[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kwin
Subject:    Re: Review Request 109795: Longer delay before delayedCheckUnredirect()
From:       Thomas_Lübking <thomas.luebking () gmail ! com>
Date:       2013-03-29 20:38:17
Message-ID: 20130329203817.26270.67673 () vidsolbach ! de
[Download RAW message or body]

[Attachment #2 (multipart/alternative)]


> On March 29, 2013, 7 p.m., Thomas Lübking wrote:
> > Do you play at a different resolution?
> > 
> > reg. the patch:
> > it can in theory delay the check infinitely since everytime ::checkUnredirect() \
> > is called (to figure whether the last state change requires this) it's \
> > unconditionally scheduled by by another 100ms 
> > Covering more than "this event loop cycle" is hardly an issue, though.
> > 
> > Reg. the comment: i doubt this is about the game - more about the driver or pot. \
> > in conflict with Xrandr.
> 
> Alexander Mezin wrote:
> No, resolution isn't changed.
> 
> Seems that it's about the game, because:
> I tried SNA and UXA accelleration (intel driver) - no change
> Other games work fine (regardless of whether they change resolution or not, and \
> regardless of accelleration architecture) Also, it's not problem about screen not \
> redrawn: I don't hear any sound too. 
> Thomas Lübking wrote:
> ah, the meat clops is a handcrafted engine - i understand it happens for you with \
> xonotic as well (w/o res. change) and increasing the timeout from 0 to 1 would fix \
> this? 
> Alexander Mezin wrote:
> The problem is only with super meat boy. All other games, including Xonotic, work \
> fine without this patch. Increasing timeout from 0 to 1 fixes the problem.
> Removing "if" without increasing timeout fixes the problem.
> 
> I just thought that there may be other games with similar problem. 1 ms timeout \
> could be a hack that works only for this game and only on my machine.

Ok, it's actually quite really the game then.
I assume that it will cause sth. that triggers and then breaks (eventually several \
times) the unredirection state (thus bumping the timer on and on w/o the if cause)

I'd therefore suggest to look into some "smoothing" routine that prevents subsequent \
un/redirections within a given timeframe.

Are you interested into looking into that or do you want to try a patch?


- Thomas


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/109795/#review30083
-----------------------------------------------------------


On March 29, 2013, 5:12 p.m., Alexander Mezin wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/109795/
> -----------------------------------------------------------
> 
> (Updated March 29, 2013, 5:12 p.m.)
> 
> 
> Review request for kwin.
> 
> 
> Description
> -------
> 
> Super Meat Boy freezes when I start it full-screen with OpenGL compositing engine. \
> Seems that kwin just does its work unexpectedly fast. 
> Actually, even 1 ms delay or removing "if" fixes the problem for me.
> 
> But to be sure I think it's better to:
> 1) increase delay up to 100 ms. I think it can't be noticed by user
> 2) remove the "if", so successive calls won't be handled faster if the timer wasn't \
> timed out before these calls was made. 
> Could this patch cause any problems?
> 
> 
> Diffs
> -----
> 
> kwin/composite.cpp ee885dc 
> 
> Diff: http://git.reviewboard.kde.org/r/109795/diff/
> 
> 
> Testing
> -------
> 
> Tested on KDE 4.10.1
> Xonotic, Super Meat Boy, Battle for Wesnoth runs fine.
> 
> 
> Thanks,
> 
> Alexander Mezin
> 
> 


[Attachment #5 (text/html)]

<html>
 <body>
  <div style="font-family: Verdana, Arial, Helvetica, Sans-Serif;">
   <table bgcolor="#f9f3c9" width="100%" cellpadding="8" style="border: 1px #c9c399 \
solid;">  <tr>
     <td>
      This is an automatically generated e-mail. To reply, visit:
      <a href="http://git.reviewboard.kde.org/r/109795/">http://git.reviewboard.kde.org/r/109795/</a>
  </td>
    </tr>
   </table>
   <br />





<blockquote style="margin-left: 1em; border-left: 2px solid #d0d0d0; padding-left: \
10px;">  <p style="margin-top: 0;">On March 29th, 2013, 7 p.m. UTC, <b>Thomas \
Lübking</b> wrote:</p>  <blockquote style="margin-left: 1em; border-left: 2px solid \
#d0d0d0; padding-left: 10px;">  <pre style="white-space: pre-wrap; white-space: \
-moz-pre-wrap; white-space: -pre-wrap; white-space: -o-pre-wrap; word-wrap: \
break-word;">Do you play at a different resolution?

reg. the patch:
it can in theory delay the check infinitely since everytime ::checkUnredirect() is \
called (to figure whether the last state change requires this) it&#39;s \
unconditionally scheduled by by another 100ms

Covering more than &quot;this event loop cycle&quot; is hardly an issue, though.

Reg. the comment: i doubt this is about the game - more about the driver or pot. in \
conflict with Xrandr.</pre>  </blockquote>




 <p>On March 29th, 2013, 7:54 p.m. UTC, <b>Alexander Mezin</b> wrote:</p>
 <blockquote style="margin-left: 1em; border-left: 2px solid #d0d0d0; padding-left: \
10px;">  <pre style="white-space: pre-wrap; white-space: -moz-pre-wrap; white-space: \
-pre-wrap; white-space: -o-pre-wrap; word-wrap: break-word;">No, resolution isn&#39;t \
changed.

