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 5A4AF3858430; Tue, 16 Nov 2021 12:17:35 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 5A4AF3858430
From: "aldyh at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/103192] [12 Regression] ICE on libgomp
 target-in-reduction-2.{C,c}
Date: Tue, 16 Nov 2021 12:17:34 +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: 12.0
X-Bugzilla-Keywords: ice-on-valid-code, wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: aldyh at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P1
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 12.0
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-103192-4-Dlo1DOrexc@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-103192-4@http.gcc.gnu.org/bugzilla/>
References: <bug-103192-4@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: Tue, 16 Nov 2021 12:17:35 -0000

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D103192

--- Comment #18 from Aldy Hernandez <aldyh at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #17)
> iftmp.2373_1515 is defined earlier as:
>   iftmp.2373_1515 =3D code_1387(D) !=3D 181 ? ctx_1386 : outer_ctx_1389;
> so the transformation by dom3? from
>   if (outer_ctx_1389 !=3D 0B)
> to
>   if (iftmp.2373_1515 !=3D 0B)
> is sound (even though it undoes some previously done optimization), but
> perhaps
> we need to reset_flow_sensitive_info on it somewhere along the way?  Thou=
gh,
> the definition is always not in code guarded by outer_ctx_1389 !=3D 0B ch=
eck,
> so I wonder why it would be marked as non-NULL, only the uses or some of =
the
> uses of the SSA_NAME can be non-NULL.

It looks like DOM uses propagate_value() which makes the substitution outri=
ght,
whereas the copy-prop pass is more careful with global ranges:

              /* Points-to information is cfg insensitive,
                 but [E]VRP might record context sensitive alignment
                 info, non-nullness, etc.  So reset context sensitive
                 info if the two SSA_NAMEs aren't defined in the same
                 basic block.  */
              if (var_bb !=3D copy_of_bb)
                reset_flow_sensitive_info (copy_of[i].value);

Perhaps we need something similar in DOM's cprop_operand?=