On Sun 17 Mar 2024 at 18:19, Jan Dakinevich <jan.dakinevich@xxxxxxxxxxxxxxxxx> wrote: > On 3/15/24 16:33, Mark Brown wrote: >> On Fri, Mar 15, 2024 at 02:21:45AM +0300, Jan Dakinevich wrote: >> >>> static const struct regmap_config g12a_toacodec_regmap_cfg = { >>> - .reg_bits = 32, >>> - .val_bits = 32, >>> - .reg_stride = 4, >>> + .reg_bits = 32, >>> + .val_bits = 32, >>> + .reg_stride = 4, >>> + .max_register = TOACODEC_CTRL0, >>> + .max_register_is_0 = true, >> >> If the maximum register is 0 how does the regmap have a stride? > > reg_stride inherited from existing code. Apparently, it was meaningless > even before my modifications (the hardware has single register > regardless of max_register declaration) and it should be dropped. But, > is it okay to remove it in the same commit? Yes it has a single register, for now. Still the stride is 4. And assuming the mmio region passed from DT is correct, I'm not sure the hunk is useful at all. -- Jerome