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="#mailing">Mailing 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 </ul>
  29
  30 <h2 id="guidelines">Basic guidelines</h2>
  31
  32 <ul>
  33 <li>Patches should not mix code changes with code formatting changes (except,
  34 perhaps, in very trivial cases.)
  35 <li>Code patches should follow Mesa
  36 <a href="codingstyle.html" target="_parent">coding conventions</a>.
  37 <li>Whenever possible, patches should only effect individual Mesa/Gallium
  38 components.
  39 <li>Patches should never introduce build breaks and should be bisectable (see
  40 <code>git bisect</code>.)
  41 <li>Patches should be properly <a href="#formatting">formatted</a>.
  42 <li>Patches should be sufficiently <a href="#testing">tested</a> before submitting.
  43 <li>Patches should be submitted to <a href="#mailing">submitted to mesa-dev</a>
  44 for <a href="#reviewing">review</a> using <code>git send-email</code>.
  45
  46 </ul>
  47
  48 <h2 id="formatting">Patch formatting</h2>
  49
  50 <ul>
  51 <li>Lines should be limited to 75 characters or less so that git logs
  52 displayed in 80-column terminals avoid line wrapping.  Note that git
  53 log uses 4 spaces of indentation (4 + 75 &lt; 80).
  54 <li>The first line should be a short, concise summary of the change prefixed
  55 with a module name.  Examples:
  56 <pre>
  57     mesa: Add support for querying GL_VERTEX_ATTRIB_ARRAY_LONG
  58
  59     gallium: add PIPE_CAP_DEVICE_RESET_STATUS_QUERY
  60
  61     i965: Fix missing type in local variable declaration.
  62 </pre>
  63 <li>Subsequent patch comments should describe the change in more detail,
  64 if needed.  For example:
  65 <pre>
  66     i965: Remove end-of-thread SEND alignment code.
  67
  68     This was present in Eric's initial implementation of the compaction code
  69     for Sandybridge (commit 077d01b6). There is no documentation saying this
  70     is necessary, and removing it causes no regressions in piglit on any
  71     platform.
  72 </pre>
  73 <li>A "Signed-off-by:" line is not required, but not discouraged either.
  74 <li>If a patch address a bugzilla issue, that should be noted in the
  75 patch comment.  For example:
  76 <pre>
  77    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89689
  78 </pre>
  79 <li>If there have been several revisions to a patch during the review
  80 process, they should be noted such as in this example:
  81 <pre>
  82     st/mesa: add ARB_texture_stencil8 support (v4)
  83
  84     if we support stencil texturing, enable texture_stencil8
  85     there is no requirement to support native S8 for this,
  86     the texture can be converted to x24s8 fine.
  87
  88     v2: fold fixes from Marek in:
  89        a) put S8 last in the list
  90        b) fix renderable to always test for d/s renderable
  91         fixup the texture case to use a stencil only format
  92         for picking the format for the texture view.
  93     v3: hit fallback for getteximage
  94     v4: put s8 back in front, it shouldn't get picked now (Ilia)
  95 </pre>
  96 <li>If someone tested your patch, document it with a line like this:
  97 <pre>
  98     Tested-by: Joe Hacker &lt;jhacker@foo.com&gt;
  99 </pre>
 100 <li>If the patch was reviewed (usually the case) or acked by someone,
 101 that should be documented with:
 102 <pre>
 103     Reviewed-by: Joe Hacker &lt;jhacker@foo.com&gt;
 104     Acked-by: Joe Hacker &lt;jhacker@foo.com&gt;
 105 </pre>
 106 </ul>
 107
 108
 109
 110 <h2 id="testing">Testing Patches</h2>
 111
 112 <p>
 113 It should go without saying that patches must be tested.  In general,
 114 do whatever testing is prudent.
 115 </p>
 116
 117 <p>
 118 You should always run the Mesa test suite before submitting patches.
 119 The test suite can be run using the 'make check' command. All tests
 120 must pass before patches will be accepted, this may mean you have
 121 to update the tests themselves.
 122 </p>
 123
 124 <p>
 125 Whenever possible and applicable, test the patch with
 126 <a href="http://piglit.freedesktop.org">Piglit</a> and/or
 127 <a href="https://android.googlesource.com/platform/external/deqp/">dEQP</a>
 128 to check for regressions.
