Re: [PATCH 1/1] x86: Use resource_set_{range,size}() helpers

[Ilpo Järvinen]

On Fri, 18 Apr 2025, Ingo Molnar wrote:

> 
> * Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx> wrote:
> 
> > On Wed, 16 Apr 2025, Andy Shevchenko wrote:
> > 
> > > On Wed, Apr 16, 2025 at 01:13:18PM +0300, Ilpo Järvinen wrote:
> > > > Convert open coded resource size calculations to use
> > > > resource_set_{range,size}() helpers.
> > > > 
> > > > While at it, use SZ_* for size parameter which makes the intent of code
> > > > more obvious.
> > > 
> > > ...
> > > 
> > > > +	resource_set_range(res, base, 1ULL << (segn_busn_bits + 20));
> > > 
> > > Then probably
> > > 
> > > 	resource_set_range(res, base, BIT_ULL(segn_busn_bits) * SZ_1M);
> > > 
> > > to follow the same "While at it"?
> > 
> > I'll change that now since you brought it up. It did cross my mind to 
> > convert that to * SZ_1M but it seemed to go farther than I wanted with a 
> > simple conversion patch.
> > 
> > I've never liked the abuse of BIT*() for size related shifts though, 
> > I recall I saw somewhere a helper that was better named for size 
> > related operations but I just cannot recall its name and seem to not 
> > find that anymore :-(. But until I come across it once again, I guess 
> > I'll have to settle to BIT*().
> 
> BITS_TO_LONGS()?

Hi Ingo,

I'm not entiry sure if you're referring to my BIT*() matching unrelated
macros such as BITS_TO_LONGS() (I only meant BIT() and BIT_ULL() which I 
thought was clear from the context), or that BITS_TO_LONGS() would be the 
solution what I'm looking for.

In case you meant the latter, BITS_TO_LONGS() is not what I'm after. 
BIT(n) sets nth bit and what I'm looking for is converting n to power of 
two size. Obviously, both are mathematically doing 2^n (or 1 << n) but 
they feel conceptually very different things.

-- 
 i.