On Thursday, June 30, 2011 05:48:54 PM Stefano Babic wrote: > On 06/30/2011 11:57 AM, Matthias Weisser wrote: > > Adding support for USB host on imx25 using the internal PHY > > > > Signed-off-by: Matthias Weisser <weiss...@arcor.de> > > --- > > Hi Matthias, > > > +#ifdef CONFIG_MX25 > > +#define MX25_USB_CTRL_IP_PUE_DOWN_BIT (1<<6) > > +#define MX25_USB_CTRL_HSTD_BIT (1<<5) > > +#define MX25_USB_CTRL_USBTE_BIT (1<<4) > > +#define MX25_USB_CTRL_OCPOL_OTG_BIT (1<<3) > > +#endif > > + > > +#ifdef CONFIG_MX31 > > > > #define MX31_OTG_SIC_SHIFT 29 > > #define MX31_OTG_SIC_MASK (0x3 << MX31_OTG_SIC_SHIFT) > > #define MX31_OTG_PM_BIT (1 << 24) > > > > @@ -42,10 +50,19 @@ > > > > #define MX31_H1_SIC_MASK (0x3 << MX31_H1_SIC_SHIFT) > > #define MX31_H1_PM_BIT (1 << 8) > > #define MX31_H1_DT_BIT (1 << 4) > > > > +#endif > > Can we try to get rid of nasty #ifdef ? I am sure, we would have in > future support for at least MX5, and if we add a special handling for > each SOC this file will become unreadable.
... we would have ... if the developer won't hang himself first ;-) Otherwise, we should probably unify this IMX_USB_BASE stuff indeed. Maybe even compute USBCTRL_OTGBASE_OFFSET on the fly (as imx51 has different offset there). > > It seems to me that in most cases we can do it using the same names. > > > /* Take USB2 */ > > > > +#ifdef CONFIG_MX25 > > + ehci = (struct usb_ehci *)(IMX_USB_BASE + > > +#else > > > > ehci = (struct usb_ehci *)(MX31_OTG_BASE_ADDR + > > > > +#endif > > For example, here we can have a CONFIG_MXB_USB_BASE (or something like > this), without adding #ifdef. I do not know if we can get completely get > rid of them, but it is worth to try. > > Best regards, > Stefano Babic _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot