[mythtv] dvb zero length record

Tj NG htpc at treblid.dyndns.org
Sat May 28 13:15:17 UTC 2005


Tom Hughes wrote:

>I don't see how that patch will help anything either. You've moved
>the Reset call that stops the SIParser to after tuning is complete
>which shouldn't help - if anything it will makes things worse.
>  
>
In a sense I'm hoping it would force the NewPMT signal to emit...Adding 
the 2nd Reset() will eventually make channel.PMTSet == true for both cases.

Without that call, and if  "TuneXXXX" functions in TuneTransport set 
"havetuned" to true, the TuneTransport  function will eventually return 
from the lower part of the code block.. Then "Tune" will always return 
false with that "Timeout Getting PMT" message because the value PMTSet 
is never set to true.

Hopefully I havn't misread the code.. At least this patch doesn't seem 
to break anything for me, still not sure if it actually solves any zero 
byte recordings, as I'm still testing....


>The problem (assuming you have the same problem as me) is picking up
>the PMT from the old multiplex - do you have a siparser log of it
>going wrong? Because that should prove whether it is getting the PMT
>from the old multiplex.
>  
>
I'm now running my backend with "-v siparser"....

>Given that you need to make sure that tuning is complete before
>restarting the SIParser, but the code already seems to do that
>assuming that the lock detection is correct. My theory was that
>the lock detection was broken which is what my patch attempted
>to fix.
>  
>

>I've been running with my patch for two weeks now and haven't lost
>anything yet...
>  
>
Your patch improve things dramatically, but I stil got 1 or 2 zero byte 
recordings in around 15+ hrs of recording... I have been using your 
patch for a while now and it doesn't seem to break anything, perhaps you 
should get Taylor to review your code and commit it to CVS??

Anyway, I'm still actively recording (using ur patch and mine) and will 
post my logs if I got more 0 byte files.

>They need to be channels on different multiplexes as well so the
>card has to retune to a new frequency.
>
>  
>
Yup I am doing that... I live in Australia, every channel is in their 
own multiplex (i.e. every multiplex is showing the same thing in lousy 
SD and crappy HD).. So when I tune to a different channel, I'm tuning to 
a different multiplex...

It seems that nobody else is seeing this problem. Are you using a 
TwinHan card? Could this be a card problem?




More information about the mythtv-dev mailing list