From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-297233-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 15925 invoked by alias); 19 Jul 2011 12:08:53 -0000
Received: (qmail 15769 invoked by uid 22791); 19 Jul 2011 12:08:46 -0000
X-SWARE-Spam-Status: No, hits=-2.2 required=5.0	tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,RCVD_IN_DNSWL_LOW,TW_TM
X-Spam-Check-By: sourceware.org
Received: from mail-yi0-f47.google.com (HELO mail-yi0-f47.google.com) (209.85.218.47)    by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 19 Jul 2011 12:08:11 +0000
Received: by yib18 with SMTP id 18so1938235yib.20        for <gcc-patches@gcc.gnu.org>; Tue, 19 Jul 2011 05:08:10 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.150.67.21 with SMTP id p21mr2876553yba.273.1311077290737; Tue, 19 Jul 2011 05:08:10 -0700 (PDT)
Received: by 10.150.205.2 with HTTP; Tue, 19 Jul 2011 05:08:10 -0700 (PDT)
In-Reply-To: <CAEwic4bPxnFK6ezMgQw-k3JHqhBM72+GBA7vO995MviRTe4R4Q@mail.gmail.com>
References: <CAEwic4YkA+GT-KDjb-65=8whqVXevNPOu+KHo4a-4ze-jasxQA@mail.gmail.com>	<CAFiYyc3QEF3grA0kxrjUmsBT59EtAzyFmbbRfQ+CEAFqJAg3yg@mail.gmail.com>	<CAEwic4Yo3fPodMJH6hVkZRYpE8edib5O7E3u_Z56dtZWjgZ3Hw@mail.gmail.com>	<CAEwic4YDkH+pLLZvmVBNv9s2-5n6s2cztSsXpj10hjO7G0AFQg@mail.gmail.com>	<Pine.LNX.4.64.1107081825110.14485@wotan.suse.de>	<CAEwic4bZQME0L1ZVOmc7WMCWLsbn_xaROF4OEv37oii2A_FBFw@mail.gmail.com>	<CAEwic4YeC1jivgWOyu1R0gNqOYsT3jzrFS1=Nj7gL-GWA7kFnQ@mail.gmail.com>	<CAFiYyc0=oJ543eAf=_shr2MYugsR7GK==L+bzdueaAKqYZZHPQ@mail.gmail.com>	<CAEwic4bPxnFK6ezMgQw-k3JHqhBM72+GBA7vO995MviRTe4R4Q@mail.gmail.com>
Date: Tue, 19 Jul 2011 12:24:00 -0000
Message-ID: <CAFiYyc0KBTFUWbHvCi+_=9KTNPL5ivuVUBXrgrkjF9nUY2ashA@mail.gmail.com>
Subject: Re: [patch tree-optimization]: [3 of 3]: Boolify compares & more
From: Richard Guenther <richard.guenther@gmail.com>
To: Kai Tietz <ktietz70@googlemail.com>
Cc: GCC Patches <gcc-patches@gcc.gnu.org>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
X-IsSubscribed: yes
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-patches.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
X-SW-Source: 2011-07/txt/msg01502.txt.bz2

