[PATCH v2 07/10] dts/ls2085a: Update DTSI to add support of various peripherals

[Sharma Bhupesh]

> From: Rob Herring [mailto:robherring2 at gmail.com]
> Sent: Tuesday, September 08, 2015 2:28 AM
> On Sun, Sep 6, 2015 at 4:03 PM, Sharma Bhupesh
> <bhupesh.sharma at freescale.com> wrote:
> >> From: Rob Herring [mailto:robherring2 at gmail.com]
> >> Sent: Monday, September 07, 2015 2:27 AM On Sat, Sep 5, 2015 at 3:37
> >> AM, Sharma Bhupesh <bhupesh.sharma at freescale.com> wrote:
> >> >> From: pku.leo at gmail.com [mailto:pku.leo at gmail.com]
> >> >> Sent: Saturday, September 05, 2015 2:33 AM On Fri, Sep 4, 2015 at
> >> >> 2:05 AM, Bhupesh Sharma <bhupesh.sharma at freescale.com> wrote:
> >> >> > This patch updates the LS2085a DTSI (DTS Include) file to add
> >> >> > support for various peripherals supported by FSL LS2085a SoC,
> >> >> > for
> >> e.g.:
> >> >>
> >> >> The title and description here are still using the old LS2085 name.
> >> >
> >> > Ok.
> >> >
> >> > [snip] ..
> >> >
> >> >> > +       amba {
> >> >> > +               compatible = "arm,amba-bus";
> >> >> > +               #address-cells = <2>;
> >> >> > +               #size-cells = <2>;
> >> >> > +               ranges;
> >> >> > +
> >> >> > +               cluster1_core0_watchdog: wdt at c000000 {
> >> >> > +                       compatible = "arm,primecell";
> >> >>
> >> >> Binding of Primecell requires a specific name for the device
> >> >> instead of just a "arm,primecell".
> >> >
> >> > The sp805 driver doesn't have a compatible string like arm, sp805.
> >> > You can refer to 'Documentation/devicetree/bindings/wdt' for
> >> > details -
> >> there is no entry for sp805 here.
> >>
> >> Because no one wrote it yet. Please read
> >> Documentation/devicetree/bindings/arm/primecell.txt. If sp805 has no
> >> binding doc, please write one.
> >>
> >> > Instead it relies on the 'drivers/amba/bus.c' framework to invoke
> >> > the probe function of the
> >> > sp805 wdt driver using the unique PRIMECELL identifiers of the
> >> > sp805
> >> wdt IP.
> >>
> >> Whether the kernel chooses to bind with the ID registers or
> >> compatible strings is up to the kernel. That is independent of the
> bindings.
> >>
> >
> > Right. But wouldn't the driver require DTS awareness to handle the
> compatible string then.
> > Shouldn't the DTS awareness added both in terms of bindings and driver
> handling?
> 
> No. The OS is free to do device enumeration however it sees fit and that
> could change independent of the DT.
 
Ok. Will address this in v3 of this patchset then.

Regards,
Bhupesh