linux-kernel - Re: [PATCH v5 2/6] dt-bindings: PCI: qcom,pcie-sa8775p: document qcs8300

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <574a67b7-bd57-4cf4-9ecb-cdcefafeb791@kernel.org>
Date: Wed, 7 May 2025 12:03:14 +0200
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Qiang Yu <quic_qianyu@...cinc.com>,
 Ziyue Zhang <quic_ziyuzhan@...cinc.com>
Cc: vkoul@...nel.org, kishon@...nel.org, robh@...nel.org, krzk+dt@...nel.org,
 conor+dt@...nel.org, dmitry.baryshkov@...aro.org, neil.armstrong@...aro.org,
 abel.vesa@...aro.org, manivannan.sadhasivam@...aro.org,
 lpieralisi@...nel.org, kw@...ux.com, bhelgaas@...gle.com,
 andersson@...nel.org, konradybcio@...nel.org, linux-phy@...ts.infradead.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
 linux-arm-msm@...r.kernel.org, linux-pci@...r.kernel.org,
 quic_krichai@...cinc.com, quic_vbadigan@...cinc.com
Subject: Re: [PATCH v5 2/6] dt-bindings: PCI: qcom,pcie-sa8775p: document
 qcs8300

On 07/05/2025 11:56, Qiang Yu wrote:
> 
> On 5/7/2025 4:25 PM, Krzysztof Kozlowski wrote:
>> On 07/05/2025 10:19, Ziyue Zhang wrote:
>>> On 5/7/2025 1:10 PM, Krzysztof Kozlowski wrote:
>>>> On Wed, May 07, 2025 at 11:10:15AM GMT, Ziyue Zhang wrote:
>>>>> Add compatible for qcs8300 platform, with sa8775p as the fallback.
>>>>>
>>>>> Signed-off-by: Ziyue Zhang <quic_ziyuzhan@...cinc.com>
>>>>> ---
>>>>>    .../bindings/pci/qcom,pcie-sa8775p.yaml       | 26 ++++++++++++++-----
>>>>>    1 file changed, 19 insertions(+), 7 deletions(-)
>>>>>
>>>>> diff --git a/Documentation/devicetree/bindings/pci/qcom,pcie-sa8775p.yaml b/Documentation/devicetree/bindings/pci/qcom,pcie-sa8775p.yaml
>>>>> index efde49d1bef8..154bb60be402 100644
>>>>> --- a/Documentation/devicetree/bindings/pci/qcom,pcie-sa8775p.yaml
>>>>> +++ b/Documentation/devicetree/bindings/pci/qcom,pcie-sa8775p.yaml
>>>>> @@ -16,7 +16,12 @@ description:
>>>>>    
>>>>>    properties:
>>>>>      compatible:
>>>>> -    const: qcom,pcie-sa8775p
>>>>> +    oneOf:
>>>>> +      - const: qcom,pcie-sa8775p
>>>>> +      - items:
>>>>> +          - enum:
>>>>> +              - qcom,pcie-qcs8300
>>>>> +          - const: qcom,pcie-sa8775p
>>>>>    
>>>>>      reg:
>>>>>        minItems: 6
>>>>> @@ -45,7 +50,7 @@ properties:
>>>>>    
>>>>>      interrupts:
>>>>>        minItems: 8
>>>>> -    maxItems: 8
>>>>> +    maxItems: 9
>>>> I don't understand why this is flexible for sa8775p. I assume this
>>>> wasn't tested or finished, just like your previous patch suggested.
>>>>
>>>> Please send complete bindings once you finish them or explain what
>>>> exactly changed in the meantime.
>>>>
>>>> Best regards,
>>>> Krzysztof
>>> Hi Krzysztof
>>> Global interrupt is optional in the PCIe driver. It is not present in
>>> the SA8775p PCIe device tree node, but it is required for the QCS8300
>> And hardware?
> 
> The PCIe controller on the SA8775p is also capable of generating a global
> interrupt.
>>> I did the DTBs and yaml checks before pushing this patch. This is how
>>> I became aware that `maxItem` needed to be changed to 9.
>> If it is required for QCS8300, then you are supposed to make it required
>> in the binding for this device. Look at other bindings.
> 
> The global interrupt is not mandatory. The PCIe driver can still function
> without this interrupt, but it will offer a better user experience when
> the device is plugged in or removed. On other platforms, the global
> interrupt is also optional, and `minItems` and `maxItems` are set to 8 and
> 9 respectively. Please refer to `qcom,pcie - sm8550.yaml`,
> `qcom,pcie - sm8450.yaml`, and `qcom,pcie - x1e80100.yaml`.
I don't know what does it prove. You cannot add requirement of global
interrupt to existing devices because it would be an ABI break.

Best regards,
Krzysztof