After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 794106 - Disable wayland when mgag200 graphic card is detected.
Disable wayland when mgag200 graphic card is detected.
Status: RESOLVED OBSOLETE
Product: gdm
Classification: Core
Component: general
unspecified
Other Linux
: Normal normal
: ---
Assigned To: GDM maintainers
GDM maintainers
Depends on:
Blocks:
 
 
Reported: 2018-03-06 08:55 UTC by Yifan J
Modified: 2018-05-24 11:47 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Patch to disable wayland on runtime. (2.18 KB, patch)
2018-03-06 08:58 UTC, Yifan J
none Details | Review

Description Yifan J 2018-03-06 08:55:42 UTC
In our SUSE's SLE 15 beta tseting, we received quite some GNOME issues caused by Wayland + mgag200 combination. Quote:

> "SLES 15 Beta 5. The gnome desktop keyboard and mouse on the server is
> extremely slow responding with all Lenovo systems I have tried. (The most
> recent systems I have tested are ST550 and x3650 M5). This includes several
> new purley-based systems and older broadwell/haswell systems as well. All of
> these system use Matrox graphics controllers. "

> "With sles15 beta4 I am observing both functionality and performance issues on
> the HPE ilo graphical console when graphical.target is chosen.  Interaction is
> very sluggish, mouse very slow to follow physical mouse movement, keyboard is
> subject to rapid, repeated keystrokes.  For example, I noticed the
> sluggishness, opened a terminal window, intending to type "ps" and got
> "pppppppps".  The first character typed after a pause seems most likely to
> repeat. "

> "sles 15 beta 5 gnome keyboard and mouse are extremely slow with Matrox
> graphics controllers"

So it should make sense to disable wayland by now on those chips using similar code path in:

bgo#789081 gdm should never try wayland on Cirrus chipset
Comment 1 Yifan J 2018-03-06 08:58:34 UTC
Created attachment 369373 [details] [review]
Patch to disable wayland  on runtime.
Comment 2 Ray Strode [halfline] 2018-03-07 14:38:05 UTC
hmm comment 0 doesn't sounds like a mga specific problem though.  some sort of mutter bug? maybe bug 774989 coming back?

i think we should fix the bug instead of just cutting out slow hardware that happen to see it more frequently
Comment 3 Yifan J 2018-03-12 04:11:14 UTC
Hi Ray,

Thanks for the tracking on it. While I didn't observe things like bug 774989 in machines, except received similar issues with mgag200 at the moment. So I've been suspecting it's hardware specific. Do you have ideas to gather ad-hoc logs which might be more helpful for debugging from mutter/clutter side?

Meanwhile I saw an issue for the keyboard/mouse slow issue is opened in gitlab:

https://gitlab.gnome.org/GNOME/mutter/issues/59

Maybe it's better to track the issue there?
Comment 4 Felix Zhang 2018-04-08 12:16:08 UTC
For another datapoint, same problem is also reported against ASPEED ast drivers:
https://gitlab.gnome.org/GNOME/mutter/issues/106
Comment 5 GNOME Infrastructure Team 2018-05-24 11:47:10 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to GNOME's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.gnome.org/GNOME/gdm/issues/364.