[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87a0d129-163b-18ae-eb09-cf7df4048dad@baylibre.com>
Date: Fri, 13 Mar 2020 16:29:58 +0100
From: Neil Armstrong <narmstrong@...libre.com>
To: Laurent Pinchart <laurent.pinchart@...asonboard.com>
Cc: Phong LE <ple@...libre.com>, airlied@...ux.ie, daniel@...ll.ch,
robh+dt@...nel.org, mark.rutland@....com, a.hajda@...sung.com,
jonas@...boo.se, jernej.skrabec@...l.net, sam@...nborg.org,
mripard@...nel.org, heiko.stuebner@...obroma-systems.com,
linus.walleij@...aro.org, stephan@...hold.net, icenowy@...c.io,
broonie@...nel.org, mchehab+samsung@...nel.org,
davem@...emloft.net, gregkh@...uxfoundation.org,
Jonathan.Cameron@...wei.com, andriy.shevchenko@...ux.intel.com,
dri-devel@...ts.freedesktop.org, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 2/4] dt-bindings: display: bridge: add it66121 bindings
On 13/03/2020 15:17, Laurent Pinchart wrote:
> Hi Neil,
>
> On Fri, Mar 13, 2020 at 03:12:13PM +0100, Neil Armstrong wrote:
>> On 13/03/2020 14:40, Laurent Pinchart wrote:
>>> On Wed, Mar 11, 2020 at 01:51:33PM +0100, Phong LE wrote:
>>>> Add the ITE bridge HDMI it66121 bindings.
>>>>
>>>> Signed-off-by: Phong LE <ple@...libre.com>
>>>> ---
>>>> .../bindings/display/bridge/ite,it66121.yaml | 98 +++++++++++++++++++
>>>> 1 file changed, 98 insertions(+)
>>>> create mode 100644 Documentation/devicetree/bindings/display/bridge/ite,it66121.yaml
>>>>
[...]
>>>> +
>>>> + pclk-dual-edge:
>>>> + maxItems: 1
>>>> + description: enable pclk dual edge mode.
>>>
>>> I'm having a bit of trouble understanding how this operates. Looking at
>>> the driver code the property is only taken into account to calculate the
>>> maximum allowed frequency. How is the IT66121 configured for single vs.
>>> dual pixel clock edge mode ?
>>
>> Dual edge mode is Dual-Data-Rate mode, the normal mode is MEDIA_BUS_FMT_RGB888_1X24 and dual edge is
>> MEDIA_BUS_FMT_RGB888_2X12_LE (or MEDIA_BUS_FMT_RGB888_2X12_BE, not sure) on a single clock period.
>>
>> This should be negociated at runtime, but the bus width should be specified somewhere to select
>> one of the modes.
>
> How about replacing this property by bus-width to report the connected
> bus width ? It should then become an endpoint property.
It was my thought.
The mediatek dpi also uses this property, which is also very wrong in the same way.
Neil
[...]
Powered by blists - more mailing lists