docs/submittingpatches.html

   1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
   2 <html lang="en">
   3 <head>
   4   <meta http-equiv="content-type" content="text/html; charset=utf-8">
   5   <title>Submitting patches</title>
   6   <link rel="stylesheet" type="text/css" href="mesa.css">
   7 </head>
   8 <body>
   9
  10 <div class="header">
  11   <h1>The Mesa 3D Graphics Library</h1>
  12 </div>
  13
  14 <iframe src="contents.html"></iframe>
  15 <div class="content">
  16
  17 <h1>Submitting patches</h1>
  18
  19
  20 <ul>
  21 <li><a href="#guidelines">Basic guidelines</a>
  22 <li><a href="#formatting">Patch formatting</a>
  23 <li><a href="#testing">Testing Patches</a>
  24 <li><a href="#submit">Submitting Patches</a>
  25 <li><a href="#reviewing">Reviewing Patches</a>
  26 <li><a href="#nominations">Nominating a commit for a stable branch</a>
  27 <li><a href="#criteria">Criteria for accepting patches to the stable branch</a>
  28 <li><a href="#backports">Sending backports for the stable branch</a>
  29 <li><a href="#gittips">Git tips</a>
  30 </ul>
  31
  32 <h2 id="guidelines">Basic guidelines</h2>
  33
  34 <ul>
  35 <li>Patches should not mix code changes with code formatting changes (except,
  36 perhaps, in very trivial cases.)
  37 <li>Code patches should follow Mesa
  38 <a href="codingstyle.html" target="_parent">coding conventions</a>.
  39 <li>Whenever possible, patches should only affect individual Mesa/Gallium
  40 components.
  41 <li>Patches should never introduce build breaks and should be bisectable (see
  42 <code>git bisect</code>.)
  43 <li>Patches should be properly <a href="#formatting">formatted</a>.
  44 <li>Patches should be sufficiently <a href="#testing">tested</a> before submitting.
  45 <li>Patches should be <a href="#submit">submitted</a>
  46 to <a href="#mailing">mesa-dev</a> or with
  47 a <a href="#merge-request">merge request</a>
  48 for <a href="#reviewing">review</a>.
  49
  50 </ul>
  51
  52 <h2 id="formatting">Patch formatting</h2>
  53
  54 <ul>
  55 <li>Lines should be limited to 75 characters or less so that git logs
  56 displayed in 80-column terminals avoid line wrapping.  Note that git
  57 log uses 4 spaces of indentation (4 + 75 &lt; 80).
  58 <li>The first line should be a short, concise summary of the change prefixed
  59 with a module name.  Examples:
  60 <pre>
  61     mesa: Add support for querying GL_VERTEX_ATTRIB_ARRAY_LONG
  62
  63     gallium: add PIPE_CAP_DEVICE_RESET_STATUS_QUERY
  64
  65     i965: Fix missing type in local variable declaration.
  66 </pre>
  67 <li>Subsequent patch comments should describe the change in more detail,
  68 if needed.  For example:
  69 <pre>
  70     i965: Remove end-of-thread SEND alignment code.
  71
  72     This was present in Eric's initial implementation of the compaction code
  73     for Sandybridge (commit 077d01b6). There is no documentation saying this
  74     is necessary, and removing it causes no regressions in piglit on any
  75     platform.
  76 </pre>
  77 <li>A "Signed-off-by:" line is not required, but not discouraged either.
  78 <li>If a patch addresses a bugzilla issue, that should be noted in the
  79 patch comment.  For example:
  80 <pre>
  81    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89689
  82 </pre>
  83 <li>If a patch addresses a issue introduced with earlier commit, that should be
  84 noted in the patch comment.  For example:
  85 <pre>
  86    Fixes: d7b3707c612 "util/disk_cache: use stat() to check if entry is a directory"
  87 </pre>
  88 <li>If there have been several revisions to a patch during the review
  89 process, they should be noted such as in this example:
  90 <pre>
  91     st/mesa: add ARB_texture_stencil8 support (v4)
  92
  93     if we support stencil texturing, enable texture_stencil8
  94     there is no requirement to support native S8 for this,
  95     the texture can be converted to x24s8 fine.
  96
  97     v2: fold fixes from Marek in:
  98        a) put S8 last in the list
  99        b) fix renderable to always test for d/s renderable
 100         fixup the texture case to use a stencil only format
 101         for picking the format for the texture view.
