[Openchrome-users] CN700 problems, some details.

Christiaan van Dijk dvbmail
Sun Dec 17 03:34:56 PST 2006


Looked a bit deeper into the changed behavior and I found the following:

Openchrome rev. 250:
Seems to work fine but sometimes X takes full CPU while playing video, 
can be solved by pausing/resuming, if mplayer drops behind way too far X 
crashes. Also the combination of heavy disc IO and screen IO crashes X.

Openchrome rev. 252
Same story as 250 but did not test in all the details, mplayer works but 
need to pause when frame starts dropping to let X catchup?

Openchrome rev. 255:
System performance is poor, starting a shell can take up to 30 seconds 
some times. Playing a video with mplayer crashes X as soon as frames 
begin to drop (which indicates X is taking too much CPU time).

Looking at the changes between rev. 252 and 255 I can only see the 
following change which could relate to this issue:

via_driver.c:

#ifdef XF86DRI
    if (pVia->directRenderingEnabled) {
          kickVblank(pScrn);
++     VIADRIRingBufferInit(pScrn);
          viaDRIOffscreenRestore(pScrn);
--       VIADRIRingBufferInit(pScrn);
    }
#endif 
  
This small change has a big effect on the problem, also in line with the 
disabling of DRI which helps improve the situation.
Will look further into this one. Also found a dmesg entry but I'm not 
sure what to think of it:

PCI: Unable to reserve mem region #1:4000000 at f4000000 for device 
0000:01:00.0

The device is the VT8237 PCI bridge, really don't know what to think of 
this, starting to get a bit paranoid :-)

Regds,
Christiaan van Dijk.






More information about the Openchrome-users mailing list