Re: [PATCH v11 00/74] Add HDMI 2.0 support to DW HDMI QP TX

Hi,

On Tue, Sep 01, 2026 at 09:50:24PM +0300, Cristian Ciocaltea wrote:

Enable HDMI 2.0 display modes (e.g. 4K@60Hz) on the Synopsys DW HDMI QP
TX controller, as found in Rockchip RK3576 & RK3588 SoCs, by adding SCDC
management for high TMDS clock ratio and scrambling.  Since SCDC state
is lost on sink disconnects, the bridge driver needs to trigger a CRTC
reset during connector detection.

To support this at the DRM infrastructure level, the series first
introduces the HDMI version enum, then prepares for changing the
signature of drmm_connector_hdmi_init(), i.e. dropping the vendor,
product, supported_formats and max_bpc arguments, which are being moved
into struct drm_connector_hdmi_funcs, by temporarily renaming the helper
to drmm_connector_hdmi_ini2().  This lets the new signature be
introduced under the original name while callers are converted
incrementally.  Note the transitional name matches the original's length
so continuation-line arguments stay aligned to the opening parenthesis,
keeping the diff to the identifier itself and avoiding re-alignment
churn.

Appending more HDMI-specific arguments to the init function would not
scale well, hence the hdmi_funcs struct is extended with new fields:
supported_hdmi_ver, supported_tmds_char_rate.  These are used to
infer and/or limit the maximum TMDS character rate permitted for the
connector.

Patches 6-11 build the connector and bridge scrambling infrastructure on
top: the connector scrambler callbacks/flags, the scdc-helper
additions (connector-prefixed debug macro and SCDC version helper), and
the HDMI scrambling management helpers including SCDC source-version
advertisement.

Patches 12-25 wires this up through the hdmi-state-helper and bridge
connector layers: source TMDS rate validation, hotplug SCDC state sync
and the scrambling requirement, new source-side scrambling bridge ops,
the switch to a cached-status, atomic-aware .detect_ctx() connector
helper, and finally hooking up the HDMI 2.0 scrambler callbacks.

The SCDC scrambling feature itself is implemented in the DW HDMI QP
bridge driver, alongside i2c error-message rate limiting,
.enable_hpd()/.disable_hpd() PHY ops and a dw_hdmi_qp_hpd_notify()
helper (patches 26-30).

Patches 31-41 cover the Rockchip platform driver and HPD handling: bug
fixes, minor cleanups, avoiding spurious HDP IRQ wakeups, masking the
RK3576 HPD IRQ in io_init, implementing the .{enable|disable}_hpd() PHY
ops, switching HPD reports to dw_hdmi_qp_hpd_notify() to restrict events
to the affected connector, dropping the now-unused .setup_hpd() PHY op.

Patches 42-48 convert VC4 HDMI to the common infrastructure as a proof
of reuse: adopting the shared TMDS char rate constants, switching to
drm_hdmi_mode_needs_scrambling() and force_ctx(), proper -EDEADLK
handling, and replacing the driver-local scrambling implementation with
the common SCDC scrambling helpers.

Patches 50-57 advertise supported version for all HDMI
connectors/bridges, except the MediaTek HDMI driver (for now), as it
requires non-trivial changes: there's a core library shared between two
drivers, one being HDMI 2.0 compliant.

Patches 58-67 add KUnit tests: HDMI caps-based init coverage and
source-side scrambler validation for the connector, conversion of
hdmi_state_helper to the new drmm_connector_hdmi_init() with
max_tmds_char_rate validation, a new 4K@60Hz 600MHz TMDS EDID,
source-side scrambling decision coverage, and conformity fixes for the
existing 1080p+4K YUV420 200MHz and 4K RGB/YUV 340MHz test EDID blobs.

Patches 68-71 remove the now unused drmm_connector_hdmi_ini2() helper
and redundant hdmi fields from connector structs: vendor, product,
supported_formats and max_bpc.

Eventually patches 72-74 consolidate SPD InfoFrame handling of
vendor/product.

This has been tested on the following boards:

* Radxa ROCK 5B (RK3588)
* Radxa ROCK 4D (RK3576)
* Raspberry Pi 5 Model B Rev 1.1 (BCM2712 D0)

This series is getting a bit insane. You have preliminary patches that
have been reviewed/acked already, you should merge them, and possibly
split the rest into preliminary series.

Maxime