// -*- mode:doc; -*- // vim: set syntax=asciidoc: == Contributing to Buildroot There are many ways in which you can contribute to Buildroot: analyzing and fixing bugs, analyzing and fixing package build failures detected by the autobuilders, testing and reviewing patches sent by other developers, working on the items in our TODO list and sending your own improvements to Buildroot or its manual. The following sections give a little more detail on each of these items. If you are interested in contributing to Buildroot, the first thing you should do is to subscribe to the Buildroot mailing list. This list is the main way of interacting with other Buildroot developers and to send contributions to. If you aren't subscribed yet, then refer to xref:community-resources[] for the subscription link. If you are going to touch the code, it is highly recommended to use a git repository of Buildroot, rather than starting from an extracted source code tarball. Git is the easiest way to develop from and directly send your patches to the mailing list. Refer to xref:getting-buildroot[] for more information on obtaining a Buildroot git tree. === Reproducing, analyzing and fixing bugs A first way of contributing is to have a look at the open bug reports in the https://bugs.buildroot.org/buglist.cgi?product=buildroot[Buildroot bug tracker]. As we strive to keep the bug count as small as possible, all help in reproducing, analyzing and fixing reported bugs is more than welcome. Don't hesitate to add a comment to bug reports reporting your findings, even if you don't yet see the full picture. === Analyzing and fixing autobuild failures The Buildroot autobuilders are a set of build machines that continuously run Buildroot builds based on random configurations. This is done for all architectures supported by Buildroot, with various toolchains, and with a random selection of packages. With the large commit activity on Buildroot, these autobuilders are a great help in detecting problems very early after commit. All build results are available at http://autobuild.buildroot.org[], statistics are at http://autobuild.buildroot.org/stats.php[]. Every day, an overview of all failed packages is sent to the mailing list. Detecting problems is great, but obviously these problems have to be fixed as well. Your contribution is very welcome here! There are basically two things that can be done: - Analyzing the problems. The daily summary mails do not contain details about the actual failures: in order to see what's going on you have to open the build log and check the last output. Having someone doing this for all packages in the mail is very useful for other developers, as they can make a quick initial analysis based on this output alone. - Fixing a problem. When fixing autobuild failures, you should follow these steps: . Check if you can reproduce the problem by building with the same configuration. You can do this manually, or use the http://git.buildroot.org/buildroot-test/tree/utils/br-reproduce-build[br-reproduce-build] script that will automatically clone a Buildroot git repository, checkout the correct revision, download and set the right configuration, and start the build. . Analyze the problem and create a fix. . Verify that the problem is really fixed by starting from a clean Buildroot tree and only applying your fix. . Send the fix to the Buildroot mailing list (see xref:submitting-patches[]). In case you created a patch against the package sources, you should also send the patch upstream so that the problem will be fixed in a later release, and the patch in Buildroot can be removed. In the commit message of a patch fixing an autobuild failure, add a reference to the build result directory, as follows: --------------------- Fixes: http://autobuild.buildroot.org/results/51000a9d4656afe9e0ea6f07b9f8ed374c2e4069 --------------------- === Reviewing and testing patches With the amount of patches sent to the mailing list each day, the maintainer has a very hard job to judge which patches are ready to apply and which ones aren't. Contributors can greatly help here by reviewing and testing these patches. In the review process, do not hesitate to respond to patch submissions for remarks, suggestions or anything that will help everyone to understand the patches and make them better. Please use internet style replies in plain text emails when responding to patch submissions. To indicate approval of a patch, there are three formal tags that keep track of this approval. To add your tag to a patch, reply to it with the approval tag below the original author's Signed-off-by line. These tags will be picked up automatically by patchwork (see xref:apply-patches-patchwork[]) and will be part of the commit log when the patch is accepted. Tested-by:: Indicates that the patch has been tested successfully. You are encouraged to specify what