On 2023-11-16 4:10 PM, Jaroslav Kysela wrote:
On 16. 11. 23 12:22, Cezary Rojewski wrote:
Introduce a set of functions that ultimately facilite SDxFMT-related calculations in atomic manner:
First, introduce snd_pcm_subformat_width() and snd_pcm_hw_params_bps() helpers that separate the base functionality from the HDAudio-specific one.
I think, snd_pcm_subformat_width() should be implemented only as static helper for now. There are no users and only snd_pcm_hw_params_bps() makes sense for callers (format & subformat must be passed together).
Ack. Inlining into its only user.
Also note that _bps is mostly used for bits-per-second in the computer world, so maybe _bpsam may be used here to pay more attention by readers, but it's just my optional proposal. Maybe my head has some "fixed wiring" for this :)
The current shortcut is taken from existing intel/ code which utilizes 'bps' but I agree that it can be misleading.
What I did to solve the problem: opened both, HDAudio (bare) and HDAudio+ADSP specs and in both the relevant fields (the fields that relate to bits-per-sample) are called "BITS". Also searched google for the 'bits per sample' phrase and one of the first thing that came up was audacity doc which used the exact same phrase (bits).
All occurrences of 'bps' will be replaced by 'bits'.
Kind regards, Czarek