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 07E233858C39; Sun,  9 Jul 2023 22:34:51 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 07E233858C39
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1688942091;
	bh=l1EYUs6bI+D31PCYwZ4uHppUnVGV3pRTr4gMj2fjzqU=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=ITFwd2zN194VPtgyTnSv+hq5xMRFvkdNj/520ZD72xjI78UTwZNOpBO8dOltEyoBD
	 Geh6fdWkShpcUJ3s6xh2mJ00ou9pdVgeHhNLxLGVrvLmMR8Kasjqvkk7uvzel+7U1C
	 oAiYNDZqM3eTLVomOa9VVm+oIt2mDJoyryd4NerI=
From: "roger at nextmovesoftware dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/110598] [14 Regression] wrong code on llvm-14.0.6 due to
 memcmp being miscompiled
Date: Sun, 09 Jul 2023 22:34:50 +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: 14.0
X-Bugzilla-Keywords: wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: roger at nextmovesoftware dot com
X-Bugzilla-Status: ASSIGNED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: roger at nextmovesoftware dot com
X-Bugzilla-Target-Milestone: 14.0
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: assigned_to bug_status
Message-ID: <bug-110598-4-YWTBq3GHql@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-110598-4@http.gcc.gnu.org/bugzilla/>
References: <bug-110598-4@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=3D110598

Roger Sayle <roger at nextmovesoftware dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |roger at nextmoveso=
ftware dot com
             Status|NEW                         |ASSIGNED

--- Comment #4 from Roger Sayle <roger at nextmovesoftware dot com> ---
Mine.  The new peephole2 is misbehaving with the (odd) RTL sequence:
(insn 62 61 63 2 (set (reg:DI 1 dx [111])
        (const_int 0 [0])) "pr110598.c":10:10 90 {*movdi_internal}

(insn 63 62 64 2 (parallel [
            (set (reg:DI 1 dx [110])
                (xor:DI (reg:DI 1 dx [111])
                    (reg:DI 1 dx)))
            (clobber (reg:CC 17 flags))
        ]) "pr110598.c":10:10 626 {*xordi_1}
     (expr_list:REG_UNUSED (reg:CC 17 flags)
        (nil)))

Doh!  I need to confirm/check that rega !=3D regb.  Interesting that DF doe=
sn't
consider insn 62 as REG_UNUSED (i.e. dead), a classic false dependency.=