Moreover this could possibly unveils a bug, as the behaviour has nothing rational.
I was able to reproduce the same behaviours with late january CVS, 0.16, and now I'm working with the sources of 0.17.
My system is Debian pure64 (CPU is AMD64) with kernel 2.6.10, ivtv is 0.2.0-rc3f (I also tried with 0.1 driver, 0.3 driver, and found nothing usefull)
The description: basically, I'm *almost* able to get LiveTV working. Most often, I can have it (everything fine with sound and channel switching) after a cold boot.
Starting b/e then f/e in verbose mode gives the following :
2005-02-16 04:18:56.079 Changing from None to WatchingLiveTV
2005-02-16 04:18:56.087 Using profile 'Live TV' to record
2005-02-16 04:18:56.088 Channel(/dev/video0)::CheckSignalFull(): input = 4
2005-02-16 04:18:56.116 2 ok
2005-02-16 04:18:56.148 135 []:[] []:[] []:[] []:[] []:[] []:[]F2[]:[]/hom...
2005-02-16 04:18:57.098 6 128000
2005-02-16 04:18:57.103 7 0[]:[]0
2005-02-16 04:18:57.111 6 128000
2005-02-16 04:18:57.301 6 128000
2005-02-16 04:18:57.446 6 128000
2005-02-16 04:18:57.662 6 128000
2005-02-16 04:18:57.792 6 128000
2005-02-16 04:18:57.925 6 128000
2005-02-16 04:18:58.132 6 128000
2005-02-16 04:18:58.262 6 128000
2005-02-16 04:18:58.268 8 0[]:[]34
2005-02-16 04:18:58.278 78 0[]:[]0[]:[]0[]:[]38[]:[]0[]:[]1[]:[]0[]:[]3441...
2005-02-16 04:18:58.406 6 128000
2005-02-16 04:18:58.464 1 1
2005-02-16 04:18:58.471 1 1
2005-02-16 04:18:58.478 2 25
2005-02-16 04:18:58.485 7 Tuner 0
2005-02-16 04:18:58.492 2 ok
2005-02-16 04:18:58.596 8 0[]:[]41
2005-02-16 04:18:58.601 25 0[]:[]3[]:[]0[]:[]1130534
2005-02-16 04:18:58.618 6 128000
2005-02-16 04:18:58.742 6 128000
2005-02-16 04:18:58.886 6 128000
2005-02-16 04:18:59.019 6 128000
And frontend seems to do its job.
--> Now when the strange behaviour occurs, starting LiveTV gets nothing but a black screen. Still, my logs gets filled by ivtv so I guess the system is using it.
In that case, the backend verbose is a bit different. It looks like this:
2005-02-16 04:23:51.914 Changing from None to WatchingLiveTV
2005-02-16 04:23:51.927 Using profile 'Live TV' to record
2005-02-16 04:23:51.928 Channel(/dev/video0)::CheckSignalFull(): input = 4
2005-02-16 04:23:51.960 2 ok
2005-02-16 04:23:51.992 135 []:[] []:[] []:[] []:[] []:[] []:[]F2[]:[]/hom...
2005-02-16 04:23:52.936 6 128000
2005-02-16 04:23:53.140 6 128000
2005-02-16 04:23:53.284 6 128000
2005-02-16 04:23:53.489 6 128000
2005-02-16 04:23:53.619 6 128000
2005-02-16 04:23:53.764 6 128000
2005-02-16 04:23:53.896 6 128000
2005-02-16 04:23:54.089 6 128000
2005-02-16 04:23:54.218 6 128000
2005-02-16 04:23:54.376 6 128000
2005-02-16 04:23:54.570 6 128000
2005-02-16 04:23:54.698 6 128000
2005-02-16 04:23:54.856 6 128000
2005-02-16 04:23:55.051 6 128000
2005-02-16 04:23:55.168 6 128000
2005-02-16 04:23:55.300 6 128000
2005-02-16 04:23:55.457 6 128000
2005-02-16 04:23:55.648 6 128000
2005-02-16 04:23:55.780 6 128000
2005-02-16 04:23:55.936 6 128000
Then NOTHING more. It seems to hang. Well actually it does not completely as I can see the QueueJob messages every minutes.
Notice the difference between the OK and the noOK : mythbackend does not show such line as
2005-02-16 04:18:58.478 2 25 2005-02-16 04:18:58.485 7 Tuner 0 2005-02-16 04:18:58.492 2 ok 2005-02-16 04:18:58.596 8 0[]:[]41 2005-02-16 04:18:58.601 25 0[]:[]3[]:[]0[]:[]1130534
when it fails. I guess it struggles to initialize tuning the channel... or?
Exiting b/e end f/e then starting mplayer /dev/video shows the correct LiveTV.
I don't know if this relies on the amd64 architecture. I found no other topic describing the same issue. Now after more that a week hunting the bug (or is it?), I'm out of clue.
Can anyone give me some new direction to look at?
PS : I may add that unfortunately I was totally unable to catch any difference in the system use between OK and noOK. It's really "sometimes if works, sometimes it doesn't" :-/
_______________________________________________
mythtv-dev mailing list
[email protected]
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
