Menu

#260 Xorg loading the wacom driver under kernel-3.17.1 crashes X

closed-works-for-me
nobody
None
2015-12-04
2014-10-22
Mike Auty
No

Hiya,

So I'm trying to get the 3.17 kernel working on my tablet (having already run into https://bugzilla.kernel.org/show_bug.cgi?id=85891 on my desktop). I've now found that the hid_wacom module will cause Xorg to become a "hung task" as detected by the kernel with DETECT_HUNG_TASK enabled (enters the uninterruptable D state indefinitely). No usable stacktrace is reported. The attached file is the systemd log output after starting X successfully and then running "modprobe wacom". I believe this is an Xorg issue, but if it's a kernel driver issue, please let me know and I'll report it on the kernel bugzilla like the other one.

I can still SSH into the machine but the screen is completely frozen (as is console switching). I've no idea how to debug this more accurately, but I'm happy to run tests or test patches if any of that may help, just let me know what I can do...

1 Attachments

Discussion

  • Jason Gerecke

    Jason Gerecke - 2014-11-14

    Freezing in the 'D' state sounds more likely to be an issue with the kernel driver, but more information is required to nail down exactly where things are going wrong. Since I've not had any luck reproducing this, please try creating an /etc/X11/xorg.conf.d/99-wacom-debug.conf file with the following contents:

    Section "InputClass"
        Identifier "Wacom USB debug class"
        MatchDriver "wacom"
        Option "DebugLevel" "11"
        Option "CommonDBG" "11"
    EndSection
    

    This will cause the driver to emit significantly more information into the logs and should give a better idea of where the bad call is occurring. One the file is in place, restart the X server and grab the logs over SSH as you did before.

     
  • Jason Gerecke

    Jason Gerecke - 2015-06-22
    • status: new --> closed-works-for-me
     
  • Jason Gerecke

    Jason Gerecke - 2015-06-22

    Closing as works-for-me since the bug was not reproducible and a number of modifications have been made to the kernel driver since the bug was reported. Please re-open if this is still an issue.