Kernel Repository Guidelines ============================ Table Of Contents ================= Getting Started Patch Headers Before You Commit -- Things To Check Config Option Changes Committing and Log Messages What Is The Kernel ABI? Kernel ABI Changes Embargoed Patches Related Information Getting Started =============== Make sure you have the git and quilt tools installed. Introduce yourself if you haven't done so already: $ git config --global user.name "Your Name" $ git config --global user.email your@email If you omit the --global option, the setting will only apply to this repository clone. Set up some Git hooks and helpers: $ ./scripts/install-git-hooks To hack on the kernel sources: $ ./scripts/sequence-patch.sh $ cd tmp/linux-$version-$branch $ quilt new patches.suse/fix-foo-and-bar.patch $ quilt edit some/file.c $ ./refresh_patch.sh $ quilt header -e # see next chapter Refer to the Quilt documentation for details. When you are done, add the new patch to an appropriate place in the series.conf file and run ./scripts/log to commit it. Patches should be named such that they consist of alphanumeric characters, '-' and '.'. Typically, patches are named by filtering the Subject of the patch to a lower-case, dash-separated form like the one in the example above. To build RPM packages: $ ./scripts/tar-up.sh This creates a source package in the kernel-source directory. Use $ ./scripts/osc_wrapper [kernel-source/kernel-$flavor.spec] to build a kernel package locally, or $ ./scripts/osc_wrapper upload [--ibs] to have all flavors and architectures built by the Open Build Service. The --ibs option uses the SUSE internal instance. Patch Headers ============= Each patch must have an RFC822-style header that at a minimum describes what the patch does, who wrote it and who inside SUSE can be contacted about problems with the patch. The rules for patch headers are: * Each patch must have a From tag that identifies the author of the patch. * Each patch must have a Subject tag that briefly describes what the patch does. A brief summary that could appear in a change log makes the most sense in most cases. * Unless the author specified in the From tag has a @suse.de, @suse.com or @suse.cz address, the patch must include a Signed-off-by, Acked-by or Reviewed-by header which identifies the person in one of these domains who feels responsible for the patch inside the company. * The patch must include a Patch-mainline tag that identifies where the patch came from (for backports from mainline) or when it is expected to be added to mainline. The format is one of: For backports from mainline: Patch-mainline: <upstream version, for instance, "v6.5-rc7"> Git-commit: <git hash> If the commit is from a maintainer repository or some other repository that isn't Linus's: Patch-mainline: Queued in subsystem maintainer repository Git-repo: <url> Git-commit: <git hash> If the patch is not upstream, depending on the situation: Patch-mainline: Submitted, <timestamp - destination> Patch-mainline: Not yet, <reason> Patch-mainline: Never, <reason> * The patch should include a References tag that identifies the Bugzilla bug number, JIRA issue ID, etc. where the patch is discussed. Please prefix bugzilla.suse.com bug numbers with bsc# and JIRA issue IDs with jsc#. Please make sure you specify a JIRA Implementation task when referencing JIRA features, not its Epic ID. Have a look at https://en.opensuse.org/openSUSE:Packaging_Patches_guidelines#Current_set_of_abbreviations for a full list of abbreviations. * The patch header should include a more extensive description of what the patch does, why and how. The idea is to allow others to quickly identify what each patch is about and to give enough information for reviewing. More details about valid patch headers can be found in scripts/patch-tag-template. The helper script scripts/patch-tag can be used for managing these tags. Documentation for patch-tag can be found at the top of the script itself. Example usage of scripts/patch-tag-template: $ cp scripts/patch-tag-template ~/.patchtag Edit ~/.patchtag with any default values you want $ patch-tag -e file.diff Example patch header: | From: Pablo Neira Ayuso <pablo@netfilter.org> | Date: Tue, 15 Aug 2023 15:39:01 +0200 | Subject: netfilter: nf_tables: GC transaction race with netns dismantle | Patch-mainline: v6.5-rc7 | Git-commit: 02c6c24402bf1c1e986899c14ba22a10b510916b | References: CVE-2023-4563 bsc#1214727 | | Use maybe_get_net() since GC workqueue might race with netns exit path. | | Fixes: 5f68718b34a5 ("netfilter: nf_tables: GC transaction