1 .. SPDX-License-Identifier: GPL-2.0
8 KVM strives to be a welcoming community; contributions from newcomers are
9 valued and encouraged. Please do not be discouraged or intimidated by the
10 length of this document and the many rules/guidelines it contains. Everyone
11 makes mistakes, and everyone was a newbie at some point. So long as you make
12 an honest effort to follow KVM x86's guidelines, are receptive to feedback,
13 and learn from any mistakes you make, you will be welcomed with open arms, not
14 torches and pitchforks.
18 Testing is mandatory. Be consistent with established styles and patterns.
22 KVM x86 is currently in a transition period from being part of the main KVM
23 tree, to being "just another KVM arch". As such, KVM x86 is split across the
24 main KVM tree, ``git.kernel.org/pub/scm/virt/kvm/kvm.git``, and a KVM x86
25 specific tree, ``github.com/kvm-x86/linux.git``.
27 Generally speaking, fixes for the current cycle are applied directly to the
28 main KVM tree, while all development for the next cycle is routed through the
29 KVM x86 tree. In the unlikely event that a fix for the current cycle is routed
30 through the KVM x86 tree, it will be applied to the ``fixes`` branch before
31 making its way to the main KVM tree.
33 Note, this transition period is expected to last quite some time, i.e. will be
34 the status quo for the foreseeable future.
38 The KVM x86 tree is organized into multiple topic branches. The purpose of
39 using finer-grained topic branches is to make it easier to keep tabs on an area
40 of development, and to limit the collateral damage of human errors and/or buggy
41 commits, e.g. dropping the HEAD commit of a topic branch has no impact on other
42 in-flight commits' SHA1 hashes, and having to reject a pull request due to bugs
43 delays only that topic branch.
45 All topic branches, except for ``next`` and ``fixes``, are rolled into ``next``
46 via a Cthulhu merge on an as-needed basis, i.e. when a topic branch is updated.
47 As a result, force pushes to ``next`` are common.
51 Fixes that target the current release, a.k.a. mainline, are typically applied
52 directly to the main KVM tree, i.e. do not route through the KVM x86 tree.
54 Changes that target the next release are routed through the KVM x86 tree. Pull
55 requests (from KVM x86 to main KVM) are sent for each KVM x86 topic branch,
56 typically the week before Linus' opening of the merge window, e.g. the week
57 following rc7 for "normal" releases. If all goes well, the topic branches are
58 rolled into the main KVM pull request sent during Linus' merge window.
60 The KVM x86 tree doesn't have its own official merge window, but there's a soft
61 close around rc5 for new features, and a soft close around rc6 for fixes (for
62 the next release; see above for fixes that target the current release).
66 Submissions are typically reviewed and applied in FIFO order, with some wiggle
67 room for the size of a series, patches that are "cache hot", etc. Fixes,
68 especially for the current release and or stable trees, get to jump the queue.
69 Patches that will be taken through a non-KVM tree (most often through the tip
70 tree) and/or have other acks/reviews also jump the queue to some extent.
72 Note, the vast majority of review is done between rc1 and rc6, give or take.
73 The period between rc6 and the next rc1 is used to catch up on other tasks,
74 i.e. radio silence during this period isn't unusual.
76 Pings to get a status update are welcome, but keep in mind the timing of the
77 current release cycle and have realistic expectations. If you are pinging for
78 acceptance, i.e. not just for feedback or an update, please do everything you
79 can, within reason, to ensure that your patches are ready to be merged! Pings
80 on series that break the build or fail tests lead to unhappy maintainers!
87 Fixes that target the current release, a.k.a. mainline, should be based on
88 ``git://git.kernel.org/pub/scm/virt/kvm/kvm.git master``. Note, fixes do not
89 automatically warrant inclusion in the current release. There is no singular
90 rule, but typically only fixes for bugs that are urgent, critical, and/or were
91 introduced in the current release should target the current release.
93 Everything else should be based on ``kvm-x86/next``, i.e. there is no need to
94 select a specific topic branch as the base. If there are conflicts and/or
95 dependencies across topic branches, it is the maintainer's job to sort them
98 The only exception to using ``kvm-x86/next`` as the base is if a patch/series
99 is a multi-arch series, i.e. has non-trivial modifications to common KVM code
100 and/or has more than superficial changes to other architectures' code. Multi-
101 arch patch/series should instead be based on a common, stable point in KVM's
102 history, e.g. the release candidate upon which ``kvm-x86 next`` is based. If
103 you're unsure whether a patch/series is truly multi-arch, err on the side of
104 caution and treat it as multi-arch, i.e. use a common base.
