Marty wrote:
> I'm seeing the same DMA errors reported by other readers.  Common factors 
> include the K8T800 chipset and dual tuners.  In one test I forgot to install 
> the 
> ivtv module and got a kernel oops when I started to display a video in 
> mythtv. 
> It was repeatable.  This may be a major bug in the kernel.  I know it's a 
> long 
> shot, but the bug may have first appeared in 2.6.18, since previous kernels 
> seemed more stable with ivtv.  The changelog for 2.6.18 mentions several VIA 
> chipset-related changes.

Today I am trying linux 2.6.24-rc6 for the first time.  After nearly two hours 
of extreme load, I've seen no problems except for rare, brief video and audio 
breaks, and the following occasional errors:

kernel: ivtv0: All encoder VBI stream buffers are full. Dropping data.
kernel: ivtv0: Cause: the application is not reading fast enough.

There errors tend to come in very short bursts, a few second at most, separated 
by randomly sized periods averaging about 10 minutes.  Otherwise VBI seems to 
be 
working fine.

It's possible. though unlikely, that these are caused by heavier CPU and I/O 
stress which I am testing with now, compared to previous tests.  On the other 
hand, if this is a new error, then it increases my suspicion that VBI is 
related 
to the DMA errors, since VBI only started working after the 2.6.17 series 
kernels.

For the record, my previous theory about the kernel oops was wrong.  It's 
probably a bug in the 2.6.18 unichrome driver.

_______________________________________________
ivtv-users mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-users

Reply via email to