kind of testing you performed (compile-test on architecture X and Y, runtime test on target A, ...). This additional information helps other testers and the maintainer. Reviewed-by:: Indicates that you code-reviewed the patch and did your best in spotting problems, but you are not sufficiently familiar with the area touched to provide an Acked-by tag. This means that there may be remaining problems in the patch that would be spotted by someone with more experience in that area. Should such problems be detected, your Reviewed-by tag remains appropriate and you cannot be blamed. Acked-by:: Indicates that you code-reviewed the patch and you are familiar enough with the area touched to feel that the patch can be committed as-is (no additional changes required). In case it later turns out that something is wrong with the patch, your Acked-by could be considered inappropriate. The difference between Acked-by and Reviewed-by is thus mainly that you are prepared to take the blame on Acked patches, but not on Reviewed ones. If you reviewed a patch and have comments on it, you should simply reply to the patch stating these comments, without providing a Reviewed-by or Acked-by tag. These tags should only be provided if you judge the patch to be good as it is. It is important to note that neither Reviewed-by nor Acked-by imply that testing has been performed. To indicate that you both reviewed and tested the patch, provide two separate tags (Reviewed/Acked-by and Tested-by). Note also that _any developer_ can provide Tested/Reviewed/Acked-by tags, without exception, and we encourage everyone to do this. Buildroot does not have a defined group of _core_ developers, it just so happens that some developers are more active than others. The maintainer will value tags according to the track record of their submitter. Tags provided by a regular contributor will naturally be trusted more than tags provided by a newcomer. As you provide tags more regularly, your 'trustworthiness' (in the eyes of the maintainer) will go up, but _any_ tag provided is valuable. Buildroot's Patchwork website can be used to pull in patches for testing purposes. Please see xref:apply-patches-patchwork[] for more information on using Buildroot's Patchwork website to apply patches. [[apply-patches-patchwork]] ==== Applying Patches from Patchwork The main use of Buildroot's Patchwork website for a developer is for pulling in patches into their local git repository for testing purposes. When browsing patches in the patchwork management interface, an +mbox+ link is provided at the top of the page. Copy this link address and run the following commands: --------------------- $ git checkout -b $ wget -O - | git am --------------------- Another option for applying patches is to create a bundle. A bundle is a set of patches that you can group together using the patchwork interface. Once the bundle is created and the bundle is made public, you can copy the +mbox+ link for the bundle and apply the bundle using the above commands. === Work on items from the TODO list If you want to contribute to Buildroot but don't know where to start, and you don't like any of the above topics, you can always work on items from the http://elinux.org/Buildroot#Todo_list[Buildroot TODO list]. Don't hesitate to discuss an item first on the mailing list or on IRC. Do edit the wiki to indicate when you start working on an item, so we avoid duplicate efforts. [[submitting-patches]] === Submitting patches [NOTE] _Please, do not attach patches to bugs, send them to the mailing list instead_. If you made some changes to Buildroot and you would like to contribute them to the Buildroot project, proceed as follows. ==== The formatting of a patch We expect patches to be formatted in a specific way. This is necessary to make it easy to review patches, to be able to apply them easily to the git repository, to make it easy to find back in the history how and why things have changed, and to make it possible to use +git bisect+ to locate the origin of a problem. First of all, it is essential that the patch has a good commit message. The commit message should start with a separate line with a brief summary of the change, prefixed by the area touched by the patch. A few examples of good commit titles: * +package/linuxptp: bump version to 2.0+ * +configs/imx23evk: bump Linux version to 4.19+ * +package/pkg-generic: postpone evaluation of dependency conditions+ * +boot/uboot: needs host-{flex,bison}+ * +support/testing: add python-ubjson tests+ The description that follows the prefix should start with a lower case letter (i.e "bump", "needs", "postpone", "add" in the above examples). Second, the body of the commit message should describe _why_ this change is needed, and if necessary also give details about _how_ it was done. When writing the commit message, think of how the reviewers will