[PATCH v3 3/3] drm/bridge: add Silicon Image SiI8620 driver

[Daniel Vetter]

On Thu, Oct 06, 2016 at 11:02:58AM +0200, Andrzej Hajda wrote:
> Hi Daniel, Archit,
> 
> On 30.09.2016 12:33, Andrzej Hajda wrote:
> > On 30.09.2016 12:07, Daniel Vetter wrote:
> >> On Fri, Sep 30, 2016 at 09:30:16AM +0530, Archit Taneja wrote:
> >>> Hi Andrezj,
> >>>
> >>> On 09/26/2016 07:10 PM, Andrzej Hajda wrote:
> >>>> SiI8620 transmitter converts eTMDS/HDMI signal to MHL 3.0.
> >>>> It is controlled via I2C bus. Its interaction with other
> >>>> devices in video pipeline is performed mainly on HW level.
> >>>> The only interaction it does on device driver level is
> >>>> filtering-out unsupported video modes, it exposes drm_bridge
> >>>> interface to perform this operation.
> >>> The patchset looks good to me. Is the MHL header patch
> >>> accepted? I was wondering how we pull this in.
> >>>
> >>> +Daniel
> >> I think someone with real clue about what MHL is needs to review that
> >> header. Also I have no idea why that's under video/, is there another
> >> driver in media we want to share this with?
> >> -Daniel
> >>
> > I have put it into include/linux as MHL could be used to transmit:
> > - video,
> > - audio,
> > - remote control protocol (input device),
> > - ... embed other protocols (USB for example),
> >
> > But since I am not aware of other MHL users in near future
> > I can put the header together with the bridge driver.
> 
> These patches are hanging on the list for almost year,
> since Archit decided to review it (thanks Archit), I would
> like to end this process.
> 
> The options I see:
> 1. Leave it as is, mhl.h is like hdmi.h - it can server for
>     multiple subsystems. I guess it can be hard to find
>     MHL specialist to review it as it does not seems
>     to be popular subject, on the other side it is only
>     in-kernel header so it should pose serious danger.
> 2. Move the header to some of drm dirs:
>     a) drivers/gpu/drm/bridge/
>     b) include/drm/bridge/
>     c) include/drm/
>     ...
> 3. Incorporate it into drivers/gpu/drm/bridge/sil-sii8620.h
>     This is the least problematic solution, but possible
>     future abstraction of MHL will be more noisy.
> 
> Daniel, which option do you prefer? For me any option
> is OK, I just want to end this little bit frustrating process.

I just brought this up as a question, I don't personally care all that
much. Except for option 2a) I think they are all ok (we only have internal
headers outside of include/). Whatever you&Archit can agree on is fine
with me (and then Archit can all push it directly to drm-misc).
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch