<br><br><div class="gmail_quote">On Thu, Feb 9, 2012 at 11:25 AM, william otten <span dir="ltr">&lt;<a href="mailto:williammotten@gmail.com">williammotten@gmail.com</a>&gt;</span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br><br><div class="gmail_quote"><div class="im">On Thu, Feb 9, 2012 at 10:24 AM, Mark Kendall <span dir="ltr">&lt;<a href="mailto:mark.kendall@gmail.com" target="_blank">mark.kendall@gmail.com</a>&gt;</span> wrote:<br></div>
<div><div class="h5"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF"><div><div><div><br><span>On 9 Feb 2012, at 12:12, william otten &lt;<a href="mailto:williammotten@gmail.com" target="_blank">williammotten@gmail.com</a>&gt; wrote:</span><br></div><div><br>
</div><div></div><blockquote type="cite"><div><br><br><div class="gmail_quote">On Wed, Feb 8, 2012 at 8:17 AM, William Otten <span dir="ltr">&lt;<a href="mailto:williammotten@gmail.com" target="_blank"></a><a href="mailto:williammotten@gmail.com" target="_blank">williammotten@gmail.com</a>&gt;</span> wrote:<br>

<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">

  
    
  
  <div bgcolor="#FFFFFF" text="#000000"><div><div>
    On 2/7/2012 5:53 AM, William Otten wrote:
    <blockquote type="cite">
      
      On 2/6/2012 2:23 PM, William Otten wrote:
      <blockquote type="cite">
        
        On 2/6/2012 10:56 AM, william otten wrote:
        <blockquote type="cite"><br>
          <br>
          <div class="gmail_quote">On Sat, Feb 4, 2012 at 8:58 AM,
            William Otten <span dir="ltr">&lt;<a href="mailto:williammotten@gmail.com" target="_blank"></a><a href="mailto:williammotten@gmail.com" target="_blank">williammotten@gmail.com</a>&gt;</span>
            wrote:<br>
            <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
              <div bgcolor="#FFFFFF" text="#000000">
                <div> On 2/3/2012 9:59 PM, william otten
                  wrote: </div>
                <div>
                  <div>
                    <blockquote type="cite"><br>
                      <br>
                      <div class="gmail_quote">On Fri, Feb 3, 2012 at
                        4:01 PM, Matt Garman <span dir="ltr">&lt;<a href="mailto:matthew.garman@gmail.com" target="_blank"></a><a href="mailto:matthew.garman@gmail.com" target="_blank">matthew.garman@gmail.com</a>&gt;</span>
                        wrote:<br>
                        <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
                          <p>Hi William,</p>
                          <p>The only way I know to verify the patch is
                            in place (other than by testing in myth
                            using opengl) is to look at the source. Look
                            at that function I mentioned in the earlier
                            post, and see if it includes the additional
                            null pointer check. Or, I suppose if you are
                            already in contact with the xorg devs, just
                            ask them directly.</p>
                          <p>I&#39;m afraid I can&#39;t think of any other way
                            to check. Hope that helps!<span><font color="#888888"><br>
                                Matt<br>
                              </font></span></p>
                          <div>
                            <div>
                              <div class="gmail_quote">On Feb 3, 2012
                                6:47 AM, &quot;William Otten&quot; &lt;<a href="mailto:williammotten@gmail.com" target="_blank"></a><a href="mailto:williammotten@gmail.com" target="_blank">williammotten@gmail.com</a>&gt;



