Re: [PATCH v2 2/3] iio: dac: ad5706r: Add support for AD5706R DAC

[Andy Shevchenko <andriy.shevchenko@intel.com>]

On Thu, Mar 12, 2026 at 02:32:21AM +0000, Torreno, Alexis Czezar wrote:

...

> > > +	st->tx_buf = cpu_to_be32((((u32)reg) << 16) |
> > > +			((u32)val << (16 - (num_bytes * 8))));
> > 
> > What the heck is this?
> 
> I'll make this more readable.

The problems with the above are:
- macro style (it's not a macro, no need so many parentheses)
- unneeded castings (C promotes types to int)
- "clever" compression of the standard patterns (see below)

> but the format goal is:
> 
> if num_bytes = 1
>     tx_buf [31:0] = [reg 31:16] [val 15:8] [ XXXX]
> if num_bytes = 2
>     tx_buf [31:0] = [reg 31:16] [      val 15:0        ]

Assuming tx->buf is of u8:

	put_unaligned_be16(reg, &st->tx_buf[0]);
	if ( == 1)
		st->tx_buf[2] = val;
	else if ( == 2)
		put_unaligned_be16(val, &st->tx_buf[2]);
	else
		ERR!

(can also be switch-case, but it will be more verbose).

-- 
With Best Regards,
Andy Shevchenko