Re: [PATCH 1/2] dt: bindings: add new dt entry to indentify external FEM

[Bhagavathi Perumal S <bperumal@codeaurora.org>]

On 2018-10-30 21:02, Rob Herring wrote:
> On Tue, Oct 30, 2018 at 3:41 AM <bperumal@codeaurora.org> wrote:
>> 
>> On 2018-10-16 00:54, Rob Herring wrote:
>> > On Thu, Oct 04, 2018 at 05:12:43PM +0530, Bhagavathi Perumal S wrote:
>> >> This adds new dt entry ext-fem-name, it is used by ath10k driver
>> >> to select correct timing parameters and configure it in target wifi
>> >> hardware.
>> >> The Front End Module(FEM) normally includes tx power amplifier(PA) and
>> >> rx low noise amplifier(LNA). The default timing parameters like tx end
>> >> to
>> >> PA off timing values were fine tuned for internal FEM used in
>> >> reference
>> >> design. And these timing values can not be same if ODM modifies
>> >> hardware
>> >> design with different external FEM. This DT entry helps to choose
>> >> correct
>> >> timing values in driver if different external FEM hardware used.
>> >
>> > 'dt-bindings: net: ath10k: ...' for the subject please.
>> Sure, I will change and send v2 patch.
>> 
>> >
>> >>
>> >> Signed-off-by: Bhagavathi Perumal S <bperumal@codeaurora.org>
>> >> ---
>> >>  .../bindings/net/wireless/qcom,ath10k.txt          | 22
>> >> ++++++++++++++++++++++
>> >>  1 file changed, 22 insertions(+)
>> >>
>> >> diff --git
>> >> a/Documentation/devicetree/bindings/net/wireless/qcom,ath10k.txt
>> >> b/Documentation/devicetree/bindings/net/wireless/qcom,ath10k.txt
>> >> index 7fd4e8c..fbaf309 100644
>> >> --- a/Documentation/devicetree/bindings/net/wireless/qcom,ath10k.txt
>> >> +++ b/Documentation/devicetree/bindings/net/wireless/qcom,ath10k.txt
>> >> @@ -56,6 +56,7 @@ Optional properties:
>> >>                                   the length can vary between hw versions.
>> >>  - <supply-name>-supply: handle to the regulator device tree node
>> >>                         optional "supply-name" is "vdd-0.8-cx-mx".
>> >> +- ext-fem-name: name of external front end module used.
>> >
>> > What are valid names? What if the OS doesn't recognize the name?
>> > Perhaps
>> Some valid FEM devices are "microsemi-lx5586", "sky85703-11" and
>> "sky85803" etc.
>> Currently driver accepts only "microsemi-lx5586", Since this FEM
>> requires different timing settings,
> 
> Assuming you keep this, then you need to enumerate what are valid
> values in the binding. Otherwise, 'ext-fem-name = "rob"' is valid.

Sure, Will enumerate valid values in v2 patch.
> 
>> And it can be extended for other FEM devices in future.
>> If OS doesn't recognize the name, it uses predefined default timing
>> settings.
>> 
>> > this should be a compatible string for the particular module instead?
>> > Then it could cover any differences, not just the FEM.
>> >
>> No, it is specific to FEM device.
> 
> You will be better off having either a specific compatible string or
> using VID/PID as this is a PCI device. Then you can handle other
> differences you may find between h/w versions without a DT change.
> 

FEM is a part of wifi device, and not a separate PCI device. Different 
external
FEM can be used in a same wifi device, so it might not be feasible to 
define
a compatible string to cover any other differences. IMO we can go with
the property name instead of defining compatible string specific to
every FEM device.

>> >>
>> >>  Example (to supply the calibration data alone):
>> >>
>> >> @@ -150,3 +151,24 @@ wifi@18000000 {
>> >>                         <0 141 0 /* CE11 */ >;
>> >>              vdd-0.8-cx-mx-supply = <&pm8998_l5>;
>> >>  };
>> >> +
>> >> +Example (to supply the external front end module name):
>> >> +
>> >> +In this example, the front end module is defined as a property of the
>> >> ath10k
>> >> +device node.
>> >
>> > Really need a whole new example for 1 property?
>> Will add this property in existing example.
>> 
>> >
>> >> +
>> >> +pci {
>> >> +    pcie@0 {
>> >> +            reg = <0 0 0 0 0>;
>> >> +            #interrupt-cells = <1>;
>> >> +            #size-cells = <2>;
>> >> +            #address-cells = <3>;
>> >> +            device_type = "pci";
>> >> +
>> >> +            ath10k@0,0 {
>> >
>> > wifi@0,0
>> Added in ath10k@0,0 to make consistent with existing property
>> "qcom,ath10k-calibration-data" below,
>> "
> 
> I don't see how that matters.
> 
>> pci {
>>          pcie@0 {
>>                  reg = <0 0 0 0 0>;
>>                  #interrupt-cells = <1>;
>>                  #size-cells = <2>;
>>                  #address-cells = <3>;
>>                  device_type = "pci";
>> 
>>                  ath10k@0,0 {
>>                          reg = <0 0 0 0 0>;
>>                          device_type = "pci";
>>                          qcom,ath10k-calibration-data = [ 01 02 03 ... 
>> ];
>>                  };
>>          };
>> };
>> "
>> 
>> If wifi@0,0 is preferable, then I will add new entry "wifi@0,0" and 
>> add
>> ext-fem-name into it.
> 
> Node names are supposed to reflect the class of device, not the model.
> See the DT spec for a list.

Will change accordingly.

> 
>> 
>> >
>> >> +                    reg = <0 0 0 0 0>;
>> >> +                    device_type = "pci";
> 
> Also, this is wrong. Only PCI bridges should have this property. dtc
> will give you a warning on this.

Will remove it in v2 patch.

> 
>> >> +                    ext-fem-name = "microsemi-lx5586";
>> >> +            };
>> >> +    };
>> >> +};
>> >> --
>> >> 1.9.1
>> >>
>> 
>> Thanks, Sorry for the delayed response,
>> Bhagavathi Perumal S.