From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 48) id 35D0D3861810; Tue, 30 Mar 2021 23:20:12 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 35D0D3861810 From: "bergner at gcc dot gnu.org" To: gcc-bugs@gcc.gnu.org Subject: [Bug target/99842] MMA test case ICEs using -O3 Date: Tue, 30 Mar 2021 23:20:12 +0000 X-Bugzilla-Reason: CC X-Bugzilla-Type: changed X-Bugzilla-Watch-Reason: None X-Bugzilla-Product: gcc X-Bugzilla-Component: target X-Bugzilla-Version: 11.0 X-Bugzilla-Keywords: X-Bugzilla-Severity: normal X-Bugzilla-Who: bergner at gcc dot gnu.org X-Bugzilla-Status: ASSIGNED X-Bugzilla-Resolution: X-Bugzilla-Priority: P3 X-Bugzilla-Assigned-To: bergner at gcc dot gnu.org X-Bugzilla-Target-Milestone: 11.0 X-Bugzilla-Flags: X-Bugzilla-Changed-Fields: cf_gcctarget everconfirmed cf_known_to_fail assigned_to cf_reconfirmed_on bug_status target_milestone Message-ID: In-Reply-To: References: 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 List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 30 Mar 2021 23:20:12 -0000 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D99842 Peter Bergner changed: What |Removed |Added ---------------------------------------------------------------------------- Target| |powerpc64le-linux Ever confirmed|0 |1 Known to fail| |11.0 Assignee|unassigned at gcc dot gnu.org |bergner at gcc dot = gnu.org Last reconfirmed| |2021-03-30 Status|UNCONFIRMED |ASSIGNED Target Milestone|--- |11.0 --- Comment #1 from Peter Bergner --- Mine. The problem is that the mma_assemble_input_operand predicate is rejecting valid reg+reg indexed addresses in the MEMs in the above rtl. The predicate is calling quad_address_p() which accepts reg+offset addresses (w= ith constrained offset values), but doesn't allow reg+reg addresses, which are valid. Replacing the MEM_P() && quad_address_p() test with a call to memory_operan= d() fixes the ICE, since it calls down to rs6000_legitimate_address_p(), which calls quad_address_p() to validate reg+offset addresses, but also allows reg+reg addresses. I'll submit a patch with that fix.=