Re: [PATCH V1 2/3] arm64: dts: qcom: sm8650: Enable MCQ support for UFS controller

[Krzysztof Kozlowski <krzk@kernel.org>]

On 31/07/2025 10:34, Ram Kumar Dwivedi wrote:
> 
> 
> On 31-Jul-25 12:15 PM, Krzysztof Kozlowski wrote:
>> On 30/07/2025 10:22, Ram Kumar Dwivedi wrote:
>>> Enable Multi-Circular Queue (MCQ) support for the UFS host controller
>>> on the Qualcomm SM8650 platform by updating the device tree node. This
>>> includes adding new register regions and specifying the MSI parent
>>> required for MCQ operation.
>>>
>>> MCQ is a modern queuing model for UFS that improves performance and
>>> scalability by allowing multiple hardware queues. 
>>>
>>> Changes:
>>> - Add reg entries for mcq_sqd and mcq_vs regions.
>>> - Define reg-names for the new regions.
>>> - Specify msi-parent for interrupt routing.
>>>
>>> Signed-off-by: Ram Kumar Dwivedi <quic_rdwivedi@quicinc.com>
>>> ---
>>>  arch/arm64/boot/dts/qcom/sm8650.dtsi | 9 ++++++++-
>>>  1 file changed, 8 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/arch/arm64/boot/dts/qcom/sm8650.dtsi b/arch/arm64/boot/dts/qcom/sm8650.dtsi
>>> index e14d3d778b71..5d164fe511ba 100644
>>> --- a/arch/arm64/boot/dts/qcom/sm8650.dtsi
>>> +++ b/arch/arm64/boot/dts/qcom/sm8650.dtsi
>>> @@ -3982,7 +3982,12 @@ ufs_mem_phy: phy@1d80000 {
>>>  
>>>  		ufs_mem_hc: ufshc@1d84000 {
>>>  			compatible = "qcom,sm8650-ufshc", "qcom,ufshc", "jedec,ufs-2.0";
>>> -			reg = <0 0x01d84000 0 0x3000>;
>>> +			reg = <0 0x01d84000 0 0x3000>,
>>> +			      <0 0x01da5000 0 0x2000>,
>>> +			      <0 0x01da4000 0 0x0010>;
>>
>>
>> These are wrong address spaces. Open your datasheet and look there.
>>
> Hi Krzysztof,
> 
> I’ve reviewed it again, and it is correct and functioning as expected both on our upstream and downstream codebase.
> I think it is probably overlooked by you. Can you please double check from your end?
> 

No, it is not overlooked. There is no address space of length 0x10 at
0x01da4000 in qcom doc/datasheet system. Just open the doc and look
there by yourself. The size is 0x15000.

You are talking now about driver and this proves my point here:
https://lore.kernel.org/linux-devicetree/1547e339-5be2-4d87-ab35-98a9be0d250e@kernel.org/



Best regards,
Krzysztof