2024-01-31 10:28:12

by Chukun Pan

[permalink] [raw]
Subject: [PATCH 1/3] arm64: dts: qcom: ipq6018: separate CPU OPP tables

Some IPQ6000 SoCs do not come with PMIC (MP5496) chip,
which causes cpufreq to be unavailable due to lack of
cpu-supply. Separate CPU OPP tables from soc.dtsi to
support versions with and without PMIC chip.

Signed-off-by: Chukun Pan <[email protected]>
---
arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts | 1 +
arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi | 74 ++++++++++++++++++++
arch/arm64/boot/dts/qcom/ipq6018.dtsi | 56 ---------------
3 files changed, 75 insertions(+), 56 deletions(-)
create mode 100644 arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi

diff --git a/arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts b/arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts
index f5f4827c0e17..06dfc2cb6b7f 100644
--- a/arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts
+++ b/arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts
@@ -8,6 +8,7 @@
/dts-v1/;

#include "ipq6018.dtsi"
+#include "ipq6018-opp.dtsi"

/ {
model = "Qualcomm Technologies, Inc. IPQ6018/AP-CP01-C1";
diff --git a/arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi b/arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi
new file mode 100644
index 000000000000..9c0bed2d8bf5
--- /dev/null
+++ b/arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi
@@ -0,0 +1,74 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
+/*
+ * IPQ60xx with PMIC (MP5496) CPU OPP tables
+ */
+
+/ {
+ cpu_opp_table: opp-table-cpu {
+ compatible = "operating-points-v2-kryo-cpu";
+ nvmem-cells = <&cpu_speed_bin>;
+ opp-shared;
+
+ opp-864000000 {
+ opp-hz = /bits/ 64 <864000000>;
+ opp-microvolt = <725000>;
+ opp-supported-hw = <0xf>;
+ clock-latency-ns = <200000>;
+ };
+
+ opp-1056000000 {
+ opp-hz = /bits/ 64 <1056000000>;
+ opp-microvolt = <787500>;
+ opp-supported-hw = <0xf>;
+ clock-latency-ns = <200000>;
+ };
+
+ opp-1320000000 {
+ opp-hz = /bits/ 64 <1320000000>;
+ opp-microvolt = <862500>;
+ opp-supported-hw = <0x3>;
+ clock-latency-ns = <200000>;
+ };
+
+ opp-1440000000 {
+ opp-hz = /bits/ 64 <1440000000>;
+ opp-microvolt = <925000>;
+ opp-supported-hw = <0x3>;
+ clock-latency-ns = <200000>;
+ };
+
+ opp-1608000000 {
+ opp-hz = /bits/ 64 <1608000000>;
+ opp-microvolt = <987500>;
+ opp-supported-hw = <0x1>;
+ clock-latency-ns = <200000>;
+ };
+
+ opp-1800000000 {
+ opp-hz = /bits/ 64 <1800000000>;
+ opp-microvolt = <1062500>;
+ opp-supported-hw = <0x1>;
+ clock-latency-ns = <200000>;
+ };
+ };
+};
+
+&CPU0 {
+ operating-points-v2 = <&cpu_opp_table>;
+ cpu-supply = <&ipq6018_s2>;
+};
+
+&CPU1 {
+ operating-points-v2 = <&cpu_opp_table>;
+ cpu-supply = <&ipq6018_s2>;
+};
+
+&CPU2 {
+ operating-points-v2 = <&cpu_opp_table>;
+ cpu-supply = <&ipq6018_s2>;
+};
+
+&CPU3 {
+ operating-points-v2 = <&cpu_opp_table>;
+ cpu-supply = <&ipq6018_s2>;
+};
diff --git a/arch/arm64/boot/dts/qcom/ipq6018.dtsi b/arch/arm64/boot/dts/qcom/ipq6018.dtsi
index 5e1277fea725..ea72fd5739ac 100644
--- a/arch/arm64/boot/dts/qcom/ipq6018.dtsi
+++ b/arch/arm64/boot/dts/qcom/ipq6018.dtsi
@@ -41,8 +41,6 @@ CPU0: cpu@0 {
next-level-cache = <&L2_0>;
clocks = <&apcs_glb APCS_ALIAS0_CORE_CLK>;
clock-names = "cpu";
- operating-points-v2 = <&cpu_opp_table>;
- cpu-supply = <&ipq6018_s2>;
};