On Fri, Jul 15, 2011 at 9:42 AM, Kai Tietz <ktietz70@googlemail.com> wrote:
> Hello,
>
> This patch removes from tree-vrp the use of TRUTH-bitwise expression code=
s. Also
> it merges the handling for boolean compatible and non-boolean typed
> bitwise-binary
> expressions.
> Additional it adds primitive checks for bitwise-not expression on
> boolean-compatible
> types.
> In substitute_and_fold the scan-direction of statements within a BB is
> controlled now
> by its do_dce flag. =A0This provides better results in vrp-pass.
>
> ChangeLog gcc
>
> 2011-07-15 =A0Kai Tietz =A0<ktietz@redhat.com>
>
> =A0 =A0 =A0 =A0* tree-ssa-propagate.c (substitute_and_fold): Use
> =A0 =A0 =A0 =A0do_dce flag to deside, if BB's statements are scanned
> =A0 =A0 =A0 =A0in last to first, or first to last order.
> =A0 =A0 =A0 =A0* tree-vrp.c (extract_range_from_binary_expr):
> =A0 =A0 =A0 =A0Remove TRUTH-binary checks. And unify bitwise-binary
> =A0 =A0 =A0 =A0cases.
> =A0 =A0 =A0 =A0(register_edge_assert_for_1): Add handling boolean-compati=
ble
> =A0 =A0 =A0 =A0typed BIT_IOR_EXPR and BIT_NOT_EXPR.
> =A0 =A0 =A0 =A0(extract_range_from_unary_expr): Add support for 1-bit
> =A0 =A0 =A0 =A0integral typed BIT_NOT_EXPR expression.
> =A0 =A0 =A0 =A0(extract_range_from_assignment): Remove TRUTH-binary check=
s.
> =A0 =A0 =A0 =A0Add handling for 1-bit integral typed BIT_NOT_EXPR express=
ion.
> =A0 =A0 =A0 =A0(build_assert_expr_for): Likewise.
> =A0 =A0 =A0 =A0(register_edge_assert_for_1): Likewise.
> =A0 =A0 =A0 =A0(simplify_stmt_using_ranges): Likewise.
> =A0 =A0 =A0 =A0(ssa_name_get_inner_ssa_name_p): New helper function.
> =A0 =A0 =A0 =A0(ssa_name_get_cast_to_p): New helper function.
> =A0 =A0 =A0 =A0(simplify_truth_ops_using_ranges): Handle prefixed
> =A0 =A0 =A0 =A0cast instruction for result. =A0Remove TRUTH-binary checks.
> =A0 =A0 =A0 =A0Add handling for 1-bit integral typed BIT_NOT_EXPR express=
ion.
> =A0 =A0 =A0 =A0and BIT_NOT_EXPR.
> =A0 =A0 =A0 =A0 Add handling for one bit
>
> ChangeLog gcc/testsuite
>
> 2011-07-15 =A0Kai Tietz =A0<ktietz@redhat.com>
>
> =A0 =A0 =A0 =A0* gcc.dg/tree-ssa/vrp47.c: Test no longer needs
> =A0 =A0 =A0 =A0dom dump.
>
> Bootstrapped and regression tested for all standard languages (plus
> Ada & Obj-C++) on x86_64-pc-linux-gnu. =A0Ok for apply?
>
> Regards,
> Kai
>
> Index: gcc/gcc/testsuite/gcc.dg/tree-ssa/vrp47.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc.orig/gcc/testsuite/gcc.dg/tree-ssa/vrp47.c =A0 =A0 =A02011-07-13
> 12:57:46.869620200 +0200
> +++ gcc/gcc/testsuite/gcc.dg/tree-ssa/vrp47.c =A0 2011-07-13
> 22:29:53.221967000 +0200
> @@ -4,7 +4,7 @@
> =A0 =A0jumps when evaluating an && condition. =A0VRP is not able to optim=
ize
> =A0 =A0this. =A0*/
> =A0/* { dg-do compile { target { ! "mips*-*-* s390*-*-* =A0avr-*-*
> mn10300-*-*" } } } */
> -/* { dg-options "-O2 -fdump-tree-vrp -fdump-tree-dom" } */
> +/* { dg-options "-O2 -fdump-tree-vrp" } */
> =A0/* { dg-options "-O2 -fdump-tree-vrp -fdump-tree-dom -march=3Di586" {
> target { i?86-*-* && ilp32 } } } */
>
> =A0int h(int x, int y)
> @@ -36,13 +36,10 @@ int f(int x)
> =A0 =A00 or 1. =A0*/
> =A0/* { dg-final { scan-tree-dump-times "\[xy\]\[^ \]* !=3D" 0 "vrp1" } }=
 */
>
> -/* This one needs more copy propagation that only happens in dom1. =A0*/
> -/* { dg-final { scan-tree-dump-times "x\[^ \]* & y" 1 "dom1" } } */
> -/* { dg-final { scan-tree-dump-times "x\[^ \]* & y" 1 "vrp1" { xfail
> *-*-* } } } */
> +/* { dg-final { scan-tree-dump-times "x\[^ \]* & y" 1 "vrp1" } } */
>
> =A0/* These two are fully simplified by VRP. =A0*/
> =A0/* { dg-final { scan-tree-dump-times "x\[^ \]* \[|\] y" 1 "vrp1" } } */
> =A0/* { dg-final { scan-tree-dump-times "x\[^ \]* \\^ 1" 1 "vrp1" } } */
>
> =A0/* { dg-final { cleanup-tree-dump "vrp\[0-9\]" } } */
> -/* { dg-final { cleanup-tree-dump "dom\[0-9\]" } } */
> Index: gcc/gcc/tree-ssa-propagate.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc.orig/gcc/tree-ssa-propagate.c =A0 2011-07-13 12:57:46.870620200 +=
0200
> +++ gcc/gcc/tree-ssa-propagate.c =A0 =A0 =A0 =A02011-07-13 22:29:53.25397=
1100 +0200
> @@ -979,6 +979,9 @@ replace_phi_args_in (gimple phi, ssa_pro
>
> =A0 =A0DO_DCE is true if trivially dead stmts can be removed.
>
> + =A0 If DO_DCE is true, the statements within a BB are walked from
> + =A0 last to first element. =A0Otherwise we scan from first to last elem=
ent.
> +
> =A0 =A0Return TRUE when something changed. =A0*/
>
> =A0bool
> @@ -1059,9 +1062,10 @@ substitute_and_fold (ssa_prop_get_value_
> =A0 =A0 =A0 =A0for (i =3D gsi_start_phis (bb); !gsi_end_p (i); gsi_next (=
&i))
> =A0 =A0 =A0 =A0 =A0replace_phi_args_in (gsi_stmt (i), get_value_fn);
>
> - =A0 =A0 =A0/* Propagate known values into stmts. =A0Do a backward walk =
to expose
> - =A0 =A0 =A0 =A0more trivially deletable stmts. =A0*/
> - =A0 =A0 =A0for (i =3D gsi_last_bb (bb); !gsi_end_p (i);)
> + =A0 =A0 =A0/* Propagate known values into stmts. =A0Do a backward walk =
if
> + =A0 =A0 =A0 =A0 do_dce is true. In some case it exposes
> + =A0 =A0 =A0 =A0more trivially deletable stmts to walk backward. =A0*/
> + =A0 =A0 =A0for (i =3D (do_dce ? gsi_last_bb (bb) : gsi_start_bb (bb));
> !gsi_end_p (i);)
> =A0 =A0 =A0 =A0{
> =A0 =A0 =A0 =A0 =A0 bool did_replace;
> =A0 =A0 =A0 =A0 =A0gimple stmt =3D gsi_stmt (i);
> @@ -1070,7 +1074,10 @@ substitute_and_fold (ssa_prop_get_value_
> =A0 =A0 =A0 =A0 =A0gimple_stmt_iterator oldi;
>
> =A0 =A0 =A0 =A0 =A0oldi =3D i;
> - =A0 =A0 =A0 =A0 gsi_prev (&i);
> + =A0 =A0 =A0 =A0 if (do_dce)
> + =A0 =A0 =A0 =A0 =A0 gsi_prev (&i);
> + =A0 =A0 =A0 =A0 else
> + =A0 =A0 =A0 =A0 =A0 gsi_next (&i);
>
> =A0 =A0 =A0 =A0 =A0/* Ignore ASSERT_EXPRs. =A0They are used by VRP to gen=
erate
> =A0 =A0 =A0 =A0 =A0 =A0 range information for names and they are discarded

