On Tue, May 19, 2026 at 07:32:33PM -0300, Val Packett wrote: > On 5/19/26 1:06 PM, Mark Brown wrote:
> > In spite of the changelog this doesn't actually seem to use slot_width? > > It should. > Right, I hadn't exactly figured out where that slots in (ha).. > The datasheet seems to say that I2SBCK is slot width, but there are only 3 > options for it (32/48/64) and I currently select the one that matches > 2*physical_width. > Grepping the tree I'm seeing that slot_width is basically the same thing as > physical width. Should slot_width from set_tdm_slot take precedence over the > params_physical_width() from hw_params? Should I just have both callbacks > set the variable and whichever was called last "wins"? If TDM is configured then hw_params should not cause the number of bits on the bus to vary at all, the clocking of the bus is fixed by the TDM configuration. > > Note also that we should disable TDM if slots == 0. > This is pretty much already the case.. but I guess I should also reset the > masks specifically to the i2s default (2nd channel to R). Are masks also > zero when slots == 0? They should be.
signature.asc
Description: PGP signature
