Re: [PATH v3 1/2] dt: bindings: add new DT entry for ath11k PCI device support

[Anilkumar Kolli <akolli@codeaurora.org>]

On 2021-11-19 04:41, Rob Herring wrote:
> On Thu, Nov 18, 2021 at 07:44:51PM +0530, Anilkumar Kolli wrote:
>> Ath11k driver supports PCI devices such as QCN9074/QCA6390.
>> Ath11k firmware uses host DDR memory, DT entry is used to reserve
>> these host DDR memory regions, send these memory base
>> addresses using DT entries.
>> 
>> Signed-off-by: Anilkumar Kolli <akolli@codeaurora.org>
>> ---
>> V2:
>>   - Use reserved-memory (Rob)
>> 
>>  .../bindings/net/wireless/qcom,ath11k.yaml         | 48 
>> ++++++++++++++++++++++
>>  1 file changed, 48 insertions(+)
>> 
>> diff --git 
>> a/Documentation/devicetree/bindings/net/wireless/qcom,ath11k.yaml 
>> b/Documentation/devicetree/bindings/net/wireless/qcom,ath11k.yaml
>> index 85c2f699d602..5a8994f6cb10 100644
>> --- a/Documentation/devicetree/bindings/net/wireless/qcom,ath11k.yaml
>> +++ b/Documentation/devicetree/bindings/net/wireless/qcom,ath11k.yaml
>> @@ -150,6 +150,12 @@ properties:
>>        string to uniquely identify variant of the calibration data in 
>> the
>>        board-2.bin for designs with colliding bus and device specific 
>> ids
>> 
>> +  memory-region:
>> +    maxItems: 1
>> +    description:
>> +      phandle to a node describing reserved memory (System RAM 
>> memory)
>> +      used by ath11k firmware (see 
>> bindings/reserved-memory/reserved-memory.txt)
>> +
>>  required:
>>    - compatible
>>    - reg
>> @@ -279,3 +285,45 @@ examples:
>>                            "tcl2host-status-ring";
>>          qcom,rproc = <&q6v5_wcss>;
>>      };
>> +
> 
> This looks like a separate example. Please split to its own entry.
> 
>> +    memory {
>> +        device_type = "memory";
>> +        reg = <0x0 0x40000000 0x0 0x20000000>;
>> +    };
> 
> Outside the scope of what's needed in the example.
> 
Yes, memory entry is available in 
"arch/arm64/boot/dts/qcom/ipq8074-hk10.dtsi"
Since its used in ath11k patch, added example.
I will remove in next version.

>> +
>> +    reserved-memory {
>> +        #address-cells = <1>;
>> +        #size-cells = <1>;
>> +        ranges;
>> +
>> +        qcn9074_pcie0: qcn9074_pcie0@51100000 {
>> +            no-map;
>> +            reg = <0x0 0x51100000 0x0 0x03500000>;
>> +        };
>> +
>> +        qcn9074_pcie1: qcn9074_pcie1@54600000 {
>> +            no-map;
>> +            reg = <0x0 0x54600000 0x0 0x03500000>;
>> +        };
>> +    };
> 
> As is this really.
> 
ipq8074-hk10.dtsi board supports two PCI bus and QCN9074 on each PCI.
So added two separate entries to reserves memory for each QCN9074.

>> +
>> +    pcie0_rp: pcie0_rp {
>> +        reg = <0 0 0 0 0>;
> 
> This isn't a valid PCI bus binding.
> 
Got it, let me rework in next patch.

>> +
>> +        status = "ok";
> 
> Don't need status.

Sure, will remove in next patch.

> 
>> +        ath11k_0: ath11k@0 {
> 
> wifi@0
> 
>> +            reg = <0 0 0 0 0 >;
>> +            memory-region = <&qcn9074_pcie0>;
>> +        };
>> +    };
>> +
>> +    pcie1_rp: pcie1_rp {
>> +        reg = <0 0 0 0 0>;
>> +
>> +        status = "ok";
>> +        ath11k_1: ath11k@1 {
>> +            reg = <0 0 0 0 0 >;
> 
> unit-address and reg don't match.
> 
will update in next patch.

> Why do we need 2 nodes in the first place?
> 
ipq8074-hk10.dtsi board supports two PCI bus and QCN9074 on each PCI.
So added two separate entries to reserves memory for each QCN9074.

Since its example, Shall I remove one ?

>> +            memory-region = <&qcn9074_pcie1>;
>> +        };
>> +    };
>> +
>> --
>> 2.7.4
>> 

- Anil.