On Tue, Dec 26, 2017 at 03:30:17PM +0100, Mark Kettenis wrote:
> > Date: Tue, 26 Dec 2017 17:16:47 +0800
> > From: Kevin Lo <ke...@openbsd.org>
> > 
> > On Tue, Dec 26, 2017 at 09:52:36AM +0100, Mark Kettenis wrote:
> > > > Date: Tue, 26 Dec 2017 15:59:12 +0800
> > > > From: Kevin Lo <ke...@openbsd.org>
> > > > 
> > > > On Tue, Dec 26, 2017 at 05:12:39PM +1100, Jonathan Gray wrote:
> > > > > 
> > > > > On Tue, Dec 26, 2017 at 01:15:35PM +0800, Kevin Lo wrote:
> > > > > > The diff below adds support for the "next-generation" clock and 
> > > > > > pinctrl
> > > > > > bindings for the Allwinner A33.
> > > > > > 
> > > > > > Tested on my Banana Pi M2 Magic.  dmesg: http://ix.io/DoB
> > > > > > ok?
> > > > > > 
> > > > > > Index: sys/dev/fdt/ehci_fdt.c
> > > > > > ===================================================================
> > > > > > RCS file: /cvs/src/sys/dev/fdt/ehci_fdt.c,v
> > > > > > retrieving revision 1.2
> > > > > > diff -u -p -u -p -r1.2 ehci_fdt.c
> > > > > > --- sys/dev/fdt/ehci_fdt.c  17 Dec 2017 13:23:03 -0000      1.2
> > > > > > +++ sys/dev/fdt/ehci_fdt.c  25 Dec 2017 09:43:48 -0000
> > > > > > @@ -273,9 +273,10 @@ sun4i_phy_init(struct ehci_fdt_softc *sc
> > > > > >  
> > > > > >     /*
> > > > > >      * We need to poke an undocumented register to make the PHY
> > > > > > -    * work on Allwinner H3/H5/A64.
> > > > > > +    * work on Allwinner H3/H5/A33/A64.
> > > > > >      */
> > > > > > -   if (OF_is_compatible(node, "allwinner,sun8i-h3-usb-phy") ||
> > > > > > +   if (OF_is_compatible(node, "allwinner,sun8i-a33-usb-phy") ||
> > > > > > +       OF_is_compatible(node, "allwinner,sun8i-h3-usb-phy") ||
> > > > > >         OF_is_compatible(node, "allwinner,sun50i-a64-usb-phy")) {
> > > > > >             val = bus_space_read_4(sc->sc.iot, sc->sc.ioh, 0x810);
> > > > > >             val &= ~(1 << 1);
> > > > > 
> > > > > Is this actually required?  U-Boot only does it for h3/h5.
> > > > > 
> > > > > If so should "allwinner,sun8i-a23-usb-phy" and
> > > > > "allwinner,sun8i-v3s-usb-phy" also be added?
> > > > 
> > > > Thanks for catching this, it's not required.  Here is an updated diff:
> > > 
> > > Hi Kevin,
> > 
> > Hi Mark,
> > 
> > > The clock bits look good to me; ok kettenis@ for this.
> > > 
> > > The pin stuff is actually automatically generated by some code I
> > > maintain outside of the tree.  I'll add the A33 bits to my code and
> > > commit that for you if you don't mind.
> > 
> > I don't mind, thank you :)
> 
> Done.
> 
> Looks like your board has an AXP223 PMIC on it.

Right.

> Shouldn't be too difficult to add support for that one to axppmic(4).
> Do you want me to create a diff for you to test?

That'd be great, thanks.  Testing is easier than writing code :)

> Cheers,
> 
> Mark

Regards,
Kevin

Reply via email to