read it, but also think about how you will read it when you look at this change again a few years down the line. Third, the patch itself should do only one change, but do it completely. Two unrelated or weakly related changes should usually be done in two separate patches. This usually means that a patch affects only a single package. If several changes are related, it is often still possible to split them up in small patches and apply them in a specific order. Small patches make it easier to review, and often make it easier to understand afterwards why a change was done. However, each patch must be complete. It is not allowed that the build is broken when only the first but not the second patch is applied. This is necessary to be able to use +git bisect+ afterwards. Of course, while you're doing your development, you're probably going back and forth between packages, and certainly not committing things immediately in a way that is clean enough for submission. So most developers rewrite the history of commits to produce a clean set of commits that is appropriate for submission. To do this, you need to use _interactive rebasing_. You can learn about it https://git-scm.com/book/en/v2/Git-Tools-Rewriting-History[in the Pro Git book]. Sometimes, it is even easier to discard you history with +git reset --soft origin/master+ and select individual changes with +git add -i+ or +git add -p+. Finally, the patch should be signed off. This is done by adding +Signed-off-by: Your Real Name + at the end of the commit message. +git commit -s+ does that for you, if configured properly. The +Signed-off-by+ tag means that you publish the patch under the Buildroot license (i.e. GPL-2.0+, except for package patches, which have the upstream license), and that you are allowed to do so. See http://developercertificate.org/[the Developer Certificate of Origin] for details. When adding new packages, you should submit every package in a separate patch. This patch should have the update to +package/Config.in+, the package +Config.in+ file, the +.mk+ file, the +.hash+ file, any init script, and all package patches. If the package has many sub-options, these are sometimes better added as separate follow-up patches. The summary line should be something like +: new package+. The body of the commit message can be empty for simple packages, or it can contain the description of the package (like the Config.in help text). If anything special has to be done to build the package, this should also be explained explicitly in the commit message body. When you bump a package to a new version, you should also submit a separate patch for each package. Don't forget to update the +.hash+ file, or add it if it doesn't exist yet. Also don't forget to check if the +_LICENSE+ and +_LICENSE_FILES+ are still valid. The summary line should be something like +: bump to version +. If the new version only contains security updates compared to the existing one, the summary should be +: security bump to version + and the commit message body should show the CVE numbers that are fixed. If some package patches can be removed in the new version, it should be explained explicitly why they can be removed, preferably with the upstream commit ID. Also any other required changes should be explained explicitly, like configure options that no longer exist or are no longer needed. If you are interested in getting notified of build failures and of further changes in the packages you added or modified, please add yourself to the DEVELOPERS file. This should be done in the same patch creating or modifying the package. See xref:DEVELOPERS[the DEVELOPERS file] for more information. Buildroot provides a handy tool to check for common coding style mistakes on files you created or modified, called +check-package+ (see xref:check-package[] for more information). ==== Preparing a patch series Starting from the changes committed in your local git view, _rebase_ your development branch on top of the upstream tree before generating a patch set. To do so, run: --------------------- $ git fetch --all --tags $ git rebase origin/master --------------------- Now, you are ready to generate then submit your patch set. To generate it, run: --------------------- $ git format-patch -M -n -s -o outgoing origin/master --------------------- This will generate patch files in the +outgoing+ subdirectory, automatically adding the +Signed-off-by+ line. Once patch files are generated, you can review/edit the commit message before submitting them, using your favorite text editor. Buildroot provides a handy tool to know to whom your patches should be sent, called +get-developers+ (see xref:DEVELOPERS[] for more information). This tool reads your patches and outputs the appropriate +git send-email+ command to use: --------------------- $ ./utils/get-developers outgoing/* --------------------- Use