Re: [PATCH 3/3] dt-bindings: clock: amlogic,gxbb-aoclkc: Update bindings

[Jerome Brunet <jbrunet@baylibre.com>]

On Mon, 2017-07-24 at 13:47 +0200, Neil Armstrong wrote:
> On 07/21/2017 10:44 PM, Stephen Boyd wrote:
> > On 07/09, Rob Herring wrote:
> > > On Thu, Jul 06, 2017 at 12:24:23PM +0200, Neil Armstrong wrote:
> > > > On the first revision of the bindings, only the gates + resets were
> > > > known
> > > > in the AO Clock HW, but more registers used to configures AO clock are
> > > > known
> > > > to be spread among the AO register space.
> > > > This patch adds these registers to the Ao Clock bindings with direct
> > > > access
> > > > and shared extcon access.
> > > > 
> > > > Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
> > > > ---
> > > >  .../devicetree/bindings/clock/amlogic,gxbb-aoclkc.txt         | 11
> > > > +++++++++--
> > > >  1 file changed, 9 insertions(+), 2 deletions(-)
> > > 
> > > This looks like the binding might be too specific with a reg list of 
> > > single registers, and you should define a system controller node 
> > > instead. Depends on what else is in the "A0" block.
> > > 
> > 
> > Agreed. Why can't we expand the size in DT and then access the
> > registers directly in the driver. Hopefully it keeps working to
> > apply the dts patch without the driver patch too (and
> > vice-versa), because the kernel can only make a mapping as small
> > as a page which would cover these newly added reg properties
> > anyway.
> > 
> 
> Hi Rob, Stephen,
> 
> Thanks for your feedback, but on these Amlogic platforms, the AO register bank
> is
> filled with interleaved registers used for various purposes.
> 
> For instance, the first 1k of registers are :
> 
> 0-c RTI_STATUS
> c-14 RTI_PWR_CNTL
> 14-1c PIN_MUX
> 1c RTI_STATUS
> 24-30 GPIO
> 30 JTAG_CONFIG
> 34 WD
> 38-40 CPU_CTRL
> 40 RTI_GEN_CTRL
> 44 CPU_CTRL
> 4c-58 TIMER
> 58 OSCIN
> 60 AHB2DDR
> ...
> 
> 
> And so on, and the clock related registers are split among this space.
> 
> For sure, this could be declared as an system controller node, but this would
> imply completely
> re-designing the actual clock driver and drop the actual bindings.
> (and with re-writting the clock gate code to use regmap registers access)

Maybe it is time to investigate having the regmap clock from qcom available to
every other platform ?

> 
> Anyway, I'm open to suggestions...
> 
> Neil