Re: [PATCH 2/5] bits_per_long.h: introduce SMALL_CONST() macro

From: Yury Norov <yury.norov@gmail.com>
To: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: linux-m68k <linux-m68k@lists.linux-m68k.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Linux-SH <linux-sh@vger.kernel.org>,
	Linux-Arch <linux-arch@vger.kernel.org>,
	Geert Uytterhoeven <geert@linux-m68k.org>,
	Yoshinori Sato <ysato@users.sourceforge.jp>,
	Rich Felker <dalias@libc.org>, Arnd Bergmann <arnd@arndb.de>,
	Dennis Zhou <dennis@kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Wolfram Sang <wsa+renesas@sang-engineering.com>,
	David Sterba <dsterba@suse.com>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	Stefano Brivio <sbrivio@redhat.com>,
	"Ma, Jianpeng" <jianpeng.ma@intel.com>,
	Wei Yang <richard.weiyang@linux.alibaba.com>,
	Josh Poimboeuf <jpoimboe@redhat.com>,
	John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Subject: Re: [PATCH 2/5] bits_per_long.h: introduce SMALL_CONST() macro
Date: Fri, 29 Jan 2021 13:28:20 -0800	[thread overview]
Message-ID: <CAAH8bW_Gt+0bC=S2HtR_B3cH-KGJQQaUQ0z0xn=aZCtBzy43yQ@mail.gmail.com> (raw)
In-Reply-To: <CAHp75VcSc=myrcvyBOkaUDguR6aPjJAFFXi2iSvmU21+1664Hw@mail.gmail.com>

On Fri, Jan 29, 2021 at 1:11 PM Andy Shevchenko
<andy.shevchenko@gmail.com> wrote:
>
> On Fri, Jan 29, 2021 at 10:49 PM Yury Norov <yury.norov@gmail.com> wrote:
> >
> > Many algorithms become simplier if they are passed with relatively small
>
> simpler
>
> > input values. One example is bitmap operations when the whole bitmap fits
> > into one word. To implement such simplifications, linux/bitmap.h declares
> > small_const_nbits() macro.
> >
> > Other subsystems may also benefit from optimizations of this sort, like
> > find_bit API in the following patches. So it looks helpful to generalize
> > the macro and extend it's visibility.
>
> > It should probably go to linux/kernel.h, but doing that creates circular
> > dependencies. So put it in asm-generic/bitsperlong.h.
>
> No, no, please leave kernel.h alone. It's already quite a mess.
>
> And this shouldn't be in the commit message either.
>
> ...
>
> > -       if (small_const_nbits(nbits))
> > +       if (SMALL_CONST(nbits - 1))
>
> Not sure if we need to rename it.

Lower case for generic macro kind of breaking the rules.

Behaviour has changed too. Before it was:
0 < VAL <= 32
Now it's
0 <= VAL < 32
Which is better for generic macro.

So changing the name looks reasonable to me.

> ...
>
> > --- a/include/linux/bits.h
> > +++ b/include/linux/bits.h
> > @@ -37,7 +37,7 @@
> >  #define GENMASK(h, l) \
> >         (GENMASK_INPUT_CHECK(h, l) + __GENMASK(h, l))
> >
> > -#define BITS_FIRST(nr)         GENMASK(nr), 0)
> > +#define BITS_FIRST(nr)         GENMASK((nr), 0)
>
> How come?!

git send-email wrong_dir/000*
Will resubmit today later.

> ...
>
> > diff --git a/tools/include/asm-generic/bitsperlong.h b/tools/include/asm-generic/bitsperlong.h
> > index 8f2283052333..432d272baf27 100644
> > --- a/tools/include/asm-generic/bitsperlong.h
> > +++ b/tools/include/asm-generic/bitsperlong.h
>
> I think a tools update would be better to have in a separate patch.
>
> --
> With Best Regards,
> Andy Shevchenko