Hi Ricardo,

On Sunday 11 May 2008, Ricardo Ferreira wrote:
> Hello,
>
> On Sunday 11 May 2008 11:06:46 Laurent Pinchart wrote:
> > Hi Ricardo,
> >
> > On Sunday 11 May 2008, Ricardo Ferreira wrote:
> > > Hi, thanks for the response,
> > >
> > > On Sunday 11 May 2008 09:15:05 Laurent Pinchart wrote:
> > > > Hi Ricardo,
> > > >
> > > > On Wednesday 07 May 2008, Ricardo Ferreira wrote:
> > > > > Hello all,
> > > > >
> > > > > I stumbled upon a problem when using your driver. It seems that it
> > > > > blocks on the first VIDIOC_DQBUF ioctl.
> > > > >
> > > > > I first noticed it in my code using the latest debian
> > > > > linux-uvc-source on an amd64 2.6.24 kernel. I assumed it was my
> > > > > problem so I tried luvcview and noticed that there was always a 2
> > > > > second block the first time that same ioctl was called, but it
> > > > > worked flawlessly after that.
> > > > >
> > > > > After grabing your latest svn driver, luvcview stopped working,
> > > > > blocking on the first call to VIDIOC_DQBUF.
> > > > >
> > > > > I have an integrated webcam:
> > > > > Bus 006 Device 003: ID 05c8:0103 Cheng Uei Precision Industry Co.,
> > > > > Ltd (Foxlink)
> > > > >
> > > > > since I never heard of this manufacturer, I tried a phillips
> > > > > spc1000nc freshly bought. The results were exactly the same.
> > > > >
> > > > > Do you think it has to do with the driver, this particular kernel,
> > > > > amd64 (had some other compatibility issues because of this in the
> > > > > past), or anything else?
> > > >
> > > > Could you please retry with r206 and r207 and compare the results ?
> > > > r207 introduces a patch to drop incomplete frames.
> > >
> > > Yes, that is exactly were it breaks. r206 has that 1-2 second delay the
> > > first time DQBUF is called, but after that it is perfect. in r207 the
> > > first call to DQBUF never returns.
> >
> > Thanks for your fast reply.
> >
> > > > If reverting to r206 helps, I'll ask you to try a patch to get more
> > > > information about the issue.
> > >
> > > sure, whatever helps.
> >
> > I've attached a patch to this e-mail. Could you please apply it, load the
> > driver with the trace parameter set to 143 and report the kernel log
> > messages ?
>
> many of these keep repeating:
>
> uvcvideo: Frame complete (EOF found).
> uvcvideo: EOF in empty payload.
> uvcvideo: incomplete buffer (153600 bytes, expected 168960)
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Dropping payload (out of sync).
> uvcvideo: Frame complete (EOF found).
> uvcvideo: EOF in empty payload.
> uvcvideo: incomplete buffer (153600 bytes, expected 168960)
>
> Anything else I can do?

Looks like a bug in the camera :-/

Could you try with different resolutions and report the number of bytes 
received and expected number of bytes for each resolution supported by the 
camera ? I'd appreciate if you could perform the test for both the Foxlink 
and Philips cameras.

Please also post the output of lsusb -v for both cameras.

Best regards,

Laurent Pinchart
_______________________________________________
Linux-uvc-devel mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/linux-uvc-devel

Reply via email to