Bug 56931 - Using xbmc disconnects screen
Summary: Using xbmc disconnects screen
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: 7.6 (2010.12)
Hardware: Other All
: medium normal
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-11-09 18:11 UTC by Tomi Pieviläinen
Modified: 2016-02-25 07:14 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
dmesg (79.04 KB, text/plain)
2012-11-09 18:11 UTC, Tomi Pieviläinen
no flags Details
Xorg.log (136.60 KB, text/plain)
2012-11-09 18:11 UTC, Tomi Pieviläinen
no flags Details

Description Tomi Pieviläinen 2012-11-09 18:11:04 UTC
Created attachment 69831 [details]
dmesg

Trying to use xbmc causes the screen to lose signal after a few seconds (not immediate). At this point I'm not sure what makes it special, I don't even have to play a video, just browse the menus. Trying to re-enable the screen with xrandr does nothing but repeat the error in dmesg:

[117133.469521] [drm:radeon_dp_link_train_cr] *ERROR* displayport link status failed
[117133.469527] [drm:radeon_dp_link_train_cr] *ERROR* clock recovery failed

I'm using A6-3500 with F1A75-V_PRO motherboard, screens on DVI and DP (attached to active DP-DLDVI adapter and to a DLDVI screen).
Comment 1 Tomi Pieviläinen 2012-11-09 18:11:34 UTC
Created attachment 69832 [details]
Xorg.log
Comment 2 Alex Deucher 2012-11-09 18:15:46 UTC
What version of mesa are you using?  Does disabling acceleration fix the issue?
Option "NoAccel" "True"
in the device section of your xorg.conf
Comment 3 Alex Deucher 2012-11-09 18:17:18 UTC
Also, does a 3.7 kernel work any better?
Comment 4 Tomi Pieviläinen 2012-11-09 19:07:12 UTC
Mesa seems to a mix, all packages installed from ubuntu quantal repos:

$ dpkg -l |grep mesa|grep ii
ii  libegl1-mesa:amd64                        9.1~git1210191809.c2e37b~gd~p              amd64        free implementation of the EGL API -- runtime
ii  libegl1-mesa-drivers:amd64                9.1~git1210191809.c2e37b~gd~p              amd64        free implementation of the EGL API -- hardware drivers
ii  libgl1-mesa-dev                           9.1~git1210191809.c2e37b~gd~p              amd64        free implementation of the OpenGL API -- GLX development files
ii  libgl1-mesa-dri:amd64                     9.1~git1210191809.c2e37b~gd~p              amd64        free implementation of the OpenGL API -- DRI modules
ii  libgl1-mesa-glx:amd64                     9.1~git1210191809.c2e37b~gd~p              amd64        free implementation of the OpenGL API -- GLX runtime
ii  libglapi-mesa:amd64                       9.1~git1210191809.c2e37b~gd~p              amd64        free implementation of the GL API -- shared library
ii  libglu1-mesa:amd64                        9.0.0-0ubuntu1                             amd64        Mesa OpenGL utility library (GLU)
ii  libopenvg1-mesa:amd64                     9.1~git1210191809.c2e37b~gd~p              amd64        free implementation of the OpenVG API -- runtime
ii  mesa-common-dev                           9.1~git1210191809.c2e37b~gd~p              amd64        Developer documentation for Mesa
ii  mesa-utils                                8.0.1+git20110129+d8f7d6b-0ubuntu2         amd64        Miscellaneous Mesa GL utilities

I think xbmc requires acceleration, so NoAccel wouldn't help. I'll try 3.7.
Comment 5 Tomi Pieviläinen 2012-11-09 19:32:22 UTC
Hmh, trying to install 3.7 from ubuntu kernel team just gave "PANIC: early exception 08 rip 246:10 error 810440a6 cr2 0" after reboot. Will have to wait until that is resolved to test 3.7.

Meanwhile I tried booting to 3.5, but the DP screen never activated... same error in dmesg. Will try cold boot next to get the screen back.
Comment 6 Alex Deucher 2012-11-09 19:46:45 UTC
(In reply to comment #4)
> 
> I think xbmc requires acceleration, so NoAccel wouldn't help. I'll try 3.7.

