> > +#define REG_SIZE 0x8000
> > +
> > +#define I2C_REGMAP_VAL 8
> > +#define I2C_REGMAP_REG 16
> > +
> > +#define KSZ_REGMAP_COMMON(n, width)
> \
> > +{ \
> > + .name = n, \
> > + .max_register = REG_SIZE - (width), \
> > + .reg_bits = I2C_REGMAP_REG, \
> > + .val_bits = I2C_REGMAP_VAL * (width), \
> > + .reg_stride = (width), \
> > + .reg_format_endian = REGMAP_ENDIAN_BIG,
> \
> > + .val_format_endian = REGMAP_ENDIAN_BIG,
> \
> > +}
> This comes from my regmap RFC series, right ? Why do we need to redefine
> it here again ?
This is different from the SPI version. The variables pad_bits,
read_flag_mask, and write_flag_mask are not used. reg_bits is also different.
For KSZ8795 SPI the differences are SPI_ADDR_S, SPI_TURNAROUND_S, and
SPI_CMD_LEN.