 129 </p>
 130
 131
 132 <h2 id="mailing">Mailing Patches</h2>
 133
 134 <p>
 135 Patches should be sent to the mesa-dev mailing list for review:
 136 <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev">
 137 mesa-dev@lists.freedesktop.org</a>.
 138 When submitting a patch make sure to use
 139 <a href="https://git-scm.com/docs/git-send-email">git send-email</a>
 140 rather than attaching patches to emails. Sending patches as
 141 attachments prevents people from being able to provide in-line review
 142 comments.
 143 </p>
 144
 145 <p>
 146 When submitting follow-up patches you can use --in-reply-to to make v2, v3,
 147 etc patches show up as replies to the originals. This usually works well
 148 when you're sending out updates to individual patches (as opposed to
 149 re-sending the whole series). Using --in-reply-to makes
 150 it harder for reviewers to accidentally review old patches.
 151 </p>
 152
 153 <p>
 154 When submitting follow-up patches you should also login to
 155 <a href="https://patchwork.freedesktop.org">patchwork</a> and change the
 156 state of your old patches to Superseded.
 157 </p>
 158
 159 <h2 id="reviewing">Reviewing Patches</h2>
 160
 161 <p>
 162 When you've reviewed a patch on the mailing list, please be unambiguous
 163 about your review.  That is, state either
 164 </p>
 165 <pre>
 166     Reviewed-by: Joe Hacker &lt;jhacker@foo.com&gt;
 167 </pre>
 168 or
 169 <pre>
 170     Acked-by: Joe Hacker &lt;jhacker@foo.com&gt;
 171 </pre>
 172 <p>
 173 Rather than saying just "LGTM" or "Seems OK".
 174 </p>
 175
 176 <p>
 177 If small changes are suggested, it's OK to say something like:
 178 </p>
 179 <pre>
 180    With the above fixes, Reviewed-by: Joe Hacker &lt;jhacker@foo.com&gt;
 181 </pre>
 182 <p>
 183 which tells the patch author that the patch can be committed, as long
 184 as the issues are resolved first.
 185 </p>
 186
 187
 188 <h2 id="nominations">Nominating a commit for a stable branch</h2>
 189
 190 <p>
 191 There are three ways to nominate patch for inclusion of the stable branch and
 192 release.
 193 </p>
 194 <ul>
 195 <li> By adding the Cc: mesa-stable@ tag as described below.
 196 <li> Sending the commit ID (as seen in master branch) to the mesa-stable@ mailing list.
 197 <li> Forwarding the patch from the mesa-dev@ mailing list.
 198 </li>
 199 </ul>
 200 <p>
 201 Note: resending patch identical to one on mesa-dev@ or one that differs only
 202 by the extra mesa-stable@ tag is <strong>not</strong> recommended.
 203 </p>
 204
 205
 206 <h3 id="thetag">The stable tag</h3>
 207
 208 <p>
 209 If you want a commit to be applied to a stable branch,
 210 you should add an appropriate note to the commit message.
 211 </p>
 212
 213 <p>
 214 Here are some examples of such a note:
 215 </p>
 216 <ul>
 217   <li>CC: &lt;mesa-stable@lists.freedesktop.org&gt;</li>
 218   <li>CC: "9.2 10.0" &lt;mesa-stable@lists.freedesktop.org&gt;</li>
 219   <li>CC: "10.0" &lt;mesa-stable@lists.freedesktop.org&gt;</li>
 220 </ul>
 221
 222 Simply adding the CC to the mesa-stable list address is adequate to nominate
 223 the commit for the most-recently-created stable branch. It is only necessary
 224 to specify a specific branch name, (such as "9.2 10.0" or "10.0" in the
 225 examples above), if you want to nominate the commit for an older stable
 226 branch. And, as in these examples, you can nominate the commit for the older
 227 branch in addition to the more recent branch, or nominate the commit
 228 exclusively for the older branch.
 229
 230 This "CC" syntax for patch nomination will cause patches to automatically be
 231 copied to the mesa-stable@ mailing list when you use "git send-email" to send
 232 patches to the mesa-dev@ mailing list. If you prefer using --suppress-cc that
 233 won't have any effect negative effect on the patch nomination.
