1.. SPDX-License-Identifier: GPL-2.0
2
3====================
4Rebasing and merging
5====================
6
7Maintaining a subsystem, as a general rule, requires a familiarity with the
8Git source-code management system.  Git is a powerful tool with a lot of
9features; as is often the case with such tools, there are right and wrong
10ways to use those features.  This document looks in particular at the use
11of rebasing and merging.  Maintainers often get in trouble when they use
12those tools incorrectly, but avoiding problems is not actually all that
13hard.
14
15One thing to be aware of in general is that, unlike many other projects,
16the kernel community is not scared by seeing merge commits in its
17development history.  Indeed, given the scale of the project, avoiding
18merges would be nearly impossible.  Some problems encountered by
19maintainers result from a desire to avoid merges, while others come from
20merging a little too often.
21
22Rebasing
23========
24
25"Rebasing" is the process of changing the history of a series of commits
26within a repository.  There are two different types of operations that are
27referred to as rebasing since both are done with the ``git rebase``
28command, but there are significant differences between them:
29
30 - Changing the parent (starting) commit upon which a series of patches is
31   built.  For example, a rebase operation could take a patch set built on
32   the previous kernel release and base it, instead, on the current
33   release.  We'll call this operation "reparenting" in the discussion
34   below.
35
36 - Changing the history of a set of patches by fixing (or deleting) broken
37   commits, adding patches, adding tags to commit changelogs, or changing
38   the order in which commits are applied.  In the following text, this
39   type of operation will be referred to as "history modification"
40
41The term "rebasing" will be used to refer to both of the above operations.
42Used properly, rebasing can yield a cleaner and clearer development
43history; used improperly, it can obscure that history and introduce bugs.
44
45There are a few rules of thumb that can help developers to avoid the worst
46perils of rebasing:
47
48 - History that has been exposed to the world beyond your private system
49   should usually not be changed.  Others may have pulled a copy of your
50   tree and built on it; modifying your tree will create pain for them.  If
51   work is in need of rebasing, that is usually a sign that it is not yet
52   ready to be committed to a public repository.
53
54   That said, there are always exceptions.  Some trees (linux-next being
55   a significant example) are frequently rebased by their nature, and
56   developers know not to base work on them.  Developers will sometimes
57   expose an unstable branch for others to test with or for automated
58   testing services.  If you do expose a branch that may be unstable in
59   this way, be sure that prospective users know not to base work on it.
60
61 - Do not rebase a branch that contains history created by others.  If you
62   have pulled changes from another developer's repository, you are now a
63   custodian of their history.  You should not change it.  With few
64   exceptions, for example, a broken commit in a tree like this should be
65   explicitly reverted rather than disappeared via history modification.
66
67 - Do not reparent a tree without a good reason to do so.  Just being on a
68   newer base or avoiding a merge with an upstream repository is not
69   generally a good reason.
70
71 - If you must reparent a repository, do not pick some random kernel commit
72   as the new base.  The kernel is often in a relatively unstable state
73   between release points; basing development on one of those points
74   increases the chances of running into surprising bugs.  When a patch
75   series must move to a new base, pick a stable point (such as one of
76   the -rc releases) to move to.
77
78 - Realize that reparenting a patch series (or making significant history
79   modifications) changes the environment in which it was developed and,
80   likely, invalidates much of the testing that was done.  A reparented
81   patch series should, as a general rule, be treated like new code and
82   retested from the beginning.
83
84A frequent cause of merge-window trouble is when Linus is presented with a
85patch series that has clearly been reparented, often to a random commit,
86shortly before the pull request was sent.  The chances of such a series
87having been adequately tested are relatively low - as are the chances of
88the pull request being acted upon.
89
90If, instead, rebasing is limited to private trees, commits are based on a
91well-known starting point, and they are well tested, the potential for
92trouble is low.
93
94Merging
95=======
96
97Merging is a common operation in the kernel development process; the 5.1
98development cycle included 1,126 merge commits - nearly 9% of the total.
99Kernel work is accumulated in over 100 different subsystem trees, each of
100which may contain multiple topic branches; each branch is usually developed
101independently of the others.  So naturally, at least one merge will be
102required before any given branch finds its way into an upstream repository.
103
104Many projects require that branches in pull requests be based on the
105current trunk so that no merge commits appear in the history.  The kernel
106is not such a project; any rebasing of branches to avoid merges will, most
107likely, lead to trouble.
108
109Subsystem maintainers find themselves having to do two types of merges:
110from lower-level subsystem trees and from others, either sibling trees or
111the mainline.  The best practices to follow differ in those two situations.
