Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <3fe52061d5d9e3701f598add5c64f8b3583c31df.camel@sipsolutions.net>
Subject: Re: [PATCH] bitfield: add FIELD_PREP_CONST()
From:   Johannes Berg <johannes@sipsolutions.net>
To:     Alexander Lobakin <alexandr.lobakin@intel.com>
Cc:     linux-wireless@vger.kernel.org
Date:   Wed, 18 Jan 2023 17:22:30 +0100
In-Reply-To: <92209c00-7fcf-1592-e97c-f49e7734fdfc@intel.com>
References: <20230118142652.53f20593504b.Iaeea0aee77a6493d70e573b4aa55c91c00e01e4b@changeid>
         <92209c00-7fcf-1592-e97c-f49e7734fdfc@intel.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.46.3 (3.46.3-1.fc37) 
MIME-Version: 1.0
Precedence: bulk

On Wed, 2023-01-18 at 17:10 +0100, Alexander Lobakin wrote:
>=20
> Have you tried combining it with FIELD_PREP() using
> __builtin_choose_expr() + __builtin_is_constexpr() (or
> __builtin_constant_p() depending on which will satisfy the compiler)?
> I'm not saying it's 100% possible, but worth trying.
>=20

I haven't tried it that way, but I tried rewriting FIELD_PREP() itself
to be constant-compatible, and as soon as the compiler saw
__builtin_constant_p() in the initializer it already complained that it
was non-constant...

I didn't think of __builtin_choose_expr, but it doesn't work either
because it only promises that the unused expression is not *evaluated*,
not that it's not "looked at", so it still complains both ways (in the
constant case that you can't have ({ }) braced groups, and in the non-
constant case that the _CONST version is bad...

johannes