It would be helpful in the future, if the bug went through SEG so that the problem can be triaged. The bug is missing information about the version of the various libraries involved (gnome-settings-daemon and gnome-desktop versions at the very minimum). It would also have been helpful to know which graphics driver the customer was using. I'm guessing it's one that doesn't support XRandR.

It seems that the customer is pressing the "Fn+F7" key (that's usually a way to trigger video outputs to be checked), so the immediate mitigation would be to stop doing that, and use xrandr in the meantime.

Created attachment 1009432 [details]
Install rpm packages info on that system

Created attachment 1009433 [details]
lspci

Created attachment 1009434 [details]
modinfo from that system

Created attachment 1009435 [details]
Xorg

Request info:  These informations are comes from the sosreport of that system which encounter this problem.
1. The gnome-settings-daemon and gnome-desktop version is here:
gnome-settings-daemon-2.28.2-30.el6.x86_64 
gnome-desktop-2.28.2-11.el6.x86_64 

For all other packages versions, please refer to the file attachment "installed-rpms".

2. The customer is using Matrox graphics card.
01:00.1 VGA compatible controller: Matrox Electronics Systems Ltd. MGA G200EH

See the file attachment "lspci" for detail.

3. This Matrox graphics card driver is xorg-x11-drv-mga-1.6.1-7.el6.x86_64.

4. I attach the xorg log file to this case also, for your reference.

5. ccpp-2015-01-20-14_3A19_3A45-62345.zip


Reproducible Steps:
1. While this error message occur, customer is remoting into this system by a XManager client. This maybe relative.
2. We could not reproduce this problem at this moment, even we remote into this system, and Press "Fn + F7".
3. This error message abrtd info only once so far as I know.


BR,
Nolan Wai

I'm guessing there's no support for XRandR on that device. Nevertheless, it is indeed possible for the D-Bus objects to be registered (and thus called) when screens cannot be detected.

In _new():
2092                 if (!register_manager_dbus (manager_object)) {•

And in start():
1897         manager->priv->rw_screen = gnome_rr_screen_new (•
1898                 gdk_screen_get_default (), on_randr_event, manager, error);•
1899 •       
1900         if (manager->priv->rw_screen == NULL)•
1901                 return FALSE;•

handle_fn_f7() will then be called, with manager->priv->rw_screen == NULL.

We need to review all of the plugin's entry points and ensure that we bail when rw_screen is NULL.

Hi Bastien and Bugzilla Engineer team,

Any update after you review all these data/information I provided? I hope we can more this bugzilla ticket forward.   Some suggestion/findings are welcomed.

BR,
Nolan Wai

(In reply to Nolan Wei from comment #12)
> Hi Bastien and Bugzilla Engineer team,
> 
> Any update after you review all these data/information I provided? I hope we
> can more this bugzilla ticket forward.   Some suggestion/findings are
> welcomed.

There's nothing to add. The crash happens because the user is pressing the "Fn+F7" key (the "check for external displays" key on non-Thinkpad laptops), and the graphics card doesn't have XRandR support.

You should verify that this is the case, and possibly disable the "xrandr" gnome-settings-daemon plugin if that's the case.

Thanks for the reply. I already sent the steps to user, about disable the "xrandr" gnome-settings-daemon plugin.

But, for the condition/steps to trigger this issue, "the 'Fn+F7' key and graphics card doesn't have XRandR support", could this define as BUG? Or this is a work as design trigger condition?

Because user want to know the detail date of how this BUG(if yes) could be fixed.

It will be fixed in RHEL 6.7 if Product Management thinks it's a big enough problem.

The right way to fix this is to ask NVidia to support XRandR. Whatever we do is going to make those functionalities a stub.

(In reply to Nolan Wei from comment #14)
> Thanks for the reply. I already sent the steps to user, about disable the
> "xrandr" gnome-settings-daemon plugin.
> 

how to setting?

> But, for the condition/steps to trigger this issue, "the 'Fn+F7' key and
> graphics card doesn't have XRandR support", could this define as BUG? Or
> this is a work as design trigger condition?
> 
> Because user want to know the detail date of how this BUG(if yes) could be
> fixed.

Use this command, to disable the xrandr" gnome-settings-daemon plugin:
# gconftool-2 -s /apps/gnome_settings_daemon/plugins/xrandr/active -t bool false

To make sure whether the "xrandr" plugin has been disabled, use this command:
# gconftool-2 --get /apps/gnome_settings_daemon/plugins/xrandr/active
false

If system rebooted, this setting will persistent.


On the other hand, NVidia support XRandR will be another way to fix this issue, then while Fn+F7 to make gnome switch mode will not trigger any problem.

gnome-settings-daemon-2.28.2-33.el6 should fix the crash.

Waiting on feedback about this...

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2016-0962.html