the output of +get-developers+ to send your patches: --------------------- $ git send-email --to buildroot@buildroot.org --cc bob --cc alice outgoing/* --------------------- Alternatively, +get-developers -e+ can be used directly with the +--cc-cmd+ argument to +git send-email+ to automatically CC the affected developers: --------------------- $ git send-email --to buildroot@buildroot.org \ --cc-cmd './utils/get-developers -e' origin/master --------------------- +git+ can be configured to automatically do this out of the box with: --------------------- $ git config sendemail.to buildroot@buildroot.org $ git config sendemail.ccCmd "$(pwd)/utils/get-developers -e" --------------------- And then just do: --------------------- $ git send-email origin/master --------------------- Note that +git+ should be configured to use your mail account. To configure +git+, see +man git-send-email+ or google it. If you do not use +git send-email+, make sure posted *patches are not line-wrapped*, otherwise they cannot easily be applied. In such a case, fix your e-mail client, or better yet, learn to use +git send-email+. ==== Cover letter If you want to present the whole patch set in a separate mail, add +--cover-letter+ to the +git format-patch+ command (see +man git-format-patch+ for further information). This will generate a template for an introduction e-mail to your patch series. A 'cover letter' may be useful to introduce the changes you propose in the following cases: * large number of commits in the series; * deep impact of the changes in the rest of the project; * RFC footnote:[RFC: (Request for comments) change proposal]; * whenever you feel it will help presenting your work, your choices, the review process, etc. ==== Patches for maintenance branches When fixing bugs on a maintenance branch, bugs should be fixed on the master branch first. The commit log for such a patch may then contain a post-commit note specifying what branches are affected: ---- package/foo: fix stuff Signed-off-by: Your Real Name --- Backport to: 2020.02.x, 2020.05.x (2020.08.x not affected as the version was bumped) ---- Those changes will then be backported by a maintainer to the affected branches. However, some bugs may apply only to a specific release, for example because it is using an older version of a package. In that case, patches should be based off the maintenance branch, and the patch subject prefix must include the maintenance branch name (for example "[PATCH 2020.02.x]"). This can be done with the +git format-patch+ flag +--subject-prefix+: --------------------- $ git format-patch --subject-prefix "PATCH 2020.02.x" \ -M -s -o outgoing origin/2020.02.x --------------------- Then send the patches with +git send-email+, as described above. ==== Patch revision changelog When improvements are requested, the new revision of each commit should include a changelog of the modifications between each submission. Note that when your patch series is introduced by a cover letter, an overall changelog may be added to the cover letter in addition to the changelog in the individual commits. The best thing to rework a patch series is by interactive rebasing: +git rebase -i origin/master+. Consult the git manual for more information. When added to the individual commits, this changelog is added when editing the commit message. Below the +Signed-off-by+ section, add +---+ and your changelog. Although the changelog will be visible for the reviewers in the mail thread, as well as in http://patchwork.buildroot.org[patchwork], +git+ will automatically ignores lines below +---+ when the patch will be merged. This is the intended behavior: the changelog is not meant to be preserved forever in the +git+ history of the project. Hereafter the recommended layout: --------------- Patch title: short explanation, max 72 chars A paragraph that explains the problem, and how it manifests itself. If the problem is complex, it is OK to add more paragraphs. All paragraphs should be wrapped at 72 characters. A paragraph that explains the root cause of the problem. Again, more than one paragraph is OK. Finally, one or more paragraphs that explain how the problem is solved. Don't hesitate to explain complex solutions in detail. Signed-off-by: John DOE --- Changes v2 -> v3: - foo bar (suggested by Jane) - bar buz Changes v1 -> v2: - alpha bravo (suggested by John) - charly delta --------------- Any patch revision should include the version number. The version number is simply composed of the letter +v+ followed by an +integer+ greater or equal to two (i.e. "PATCH v2", "PATCH v3" ...). This can be easily handled with +git format-patch+ by using the option +--subject-prefix+: --------------------- $ git format-patch --subject-prefix "PATCH v4" \ -M -s -o outgoing origin/master --------------------- Since git version 1.8.1, you can also use +-v + (where is