On Thu, 14 Sep 2017, Lee Jones wrote: > On Thu, 14 Sep 2017, Vadim Pasternak wrote: > > > The mlxreg a multifunction device driver handling LEDs, events, exposing > > through sysfs reset signal, and reset causes info. These components share > > a common register space. > > > > Signed-off-by: Vadim Pasternak <vadimp@xxxxxxxxxxxx> > > --- > > v4->v5: > > Comments pointed out by Rob: > > - Avoid duplications in reg; > > - Remove unnecessary details in description; > > - Do not use names like phandels; > > Changes added by Vadim: > > - Combine hotplug nodes interrupt aggregation, register offset and mask > > properties into hotplug-spec properties; > > - Combine reset, cause, etc subnodes register offset, mask and effective > > bit properties into attr-spec properties; > > v3->v4: > > Comments pointed out by Rob: > > - Make a separate patch /devicetree/bindings/vendor-prefixes.txt; > > - Add .txt to Documentation/devicetree/bindings/mfd/mellanox,mlxreg-core > > and send it within this series; > > - Modify "compatible" property; > > - Modify explanation for "deferred" property; > > - Describe each subnode by its own section; > > - Don't use underscore in attribute names; > > --- > > .../bindings/mfd/mellanox,mlxreg-core.txt | 165 +++++++++++++++++++++ > > 1 file changed, 165 insertions(+) > > create mode 100644 Documentation/devicetree/bindings/mfd/mellanox,mlxreg-core.txt > > Wow! I have never seen a binding like this before. > > > diff --git a/Documentation/devicetree/bindings/mfd/mellanox,mlxreg-core.txt b/Documentation/devicetree/bindings/mfd/mellanox,mlxreg-core.txt > > new file mode 100644 > > index 0000000..1e3cce5 > > --- /dev/null > > +++ b/Documentation/devicetree/bindings/mfd/mellanox,mlxreg-core.txt > > @@ -0,0 +1,165 @@ > > +Mellanox programmable device control. > > +------------------------------------- > > +This binding defines the device control interface over for Mellanox BMC based > > +switches. > > + > > +Required properties: > > +- compatible = "mellanox,mlxreg-i2c" or > > + "mellanox,mlxreg-i2c-16" > > + > > +- #address-cells : must be 1; > > +- #size-cells : must be 0; > > +- reg : I2C address; > > + > > +Optional properties: > > +- interrupt-parent : phandle of parent interrupt controller; > > +- interrupts : interrupt line; > > +- deferred : I2C deferred bus phandle; > > + I2C bus activation order enforce for the cases when hot-plug > > + devices are attached to I2C bus, which is initialized after the > > + I2C bus, where programmable device is attached; > > Dependencies aren't usually expressed in DT. > > Why can't this be done pragmatically? > > Usually we try to interrogate a device that we depend on and if it's > not present, we return -EPROBE_DEFER and try again latter. Actually, thinking about this. Are these hierarchical? If so, they would be better represented as a sub-node. -- Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
