[PATCH v7 1/2] drm/rockchip: inno-hdmi: Convert to drm bridge

[Dmitry Baryshkov]

On Fri, Sep 05, 2025 at 08:40:31AM +0800, Andy Yan wrote:
> 
> Hello Dmitry，
> 
> At 2025-09-05 08:05:06, "Dmitry Baryshkov" <dmitry.baryshkov at oss.qualcomm.com> wrote:
> >On Wed, Sep 03, 2025 at 07:07:38PM +0800, Andy Yan wrote:
> >> From: Andy Yan <andy.yan at rock-chips.com>
> >> 
> >> Convert it to drm bridge driver, it will be convenient for us to
> >> migrate the connector part to the display driver later.
> >> 
> >> Signed-off-by: Andy Yan <andy.yan at rock-chips.com>
> >> 
> >> ---
> >> 
> >> Changes in v7:
> >> - Rebase on latest drm-misc-next
> >> 
> >> Changes in v6:
> >> - Rebase on latest drm-misc-next
> >> - Link to V5: https://lore.kernel.org/linux-rockchip/20250512124615.2848731-1-andyshrk@163.com/
> >> 
> >> Changes in v5:
> >> - Split cleanup code to seperate patch
> >> - Switch to devm_drm_bridge_alloc() API
> >> - Link to V4: https://lore.kernel.org/linux-rockchip/20250422070455.432666-1-andyshrk@163.com/
> >> 
> >> Changes in v4:
> >> - Do not store colorimetry within inno_hdmi struct
> >> - Link to V3: https://lore.kernel.org/linux-rockchip/20250402123150.238234-1-andyshrk@163.com/
> >> 
> >> Changes in v3:
> >> - First included in v3
> >> - Link to V2: https://lore.kernel.org/dri-devel/20250325132944.171111-1-andyshrk@163.com/
> >> 
> >>  drivers/gpu/drm/bridge/Kconfig                |   7 +
> >>  drivers/gpu/drm/bridge/Makefile               |   1 +
> >>  .../inno_hdmi.c => bridge/inno-hdmi.c}        | 502 +++++-------------
> >>  drivers/gpu/drm/rockchip/Kconfig              |   1 +
> >>  drivers/gpu/drm/rockchip/Makefile             |   2 +-
> >>  drivers/gpu/drm/rockchip/inno_hdmi-rockchip.c | 188 +++++++
> >>  include/drm/bridge/inno_hdmi.h                |  33 ++
> >>  7 files changed, 366 insertions(+), 368 deletions(-)
> >>  rename drivers/gpu/drm/{rockchip/inno_hdmi.c => bridge/inno-hdmi.c} (69%)
> >>  create mode 100644 drivers/gpu/drm/rockchip/inno_hdmi-rockchip.c
> >>  create mode 100644 include/drm/bridge/inno_hdmi.h
> >> 
> >> @@ -637,14 +584,13 @@ static void inno_hdmi_init_hw(struct inno_hdmi *hdmi)
> >>  	hdmi_modb(hdmi, HDMI_STATUS, m_MASK_INT_HOTPLUG, v_MASK_INT_HOTPLUG(1));
> >>  }
> >>  
> >> -static int inno_hdmi_disable_frame(struct drm_connector *connector,
> >> -				   enum hdmi_infoframe_type type)
> >> +static int inno_hdmi_bridge_clear_infoframe(struct drm_bridge *bridge,
> >> +					    enum hdmi_infoframe_type type)
> >>  {
> >> -	struct inno_hdmi *hdmi = connector_to_inno_hdmi(connector);
> >> +	struct inno_hdmi *hdmi = bridge_to_inno_hdmi(bridge);
> >>  
> >>  	if (type != HDMI_INFOFRAME_TYPE_AVI) {
> >> -		drm_err(connector->dev,
> >> -			"Unsupported infoframe type: %u\n", type);
> >> +		drm_err(bridge->dev, "Unsupported infoframe type: %u\n", type);
> >>  		return 0;
> >>  	}
> >>  
> >> @@ -653,20 +599,19 @@ static int inno_hdmi_disable_frame(struct drm_connector *connector,
> >>  	return 0;
> >>  }
> >>  
> >> -static int inno_hdmi_upload_frame(struct drm_connector *connector,
> >> -				  enum hdmi_infoframe_type type,
> >> -				  const u8 *buffer, size_t len)
> >> +static int inno_hdmi_bridge_write_infoframe(struct drm_bridge *bridge,
> >> +					    enum hdmi_infoframe_type type,
> >> +					    const u8 *buffer, size_t len)
> >>  {
> >> -	struct inno_hdmi *hdmi = connector_to_inno_hdmi(connector);
> >> +	struct inno_hdmi *hdmi = bridge_to_inno_hdmi(bridge);
> >>  	ssize_t i;
> >>  
> >>  	if (type != HDMI_INFOFRAME_TYPE_AVI) {
> >> -		drm_err(connector->dev,
> >> -			"Unsupported infoframe type: %u\n", type);
> >> +		drm_err(bridge->dev, "Unsupported infoframe type: %u\n", type);
> >>  		return 0;
> >>  	}
> >>  
> >> -	inno_hdmi_disable_frame(connector, type);
> >> +	inno_hdmi_bridge_clear_infoframe(bridge, type);
> >>  
> >>  	for (i = 0; i < len; i++)
> >>  		hdmi_writeb(hdmi, HDMI_CONTROL_PACKET_ADDR + i, buffer[i]);
> >
> >It's not an issue for this patch (and I think it can be fixed after this
> >series is merged). I took a quick glance at frame programming. It feels
> >like the clear_infoframe should be poking at registers 0x9c / 0x9d. And
> >write_infoframe then can support HDMI, SPD and Audio infoframes in
> >addition to the AVI. I don't have hardware to experiment (nor time :-)),
> >but would there be a chance to improve this?
> 
> Okay, I'll keep your suggestions in mind and look for an opportunity to try them out later. 
> The hardware for this board is indeed very scarce at the moment—I put in a lot of effort just
> to get my hands on one. The main reason I'm modifying this code is to convert all of Rockchip's
> display interface drivers into bridge mode, which will make it easier to separate the connector
> part into the display driver side in the future.

No worries, it was just a quick observation from some TRM excerpts that
I found.

-- 
With best wishes
Dmitry