                                wrote:<br type="attribution">
                                <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> On 1/29/2012
                                  7:34 PM, Matt Garman wrote:<br>
                                  <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> On Sat, Jan
                                    28, 2012 at 01:55:31PM -0600,
                                    william otten wrote:<br>
                                    <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> The patch
                                      is quite simple and involves one
                                      function in<br>
mesa/drivers/dri/intel/intel_screen.c<br>
                                      Another person on this list<br>
                                      reported that the patch solved an
                                      identical crash for him.  Before<br>
                                      diving into much debugging, I
                                      suggest you apply this patch to
                                      the mesa<br>
                                      7.11.2 code, recompile it, and
                                      give it a test.<br>
                                    </blockquote>
                                    I may be the &quot;other person&quot; who has
                                    been referred to several times<br>
                                    in this thread, as I did exactly
                                    that: I patched the intel_screen.c<br>
                                    file and rebuilt mesa, and now have
                                    a nicely working OpenGL-based<br>
                                    SNB frontend (which I&#39;m proud to say
                                    is also quite small and<br>
                                    low-power :).<br>
                                    <br>
                                    &gt; From my skimming of this
                                    thread, it looks like the individual
                                    with<br>
                                    the patching issues is running
                                    Ubuntu.  I&#39;m running CentOS plus<br>
                                    ATrpms (for MythTV) and elrepo for
                                    newer Intel video drivers and<br>
                                    supporting packages (in particular,
                                    mesa).<br>
                                    <br>
                                    What I did was simply download the
                                    source RPM for mesa from elrepo,<br>
                                    manually hack the intel_screen.c
                                    file according to the patch, and<br>
                                    rebuild.  I don&#39;t know how to do
                                    this with Ubuntu/debian packages,<br>
                                    but I can&#39;t imagine the process is
                                    dramatically different from an<br>
                                    RPM-based distro.<br>
                                    <br>
                                    FWIW, the patch that has been
                                    mentioned doesn&#39;t apply cleanly to
                                    the<br>
                                    mesa source from elrepo, because
                                    it&#39;s a slightly older version:<br>
                                    looks like the mesa-20111103
                                    snapshot.<br>
                                    <br>
                                    But the patch is super simple.  Just
                                    look for the function<br>
                                    &quot;intelDRI2Flush()&quot; in the file
                                    mentioned above.  It should be near<br>
                                    the top of the file.  Just replace
                                    that function definition with<br>
                                    this one:<br>
                                    <br>
                                    static void<br>
                                    intelDRI2Flush(__DRIdrawable
                                    *drawable)<br>
                                    {<br>
                                       GET_CURRENT_CONTEXT(ctx);<br>
                                       struct intel_context *intel =
                                    intel_context(ctx);<br>
                                    <br>
                                       if (intel != NULL) { /* add
                                    this... */<br>
                                           if (intel-&gt;gen&lt;  4)<br>
                                             
                                     INTEL_FIREVERTICES(intel);<br>
                                    <br>
                                           intel-&gt;need_throttle =
                                    GL_TRUE;<br>
                                    <br>
                                           if (intel-&gt;batch.used)<br>
                                             
                                     intel_batchbuffer_flush(intel);<br>
                                       } /* ...and this */<br>
                                    }<br>
                                    <br>
                                    As I commented above, the patch just
                                    adds two lines, which just do a<br>
                                    NULL pointer safety check.<br>
                                    <br>
                                    So, generic instructions would be:<br>
                                    <br>
                                        - get your distro&#39;s newest mesa
                                    source package<br>
                                        - unpack it (e.g. RPMs are
                                    really cpio archives)<br>
                                        - track down the intel_screen.c
                                    file and patch (or manually<br>
                                          edit) it as described above<br>
                                        - re-package the mesa source
                                    package<br>
                                        - use your distro&#39;s source
                                    package build tool to build a binary<br>
                                          package (e.g. rpmrebuild for
                                    RPM-based distros)<br>
                                    <br>
                                    Working with your distro&#39;s source
                                    packages should be well-documented<br>
                                    (at least for major distros like
                                    Ubuntu and Fedora).<br>
                                    <br>
                                    The other option is to wait: I also
                                    reported my success with this on<br>
                                    the elrepo mailing list.  The elrepo
                                    mesa maintainer said he didn&#39;t<br>
                                    want to update the package with a
                                    patch that wasn&#39;t committed<br>
                                    (according to this thread, it is
                                    now, but wasn&#39;t when I made the<br>
                                    comment).  So I would imagine that
                                    package maintainers will probably<br>
                                    be open to incorporating this patch
                                    into their builds now that it&#39;s<br>
                                    officially committed... I might
                                    request it for elrepo if I get<br>
                                    around to it.<br>
                                    <br>
                                    In other words, with this patch
                                    officially committed, it will<br>
                                    eventually make its way into
                                    different distro&#39;s mesa builds.<br>
                                    <br>
                                    HTH,<br>
                                    Matt<br>
                                    <br>
_______________________________________________<br>
                                    mythtv-users mailing list<br>
                                    <a href="mailto:mythtv-users@mythtv.org" target="_blank"></a><a href="mailto:mythtv-users@mythtv.org" target="_blank">mythtv-users@mythtv.org</a><br>
                                    <a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank"></a><a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank">http://www.mythtv.org/mailman/listinfo/mythtv-users</a><br>