API to avoid race with control plane") | Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> | Signed-off-by: Florian Westphal <fw@strlen.de> | Acked-by: Michal Kubecek <mkubecek@suse.cz> Patch sorting ============= Patches added to the "sorted patches" section of series.conf must be sorted according to the upstream order of the commit that they backport. After you've added a patch file to the main patches.suse/ or a different patch directory, and supplemented the required tags described in the section "Patch Headers", run $ ./scripts/git_sort/series_insert.py <patch> to insert an entry for a new patch file to the sorted section of series.conf. For more information, please read "scripts/git_sort/README.md". Before You Commit -- Things To Check ==================================== Make sure that all patches still apply after your changes. One way of doing this is using scripts/sequence-patch.sh: $ export SCRATCH_AREA=/var/tmp/scratch $ ./scripts/sequence-patch.sh Creating tree in /var/tmp/scratch/linux-5.14-SLE15-SP5 Cleaning up from previous run Linking from /var/tmp/scratch/linux-5.14.orig ... [ Tree: /var/tmp/scratch/linux-5.14-SLE15-SP5 ] [ Generating Module.supported ] [ Copying config/x86_64/default ] Note the "Tree:" line output by the sequence-patch.sh script which specifies the location of the expanded kernel tree that is configured for local build. Please test-compile the kernel or even test-build kernel packages, depending on the impact of your changes. Use scripts/tar-up.sh for creating an OBS package directory. The kernel source tree that scripts/sequence-patch.sh creates can be test-compiled. Before that, make sure all prerequisites are installed. These include libopenssl-devel, libelf-devel and dwarves. Have a look into rpm/kernel-binary.spec.in for a complete list. Then, the compilation can be done as follows: $ cd /var/tmp/scratch/linux-5.14-SLE15-SP5 $ make oldconfig $ make When committing a patch series, try to make the series easily bisectable. In other words, when applying only the first x patches (1 <= x <= n, n being the number of patches in the series), the kernel should be still buildable and functional. This means especially that just adding upstream patches unmodified to a series and doing a cleanup patch at the end of the series to ensure the kernel is buildable and functional is to be avoided. Each patch from upstream should be modified as required to fit into the kernel it is backported to, both for build time and runtime. Applying all patches in the tree with scripts/sequence-patch.sh can take a significant amount of time. The --rapid option is present to speed up the process and tells the script to use Rapidquilt instead of the regular Quilt. The Rapidquilt implementation applies patches in parallel and typically produces an expanded tree in a fraction of the original time. A Rapidquilt package can be obtained from https://download.opensuse.org/repositories/Kernel:/tools/. Config Option Changes ===================== SUSE kernel packages for various architectures and configurations are built from the same sources. Each such kernel has its own configuration file in config/$ARCH/$FLAVOR. Checks are in place that abort the kernel build when those configuration files are missing necessary config options. When adding patches that introduce new kernel config options, please also update all config files as follows: $ scripts/sequence-patch.sh $ cd /var/tmp/scratch/linux-5.14-SLE15-SP5 $ patches/scripts/run_oldconfig.sh Committing and Log Messages =========================== Any commit which affects the kernel package (rather than internals of the repository such as helper scripts) should be documented in kernel-source.changes. The log entry must include the timestamp (`date`), email address of the committer, and a change summary, and should include the names of the affected files, as in the following example: | ------------------------------------------------------------------- | Wed Dec 1 18:29:44 CET 2004 - agruen@suse.de | | - patches.fixes/serialize-dgram-read.diff: Serialize dgram read | using semaphore just like stream (#48427). | There is a simple helper script for creating changelog entries in this format (/work/src/bin/vc). An advanced script (scripts/log) for creating changelog entries from patch headers, and then automatically committing the change, exists as well. This script extracts Subject: and References: headers from added or modified patches and generates a changelog entry proposal that the user can further modify. This approach works well for new or removed patches. When