CPU1: cpu@1 {
@@ -53,8 +51,6 @@ CPU1: cpu@1 {
next-level-cache = <&L2_0>;
clocks = <&apcs_glb APCS_ALIAS0_CORE_CLK>;
clock-names = "cpu";
- operating-points-v2 = <&cpu_opp_table>;
- cpu-supply = <&ipq6018_s2>;
};

CPU2: cpu@2 {
@@ -65,8 +61,6 @@ CPU2: cpu@2 {
next-level-cache = <&L2_0>;
clocks = <&apcs_glb APCS_ALIAS0_CORE_CLK>;
clock-names = "cpu";
- operating-points-v2 = <&cpu_opp_table>;
- cpu-supply = <&ipq6018_s2>;
};

CPU3: cpu@3 {
@@ -77,8 +71,6 @@ CPU3: cpu@3 {
next-level-cache = <&L2_0>;
clocks = <&apcs_glb APCS_ALIAS0_CORE_CLK>;
clock-names = "cpu";
- operating-points-v2 = <&cpu_opp_table>;
- cpu-supply = <&ipq6018_s2>;
};

L2_0: l2-cache {
@@ -95,54 +87,6 @@ scm {
};
};

- cpu_opp_table: opp-table-cpu {
- compatible = "operating-points-v2-kryo-cpu";
- nvmem-cells = <&cpu_speed_bin>;
- opp-shared;
-
- opp-864000000 {
- opp-hz = /bits/ 64 <864000000>;
- opp-microvolt = <725000>;
- opp-supported-hw = <0xf>;
- clock-latency-ns = <200000>;
- };
-
- opp-1056000000 {
- opp-hz = /bits/ 64 <1056000000>;
- opp-microvolt = <787500>;
- opp-supported-hw = <0xf>;
- clock-latency-ns = <200000>;
- };
-
- opp-1320000000 {
- opp-hz = /bits/ 64 <1320000000>;
- opp-microvolt = <862500>;
- opp-supported-hw = <0x3>;
- clock-latency-ns = <200000>;
- };
-
- opp-1440000000 {
- opp-hz = /bits/ 64 <1440000000>;
- opp-microvolt = <925000>;
- opp-supported-hw = <0x3>;
- clock-latency-ns = <200000>;
- };
-
- opp-1608000000 {
- opp-hz = /bits/ 64 <1608000000>;
- opp-microvolt = <987500>;
- opp-supported-hw = <0x1>;
- clock-latency-ns = <200000>;
- };
-
- opp-1800000000 {
- opp-hz = /bits/ 64 <1800000000>;
- opp-microvolt = <1062500>;
- opp-supported-hw = <0x1>;
- clock-latency-ns = <200000>;
- };
- };
-
pmuv8: pmu {
compatible = "arm,cortex-a53-pmu";
interrupts = <GIC_PPI 7 (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_HIGH)>;
--
2.25.1



2024-01-31 13:04:56

by Dmitry Baryshkov

[permalink] [raw]
Subject: Re: [PATCH 1/3] arm64: dts: qcom: ipq6018: separate CPU OPP tables

On Wed, 31 Jan 2024 at 12:22, Chukun Pan <[email protected]> wrote:
>
> Some IPQ6000 SoCs do not come with PMIC (MP5496) chip,
> which causes cpufreq to be unavailable due to lack of
> cpu-supply. Separate CPU OPP tables from soc.dtsi to
> support versions with and without PMIC chip.

I went on and checked ipq6018.dtsi. It will need to be reworked before
we can continue with PMIC-less devices.

Obviously, the PMIC is not a part of the SoC. So please move the
"qcom,rpm-mp5496-regulators" node to the board files together with the
cpu-supply properties that reference that regulator.

The SoC itself supports all listed frequencies, so it is incorrect to
split the opp tables from the ipq6018.dtsi. Instead please patch the
PMIC-less boards in the following way:

#include "ipq6018.dtsi"
&cpu_opp_table {
/* the board doesn't have a PMIC, disable CPU frequencies which
require higher voltages */
/delete-node/ opp-1320000000;
/delete-node/ opp-1440000000;
};

>
> Signed-off-by: Chukun Pan <[email protected]>
> ---
> arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts | 1 +
> arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi | 74 ++++++++++++++++++++
> arch/arm64/boot/dts/qcom/ipq6018.dtsi | 56 ---------------
> 3 files changed, 75 insertions(+), 56 deletions(-)
> create mode 100644 arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi
>
> diff --git a/arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts b/arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts
> index f5f4827c0e17..06dfc2cb6b7f 100644
> --- a/arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts
> +++ b/arch/arm64/boot/dts/qcom/ipq6018-cp01-c1.dts
> @@ -8,6 +8,7 @@
> /dts-v1/;
>
> #include "ipq6018.dtsi"
> +#include "ipq6018-opp.dtsi"
>
> / {
> model = "Qualcomm Technologies, Inc. IPQ6018/AP-CP01-C1";
> diff --git a/arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi b/arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi
> new file mode 100644
> index 000000000000..9c0bed2d8bf5
> --- /dev/null
> +++ b/arch/arm64/boot/dts/qcom/ipq6018-opp.dtsi
> @@ -0,0 +1,74 @@
> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
> +/*
> + * IPQ60xx with PMIC (MP5496) CPU OPP tables
> + */
> +
> +/ {
> + cpu_opp_table: opp-table-cpu {
> + compatible = "operating-points-v2-kryo-cpu";
> + nvmem-cells = <&cpu_speed_bin>;
> + opp-shared;
> +
> + opp-864000000 {
> + opp-hz = /bits/ 64 <864000000>;
> + opp-microvolt = <725000>;
> + opp-supported-hw = <0xf>;
> + clock-latency-ns = <200000>;
> + };
> +
> + opp-1056000000 {
> + opp-hz = /bits/ 64 <1056000000>;
> + opp-microvolt = <787500>;
> + opp-supported-hw = <0xf>;
> + clock-latency-ns = <200000>;
> + };
> +
> + opp-1320000000 {
> + opp-hz = /bits/ 64 <1320000000>;
> + opp-microvolt = <862500>;
> + opp-supported-hw = <0x3>;
> + clock-latency-ns = <200000>;
> + };
> +
> + opp-1440000000 {
> + opp-hz = /bits/ 64 <1440000000>;
> + opp-microvolt = <925000>;
> + opp-supported-hw = <0x3>;
> + clock-latency-ns = <200000>;
> + };
> +
> + opp-1608000000 {
> + opp-hz = /bits/ 64 <1608000000>;
> + opp-microvolt = <987500>;
> + opp-supported-hw = <0x1>;
> + clock-latency-ns = <200000>;
> + };
> +
> + opp-1800000000 {
> + opp-hz = /bits/ 64 <1800000000>;
> + opp-microvolt = <1062500>;
> + opp-supported-hw = <0x1>;
> + clock-latency-ns = <200000>;
> + };
> + };
> +};
> +
> +&CPU0 {
> + operating-points-v2 = <&cpu_opp_table>;
> + cpu-supply = <&ipq6018_s2>;
> +};
> +
> +&CPU1 {
> + operating-points-v2 = <&cpu_opp_table>;
> + cpu-supply = <&ipq6018_s2>;
> +};
> +
> +&CPU2 {
> + operating-points-v2 = <&cpu_opp_table>;
> + cpu-supply = <&ipq6018_s2>;
> +};
> +
> +&CPU3 {
> + operating-points-v2 = <&cpu_opp_table>;
> + cpu-supply = <&ipq6018_s2>;
> +};
> diff --git a/arch/arm64/boot/dts/qcom/ipq6018.dtsi b/arch/arm64/boot/dts/qcom/ipq6018.dtsi
> index 5e1277fea725..ea72fd5739ac 100644
> --- a/arch/arm64/boot/dts/qcom/ipq6018.dtsi
> +++ b/arch/arm64/boot/dts/qcom/ipq6018.dtsi
> @@ -41,8 +41,6 @@ CPU0: cpu@0 {
> next-level-cache = <&L2_0>;
> clocks = <&apcs_glb APCS_ALIAS0_CORE_CLK>;
> clock-names = "cpu";
> - operating-points-v2 = <&cpu_opp_table>;
> - cpu-supply = <&ipq6018_s2>;
> };
>
> CPU1: cpu@1 {
> @@ -53,8 +51,6 @@ CPU1: cpu@1 {
> next-level-cache = <&L2_0>;
> clocks = <&apcs_glb APCS_ALIAS0_CORE_CLK>;
> clock-names = "cpu";
> - operating-points-v2 = <&cpu_opp_table>;
> - cpu-supply = <&ipq6018_s2>;
> };
>
> CPU2: cpu@2 {
> @@ -65,8 +61,6 @@ CPU2: cpu@2 {
> next-level-cache = <&L2_0>;
> clocks = <&apcs_glb APCS_ALIAS0_CORE_CLK>;
> clock-names = "cpu";
> - operating-points-v2 = <&cpu_opp_table>;
> - cpu-supply = <&ipq6018_s2>;
> };
>
> CPU3: cpu@3 {
> @@ -77,8 +71,6 @@ CPU3: cpu@3 {
> next-level-cache = <&L2_0>;
> clocks = <&apcs_glb APCS_ALIAS0_CORE_CLK>;
> clock-names = "cpu";
> - operating-points-v2 = <&cpu_opp_table>;
> - cpu-supply = <&ipq6018_s2>;
> };
>
> L2_0: l2-cache {
> @@ -95,54 +87,6 @@ scm {
> };
> };
>
> - cpu_opp_table: opp-table-cpu {
> - compatible = "operating-points-v2-kryo-cpu";
> - nvmem-cells = <&cpu_speed_bin>;
> - opp-shared;
> -
> - opp-864000000 {
> - opp-hz = /bits/ 64 <864000000>;
> - opp-microvolt = <725000>;
> - opp-supported-hw = <0xf>;
> - clock-latency-ns = <200000>;
> - };
> -
> - opp-1056000000 {
> - opp-hz = /bits/ 64 <1056000000>;
> - opp-microvolt = <787500>;
> - opp-supported-hw = <0xf>;
> - clock-latency-ns = <200000>;
> - };
> -
> - opp-1320000000 {
> - opp-hz = /bits/ 64 <1320000000>;
> - opp-microvolt = <862500>;
> - opp-supported-hw = <0x3>;
> - clock-latency-ns = <200000>;
> - };
> -
> - opp-1440000000 {
> - opp-hz = /bits/ 64 <1440000000>;
> - opp-microvolt = <925000>;
> - opp-supported-hw = <0x3>;
> - clock-latency-ns = <200000>;
> - };
> -
> - opp-1608000000 {
> - opp-hz = /bits/ 64 <1608000000>;
> - opp-microvolt = <987500>;
> - opp-supported-hw = <0x1>;
> - clock-latency-ns = <200000>;
> - };
> -
> - opp-1800000000 {
> - opp-hz = /bits/ 64 <1800000000>;
> - opp-microvolt = <1062500>;
> - opp-supported-hw = <0x1>;
> - clock-latency-ns = <200000>;
> - };
> - };
> -
> pmuv8: pmu {
> compatible = "arm,cortex-a53-pmu";
> interrupts = <GIC_PPI 7 (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_HIGH)>;
> --
> 2.25.1
>
>


--
With best wishes
Dmitry

2024-02-01 03:09:04

by Chukun Pan

[permalink] [raw]
Subject: Re: [PATCH 1/3] arm64: dts: qcom: ipq6018: separate CPU OPP tables

Hi, Dmitry
> I went on and checked ipq6018.dtsi. It will need to be reworked before
> we can continue with PMIC-less devices.

> Obviously, the PMIC is not a part of the SoC. So please move the
> "qcom,rpm-mp5496-regulators" node to the board files together with the
> cpu-supply properties that reference that regulator.

Thanks a lot for your advice, now things are clearer.
My idea is as follows:

1. Add all frequencies supported by SoCs in ipq6018.dtsi

2. Move cpu-supply and mp5496 nodes to ipq6018-mp5496.dtsi

&CPU0 {
cpu-supply = <&ipq6018_s2>;
};
..

&rpm_requests {
regulators {
rpm-mp5496...
ipq6018_s2...
};
};

> The SoC itself supports all listed frequencies, so it is incorrect to
> split the opp tables from the ipq6018.dtsi. Instead please patch the
> PMIC-less boards in the following way:

> #include "ipq6018.dtsi"
> &cpu_opp_table {
> /* the board doesn't have a PMIC, disable CPU frequencies which
> require higher voltages */
> /delete-node/ opp-1320000000;
> /delete-node/ opp-1440000000;
>};

Thank you but no need. The CPUFreq NVMEM driver will give the CPU
maximum frequency based on the cpu_speed_bin and opp-supported-hw.

Thanks,
Chukun

--
2.25.1


2024-02-01 03:16:34

by Dmitry Baryshkov

[permalink] [raw]
Subject: Re: [PATCH 1/3] arm64: dts: qcom: ipq6018: separate CPU OPP tables

On Thu, 1 Feb 2024 at 05:08, Chukun Pan <[email protected]> wrote:
>
> Hi, Dmitry
> > I went on and checked ipq6018.dtsi. It will need to be reworked before
> > we can continue with PMIC-less devices.
>
> > Obviously, the PMIC is not a part of the SoC. So please move the
> > "qcom,rpm-mp5496-regulators" node to the board files together with the
> > cpu-supply properties that reference that regulator.
>
> Thanks a lot for your advice, now things are clearer.
> My idea is as follows:
>
> 1. Add all frequencies supported by SoCs in ipq6018.dtsi
>
> 2. Move cpu-supply and mp5496 nodes to ipq6018-mp5496.dtsi

Straight to the board files, please, no need for additional includes.

>
> &CPU0 {
> cpu-supply = <&ipq6018_s2>;
> };
> ...
>
> &rpm_requests {
> regulators {
> rpm-mp5496...
> ipq6018_s2...
> };
> };
>
> > The SoC itself supports all listed frequencies, so it is incorrect to
> > split the opp tables from the ipq6018.dtsi. Instead please patch the
> > PMIC-less boards in the following way:
>
> > #include "ipq6018.dtsi"
> > &cpu_opp_table {
> > /* the board doesn't have a PMIC, disable CPU frequencies which
> > require higher voltages */
> > /delete-node/ opp-1320000000;
> > /delete-node/ opp-1440000000;
> >};
>
> Thank you but no need. The CPUFreq NVMEM driver will give the CPU
> maximum frequency based on the cpu_speed_bin and opp-supported-hw.

From your patches I had the feeling that you still want to limit the
high-frequency OPP entries if there is no PMIC.

>
> Thanks,
> Chukun
>
> --
> 2.25.1
>


--
With best wishes
Dmitry

2024-02-01 04:01:11

by Chukun Pan

[permalink] [raw]
Subject: Re: [PATCH 1/3] arm64: dts: qcom: ipq6018: separate CPU OPP tables

Hi, Dmitry
> Straight to the board files, please, no need for additional includes.

Is it possible to move the mp5496 node to mp5496.dtsi?
Because ipq9574 also uses this mp5496 pmic.

&rpm_requests {
regulators {
compatible = "qcom,rpm-mp5496-regulators";

mp5496_s1: s1 {
regulator-min-microvolt = <725000>;
regulator-max-microvolt = <1075000>;
status = "disabled";
};

mp5496_s2: s2 {
regulator-min-microvolt = <725000>;
regulator-max-microvolt = <1062500>;
status = "disabled";
};

mp5496_l2: l2 {
regulator-min-microvolt = <1800000>;
regulator-max-microvolt = <3300000>;
status = "disabled";
};
};
};

> From your patches I had the feeling that you still want to limit the
> high-frequency OPP entries if there is no PMIC.

Sorry for this misunderstanding, the cpu max frequency is determined
by the cpu_speed_bin. It's just that the efuse of the board I have
without pmic are all 1.2GHz.

Thanks,
Chukun

--
2.25.1


2024-02-01 04:38:41

by Dmitry Baryshkov

[permalink] [raw]
Subject: Re: [PATCH 1/3] arm64: dts: qcom: ipq6018: separate CPU OPP tables

On Thu, 1 Feb 2024 at 06:00, Chukun Pan <[email protected]> wrote:
>
> Hi, Dmitry
> > Straight to the board files, please, no need for additional includes.
>
> Is it possible to move the mp5496 node to mp5496.dtsi?
> Because ipq9574 also uses this mp5496 pmic.
>
> &rpm_requests {
> regulators {
> compatible = "qcom,rpm-mp5496-regulators";
>
> mp5496_s1: s1 {
> regulator-min-microvolt = <725000>;
> regulator-max-microvolt = <1075000>;
> status = "disabled";
> };
>
> mp5496_s2: s2 {
> regulator-min-microvolt = <725000>;
> regulator-max-microvolt = <1062500>;
> status = "disabled";
> };
>
> mp5496_l2: l2 {
> regulator-min-microvolt = <1800000>;
> regulator-max-microvolt = <3300000>;
> status = "disabled";
> };
> };
> };

Usually this is a bad idea, the regulator boundaries are board specific.

>
> > From your patches I had the feeling that you still want to limit the
> > high-frequency OPP entries if there is no PMIC.
>
> Sorry for this misunderstanding, the cpu max frequency is determined
> by the cpu_speed_bin. It's just that the efuse of the board I have
> without pmic are all 1.2GHz.

Ack, sounds good then.

--
With best wishes
Dmitry