Merge lp:~timchen119/unity-settings-daemon/unity-settings-daemon.1471708.trusty into lp:unity-settings-daemon/14.04
| Status: | Rejected |
|---|---|
| Rejected by: | Iain Lane on 2015-09-02 |
| Proposed branch: | lp:~timchen119/unity-settings-daemon/unity-settings-daemon.1471708.trusty |
| Merge into: | lp:unity-settings-daemon/14.04 |
| Diff against target: |
12 lines (+1/-1) 1 file modified
plugins/xrandr/gsd-xrandr-manager.c (+1/-1) |
| To merge this branch: | bzr merge lp:~timchen119/unity-settings-daemon/unity-settings-daemon.1471708.trusty |
| Related bugs: |
| Reviewer | Review Type | Date Requested | Status |
|---|---|---|---|
| Iain Lane | 2015-07-08 | Needs Fixing on 2015-07-09 | |
|
Review via email:
|
|||
Description of the Change
This one-line change works around a race condition in the intel driver, that causes XOpenDisplay() to hang on resume from S3. Killing input mapping after 3 seconds seems to be enough.
| Alberto Milone (albertomilone) wrote : | # |
| Iain Lane (laney) wrote : | # |
Did anyone try to figure out the underlying cause? Please report a bug (upstream?) and reference it in a comment. We should get this fixed properly. We also need to fix this in vivid and wily.
Onto the code. Does this have to be done synchronously? If not then please use g_spawn_async or similar and a timeout to kill the process. This will avoid blocking. If it fails (hangs) then could retrying succeed? If so then we should do that. We should also log a message if it's killed so that the user has a chance to know what's going on.
| Tim Chen (timchen119) wrote : | # |
Thanks Iain, it was originally a quick workaround for the possible race condition, totally agree that making it asynchronous, adding more verbose log and finding out the root cause are better ways to fix this issue.
| Iain Lane (laney) wrote : | # |
tseliot re-did this


This for LP: #1471708