xref: /linux/Documentation/maintainer/feature-and-driver-maintainers.rst (revision 3d0fe49454652117522f60bfbefb978ba0e5300b)
1.. SPDX-License-Identifier: GPL-2.0
2
3==============================
4Feature and driver maintainers
5==============================
6
7The term "maintainer" spans a very wide range of levels of engagement
8from people handling patches and pull requests as almost a full time job
9to people responsible for a small feature or a driver.
10
11Unlike most of the chapter, this section is meant for the latter (more
12populous) group. It provides tips and describes the expectations and
13responsibilities of maintainers of a small(ish) section of the code.
14
15Drivers and alike most often do not have their own mailing lists and
16git trees but instead send and review patches on the list of a larger
17subsystem.
18
19Responsibilities
20================
21
22The amount of maintenance work is usually proportional to the size
23and popularity of the code base. Small features and drivers should
24require relatively small amount of care and feeding. Nonetheless
25when the work does arrive (in form of patches which need review,
26user bug reports etc.) it has to be acted upon promptly.
27Even when a particular driver only sees one patch a month, or a quarter,
28a subsystem could well have a hundred such drivers. Subsystem
29maintainers cannot afford to wait a long time to hear from reviewers.
30
31The exact expectations on the response time will vary by subsystem.
32The patch review SLA the subsystem had set for itself can sometimes
33be found in the subsystem documentation. Failing that as a rule of thumb
34reviewers should try to respond quicker than what is the usual patch
35review delay of the subsystem maintainer. The resulting expectations
36may range from two working days for fast-paced subsystems (e.g. networking)
37to as long as a few weeks in slower moving parts of the kernel.
38
39Mailing list participation
40--------------------------
41
42Linux kernel uses mailing lists as the primary form of communication.
43Maintainers must be subscribed and follow the appropriate subsystem-wide
44mailing list. Either by subscribing to the whole list or using more
45modern, selective setup like
46`lei <https://people.kernel.org/monsieuricon/lore-lei-part-1-getting-started>`_.
47
48Maintainers must know how to communicate on the list (plain text, no invasive
49legal footers, no top posting, etc.)
50
51Reviews
52-------
53
54Maintainers must review *all* patches touching exclusively their drivers,
55no matter how trivial. If the patch is a tree wide change and modifies
56multiple drivers - whether to provide a review is left to the maintainer.
57
58When there are multiple maintainers for a piece of code an ``Acked-by``
59or ``Reviewed-by`` tag (or review comments) from a single maintainer is
60enough to satisfy this requirement.
61
62If the review process or validation for a particular change will take longer
63than the expected review timeline for the subsystem, maintainer should
64reply to the submission indicating that the work is being done, and when
65to expect full results.
66
67Refactoring and core changes
68----------------------------
69
70Occasionally core code needs to be changed to improve the maintainability
71of the kernel as a whole. Maintainers are expected to be present and
72help guide and test changes to their code to fit the new infrastructure.
73
74Bug reports
75-----------
76
77Maintainers must ensure severe problems in their code reported to them
78are resolved in a timely manner: regressions, kernel crashes, kernel warnings,
79compilation errors, lockups, data loss, and other bugs of similar scope.
80
81Maintainers furthermore should respond to reports about other kinds of
82bugs as well, if the report is of reasonable quality or indicates a
83problem that might be severe -- especially if they have *Supported*
84status of the codebase in the MAINTAINERS file.
85
86Selecting the maintainer
87========================
88
89The previous section described the expectations of the maintainer,
90this section provides guidance on selecting one and describes common
91misconceptions.
92
93The author
94----------
95
96Most natural and common choice of a maintainer is the author of the code.
97The author is intimately familiar with the code, so it is the best person
98to take care of it on an ongoing basis.
99
100That said, being a maintainer is an active role. The MAINTAINERS file
101is not a list of credits (in fact a separate CREDITS file exists),
102it is a list of those who will actively help with the code.
103If the author does not have the time, interest or ability to maintain
104the code, a different maintainer must be selected.
105
106Multiple maintainers
107--------------------
108
109Modern best practices dictate that there should be at least two maintainers
110for any piece of code, no matter how trivial. It spreads the burden, helps
111people take vacations and prevents burnout, trains new members of
112the community etc. etc. Even when there is clearly one perfect candidate,
113another maintainer should be found.
114
115Maintainers must be human, therefore, it is not acceptable to add a mailing
116list or a group email as a maintainer. Trust and understanding are the
117foundation of kernel maintenance and one cannot build trust with a mailing
118list. Having a mailing list *in addition* to humans is perfectly fine.
119
120Corporate structures
121--------------------
122
123To an outsider the Linux kernel may resemble a hierarchical organization
124with Linus as the CEO. While the code flows in a hierarchical fashion,
125the corporate template does not apply here. Linux is an anarchy held
126together by (rarely expressed) mutual respect, trust and convenience.
127
128All that is to say that managers almost never make good maintainers.
129The maintainer position more closely matches an on-call rotation
130than a position of power.
131
132The following characteristics of a person selected as a maintainer
133are clear red flags:
134
135 - unknown to the community, never sent an email to the list before
136 - did not author any of the code
137 - (when development is contracted) works for a company which paid
138   for the development rather than the company which did the work
139
140Non compliance
141==============
142
143Subsystem maintainers may remove inactive maintainers from the MAINTAINERS
144file. If the maintainer was a significant author or played an important
145role in the development of the code, they should be moved to the CREDITS file.
146
147Removing an inactive maintainer should not be seen as a punitive action.
148Having an inactive maintainer has a real cost as all developers have
149to remember to include the maintainers in discussions and subsystem
150maintainers spend brain power figuring out how to solicit feedback.
151
152Subsystem maintainers may remove code for lacking maintenance.
153
154Subsystem maintainers may refuse accepting code from companies
155which repeatedly neglected their maintainership duties.
156