docs: move `Fixes:` tag explanation to its own section
[mesa.git] / docs / submittingpatches.rst
1 Submitting Patches
2 ==================
3
4 Basic guidelines
5 ----------------
6
7 - Patches should not mix code changes with code formatting changes
8 (except, perhaps, in very trivial cases.)
9 - Code patches should follow Mesa `coding
10 conventions <codingstyle.rst>`__.
11 - Whenever possible, patches should only affect individual Mesa/Gallium
12 components.
13 - Patches should never introduce build breaks and should be bisectable
14 (see ``git bisect``.)
15 - Patches should be properly `formatted <#formatting>`__.
16 - Patches should be sufficiently `tested <#testing>`__ before
17 submitting.
18 - Patches should be `submitted <#submit>`__ via a merge request for
19 `review <#reviewing>`__.
20
21 .. _formatting:
22
23 Patch formatting
24 ----------------
25
26 - Lines should be limited to 75 characters or less so that git logs
27 displayed in 80-column terminals avoid line wrapping. Note that git
28 log uses 4 spaces of indentation (4 + 75 < 80).
29 - The first line should be a short, concise summary of the change
30 prefixed with a module name. Examples:
31
32 ::
33
34 mesa: Add support for querying GL_VERTEX_ATTRIB_ARRAY_LONG
35
36 gallium: add PIPE_CAP_DEVICE_RESET_STATUS_QUERY
37
38 i965: Fix missing type in local variable declaration.
39
40 - Subsequent patch comments should describe the change in more detail,
41 if needed. For example:
42
43 ::
44
45 i965: Remove end-of-thread SEND alignment code.
46
47 This was present in Eric's initial implementation of the compaction code
48 for Sandybridge (commit 077d01b6). There is no documentation saying this
49 is necessary, and removing it causes no regressions in piglit on any
50 platform.
51
52 - A "Signed-off-by:" line is not required, but not discouraged either.
53 - If a patch addresses an issue in gitlab, use the Closes: tag For
54 example:
55
56 ::
57
58 Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/1
59
60 Prefer the full url to just ``Closes: #1``, since the url makes it
61 easier to get to the bug page from ``git log``
62
63 **Do not use the ``Fixes:`` tag for this!** Mesa already uses
64 ``Fixes:`` for something else.
65 See `below <#fixes>`__.
66
67 - If there have been several revisions to a patch during the review
68 process, they should be noted such as in this example:
69
70 ::
71
72 st/mesa: add ARB_texture_stencil8 support (v4)
73
74 if we support stencil texturing, enable texture_stencil8
75 there is no requirement to support native S8 for this,
76 the texture can be converted to x24s8 fine.
77
78 v2: fold fixes from Marek in:
79 a) put S8 last in the list
80 b) fix renderable to always test for d/s renderable
81 fixup the texture case to use a stencil only format
82 for picking the format for the texture view.
83 v3: hit fallback for getteximage
84 v4: put s8 back in front, it shouldn't get picked now (Ilia)
85
86 - If someone tested your patch, document it with a line like this:
87
88 ::
89
90 Tested-by: Joe Hacker <jhacker@foo.com>
91
92 - If the patch was reviewed (usually the case) or acked by someone,
93 that should be documented with:
94
95 ::
96
97 Reviewed-by: Joe Hacker <jhacker@foo.com>
98 Acked-by: Joe Hacker <jhacker@foo.com>
99
100 - When updating a merge request add all the tags (``Acked-by:``, ``Reviewed-by:``,
101 ``Fixes:``, ``Cc: mesa-stable`` and/or other) to the commit messages.
102 This provides reviewers with quick feedback if the patch has already
103 been reviewed.
104
105 .. _fixes:
106
107 The ``Fixes:`` tag
108 ------------------
109
110 If a patch addresses a issue introduced with earlier commit, that
111 should be noted in the commit message. For example::
112
113 Fixes: d7b3707c612 "util/disk_cache: use stat() to check if entry is a directory"
114
115 You can produce those fixes lines by running this command once::
116
117 git config --global alias.fixes "show -s --pretty='format:Fixes: %h (\"%s\")'"
118
119 After that, using ``git fixes <sha1>`` will print the full line for you.
120
121 .. _testing:
122
123 Testing Patches
124 ---------------
125
126 It should go without saying that patches must be tested. In general, do
127 whatever testing is prudent.
