[Intel-gfx] CRT not detected via hotplug on resume

[Rainer Dorsch]

Am Donnerstag 22 September 2011, 16:33:46 schrieb Keith Packard:
> On Thu, 22 Sep 2011 10:36:12 +0200, Rainer Dorsch <ml at bokomoko.de> wrote:
> > This kind of output I do not see in the syslog anymore during regular
> 
> > operation now (is that the 10-second polling?):
> Yes. So, the patch is definitely included.

Thanks for confirming this.
 
> > > Let me know if it actually helps your suspend/resume issues.
> > 
> > So far the patch at least did not break anything visible to me :-) Since
> > I cannot reproduce the suspend/resume issues reliably, this takes now
> > some time. If I will have not seen it within the next two weeks, the
> > problem is fixed or at least improved.
> 
> Thanks for testing this patch.

The patch runs stable for me so far. But I hit the suspend/resume issue again. 
When I ssh into the machine, I get through xrandr

rd at blackbox:~$ xrandr -display :0
Screen 0: minimum 320 x 200, current 1280 x 1024, maximum 8192 x 8192
VGA1 disconnected (normal left inverted right x axis y axis)
HDMI1 connected 1280x1024+0+0 (normal left inverted right x axis y axis) 376mm 
x 301mm
   1280x1024      60.0*+
   1024x768       60.0  
   800x600        60.3  
   640x480        60.0  
   720x400        70.1  
rd at blackbox:~$

which looks ok fo me. I tried to switch to another resolution, but the machine 
autosuspended, before I was able to do so (I took me a few minutes to figure 
out a copy and paste problem :-( ). After the second resume, HDMI was working 
again.

xrandr shows the same data after the second resume:

rd at blackbox:~$ xrandr -display :0
Screen 0: minimum 320 x 200, current 1280 x 1024, maximum 8192 x 8192
VGA1 disconnected (normal left inverted right x axis y axis)
HDMI1 connected 1280x1024+0+0 (normal left inverted right x axis y axis) 376mm 
x 301mm
   1280x1024      60.0*+
   1024x768       60.0  
   800x600        60.3  
   640x480        60.0  
   720x400        70.1  
rd at blackbox:~$ 

Are there any other data I could gather to figure out what is going wrong?

Thanks,
Rainer

BTW, I start thinking on workarounds. One workaround would be to immediately 
suspend the system again, when I hit the problem. Since only graphics seems to 
be affected, I am thinking of "blindly" switching to a VT enter the root 
password and run pm-suspend. What I am worried about is that I have to enter 
the root password blindly. Are there better ways to manually suspend (without 
booting up another system and sshing in)? Could a cron script (e.g. running 
once a minute) detect the situation (but how, xrandr output seems to be ok)?

Many thanks,
Rainer