Re: [PATCH v1 2/7] dt-bindings: display: mediatek: Add binding for MT8195 HDMI-TX v2

[AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>]

Il 29/11/24 03:51, CK Hu (胡俊光) ha scritto:
> On Thu, 2024-11-28 at 11:32 +0100, AngeloGioacchino Del Regno wrote:
>> External email : Please do not click links or open attachments until you have verified the sender or the content.
>>
>>
>> Il 28/11/24 07:02, CK Hu (胡俊光) ha scritto:
>>> Hi, Angelo:
>>>
>>> On Wed, 2024-11-20 at 13:45 +0100, AngeloGioacchino Del Regno wrote:
>>>> External email : Please do not click links or open attachments until you have verified the sender or the content.
>>>>
>>>>
>>>> Add a binding for the HDMI TX v2 Encoder found in MediaTek MT8195
>>>> and MT8188 SoCs.
>>>>
>>>> This fully supports the HDMI Specification 2.0b, hence it provides
>>>> support for 3D-HDMI, Polarity inversion, up to 16 bits Deep Color,
>>>> color spaces including RGB444, YCBCR420/422/444 (ITU601/ITU709) and
>>>> xvYCC, with output resolutions up to 3840x2160p@60Hz.
>>>>
>>>> Moreover, it also supports HDCP 1.4 and 2.3, Variable Refresh Rate
>>>> (VRR) and Consumer Electronics Control (CEC).
>>>>
>>>> This IP also includes support for HDMI Audio, including IEC60958
>>>> and IEC61937 SPDIF, 8-channel PCM, DSD, and other lossless audio
>>>> according to HDMI 2.0.
>>>>
>>>> Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
>>>> ---
>>>>    .../mediatek/mediatek,mt8195-hdmi.yaml        | 150 ++++++++++++++++++
>>>>    1 file changed, 150 insertions(+)
>>>>    create mode 100644 Documentation/devicetree/bindings/display/mediatek/mediatek,mt8195-hdmi.yaml
>>>>
>>>> diff --git a/Documentation/devicetree/bindings/display/mediatek/mediatek,mt8195-hdmi.yaml b/Documentation/devicetree/bindings/display/mediatek/mediatek,mt8195-hdmi.yaml
>>>> new file mode 100644
>>>> index 000000000000..273a8871461e
>>>> --- /dev/null
>>>> +++ b/Documentation/devicetree/bindings/display/mediatek/mediatek,mt8195-hdmi.yaml
>>>> @@ -0,0 +1,150 @@
>>>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>>>> +%YAML 1.2
>>>> +---
>>>> +$id: https://urldefense.com/v3/__http://devicetree.org/schemas/display/mediatek/mediatek,mt8195-hdmi.yaml*__;Iw!!CTRNKA9wMg0ARbw!lu0D_C3TwQ2-02jWYABnMIQ8vEoUwP0O4gbQndJnPUMpdi6wXdAHra9ivCfB7zoelDI7qsS20YdRlmP4bEKAABletXFX$
>>>> +$schema: https://urldefense.com/v3/__http://devicetree.org/meta-schemas/core.yaml*__;Iw!!CTRNKA9wMg0ARbw!lu0D_C3TwQ2-02jWYABnMIQ8vEoUwP0O4gbQndJnPUMpdi6wXdAHra9ivCfB7zoelDI7qsS20YdRlmP4bEKAAFlnY-KY$
>>>> +
>>>> +title: MediaTek HDMI-TX v2 Encoder
>>>> +
>>>> +maintainers:
>>>> +  - AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
>>>> +  - CK Hu <ck.hu@mediatek.com>
>>>> +
>>>> +description: |
>>>> +  The MediaTek HDMI-TX v2 encoder can generate HDMI format data based on
>>>> +  the HDMI Specification 2.0b.
>>>> +
>>>> +properties:
>>>> +  compatible:
>>>> +    enum:
>>>> +      - mediatek,mt8188-hdmi-tx
>>>> +      - mediatek,mt8195-hdmi-tx
>>>> +
>>>> +  reg:
>>>> +    maxItems: 1
>>>> +
>>>> +  interrupts:
>>>> +    maxItems: 1
>>>> +
>>>> +  clocks:
>>>> +    items:
>>>> +      - description: HDMI APB clock
>>>> +      - description: HDCP top clock
>>>> +      - description: HDCP reference clock
>>>> +      - description: VPP HDMI Split clock
>>>
>>> I would like to know more about HDMI v2.
>>> Would you map each v2 clock to v1 clock?
>>> If one clock has no mapping, is it a new feature that v1 does not has?
>>>
>>
>> The HDMIv2 HW block seems to be almost completely different from the v1, and
>> it is also interconnected in a different way compared to MT8173 (the path goes
>> through VPP1, while the v1 is just direct to DPI/MMSYS).
>>
>> The v1 block had specific clocks for the audio (i2s, I believe) and for the SPDIF,
>> and I have no idea how v1 does HDCP, but I don't see any specific clock for that.
>>
>> The v2 block is clocked from the HDCP clock, the (apb) bus has its own clock, and
>> the video out needs the vpp split clock.
>>
>> It's just different, and we can't shove the v2 binding inside of the v1 one, but
>> even if we could, since the v2 block is *that much* different from v1, it'd be a
>> mistake to do so.
>>
>> Since the binding describes hardware, and since this v2 HW is *very* different
>> from v1, it needs a new binding document, that is true even if you find a way to
>> get the clocks to match (which is not possible, anyway).
> 

Hey. Sorry about the late reply, it's been quite busy days here :-)

> v2 indeed is very different from v1, so it's not necessary to merge binding document.
> I would like to have more information about the difference in binding document,
> so that we could clearly understand that v1 and v2 are so different.
> 

Sounds reasonable.

> I think pixel clock is important for HDMI hardware, but I do not see it in HDMI v2.
> It is better has some documentation about why pixel clock disappear in HDMI v2.

Right. I'll write it in the PHY description.

   phys:
     maxItems: 1
     description:
        The HDMI TX PHY, clocking TMDS and pixel to this controller

> 
> I've some 'WHY' about v2.
> Why no audio clock in v2?

It's clocked by the HDCP CGs...

> Audio control part is moved out of HDMI block?
> 
> For HDCP, maybe v1 driver has not implement it so forget to add it in binding document.
> So just skip the HDCP.

...and the controller needs HDCP clocks for register access, so these are
main clocks and cannot be skipped.

> 
> The four clock in v2 does not exist in v1, so what is the function of each one?
> If possible, ask MediaTek staff for more information.

I'll try to ask to shove more information in this binding, let's see.

Cheers,
Angelo

> 
> Regards,
> CK
> 
>>
>> Cheers,
>> Angelo
>>
>>> Regards,
>>> CK
>>>
>>>> +
>>>> +  clock-names:
>>>> +    items:
>>>> +      - const: bus
>>>> +      - const: hdcp
>>>> +      - const: hdcp24m
>>>> +      - const: hdmi-split
>>>> +
>>>>
>>>> --
>>>> 2.47.0
>>>>
>>
>>