On Thu, May 04, 2000, Alan Cox <[EMAIL PROTECTED]> wrote:
> > I'm almost fully finished implementing a patch which provides a device node
> > per endpoint which would remove many of the ioctl's.
>
> You should only need one ioctl() which is the 'what device' you issue first.
Yes, but that extends the already "broken" API. Some people want it changed,
and I can see their point of view. Using ioctl() for everything is ugly.
> > So, this is a no win situation. Keep support for 2.2 in the form of ioctl's
> > and get flamed for making everything an ioctl, or move forward and fix the
> > deficiencies at the cost of making 2.2 support exceedingly difficult.
> >
> > Once again, my opinion is to not hold down USB with legacy baggage in the
> > form of 2.2 support.
>
> I'd like to keep minimal baggage but encourage people to do it the right
> way. The single ioctl then the same API is as close as I can see to avoiding
> the two API problem
If you want to keep usbdevfs only for 2.2, that could be a solution.
JE
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
