Re: [PATCH v2 05/17] staging:iio:adc:ad7280a: Use bitfield ops to managed fields in transfers.

[Jonathan Cameron <jic23@kernel.org>]

On Tue, 14 Dec 2021 14:12:24 -0300
Marcelo Schmitt <marcelo.schmitt1@gmail.com> wrote:

> Minor bit inline.
> 
> On 12/05, Jonathan Cameron wrote:
> > From: Jonathan Cameron <Jonathan.Cameron@huawei.com>
> > 
> > The write and two types of read transfer are sufficiently complex that
> > they benefit from the clarity of using FIELD_PREP() and FIELD_GET()
> > 
> > Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>  
> Reviewed-by: Marcelo Schmitt <marcelo.schmitt1@gmail.com>

Those hunks you highlighted in patch 2 now in this patch and I've added a note
to the description to say a similar arguement applies to the event_handler()

Reply to your comment inline. I didn't understand this one.
> 
> > ---
> >  drivers/staging/iio/adc/ad7280a.c | 46 ++++++++++++++++++++++++-------
> >  1 file changed, 36 insertions(+), 10 deletions(-)
> > 
> > diff --git a/drivers/staging/iio/adc/ad7280a.c b/drivers/staging/iio/adc/ad7280a.c
> > index 1f7ea5fb1e20..158a792c0bf8 100644
> > --- a/drivers/staging/iio/adc/ad7280a.c
> > +++ b/drivers/staging/iio/adc/ad7280a.c
> > @@ -95,6 +95,23 @@
> >  #define   AD7280A_READ_ADDR_MSK				GENMASK(7, 2)
> >  #define AD7280A_CNVST_CTRL_REG			0x1D /* D7 to D0, Read/write */
> >  
> > +/* Transfer fields */
> > +#define AD7280A_TRANS_WRITE_DEVADDR_MSK		GENMASK(31, 27)
> > +#define AD7280A_TRANS_WRITE_ADDR_MSK		GENMASK(26, 21)
> > +#define AD7280A_TRANS_WRITE_VAL_MSK		GENMASK(20, 13)
> > +#define AD7280A_TRANS_WRITE_ALL_MSK		BIT(12)
> > +#define AD7280A_TRANS_WRITE_CRC_MSK		GENMASK(10, 3)
> > +#define AD7280A_TRANS_WRITE_RES_PATTERN		0x2
> > +
> > +/* Layouts differ for channel vs other registers */
> > +#define AD7280A_TRANS_READ_DEVADDR_MSK		GENMASK(31, 27)
> > +#define AD7280A_TRANS_READ_CONV_CHANADDR_MSK	GENMASK(26, 23)
> > +#define AD7280A_TRANS_READ_CONV_DATA_MSK	GENMASK(22, 11)
> > +#define AD7280A_TRANS_READ_REG_REGADDR_MSK	GENMASK(26, 21)
> > +#define AD7280A_TRANS_READ_REG_DATA_MSK		GENMASK(20, 13)
> > +#define AD7280A_TRANS_READ_WRITE_ACK_MSK	BIT(10)
> > +#define AD7280A_TRANS_READ_CRC_MSK		GENMASK(9, 2)
> > +
> >  /* Magic value used to indicate this special case */
> >  #define AD7280A_ALL_CELLS				(0xAD << 16)
> >  
> > @@ -216,10 +233,16 @@ static int __ad7280_read32(struct ad7280_state *st, unsigned int *val)
> >  static int ad7280_write(struct ad7280_state *st, unsigned int devaddr,
> >  			unsigned int addr, bool all, unsigned int val)
> >  {
> > -	unsigned int reg = devaddr << 27 | addr << 21 |
> > -			(val & 0xFF) << 13 | all << 12;
> > +	unsigned int reg = FIELD_PREP(AD7280A_TRANS_WRITE_DEVADDR_MSK, devaddr) |
> > +		FIELD_PREP(AD7280A_TRANS_WRITE_ADDR_MSK, addr) |
> > +		FIELD_PREP(AD7280A_TRANS_WRITE_VAL_MSK, val) |
> > +		FIELD_PREP(AD7280A_TRANS_WRITE_ALL_MSK, all);  
> Does reg get initialized to 0? If not, we should take care of the reserved bit D11
> 
You have lost me here.  We are writing the whole register and this first
assignment of reg uses = not |= so any bits not set explicitly will be zero.
We could do a FIELD_PREP(BIT(11), 0) with an appropriate name for the BIT(11) define
but that would be a bit odd.

> > +
> > +	reg |= FIELD_PREP(AD7280A_TRANS_WRITE_CRC_MSK,
> > +			ad7280_calc_crc8(st->crc_tab, reg >> 11));
> > +	/* Reserved b010 pattern not included crc calc */
> > +	reg |= AD7280A_TRANS_WRITE_RES_PATTERN;
> >  
> > -	reg |= ad7280_calc_crc8(st->crc_tab, reg >> 11) << 3 | 0x2;
> >  	st->tx = cpu_to_be32(reg);
> >  
> >  	return spi_write(st->spi, &st->tx, sizeof(st->tx));