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 A626D385702B; Wed,  5 Jul 2023 13:15:26 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org A626D385702B
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1688562926;
	bh=rckD1aNVdRmoUCQos625fMZbQ+XkvtN0AaAclhPX84c=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=l/7ub7Qm/2q9KR2P50/b6ysngw9lsV2lFvDGXNyssAQqp5p1NzmvDMOS0veyvzwcl
	 FlJgvTvuFeV4KzmQeV3KxQw5C3b06ijQZYhJ9p5NLR+tEefKCDXxcuqH6Nw58UqyY8
	 vF2hpAi99zMdxy70phYBgPE0uaCxOqxsJNOVDtbQ=
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/110557] [13/14 Regression] Wrong code for
 x86_64-linux-gnu with -O3 -mavx2: vectorized loop mishandles signed
 bit-fields
Date: Wed, 05 Jul 2023 13:15:26 +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: wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: pinskia at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 13.3
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: bug_status keywords cf_reconfirmed_on
 everconfirmed component
Message-ID: <bug-110557-4-D7nhb5x5M0@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-110557-4@http.gcc.gnu.org/bugzilla/>
References: <bug-110557-4@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=3D110557

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
           Keywords|needs-bisection             |
   Last reconfirmed|                            |2023-07-05
     Ever confirmed|0                           |1
          Component|target                      |tree-optimization

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
ifcvt produces correctly the sign extend:

  <unnamed-signed:56> _ifc__18;


  _3 =3D a_14(D) + _2;
  _ifc__22 =3D _3->D.3439;
  _ifc__18 =3D BIT_FIELD_REF <_ifc__22, 56, 8>;
  _5 =3D (long long int) _ifc__18;
  _6 =3D _5 * -4;
  _7 =3D (long unsigned int) _6;
  _15 =3D MAX_EXPR <_7, size_19>;

But then the vectorizer misses that:
```
  _ifc__18 =3D BIT_FIELD_REF <_ifc__22, 56, 8>;
  vect_patt_13.20_76 =3D vect__ifc__22.19_74 & { 18446744073709551360,
18446744073709551360, 18446744073709551360, 18446744073709551360 };
  vect_patt_10.21_77 =3D vect_patt_13.20_76 >> 8;
  vect_patt_9.22_78 =3D VIEW_CONVERT_EXPR<vector(4) long long
int>(vect_patt_10.21_77);
  _5 =3D (long long int) _ifc__18;
  vect__6.23_80 =3D vect_patt_9.22_78 * { -4, -4, -4, -4 };
  _6 =3D _5 * -4;
```

vect_patt_9.22_78 should have been sign extended ...



With the C front-end, we get:
```
  <unnamed-signed:56> _24;
  <unnamed-signed:56> _23;

  _24 =3D _25->y;
  _23 =3D _24 * -4;
  _22 =3D (long unsigned int) _23;
```

Which the vectorizer does not understand could be prompted to 64bit and then
truncated so it does not vectorize it.


r13-3219-g25413fdb2ac24933214123e24ba16502 .=