the version number): --------------------- $ git format-patch -v4 -M -s -o outgoing origin/master --------------------- When you provide a new version of a patch, please mark the old one as superseded in http://patchwork.buildroot.org[patchwork]. You need to create an account on http://patchwork.buildroot.org[patchwork] to be able to modify the status of your patches. Note that you can only change the status of patches you submitted yourself, which means the email address you register in http://patchwork.buildroot.org[patchwork] should match the one you use for sending patches to the mailing list. You can also add the +--in-reply-to + option when submitting a patch to the mailing list. The id of the mail to reply to can be found under the "Message Id" tag on http://patchwork.buildroot.org[patchwork]. The advantage of *in-reply-to* is that patchwork will automatically mark the previous version of the patch as superseded. [[reporting-bugs]] === Reporting issues/bugs or getting help Before reporting any issue, please check in xref:community-resources[the mailing list archive] whether someone has already reported and/or fixed a similar problem. However you choose to report bugs or get help, either by opening a bug in the xref:community-resources[bug tracker] or by xref:community-resources[sending a mail to the mailing list], there are a number of details to provide in order to help people reproduce and find a solution to the issue. Try to think as if you were trying to help someone else; in that case, what would you need? Here is a short list of details to provide in such case: * host machine (OS/release) * version of Buildroot * target for which the build fails * package(s) for which the build fails * the command that fails and its output * any information you think that may be relevant Additionally, you should add the +.config+ file (or if you know how, a +defconfig+; see xref:customize-store-buildroot-config[]). If some of these details are too large, do not hesitate to use a pastebin service. Note that not all available pastebin services will preserve Unix-style line terminators when downloading raw pastes. Following pastebin services are known to work correctly: - https://gist.github.com/ - http://code.bulix.org/ === Using the run-tests framework Buildroot includes a run-time testing framework called run-tests built upon Python scripting and QEMU runtime execution. There are two types of test cases within the framework, one for build time tests and another for run-time tests that have a QEMU dependency. The goals of the framework are the following: * build a well defined configuration * optionally, verify some properties of the build output * if it is a run-time test: ** boot it under QEMU ** run some test condition to verify that a given feature is working The run-tests tool has a series of options documented in the tool's help '-h' description. Some common options include setting the download folder, the output folder, keeping build output, and for multiple test cases, you can set the JLEVEL for each. Here is an example walk through of running a test case. * For a first step, let us see what all the test case options are. The test cases can be listed by executing +support/testing/run-tests -l+. These tests can all be run individually during test development from the console. Both one at a time and selectively as a group of a subset of tests. --------------------- $ support/testing/run-tests -l List of tests test_run (tests.utils.test_check_package.TestCheckPackage) Test the various ways the script can be called in a simple top to ... ok test_run (tests.toolchain.test_external.TestExternalToolchainBuildrootMusl) ... ok test_run (tests.toolchain.test_external.TestExternalToolchainBuildrootuClibc) ... ok test_run (tests.toolchain.test_external.TestExternalToolchainCCache) ... ok test_run (tests.toolchain.test_external.TestExternalToolchainCtngMusl) ... ok test_run (tests.toolchain.test_external.TestExternalToolchainLinaroArm) ... ok test_run (tests.toolchain.test_external.TestExternalToolchainSourceryArmv4) ... ok test_run (tests.toolchain.test_external.TestExternalToolchainSourceryArmv5) ... ok test_run (tests.toolchain.test_external.TestExternalToolchainSourceryArmv7) ... ok [snip] test_run (tests.init.test_systemd.TestInitSystemSystemdRoFull) ... ok test_run (tests.init.test_systemd.TestInitSystemSystemdRoIfupdown) ... ok test_run (tests.init.test_systemd.TestInitSystemSystemdRoNetworkd) ... ok test_run (tests.init.test_systemd.TestInitSystemSystemdRwFull) ... ok test_run (tests.init.test_systemd.TestInitSystemSystemdRwIfupdown) ... ok test_run (tests.init.test_systemd.TestInitSystemSystemdRwNetworkd) ... ok test_run (tests.init.test_busybox.TestInitSystemBusyboxRo) ... ok test_run (tests.init.test_busybox.TestInitSystemBusyboxRoNet) ... ok test_run (tests.init.test_busybox.TestInitSystemBusyboxRw) ... ok