[cpia] EPP (as opposed to ECP) parport support for cpia_pp ?

[Peter Pregler]

On Fri, 25 Oct 2002 22:37:43 -0400 (EDT)
Duncan Haldane <f.duncan.m.haldane@worldnet.att.net> wrote:

> I was tying to test the 2.5.44 cpia code with the Creative WebCam II
> (paralell port) camera, but the machine I set 2.5.44 up on only
> supports PCSPP and EPP paralel port modes.
> 
> The cpia_pp code only seems to support ECP and TRISTATE parport modes.

Correct.
 
> The Webcam Works fine using windows, but not under linux with ether
> the 2.4.x or 2.5.x cpia drivers, because the "paralel port is nor
> ECP-capable".
> 
> Presumably it *could* work with EPP protocol, but none of the original
> developers had an EPP parport?  is TRISTATE a subset of EPP that would
> work?

It could work. EPP is somewhat close to ECP in capabilities. The
TRISTATE is far worse and I would not recommend it for streaming but
just for grabbing single pictures. But that you can do also with
userspace-tools (at least someone was writing such stuff).

> Would it be easy to add EPP support to cpia_pp.c?  (I am not really
> interested in running an obsolete camera on an  EPP parport, but I
> could test it)

Well, the basic parport code completely changed from 2.2 to 2.4. In 2.2.
the cpia_pp has somewhat direct access to the parport. In 2.4. there is
an abstract parport-interface and the parport-code itself should
implement and automagically use what is actually available. Actually,
there should not be a message from cpia_pp that it cannot work in 2.5.
If there is such a problem the message should come from the
parport-drivers. Check out the cpia_pp_register. The capability-check
there seems to be overly strict to me. What the cpia-chip really
supports you should be able to find in the chip-documents available at
webcam.sourceforge.net. But as I said you could try the generic approach
and simply remove that and see what the parport-layer tells you.

Unfortunatly, I have neither a parport-camera available anymore. And the
parport-people did never implement all the necessary read-stuff since
they were focused more or less on writing (the printer symptom).
Therefore there is also no DMA-read available in 2.4 which has to be
reimplemented at the generic parport-read layer. And finally similar to
you I am not really interested to write new code for such old stuff.

-Peter

-- 
As an individual my scope is limited.
-------------------------------
Email: Peter_Pregler@email.com