The tree-ssa-propagate.c change is ok to apply separately.

> Index: gcc/gcc/tree-vrp.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc.orig/gcc/tree-vrp.c =A0 =A0 2011-07-13 22:25:14.690598100 +0200
> +++ gcc/gcc/tree-vrp.c =A02011-07-15 08:53:21.086266100 +0200
> @@ -2174,9 +2174,7 @@ extract_range_from_binary_expr (value_ra
> =A0 =A0 =A0 && code !=3D MIN_EXPR
> =A0 =A0 =A0 && code !=3D MAX_EXPR
> =A0 =A0 =A0 && code !=3D BIT_AND_EXPR
> - =A0 =A0 =A0&& code !=3D BIT_IOR_EXPR
> - =A0 =A0 =A0&& code !=3D TRUTH_AND_EXPR
> - =A0 =A0 =A0&& code !=3D TRUTH_OR_EXPR)
> + =A0 =A0 =A0&& code !=3D BIT_IOR_EXPR)
> =A0 =A0 {
> =A0 =A0 =A0 /* We can still do constant propagation here. =A0*/
> =A0 =A0 =A0 tree const_op0 =3D op_with_constant_singleton_value_range (op=
0);
> @@ -2231,8 +2229,7 @@ extract_range_from_binary_expr (value_ra
> =A0 =A0 =A0divisions. =A0TODO, we may be able to derive anti-ranges in
> =A0 =A0 =A0some cases. =A0*/
> =A0 if (code !=3D BIT_AND_EXPR
> - =A0 =A0 =A0&& code !=3D TRUTH_AND_EXPR
> - =A0 =A0 =A0&& code !=3D TRUTH_OR_EXPR
> + =A0 =A0 =A0&& code !=3D BIT_IOR_EXPR
> =A0 =A0 =A0 && code !=3D TRUNC_DIV_EXPR
> =A0 =A0 =A0 && code !=3D FLOOR_DIV_EXPR
> =A0 =A0 =A0 && code !=3D CEIL_DIV_EXPR
> @@ -2291,6 +2288,8 @@ extract_range_from_binary_expr (value_ra
> =A0 =A0 =A0 =A0 =A0else
> =A0 =A0 =A0 =A0 =A0 =A0set_value_range_to_varying (vr);
> =A0 =A0 =A0 =A0}
> + =A0 =A0 =A0else if (code =3D=3D BIT_IOR_EXPR)
> + =A0 =A0 =A0 =A0set_value_range_to_varying (vr);

Again, how do we arrive with a BIT_IOR_EXPR with pointer type here?
We're not supposed to have that (well, in theory, nothing verifies that).

