trying to find causes for high cpu usage

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Wolfgang Rohdewald wrote:
> I added some debug output for every 1000th call to poll()
> in cSectionHandler::Action
> 
> At 23:58 a recording started. Since then the thread for
> my 2nd device (first and primary is Nexus-S FF, 2nd is TT PCI Budget)
> also polls much more even after the recording finished.
> 
> Why is there so much activity going on for the 2nd device even when
> it is no longer needed?
> 
> Is there any way to tell vdr that no output is needed such that
> vdr can stop processing the data stream of the first device too?
> 
> Why does the first device call poll() twice as much as the second one?
> 
> Mar 22 23:57:55 mm vdr[20430]: pcount in cSectionHandler::Action:12209000
> Mar 22 23:57:59 mm vdr[20430]: pcount in cSectionHandler::Action:12210000
> Mar 22 23:58:00 mm vdr[20426]: timer 1 (28 0000-0055 'Zimmer frei!') entered VPS margin
> Mar 22 23:58:00 mm vdr[20426]: switching device 2 to channel 28
> Mar 22 23:58:02 mm vdr[20430]: pcount in cSectionHandler::Action:12211000
> Mar 22 23:58:05 mm vdr[20430]: pcount in cSectionHandler::Action:12212000
> Mar 22 23:58:08 mm vdr[20433]: pcount in cSectionHandler::Action:1000
> Mar 22 23:58:08 mm vdr[20430]: pcount in cSectionHandler::Action:12213000
> Mar 22 23:58:11 mm vdr[20430]: pcount in cSectionHandler::Action:12214000
> Mar 22 23:58:13 mm vdr[20433]: pcount in cSectionHandler::Action:2000
> Mar 22 23:58:14 mm vdr[20430]: pcount in cSectionHandler::Action:12215000
> ...
> Mar 23 00:00:00 mm vdr[20426]: timer 1 (28 0000-0055 'Zimmer frei!') start
> Mar 23 00:00:00 mm vdr[20426]: record /video0/Zimmer_frei!/2005-03-23.00:00.50.99.rec
> ...
> Mar 23 10:35:25 mm vdr[20430]: pcount in cSectionHandler::Action:24078000
> Mar 23 10:35:26 mm vdr[20433]: pcount in cSectionHandler::Action:6408000
> Mar 23 10:35:29 mm vdr[20430]: pcount in cSectionHandler::Action:24079000

The cSectionHandler::Action() has nothing to to with whether there is a
recording going on or not. It has to do with the EPG and other data.

If I'm not mistaken there is a difference in the packet sizes transferred
between FF and budget cards, maybe that's what's causing this.

Klaus


[Index of Archives]     [Linux Media]     [Asterisk]     [DCCP]     [Netdev]     [Xorg]     [Util Linux NG]     [Xfree86]     [Big List of Linux Books]     [Fedora Users]     [Fedora Women]     [ALSA Devel]     [Linux USB]

  Powered by Linux