                                  </blockquote>
                                  Matt<br>
                                  I appreciate your input on this issue.
                                  Not being a programmer, I can&#39;t get my
                                  brain wrapped around all that needs to
                                  be considered or done to test this
                                  patch. With that said, I have setup
                                  with xorg-edgers in anticipation that
                                  the patch will work it&#39;s way down
                                  stream. I contacted one of the dev&#39;s
                                  at xorg-edgers and he replied that the
                                  patch should be available within a
                                  week or two ( that was about a week
                                  ago). I have since received a couple
                                  of updates from them but I&#39;m not sure
                                  if the patch was included. Is there a
                                  way to check (other than just trying
                                  vaapi with mythtv) to see if the patch
                                  is included or has been received by
                                  me?<br>
                                  Thanks<br>
                                  William<br>
                                </blockquote>
                              </div>
                            </div>
                          </div>
                        </blockquote>
                      </div>
                      I&#39;m a fish out of water here. I tried to find the
                      source on the xorg-edgers site for over an hour
                      but no luck. The only thing I found was this:<br>
                      <br>
                      <a href="http://bazaar.launchpad.net/%7Evcs-imports/mesa/master/revision/26908" target="_blank"></a><a href="http://bazaar.launchpad.net/%7Evcs-imports/mesa/master/revision/26908" target="_blank">http://bazaar.launchpad.net/~vcs-imports/mesa/master/revision/26908</a><br>


                      <br>
                      The other url doesn&#39;t work anymore. Also I sent an
                      email to one of the dev&#39;s. <br>
                      <fieldset></fieldset>
                      <br>
                      <pre>_______________________________________________
mythtv-users mailing list
<a href="mailto:mythtv-users@mythtv.org" target="_blank"></a><a href="mailto:mythtv-users@mythtv.org" target="_blank">mythtv-users@mythtv.org</a>
<a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank"></a><a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank">http://www.mythtv.org/mailman/listinfo/mythtv-users</a>
</pre>
                    </blockquote>
                  </div>
                </div>
                <div> I tried the opengl option for render
                  and osd render and it works with ffmpeg. So the patch
                  must be working. I will try to use debug again with
                  vaapi and see if i can get errors to post since the
                  vaapi decoder option still isn&#39;t working.<br>
                </div>
              </div>
            </blockquote>
          </div>
          This the latest debug using the latest master:<br>
          <br>
          gdb.txt<br>
          <br>
          <a href="http://dl.dropbox.com/u/17574275/gdb%202-6-12.txt" target="_blank"></a><a href="http://dl.dropbox.com/u/17574275/gdb%202-6-12.txt" target="_blank">http://dl.dropbox.com/u/17574275/gdb%202-6-12.txt</a><br>


          <br>
          myth.log<br>
          <br>
          <a href="http://dl.dropbox.com/u/17574275/myth%202-6-12.log" target="_blank"></a><a href="http://dl.dropbox.com/u/17574275/myth%202-6-12.log" target="_blank">http://dl.dropbox.com/u/17574275/myth%202-6-12.log</a><br>


          <br>
          terminal capture<br>
          <br>
          <a href="http://dl.dropbox.com/u/17574275/MythtvBacktrace%202-6-12" target="_blank"></a><a href="http://dl.dropbox.com/u/17574275/MythtvBacktrace%202-6-12" target="_blank">http://dl.dropbox.com/u/17574275/MythtvBacktrace%202-6-12</a><br>


