respect to other subscribers. 453 454 455 Working with the community 456 -------------------------- 457 458 The goal of the kernel community is to provide the best possible kernel 459 there is. When you submit a patch for acceptance, it will be reviewed 460 on its technical merits and those alone. So, what should you be 461 expecting? 462 463 - criticism 464 - comments 465 - requests for change 466 - requests for justification 467 - silence 468 469 Remember, this is part of getting your patch into the kernel. You have 470 to be able to take criticism and comments about your patches, evaluate 471 them at a technical level and either rework your patches or provide 472 clear and concise reasoning as to why those changes should not be made. 473 If there are no responses to your posting, wait a few days and try 474 again, sometimes things get lost in the huge volume. 475 476 What should you not do? 477 478 - expect your patch to be accepted without question 479 - become defensive 480 - ignore comments 481 - resubmit the patch without making any of the requested changes 482 483 In a community that is looking for the best technical solution possible, 484 there will always be differing opinions on how beneficial a patch is. 485 You have to be cooperative, and willing to adapt your idea to fit within 486 the kernel. Or at least be willing to prove your idea is worth it. 487 Remember, being wrong is acceptable as long as you are willing to work 488 toward a solution that is right. 489 490 It is normal that the answers to your first patch might simply be a list 491 of a dozen things you should correct. This does **not** imply that your 492 patch will not be accepted, and it is **not** meant against you 493 personally. Simply correct all issues raised against your patch and 494 resend it. 495 496 497 Differences between the kernel community and corporate structures 498 ----------------------------------------------------------------- 499 500 The kernel community works differently than most traditional corporate 501 development environments. Here are a list of things that you can try to 502 do to avoid problems: 503 504 Good things to say regarding your proposed changes: 505 506 - "This solves multiple problems." 507 - "This deletes 2000 lines of code." 508 - "Here is a patch that explains what I am trying to describe." 509 - "I tested it on 5 different architectures..." 510 - "Here is a series of small patches that..." 511 - "This increases performance on typical machines..." 512 513 Bad things you should avoid saying: 514 515 - "We did it this way in AIX/ptx/Solaris, so therefore it must be 516 good..." 517 - "I've being doing this for 20 years, so..." 518 - "This is required for my company to make money" 519 - "This is for our Enterprise product line." 520 - "Here is my 1000 page design document that describes my idea" 521 - "I've been working on this for 6 months..." 522 - "Here's a 5000 line patch that..." 523 - "I rewrote all of the current mess, and here it is..." 524 - "I have a deadline, and this patch needs to be applied now." 525 526 Another way the kernel community is different than most traditional 527 software engineering work environments is the faceless nature of 528 interaction. One benefit of using email and irc as the primary forms of 529 communication is the lack of discrimination based on gender or race. 530 The Linux kernel work environment is accepting of women and minorities 531 because all you are is an email address. The international aspect also 532 helps to level the playing field because you can't guess gender based on 533 a person's name. A man may be named Andrea and a woman may be named Pat. 534 Most women who have worked in the Linux kernel and have expressed an 535 opinion have had positive experiences. 536 537 The language barrier can cause problems for some people who are not 538 comfortable with English. A good grasp of the language can be needed in 539 order to get ideas across properly on mailing lists, so it is 540 recommended that you check your emails to make sure they make sense in 541 English before sending them. 542 543 544 Break up your changes 545 --------------------- 546 547 The Linux kernel community does not gladly accept large chunks of code 548 dropped on it all at once. The changes need to be properly introduced, 549 discussed, and broken up into tiny, individual portions. This is almost 550 the exact opposite of what companies are used to doing. Your proposal 551 should also be introduced very early in the development process, so that 552 you can receive feedback on what you are doing. It also lets the 553 community feel that you are working with them, and not simply using them 554 as a dumping ground for your feature. However, don't send 50 emails at 555 one time to a mailing list, your patch series should be smaller than 556 that almost all of the time. 557 558 The reasons for breaking things up are the following: 559 560 1) Small patches increase the likelihood that your patches will be 561 applied, since they don't take much time or effort to verify for 562 correctness. A 5 line patch can be applied by a maintainer with 563 barely a second glance. However, a 500 line patch may take hours to 564 review for correctness (the time it takes is exponentially 565 proportional to the size of the patch, or something). 566 567 Small patches also make it very easy to debug when something goes 568 wrong. It's much easier to back out patches one by one than it is 569 to dissect a very large patch after it's been applied (and broken 570 something). 571 572 2) It's important not only to send small patches, but also to rewrite 573 and simplify (or simply re-order) patches before submitting them. 574 575 Here is an analogy from kernel developer Al Viro: 576 577 *"Think of a teacher grading homework from a math student. The 578 teacher does not want to see the student's trials and errors 579 before they came up with the solution. They want to see the 580 cleanest, most elegant answer. A good student knows this, and 581 would never submit her intermediate work before the final 582 solution.* 583 584 *The same is true of kernel development. The maintainers and 585 reviewers do not want to see the thought process behind the 586 solution to the problem one is solving. They want to see a 587 simple and elegant solution."* 588 589 It may be challenging to keep the balance between presenting an elegant 590 solution and working together with the community and discussing your 591 unfinished work. Therefore it is good to get early in the process to 592 get feedback to improve your work, but also keep your changes in small 593 chunks that they may get already accepted, even when your whole task is 594 not ready for inclusion now. 595 596 Also realize that it is not acceptable to send patches for inclusion 597 that are unfinished and will be "fixed up later." 598 599 600 Justify your change 601 ------------------- 602 603 Along with breaking up your patches, it is very important for you to let 604 the Linux community know why they should add this change. New features 605 must be justified as being needed and useful. 606 607 608 Document your change 609 -------------------- 610 611 When sending in your patches, pay special attention to what you say in 612 the text in your email. This information will become the ChangeLog 613 information for the patch, and will be preserved for everyone to see for 614 all time. It should describe the patch completely, containing: 615 616 - why the change is necessary 617 - the overall design approach in the patch 618 - implementation details 619 - testing results 620 621 For more details on what this should all look like, please see the 622 ChangeLog section of the document: 623 624 "The Perfect Patch" 625 http://www.ozlabs.org/~akpm/stuff/tpp.txt 626 627 628 All of these things are sometimes very hard to do. It can take years to 629 perfect these practices (if at all). It's a continuous process of 630 improvement that requires a lot of patience and determination. But 631 don't give up, it's possible. Many have done it before, and each had to 632 start exactly where you are now. 633 634 635 636 637 ---------- 638 639 Thanks to Paolo Ciarrocchi who allowed the "Development Process" 640 (https://lwn.net/Articles/94386/) section 641 to be based on text he had written, and to Randy Dunlap and Gerrit 642 Huizenga for some of the list of things you should and should not say. 643 Also thanks to Pat Mochel, Hanna Linder, Randy Dunlap, Kay Sievers, 644 Vojtech Pavlik, Jan Kara, Josh Boyer, Kees Cook, Andrew Morton, Andi 645 Kleen, Vadim Lobanov, Jesper Juhl, Adrian Bunk, Keri Harris, Frans Pop, 646 David A. Wheeler, Junio Hamano, Michael Kerrisk, and Alex Shepard for 647 their review, comments, and contributions. Without their help, this 648 document would not have been possible. 649 650 651 652 Maintainer: Greg Kroah-Hartman <email@example.com>