 234
 235 <h2 id="criteria">Criteria for accepting patches to the stable branch</h2>
 236
 237 Mesa has a designated release manager for each stable branch, and the release
 238 manager is the only developer that should be pushing changes to these
 239 branches. Everyone else should simply nominate patches using the mechanism
 240 described above.
 241
 242 The stable-release manager will work with the list of nominated patches, and
 243 for each patch that meets the crtieria below will cherry-pick the patch with:
 244 <code>git cherry-pick -x &lt;commit&gt;</code>. The <code>-x</code> option is
 245 important so that the picked patch references the comit ID of the original
 246 patch.
 247
 248 The stable-release manager may at times need to force-push changes to the
 249 stable branches, for example, to drop a previously-picked patch that was later
 250 identified as causing a regression). These force-pushes may cause changes to
 251 be lost from the stable branch if developers push things directly. Consider
 252 yourself warned.
 253
 254 The stable-release manager is also given broad discretion in rejecting patches
 255 that have been nominated for the stable branch. The most basic rule is that
 256 the stable branch is for bug fixes only, (no new features, no
 257 regressions). Here is a non-exhaustive list of some reasons that a patch may
 258 be rejected:
 259
 260 <ul>
 261   <li>Patch introduces a regression. Any reported build breakage or other
 262   regression caused by a particular patch, (game no longer work, piglit test
 263   changes from PASS to FAIL), is justification for rejecting a patch.</li>
 264
 265   <li>Patch is too large, (say, larger than 100 lines)</li>
 266
 267   <li>Patch is not a fix. For example, a commit that moves code around with no
 268   functional change should be rejected.</li>
 269
 270   <li>Patch fix is not clearly described. For example, a commit message
 271   of only a single line, no description of the bug, no mention of bugzilla,
 272   etc.</li>
 273
 274   <li>Patch has not obviously been reviewed, For example, the commit message
 275   has no Reviewed-by, Signed-off-by, nor Tested-by tags from anyone but the
 276   author.</li>
 277
 278   <li>Patch has not already been merged to the master branch. As a rule, bug
 279   fixes should never be applied first to a stable branch. Patches should land
 280   first on the master branch and then be cherry-picked to a stable
 281   branch. (This is to avoid future releases causing regressions if the patch
 282   is not also applied to master.) The only things that might look like
 283   exceptions would be backports of patches from master that happen to look
 284   significantly different.</li>
 285
 286   <li>Patch depends on too many other patches. Ideally, all stable-branch
 287   patches should be self-contained. It sometimes occurs that a single, logical
 288   bug-fix occurs as two separate patches on master, (such as an original
 289   patch, then a subsequent fix-up to that patch). In such a case, these two
 290   patches should be squashed into a single, self-contained patch for the
 291   stable branch. (Of course, if the squashing makes the patch too large, then
 292   that could be a reason to reject the patch.)</li>
 293
 294   <li>Patch includes new feature development, not bug fixes. New OpenGL
 295   features, extensions, etc. should be applied to Mesa master and included in
 296   the next major release. Stable releases are intended only for bug fixes.
 297
 298   Note: As an exception to this rule, the stable-release manager may accept
 299   hardware-enabling "features". For example, backports of new code to support
 300   a newly-developed hardware product can be accepted if they can be reasonably
 301   determined to not have effects on other hardware.</li>
 302
 303   <li>Patch is a performance optimization. As a rule, performance patches are
 304   not candidates for the stable branch. The only exception might be a case
 305   where an application's performance was recently severely impacted so as to
 306   become unusable. The fix for this performance regression could then be
 307   considered for a stable branch. The optimization must also be
 308   non-controversial and the patches still need to meet the other criteria of
 309   being simple and self-contained</li>
 310
 311   <li>Patch introduces a new failure mode (such as an assert). While the new
 312   assert might technically be correct, for example to make Mesa more
 313   conformant, this is not the kind of "bug fix" we want in a stable
 314   release. The potential problem here is that an OpenGL program that was
 315   previously working, (even if technically non-compliant with the
 316   specification), could stop working after this patch. So that would be a
 317   regression that is unaacceptable for the stable branch.</li>
 318 </ul>
 319
 320
 321 </div>
 322 </body>
 323 </html>