On Wed, 20 May 2026, Javier Martinez Canillas <[email protected]> wrote: > Jani Nikula <[email protected]> writes: > > Hello Jani, > >> On Tue, 19 May 2026, Javier Martinez Canillas <[email protected]> wrote: > > [...] > >>> >>> +/* HDMI spec TMDS character rate limits (in Hz) */ >>> +#define DRM_HDMI_TMDS_CHAR_RATE_MIN 25000000 >>> +#define DRM_HDMI_TMDS_CHAR_RATE_MAX_1_0 165000000 >>> +#define DRM_HDMI_TMDS_CHAR_RATE_MAX_1_3 340000000 >>> +#define DRM_HDMI_TMDS_CHAR_RATE_MAX_2_0 600000000 >> >> Usually everything in DRM is in kHz, and Hz is the exception. >> > > That is correct but in this case these constants are to be used with the > HDMI helpers. Both struct drm_connector_hdmi_state.tmds_char_rate and > the struct drm_bridge_funcs.hdmi_tmds_char_rate_valid() callback expect > the TMDS char rate to be defined in Hz.
I guess I'm surprised those things were defined in terms of Hz then. > If we define these in kHz, it means that drivers will have to * 1000 at > every call site. Yeah, not ideal. >> I'm also not sure the 1_0, 1_3, and 2_0 really help anyone reading the >> code. I won't remember what they mean in Hz or kHz, and I'll have to >> look them up every single time. >> > > I discussed this with Maxime before posting the patches since I wondered > the same. He suggested that the max TMDS character rate was linked to the > HDMI spec versions and that it would be more readable to name it using > the spec version rather than the resolution. Frankly I've never thought of the limits in terms of the spec versions, but maybe it's just me. *shrug* BR, Jani. > > The other naming I suggested was: > > /* HDMI spec TMDS character rate limits (in Hz) */ > #define DRM_HDMI_TMDS_CHAR_RATE_MIN 25000000 > #define DRM_HDMI_TMDS_CHAR_RATE_74_25MHZ 74250000 > #define DRM_HDMI_TMDS_CHAR_RATE_148_5MHZ 148500000 > #define DRM_HDMI_TMDS_CHAR_RATE_297MHZ 297000000 > #define DRM_HDMI_TMDS_CHAR_RATE_MAX_1_4 340000000 > >> BR, >> Jani. >> -- Jani Nikula, Intel
