dx-packages team mailing list archive

[Launchpad Bug Tracker <1221871@xxxxxxxxxxxxxxxxxx>]

This bug was fixed in the package indicator-power -
12.10.6+14.04.20131125-0ubuntu1

---------------
indicator-power (12.10.6+14.04.20131125-0ubuntu1) trusty; urgency=low

  [ Charles Kerr ]
  * We really only need to keep a handful of properties up-to-date for
    each upower device. Given that upower requires us to refetch all
    properties with GetAll whenever it signals that it's changed, we
    really don't need to keep bus proxies alive for each device.
    Removing the autogenerated proxy code and just using a couple of the
    direct GDBusConnection API calls cuts out some unnecessary overhead
    according to health-check. See the before & after profile logs
    attached to the ticket. (LP: #1221871)
  * fix a minor memory leak that was caused by treating the variant
    returned by g_icon_serialize() as floating.

  [ Ted Gould ]
  * Switching DBus service to an Upstart Job. (LP: #1185565)

  [ Nick Dedekind ]
  * Removed Label from brightness slider. (LP: #1240756)

  [ Ubuntu daily release ]
  * Automatic snapshot from revision 215
 -- Ubuntu daily release <ps-jenkins@xxxxxxxxxxxxxxxxxxx>   Mon, 25 Nov 2013 03:55:08 +0000

** Changed in: indicator-power (Ubuntu)
       Status: Triaged => Fix Released

-- 
You received this bug notification because you are a member of DX
Packages, which is subscribed to indicator-power in Ubuntu.
https://bugs.launchpad.net/bugs/1221871

Title:
  indicator-power-service is a lightly busy all the time

Status in The Power Indicator:
  Fix Committed
Status in “indicator-power” package in Ubuntu:
  Fix Released

Bug description:
  In a vain attempt to reduce power on these phone devices I noticed
  that indicator-power-service is using just under 1% of one CPU all the
  time.

   CPU usage:
    PID  Process                USR%   SYS%  TOTAL%
    1358 indicator-power-serv   0.05   0.04   0.08 (very light load)

  It seems to be waiting on poll() on a couple of threads which
  cumulatively add up to about 1.68 wakeups a second over a 1 hour
  analysis:

  Top polling system calls:

  PID     
    1369 indicator-power-serv poll                    1.3462 polls/sec
    1358 indicator-power-serv poll                    0.3376 polls/sec
   Total                                              1.6838 polls/sec

  This contributes to a bunch of context switching by these threads:

  Context Switches:
    PID  Process                Voluntary   Involuntary     Total
                               Ctxt Sw/Sec  Ctxt Sw/Sec  Ctxt Sw/Sec
    1369 indicator-power-serv        10.75         0.15        10.90 (moderate)
    1358 indicator-power-serv         3.53         0.07         3.60 (low)
    1372 indicator-power-serv         0.00         0.00         0.00 (idle)
   Total                             14.29         0.22        14.50

  [ See the attached log, launchpad formatting sucks ]

  I also observed a couple of brk() calls, and a small amount of heap
  growth which may indicate it has a small memory leak, but I am
  speculating on this.

  Attached is a full report from a 1 hour of monitoring indicator-power-
  service for reference.

To manage notifications about this bug go to:
https://bugs.launchpad.net/indicator-power/+bug/1221871/+subscriptions