modifying existing patches, it usually is necessary to modify the generated changelog entry by hand. (scripts/log requires the vc helper script either in the PATH, or in /work/src/bin/). What Is The Kernel ABI? ======================= All symbols that the kernel exports for use by modules, and all symbols that modules export for use by other modules, are associated with a so-called modversion, which is a checksum of the type of the symbol (including all sub-types involved). Symbols that a module imports are associated with the identical checksum. When a module is loaded, the kernel makes sure that the checksums of the exported symbols match the checksums of the imported symbols. The kernel exports a large number of symbols (in the range of 5000). We could Provide and Require all those symbols as RPM package dependencies, so that then those dependencies would make sure that all packages containing kernel modules would have a matching kernel installed. RPM does not easily handle such large amounts of dependencies though, and so instead of modeling dependencies based on individual symbols, we compute distinct classes of symbols, and we create one "kernel(...)" RPM dependency per class. Modifying, adding, or removing kernel exports will also change the kernel(...) symbols. Kernel ABI Changes ================== During kernel builds, two things related to the kernel ABI happen: * If a reference symvers file (/boot/symvers-* in kernel-$FLAVOR packages) for the particular architecture and flavor is available, we check how severe the ABI changes are compared to this reference. These reference files are located in kabi/$ARCH/symvers-$FLAVOR. Too severe changes will abort the build. See rpm/kernel-binary.spec.in and scripts/kabi-checks (SLES9 - SLES10) or rpm/symsets.pl (SLES11) for details. * We want to avoid losing kernel(...) symbols when additional symbols are added, but all previous symbols are still available: in this case, all modules will continue to load into the new kernel just fine. If a reference symsets file (/boot/symsets-* in kernel-$FLAVOR packages) for the particular architecture and flavor is available, we check which of the symbol sets in the reference file can still be exported, even though symbols have meanwhile been added. We also export the kernel(...) symbols from reference symset files. To update the reference files, use scripts/update-symvers: $ ./scripts/update-symvers kernel-default-2.6.27.25-0.1.1.x86_64.rpm \ kernel-default-2.6.27.25-0.1.1.i586.rpm ... It is also possible to update only a subset of the symbols, e.g.: $ ./scripts/update-symvers --filter=fs/xfs ... In either case, ask the branch owner and kernel packager for permission before touching the kabi files. Ignoring Kernel ABI Changes --------------------------- Sometimes we want to tolerate particular kernel ABI changes (and not abort the builds). At the same time, we do not want to update the reference symvers and symsets files, because we still want to monitor the relative changes, and we want to continue preserving all symbol sets which are still compatible. Particular kernel can be marked so that kernel ABI changes are ignored. This is done by creating kabi/$ARCH/ignore-$FLAVOR files (e.g., kabi/x86_64/ignore-xen, kabi/s390/ignore-default). The kernel ABI checks will still be produced, but the build will not be aborted. The file contents d not matter. All kernel ABI changes in all kernel packages can be ignored by creating a file called IGNORE-KABI-BADNESS in the kernel-source/ sub-directory of the repository that scripts/tar-up.sh creates. (This may be necessary for PTF kernels occasionally.) Embargoed Patches ================= At certain times during development, the kernel may include "embargoed" patches, i.e., patches that must not be mad available to parties outside of SUSE before an agreed-upon time. Such patches usually have a date of publication that has been coordinated among linux distributors, etc. Such patches must not be committed to the usual branches, because these are pushed to a public mirror, but instead to a branch named with an _EMBARGO suffix, e.g. SLE11_BRANCH_EMBARGO. The KOTD scripts will testbuild such branches, but won't publish them. Once the fix becomes public, the branch needs to be merged back info the "mainline" branch, e.g.: $ git checkout SLE11_BRANCH $ git merge SLE11_BRANCH_EMBARGO Related Information =================== Internal: https://wiki.suse.net/index.php/SUSE-Labs_Publications/Kernel_Building https://wiki.suse.net/index.php/SUSE-Labs_Publications/kernel_patches_rules Public: TBD