          <br>
          I also changed from ubuntu 11.10 to 12.4. A side note, mythtv
          plays without tearing using ffmpeg and opengl with unity.<br>
          <br>
          <fieldset></fieldset>
          <br>
          <pre>_______________________________________________
mythtv-users mailing list
<a href="mailto:mythtv-users@mythtv.org" target="_blank"></a><a href="mailto:mythtv-users@mythtv.org" target="_blank">mythtv-users@mythtv.org</a>
<a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank"></a><a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank">http://www.mythtv.org/mailman/listinfo/mythtv-users</a>
</pre>
        </blockquote>
        This statement keeps repeating in the command line but not
        showing up in <br>
        gdb.txt or myth.log:<br>
        intel_do_flush_locked failed: Input/output error<br>
        <br>
        <fieldset></fieldset>
        <br>
        <pre>_______________________________________________
mythtv-users mailing list
<a href="mailto:mythtv-users@mythtv.org" target="_blank"></a><a href="mailto:mythtv-users@mythtv.org" target="_blank">mythtv-users@mythtv.org</a>
<a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank"></a><a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank">http://www.mythtv.org/mailman/listinfo/mythtv-users</a>
</pre>
      </blockquote>
      I ran debug on the latest master v0.25pre-4398-g7090d59 and got
      the same result:<br>
      intel_do_flush_locked failed: Input/output error<br>
      The logs look the same so I&#39;m not posting them.<br>
      <br>
      <fieldset></fieldset>
      <br>
      <pre>_______________________________________________
mythtv-users mailing list
<a href="mailto:mythtv-users@mythtv.org" target="_blank"></a><a href="mailto:mythtv-users@mythtv.org" target="_blank">mythtv-users@mythtv.org</a>
<a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank"></a><a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank">http://www.mythtv.org/mailman/listinfo/mythtv-users</a>
</pre>
    </blockquote></div></div><div>
    I ran debug on the latest master v0.25pre-4422-g334f61f and ubuntu
    12.4.<br>
    I basically got the same result except when the process was finished
    this time, it ended in mythtv and not the desktop but the keyboard
    seemed to work one time (ctrl-alt-backspace and then ctrl-alt-F1 to
    reboot) but not the last time. The gdb.txt and myth.log didn&#39;t show
    any different info. I&#39;m allowing the debug to resolve itself the
    first time and using control-c the second time.<br>
  </div></div>

</blockquote></div>The latest debug with the latest master with basically the same result:<br>intel_do_flush_locked failed: Input/output error<br>Is there anything else I need to do? Do I keep posting? I haven&#39;t had any replies in a while. Is anyone else having the same or related issue? I need some guidance to move forward.<br>


<br>gdb,txt<br><br><a href="http://dl.dropbox.com/u/17574275/gdb%202-9-12.txt" target="_blank"></a><a href="http://dl.dropbox.com/u/17574275/gdb%202-9-12.txt" target="_blank">http://dl.dropbox.com/u/17574275/gdb%202-9-12.txt</a><br>

<br>myth.log<br><br><a href="http://dl.dropbox.com/u/17574275/myth%202-9-12.log" target="_blank"></a><a href="http://dl.dropbox.com/u/17574275/myth%202-9-12.log" target="_blank">http://dl.dropbox.com/u/17574275/myth%202-9-12.log</a><br>


<br>terminal capture<br><br><a href="http://dl.dropbox.com/u/17574275/MythtvBacktrace%202-9-12" target="_blank"></a><a href="http://dl.dropbox.com/u/17574275/MythtvBacktrace%202-9-12" target="_blank">http://dl.dropbox.com/u/17574275/MythtvBacktrace%202-9-12</a><br>


</div></blockquote></div></div><blockquote type="cite"><div><span>_______________________________________________</span><br><br></div></blockquote><br><div>William</div><div><br></div><div>A couple of things:-</div><div>

