[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20ec31b3-667a-0617-71d4-586789e2c2a4@quicinc.com>
Date: Mon, 26 Jun 2023 16:47:18 +0530
From: Mohammad Rafi Shaik <quic_mohs@...cinc.com>
To: Konrad Dybcio <konrad.dybcio@...aro.org>,
<krzysztof.kozlowski+dt@...aro.org>, <swboyd@...omium.org>,
<andersson@...nel.org>, <broonie@...nel.org>, <agross@...nel.org>
CC: <robh+dt@...nel.org>, <linux-arm-msm@...r.kernel.org>,
<devicetree@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
<quic_rohkumar@...cinc.com>, <srinivas.kandagatla@...aro.org>,
<dianders@...omium.org>, <judyhsiao@...omium.org>,
<quic_visr@...cinc.com>,
Srinivasa Rao Mandadapu <quic_srivasam@...cinc.com>
Subject: Re: [RESEND v6 8/8] arm64: dts: qcom: sc7280: Add qcom,adsp-pil-mode
property in clock nodes
On 6/16/2023 5:06 PM, Konrad Dybcio wrote:
> On 16.06.2023 12:35, Mohammad Rafi Shaik wrote:
>> From: Srinivasa Rao Mandadapu <quic_srivasam@...cinc.com>
>>
>> Add "qcom,adsp-pil-mode" property in clock nodes for herobrine
>> crd revision 3 board specific device tree.
>> This is to register clocks conditionally by differentiating ADSP
>> based platforms and legacy path platforms.
>> Also disable lpass_core clock, as it is creating conflict
>> with ADSP clocks and it is not required for ADSP based platforms.
>>
>> Signed-off-by: Srinivasa Rao Mandadapu <quic_srivasam@...cinc.com>
>> Signed-off-by: Mohammad Rafi Shaik <quic_mohs@...cinc.com>
>> ---
>> .../qcom/sc7280-herobrine-audioreach-wcd9385.dtsi | 12 ++++++++++++
>> 1 file changed, 12 insertions(+)
>>
>> diff --git a/arch/arm64/boot/dts/qcom/sc7280-herobrine-audioreach-wcd9385.dtsi b/arch/arm64/boot/dts/qcom/sc7280-herobrine-audioreach-wcd9385.dtsi
>> index c02ca393378f..876a29178d46 100644
>> --- a/arch/arm64/boot/dts/qcom/sc7280-herobrine-audioreach-wcd9385.dtsi
>> +++ b/arch/arm64/boot/dts/qcom/sc7280-herobrine-audioreach-wcd9385.dtsi
>> @@ -197,6 +197,14 @@ q6prmcc: clock-controller {
>> };
>> };
>>
>> +&lpass_aon {
>> + qcom,adsp-pil-mode;
> That's a whole bunch of hacky bindings that makes no sense..
>
> What should have been done from the beginning is:
>
> - all clocks should be registered inside the clock driver, unconditionally
> as far as .c code is concerned
>
> - the regmap properties within should reflect the actual max register
> range within the hardware block
>
> - device-tree should contain protected-clocks, which omits registering
> specified clks (I guess in the ADSP-less case you could probably even
> register all of them and it wouldn't hurt)
>
For AR solution, it is required to add "qcom,adsp-pil-mode" flag to
enable ahbm and ahbs clocks.
Please refer:
https://elixir.bootlin.com/linux/v6.4-rc7/source/drivers/clk/qcom/lpassaudiocc-sc7280.c
>> +};
>> +
>> +&lpass_core {
>> + status = "disabled";
> status = "reserved";
>
> Konrad
Okay, will change status flag.
Rafi
>> +};
>> +
>> &lpass_rx_macro {
>> /delete-property/ power-domains;
>> /delete-property/ power-domain-names;
>> @@ -239,3 +247,7 @@ &lpass_va_macro {
>>
>> status = "okay";
>> };
>> +
>> +&lpasscc {
>> + qcom,adsp-pil-mode;
>> +};
Powered by blists - more mailing lists