very high processor wake-ups/sec

[royden]

Using powertop, it appears that coccinella 0.9.18 causes a large number of processor wake-ups per second - on my system ca. 195 and by far the top ranker. This count is far more than that of say, sip-communicator or empathy connecting to the same server (Openfire) and is not "lap-top on a battery" friendly.

Ubuntu 10.04, 2.6.34-020634rc5-generic (kernel irrelevant IME)
Coccinella Linux binary from Sourceforge
Tcl8.5 & Tk8.5

[buzzdee (sebastia)]

on my linux desktop, it looks like this running from source:
Top causes for wakeups:
  31.7% (291.5) <interrupt> : timer
  13.9% (127.5) <interrupt> : HDA Intel, i915@pci:0000:00:02.0
  13.7% (126.3) wish : do_nanosleep (hrtimer_wakeup)
   5.8% ( 53.0) firefox : futex_wait (hrtimer_wakeup)
   5.4% ( 49.2) <interrupt> : resched

so not the highest, but anyways, not so good. need to figure out where the nanosleep is done.

[buzzdee (sebastia)]

moving the tcliaxclient extension out of the way drops the power consumption to be < 1%.
So the problem seems to be within iaxclient library or the tcliaxclient extension.

[sander (s-devrieze)]

Maybe this can be tested using the example iaxclient application?