[pvrusb2] pvrusb2-dev-0c73d0aa725b

[Mike Isely]

Sorry, I failed to notice this message earlier.  See further...

On Mon, 11 May 2009, Martin Dauskardt wrote:

> > We need to figure out if this is kernel-specific or something in the 
> > v4l-dvb repo.  Right now all I can tell you is what I think I said 
> > before: the tda9887 module appears to be reporting an I/O error when it 
> > thinks it should have transferred 4 bytes over I2C.  Obviously if the 
> > pvrusb2 driver's I2C adapter were broken we'd be seeing A LOT more 
> > problems than this.  So the tda9887 module might be doing something 
> > wrong.
> > 
> > I haven't seen this problem HOWEVER I haven't really tried that hard 
> > yet, and I'm not running under 2.6.29 yet.  This could be specific to 
> > that kernel.  Any chance you could try this with the same v4l-dvb under 
> > 2.6.27.x?
> > 
> Hi Mike,
> 
>  I tested the same hg with 2.6.27.23:
> - the tda9887 errors are still there
> - the audio does NOT disappear when switching from radio to TV
> 
> So the tda9887 i2c errors seem to be a different issue.

Not surprised to learn that.


> 
> I am very surprised that the kernel itself can have such an influence on the 
> behaviour of  v4l2 drivers. The kernel is build without any v4l2 drivers, it 
> is sure that they come from the pvrusb2-dev-0c73d0aa725b hg snapshot.

I agree that this seems inexplicable.  The pvrusb2 driver is relying on 
other pieces for routing audio but as you say this all comes from the 
v4l-dvb repo (exact module depends on the hardware you are using).


>  
> What could be the next steps to debug the problem? But I guess I won't have 
> much time until next weekend. 

I can try to reproduce it here.  I have a 2.6.29.3 kernel I can try now 
- however this kernel seems to have other issues, like crappy CPU 
scheduling behavior and file system corruption.  (Well I'm not really 
sure but the problems started when I moved up to that from 2.6.27.13 and 
after dropping back to 2.6.28.10 things so far are better.)  Anyway, 
which exact device are you using?

One thing you can try is to verify that you really aren't building 
anything v4l-dvb'ish in 2.6.29.  Especially if you're compiling stuff 
into the kernel and don't realize it - then you may have a hard time 
discovering the problem.  My reproducing the problem here would probably 
help in that regard.

  -Mike

-- 

Mike Isely
isely @ isely (dot) net
PGP: 03 54 43 4D 75 E5 CC 92 71 16 01 E2 B5 F5 C1 E8