108 When it comes to style, naming, patterns, etc., consistency is the number one
109 priority in KVM x86. If all else fails, match what already exists.
111 With a few caveats listed below, follow the tip tree maintainers' preferred
112 :ref:`maintainer-tip-coding-style`, as patches/series often touch both KVM and
113 non-KVM x86 files, i.e. draw the attention of KVM *and* tip tree maintainers.
115 Using reverse fir tree, a.k.a. reverse Christmas tree or reverse XMAS tree, for
116 variable declarations isn't strictly required, though it is still preferred.
118 Except for a handful of special snowflakes, do not use kernel-doc comments for
119 functions. The vast majority of "public" KVM functions aren't truly public as
120 they are intended only for KVM-internal consumption (there are plans to
121 privatize KVM's headers and exports to enforce this).
125 Write comments using imperative mood and avoid pronouns. Use comments to
126 provide a high level overview of the code, and/or to explain why the code does
127 what it does. Do not reiterate what the code literally does; let the code
128 speak for itself. If the code itself is inscrutable, comments will not help.
130 SDM and APM References
131 ~~~~~~~~~~~~~~~~~~~~~~
132 Much of KVM's code base is directly tied to architectural behavior defined in
133 Intel's Software Development Manual (SDM) and AMD's Architecture Programmer’s
134 Manual (APM). Use of "Intel's SDM" and "AMD's APM", or even just "SDM" or
135 "APM", without additional context is a-ok.
137 Do not reference specific sections, tables, figures, etc. by number, especially
138 not in comments. Instead, if necessary (see below), copy-paste the relevant
139 snippet and reference sections/tables/figures by name. The layouts of the SDM
140 and APM are constantly changing, and so the numbers/labels aren't stable.
142 Generally speaking, do not explicitly reference or copy-paste from the SDM or
143 APM in comments. With few exceptions, KVM *must* honor architectural behavior,
144 therefore it's implied that KVM behavior is emulating SDM and/or APM behavior.
145 Note, referencing the SDM/APM in changelogs to justify the change and provide
146 context is perfectly ok and encouraged.
150 The preferred prefix format is ``KVM: <topic>:``, where ``<topic>`` is one of::
162 **DO NOT use x86/kvm!** ``x86/kvm`` is used exclusively for Linux-as-a-KVM-guest
163 changes, i.e. for arch/x86/kernel/kvm.c. Do not use file names or complete file
164 paths as the subject/shortlog prefix.
166 Note, these don't align with the topics branches (the topic branches care much
167 more about code conflicts).
169 All names are case sensitive! ``KVM: x86:`` is good, ``kvm: vmx:`` is not.
171 Capitalize the first word of the condensed patch description, but omit ending
172 punctionation. E.g.::
174 KVM: x86: Fix a null pointer dereference in function_xyz()
178 kvm: x86: fix a null pointer dereference in function_xyz.
180 If a patch touches multiple topics, traverse up the conceptual tree to find the
181 first common parent (which is often simply ``x86``). When in doubt,
182 ``git log path/to/file`` should provide a reasonable hint.
184 New topics do occasionally pop up, but please start an on-list discussion if
185 you want to propose introducing a new topic, i.e. don't go rogue.
187 See :ref:`the_canonical_patch_format` for more information, with one amendment:
188 do not treat the 70-75 character limit as an absolute, hard limit. Instead,
189 use 75 characters as a firm-but-not-hard limit, and use 80 characters as a hard
190 limit. I.e. let the shortlog run a few characters over the standard limit if
191 you have good reason to do so.
195 Most importantly, write changelogs using imperative mood and avoid pronouns.
197 See :ref:`describe_changes` for more information, with one amendment: lead with
198 a short blurb on the actual changes, and then follow up with the context and
199 background. Note! This order directly conflicts with the tip tree's preferred
200 approach! Please follow the tip tree's preferred style when sending patches
201 that primarily target arch/x86 code that is _NOT_ KVM code.