> =A0 =A0 =A0 else
> =A0 =A0 =A0 =A0gcc_unreachable ();
>
> @@ -2299,55 +2298,7 @@ extract_range_from_binary_expr (value_ra
>
> =A0 /* For integer ranges, apply the operation to each end of the
> =A0 =A0 =A0range and see what we end up with. =A0*/
> - =A0if (code =3D=3D TRUTH_AND_EXPR
> - =A0 =A0 =A0|| code =3D=3D TRUTH_OR_EXPR)
> - =A0 =A0{
> - =A0 =A0 =A0/* If one of the operands is zero, we know that the whole
> - =A0 =A0 =A0 =A0expression evaluates zero. =A0*/
> - =A0 =A0 =A0if (code =3D=3D TRUTH_AND_EXPR
> - =A0 =A0 =A0 =A0 && ((vr0.type =3D=3D VR_RANGE
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_zerop (vr0.min)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_zerop (vr0.max))
> - =A0 =A0 =A0 =A0 =A0 =A0 || (vr1.type =3D=3D VR_RANGE
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 && integer_zerop (vr1.min)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 && integer_zerop (vr1.max))))
> - =A0 =A0 =A0 {
> - =A0 =A0 =A0 =A0 type =3D VR_RANGE;
> - =A0 =A0 =A0 =A0 min =3D max =3D build_int_cst (expr_type, 0);
> - =A0 =A0 =A0 }
> - =A0 =A0 =A0/* If one of the operands is one, we know that the whole
> - =A0 =A0 =A0 =A0expression evaluates one. =A0*/
> - =A0 =A0 =A0else if (code =3D=3D TRUTH_OR_EXPR
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& ((vr0.type =3D=3D VR_RANGE
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 && integer_onep (vr0.min)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 && integer_onep (vr0.max))
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0|| (vr1.type =3D=3D VR_RANGE
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_onep (vr1.min)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_onep (vr1.max))))
> - =A0 =A0 =A0 {
> - =A0 =A0 =A0 =A0 type =3D VR_RANGE;
> - =A0 =A0 =A0 =A0 min =3D max =3D build_int_cst (expr_type, 1);
> - =A0 =A0 =A0 }
> - =A0 =A0 =A0else if (vr0.type !=3D VR_VARYING
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& vr1.type !=3D VR_VARYING
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& vr0.type =3D=3D vr1.type
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& !symbolic_range_p (&vr0)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& !overflow_infinity_range_p (&vr0)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& !symbolic_range_p (&vr1)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& !overflow_infinity_range_p (&vr1))
> - =A0 =A0 =A0 {
> - =A0 =A0 =A0 =A0 /* Boolean expressions cannot be folded with int_const_=
binop. =A0*/
> - =A0 =A0 =A0 =A0 min =3D fold_binary (code, expr_type, vr0.min, vr1.min);
> - =A0 =A0 =A0 =A0 max =3D fold_binary (code, expr_type, vr0.max, vr1.max);
> - =A0 =A0 =A0 }
> - =A0 =A0 =A0else
> - =A0 =A0 =A0 {
> - =A0 =A0 =A0 =A0 /* The result of a TRUTH_*_EXPR is always true or false=
. =A0*/
> - =A0 =A0 =A0 =A0 set_value_range_to_truthvalue (vr, expr_type);
> - =A0 =A0 =A0 =A0 return;
> - =A0 =A0 =A0 }
> - =A0 =A0}
> - =A0else if (code =3D=3D PLUS_EXPR
> + =A0if (code =3D=3D PLUS_EXPR
> =A0 =A0 =A0 =A0 =A0 || code =3D=3D MIN_EXPR
> =A0 =A0 =A0 =A0 =A0 || code =3D=3D MAX_EXPR)
> =A0 =A0 {
> @@ -2682,71 +2633,125 @@ extract_range_from_binary_expr (value_ra
> =A0 =A0 =A0 double_int may_be_nonzero0, may_be_nonzero1;
> =A0 =A0 =A0 double_int must_be_nonzero0, must_be_nonzero1;
>
> - =A0 =A0 =A0vr0_int_cst_singleton_p =3D range_int_cst_singleton_p (&vr0);
> - =A0 =A0 =A0vr1_int_cst_singleton_p =3D range_int_cst_singleton_p (&vr1);
> - =A0 =A0 =A0int_cst_range0 =3D zero_nonzero_bits_from_vr (&vr0, &may_be_=
nonzero0,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 &must_be_nonzero0);
> - =A0 =A0 =A0int_cst_range1 =3D zero_nonzero_bits_from_vr (&vr1, &may_be_=
nonzero1,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 &must_be_nonzero1);
> -
> - =A0 =A0 =A0type =3D VR_RANGE;
> - =A0 =A0 =A0if (vr0_int_cst_singleton_p && vr1_int_cst_singleton_p)
> - =A0 =A0 =A0 min =3D max =3D int_const_binop (code, vr0.max, vr1.max);
> - =A0 =A0 =A0else if (!int_cst_range0 && !int_cst_range1)
> + =A0 =A0 =A0/* If one of the operands is zero, we know that the whole
> + =A0 =A0 =A0 =A0expression evaluates zero. =A0*/

context diffs help ... now I have to wade through +- mess :/

> + =A0 =A0 =A0if (code =3D=3D BIT_AND_EXPR
> + =A0 =A0 =A0 =A0 && ((vr0.type =3D=3D VR_RANGE
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_zerop (vr0.min)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_zerop (vr0.max))
> + =A0 =A0 =A0 =A0 =A0 =A0 || (vr1.type =3D=3D VR_RANGE
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 && integer_zerop (vr1.min)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 && integer_zerop (vr1.max))))

if you wrap all this in

 if (vr0_int_cst_singleton_p || vr1_int_cst_singleton_p)

it becomes much simpler.

> =A0 =A0 =A0 =A0{
> - =A0 =A0 =A0 =A0 set_value_range_to_varying (vr);
> - =A0 =A0 =A0 =A0 return;
> + =A0 =A0 =A0 =A0 type =3D VR_RANGE;
> + =A0 =A0 =A0 =A0 min =3D max =3D build_int_cst (expr_type, 0);

this can also be handled better via improving the existing

      if (vr0_int_cst_singleton_p && vr1_int_cst_singleton_p)
        min =3D max =3D int_const_binop (code, vr0.max, vr1.max);

handling to include the 0 and all-1s cases for AND/IOR instead
of trying to move the TRUTH_* code here.

