[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <253e7889-bef4-453c-9f39-19ba0a6dd5b0@oss.qualcomm.com>
Date: Mon, 23 Dec 2024 11:52:42 +0100
From: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
To: "Aiqun(Maria) Yu" <quic_aiquny@...cinc.com>,
Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>,
Bjorn Andersson <andersson@...nel.org>,
Konrad Dybcio <konradybcio@...nel.org>, Rob Herring <robh@...nel.org>,
Krzysztof Kozlowski <krzk+dt@...nel.org>,
Conor Dooley
<conor+dt@...nel.org>, Vinod Koul <vkoul@...nel.org>,
Dmitry Baryshkov <dmitry.baryshkov@...aro.org>,
Neil Armstrong <neil.armstrong@...aro.org>,
Abel Vesa
<abel.vesa@...aro.org>,
Sibi Sankar <quic_sibis@...cinc.com>,
Luca Weiss <luca.weiss@...rphone.com>
Cc: linux-arm-msm@...r.kernel.org, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, stable@...r.kernel.org,
Konrad Dybcio <konrad.dybcio@....qualcomm.com>
Subject: Re: [PATCH v3 00/23] arm64: dts: qcom: Fix remoteproc memory base and
length
On 23.12.2024 10:21 AM, Aiqun(Maria) Yu wrote:
> On 12/13/2024 10:53 PM, Krzysztof Kozlowski wrote:
>> Changes in v3:
>> - Add Rb tags
>> - Add four new patches (at the end) for sdx75 and sm6115
>> - Link to v2: https://lore.kernel.org/r/20241209-dts-qcom-cdsp-mpss-base-address-v2-0-d85a3bd5cced@linaro.org
>>
>> Changes in v2:
>> - arm64: dts: qcom: x1e80100: Fix ADSP...:
>> Commit msg corrections, second paragraph (Johan)
>> - Add tags
>> - Link to v1: https://lore.kernel.org/r/20241206-dts-qcom-cdsp-mpss-base-address-v1-0-2f349e4d5a63@linaro.org
>>
>> Konrad pointed out during SM8750 review, that numbers are odd, so I
>> looked at datasheets and downstream DTS for all previous platforms.
>>
>> Most numbers are odd.
>>
>> Older platforms like SM8150, SM8250, SC7280, SC8180X seem fine. I could
>> not check few like SDX75 or SM6115, due to lack of access to datasheets.
>>
>> SM8350, SM8450, SM8550 tested on hardware. Others not, but I don't
>> expect any failures because PAS drivers do not use the address space.
>> Which also explains why odd numbers did not result in any failures.
>
> In my opinion, the "QCOM_Q6V5_PAS" based Peripheral Authentication
> platforms may have the register information completely removed.
>
> There are two types of Peripheral Authentication supported:
> "QCOM_Q6V5_MSS" (self-authenticating)
> "QCOM_Q6V5_PAS" (trust-zone based Authentication)
> For "QCOM_Q6V5_PAS" based Peripheral Authentication platforms, use SCM
> calls instead of the register-based mechanism. So it is no need to
> expose the PUB reg addresses for those platforms.
(Unfortunately) not all boards using the same SoC have the same
firmware stack, and it's not obvious that self-authentication is not
useful. Plus having an accurate register space description in the
DT is "nice".
Konrad
Powered by blists - more mailing lists