203 Stating what a patch does before diving into details is preferred by KVM x86
204 for several reasons. First and foremost, what code is actually being changed
205 is arguably the most important information, and so that info should be easy to
206 find. Changelogs that bury the "what's actually changing" in a one-liner after
207 3+ paragraphs of background make it very hard to find that information.
209 For initial review, one could argue the "what's broken" is more important, but
210 for skimming logs and git archaeology, the gory details matter less and less.
211 E.g. when doing a series of "git blame", the details of each change along the
212 way are useless, the details only matter for the culprit. Providing the "what
213 changed" makes it easy to quickly determine whether or not a commit might be of
216 Another benefit of stating "what's changing" first is that it's almost always
217 possible to state "what's changing" in a single sentence. Conversely, all but
218 the most simple bugs require multiple sentences or paragraphs to fully describe
219 the problem. If both the "what's changing" and "what's the bug" are super
220 short then the order doesn't matter. But if one is shorter (almost always the
221 "what's changing), then covering the shorter one first is advantageous because
222 it's less of an inconvenience for readers/reviewers that have a strict ordering
223 preference. E.g. having to skip one sentence to get to the context is less
224 painful than having to skip three paragraphs to get to "what's changing".
228 If a change fixes a KVM/kernel bug, add a Fixes: tag even if the change doesn't
229 need to be backported to stable kernels, and even if the change fixes a bug in
232 Conversely, if a fix does need to be backported, explicitly tag the patch with
233 "Cc: stable@vger.kernel" (though the email itself doesn't need to Cc: stable);
234 KVM x86 opts out of backporting Fixes: by default. Some auto-selected patches
235 do get backported, but require explicit maintainer approval (search MANUALSEL).
239 When a function is mentioned in a comment, changelog, or shortlog (or anywhere
240 for that matter), use the format ``function_name()``. The parentheses provide
241 context and disambiguate the reference.
245 At a bare minimum, *all* patches in a series must build cleanly for KVM_INTEL=m
246 KVM_AMD=m, and KVM_WERROR=y. Building every possible combination of Kconfigs
247 isn't feasible, but the more the merrier. KVM_SMM, KVM_XEN, PROVE_LOCKING, and
248 X86_64 are particularly interesting knobs to turn.
250 Running KVM selftests and KVM-unit-tests is also mandatory (and stating the
251 obvious, the tests need to pass). The only exception is for changes that have
252 negligible probability of affecting runtime behavior, e.g. patches that only
253 modify comments. When possible and relevant, testing on both Intel and AMD is
254 strongly preferred. Booting an actual VM is encouraged, but not mandatory.
256 For changes that touch KVM's shadow paging code, running with TDP (EPT/NPT)
257 disabled is mandatory. For changes that affect common KVM MMU code, running
258 with TDP disabled is strongly encouraged. For all other changes, if the code
259 being modified depends on and/or interacts with a module param, testing with
260 the relevant settings is mandatory.
262 Note, KVM selftests and KVM-unit-tests do have known failures. If you suspect
263 a failure is not due to your changes, verify that the *exact same* failure
264 occurs with and without your changes.
266 Changes that touch reStructured Text documentation, i.e. .rst files, must build
267 htmldocs cleanly, i.e. with no new warnings or errors.
269 If you can't fully test a change, e.g. due to lack of hardware, clearly state
270 what level of testing you were able to do, e.g. in the cover letter.
274 With one exception, new features *must* come with test coverage. KVM specific
275 tests aren't strictly required, e.g. if coverage is provided by running a
276 sufficiently enabled guest VM, or by running a related kernel selftest in a VM,
277 but dedicated KVM tests are preferred in all cases. Negative testcases in
278 particular are mandatory for enabling of new hardware features as error and
279 exception flows are rarely exercised simply by running a VM.
281 The only exception to this rule is if KVM is simply advertising support for a
282 feature via KVM_GET_SUPPORTED_CPUID, i.e. for instructions/features that KVM
283 can't prevent a guest from using and for which there is no true enabling.
285 Note, "new features" does not just mean "new hardware features"! New features
286 that can't be well validated using existing KVM selftests and/or KVM-unit-tests
287 must come with tests.
289 Posting new feature development without tests to get early feedback is more
290 than welcome, but such submissions should be tagged RFC, and the cover letter
291 should clearly state what type of feedback is requested/expected. Do not abuse
292 the RFC process; RFCs will typically not receive in-depth review.
