Re: [PATCH 2/2] arm64: dts: qcom: add apq8096sg-db820c, AP8096SG variant of DB820c

[Konrad Dybcio]

On 1/14/26 1:12 AM, Dmitry Baryshkov wrote:
> On Mon, Nov 17, 2025 at 12:54:22PM +0100, Konrad Dybcio wrote:
>> On 11/15/25 1:09 AM, Dmitry Baryshkov wrote:
>>> On Fri, 14 Nov 2025 at 23:31, Konrad Dybcio
>>> <konrad.dybcio@xxxxxxxxxxxxxxxx> wrote:
>>>>
>>>> On 11/13/25 9:32 PM, Dmitry Baryshkov wrote:
>>>>> On Wed, Nov 12, 2025 at 10:44:49AM +0100, Konrad Dybcio wrote:
>>>>>> On 11/11/25 5:02 PM, Dmitry Baryshkov wrote:
>>>>>>> There has been a (rare) varint of Dragonboard 820c, utilizing Pro
>>>>>>> version of the SoC, with the major difference being CPU and GPU clock
>>>>>>> tables. Add a DT file representing this version of the board.
>>>>>>
>>>>>> So is the conclusion that both flavors were used?
>>>>>
>>>>> Yes. I have had a production unit with a non-SG flavour. Bjorn's lab
>>>>> has a standard one too. All units in Collabora lab are Pro ones.
>>>>
>>>> Pro doesn't necessarily have to == SG, this seems to be sort of
>>>
>>> My understanding was that APQ8096SG is modem-less MSM8996Pro.
>>>
>>>> a "MSM8996Pro" and "QCM8996Pro" situation.
>>>
>>>> I'm hoping that speedbin
>>>> fuse values don't have different meanings for mobilePro and SG
>>>
>>> At least downstream doesn't have separate bins for APQ versions.
>>
>> arch/arm64/boot/dts/qcom/msm8996pro.dtsi:       qcom,msm-id = <305 0x10000>;
>> arch/arm64/boot/dts/qcom/msm8996pro-v1.1.dtsi:  qcom,msm-id = <305 0x10001>;
>>
>> this is interesting, perhaps Pro==SG then
> 
> Yes. At least it matches what is written in the LK sources (or in
> MSM8996 Device Revision Guide).
> 
> The JTAG ID matches APQ8096SG, the bootloader identifies it as 8996 Pro.
> 
> Any remaining issues?

No, I think we're good

Reviewed-by: Konrad Dybcio <konrad.dybcio@xxxxxxxxxxxxxxxx>

Konrad