gnome-screensaver causes jump in compiz events

Bug #1177632 reported by Paul Larson
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Compiz
New
High
Unassigned
compiz (Ubuntu)
New
Undecided
Unassigned
gnome-screensaver (Ubuntu)
New
Undecided
Unassigned

Bug Description

I have some automated tests that run eventstat to track the average events/second for all running processes. I've been having trouble getting the data to stabilize and showed it to Colin King recently, who seems to have found the culprit. If you look at https://jenkins.qa.ubuntu.com/job/eventstat-raring-desktop-amd64-install-idle-vm/95/artifact/eventstat.csv you will see that on line 8, the events/second for compiz jumps dramatically. Move over a few columns and you'll see that this sudden rise corresponds to gnome-screensaver activating.

This is on a freshly installed raring vm, so it should be the default screensaver.

Paul Larson (pwlars)
tags: added: qa-daily-testing
Revision history for this message
Sebastien Bacher (seb128) wrote :

Thank you for your bug report, what video card/driver is on use on this machine? Does it just spike or creating events while activated? Can't you just inhibit the screensaver for your tests?

Revision history for this message
Paul Larson (pwlars) wrote :

The point of the testing is to use eventstat to find things causing lots of wakeups, which could have an impact on power consumption and battery life. So working around the bug would kind of defeat the purpose of testing.
In this case, it's running kvm so it isn't a real video card.

Revision history for this message
Sebastien Bacher (seb128) wrote :

> So working around the bug would kind of defeat the purpose of testing.

Well, in your original description you raise the issue that this leads to difficulties to get correct datas, so I thought you might want to be able to filter out a known issue... anyway that seems like a compiz issue, there had been similar bugs on some videocard in the past (bugs in the videodrivers), can you reproduce on real hardware (that would help to determine if that's indeed dependant of the video driver)

Revision history for this message
MC Return (mc-return) wrote :

@pwlars:
Could you please rerun your tests with xscreensaver, so we can determine if Compiz is the real culprit before we start investigating this issue ?

Changed in compiz:
milestone: none → 0.9.10.0
importance: Undecided → High
Changed in compiz:
milestone: 0.9.10.0 → 0.9.10.2
MC Return (mc-return)
Changed in compiz:
milestone: 0.9.10.2 → 0.9.11.0
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.