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 50E113858282; Thu, 28 Mar 2024 18:26:31 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 50E113858282
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1711650391;
	bh=d5LgW6Iy58ru2NTV7PNY+BGQEEmauUJ4wl9rc/FDwD0=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=phYnTHVE3vlpHzxtid7pvZtIuxe2NHNRzXOrKTPWOBNhPcI1BjYXBqZoIacadeaeM
	 JoQ6QAHYtCwC7d5pkr25M5Nv6D/24eKFe1oOlbnj05Bo3GVA0FaGOueExNDFcRm6Cb
	 36rS0syaLtMwaZws4crQu4v0OPyTzxfU4lv8KJw4=
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/114523] bpf: ssa-phiopt optimization generates
 unverifiable code.
Date: Thu, 28 Mar 2024 18:26:29 +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: unknown
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: pinskia at gcc dot gnu.org
X-Bugzilla-Status: RESOLVED
X-Bugzilla-Resolution: INVALID
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-114523-4-g9Mr5Xxake@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-114523-4@http.gcc.gnu.org/bugzilla/>
References: <bug-114523-4@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=3D114523

--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #7)
> The problem is the verifier:
> M=C3=A4r 26 23:57:12 H systemd[1]: 17: (ac) w0 ^=3D w3                   =
  ;
> R0_w=3Dscalar()
> R3_w=3Dscalar(smin=3Dsmin32=3D0,smax=3Dumax=3Dsmax32=3Dumax32=3D1,var_off=
=3D(0x0; 0x1))
> M=C3=A4r 26 23:57:12 H systemd[1]: 18: (57) r0 &=3D 255                  =
  ;
> R0_w=3Dscalar(smin=3Dsmin32=3D0,smax=3Dumax=3Dsmax32=3Dumax32=3D255,var_o=
ff=3D(0x0; 0xff))
>=20
>=20
> it lost track of what the input of r0 was. It had the right result after =
the
> xor but then the bit_and didn't take into account the what was before it.=
 It
> just used 0xff.

Or rather the verifier is keeping track of rN and rN_w seperately but GCC
produces code that uses both and that definitely confuses the verifier.

M=C3=A4r 26 23:57:12 H systemd[1]: 16: (77) r3 >>=3D 63                    ;
R3_w=3Dscalar(smin=3Dsmin32=3D0,smax=3Dumax=3Dsmax32=3Dumax32=3D1,var_off=
=3D(0x0; 0x1))

wait why did the verifier use r3_w here rather than just r3? that seems like
the issue and would fix this issue easier ...=