Re: [PATCH 00/14] Add initial support for the Rockchip RK3588 HDMI TX Controller

[Maxime Ripard]

On Mon, Jun 03, 2024 at 03:03:12PM GMT, Heiko Stuebner wrote:
> Am Montag, 3. Juni 2024, 14:14:17 CEST schrieb Andy Yan:
> > Hi Neil:
> > 
> > On 6/3/24 16:55, Neil Armstrong wrote:
> > > Hi Christian,
> > > 
> > > On 01/06/2024 15:12, Cristian Ciocaltea wrote:
> > >> The RK3588 SoC family integrates a Quad-Pixel (QP) variant of the
> > >> Synopsys DesignWare HDMI TX controller used in the previous SoCs.
> > >>
> > >> It is HDMI 2.1 compliant and supports the following features, among
> > >> others:
> > >>
> > > .
> > > 
> > > ..
> > > 
> > >> * SCDC I2C DDC access
> > >> * TMDS Scrambler enabling 2160p@60Hz with RGB/YCbCr4:4:4
> > >> * YCbCr4:2:0 enabling 2160p@60Hz at lower HDMI link speeds
> > >> * Multi-stream audio
> > >> * Enhanced Audio Return Channel (EARC)
> > > -> Those features were already supported by the HDMI 2.0a compliant HW, just
> > > list the _new_ features for HDMI 2.1
> > > 
> > > I did a quick review of your patchset and I don't understand why you need
> > > to add a separate dw-hdmi-qp.c since you only need simple variants of the I2C
> > > bus, infoframe and bridge setup.
> > > 
> > > Can you elaborate further ? isn't this Quad-Pixel (QP) TX controller version
> > > detectable at runtime ?
> > > 
> > > I would prefer to keep a single dw-hdmi driver if possible.
> >
> > The QP HDMI controller is a completely different variant with totally different
> > registers layout, see PATCH 13/14.
> > I think make it a separate driver will be easier for development and maintenance.
> 
> I'm with Andy here. Trying to navigate a driver for two IP blocks really
> sounds taxing especially when both are so different.

If it's a completely new controller, I agree that it needs a new driver,
but then why do we need to share code between the two?

Maxime

Attachment: signature.asc
Description: PGP signature