lkml.org 
[lkml]   [2022]   [Mar]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH v2 2/3] dt-bindings: clock: add QCOM SM6125 display clock bindings
From
On 02/03/2022 13:54, Marijn Suijten wrote:
> On 2022-02-28 10:23:19, Krzysztof Kozlowski wrote:
>> On 27/02/2022 22:43, Dmitry Baryshkov wrote:
>>> On 27/02/2022 13:03, Krzysztof Kozlowski wrote:
>>>> On 26/02/2022 21:09, Marijn Suijten wrote:
>>>>> From: Martin Botka <martin.botka@somainline.org>
>>>>>
>>>>> Add device tree bindings for display clock controller for
>>>>> Qualcomm Technology Inc's SM6125 SoC.
>>>>>
>>>>> Signed-off-by: Martin Botka <martin.botka@somainline.org>
>>>>> ---
>>>>> .../bindings/clock/qcom,dispcc-sm6125.yaml | 87 +++++++++++++++++++
>>>>> .../dt-bindings/clock/qcom,dispcc-sm6125.h | 41 +++++++++
>>>>> 2 files changed, 128 insertions(+)
>>>>> create mode 100644 Documentation/devicetree/bindings/clock/qcom,dispcc-sm6125.yaml
>>>>> create mode 100644 include/dt-bindings/clock/qcom,dispcc-sm6125.h
>>>>>
>>>>> diff --git a/Documentation/devicetree/bindings/clock/qcom,dispcc-sm6125.yaml b/Documentation/devicetree/bindings/clock/qcom,dispcc-sm6125.yaml
>>>>> new file mode 100644
>>>>> index 000000000000..3465042d0d9f
>>>>> --- /dev/null
>>>>> +++ b/Documentation/devicetree/bindings/clock/qcom,dispcc-sm6125.yaml
>>>>> @@ -0,0 +1,87 @@
>>>>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>>>>> +%YAML 1.2
>>>>> +---
>>>>> +$id: http://devicetree.org/schemas/clock/qcom,dispcc-sm6125.yaml#
>>>>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>>>>> +
>>>>> +title: Qualcomm Display Clock Controller Binding for SM6125
>>>>> +
>>>>> +maintainers:
>>>>> + - Martin Botka <martin.botka@somainline.org>
>>>>> +
>>>>> +description: |
>>>>> + Qualcomm display clock control module which supports the clocks and
>>>>> + power domains on SM6125.
>>>>> +
>>>>> + See also:
>>>>> + dt-bindings/clock/qcom,dispcc-sm6125.h
>>>>> +
>>>>> +properties:
>>>>> + compatible:
>>>>> + enum:
>>>>> + - qcom,sm6125-dispcc
>>>>> +
>>>>> + clocks:
>>>>> + items:
>>>>> + - description: Board XO source
>>>>> + - description: Byte clock from DSI PHY0
>>>>> + - description: Pixel clock from DSI PHY0
>>>>> + - description: Pixel clock from DSI PHY1
>>>>> + - description: Link clock from DP PHY
>>>>> + - description: VCO DIV clock from DP PHY
>>>>> + - description: AHB config clock from GCC
>>>>> +
>>>>> + clock-names:
>>>>> + items:
>>>>> + - const: bi_tcxo
>>>>> + - const: dsi0_phy_pll_out_byteclk
>>>>> + - const: dsi0_phy_pll_out_dsiclk
>>>>> + - const: dsi1_phy_pll_out_dsiclk
>>>>> + - const: dp_phy_pll_link_clk
>>>>> + - const: dp_phy_pll_vco_div_clk
>>>>> + - const: cfg_ahb_clk
>>>>> +
>>>>> + '#clock-cells':
>>>>> + const: 1
>>>>> +
>>>>> + '#power-domain-cells':
>>>>> + const: 1
>>>>> +
>>>>> + reg:
>>>>> + maxItems: 1
>>>>> +
>>>>> +required:
>>>>> + - compatible
>>>>> + - reg
>>>>> + - clocks
>>>>> + - clock-names
>>>>> + - '#clock-cells'
>>>>> + - '#power-domain-cells'
>>>>> +
>>>>> +additionalProperties: false
>>>>> +
>>>>> +examples:
>>>>> + - |
>>>>> + #include <dt-bindings/clock/qcom,rpmcc.h>
>>>>> + #include <dt-bindings/clock/qcom,gcc-sm6125.h>
>>>>> + clock-controller@5f00000 {
>>>>> + compatible = "qcom,sm6125-dispcc";
>>>>> + reg = <0x5f00000 0x20000>;
>>>>> + clocks = <&rpmcc RPM_SMD_XO_CLK_SRC>,
>>>>> + <&dsi0_phy 0>,
>>>>> + <&dsi0_phy 1>,
>>>>> + <0>,
>>>>
>>>> This does not look like a valid phandle. This clock is required, isn't it?
>
> I remember it being used like this before, though upon closer inspection
> only qcom,gcc-msm8998.yaml uses it as example.
>
> The clock should be optional, in that case it is perhaps desired to omit
> it from clock-names instead, or pretend there's a `dsi1_phy 1`?

I propose to omit it.

>
>>>
>>> Not, it's not required for general dispcc support.
>>> dispcc uses DSI and DP PHY clocks to provide respective pixel/byte/etc
>>> clocks. However if support for DP is not enabled, the dispcc can work
>>> w/o DP phy clock. Thus we typically add 0 phandles as placeholders for
>
> Is there any semantic difference between omitting the clock from DT (in
> clocks= /and/ clock-names=) or setting it to a 0 phandle?

Yes, there is. The DT validation does not check the meaning behind
values, so there is no difference between valid phandle/ID and 0. While
not having a clock at all is spotted by validation.

>
>>> DSI/DP clock sources and populate them as support for respective
>>> interfaces gets implemented.
>>>
>>
>> Then the clock is optional, isn't it? While not modeling it as optional?
>
> It looks like this should be modelled using minItems: then, and
> "optional" text/comment? Other clocks are optional as well, we don't
> have DSI 1 in downstream SM6125 DT sources and haven't added the DP PLL
> in our to-be-upstreamed mainline tree yet.

Are they really optional? Or maybe they should not even be provided?


Best regards,
Krzysztof

\
 
 \ /
  Last update: 2022-03-02 14:55    [W:0.072 / U:1.276 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site