Re: [PATCH v2 5/5] dt-bindings: net: ipq4019-mdio: Document ipq5332 platform

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 





On 12/15/2023 1:12 AM, Conor Dooley wrote:
On Wed, Dec 13, 2023 at 04:26:56PM +0800, Jie Luo wrote:


On 12/13/2023 12:06 AM, Conor Dooley wrote:
On Tue, Dec 12, 2023 at 07:51:50PM +0800, Luo Jie wrote:
Update the yaml file for the new DTS properties.

1. cmn-reference-clock for the CMN PLL source clock select.
2. clock-frequency for MDIO clock frequency config.
3. add uniphy AHB & SYS GCC clocks.
4. add reset-gpios for MDIO bus level reset.

Signed-off-by: Luo Jie <quic_luoj@xxxxxxxxxxx>
---
   .../bindings/net/qcom,ipq4019-mdio.yaml       | 157 +++++++++++++++++-
   1 file changed, 153 insertions(+), 4 deletions(-)

diff --git a/Documentation/devicetree/bindings/net/qcom,ipq4019-mdio.yaml b/Documentation/devicetree/bindings/net/qcom,ipq4019-mdio.yaml
index 3407e909e8a7..9546a6ad7841 100644
--- a/Documentation/devicetree/bindings/net/qcom,ipq4019-mdio.yaml
+++ b/Documentation/devicetree/bindings/net/qcom,ipq4019-mdio.yaml
@@ -20,6 +20,8 @@ properties:
             - enum:
                 - qcom,ipq6018-mdio
                 - qcom,ipq8074-mdio
+              - qcom,ipq9574-mdio
+              - qcom,ipq5332-mdio
             - const: qcom,ipq4019-mdio
     "#address-cells":
@@ -30,19 +32,71 @@ properties:
     reg:
       minItems: 1
-    maxItems: 2
+    maxItems: 5
       description:
-      the first Address and length of the register set for the MDIO controller.
-      the second Address and length of the register for ethernet LDO, this second
-      address range is only required by the platform IPQ50xx.
+      the first Address and length of the register set for the MDIO controller,
+      the optional second, third and fourth address and length of the register
+      for ethernet LDO, these three address range are required by the platform
+      IPQ50xx/IPQ5332/IPQ9574, the last address and length is for the CMN clock
+      to select the reference clock.
+
+  reg-names:
+    minItems: 1
+    maxItems: 5
     clocks:
+    minItems: 1
       items:
         - description: MDIO clock source frequency fixed to 100MHZ
+      - description: UNIPHY0 AHB clock source frequency fixed to 100MHZ
+      - description: UNIPHY1 AHB clock source frequency fixed to 100MHZ
+      - description: UNIPHY0 SYS clock source frequency fixed to 24MHZ
+      - description: UNIPHY1 SYS clock source frequency fixed to 24MHZ
     clock-names:
+    minItems: 1
       items:
         - const: gcc_mdio_ahb_clk
+      - const: gcc_uniphy0_ahb_clk
+      - const: gcc_uniphy1_ahb_clk
+      - const: gcc_uniphy0_sys_clk
+      - const: gcc_uniphy1_sys_clk

+  cmn-reference-clock:
+    oneOf:
+      - items:
+          - enum:
+              - 0   # CMN PLL reference internal 48MHZ
+              - 1   # CMN PLL reference external 25MHZ
+              - 2   # CMN PLL reference external 31250KHZ
+              - 3   # CMN PLL reference external 40MHZ
+              - 4   # CMN PLL reference external 48MHZ
+              - 5   # CMN PLL reference external 50MHZ
+              - 6   # CMN PLL reference internal 96MHZ

Why is this not represented by an element of the clocks property?

This property is for the reference clock source selection of CMN PLL,
CMN PLL generates the different clock rates for the different Ethernet
blocks, this CMN PLL configuration is not located in the GCC, so the
clock framework can't be used, which is the general hardware register
instead of RCG register for GCC.

I don't see how the clock being provided by the "GCC" (whatever that is)
or by some other clock controller or fixed clock makes a difference.
Why can't the other clock provider be represented in the devicetree?


cmn-reference-clock is for selecting the reference clock source for the
whole Ethernet block, which is just the standalone configure register.
however the clock provider has the logical register distribution, such
as for one clock tree, there is RCG, DIVIDER and branch registers in
the qcom soc chip.

The clock consumer defines the clock IDs of device tree to reference the
clocks provided by the clock controller, and these clock IDs are
provided by the header file of clock provider.

like this,
clocks = <&gcc GCC_MDIO_AHB_CLK>, <&gcc GCC_UNIPHY0_AHB_CLK>, <&gcc GCC_UNIPHY1_AHB_CLK>, <&gcc GCC_UNIPHY0_SYS_CLK>,
         <&gcc GCC_UNIPHY1_SYS_CLK>;

gcc is the device node of clock provider, and GCC_MDIO_AHB_CLK is the clock ID.


+  clock-frequency:
+    oneOf:
+      - items:
+          - enum:
+              - 12500000
+              - 6250000
+              - 3125000
+              - 1562500
+              - 781250
+              - 390625
+    description:
+      The MDIO bus clock that must be output by the MDIO bus hardware,
+      only the listed frequecies above can be configured, other frequency
+      will cause malfunction. If absent, the default hardware value is used.

Likewise.

Your commit message contains a bullet point list of what you are doing,
but there's no explanation here for why custom properties are required
to provide clock information.

This property clock-frequency is optional to configure the MDIO working
clock rate, and this is the MDIO general DT property, since the hardware
default clock rate is 390625HZ, there is requirement for higher clock rate
in the normal working case, i will update this information in the
next patch set.

I'm just realising that this particular one is not a custom property,
the unusual `oneOf: - items: - enum:` structure here threw me. This can
just be
   clock-frequency:
     enum:
       - 12500000
       - 6250000
       - 3125000
       - 1562500
       - 781250
       - 390625

but you're missing a default, given your commit about the last element
in that list being one.

Thanks,
Conor.

Ok, will update this in the next patch set, thanks for the comments.




[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux