> On 2008-09-06 at 13:31:54 [+0200], Salvatore Benedetto < > emitrax@xxxxxxxxx> > wrote: > > 2008/9/6 Michael Lotz <mmlr@xxxxxxxx>: > > > > > > As said by others, the webacm is UVC. But please note that it is > > > probalby a device using isochronous USB transfers which are not > > > really > > > tested for UHCI and not implemented for EHCI. So really I would > > > like to > > > suggest going with ACPI or wireless first. I will eventually get > > > isochronous support going for EHCI, but it's a bit involved to > > > support > > > it correctly as isochronous transfers are far less straight > > > forward to > > > implement. > > > > Actually IMHO I think this could be a good reason to go ahead with > > UVC > > support. Isochronous support hasn't been really tested due to lack > > of > > drivers for devices that actually use this transfer mode. > > +1. :-) I long for a proper UVC driver, but I suppose it is quite a > lot of > work. But even USB TV-sticks would work and all sorts of camera > devices, > even some handycams. I didn't dig UVC much yet but probably if UVC is used for more things than webcams I should fork usb_webcam to make an uvc-specific driver and leave the other one for old-style webcams. Depending how UVC works with respect to existing non-protocols (more like over-usb register read/writes and that's it), it might be cleaner to split. François.