It will still work, just slowly.  It will help narrow down where the problem is.
Comment 7 Tomi Pieviläinen 2012-11-09 20:37:46 UTC
(In reply to comment #6)
> (In reply to comment #4)
> > 
> > I think xbmc requires acceleration, so NoAccel wouldn't help. I'll try 3.7.
> 
> It will still work, just slowly.  It will help narrow down where the problem
> is.

Indeed it works. My other screen won't pivot while NoAccel is on, not sure if that has anything to do with this. But I can't make the problem manifest now. Before it didn't take much scrolling in the menus to reproduce.
Comment 8 Tomi Pieviläinen 2012-11-11 09:08:47 UTC
Just had this happen while moving a window from a screen to another in window manager, no 3D or acceleration anywhere. So seems like this isn't at least limited to fullscreen/opengl applications.
Comment 9 Tomi Pieviläinen 2012-11-11 09:44:07 UTC
After removing a discrete nvidia card from the system, the problem disappeared in all kernels. Also 3.7-kernels started to boot without panicing...

Odd though, since the card wasn't used to ouput any video, only for computation (and didn't have to be used for that either for these problems to occur).

But I would still like to debug this problem, as I would like to run cuda on that card while still using the APU to run the monitors. I just don't know what steps to take, so hints are appreciated.
Comment 10 Alex Deucher 2012-11-11 13:47:55 UTC
(In reply to comment #9)
> After removing a discrete nvidia card from the system, the problem
> disappeared in all kernels. Also 3.7-kernels started to boot without
> panicing...
> 
> Odd though, since the card wasn't used to ouput any video, only for
> computation (and didn't have to be used for that either for these problems
> to occur).
> 
> But I would still like to debug this problem, as I would like to run cuda on
> that card while still using the APU to run the monitors. I just don't know
> what steps to take, so hints are appreciated.

Ah, ok, that's the problem.  Try putting the nvidia card into another slot.  Some of the APU display lanes are shared with pcie lanes so you can only use one or the other.  On your board it looks like the DP connector is using the shared lanes.  The bios should disable the connectors served by those lanes when a pcie card is present, but it looks like your board is not disabling them properly.  You might try a newer bios revision if one is available, or moving the discrete card to another slot, or checking to see if there are any special bios options regarding the pcie slot and display lanes.
Comment 11 Tomi Pieviläinen 2012-11-11 15:02:29 UTC
 steps to take, so hints are appreciated.
> 
> Ah, ok, that's the problem.  Try putting the nvidia card into another slot. 
> Some of the APU display lanes are shared with pcie lanes so you can only use
> one or the other.  On your board it looks like the DP connector is using the
> shared lanes.  The bios should disable the connectors served by those lanes
> when a pcie card is present, but it looks like your board is not disabling
> them properly.  You might try a newer bios revision if one is available, or
> moving the discrete card to another slot, or checking to see if there are
> any special bios options regarding the pcie slot and display lanes.

That is correct, using the first PCIeX16 slot disables the DP output
completely (unless forced to use the APU output in EFI configuration,
then naturally disabling the discrete card).

But that is not what's happening here: the discrete card is in the
second PCIeX16 slot that doesn't (according to manual) conflict with
the APU and the DP display does work until "something" happens.
Browsing through menus in xbmc is a sure way to reproduce the problem,
but it does manifest in normal desktop use too.

Oh, and the obligatory "it worked on Windows", so I doubt it's a hardware problem ;).
Comment 12 Christopher M. Penalver 2016-02-25 07:14:54 UTC
Tomi Pieviläinen, Ubuntu Quantal reached EOL on May 16, 2014. For more on this, please see https://wiki.ubuntu.com/Releases .

If this is reproducible on a supported release, it will help immensely if you filed a new report with Ubuntu by ensuring you have the package xdiagnose installed, and that you click the Yes button for attaching additional debugging information running the following from a terminal:
ubuntu-bug xorg

Also, please feel free to subscribe me to it.

For more on why this is helpful, please see https://wiki.ubuntu.com/ReportingBugs.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.