[mythtv] [EXPerimental PATCH] BIG DVB PATCH V1.1

Jesper Sörensen jesper at datapartner.se
Mon Nov 8 21:51:07 UTC 2004


Taylor Jacob wrote:

>>Is it safe to ship over a PMTObject* pointer to the CAM class via a QT
>>signal? I guess I would need to send over the mutex handle too, and get
>>a lock before using the object? Or would locking the object cause
>>problems for the siparser? Any thoughts?
>>    
>>
>
>I had been thinking that there was no need to use a signal here.. If you are
>using the PMT for tuning then you need to wait for it to show up.. I haven't
>ever seen it take a second even here in my tests, so this waiting doesn't seem
>unreasonable.. It of course never beat the DB Cached version when I was testing
>it, but it never seemed to be such a slowdown that it was worth worrying about..
>
>Since Cam hangs off of channel (MAYBE it should go with recorder, but it doesn't
>exactly fit either place so i say leave it in channel).. You could just directly
>send the data to cam when its ready (via a pointer if you want).. I.E. in
>channel
>
>cam->SetPMT(PMTObject);
>  
>

Yeah, that would be even easier! But the code in dvbchannel only runs 
upon tuning, and I think we need to track PMT updates/changes somehow, 
right? Let's say I'm tuned to service 4711 and we set up the CAM after 
tuning and everything is working great, when suddenly service 4711 
changes (a new audio stream with AC3 sound, a new subtitling stream or 
whatever). Then the CAM would need a new PMT so that the new streams get 
decrypted. But maybe this is too advanced? I don't even know how the DVB 
recorder would react to something like that.

>This way there might be some way in the future for checking for authorization
>from the cam.. Which in the long run would allow you to pop a message to the
>frontend saying "Channel Not Authorized" or "Waiting to Decrypt".. I don't know
>if this is possible with the CAM code in place now, but would allow for this if
>after you set the PMT object you check for authorization..
>  
>

That would be nice indeed! I don't know if/how we get the result from 
the CAM but it's worth looking into. :-)

>Is that clear as mud?  I was quite long winded..
>  
>

No problem - I like it long winded. 8-)



More information about the mythtv-dev mailing list