<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    On 23.02.2012 18:32, Jerome Glisse wrote:
    <blockquote
cite="mid:CAH3drwaayhHk0cyK67mTdJNTC3Ox1GTPsELjGK9fqRtO9ODG2A@mail.gmail.com"
      type="cite">
      <div class="gmail_quote">2012/2/23 Mathias Fr&ouml;hlich <span
          dir="ltr">&lt;<a moz-do-not-send="true"
            href="mailto:Mathias.Froehlich@gmx.net">Mathias.Froehlich@gmx.net</a>&gt;</span><br>
        <blockquote class="gmail_quote" style="margin:0 0 0
          .8ex;border-left:1px #ccc solid;padding-left:1ex">
          <br>
          Christian,<br>
          <div class="im"><br>
            On Thursday, February 23, 2012 15:18:42 Christian K&ouml;nig
            wrote:<br>
            &gt; The function radeon_bo_list_validate can cause a<br>
            &gt; bo to move, resulting in a different sync_obj<br>
            &gt; and a dependency to wait for this move to finish.<br>
            &gt;<br>
            &gt; Signed-off-by: Christian K&ouml;nig &lt;<a
              moz-do-not-send="true"
              href="mailto:deathsimple@vodafone.de">deathsimple@vodafone.de</a>&gt;<br>
            &gt; Reviewed-by: Alex Deucher &lt;<a moz-do-not-send="true"
              href="mailto:alexander.deucher@amd.com">alexander.deucher@amd.com</a>&gt;<br>
            <br>
          </div>
          I am not sure, but to me this looks like this could fix these
          kind of gpu<br>
          lockups that I experience since some time every now and then.<br>
          The usual symptom is that I get the<br>
          <br>
          radeon 0000:01:00.0: GPU lockup CP stall for more than
          10000msec<br>
          GPU lockup (waiting for 0x00682AC3 last fence id 0x00682AC2)<br>
          [...]<br>
          <br>
          kernel message. Each time with the fence being off by one like
          in the example<br>
          above.<br>
          <br>
          If this change has the potential to fix this issue I think
          this particular<br>
          patch should be considered for the current upstream kernel
          release.<br>
          <span class="HOEnZb"><font color="#888888"><br>
              Mathias</font></span><br>
        </blockquote>
      </div>
      <br>
      No this patch doesn't. This patch is all about getting proper sync
      btw different rings.<br>
    </blockquote>
    That's unfortunately true, since we wasn't able to release any code
    that makes direct use of the different rings (yet) it shouldn't
    really matter in practice. I Just wanted to have that fix upstream
    since it is an obvious bug.<br>
    <br>
    Christian.<br>
    <blockquote
cite="mid:CAH3drwaayhHk0cyK67mTdJNTC3Ox1GTPsELjGK9fqRtO9ODG2A@mail.gmail.com"
      type="cite"><br>
      Sorry for the deception.<br>
      <br>
      Cheers,<br>
      Jerome<br>
    </blockquote>
    <br>
  </body>
</html>