On Wed, Dec 01, 2010 at 10:20:19AM -0500, Alan Stern wrote:
> On Tue, 30 Nov 2010, Sarah Sharp wrote:
>
> > On Tue, Nov 30, 2010 at 11:06:21AM -0500, Alan Stern wrote:
> > > On Mon, 29 Nov 2010, Sarah Sharp wrote:
> > > > @@ -1234,6 +1252,14 @@ struct xhci_hcd {
> > > > u32 suspended_ports[8]; /* which ports are
> > > > suspended */
> > > > unsigned long resume_done[MAX_HC_PORTS];
> > > > + /* Is each xHCI roothub port a USB 3.0, USB 2.0, or USB 1.1 port? */
> > >
> > > This comment is unclear; the port_array field doesn't answer the
> > > question asked in the comment. Furthermore, none of the ports can be
> > > USB 1.1 -- they all have to be USB-2 or USB-3.
> >
> > Ok, on closer look at this, the comment is correct. Each port_array
> > element is taken from the Major Revision field in the Extended
> > Capabilities, so it can be 0x1, 0x2, or 0x3. If the port shows up twice
> > in the Extended Capabilities registers (it could happen, although it
> > should be caught by the USB-IF compliance tests), then that element is
> > set to -1 and the port is ignored. Probably the port should be
> > disabled, but the code doesn't do that.
>
> Ah -- I misread the code. Instead of making port_array a pointer, why
> not make it an array of some maximum fixed length?
I could, but this particular patch might make it into 2.6.37 (to deal
with the USB 3.0 port disabling issue), so I'm not really inclined to
change that.
> Does xHCI allow an implementation to use separate port registers for
> high speed vs. full/low speed?
The xHCI spec doesn't disallow it. I don't see the advantage to having
a LS/FS only port, but it's possible.
Sarah Sharp
--
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
