lkml.org 
[lkml]   [2023]   [Oct]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH] arm64: dts: qcom: sc8280xp-x13s: Use the correct DP PHY compatible
From


On 10/17/23 05:28, Bjorn Andersson wrote:
> On Mon, Oct 16, 2023 at 12:10:18PM +0300, Dmitry Baryshkov wrote:
>> On Mon, 16 Oct 2023 at 12:01, Johan Hovold <johan@kernel.org> wrote:
>>>
>>> On Mon, Oct 16, 2023 at 11:51:33AM +0300, Dmitry Baryshkov wrote:
>>>> On Fri, 29 Sept 2023 at 19:03, Konrad Dybcio <konrad.dybcio@linaro.org> wrote:
>>>>>
>>>>> The DP PHY needs different settings when an eDP display is used.
>>>>> Make sure these apply on the X13s.
>>>>
>>>> Could you please clarify, is it the same PHY type, just being
>>>> repurposed for eDP or is it a different PHY type?
>>>
>>> Same PHY, just different settings AFAIK.
>>>
>>>> If the former is the case (and the same PHY can be used for both DP
>>>> and eDP), it should carry the same compatible string and use software
>>>> mechanisms (e.g. phy_set_mode_ext()) to be programmed for the correct
>>>> operation mode.
>>>
>>> Possibly, but that's not how the current binding and implementation
>>> works:
>>>
>>> 6993c079cd58 ("dt-bindings: phy: qcom-edp: Add SC8280XP PHY compatibles")
>>> 2300d1cb24b3 ("phy: qcom: edp: Introduce support for DisplayPort")
>>> 3b7267dec445 ("phy: qcom: edp: Add SC8280XP eDP and DP PHYs")
>>>
>>> https://lore.kernel.org/lkml/20220810040745.3582985-1-bjorn.andersson@linaro.org/
>>>
>>> And you'd still need to infer the mode from DT somehow.
>>
>> If it is the same hardware block, it seems incorrect to have two
>> different compat entries. For example, for PCIe RC vs PCIe EP we
>> specify the PHY mode from the host controller driver.
>> I'd say, we need to fix the bindings for both DP/eDP controller and
>> the PHY. See the `phy-mode` DT property for example.
>>
>
> It is one hardware block, supporting both eDP and DP, so I like your
> suggestion of having a single compatible instead and using some other
> means of defining the configuration. I just wasn't able to find a
> better way to do so back when I wrote the binding/driver...
Since this one is still unused, we can deprecate it (not sure if remove,
but deprecate) and add phy-type instead. I was quite surprised to see
that a new compatible was added as well :/

Konrad

\
 
 \ /
  Last update: 2023-10-17 18:23    [W:0.058 / U:0.472 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site