On Wed, 30 Oct 2024 at 11:02, Biju Das <biju.das.jz@xxxxxxxxxxxxxx> wrote:
>
> Hi Dmitry Baryshkov,
>
> > -----Original Message-----
> > From: Dmitry Baryshkov <dmitry.baryshkov@xxxxxxxxxx>
> > Sent: 30 October 2024 03:17
> > Subject: Re: [PATCH v4 00/13] Add ITE IT6263 LVDS to HDMI converter support
> >
> > On Tue, 29 Oct 2024 at 04:41, Liu Ying <victor.liu@xxxxxxx> wrote:
> > >
> > > On 10/28/2024, Dmitry Baryshkov wrote:
> > > > On Mon, Oct 28, 2024 at 11:12:00AM +0000, Biju Das wrote:
> > > >> Hi Dmitry, Liu,
> > > >>
> > > >>> -----Original Message-----
> > > >>> From: Dmitry Baryshkov <dmitry.baryshkov@xxxxxxxxxx>
> > > >>> Sent: 28 October 2024 10:20
> > > >>> Subject: Re: [PATCH v4 00/13] Add ITE IT6263 LVDS to HDMI
> > > >>> converter support
> > > >>>
> > > >>> Hi,
> > > >>>
> > > >>> On Mon, 28 Oct 2024 at 04:37, Liu Ying <victor.liu@xxxxxxx> wrote:
> > > >>>>
> > > >>>> Hi,
> > > >>>>
> > > >>>> This patch series aims to add ITE IT6263 LVDS to HDMI converter
> > > >>>> on i.MX8MP EVK. Combined with LVDS receiver and HDMI 1.4a
> > > >>>> transmitter, the IT6263 supports LVDS input and HDMI 1.4 output
> > > >>>> by conversion function. IT6263 product link can be found at [1].
> > > >>>>
> > > >>>> Patch 1 is a preparation patch to allow display mode of an
> > > >>>> existing panel to pass the added mode validation logic in patch 3.
> > > >>>>
> > > >>>> Patch 2 allows i.MX8MP LVDS Display Bridge(LDB) bridge driver to
> > > >>>> find the next non-panel bridge, that is the IT6263 in this case.
> > > >>>>
> > > >>>> Patch 3 adds mode validation logic to i.MX8MP LDB bridge driver
> > > >>>> against "ldb" clock so that it can filter out unsupported display
> > > >>>> modes read from EDID.
> > > >>>>
> > > >>>> Patch 4 adds MEDIA_BUS_FMT_RGB101010_1X7X5_{SPWG,JEIDA} support,
> > > >>>> as they are supported by IT6263(with LVDS data bit reversed order).
> > > >>>>
> > > >>>> Patch 5 makes drm_of.c use MEDIA_BUS_FMT_RGB101010_1X7X5_{JEIDA,SPWG}.
> > > >>>>
> > > >>>> Patch 6 supports getting dual-link LVDS pixel order for the sink
> > > >>>> side as needed by IT6263 driver.
> > > >>>>
> > > >>>> Patch 7 documents jeida-30 and vesa-30 data mappings in
> > > >>>> lvds-data-mapping.yaml, as needed by IT6263 DT binding.
> > > >>>>
> > > >>>> Patch 8 extracts common dual-link LVDS display properties into
> > > >>>> new lvds-dual-ports.yaml so that IT6263 DT binding can reference it.
> > > >>>>
> > > >>>> Patch 9 adds DT binding for IT6263.
> > > >>>>
> > > >>>> Patch 10 adds IT6263 bridge driver. Only video output is supported.
> > > >>>>
> > > >>>> Patch 11 adds DT overlays to support NXP adapter cards[2][3] with
> > > >>>> IT6263 populated.
> > > >>>>
> > > >>>> Patch 12 enables the IT6263 bridge driver in defconfig.
> > > >>>>
> > > >>>> Patch 13 updates MAINTAINERS to add maintainer for IT6263 driver.
> > > >>>
> > > >>> This has pretty complicated structure from the merging point of view.
> > > >>>
> > > >>> I propose we take patches 6, 8, 9 (without 30-bit formats, they
> > > >>> can be dropped while applying), 11, 12
> > > >>> (?) and 13 through drm-misc in one batch (once DT maintainers
> > > >>> review the binding parts). This looks like a minimal set, having no extra dependencies.
> > > >>
> > > >>>
> > > >>> The second set might be 4, 5 + new patch, re-adding 30-bit formats
> > > >>> to
> > > >>> IT6263 binding (no driver changes are necessary). This can go in
> > > >>> separately, after an Ack from media maintainers.
> > > >>>
> > > >>> Of course both sets can go together if linux-media maintainers
> > > >>> reacts quickly and ack merging media- formats patch through drm-misc tree.
> > >
> > > I'm fine with merging the two sets through drm-misc tree as long as
> > > linux-media and dri-devel maintainers accept this. Up to them.
> > >
> > > >>>
> > > >>> The rest of the patches don't have such strong dependencies and go in once ready / reviewed.
> > > >>>
> > > >>> WDYT?
> > > >>
> > > >> I guess, 6,8,9(without 30-bit formats), 10, 12 and 13.
> > > >>
> > > >> 11 may have dependency on 1, 2 and 3 as it is SoC specific.
> > > >
> > > > Yes, of course, 10, not 11.
> > > >
> > > >> Then 4, 5 + new patch, re-adding 30-bit formats to IT6263 binding.
> > >
> > > I think it would be good to directly support 30-bit formats in
> > > IT6263 DT binding, not re-add them to it. This way, we'll have one
> > > version of the binding, not two. So, a better first set would contain
> > > patch 6, 7(one existing A-b from Krzysztof), 8, 9, 10, 12 and 13.
> >
> > I'm not sure that 7 can go without an ack from linux-media maintainers.
>
> You mean in describing jeida-30 and vesa-30 format in
> patch#7, is valid only if patch#4 is ok with media people
> or they provide an ack for patch#7 to take it through drm tree?
The former one. I'd prefer an ack from linux-media maintainers to
accept bindings based on those names.
--
With best wishes
Dmitry