Seems that it&#39;s about the game, because:
I tried SNA and UXA accelleration (intel driver) - no change
Other games work fine (regardless of whether they change resolution or not, and \
regardless of accelleration architecture) Also, it&#39;s not problem about screen not \
redrawn: I don&#39;t hear any sound too.</pre>  </blockquote>





 <p>On March 29th, 2013, 8:12 p.m. UTC, <b>Thomas Lübking</b> wrote:</p>
 <blockquote style="margin-left: 1em; border-left: 2px solid #d0d0d0; padding-left: \
10px;">  <pre style="white-space: pre-wrap; white-space: -moz-pre-wrap; white-space: \
-pre-wrap; white-space: -o-pre-wrap; word-wrap: break-word;">ah, the meat clops is a \
handcrafted engine - i understand it happens for you with xonotic as well (w/o res. \
change) and increasing the timeout from 0 to 1 would fix this?</pre>  </blockquote>





 <p>On March 29th, 2013, 8:28 p.m. UTC, <b>Alexander Mezin</b> wrote:</p>
 <blockquote style="margin-left: 1em; border-left: 2px solid #d0d0d0; padding-left: \
10px;">  <pre style="white-space: pre-wrap; white-space: -moz-pre-wrap; white-space: \
-pre-wrap; white-space: -o-pre-wrap; word-wrap: break-word;">The problem is only with \
super meat boy. All other games, including Xonotic, work fine without this patch. \
Increasing timeout from 0 to 1 fixes the problem. Removing &quot;if&quot; without \
increasing timeout fixes the problem.

I just thought that there may be other games with similar problem. 1 ms timeout could \
be a hack that works only for this game and only on my machine.</pre>  </blockquote>








</blockquote>

<pre style="white-space: pre-wrap; white-space: -moz-pre-wrap; white-space: \
-pre-wrap; white-space: -o-pre-wrap; word-wrap: break-word;">Ok, it&#39;s actually \
quite really the game then. I assume that it will cause sth. that triggers and then \
breaks (eventually several times) the unredirection state (thus bumping the timer on \
and on w/o the if cause)

I&#39;d therefore suggest to look into some &quot;smoothing&quot; routine that \
prevents subsequent un/redirections within a given timeframe.

Are you interested into looking into that or do you want to try a patch?</pre>
<br />










<p>- Thomas</p>


<br />
<p>On March 29th, 2013, 5:12 p.m. UTC, Alexander Mezin wrote:</p>








<table bgcolor="#fefadf" width="100%" cellspacing="0" cellpadding="8" \
style="background-image: \
url('http://git.reviewboard.kde.org/static/rb/images/review_request_box_top_bg.ab6f3b1072c9.png'); \
background-position: left top; background-repeat: repeat-x; border: 1px black \
solid;">  <tr>
  <td>

<div>Review request for kwin.</div>
<div>By Alexander Mezin.</div>


<p style="color: grey;"><i>Updated March 29, 2013, 5:12 p.m.</i></p>






<h1 style="color: #575012; font-size: 10pt; margin-top: 1.5em;">Description </h1>
 <table width="100%" bgcolor="#ffffff" cellspacing="0" cellpadding="10" \
style="border: 1px solid #b8b5a0">  <tr>
  <td>
   <pre style="margin: 0; padding: 0; white-space: pre-wrap; white-space: \
-moz-pre-wrap; white-space: -pre-wrap; white-space: -o-pre-wrap; word-wrap: \
break-word;">Super Meat Boy freezes when I start it full-screen with OpenGL \
compositing engine. Seems that kwin just does its work unexpectedly fast.  
Actually, even 1 ms delay or removing &quot;if&quot; fixes the problem for me.

But to be sure I think it&#39;s better to:
1) increase delay up to 100 ms. I think it can&#39;t be noticed by user
2) remove the &quot;if&quot;, so successive calls won&#39;t be handled faster if the \
timer wasn&#39;t timed out before these calls was made.

Could this patch cause any problems?</pre>
  </td>
 </tr>
</table>


<h1 style="color: #575012; font-size: 10pt; margin-top: 1.5em;">Testing </h1>
<table width="100%" bgcolor="#ffffff" cellspacing="0" cellpadding="10" style="border: \
1px solid #b8b5a0">  <tr>
  <td>
   <pre style="margin: 0; padding: 0; white-space: pre-wrap; white-space: \
-moz-pre-wrap; white-space: -pre-wrap; white-space: -o-pre-wrap; word-wrap: \
break-word;">Tested on KDE 4.10.1 Xonotic, Super Meat Boy, Battle for Wesnoth runs \
fine.</pre>  </td>
 </tr>
</table>




<h1 style="color: #575012; font-size: 10pt; margin-top: 1.5em;">Diffs</b> </h1>
<ul style="margin-left: 3em; padding-left: 0;">

 <li>kwin/composite.cpp <span style="color: grey">(ee885dc)</span></li>

</ul>

<p><a href="http://git.reviewboard.kde.org/r/109795/diff/" style="margin-left: \
3em;">View Diff</a></p>







  </td>
 </tr>
</table>








  </div>
 </body>
</html>



_______________________________________________
kwin mailing list
kwin@kde.org
https://mail.kde.org/mailman/listinfo/kwin


[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic