Re: [PATCH] usb: gadget: udc: skip pullup() if already connected

[Alan Stern]

On Wed, Apr 22, 2026 at 07:31:16PM +0800, Xu Yang wrote:
> On Tue, Apr 21, 2026 at 10:37:53AM -0400, Alan Stern wrote:
> > On Tue, Apr 21, 2026 at 04:20:50PM +0800, Xu Yang wrote:
> > > The device controller may update vbus status via usb_udc_vbus_handler(),
> > > which tries to connect the gadget even though gadget_bind_driver() has
> > > already called usb_udc_connect_control_locked(). This causes pullup() to
> > > be called twice. Avoid this by checking if gadget->connected is true.
> > 
> > This patch is wrong.  To see why, read the comments just below the end 
> > of the patch and see also usb_gadget_activate().
> 
> OK. So it breaks usb_gadget_activate() as usb_udc_connect_control_locked()
> return early.
> 
> I think usb_gadget_activate() needs to set gadget->connected as false before
> running usb_gadget_connect_locked() just like usb_gadget_deactivate() sets
> gadget->connected as true after running usb_gadget_disconnect_locked().
> 
> Then it will work correctly, do you agree?

Yes, that should take care of it.  Don't forget to update the patch 
description as well.

> > (Also, is there really anything wrong with calling pullup() twice in a 
> > row?)
> 
> It depends on the device controller driver. But currently only a few drivers
> call usb_udc_vbus_handler(), so I think the issue hasn't shown up.
> 
> When you look at the pullup() implementation of some UDC driver, you may see
> they do more complicated thing than just pulling the data line up. Such as
> cdnsp_gadget_pullup(), dwc3_gadget_pullup(), etc.
> 
> Therefore, I think the UDC core need to handle this well to avoid calling
> pullup() twice in a row.

Okay.

Alan Stern