<br></div><div>Can you add -v playback to your log output.</div></div></blockquote><div> </div></div></div><div class="im"><div>Please be more specific on how I would do this. <br><br></div></div><div class="im"><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">

<div bgcolor="#FFFFFF"><div>Can you check whether you have any deinterlacers selected in your playback profile. If you have, try removing them.</div></div></blockquote></div><div><br><div class="im">No deinterlacers have been used.<br>
 <br></div></div>
<blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div bgcolor="#FFFFFF"><div>Can you try playing back some h264 and vc1 material. (according to your logs h264 won&#39;t work, which is slightly worrying, although it probably isn&#39;t relevant here)</div>

</div></blockquote><div><br><div class="im">Not sure what vc1 is and I think h264 is a mpg4 file? <br></div></div><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div bgcolor="#FFFFFF">
<div><br></div><div class="im"><div>And most importantly, you&#39;re not actually capturing the backtrace. If you&#39;re running directly within gdb, once mythfrontend crashes, you&#39;ll need to type in something like &#39;thread apply all bt&#39; (from memory, the correct command is in the debugging section of the docs).</div>

</div></div></blockquote><div><br><div class="im"> Here is my gdbcommands file:<br><br>handle SIGPIPE nostop noprint<br>handle SIG33 nostop noprint<br>set logging on<br>set pagination off<br>set breakpoint pending on<br>
break qFatal<br>set args -l myth.log -v important,general<br>
run<br>thread apply all bt full<br>set logging off<br><br></div></div><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div bgcolor="#FFFFFF"><div><br>
</div>
<div>Regards</div><span><font color="#888888"><div>Mark</div></font></span></div><br><div class="im">_______________________________________________<br>
mythtv-users mailing list<br>
<a href="mailto:mythtv-users@mythtv.org" target="_blank">mythtv-users@mythtv.org</a><br>
<a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank">http://www.mythtv.org/mailman/listinfo/mythtv-users</a><br>
<br></div></blockquote><div class="im"><div><br>Concerning the backtrace, I stated in the previous posts that I am allowing gdb to resolve itself the 1st time then I try using control-c the second time. I apply control-c right after I select the enter key to start the recording and then I keep pressing control-c many times. <br>

</div></div></div>
</blockquote></div><br>I decided to try different recordings / videos due to Mark&#39;s comment, lately I have been using one recording - cnbchd 1080 which never worked. I liked this for a test due to the scrolling ticker. The results:<br>
TV recordings 720P<br>I tried several recordings and got mixed results:<br>1. The recordings didn&#39;t work repeatably - If I successfully viewed a recording the next time I tried after a reboot it may or may not work.<br>
2. The startup for recordings was delayed up to 25 seconds. The usual routine was for the &quot;Please wait&quot; screen then 2 seconds later there was audio for a second then after 5 more seconds the screen would turn black and then the video and audio would start. The total time here was about 7 - 8 seconds. The other extreme was the same up to the black screen but this would happens after 25 seconds.<br>
3. Exiting a successful recording could result in a freeze. The screen would have frame of video and the keyboard was non-responsive. I would have to use a hard reset at this point.<br>Videos - AVI and MKV<br>1.  MKV videos would not work but the AVI videos did and they could be re-played again unlike the tv recordings.<br>
2. The AVI videos weren&#39;t as good a quality as the tv recordings.<br>3. The startup was normal for the AVI videos.<br>4. Exiting a successful video could result in a freeze. The screen 
would have frame of video and the keyboard was non-responsive. I would 
have to use a hard reset at this point.<br><br>Here are the log files. There is something different in the gdb.txt. I can&#39;t state when or why it happened due the numerous attempts.<br><br>gdb.txt<br><br><a href="http://dl.dropbox.com/u/17574275/gdb%202-9-12%20A.txt">http://dl.dropbox.com/u/17574275/gdb%202-9-12%20A.txt</a><br>
<br>myth.log<br><br><a href="http://dl.dropbox.com/u/17574275/myth%202-9-12%20A.log">http://dl.dropbox.com/u/17574275/myth%202-9-12%20A.log</a><br>