Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20200110165636.28035-1-will@kernel.org> <20200110165636.28035-7-will@kernel.org>
 <CAHk-=wia5ppBsfHLMx648utCjO01JAZiME0K0eSHmhWuRyL+6w@mail.gmail.com> <20200113145954.GB4458@willie-the-truck>
In-Reply-To: <20200113145954.GB4458@willie-the-truck>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Mon, 13 Jan 2020 11:31:56 -0800
Message-ID: <CAHk-=wirAWFOrfD4us1FepP0vWkZMpnqXusJyKHCqwBVsR43CA@mail.gmail.com>
Subject: Re: [RFC PATCH 6/8] READ_ONCE: Drop pointer qualifiers when reading
 from scalar types
To:     Will Deacon <will@kernel.org>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-arch <linux-arch@vger.kernel.org>,
        Android Kernel Team <kernel-team@android.com>,
        Michael Ellerman <mpe@ellerman.id.au>,
        Peter Zijlstra <peterz@infradead.org>,
        Segher Boessenkool <segher@kernel.crashing.org>,
        Christian Borntraeger <borntraeger@de.ibm.com>,
        Luc Van Oostenryck <luc.vanoostenryck@gmail.com>,
        Arnd Bergmann <arnd@arndb.de>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Jan 13, 2020 at 7:00 AM Will Deacon <will@kernel.org> wrote:
>
> I can't disagree with that, but the only option we've come up with so far
> that solves this in the READ_ONCE() macro itself is the thing from PeterZ:
>
> // Insert big fat comment here
> #define unqual_typeof(x)    typeof(({_Atomic typeof(x) ___x __maybe_unused; ___x; }))

I'm with Luc on this - that not only looks gcc-specific, it looks
fragile too, in that it's not obvious that "_Atomic typeof(x)" really
is guaranteed to do what we want.

> So I suppose my question is: how ill does this code really make you feel?

I wish the code was more obvious.

One way to do that might be to do your approach, but just write it as
a series of macros that makes it a bit more understandable what it
does.

Maybe it's just because of a "pee in the snow" effect, but I think
this is easier to explain:

  #define __pick_scalar_type(x,type,otherwise)          \
        __builtin_choose_expr(__same_type(x,type), (type)0, otherwise)

  #define __pick_integer_type(x, type, otherwise)       \
        __pick_scalar_type(x, unsigned type,            \
          __pick_scalar_type(x, signed type, otherwise))

  #define __unqual_scalar_typeof(x) typeof(             \
        __pick_integer_type(x, char,                    \
          __pick_integer_type(x, short,                 \
            __pick_integer_type(x, int,                 \
              __pick_integer_type(x, long,              \
                __pick_integer_type(x, long long, x))))))

just because you there's less repeated noise, and the repetition there
is is simpler.

So still "Eww", but maybe not quite _as_ "Eww".

             Linus