 102     v3: hit fallback for getteximage
 103     v4: put s8 back in front, it shouldn't get picked now (Ilia)
 104 </pre>
 105 <li>If someone tested your patch, document it with a line like this:
 106 <pre>
 107     Tested-by: Joe Hacker &lt;jhacker@foo.com&gt;
 108 </pre>
 109 <li>If the patch was reviewed (usually the case) or acked by someone,
 110 that should be documented with:
 111 <pre>
 112     Reviewed-by: Joe Hacker &lt;jhacker@foo.com&gt;
 113     Acked-by: Joe Hacker &lt;jhacker@foo.com&gt;
 114 </pre>
 115 <li>If sending later revision of a patch, add all the tags - ack, r-b,
 116 Cc: mesa-stable and/or other. This provides reviewers with quick feedback if the
 117 patch has already been reviewed.
 118 <li>In order for your patch to reach the prospective reviewer easier/faster,
 119 use the script scripts/get_reviewer.pl to get a list of individuals and include
 120 them in the CC list.
 121 <br>
 122 Please use common sense and do <strong>not</strong> blindly add everyone.
 123 <br>
 124 <pre>
 125     $ scripts/get_reviewer.pl --help # to get the help screen
 126     $ scripts/get_reviewer.pl -f src/egl/drivers/dri2/platform_android.c
 127     Rob Herring &lt;robh@kernel.org&gt; (reviewer:ANDROID EGL SUPPORT,added_lines:188/700=27%,removed_lines:58/283=20%)
 128     Tomasz Figa &lt;tfiga@chromium.org&gt; (reviewer:ANDROID EGL SUPPORT,authored:12/41=29%,added_lines:308/700=44%,removed_lines:115/283=41%)
 129     Emil Velikov &lt;emil.l.velikov@gmail.com&gt; (authored:13/41=32%,removed_lines:76/283=27%)
 130 </pre>
 131 </ul>
 132
 133
 134
 135 <h2 id="testing">Testing Patches</h2>
 136
 137 <p>
 138 It should go without saying that patches must be tested.  In general,
 139 do whatever testing is prudent.
 140 </p>
 141
 142 <p>
 143 You should always run the Mesa test suite before submitting patches.
 144 The test suite can be run using the 'make check' command. All tests
 145 must pass before patches will be accepted, this may mean you have
 146 to update the tests themselves.
 147 </p>
 148
 149 <p>
 150 Whenever possible and applicable, test the patch with
 151 <a href="https://piglit.freedesktop.org">Piglit</a> and/or
 152 <a href="https://android.googlesource.com/platform/external/deqp/">dEQP</a>
 153 to check for regressions.
 154 </p>
 155
 156 <p>
 157 As mentioned at the begining, patches should be bisectable.
 158 A good way to test this is to make use of the `git rebase` command,
 159 to run your tests on each commit. Assuming your branch is based off
 160 <code>origin/master</code>, you can run:
 161 </p>
 162 <pre>
 163 $ git rebase --interactive --exec "make check" origin/master
 164 </pre>
 165 <p>
 166 replacing <code>"make check"</code> with whatever other test you want to
 167 run.
 168 </p>
 169
 170
 171 <h2 id="submit">Submitting Patches</h2>
 172
 173 <p>
 174 Patches may be submitted to the Mesa project by
 175 <a href="#mailing">email</a> or with a
 176 GitLab <a href="#merge-request">merge request</a>. To prevent
 177 duplicate code review, only use one method to submit your changes.
 178 </p>
 179
 180 <h3 id="mailing">Mailing Patches</h3>
 181
 182 <p>
 183 Patches may be sent to the mesa-dev mailing list for review:
 184 <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev">
 185 mesa-dev@lists.freedesktop.org</a>.
 186 When submitting a patch make sure to use
 187 <a href="https://git-scm.com/docs/git-send-email">git send-email</a>
 188 rather than attaching patches to emails. Sending patches as
 189 attachments prevents people from being able to provide in-line review
 190 comments.
 191 </p>
 192
 193 <p>
 194 When submitting follow-up patches you can use --in-reply-to to make v2, v3,
 195 etc patches show up as replies to the originals. This usually works well
 196 when you're sending out updates to individual patches (as opposed to
 197 re-sending the whole series). Using --in-reply-to makes
 198 it harder for reviewers to accidentally review old patches.
 199 </p>
 200
 201 <p>
 202 When submitting follow-up patches you should also login to
 203 <a href="https://patchwork.freedesktop.org">patchwork</a> and change the
 204 state of your old patches to Superseded.
 205 </p>
 206
 207 <p>
 208 Some companies' mail server automatically append a legal disclaimer,
 209 usually containing something along the lines of "The information in this
 210 email is confidential" and "distribution is strictly prohibited".<br/>
 211 These legal notices prevent us from being able to accept your patch,
 212 rendering the whole process pointless. Please make sure these are
 213 disabled before sending your patches. (Note that you may need to contact
 214 your email administrator for this.)
 215 </p>
 216
 217 <h3 id="merge-request">GitLab Merge Requests</h3>
 218
 219 <p>
 220   <a href="https://gitlab.freedesktop.org/mesa/mesa">GitLab</a> Merge
 221   Requests (MR) can also be used to submit patches for Mesa.
 222 </p>
 223
 224 <p>
 225   If the MR may have interest for most of the Mesa community, you can
 226   send an email to the mesa-dev email list including a link to the MR.
