On Wed, Mar 18, 2026 at 10:46:36AM -0500, Sen Wang wrote: > Convert the IT66121 HDMI bridge driver from manually registering an > hdmi-codec platform device to using the DRM HDMI Audio Helper framework > via DRM_BRIDGE_OP_HDMI_AUDIO instead. > > The previous implementation manually allocated hdmi_codec_pdata, > registered the platform device, and implemented hdmi_codec_ops callbacks > including get_eld. The new approach sets DRM_BRIDGE_OP_HDMI_AUDIO on the > bridge, letting the framework handle the codec registration. This also > resolves some non-compliance issues with the current audio implementation, > such as HDMI audio advertising a non-functional capture stream to userspace. > > The audio callbacks are converted from hdmi_codec_ops signatures to > drm_bridge_funcs hdmi_audio callbacks: > - it66121_audio_hw_params -> it66121_hdmi_audio_prepare > - it66121_audio_startup -> it66121_hdmi_audio_startup > - it66121_audio_shutdown -> it66121_hdmi_audio_shutdown > - it66121_audio_mute -> it66121_hdmi_audio_mute_stream > > The it66121_audio_get_eld, it66121_audio_codec_ops, and > it66121_audio_codec_init functions are removed as the framework handles > these responsibilities. > > Suggested-by: Dmitry Baryshkov <[email protected]> > Signed-off-by: Sen Wang <[email protected]> > --- > Tested in beagleplay ensuring HDMI/ HDMI audio works correctly with > the new implementation, including proper ELD data and audio. > > Changelog: > > v1 -> v2: Retained no audio registered dev_info warning.
Reviewed-by: Dmitry Baryshkov <[email protected]> Also, if you have HW documentation for the bridge, would you consider implementing infoframes support? As another followup, would you consider switching the driver to drm_hdmi_acr_get_n_cts()? > drivers/gpu/drm/bridge/ite-it66121.c | 135 ++++++++++----------------- > 1 file changed, 48 insertions(+), 90 deletions(-) > -- With best wishes Dmitry
