On Tue, Jun 04, 2013 at 04:47:18PM -0500, Dan Williams wrote:
> On Tue, 2013-06-04 at 12:18 -0700, Greg KH wrote:
> > On Tue, Jun 04, 2013 at 12:50:06PM -0500, Dan Williams wrote:
> > > Greg,
> > >
> > > Just (finally?) noticed the zte_ev driver was added for v3.7 kernel; and
> > > it contains ID conflicts with 'option' for 0x05c6/0x6000, 0x19d2/0xfffe,
> > > and 0x19d2/0xffff. These devices are listed in both drivers as being
> > > ZTE devices (despite the Qualcomm vendor ID of one), so perhaps the IDs
> > > should be removed from option?
> >
> > Yes they should, sorry for not catching that myself.
> >
> > > Also, some drivers for the AC2726 that I've found also indicate that
> > > it's driven by zte_ev, these drivers have this ID list which greatly
> > > overlaps with option:
> > >
> > > { USB_DEVICE(0x19d2, 0xffff) },// AC8710
> > > { USB_DEVICE(0x19d2, 0xfffe) },// AC8700
> > > { USB_DEVICE(0x19d2, 0xfffd) },// MG880
> > > { USB_DEVICE(0x19d2, 0xfffc) },
> > > { USB_DEVICE(0x19d2, 0xfffb) },
> > > { USB_DEVICE(0x19d2, 0xfff1) },// AC2726, AC8710_V3
> > > { USB_DEVICE(0x19d2, 0xfff6) },
> > > { USB_DEVICE(0x19d2, 0xfff7) },
> > > { USB_DEVICE(0x19d2, 0xfff8) },
> > > { USB_DEVICE(0x19d2, 0xfff9) },
> > > { USB_DEVICE(0x19d2, 0xffee) },
> > > { USB_DEVICE(0x19d2, 0xffed) },// AC2716
> > > { USB_DEVICE(0x19d2, 0xffeb) },
> > > { USB_DEVICE(0x19d2, 0xffec) },
> > > { USB_DEVICE(0x05C6, 0x3197) },
> > > { USB_DEVICE(0x05C6, 0x6000) },
> > > { USB_DEVICE(0x05C6, 0x9008) },
> > >
> > > I've reviewed this ZTE driver for the AC2726 and the existing kernel
> > > zte_ev.c driver and there are no differences in the command data sent to
> > > the device. Would you like me to prepare an update that adds these IDs
> > > to zte_ev.c and removes them from option.c?
> >
> > Yes, that would be great, patches for all of this would be gladly
> > appreciated :)
>
> Next question, does this mean zte_ev is using smaller buffers than
> option, since option uses usb_wwan for most actual tx/rx logic, while
> zte_ev uses only usbserial? Using only usbserial with high-speed
> devices (Mb/s) typically restricts throughput pretty severely, so we
> might risk regression by moving these devices to zte_ev unless we
> convert zte_ev over to usb_wwan.
It shouldn't be that bad these days, as the generic code should be going
a lot faster than the old stuff did (pre 3.3 days). As I don't have one
of these devices to test with, I really don't know. Let's convert them
and see if anyone complains :)
thanks,
greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