 227   Don't send the patch to mesa-dev, just the MR link.
 228 </p>
 229 <p>
 230   Add labels to your MR to help reviewers find it. For example:
 231   <ul>
 232     <li>Mesa changes affecting all drivers: mesa
 233     <li>Hardware vendor specific code: amd, intel, nvidia, ...
 234     <li>Driver specific code: anvil, freedreno, i965, iris, radeonsi,
 235       radv, vc4, ...
 236     <li>Other tag examples: gallium, util
 237   </ul>
 238 </p>
 239 <p>
 240   If you revise your patches based on code review and push an update
 241   to your branch, you should maintain a <strong>clean</strong> history
 242   in your patches. There should not be "fixup" patches in the history.
 243   The series should be buildable and functional after every commit
 244   whenever you push the branch.
 245 </p>
 246 <p>
 247   It is your responsibility to keep the MR alive and making progress,
 248   as there are no guarantees that a Mesa dev will independently take
 249   interest in it.
 250 </p>
 251 <p>
 252   Some other notes:
 253   <ul>
 254     <li>Make changes and update your branch based on feedback
 255     <li>Old, stale MR may be closed, but you can reopen it if you
 256       still want to pursue the changes
 257     <li>You should periodically check to see if your MR needs to be
 258       rebased
 259     <li>Make sure your MR is closed if your patches get pushed outside
 260       of GitLab
 261   </ul>
 262 </p>
 263
 264 <h2 id="reviewing">Reviewing Patches</h2>
 265
 266 <p>
 267   To participate in code review, you should monitor the
 268   <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev">
 269   mesa-dev</a> email list and the GitLab
 270   Mesa <a href="https://gitlab.freedesktop.org/mesa/mesa/merge_requests">Merge
 271   Requests</a> page.
 272 </p>
 273
 274 <p>
 275 When you've reviewed a patch on the mailing list, please be unambiguous
 276 about your review.  That is, state either
 277 </p>
 278 <pre>
 279     Reviewed-by: Joe Hacker &lt;jhacker@foo.com&gt;
 280 </pre>
 281 or
 282 <pre>
 283     Acked-by: Joe Hacker &lt;jhacker@foo.com&gt;
 284 </pre>
 285 <p>
 286 Rather than saying just "LGTM" or "Seems OK".
 287 </p>
 288
 289 <p>
 290 If small changes are suggested, it's OK to say something like:
 291 </p>
 292 <pre>
 293    With the above fixes, Reviewed-by: Joe Hacker &lt;jhacker@foo.com&gt;
 294 </pre>
 295 <p>
 296 which tells the patch author that the patch can be committed, as long
 297 as the issues are resolved first.
 298 </p>
 299
 300
 301 <h2 id="nominations">Nominating a commit for a stable branch</h2>
 302
 303 <p>
 304 There are three ways to nominate a patch for inclusion in the stable branch and
 305 release.
 306 </p>
 307 <ul>
 308 <li> By adding the Cc: mesa-stable@ tag as described below.
 309 <li> Sending the commit ID (as seen in master branch) to the mesa-stable@ mailing list.
 310 <li> Forwarding the patch from the mesa-dev@ mailing list.
 311 </li>
 312 </ul>
 313 <p>
 314 Note: resending patch identical to one on mesa-dev@ or one that differs only
 315 by the extra mesa-stable@ tag is <strong>not</strong> recommended.
 316 </p>
 317 <p>
 318 If you are not the author of the original patch, please Cc: them in your
 319 nomination request.
 320 </p>
 321
 322 <p>
 323 The current patch status can be observed in the <a href="releasing.html#stagingbranch">staging branch</a>.
 324 </p>
 325
 326 <h3 id="thetag">The stable tag</h3>
 327
 328 <p>
 329 If you want a commit to be applied to a stable branch,
 330 you should add an appropriate note to the commit message.
 331 </p>
 332
 333 <p>
 334 Here are some examples of such a note:
 335 </p>
 336 <ul>
 337   <li>CC: &lt;mesa-stable@lists.freedesktop.org&gt;</li>
 338 </ul>
 339
 340 Simply adding the CC to the mesa-stable list address is adequate to nominate
 341 the commit for all the active stable branches. If the commit is not applicable
 342 for said branch the stable-release manager will reply stating so.
 343
 344 This "CC" syntax for patch nomination will cause patches to automatically be
 345 copied to the mesa-stable@ mailing list when you use "git send-email" to send
 346 patches to the mesa-dev@ mailing list. If you prefer using --suppress-cc that
 347 won't have any negative effect on the patch nomination.
