Hi,
On Wed, Aug 15, 2012 at 11:09:51AM -0400, Alan Stern wrote:
> > FYI, dwc3 would look like this:
> >
> > diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> > index e09a7c4..723a530 100644
> > --- a/drivers/usb/dwc3/gadget.c
> > +++ b/drivers/usb/dwc3/gadget.c
> > @@ -1846,6 +1846,15 @@ static void dwc3_disconnect_gadget(struct dwc3 *dwc)
> > }
> > }
> >
> > +static void dwc3_reset_gadget(struct dwc3 *dwc)
> > +{
> > + if (dwc->gadget_driver && dwc->gadget_driver->reset) {
> > + spin_unlock(&dwc->lock);
> > + dwc->gadget_driver->reset(&dwc->gadget);
> > + spin_lock(&dwc->lock);
> > + }
> > +}
>
> Should we try to maintain backward compatibility by invoking the
> disconnect callback during a reset if the gadget driver doesn't have a
> reset callback? To tell the truth, I don't know to what extent most
> gadget drivers need to know the difference between reset and
> disconnect.
good point :-) Will do that now.
> > @@ -2021,8 +2030,7 @@ static void dwc3_gadget_reset_interrupt(struct dwc3 *dwc)
> > dwc3_gadget_usb3_phy_suspend(dwc, false);
> > }
> >
> > - if (dwc->gadget.speed != USB_SPEED_UNKNOWN)
> > - dwc3_disconnect_gadget(dwc);
> > + dwc3_reset_gadget(dwc);
> >
> > reg = dwc3_readl(dwc->regs, DWC3_DCTL);
> > reg &= ~DWC3_DCTL_TSTCTRL_MASK;
>
> I'll check to see about making a similar change to net2280. dummy-hcd
> should be fairly easy.
thanks a lot. I'll do that for musb ;-)
> > When doing that, I thought we could also drop the "don't call disconnect
> > if we get a reset when gadget never enumerated" thing. What do you say ?
>
> Is there such a rule? I wasn't aware of it. Yes, it seems like a very
Unfortunately, yes. Or at least there used to be. I can't remember the
details now, sorry.
> minor sort of optimization; we could easily get rid of it.
that sounds like a plan :-) Tomorrow I can send a series of patches
starting off the split, then people can just send in more patches for
other controllers as we go. cheers
--
balbi
Attachment:
signature.asc
Description: Digital signature
