Hi Felipe, On Tue, Nov 08, 2011 at 04:20:31PM +0200, Felipe Balbi wrote: > > @@ -169,8 +169,8 @@ static int ab8500_usb_link_status_update(struct ab8500_usb *ab) > > /* TODO: Disable regulators. */ > > ab8500_usb_host_phy_dis(ab); > > ab8500_usb_peri_phy_dis(ab); > > - ab->otg.state = USB_PHY_STATE_B_IDLE; > > - ab->otg.default_a = false; > > + ab->phy.state = USB_PHY_STATE_B_IDLE; > > + ab->phy.otg->default_a = false; > > ideally users wouldn't access 'otg' directly. You should have a bunch of > reader/setter helpers which are optimized out when OTG isn't enabled. > Then we could: > > usb_otg_default_a(&ab->phy, false); > > or something similar. Ditto for all others. Can, of course, be done on a > later patch. Yes. I left it out from this set as I wasn't sure if we want later to be able to live without PHY driver, for example if you support OTG but you have UTMI+ transceiver. Or do we always want to have PHY driver? The "OTG driver" usually only needs to program the controller. > > @@ -488,19 +487,27 @@ static int __devinit ab8500_usb_probe(struct platform_device *pdev) > > if (!ab) > > return -ENOMEM; > > > > + ab->phy.otg = kzalloc(sizeof *ab->phy.otg, GFP_KERNEL); > > I don't see the point of requiring users to allocate the otg structure. > What's the rationale ? Well, I guess I was thinking about the previous idea. Eventually the PHY drivers should not be responsible of the struct otg unless they really need to be, like with CEA-2011, though that spec was cancelled if I recall. They can still hold the pointer, so they can "carry" it if needed. I'll cleanup this so the users don't allocate the structure. Thanks, -- heikki -- 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