112
113Merging from lower-level trees
114------------------------------
115
116Larger subsystems tend to have multiple levels of maintainers, with the
117lower-level maintainers sending pull requests to the higher levels.  Acting
118on such a pull request will almost certainly generate a merge commit; that
119is as it should be.  In fact, subsystem maintainers may want to use
120the --no-ff flag to force the addition of a merge commit in the rare cases
121where one would not normally be created so that the reasons for the merge
122can be recorded.  The changelog for the merge should, for any kind of
123merge, say *why* the merge is being done.  For a lower-level tree, "why" is
124usually a summary of the changes that will come with that pull.
125
126Maintainers at all levels should be using signed tags on their pull
127requests, and upstream maintainers should verify the tags when pulling
128branches.  Failure to do so threatens the security of the development
129process as a whole.
130
131As per the rules outlined above, once you have merged somebody else's
132history into your tree, you cannot rebase that branch, even if you
133otherwise would be able to.
134
135Merging from sibling or upstream trees
136--------------------------------------
137
138While merges from downstream are common and unremarkable, merges from other
139trees tend to be a red flag when it comes time to push a branch upstream.
140Such merges need to be carefully thought about and well justified, or
141there's a good chance that a subsequent pull request will be rejected.
142
143It is natural to want to merge the master branch into a repository; this
144type of merge is often called a "back merge".  Back merges can help to make
145sure that there are no conflicts with parallel development and generally
146gives a warm, fuzzy feeling of being up-to-date.  But this temptation
147should be avoided almost all of the time.
148
149Why is that?  Back merges will muddy the development history of your own
150branch.  They will significantly increase your chances of encountering bugs
151from elsewhere in the community and make it hard to ensure that the work
152you are managing is stable and ready for upstream.  Frequent merges can
153also obscure problems with the development process in your tree; they can
154hide interactions with other trees that should not be happening (often) in
155a well-managed branch.
156
157That said, back merges are occasionally required; when that happens, be
158sure to document *why* it was required in the commit message.  As always,
159merge to a well-known stable point, rather than to some random commit.
160Even then, you should not back merge a tree above your immediate upstream
161tree; if a higher-level back merge is really required, the upstream tree
162should do it first.
163
164One of the most frequent causes of merge-related trouble is when a
165maintainer merges with the upstream in order to resolve merge conflicts
166before sending a pull request.  Again, this temptation is easy enough to
167understand, but it should absolutely be avoided.  This is especially true
168for the final pull request: Linus is adamant that he would much rather see
169merge conflicts than unnecessary back merges.  Seeing the conflicts lets
170him know where potential problem areas are.  He does a lot of merges (382
171in the 5.1 development cycle) and has gotten quite good at conflict
172resolution - often better than the developers involved.
173
174So what should a maintainer do when there is a conflict between their
175subsystem branch and the mainline?  The most important step is to warn
176Linus in the pull request that the conflict will happen; if nothing else,
177that demonstrates an awareness of how your branch fits into the whole.  For
178especially difficult conflicts, create and push a *separate* branch to show
179how you would resolve things.  Mention that branch in your pull request,
180but the pull request itself should be for the unmerged branch.
181
182Even in the absence of known conflicts, doing a test merge before sending a
183pull request is a good idea.  It may alert you to problems that you somehow
184didn't see from linux-next and helps to understand exactly what you are
185asking upstream to do.
186
187Another reason for doing merges of upstream or another subsystem tree is to
188resolve dependencies.  These dependency issues do happen at times, and
189sometimes a cross-merge with another tree is the best way to resolve them;
190as always, in such situations, the merge commit should explain why the
191merge has been done.  Take a moment to do it right; people will read those
192changelogs.
193
194Often, though, dependency issues indicate that a change of approach is
195needed.  Merging another subsystem tree to resolve a dependency risks
196bringing in other bugs and should almost never be done.  If that subsystem
197tree fails to be pulled upstream, whatever problems it had will block the
198merging of your tree as well.  Preferable alternatives include agreeing
199with the maintainer to carry both sets of changes in one of the trees or
200creating a topic branch dedicated to the prerequisite commits that can be
201merged into both trees.  If the dependency is related to major
202infrastructural changes, the right solution might be to hold the dependent
203commits for one development cycle so that those changes have time to
204stabilize in the mainline.
205
206Finally
207=======
208
209It is relatively common to merge with the mainline toward the beginning of
210the development cycle in order to pick up changes and fixes done elsewhere
211in the tree.  As always, such a merge should pick a well-known release
212point rather than some random spot.  If your upstream-bound branch has
213emptied entirely into the mainline during the merge window, you can pull it
214forward with a command like::
215
216  git merge --ff-only v5.2-rc1
217
218The guidelines laid out above are just that: guidelines.  There will always
219be situations that call out for a different solution, and these guidelines
220should not prevent developers from doing the right thing when the need
221arises.  But one should always think about whether the need has truly
222arisen and be prepared to explain why something abnormal needs to be done.
223