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 C6921388E807; Thu, 17 Dec 2020 13:41:18 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org C6921388E807
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/94779] Bad optimization of simple switch
Date: Thu, 17 Dec 2020 13:41:18 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: tree-optimization
X-Bugzilla-Version: 10.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: jakub at gcc dot gnu.org
X-Bugzilla-Status: ASSIGNED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: marxin at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-94779-4-NAUoKPgdIq@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-94779-4@http.gcc.gnu.org/bugzilla/>
References: <bug-94779-4@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: Thu, 17 Dec 2020 13:41:18 -0000

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D94779

--- Comment #15 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to Martin Li=C5=A1ka from comment #14)
> Ok, I've just taken look at what EVRP pass does before SWITCHCONV pass is
> called.
> I see that EVRP can properly prune dead cases of a switch, but it's not
> perfect:
>=20
> int f1(unsigned x)
> {
>     if (x =3D=3D 2 || x =3D=3D 3 || x >=3D 5)
>       __builtin_unreachable ();
>     switch (x)
>     {
>       case 0: return 1;
>       case 1: return 2;
>       case 2: return 3;
>       case 3 ... 8: return 4;
>     }
> }

The old VRP/EVRP only tracks simple ranges and anti-ranges, so can't deal w=
ith
what you have above, the new ranger code can deal with multiple subranges, =
but
the question is if all the interfaces deal with those.

> This seems to me like a strange EVRP transformation :/

Why?  And, the user could have written it that way too.=