Project Home
Project Home
Documents
Documents
Wiki
Wiki
Discussion Forums
Discussions
Project Information
Project Info
Forum Topic - Photon can't switch video mode when at least one GLES application is running: (5 Items)
   
Photon can't switch video mode when at least one GLES application is running  
Photon can't switch video mode when at least one GLES application is running, to reproduce this run gf-ph-3d and try to 
switch video mode resolution or color depth on the same resolution. Ideally AFAIU application should receive EXPOSE 
event and re-create all GLES stuff, but in practice after attempt of video mode switch, Display Settings utility reports
 about an unknown error.
RE: Photon can't switch video mode when at least one GLES application is running  
I have filed a PR67213 for investigation, but just to be clear, it is io-display that would be preventing the mode 
switch.

________________________________

From: Mike Gorchak [mailto:community-noreply@qnx.com]
Sent: Thu 09/04/2009 12:40 AM
To: photon-graphics
Subject: Photon can't switch video mode when at least one GLES application is running



Photon can't switch video mode when at least one GLES application is running, to reproduce this run gf-ph-3d and try to 
switch video mode resolution or color depth on the same resolution. Ideally AFAIU application should receive EXPOSE 
event and re-create all GLES stuff, but in practice after attempt of video mode switch, Display Settings utility reports
 about an unknown error.


_______________________________________________
Photon microGUI
http://community.qnx.com/sf/go/post26482



Attachment: Text winmail.dat 3.71 KB
Re: RE: Photon can't switch video mode when at least one GLES application is running  
Thanks, Derek. Take a look into sloginfo information after that, it contains a lot of errors, concerning failed to 
render command 0x4E or somthing like this.
RE: RE: Photon can't switch video mode when at least one GLES application is running  
Yes, that is no surprise, because offscreen is invalidated during modeswitch, and the application is trying to draw 
using invalidated surface IDs.  That said, the demo you reference uses a PtRaw for rendering, and not a PtOSContainer, 
so the application would have to intercept the expose event itself, and recreate the osc it is using.
 
-Derek

________________________________

From: Mike Gorchak [mailto:community-noreply@qnx.com]
Sent: Thu 09/04/2009 5:05 AM
To: photon-graphics
Subject: Re: RE: Photon can't switch video mode when at least one GLES application is running



Thanks, Derek. Take a look into sloginfo information after that, it contains a lot of errors, concerning failed to 
render command 0x4E or somthing like this.


_______________________________________________
Photon microGUI
http://community.qnx.com/sf/go/post26499



Attachment: Text winmail.dat 3.91 KB
Re: RE: Photon can't switch video mode when at least one GLES application is running  
> I have filed a PR67213 for investigation, but just to be clear, it is io-
> display that would be preventing the mode switch.
> 
> ________________________________
> 
> From: Mike Gorchak [mailto:community-noreply@qnx.com]
> Sent: Thu 09/04/2009 12:40 AM
> To: photon-graphics
> Subject: Photon can't switch video mode when at least one GLES application is 
> running
> 
> 
> 
> Photon can't switch video mode when at least one GLES application is running, 
> to reproduce this run gf-ph-3d and try to switch video mode resolution or 
> color depth on the same resolution. Ideally AFAIU application should receive 
> EXPOSE event and re-create all GLES stuff, but in practice after attempt of 
> video mode switch, Display Settings utility reports about an unknown error.
> 
> 
> _______________________________________________
> Photon microGUI
> http://community.qnx.com/sf/go/post26482
> 
> 
> 


I have finished the investigation of PR67213.

It is because a hybrid GF/Photon application does a gf_dev_attach(), and io-graphics cannot umount the device, which 
happens during a grafx restart service message.  This is expected behaviour, and part of the architectural compromises 
of GF/Photon co-existance.

Though it is up for debate whether is failing of a mode switch should occur, then reason it happens is now understood.