Re: [PATCH v2 04/11] iio: adc: ad_sigma_delta: use BITS_TO_BYTES() macro

[Jonathan Cameron <jic23@kernel.org>]

On Tue, 1 Jul 2025 16:47:02 +0300
Andy Shevchenko <andriy.shevchenko@intel.com> wrote:

> On Mon, Jun 30, 2025 at 08:33:59AM -0500, David Lechner wrote:
> > On 6/30/25 3:59 AM, Andy Shevchenko wrote:  
> > > On Sat, Jun 28, 2025 at 03:56:43PM +0100, Jonathan Cameron wrote:  
> > >> On Fri, 27 Jun 2025 18:40:00 -0500
> > >> David Lechner <dlechner@baylibre.com> wrote:  
> 
> ...
> 
> > >>> -	samples_buf_size = ALIGN(slot * indio_dev->channels[0].scan_type.storagebits / 8, 8);
> > >>> +	samples_buf_size = ALIGN(slot * BITS_TO_BYTES(scan_type->storagebits), 8);  
> > >>
> > >> Ah. You do it here. Fair enough and no problem wrt to patch 1 then.  
> > > 
> > > Hmm... Should the second 8 be something like sizeof(unsigned long lone) for
> > > semantic distinguishing with 8-bit bytes?  
> > 
> > Yeah, I considered to use sizeof(s64) to match the next line, but it
> > it seems like a separate change, so in the end I decided against doing
> > it in this patch and it seems too small of a thing for a separate patch.  
> 
> The problem in not the size of the change, the problem is that semantically
> those 8:s are _different_ and code readability will be much better if we make
> them so explicitly.
Agreed. It's not so bad once we are down to just one magic 8 (ball :) but
definitely makes sense to give them both explicit meaning.

A tiny follow up patch, or rolling it in here with a comment in the patch
description would both be fine.

Jonathan

>