[linux-dvb] Re: alps_tdmb7 / WinTV Nova-T USB Driver Regression?

[Holger Waechtler <holger@qanu.de>]

Richard Taylor wrote:

> On Wed, Oct 06, 2004 at 11:35:59AM +0200, Holger Waechtler wrote:
>  
> > Patrick Boettcher wrote:
> >
> >    
> > > On Mon, 4 Oct 2004, Richard Taylor wrote:
> > >
> > >
> > >
> > >      
> > > > Certain multiplexes cease to work completely, all have degraded.
> > > > mythbackend gives me a stream of "WARNING - Transport Stream Continuity
> > > > Error. PID = nnn" errors.
> > > >  
> > > >        
> > > Please have in mind that USB is a serial bus, so 1.1 device have to share
> > > the 12MBit/s. If you use them at the same time it is just normal, that
> > > URBs are droped.
> > >
> > >
> > >      
> > Be sure that your USB1.1 devices don't share the same bus, most PCs have 
> > multiple independent USB ports.
> >
> >    
> The devices are on separate controllers (my box has three)
>
> Also, the problem with the WinTV Nova-T is there regardless of whether the second USB receiver is connected or not. 
> The second receiver is a red herring, as it was only the trigger for me to upgrade to the latest driver, not the cause of the problems. 
> Going back to the drivers in kernel 2.6.7, the WinTV card has a few minor errors on the BBC mplex, and is perfect on the ITV one.
> Using the CVS / latest release drivers, the BBC mplex is un-watchable, and the ITV one has lots of glitches.
>
> So there is definitely a change for the worse in the dvb-kernel code between the 2.6.7 kernel drivers, and the CVS / latest release.
>  
Can you please go back in CVS history and try to find out when exactly 
the regression was introduced? You can check out old versions using

$ cvs up -dPD"<date>"

where <data> is a cleartext or symbolic date like "3 days ago" or "4 
months ago". Please check the CVS documentation for more help. thenafter 
you can try to isolate the exact problem using 'cvs diff'...

Holger