[mythtv] Random blockiness that probably isn't CPU or PCI bus bandwidth-related

John Patrick Poet john at BlueSkyTours.com
Wed Jan 19 02:26:59 EST 2005


Eric Anderson wrote:

> BTW - With my version I don't get any buffer overruns at the end of a 
> recording. (I have the
> reader thread close() the device so we don't have the case where the 
> main thread tells the
> reader to stop reading, process remaining data, then close -- during 
> which time the device
> driver hits an overrun.) I *do* still sometimes get a buffer overrun 
> when multiple recordings
> start up. But it's normally a one-time thing.



Wow!  That is so obvious.  I just added a close to the end of the 
fill_ringbuffer function, and hammered on it.  I cannot get it to 
produce an overrun message now.  Thanks.

>
> I'm wondering if *any* buffer overrun poisons the device driver for 
> later. (Just a theory...)
> Because with the CVS version, I pretty much always get a buffer 
> overrun at the end of
> the recording. Like you said (below), maybe the 1.6 driver fixes some 
> stuff.


I did too, until you told me to add the close() :-[

>
> Have you tried the v1.6 driver?  It seems to work well for me -- 
> although it may be why I see an occasional overrun now, when I didn't 
> a couple of weeks ago (I was using the 1.5 driver).
>
>>
>
> I haven't. But I think I reviewed diffs of the code from 1.4 to 1.5 to 
> 1.6 and I didn't see anything related
> to the CX8800 stuff. I am planning to try that out, though. Just 
> haven't gotten to it yet.


I did not compare the versions, I just know that I have never had a show 
become corrupt since switching to 1.6.  Very possible that it is just a 
coincidence, though.

After reading the forums on the pcHDTV site, as well as the Myth mailing 
list, I have become convinced that the motherboard is a critical factor 
in how well the pcHDTV cards work.  I am thinking that the reason our 
different versions of the ringbuffer code behave different for each of 
us, has to do with the other hardware in our systems.  Although, with 
three cards going, I would expect my system to be more demanding than yours.

Thanks for pointing out the missing close.

John



More information about the mythtv-dev mailing list