Re: [PATCH v2] minmax: clamp more efficiently by avoiding extra comparison

[Andy Shevchenko]

On Sat, Sep 24, 2022 at 12:37:26PM +0200, Jason A. Donenfeld wrote:
> On Fri, Sep 23, 2022 at 03:54:12PM -0700, Andrew Morton wrote:
> > On Fri, 23 Sep 2022 17:40:01 +0200 "Jason A. Donenfeld" <Jason@xxxxxxxxx> wrote:

...

> Worth noting, by the way, is that the input validation check already
> caught a bug when 0day test bot choked:
> 
> https://lore.kernel.org/linux-hwmon/20220924101151.4168414-1-Jason@xxxxxxxxx/

Hooray, it was a good idea! :-)

> So, options:
> 1) Keep this patch as-is, because it is useful on modern compilers.
> 2) Add an ifdef on compiler version, so we generate the best code in
>    each case.
> 3) Go back to testing twice, but keep the checker macro because it's
>    apparently useful.
> 4) Do nothing and discard this series.
> 
> Any of those are okay with me. Opinions?

I tend to case 3) (I believe you typo'ed double 2) cases) and apply the rest
as a separate change with all downsides explained (kinda 1) approach).

-- 
With Best Regards,
Andy Shevchenko