Mailing List archive

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

[linux-dvb] Re: Fw: DVB audio/video decoder API



On Friday 01 November 2002 02:18, Ralph Metzler wrote:
> Dennis writes:
>  > 00000000 1965     8.01648     DmxDevBufferRead
>  > 00000710 2150     8.77121     DmxDevTSCallback
>  > 00001ae0 2526     10.3052     DvbDmxSWFilterPacket
>  > 00001740 13985    57.0537     DmxDevBufferWrite

> What context switches? It's all in kernel space.
> DmxDevBufWrite writes the TS packets belonging to each selected PID into
> the corresponding buffer where it then can be fetched by the user process.
> That's the actual demuxing process.

You're right of course, (as usual :P). 

So if I understand things right, the new firmware does create a TS itself? 
Does the driver receive the "full" TS (i.e. all PID's available) or only the 
ones the application asked for? Since VDR asks a TS from the driver, and the 
AV7111 provides a TS (which it created itself) to the driver, what do you 
mean by the actual demuxing process?

I feel like a broken record with this issue, for which I apologize!

What I am trying to find out is if there a reason (i.e. added complexity, 
software (de)muxing, etc) why the HEAD drivers take so much more CPU than the 
old drivers, in the same scenario's (only recording one channel on a 
secondary card, while watching non-timeshifted live TV on the primary card).

I owe a lot of people a lot of beers :)
Thanks again,
Dennis




--
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe linux-dvb" as subject.



Home | Main Index | Thread Index