128
129 You should always run the Mesa test suite before submitting patches. The
130 test suite can be run using the 'meson test' command. All tests must
131 pass before patches will be accepted, this may mean you have to update
132 the tests themselves.
133
134 Whenever possible and applicable, test the patch with
135 `Piglit <https://piglit.freedesktop.org>`__ and/or
136 `dEQP <https://android.googlesource.com/platform/external/deqp/>`__ to
137 check for regressions.
138
139 As mentioned at the beginning, patches should be bisectable. A good way
140 to test this is to make use of the \`git rebase\` command, to run your
141 tests on each commit. Assuming your branch is based off
142 ``origin/master``, you can run:
143
144 ::
145
146 $ git rebase --interactive --exec "meson test -C build/" origin/master
147
148 replacing ``"meson test"`` with whatever other test you want to run.
149
150 .. _submit:
151
152 Submitting Patches
153 ------------------
154
155 Patches are submitted to the Mesa project via a
156 `GitLab <https://gitlab.freedesktop.org/mesa/mesa>`__ Merge Request.
157
158 Add labels to your MR to help reviewers find it. For example:
159
160 - Mesa changes affecting all drivers: mesa
161 - Hardware vendor specific code: amd, intel, nvidia, ...
162 - Driver specific code: anvil, freedreno, i965, iris, radeonsi, radv,
163 vc4, ...
164 - Other tag examples: gallium, util
165
166 Tick the following when creating the MR. It allows developers to rebase
167 your work on top of master.
168
169 ::
170
171 Allow commits from members who can merge to the target branch
172
173 If you revise your patches based on code review and push an update to
174 your branch, you should maintain a **clean** history in your patches.
175 There should not be "fixup" patches in the history. The series should be
176 buildable and functional after every commit whenever you push the
177 branch.
178
179 It is your responsibility to keep the MR alive and making progress, as
180 there are no guarantees that a Mesa dev will independently take interest
181 in it.
182
183 Some other notes:
184
185 - Make changes and update your branch based on feedback
186 - After an update, for the feedback you handled, close the feedback
187 discussion with the "Resolve Discussion" button. This way the
188 reviewers know which feedback got handled and which didn't.
189 - Old, stale MR may be closed, but you can reopen it if you still want
190 to pursue the changes
191 - You should periodically check to see if your MR needs to be rebased
192 - Make sure your MR is closed if your patches get pushed outside of
193 GitLab
194 - Please send MRs from a personal fork rather than from the main Mesa
195 repository, as it clutters it unnecessarily.
196
197 .. _reviewing:
198
199 Reviewing Patches
200 -----------------
201
202 To participate in code review, you can monitor the GitLab Mesa `Merge
203 Requests <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests>`__
204 page, and/or register for notifications in your gitlab settings.
205
206 When you've reviewed a patch, please be unambiguous about your review.
207 That is, state either
208
209 ::
210
211 Reviewed-by: Joe Hacker <jhacker@foo.com>
212
213 or
214
215 ::
216
217 Acked-by: Joe Hacker <jhacker@foo.com>
218
219 Rather than saying just "LGTM" or "Seems OK".
220
221 If small changes are suggested, it's OK to say something like:
222
223 ::
224
225 With the above fixes, Reviewed-by: Joe Hacker <jhacker@foo.com>
226
227 which tells the patch author that the patch can be committed, as long as
228 the issues are resolved first.
229
230 These Reviewed-by, Acked-by, and Tested-by tags should also be amended
231 into commits in a MR before it is merged.
232
233 When providing a Reviewed-by, Acked-by, or Tested-by tag in a gitlab MR,
234 enclose the tag in backticks:
235
236 ::
237
238 `Reviewed-by: Joe Hacker <jhacker@example.com>`
239
240 This is the markdown format for literal, and will prevent gitlab from
241 hiding the < and > symbols.
242
243 Review by non-experts is encouraged. Understanding how someone else goes
244 about solving a problem is a great way to learn your way around the
245 project. The submitter is expected to evaluate whether they have an
246 appropriate amount of review feedback from people who also understand
247 the code before merging their patches.
248
249 Nominating a commit for a stable branch
250 ---------------------------------------
251
252 There are several ways to nominate a patch for inclusion in the stable
253 branch and release. In order or preference:
254
255 - By adding the ``Fixes:`` tag in the commit message as described above, if you are fixing
256 a specific commit.
257 - By adding the ``Cc: mesa-stable`` tag in the commit message as described below.
