On Fri, Jan 20, 2023, Elson Serrao wrote: > > > On 1/19/2023 5:15 PM, Thinh Nguyen wrote: > > On Thu, Jan 19, 2023, Elson Serrao wrote: > > > > > > > > > On 1/18/2023 5:44 PM, Thinh Nguyen wrote: > > > > On Tue, Jan 17, 2023, Elson Roy Serrao wrote: > > > > > Add a flag to indicate whether the gadget is capable > > > > > of sending remote wakeup to the host. > > > > > > > > > > Signed-off-by: Elson Roy Serrao <quic_eserrao@xxxxxxxxxxx> > > > > > --- > > > > > drivers/usb/gadget/composite.c | 3 +++ > > > > > include/linux/usb/gadget.h | 2 ++ > > > > > 2 files changed, 5 insertions(+) > > > > > > > > > > diff --git a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c > > > > > index 403563c..b83963a 100644 > > > > > --- a/drivers/usb/gadget/composite.c > > > > > +++ b/drivers/usb/gadget/composite.c > > > > > @@ -965,6 +965,9 @@ static int set_config(struct usb_composite_dev *cdev, > > > > > else > > > > > usb_gadget_clear_selfpowered(gadget); > > > > > + if (USB_CONFIG_ATT_WAKEUP & c->bmAttributes) > > > > > + gadget->rw_capable = 1; > > > > > > > > Some device may not support remote wakeup. gadget->rw_capable should be > > > > set and reported by the UDC. May need a gadget ops to enable remote > > > > wakeup here. > > > > > > > > Thanks, > > > > Thinh > > > > > > > Not exactly clear on which parameter in UDC decides whether a device > > > supports remote wakeup. Here I have this flag just to indicate whether the > > > connected device is rw capable based on the bmAttributes populated in the > > > config descriptor. If the UDC doesnt have a callback for remote wakeup we > > > have that check when calling the gadget op in udc/core.c (have added a > > > similar check in usb_func_wakeup() also ). > > > > That flag describes the gadget's capability, not the device > > configuration. However, it's being used as a configuration flag. > > > > Thanks, > > Thinh > > > > Thank you for the clarification. Please let me know if below approach where > we consider both gadget's capability and device configuration fine? > > if (gadget->ops->wakeup || gadget->ops->func_wakeup) > gadget->rw_capable = USB_CONFIG_ATT_WAKEUP & c->bmAttributes ? 1: 0; > The way gadget->rw_capable is named and described, it's a capability flag. That is, its value shouldn't change from the user config. Perhaps we don't need that in the usb_gadget, and we can have something that looks like this: if (gadget->ops->wakeup && (c->bmAttributes & USB_CONFIG_ATT_WAKEUP)) usb_gadget_enable_remote_wakeup(g); else usb_gadget_disable_remote_wakeup(g); The setting of the remote wakeup configuration can be tracked internally by the dwc3 driver based on the usb_gadget_enable_remote_wakeup call. Thanks, Thinh
