Hi Simon, First RK3588 report for this patch, with register-level before and after, both HDMI controllers driven at once:
Tested-by: Igor Paunovic <[email protected]> # RK3588 Orange Pi 5 Plus, both HDMI controllers Base: drm-misc-next 47b823940e38 with the HDMI 2.0 scrambling series v11 on top. The patch needs one trivial context resolution there (v11 still carries the open-coded tables next to its new scrambling defines); the result is your intent verbatim, -206/+3. Before (table code): at TMDS 297 MHz / 48 kHz the controller ran with AUDPKT_ACR_CONTROL1 = 0, CTS_OVR_EN clear - the auto-CTS fallback your commit message describes. Audio still played here because both my sinks are permissive. After (this patch), read back from AUDPKT_ACR_CONTROL0/1 while audio was playing: TMDS sink N CTS spec CTS 297 MHz Sony (300 MHz max) 5120 247500 247500 594 MHz Samsung (HF-VSDB 2.0) 6144 594000 594000 Both values land exactly on the HDMI spec numbers, CTS_OVR_EN set, and audio is audible on both sinks simultaneously. The 594 MHz row runs on a scrambled HDMI 2.0 link (4K60 RGB), which I believe no one had exercised with this patch yet. Note N at 594 MHz changes from the old table's 10240 to the helper's spec value 6144 - no audible difference on my sinks. Igor
