<html>
    <head>
      <base href="https://bugs.freedesktop.org/">
    </head>
    <body><span class="vcard"><a class="email" href="mailto:kenneth@whitecape.org" title="Kenneth Graunke <kenneth@whitecape.org>"> <span class="fn">Kenneth Graunke</span></a>
</span> changed
          <a class="bz_bug_link 
          bz_status_RESOLVED  bz_closed"
   title="RESOLVED FIXED - [IVB,BYT] GPU hang on deqp-gles31.functional.separate.shader.random"
   href="https://bugs.freedesktop.org/show_bug.cgi?id=97083">bug 97083</a>
          <br>
             <table border="1" cellspacing="0" cellpadding="8">
          <tr>
            <th>What</th>
            <th>Removed</th>
            <th>Added</th>
          </tr>

         <tr>
           <td style="text-align:right;">Resolution</td>
           <td>---
           </td>
           <td>FIXED
           </td>
         </tr>

         <tr>
           <td style="text-align:right;">Status</td>
           <td>NEW
           </td>
           <td>RESOLVED
           </td>
         </tr></table>
      <p>
        <div>
            <b><a class="bz_bug_link 
          bz_status_RESOLVED  bz_closed"
   title="RESOLVED FIXED - [IVB,BYT] GPU hang on deqp-gles31.functional.separate.shader.random"
   href="https://bugs.freedesktop.org/show_bug.cgi?id=97083#c2">Comment # 2</a>
              on <a class="bz_bug_link 
          bz_status_RESOLVED  bz_closed"
   title="RESOLVED FIXED - [IVB,BYT] GPU hang on deqp-gles31.functional.separate.shader.random"
   href="https://bugs.freedesktop.org/show_bug.cgi?id=97083">bug 97083</a>
              from <span class="vcard"><a class="email" href="mailto:kenneth@whitecape.org" title="Kenneth Graunke <kenneth@whitecape.org>"> <span class="fn">Kenneth Graunke</span></a>
</span></b>
        <pre>Curro figured this one out.  The problem was the NoDDClear/NoDDCheck flags on
the PLN instructions.  Normally, we do a (-f0.0) PLN with NoDDClear (for helper
pixels) and a (+f0.0) PLN with NoDDCheck.  This works fine at the top level,
because at least one pixel won't be a helper invocation, so the (-f0.0) PLN
with NoDDClear will actually execute.

But with my (broken) rework, we started emitting PLNs inside control flow.  If
only helper pixels take a certain control flow path, the (-f0.0) PLN with
NoDDClear might not ever execute, leaving the register dependency set, so the
next instruction that tries to touch that register would hang.

So, move to top is a real solution to this problem, not just papering over it.
I've also sent a patch to the mailing list that eliminates the PLN dependency
hints, which will allow us to move PLNs into control flow safely (if we want
to):
<a href="https://lists.freedesktop.org/archives/mesa-dev/2016-August/124946.html">https://lists.freedesktop.org/archives/mesa-dev/2016-August/124946.html</a></pre>
        </div>
      </p>


      <hr>
      <span>You are receiving this mail because:</span>

      <ul>
          <li>You are the QA Contact for the bug.</li>
      </ul>
    </body>
</html>