From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
 id C41B53858D1E; Wed,  6 Apr 2022 07:10:50 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org C41B53858D1E
From: "peterz at infradead dot org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug ipa/105160] [12 regression] ipa modref marks functions with asm
 volatile as const or pure
Date: Wed, 06 Apr 2022 07:10:50 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: ipa
X-Bugzilla-Version: 12.0
X-Bugzilla-Keywords: wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: peterz at infradead dot org
X-Bugzilla-Status: ASSIGNED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P1
X-Bugzilla-Assigned-To: hubicka at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 12.0
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-105160-4-Op6nKKGk5l@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-105160-4@http.gcc.gnu.org/bugzilla/>
References: <bug-105160-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-BeenThere: gcc-bugs@gcc.gnu.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Gcc-bugs mailing list <gcc-bugs.gcc.gnu.org>
List-Unsubscribe: <https://gcc.gnu.org/mailman/options/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=unsubscribe>
List-Archive: <https://gcc.gnu.org/pipermail/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-request@gcc.gnu.org?subject=help>
List-Subscribe: <https://gcc.gnu.org/mailman/listinfo/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=subscribe>
X-List-Received-Date: Wed, 06 Apr 2022 07:10:50 -0000

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D105160

--- Comment #7 from peterz at infradead dot org ---
(In reply to peterz from comment #6)
> Happy accident; I've been wanting to allow doing something like:
>=20
> static __always_inline __pure bool arch_static_branch(struct static_key *
> const key)
> {
>     asm_volatile_goto("1:"
>                       ".byte " __stringify(BYTES_NOP5) "\n\t"
>                       ".pushsection __jump_table, \"aw\" \n\t"
>                       _ASM_ALIGN "\n\t"
>                       ".long 1b - ., %l[l_yes] - . \n\t"
>                       _ASM_PTR "%c0 - . \n\t"
>                       ".popsecion \n\t"
>                       : : "i" (key) : : l_yes);
>     return false;
> l_yes:
>     return true;
> }
>=20
>=20
> Now, since asm-goto is implicitly volatile, and asm volatile (per this
> thread) destroys __pure due to side-effects, this doesn't actually work. =
But
> I'd like to still have the explicit pure attribute override this. If the
> programmer gets it wrong like this, he/she gets to keep the pieces.
>=20
> Specifically in this case, he result of the function really only depends =
on
> the @key argument. Any call to this will have the exact same outcome and
> merging multiple instances is *good*.
>=20
> Also see this thread on linux-toolchains:
>=20
>=20=20
> https://lore.kernel.org/all/YG80wg%2F2iZjXfCDJ@hirez.programming.kicks-as=
s.
> net/

Possible test case here:

  https://lore.kernel.org/all/YIb0vYQMEs9HRjPl@hirez.programming.kicks-ass.=
net/=