On Mon, Jul 01, 2019 at 02:27:51PM +0200, Andrzej Hajda wrote:
> On 01.07.2019 11:58, Maxime Ripard wrote:
> > On Fri, Jun 28, 2019 at 12:39:32PM +0200, Andrzej Hajda wrote:
> >> On 12.06.2019 17:20, Maxime Ripard wrote:
> >>>> I am not sure if I understand whole discussion here, but I also do not
> >>>> understand whole edp-connector thing.
> >>> The context is this one:
> >>> https://patchwork.freedesktop.org/patch/257352/?series=51182&rev=1
> >>> https://patchwork.freedesktop.org/patch/283012/?series=56163&rev=1
> >>> https://patchwork.freedesktop.org/patch/286468/?series=56776&rev=2
> >>>
> >>> TL;DR: This bridge is being used on ARM laptops that can come with
> >>> different eDP panels. Some of these panels require a regulator to be
> >>> enabled for the panel to work, and this is obviously something that
> >>> should be in the DT.
> >>>
> >>> However, we can't really describe the panel itself, since the vendor
> >>> uses several of them and just relies on the eDP bus to do its job at
> >>> retrieving the EDIDs. A generic panel isn't really working either
> >>> since that would mean having a generic behaviour for all the panels
> >>> connected to that bus, which isn't there either.
> >>>
> >>> The connector allows to expose this nicely.
> >> As VESA presentation says[1] eDP is based on DP but is much more
> >> flexible, it is up to integrator (!!!) how the connection, power
> >> up/down, initialization sequence should be performed. Trying to cover
> >> every such case in edp-connector seems to me similar to panel-simple
> >> attempt failure. Moreover there is no such thing as physical standard
> >> eDP connector. Till now I though DT connector should describe physical
> >> connector on the device, now I am lost, are there some DT bindings
> >> guidelines about definition of a connector?
> > This might be semantics but I guess we're in some kind of grey area?
> >
> > Like, for eDP, if it's soldered I guess we could say that there's no
> > connector. But what happens if for some other board, that signal is
> > routed through a ribbon?
> >
> > You could argue that there's no physical connector in both cases, or
> > that there's one in both, or one for the ribbon and no connector for
> > the one soldered in.
>
> This is not about ribbon vs soldering. It is about usage: this
> connection is static across the whole life of the device (except
> exceptional things: repair, non-standard usage, etc).

It doesn't have to be.

> And "the real connector" is (at least for me) something where
> end-user can connect/disconnect different things: USB, HDMI,
> ethernet, etc. And obviously to be functional it should be somehow
> standardized. So even if there could be some grey area, I do not see
> it here.

Well, if there's a ribbon connector, then you have a physical
connector, with the end user being able to connect / disconnect
various displays. It might not be the case with actual products, but
it's pretty common with SBCs to have that signal routed through a
connector, and the user has several options to connect a display to
it.

The line really is blurred.

> >> Maybe instead of edp-connector one would introduce integrator's specific
> >> connector, for example with compatible "olimex,teres-edp-connector"
> >> which should follow edp abstract connector rules? This will be at least
> >> consistent with below presentation[1] - eDP requirements depends on
> >> integrator. Then if olimex has standard way of dealing with panels
> >> present in olimex/teres platforms the driver would then create
> >> drm_panel/drm_connector/drm_bridge(?) according to these rules, I guess.
> >> Anyway it still looks fishy for me :), maybe because I am not
> >> familiarized with details of these platforms.
>
> > That makes sense yes
>
> And what if some panel can be used with this pseudo-connecter and in
> some different hw directly? Code duplication? DT overlays?

Overlays are a solution, but I would advocate to always have the
connector.

Maxime

--
Maxime Ripard, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com