> =A0 =A0 =A0 =A0}
> - =A0 =A0 =A0else if (code =3D=3D BIT_AND_EXPR)
> + =A0 =A0 =A0/* If one of the operands has all bits set to one, we know
> + =A0 =A0 =A0 =A0 that the whole expression evaluates to this one. =A0*/
> + =A0 =A0 =A0else if (code =3D=3D BIT_IOR_EXPR
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0&& (vr0.type =3D=3D VR_RANGE
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_all_onesp (vr0.min)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_all_onesp (vr0.max)))
> =A0 =A0 =A0 =A0{
> - =A0 =A0 =A0 =A0 min =3D double_int_to_tree (expr_type,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 dou=
ble_int_and (must_be_nonzero0,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 must_be_nonzero1));
> - =A0 =A0 =A0 =A0 max =3D double_int_to_tree (expr_type,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 dou=
ble_int_and (may_be_nonzero0,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 may_be_nonzero1));
> - =A0 =A0 =A0 =A0 if (TREE_OVERFLOW (min) || tree_int_cst_sgn (min) < 0)
> - =A0 =A0 =A0 =A0 =A0 min =3D NULL_TREE;
> - =A0 =A0 =A0 =A0 if (TREE_OVERFLOW (max) || tree_int_cst_sgn (max) < 0)
> - =A0 =A0 =A0 =A0 =A0 max =3D NULL_TREE;
> - =A0 =A0 =A0 =A0 if (int_cst_range0 && tree_int_cst_sgn (vr0.min) >=3D 0)
> - =A0 =A0 =A0 =A0 =A0 {
> - =A0 =A0 =A0 =A0 =A0 =A0 if (min =3D=3D NULL_TREE)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 min =3D build_int_cst (expr_type, 0);
> - =A0 =A0 =A0 =A0 =A0 =A0 if (max =3D=3D NULL_TREE || tree_int_cst_lt (vr=
0.max, max))
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 max =3D vr0.max;
> + =A0 =A0 =A0 =A0 type =3D VR_RANGE;
> + =A0 =A0 =A0 =A0 min =3D max =3D fold_convert (expr_type, vr0.min);
> + =A0 =A0 =A0 }
> + =A0 =A0 =A0else if (code =3D=3D BIT_IOR_EXPR
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0&& (vr1.type =3D=3D VR_RANGE
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_all_onesp (vr1.min)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& integer_all_onesp (vr1.max)))
> + =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 type =3D VR_RANGE;
> + =A0 =A0 =A0 =A0 min =3D max =3D fold_convert (expr_type, vr1.min);
> + =A0 =A0 =A0 }
> + =A0 =A0 =A0else if (TYPE_PRECISION (TREE_TYPE (op1)) =3D=3D 1)
> + =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 if (vr0.type !=3D VR_VARYING
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& vr1.type !=3D VR_VARYING
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& vr0.type =3D=3D vr1.type
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& !symbolic_range_p (&vr0)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& !overflow_infinity_range_p (&vr0)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& !symbolic_range_p (&vr1)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0&& !overflow_infinity_range_p (&vr1))
> + =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 /* Boolean expressions cannot be folded with in=
t_const_binop. =A0*/
> + =A0 =A0 =A0 =A0 =A0 =A0 min =3D fold_binary (code, expr_type, vr0.min, =
vr1.min);
> + =A0 =A0 =A0 =A0 =A0 =A0 max =3D fold_binary (code, expr_type, vr0.max, =
vr1.max);
> =A0 =A0 =A0 =A0 =A0 =A0}
> - =A0 =A0 =A0 =A0 if (int_cst_range1 && tree_int_cst_sgn (vr1.min) >=3D 0)
> + =A0 =A0 =A0 =A0 else
> =A0 =A0 =A0 =A0 =A0 =A0{
> - =A0 =A0 =A0 =A0 =A0 =A0 if (min =3D=3D NULL_TREE)
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 min =3D build_int_cst (expr_type, 0);
> - =A0 =A0 =A0 =A0 =A0 =A0 if (max =3D=3D NULL_TREE || tree_int_cst_lt (vr=
1.max, max))
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 max =3D vr1.max;
> + =A0 =A0 =A0 =A0 =A0 =A0 set_value_range_to_varying (vr);
> + =A0 =A0 =A0 =A0 =A0 =A0 return;
> =A0 =A0 =A0 =A0 =A0 =A0}
> =A0 =A0 =A0 =A0}
> - =A0 =A0 =A0else if (!int_cst_range0
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0|| !int_cst_range1
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0|| tree_int_cst_sgn (vr0.min) < 0
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0|| tree_int_cst_sgn (vr1.min) < 0)
> - =A0 =A0 =A0 {
> - =A0 =A0 =A0 =A0 set_value_range_to_varying (vr);
> - =A0 =A0 =A0 =A0 return;
> - =A0 =A0 =A0 }
> =A0 =A0 =A0 else
> - =A0 =A0 =A0 {
> - =A0 =A0 =A0 =A0 min =3D double_int_to_tree (expr_type,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 dou=
ble_int_ior (must_be_nonzero0,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 must_be_nonzero1));
> - =A0 =A0 =A0 =A0 max =3D double_int_to_tree (expr_type,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 dou=
ble_int_ior (may_be_nonzero0,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 may_be_nonzero1));
> - =A0 =A0 =A0 =A0 if (TREE_OVERFLOW (min) || tree_int_cst_sgn (min) < 0)
> - =A0 =A0 =A0 =A0 =A0 min =3D vr0.min;
> + =A0 =A0 =A0 =A0{
> + =A0 =A0 =A0 =A0 vr0_int_cst_singleton_p =3D range_int_cst_singleton_p (=
&vr0);
