GigEVision support

Falco Hirschenberger falco.hirschenberger at itwm.fraunhofer.de
Thu Jul 9 15:13:53 UTC 2009


Greg KH schrieb:
> On Thu, Jul 09, 2009 at 03:11:35PM +0200, Falco Hirschenberger wrote:
>> Matthias Urlichs schrieb:
>>> Hi,
>>>
>>> Falco Hirschenberger:
>>>>> Or are you referring to the fact that the device should show up as a
>>>>> proper v4l2 device to userspace, and not need a custom userspace program
>>>>> interface?
>>>> This would also be an option, but I suppose the v4l2 Interface is not  
>>>> flexible enough for these complex cameras.
>>>>
>>> Why shouldn't it be? You can create your own control commands if the
>>> ones in the standard are not sufficient, and it supports a variety of
>>> video standards.
>>>
>> I must confess I'm have not much experience with the v4l2 API. Perhaps 
>> it's really an option. But that's only a question of the user interface. 
>> it's more important to write a reliable driver and think about the 
>> interface afterwards.
> 
> Heh, no, please don't.  The interface influences how the driver is
> written in many different ways.  So please consider it.

Ok, you're right, that should be considered.

> And I agree with the others, v4l2 is the correct interface, and it
> should be able to handle this type of camera just fine.

Is it correct that, if it is implemented with the v4l2 interface, it is 
necessarily a kernel module, as it will use a /dev/video device? I must 
confess that I am no kernel-hacker or systems C programmer. My lingua 
franca is C++ and the prototype is using boost for network stuff.
So I'll be no great help when implementing it as kernel module with v4il.

Btw. GenICam is something similar to v4il, it is not only a 
configuration, but also an acquisition interface which decouples the 
protocol (firewire, usb, gigevision) from the API [1].


[1] http://en.wikipedia.org/wiki/GenICam

Regards, Falco




More information about the devel mailing list