Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:6 as permitted sender) client-ip=2620:137:e000::3:6;
MIME-Version: 1.0
References: <20231024161931.78567-1-sebastian.reichel@collabora.com>
 <20231024161931.78567-2-sebastian.reichel@collabora.com> <CAHk-=whYDbZ29fx_xeSxtYSjtF8WJkaLjzyB8RN5_Rk9Sh-YyQ@mail.gmail.com>
 <CAHk-=wjO5ivM6k7iMiThO9JfxH0dhLe=mcC4TQwReU0nBCnWpg@mail.gmail.com>
 <your-ad-here.call-01698246313-ext-3263@work.hours> <CAHk-=wgs2DDdckcONG+YbB-GDH2QFCoZJ=Vm+YXxb1moZzuDgQ@mail.gmail.com>
 <055dd714f98f4e45b2add561b7663378@AcuMS.aculab.com>
In-Reply-To: <055dd714f98f4e45b2add561b7663378@AcuMS.aculab.com>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Thu, 26 Oct 2023 06:54:07 -1000
Message-ID: <CAHk-=wgBM7Swi7wmvi0M+sBFjzn-vEv1KzgM-hbJ4h3J3BR+UA@mail.gmail.com>
Subject: Re: [PATCH v4 1/3] math.h: add DIV_ROUND_UP_NO_OVERFLOW
To:     David Laight <David.Laight@aculab.com>
Cc:     Vasily Gorbik <gor@linux.ibm.com>,
        Sebastian Reichel <sebastian.reichel@collabora.com>,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        Michael Turquette <mturquette@baylibre.com>,
        Stephen Boyd <sboyd@kernel.org>,
        "linux-clk@vger.kernel.org" <linux-clk@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Niklas Schnelle <schnelle@linux.ibm.com>,
        "kernel@collabora.com" <kernel@collabora.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, 25 Oct 2023 at 22:57, David Laight <David.Laight@aculab.com> wrote:
>
> Doesn't that version end up calling inline functions?
> So won't be usable in static initialisers - the same as statement functions.

David, please either read what I write, or actually test things out,
instead of just adding noise.

It is not a problem to have a _Generic() - or a
__builtin_chooise_expr() - that has an inline function that doesn't
get used. The compiler will remove it.

The problem is literally that you can't use "__auto_type", because
*STATEMENT EXPRESSIONS* do not work outside of function definitions,
because gcc won't even *parse* them, and just says

  error: braced-group within expression allowed only inside a function

exactly like I explained in the comment of the patch that does the _Generic().

Which is why my solution ended up using those inline functions, and
splitting manually by type, instead of an expression statement and
__auto_type.

Whether you then split into different expressions with _Generic() or
by using __builtin_choose_expr() is irrelevant, although once you do
"split by type", _Generic() ends up being much more convenient because
that's what it's designed for (and you don't end up trying to figure
out the size with sizeof() and typeof tricks).

               Linus