> + =A0 =A0 =A0 =A0 vr1_int_cst_singleton_p =3D range_int_cst_singleton_p (=
&vr1);
> + =A0 =A0 =A0 =A0 int_cst_range0 =3D zero_nonzero_bits_from_vr (&vr0, &ma=
y_be_nonzero0,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 &must_be_nonzero0);
> + =A0 =A0 =A0 =A0 int_cst_range1 =3D zero_nonzero_bits_from_vr (&vr1, &ma=
y_be_nonzero1,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 &must_be_nonzero1);
> +
> + =A0 =A0 =A0 =A0 type =3D VR_RANGE;
> + =A0 =A0 =A0 =A0 if (vr0_int_cst_singleton_p && vr1_int_cst_singleton_p)
> + =A0 =A0 =A0 =A0 =A0 min =3D max =3D int_const_binop (code, vr0.max, vr1=
.max);
> + =A0 =A0 =A0 =A0 else if (!int_cst_range0 && !int_cst_range1)
> + =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 set_value_range_to_varying (vr);
> + =A0 =A0 =A0 =A0 =A0 =A0 return;
> + =A0 =A0 =A0 =A0 =A0 }
> + =A0 =A0 =A0 =A0 else if (code =3D=3D BIT_AND_EXPR)
> + =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 min =3D double_int_to_tree (expr_type,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 double_int_and (must_be_nonzero0,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 must_be_nonzero1));
> + =A0 =A0 =A0 =A0 =A0 =A0 max =3D double_int_to_tree (expr_type,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 double_int_and (may_be_nonzero0,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 may_be_nonzero1));
> + =A0 =A0 =A0 =A0 =A0 =A0 if (TREE_OVERFLOW (min) || tree_int_cst_sgn (mi=
n) < 0)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 min =3D NULL_TREE;
> + =A0 =A0 =A0 =A0 =A0 =A0 if (TREE_OVERFLOW (max) || tree_int_cst_sgn (ma=
x) < 0)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 max =3D NULL_TREE;
> + =A0 =A0 =A0 =A0 =A0 =A0 if (int_cst_range0 && tree_int_cst_sgn (vr0.min=
) >=3D 0)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (min =3D=3D NULL_TREE)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 min =3D build_int_cst (expr_type, 0=
);
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (max =3D=3D NULL_TREE || tree_int_cs=
t_lt (vr0.max, max))
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 max =3D vr0.max;
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 }
> + =A0 =A0 =A0 =A0 =A0 =A0 if (int_cst_range1 && tree_int_cst_sgn (vr1.min=
) >=3D 0)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (min =3D=3D NULL_TREE)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 min =3D build_int_cst (expr_type, 0=
);
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (max =3D=3D NULL_TREE || tree_int_cs=
t_lt (vr1.max, max))
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 max =3D vr1.max;
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 }
> + =A0 =A0 =A0 =A0 =A0 }
> + =A0 =A0 =A0 =A0 else if (!int_cst_range0
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0|| !int_cst_range1
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0|| tree_int_cst_sgn (vr0.min) < 0
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0|| tree_int_cst_sgn (vr1.min) < 0)
> + =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 set_value_range_to_varying (vr);
> + =A0 =A0 =A0 =A0 =A0 =A0 return;
> + =A0 =A0 =A0 =A0 =A0 }
> =A0 =A0 =A0 =A0 =A0else
> - =A0 =A0 =A0 =A0 =A0 min =3D vrp_int_const_binop (MAX_EXPR, min, vr0.min=
);
> - =A0 =A0 =A0 =A0 if (TREE_OVERFLOW (max) || tree_int_cst_sgn (max) < 0)
> - =A0 =A0 =A0 =A0 =A0 max =3D NULL_TREE;
> - =A0 =A0 =A0 =A0 min =3D vrp_int_const_binop (MAX_EXPR, min, vr1.min);
> + =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 min =3D double_int_to_tree (expr_type,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 double_int_ior (must_be_nonzero0,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 must_be_nonzero1));
> + =A0 =A0 =A0 =A0 =A0 =A0 max =3D double_int_to_tree (expr_type,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 double_int_ior (may_be_nonzero0,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 may_be_nonzero1));
> + =A0 =A0 =A0 =A0 =A0 =A0 if (TREE_OVERFLOW (min) || tree_int_cst_sgn (mi=
n) < 0)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 min =3D vr0.min;
> + =A0 =A0 =A0 =A0 =A0 =A0 else
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 min =3D vrp_int_const_binop (MAX_EXPR, min,=
 vr0.min);
> + =A0 =A0 =A0 =A0 =A0 =A0 if (TREE_OVERFLOW (max) || tree_int_cst_sgn (ma=
x) < 0)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 max =3D NULL_TREE;
> + =A0 =A0 =A0 =A0 =A0 =A0 min =3D vrp_int_const_binop (MAX_EXPR, min, vr1=
.min);
> + =A0 =A0 =A0 =A0 =A0 }
> =A0 =A0 =A0 =A0}
> =A0 =A0 }
> =A0 else
> @@ -2809,7 +2814,7 @@ extract_range_from_unary_expr (value_ran
> =A0 =A0 =A0cannot easily determine a resulting range. =A0*/
> =A0 if (code =3D=3D FIX_TRUNC_EXPR
> =A0 =A0 =A0 || code =3D=3D FLOAT_EXPR
> - =A0 =A0 =A0|| code =3D=3D BIT_NOT_EXPR
> + =A0 =A0 =A0|| (code =3D=3D BIT_NOT_EXPR && TYPE_PRECISION (type) !=3D 1)

