[linux-dvb] Regression in cx88_dvb relative 2.6.25?

[linux-dvb] Regression in cx88_dvb relative 2.6.25?

[Sneake]

Has anybody else seen a regression with the recent v4l-dvb code? I have a Pinnacle PCTV HD PCI card:
00:07.0 Multimedia video controller: Conexant CX23880/1/2/3 PCI Video and Audio Decoder (rev 05)
00:07.1 Multimedia controller: Conexant CX23880/1/2/3 PCI Video and Audio Decoder [Audio Port] (rev 05)
00:07.2 Multimedia controller: Conexant CX23880/1/2/3 PCI Video and Audio Decoder [MPEG Port] (rev 05)

It used to work flawlessly under the v4l-dvb branch, and when the code was merged into 2.6.25, then the stock kernel also worked flawlessly.

(all of this is for ATSC streams, not analog NTSC streams).

I then added 2 USB tuners (WinDVR-950 and Pinnacle PCTV-HD 800e), and so I pulled the latest v4l-dvb branch, built it, and installed it: and the PCI card stopped working. Now, trying to capture from the PCI card gives me a large number of errors in the recovered data stream.

Reverting to the stock 2.6.25 kernel modules will restore the PCI card's operation, so that would tend to eliminate user space changes, signal changes, and pretty much everything but the kernel drivers, as far as I can see.

So: has anybody else seen this?

_______________________________________________
linux-dvb mailing list
linux-dvb@linuxtv.org
http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb

Re: [linux-dvb] Regression in cx88_dvb relative 2.6.25?

[Sneake]

On Wed, 2008-06-11 at 22:39 -0400, Devin Heitmueller
devin.heitmueller-at-......... |linux-dvb| wrote:
> Hello there,
> 
> Not that I have reason to believe the problem is with the em28xx
> driver, but could you please reboot the PC without the USB devices
> connected (so we know for sure the driver did not load), and make sure
> you still have the issue?
> 
> I just want to be sure it's the V4L code in general and not some
> interaction between your PCI device and the USB devices.

GOOD GUESS: I did exactly that (removed the two USB devices) and now the
PCI card works.

Now, how the heck can a USB device that isn't even active at the time
screw up the PCI device?

Note: when I went back to the stock 2.6.25 drivers the devices were
still plugged in, they just did not have drivers active. That would tend
to exclude weird stuff like ground loops (which would have still been
present).

Are you thinking an electrical level interaction (noise) or some
software level interaction?



_______________________________________________
linux-dvb mailing list
linux-dvb@linuxtv.org
http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb