[PATCH] gpio: dt-bindings: add ibm,ppc4xx-gpio binding
robh at kernel.org
Tue May 17 02:56:18 AEST 2016
On Mon, May 16, 2016 at 10:27:46AM -0500, Rob Herring wrote:
> On Thu, May 12, 2016 at 12:07:48AM +0200, Christian Lamparter wrote:
> > This patch adds binding information for IBM/AMCC/APM GPIO
> > Controllers of the PowerPC 4XX series and compatible SoCs.
> > The "PowerPC 405EP Embedded Processor Data Sheet" has the
> > following to say about the GPIO controllers: "
> > - Controller functions and GPIO registers are programmed
> > and accessed via memory-mapped OPB bus master accesses
> > - All GPIOs are pin-shared with other functions. DCRs control
> > whether a particular pin that has GPIO capabilities acts
> > as a GPIO or is used for another purpose.
> > - Each GPIO outputs is separately programmable to emulate
> > an open-drain driver (i.e. drives to zero, threestated if
> > output bit is 1)
> > "
> > The ppc4xx_gpio.c driver is part of the platform/sysdev drivers
> > in arch/powerpc/sysdev.
> > Signed-off-by: Christian Lamparter <chunkeey at googlemail.com>
> > ---
> > I looked into arch/powerpc/sysdev/ppc4xx_gpio.c driver and
> > it doesn't have support for the tri-state logic (open drain
> > is disabled), but the hardware would support it.
> > (the #gpio-cells description suffers because of this, since
> > the high-z option isn't there).
> > Also there's another problem: There's no DCR pinmux driver?!
> > So sadly, there's not much information on how to use the DCRs
> > to control the which pin is muxed to the GPIO or to a SoC
> > function like the i2c.
> > This was all the valuable information I could find about the
> > hardware, so it is included it in the binding text, even
> > though there's no support for it... Is there anything else
> > to add?
> > ---
> > .../devicetree/bindings/gpio/ibm,ppc4xx-gpio.txt | 24 ++++++++++++++++++++++
> > 1 file changed, 24 insertions(+)
> > create mode 100644 Documentation/devicetree/bindings/gpio/ibm,ppc4xx-gpio.txt
> Normally, we don't allow wildcards for compatible strings, but given
> this is an established binding it is okay.
> Acked-by: Rob Herring <robh at kernel.org>
As this is just a binding doc, I've applied it.
More information about the Linuxppc-dev