[PATCH v1 1/1] video: bridge-uclass: add get_display_timing ops

[Simon Glass]

Hi Svyatoslav,

On Sun, 16 Feb 2025 at 06:00, Svyatoslav Ryhel <clamor95 at gmail.com> wrote:
>
> нд, 16 лют. 2025 р. о 14:56 Simon Glass <sjg at chromium.org> пише:
> >
> > Hi Svyatoslav,
> >
> > On Sat, 15 Feb 2025 at 11:12, Svyatoslav Ryhel <clamor95 at gmail.com> wrote:
> > >
> > > сб, 15 лют. 2025 р. о 20:07 Simon Glass <sjg at chromium.org> пише:
> > > >
> > > > Hi Svyatoslav,
> > > >
> > > > On Sat, 15 Feb 2025 at 10:42, Svyatoslav Ryhel <clamor95 at gmail.com> wrote:
> > > > >
> > > > > сб, 15 лют. 2025 р. о 19:24 Simon Glass <sjg at chromium.org> пише:
> > > > > >
> > > > > > Hi Svyatoslav,
> > > > > >
> > > > > > On Sat, 15 Feb 2025 at 08:52, Svyatoslav Ryhel <clamor95 at gmail.com> wrote:
> > > > > > >
> > > > > > > Add get_display_timing ops for internal bridges linked to
> > > > > > > panels that do not support EDID (MIPI-DSI panels for example)
> > > > > > > or have EDID not routed.
> > > > > > >
> > > > > > > Signed-off-by: Svyatoslav Ryhel <clamor95 at gmail.com>
> > > > > > > ---
> > > > > > >  drivers/video/bridge/video-bridge-uclass.c | 11 +++++++++++
> > > > > > >  include/video_bridge.h                     | 18 ++++++++++++++++++
> > > > > > >  2 files changed, 29 insertions(+)
> > > > > >
> > > > > > This should have a test, e.g. in test/dm/video.c or similar.
> > > > > >
> > > > >
> > > > > I totally agree, but it seems that there is no tests for
> > > > > UCLASS_VIDEO_BRIDGE unless I am missing smth. May you pls point me to
> > > > > it? Maybe I can add simple bridge analog of linux driver as sandbox
> > > > > test case?
> > > >
> > > > Oh dear, that's my fault. Yes, please add something simple.
> > > >
> > >
> > > Surely, but I would need ofnode graph merged in first. Then I can add
> > > simple bridge and lvds encoder in linux style. And convert simple
> > > panel to be able to work with graph. I assume that would be correct
> > > course.
> >
> > Ultimately yes, but you could add a very simple video-bridge test to
> > start, that just gets a device and calls a driver method?
> >
>
> That would be double job. I was thinking about adding lvds-encoder
> and/or simple bridge, which I can use in sandbox to make a test and
> then it can be used on actual devices. For example transformers are
> modeled with lvds-encoder.

That sounds great to me.

Regards,
Simon