On Sun, Mar 24, 2019 at 6:12 AM William Breathitt Gray
<vilhelm.gray@xxxxxxxxx> wrote:
>
> On Fri, Mar 22, 2019 at 09:12:02PM +0200, Andy Shevchenko wrote:
> > On Thu, Mar 14, 2019 at 09:29:32PM +0900, William Breathitt Gray wrote:
> > > Changes in v10:
> > > - Fix off-by-one error in bitmap initialization in the
> > > test_for_each_set_clump8 function
> > > - Fix typos in clump_exp array definition in test_bitmap.c ("0x28"
> > > should have been "0x38")
> > > - Utilize for_each_set_clump8 macro in intel_soc_dts_iosf.c
> >
> > One more, can you look at gen_74x164_set_multiple() ? It seems a candidate as
> > well, if I'm not mistaken.
>
> We can utilize the for_each_set_clump8 macro in the
> gen_74x164_set_multiple function, but I skipped over it earlier since I
> noticed it used the BITS_PER_BYTE define rather than a hardcoded 8. If
> it always loops 8 bits at a time, then we can use the
> for_each_set_clump8 macro;
Yes, see below.
> otherwise we would need the more generic
> for_each_set_clump macro to handle the non-8-bit looping cases.
> Will BITS_PER_BYTE always be defined as 8 bits?
It's not correct question, the right one "is the hardware always in
8-bit chunks". And datasheet is crystal clear about this.
This very old and famous IC has 8-bit from the 70-s (IIRC the epoch of
the design).
So, it is always in 8-bit chunks.
--
With Best Regards,
Andy Shevchenko
