On Tue, Apr 30, 2019 at 10:05:06AM -0500, Bin Liu wrote: > Hi Paul, > > On Thu, Apr 04, 2019 at 11:10:46AM -0500, Bin Liu wrote: > > On Thu, Apr 04, 2019 at 11:37:43AM +0200, Paul Cercueil wrote: > > > The pullup may be already enabled before the driver is initialized. > > > It has to be disabled at init time, as we cannot guarantee that a gadget > > > driver will be bound to the UDC. > > > > > > Signed-off-by: Paul Cercueil <paul@xxxxxxxxxxxxxxx> > > > --- > > > drivers/usb/musb/musb_gadget.c | 1 + > > > 1 file changed, 1 insertion(+) > > > > > > diff --git a/drivers/usb/musb/musb_gadget.c b/drivers/usb/musb/musb_gadget.c > > > index ffe462a657b1..094dc59c834c 100644 > > > --- a/drivers/usb/musb/musb_gadget.c > > > +++ b/drivers/usb/musb/musb_gadget.c > > > @@ -1794,6 +1794,7 @@ int musb_gadget_setup(struct musb *musb) > > > > > > musb->is_active = 0; > > > musb_platform_try_idle(musb, 0); > > > + musb_pullup(musb, 0); > > > > > > status = usb_add_gadget_udc(musb->controller, &musb->g); > > > if (status) > > > > Applied. > > Sorry, I have dropped this patch from my tree. I realized the gadget > core already defines callbacks to control the pullup. I think the > proper fix should be calling usb_udc_vbus_handler() in musb gadget or > glue layers. Basically besides the problem your patch tries to fix, there is another problem related to the pullup, which is enabled right after bind a gadget driver, regardless of vbus status. This violates the USB Spec. Implementing the udc core callbacks should fix both issues. -Bin.
