Re: usb: gadget: composite: Return bcdUSB 0x0310

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Mon, Apr 11, 2016 at 02:28:08PM -0700, John Youn wrote:
> On 4/8/2016 5:58 PM, Greg Kroah-Hartman wrote:
> > On Fri, Apr 08, 2016 at 04:19:40PM -0700, John Youn wrote:
> >> Hi,
> >>
> >> I'd like to request the following be applied to stable trees 3.2+.
> >>
> >> 1a85329171094951956a37acc8abb7e51c1e742e ("usb: gadget: composite:
> >> Return bcdUSB 0x0310")
> >>
> >> The USB 3.1 specification replaces the USB 3.0 specification and all new
> >> devices that are running at SuperSpeed or higher speeds must report a
> >> bcdUSB of 0x0310.
> > 
> > That doesn't seem very "backwards compatible" of the spec to me, what
> > happens if you return a 3.0 string, when you really are only handling
> > 3.0 things like we do in older kernels?
> > 
> > thanks,
> > 
> > greg k-h
> > 
> 
> Hi Greg,
> 
> I believe there would be no backward-incompatible change as all 3.0
> things are now part of the 3.1 spec, and the new stuff in 3.1 (such as
> SuperSpeedPlus) are exposed through new descriptors. And I don't
> believe anything currently looks at bcdUSB 3.0 vs 3.1 (unlike with USB
> 2.x). But I can't say for sure.
> 
> This is only needed because the CV tool has started requiring this and
> I believe the compliance process will also start requiring it for all
> new devices. We plan to certify new devices with kernels as old as
> 3.18, hence this backport.

For "new" devices, why are you using such an old and obsolete kernel as
3.18?

For an "old" device, a bcdUSB of 3.0 should just be fine, right?

> If you like, I can solicit comments on this from the linux-usb list.

Please do, you need the maintainer of the subsystem to agree to this at
the least :)

thanks,

greg k-h