Huh?  That doesn't look worthwhile.  Please instead provide true support
for BIT_NOT_EXPR, as a separate patch.

> =A0 =A0 =A0 || code =3D=3D CONJ_EXPR)
> =A0 =A0 {
> =A0 =A0 =A0 /* We can still do constant propagation here. =A0*/
> @@ -3303,10 +3308,7 @@ extract_range_from_assignment (value_ran
> =A0 =A0 extract_range_from_assert (vr, gimple_assign_rhs1 (stmt));
> =A0 else if (code =3D=3D SSA_NAME)
> =A0 =A0 extract_range_from_ssa_name (vr, gimple_assign_rhs1 (stmt));
> - =A0else if (TREE_CODE_CLASS (code) =3D=3D tcc_binary
> - =A0 =A0 =A0 =A0 =A0|| code =3D=3D TRUTH_AND_EXPR
> - =A0 =A0 =A0 =A0 =A0|| code =3D=3D TRUTH_OR_EXPR
> - =A0 =A0 =A0 =A0 =A0|| code =3D=3D TRUTH_XOR_EXPR)
> + =A0else if (TREE_CODE_CLASS (code) =3D=3D tcc_binary)
> =A0 =A0 extract_range_from_binary_expr (vr, gimple_assign_rhs_code (stmt),
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0gi=
mple_expr_type (stmt),
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0gi=
mple_assign_rhs1 (stmt),
> @@ -3976,7 +3978,9 @@ build_assert_expr_for (tree cond, tree v
> =A0 =A0 =A0 tree a =3D build2 (ASSERT_EXPR, TREE_TYPE (v), v, cond);
> =A0 =A0 =A0 assertion =3D gimple_build_assign (n, a);
> =A0 =A0 }
> - =A0else if (TREE_CODE (cond) =3D=3D TRUTH_NOT_EXPR)
> + =A0else if (TREE_CODE (cond) =3D=3D TRUTH_NOT_EXPR
> + =A0 =A0 =A0 =A0 =A0|| (TREE_CODE (cond) =3D=3D BIT_NOT_EXPR
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0&& TYPE_PRECISION (TREE_TYPE (cond)) =3D=3D =
1))

I dont' think we arrive with TRUTH_NOT_EXPR here either - look at the
single caller please.

> =A0 =A0 {
> =A0 =A0 =A0 /* Given !V, build the assignment N =3D false. =A0*/
> =A0 =A0 =A0 tree op0 =3D TREE_OPERAND (cond, 0);
> @@ -4519,11 +4523,9 @@ register_edge_assert_for_1 (tree op, enu
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0 =A0 =A0invert);
> =A0 =A0 }
> =A0 else if ((code =3D=3D NE_EXPR
> - =A0 =A0 =A0 =A0 =A0 && (gimple_assign_rhs_code (op_def) =3D=3D TRUTH_AN=
D_EXPR
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 || gimple_assign_rhs_code (op_def) =3D=3D B=
IT_AND_EXPR))
> + =A0 =A0 =A0 =A0 =A0 && gimple_assign_rhs_code (op_def) =3D=3D BIT_AND_E=
XPR)
> =A0 =A0 =A0 =A0 =A0 || (code =3D=3D EQ_EXPR
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0&& (gimple_assign_rhs_code (op_def) =3D=3D T=
RUTH_OR_EXPR
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0|| gimple_assign_rhs_code (op_def) =
=3D=3D BIT_IOR_EXPR)))
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0&& gimple_assign_rhs_code (op_def) =3D=3D BI=
T_IOR_EXPR))
> =A0 =A0 {
> =A0 =A0 =A0 /* Recurse on each operand. =A0*/
> =A0 =A0 =A0 retval |=3D register_edge_assert_for_1 (gimple_assign_rhs1 (o=
p_def),
> @@ -4531,7 +4533,9 @@ register_edge_assert_for_1 (tree op, enu
> =A0 =A0 =A0 retval |=3D register_edge_assert_for_1 (gimple_assign_rhs2 (o=
p_def),
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0 =A0code, e, bsi);
> =A0 =A0 }
> - =A0else if (gimple_assign_rhs_code (op_def) =3D=3D TRUTH_NOT_EXPR)
> + =A0else if (gimple_assign_rhs_code (op_def) =3D=3D TRUTH_NOT_EXPR
> + =A0 =A0 =A0 =A0 =A0|| (gimple_assign_rhs_code (op_def) =3D=3D BIT_NOT_E=
XPR
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0&& TYPE_PRECISION (TREE_TYPE (op)) =3D=3D 1))