test_run (tests.init.test_busybox.TestInitSystemBusyboxRwNet) ... ok Ran 157 tests in 0.021s OK --------------------- Those runtime tests are regularly executed by Buildroot Gitlab CI infrastructure, see .gitlab.yml and https://gitlab.com/buildroot.org/buildroot/-/jobs. ==== Creating a test case The best way to get familiar with how to create a test case is to look at a few of the basic file system +support/testing/tests/fs/+ and init +support/testing/tests/init/+ test scripts. Those tests give good examples of a basic build and build with run type of tests. There are other more advanced cases that use things like nested +br2-external+ folders to provide skeletons and additional packages. The test cases by default use a br-arm-full-* uClibc-ng toolchain and the prebuild kernel for a armv5/7 cpu. It is recommended to use the default defconfig test configuration except when Glibc/musl or a newer kernel are necessary. By using the default it saves build time and the test would automatically inherit a kernel/std library upgrade when the default is updated. The basic test case definition involves * Creation of a new test file * Defining a unique test class * Determining if the default defconfig plus test options can be used * Implementing a +def test_run(self):+ function to optionally startup the emulator and provide test case conditions. Beyond creating the test script, there are a couple of additional steps that should be taken once you have your initial test case script. The first is to add yourself to the +DEVELOPERS+ file to be the maintainer of that test case. The second is to update the Gitlab CI yml by executing +make .gitlab-ci.yml+. ==== Debugging a test case Within the Buildroot repository, the testing framework is organized at the top level in +support/testing/+ by folders of +conf+, +infra+ and +tests+. All the test cases live under the +test+ folder and are organized in various folders representing the catagory of test. Lets walk through an example. * Using the Busybox Init system test case with a read/write rootfs +tests.init.test_busybox.TestInitSystemBusyboxRw+ * A minimal set of command line arguments when debugging a test case would include '-d' which points to your dl folder, '-o' to an output folder, and '-k' to keep any output on both pass/fail. With those options, the test will retain logging and build artifacts providing status of the build and execution of the test case. --------------------- $ support/testing/run-tests -d dl -o output_folder -k tests.init.test_busybox.TestInitSystemBusyboxRw 15:03:26 TestInitSystemBusyboxRw Starting 15:03:28 TestInitSystemBusyboxRw Building 15:08:18 TestInitSystemBusyboxRw Building done 15:08:27 TestInitSystemBusyboxRw Cleaning up . Ran 1 test in 301.140s OK --------------------- * For the case of a successful build, the +output_folder+ would contain a folder with the Buildroot build, build log and run-time log. If the build failed, the console output would show the stage at which it failed (setup / build / run). Depending on the failure stage, the build/run logs and/or Buildroot build artifacts can be inspected and instrumented. If the QEMU instance needs to be launched for additional testing, the first few lines of the run-time log capture it and it would allow some incremental testing without re-running +support/testing/run-tests+. * You can also make modifications to the current sources inside the +output_folder+ (e.g. for debug purposes) and rerun the standard Buildroot make targets (in order to regenerate the complete image with the new modifications) and then rerun the test. Modifying the sources directly can speed up debugging compared to adding patch files, wiping the output directoy, and starting the test again. --------------------- $ ls output_folder/ TestInitSystemBusyboxRw/ TestInitSystemBusyboxRw-build.log TestInitSystemBusyboxRw-run.log --------------------- * The source file used to implement this example test is found under +support/testing/tests/init/test_busybox.py+. This file outlines the minimal defconfig that creates the build, QEMU configuration to launch the built images and the test case assertions. To test an existing or new test case within Gitlab CI, there is a method of invoking a specific test by creating a Buildroot fork in Gitlab under your account. This can be handy when adding/changing a run-time test or fixing a bug on a use case tested by a run-time test case. In the examples below, the component of the branch name is a unique string you choose to identify this specific job being created. * to trigger all run-test test case jobs: --------------------- $ git push gitlab HEAD:-runtime-tests --------------------- * to trigger one test case job, a specific branch naming string is used that includes the full test case name. --------------------- $ git push gitlab HEAD:- ---------------------