Re: Re(2): Re(2): Re(2): Quickcam device driver for Linux
> Is the switch to kernel space necessary in Linux? I assumed that the
> ioperm() call gave permissions to the access the parallel port i/o addresses
> to the user-mode process.
It would certainly appear that way from the generated asm code. The five
x86 Unixes that cqcam runs on all appear to let userspace programs do I/O.
The I/O itself is actually about as efficient as polling a PS/2-era
parallel port can get.
> interface. I doubt that there is enough in common between them
> and the Quickcam that you will be able to write a single /dev/camera
The device driver is either compatible or minimal in this case, but not
both. To provide a uniform interface, it's going to have to translate
(and even resize or quantize!) the image before providing it as "raw"
(since each device defines "raw" differently) output.
If you must write camera drivers, I would be in favor of making them
device-specific. If you want to provide a common interface to
applications, write a userspace daemon (or plug-in) for each device.