 348
 349 <p>
 350 Note: by removing the tag [as the commit is pushed] the patch is
 351 <strong>explicitly</strong> rejected from inclusion in the stable branch(es).
 352 <br>
 353 Thus, drop the line <strong>only</strong> if you want to cancel the nomination.
 354 </p>
 355
 356 Alternatively, if one uses the "Fixes" tag as described in the "Patch formatting"
 357 section, it nominates a commit for all active stable branches that include the
 358 commit that is referred to.
 359
 360 <h2 id="criteria">Criteria for accepting patches to the stable branch</h2>
 361
 362 Mesa has a designated release manager for each stable branch, and the release
 363 manager is the only developer that should be pushing changes to these branches.
 364 Everyone else should nominate patches using the mechanism described above.
 365
 366 The following rules define which patches are accepted and which are not. The
 367 stable-release manager is also given broad discretion in rejecting patches
 368 that have been nominated.
 369
 370 <ul>
 371   <li>Patch must conform with the <a href="#guidelines">Basic guidelines</a></li>
 372
 373   <li>Patch must have landed in master first. In case where the original
 374   patch is too large and/or otherwise contradicts with the rules set within, a
 375   backport is appropriate.</li>
 376
 377   <li>It must not introduce a regression - be that build or runtime wise.
 378
 379   Note:  If the regression is due to faulty piglit/dEQP/CTS/other test the
 380   latter must be fixed first. A reference to the offending test(s) and
 381   respective fix(es) should be provided in the nominated patch.</li>
 382
 383   <li>Patch cannot be larger than 100 lines.</li>
 384
 385   <li>Patches that move code around with no functional change should be
 386   rejected.</li>
 387
 388   <li>Patch must be a bug fix and not a new feature.
 389
 390   Note: An exception to this rule, are hardware-enabling "features". For
 391   example, <a href="#backports">backports</a> of new code to support a
 392   newly-developed hardware product can be accepted if they can be reasonably
 393   determined not to have effects on other hardware.</li>
 394
 395   <li>Patch must be reviewed, For example, the commit message has Reviewed-by,
 396   Signed-off-by, or Tested-by tags from someone but the author.</li>
 397
 398   <li>Performance patches are considered only if they provide information
 399   about the hardware, program in question and observed improvement. Use numbers
 400   to represent your measurements.</li>
 401 </ul>
 402
 403 If the patch complies with the rules it will be
 404 <a href="releasing.html#pickntest">cherry-picked</a>. Alternatively the release
 405 manager will reply to the patch in question stating why the patch has been
 406 rejected or would request a backport.
 407
 408 A summary of all the picked/rejected patches will be presented in the
 409 <a href="releasing.html#prerelease">pre-release</a> announcement.
 410
 411 The stable-release manager may at times need to force-push changes to the
 412 stable branches, for example, to drop a previously-picked patch that was later
 413 identified as causing a regression). These force-pushes may cause changes to
 414 be lost from the stable branch if developers push things directly. Consider
 415 yourself warned.
 416
 417 <h2 id="backports">Sending backports for the stable branch</h2>
 418 By default merge conflicts are resolved by the stable-release manager. In which
 419 case he/she should provide a comment about the changes required, alongside the
 420 <code>Conflicts</code> section. Summary of which will be provided in the
 421 <a href="releasing.html#prerelease">pre-release</a> announcement.
 422 <br>
 423 Developers are interested in sending backports are recommended to use either a
 424 <code>[BACKPORT #branch]</code> subject prefix or provides similar information
 425 within the commit summary.
 426
 427 <h2 id="gittips">Git tips</h2>
 428
 429 <ul>
 430 <li><code>git rebase -i ...</code> is your friend. Don't be afraid to use it.
 431 <li>Apply a fixup to commit FOO.
 432 <pre>
 433     git add ...
 434     git commit --fixup=FOO
 435     git rebase -i --autosquash ...
 436 </pre>
 437 <li>Test for build breakage between patches e.g last 8 commits.
 438 <pre>
 439     git rebase -i --exec="make -j4" HEAD~8
 440 </pre>
 441 <li>Sets the default mailing address for your repo.
 442 <pre>
 443     git config --local sendemail.to mesa-dev@lists.freedesktop.org
 444 </pre>
 445 <li> Add version to subject line of patch series in this case for the last 8
 446 commits before sending.
 447 <pre>
 448     git send-email --subject-prefix="PATCH v4" HEAD~8
 449     git send-email -v4 @~8 # shorter version, inherited from git format-patch
 450 </pre>
 451 <li> Configure git to use the get_reviewer.pl script interactively. Thus you
 452 can avoid adding the world to the CC list.
 453 <pre>
 454     git config sendemail.cccmd "./scripts/get_reviewer.pl -i"
 455 </pre>
 456 </ul>
 457
 458
 459 </div>
 460 </body>
 461 </html>