258 - By submitting a merge request against the ``staging/year.quarter``
259 branch on gitlab.
260
261 Please **DO NOT** send patches to mesa-stable@lists.freedesktop.org, it
262 is not monitored actively and is a historical artifact.
263
264 If you are not the author of the original patch, please Cc: them in your
265 nomination request.
266
267 The current patch status can be observed in the `staging
268 branch <releasing.rst#stagingbranch>`__.
269
270 The stable tag
271 ~~~~~~~~~~~~~~
272
273 If you want a commit to be applied to a stable branch, you should add an
274 appropriate note to the commit message.
275
276 Using a ``Fixes:`` tag as described in `Patch formatting <#formatting>`__
277 is the preferred way to nominate a commit that should be backported.
278 There are scripts that will figure out which releases to apply the patch
279 to automatically, so you don't need to figure it out.
280
281 Alternatively, you may use a "CC:" tag. Here are some examples of such a
282 note:
283
284 ::
285
286 CC: 20.0 19.3 <mesa-stable>
287
288 Using the CC tag **should** include the stable branches you want to
289 nominate the patch to. If you do not provide any version it is nominated
290 to all active stable branches.
291
292 .. _criteria:
293
294 Criteria for accepting patches to the stable branch
295 ---------------------------------------------------
296
297 Mesa has a designated release manager for each stable branch, and the
298 release manager is the only developer that should be pushing changes to
299 these branches. Everyone else should nominate patches using the
300 mechanism described above. The following rules define which patches are
301 accepted and which are not. The stable-release manager is also given
302 broad discretion in rejecting patches that have been nominated.
303
304 - Patch must conform with the `Basic guidelines <#guidelines>`__
305 - Patch must have landed in master first. In case where the original
306 patch is too large and/or otherwise contradicts with the rules set
307 within, a backport is appropriate.
308 - It must not introduce a regression - be that build or runtime wise.
309
310 .. note::
311 If the regression is due to faulty piglit/dEQP/CTS/other test
312 the latter must be fixed first. A reference to the offending test(s)
313 and respective fix(es) should be provided in the nominated patch.
314
315 - Patch cannot be larger than 100 lines.
316 - Patches that move code around with no functional change should be
317 rejected.
318 - Patch must be a bug fix and not a new feature.
319
320 .. note::
321 An exception to this rule, are hardware-enabling "features". For
322 example, `backports <#backports>`__ of new code to support a
323 newly-developed hardware product can be accepted if they can be
324 reasonably determined not to have effects on other hardware.
325
326 - Patch must be reviewed, For example, the commit message has
327 Reviewed-by, Signed-off-by, or Tested-by tags from someone but the
328 author.
329 - Performance patches are considered only if they provide information
330 about the hardware, program in question and observed improvement. Use
331 numbers to represent your measurements.
332
333 If the patch complies with the rules it will be
334 `cherry-picked <releasing.rst#pickntest>`__. Alternatively the release
335 manager will reply to the patch in question stating why the patch has
336 been rejected or would request a backport. The stable-release manager
337 may at times need to force-push changes to the stable branches, for
338 example, to drop a previously-picked patch that was later identified as
339 causing a regression). These force-pushes may cause changes to be lost
340 from the stable branch if developers push things directly. Consider
341 yourself warned.
342
343 .. _backports:
344
345 Sending backports for the stable branch
346 ---------------------------------------
347
348 By default merge conflicts are resolved by the stable-release manager.
349 The release maintainer should resolve trivial conflicts, but for complex
350 conflicts they should ask the original author to provide a backport or
351 de-nominate the patch.
352
353 For patches that either need to be nominated after they've landed in
354 master, or that are known ahead of time to not not apply cleanly to a
355 stable branch (such as due to a rename), using a gitlab MR is most
356 appropriate. The MR should be based on and target the
357 staging/year.quarter branch, not on the year.quarter branch, per the
358 stable branch policy. Assigning the MR to release maintainer for said
359 branch or mentioning them is helpful, but not required.
360
361 Git tips
362 --------
363
364 - ``git rebase -i ...`` is your friend. Don't be afraid to use it.
365 - Apply a fixup to commit FOO.
366
367 .. code-block:: console
368
369 git add ...
370 git commit --fixup=FOO
371 git rebase -i --autosquash ...
372
373 - Test for build breakage between patches e.g last 8 commits.
374
375 .. code-block:: console
376
377 git rebase -i --exec="ninja -C build/" HEAD~8