296 Except for "obvious" found-by-inspection bugs, fixes must be accompanied by a
297 reproducer for the bug being fixed. In many cases the reproducer is implicit,
298 e.g. for build errors and test failures, but it should still be clear to
299 readers what is broken and how to verify the fix. Some leeway is given for
300 bugs that are found via non-public workloads/tests, but providing regression
301 tests for such bugs is strongly preferred.
303 In general, regression tests are preferred for any bug that is not trivial to
304 hit. E.g. even if the bug was originally found by a fuzzer such as syzkaller,
305 a targeted regression test may be warranted if the bug requires hitting a
306 one-in-a-million type race condition.
308 Note, KVM bugs are rarely urgent *and* non-trivial to reproduce. Ask yourself
309 if a bug is really truly the end of the world before posting a fix without a
317 Do not explicitly reference bug reports, prior versions of a patch/series, etc.
318 via ``In-Reply-To:`` headers. Using ``In-Reply-To:`` becomes an unholy mess
319 for large series and/or when the version count gets high, and ``In-Reply-To:``
320 is useless for anyone that doesn't have the original message, e.g. if someone
321 wasn't Cc'd on the bug report or if the list of recipients changes between
324 To link to a bug report, previous version, or anything of interest, use lore
325 links. For referencing previous version(s), generally speaking do not include
326 a Link: in the changelog as there is no need to record the history in git, i.e.
327 put the link in the cover letter or in the section git ignores. Do provide a
328 formal Link: for bug reports and/or discussions that led to the patch. The
329 context of why a change was made is highly valuable for future readers.
333 If you are using git version 2.9.0 or later (Googlers, this is all of you!),
334 use ``git format-patch`` with the ``--base`` flag to automatically include the
335 base tree information in the generated patches.
337 Note, ``--base=auto`` works as expected if and only if a branch's upstream is
338 set to the base topic branch, e.g. it will do the wrong thing if your upstream
339 is set to your personal repository for backup purposes. An alternative "auto"
340 solution is to derive the names of your development branches based on their
341 KVM x86 topic, and feed that into ``--base``. E.g. ``x86/pmu/my_branch_name``,
342 and then write a small wrapper to extract ``pmu`` from the current branch name
343 to yield ``--base=x/pmu``, where ``x`` is whatever name your repository uses to
344 track the KVM x86 remote.
348 KVM selftests that are associated with KVM changes, e.g. regression tests for
349 bug fixes, should be posted along with the KVM changes as a single series. The
350 standard kernel rules for bisection apply, i.e. KVM changes that result in test
351 failures should be ordered after the selftests updates, and vice versa, new
352 tests that fail due to KVM bugs should be ordered after the KVM fixes.
354 KVM-unit-tests should *always* be posted separately. Tools, e.g. b4 am, don't
355 know that KVM-unit-tests is a separate repository and get confused when patches
356 in a series apply on different trees. To tie KVM-unit-tests patches back to
357 KVM patches, first post the KVM changes and then provide a lore Link: to the
358 KVM patch/series in the KVM-unit-tests patch(es).
362 When a patch/series is officially accepted, a notification email will be sent
363 in reply to the original posting (cover letter for multi-patch series). The
364 notification will include the tree and topic branch, along with the SHA1s of
365 the commits of applied patches.
367 If a subset of patches is applied, this will be clearly stated in the
368 notification. Unless stated otherwise, it's implied that any patches in the
369 series that were not accepted need more work and should be submitted in a new
372 If for some reason a patch is dropped after officially being accepted, a reply
373 will be sent to the notification email explaining why the patch was dropped, as
374 well as the next steps.
378 SHA1s are not 100% guaranteed to be stable until they land in Linus' tree! A
379 SHA1 is *usually* stable once a notification has been sent, but things happen.
380 In most cases, an update to the notification email be provided if an applied
381 patch's SHA1 changes. However, in some scenarios, e.g. if all KVM x86 branches
382 need to be rebased, individual notifications will not be given.
386 Bugs that can be exploited by the guest to attack the host (kernel or
387 userspace), or that can be exploited by a nested VM to *its* host (L2 attacking
388 L1), are of particular interest to KVM. Please follow the protocol for
389 :ref:`securitybugs` if you suspect a bug can lead to an escape, data leak, etc.
projects / linux-modified.git / blob