On Tue, Nov 20, 2012 at 01:54:48PM +0800, Peter Chen wrote:
> On Mon, Nov 19, 2012 at 08:17:34PM +0100, Sascha Hauer wrote:
> > Hi Peter,
> >
> > > @@ -139,6 +139,7 @@ struct ci13xxx {
> > > enum ci_role role;
> > > bool is_otg;
> > > struct work_struct work;
> > > + struct delayed_work dwork;
> > > struct workqueue_struct *wq;
> > >
> > > struct dma_pool *qh_pool;
> > > @@ -164,6 +165,11 @@ struct ci13xxx {
> > > bool global_phy;
> > > struct usb_phy *transceiver;
> > > struct usb_hcd *hcd;
> > > + /* events handled at ci_role_work */
> > > +#define ID 0
> > > +#define B_SESS_VALID 1
> > > + unsigned long events;
> > > + struct usb_otg otg;
> >
> > I looked into implementing ULPI support for the chipidea driver. This
> > does not integrate very well with having a struct usb_otg here. Instead
> > it should be a pointer. Look into drivers/usb/otg/ulpi.c. This
> > allocates the struct usb_otg itself and we have to feed the pointer into
> > struct ci13xxx.
>
> We discussed before that the otg is not related to phy.
Maybe the real problem is that the usb_otg instead of the phy provides
the set_vbus callback. It's a bit strange that the phy has the
set_power callback to draw current and the otg has the set_vbus callback
to provide vbus.
Moving set_vbus to the phy would solve things, then the ulpi driver
would have no business creating a struct usb_otg and this could be
private to the chipidea driver. We would also have the possibility to
provide a set_vbus function without (needed for ulpi even for host-only
mode) without having a struct usb_otg.
Sascha
--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
--
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
