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 2EB8F3858C54; Fri, 15 Sep 2023 22:45:46 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 2EB8F3858C54
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1694817946;
	bh=3/nfTZ0D0EhIq3r+fG1zL1i8Q7an5MDU3lKWLVMXZts=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=v1agvxw0lkfofAvnnIlHx74r3rMEntceu089g9FBKeqoWbiBstLj299ScFqhQ035J
	 HVf0XRs/PDU1vdlHfmH2CS+5N1+pRvS770gSq/17cK1wApNaVIMGzKqK+5ifckmMJY
	 Jl/wwqpQCETyd+xs09QQ+E2g9WdUZwga/vrqQyOo=
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/110992] [13/14 Regression] missed VRP
 optimization due to transformation of `a & -zero_one_valued_p` into `a *
 zero_one_valued_p`
Date: Fri, 15 Sep 2023 22:45:45 +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: 14.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: pinskia at gcc dot gnu.org
X-Bugzilla-Status: ASSIGNED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: pinskia at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 13.3
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-110992-4-hb5H1KV78m@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-110992-4@http.gcc.gnu.org/bugzilla/>
References: <bug-110992-4@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
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D110992

--- Comment #10 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
This is what I came up with in the end:
/* (a * b@[0,1]) =3D=3D CST
     ->=20
       CST =3D=3D 0 ? (a =3D=3D CST | b =3D=3D 0) : (a =3D=3D CST & b !=3D =
0)
   (a * b@[0,1]) !=3D CST
     ->=20
       CST !=3D 0 ? (a !=3D CST | b =3D=3D 0) : (a !=3D CST & b !=3D 0)  */
(for cmp (ne eq)
 (simplify
  (cmp (mult:cs @0 zero_one_valued_p@1) INTEGER_CST@2)
  (if ((cmp =3D=3D EQ_EXPR) ^ (wi::to_wide (@2) !=3D 0))
   (bit_ior
    (cmp @0 @2)
    (convert (bit_xor @1 { build_one_cst (TREE_TYPE (@1)); })))
   (bit_and (cmp @0 @2) (convert @1)))))


This allows more than just CST =3D=3D 0 even. I will write up some testcase=
s now
...=