Now without the TRUTH_NOT_EXPR handling.  Also elsewhere I guess.

> =A0 =A0 {
> =A0 =A0 =A0 /* Recurse, flipping CODE. =A0*/
> =A0 =A0 =A0 code =3D invert_tree_comparison (code, false);
> @@ -4588,8 +4592,8 @@ register_edge_assert_for (tree name, edg
> =A0 =A0 =A0the value zero or one, then we may be able to assert values
> =A0 =A0 =A0for SSA_NAMEs which flow into COND. =A0*/
>
> - =A0/* In the case of NAME =3D=3D 1 or NAME !=3D 0, for TRUTH_AND_EXPR d=
efining
> - =A0 =A0 statement of NAME we can assert both operands of the TRUTH_AND_=
EXPR
> + =A0/* In the case of NAME =3D=3D 1 or NAME !=3D 0, for BIT_AND_EXPR def=
ining
> + =A0 =A0 statement of NAME we can assert both operands of the BIT_AND_EX=
PR
> =A0 =A0 =A0have nonzero value. =A0*/
> =A0 if (((comp_code =3D=3D EQ_EXPR && integer_onep (val))
> =A0 =A0 =A0 =A0|| (comp_code =3D=3D NE_EXPR && integer_zerop (val))))
> @@ -4597,8 +4601,7 @@ register_edge_assert_for (tree name, edg
> =A0 =A0 =A0 gimple def_stmt =3D SSA_NAME_DEF_STMT (name);
>
> =A0 =A0 =A0 if (is_gimple_assign (def_stmt)
> - =A0 =A0 =A0 =A0 && (gimple_assign_rhs_code (def_stmt) =3D=3D TRUTH_AND_=
EXPR
> - =A0 =A0 =A0 =A0 =A0 =A0 || gimple_assign_rhs_code (def_stmt) =3D=3D BIT=
_AND_EXPR))
> + =A0 =A0 =A0 =A0 && gimple_assign_rhs_code (def_stmt) =3D=3D BIT_AND_EXP=
R)
> =A0 =A0 =A0 =A0{
> =A0 =A0 =A0 =A0 =A0tree op0 =3D gimple_assign_rhs1 (def_stmt);
> =A0 =A0 =A0 =A0 =A0tree op1 =3D gimple_assign_rhs2 (def_stmt);
> @@ -4607,8 +4610,8 @@ register_edge_assert_for (tree name, edg
> =A0 =A0 =A0 =A0}
> =A0 =A0 }
>
> - =A0/* In the case of NAME =3D=3D 0 or NAME !=3D 1, for TRUTH_OR_EXPR de=
fining
> - =A0 =A0 statement of NAME we can assert both operands of the TRUTH_OR_E=
XPR
> + =A0/* In the case of NAME =3D=3D 0 or NAME !=3D 1, for BIT_IOR_EXPR def=
ining
> + =A0 =A0 statement of NAME we can assert both operands of the BIT_IOR_EX=
PR
> =A0 =A0 =A0have zero value. =A0*/
> =A0 if (((comp_code =3D=3D EQ_EXPR && integer_zerop (val))
> =A0 =A0 =A0 =A0|| (comp_code =3D=3D NE_EXPR && integer_onep (val))))
> @@ -4616,11 +4619,12 @@ register_edge_assert_for (tree name, edg
> =A0 =A0 =A0 gimple def_stmt =3D SSA_NAME_DEF_STMT (name);
>
> =A0 =A0 =A0 if (is_gimple_assign (def_stmt)
> - =A0 =A0 =A0 =A0 && (gimple_assign_rhs_code (def_stmt) =3D=3D TRUTH_OR_E=
XPR
> + =A0 =A0 =A0 =A0 && ((gimple_assign_rhs_code (def_stmt) =3D=3D BIT_IOR_E=
XPR
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0&& TYPE_PRECISION (TREE_TYPE (name)) =3D=3D =
1)
> =A0 =A0 =A0 =A0 =A0 =A0 =A0/* For BIT_IOR_EXPR only if NAME =3D=3D 0 both=
 operands have
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 necessarily zero value. =A0*/

The comment needs updating and the condition wants to be re-structured.

> =A0 =A0 =A0 =A0 =A0 =A0 =A0|| (comp_code =3D=3D EQ_EXPR
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 && (gimple_assign_rhs_code (def_stmt) =
=3D=3D BIT_IOR_EXPR))))
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 && gimple_assign_rhs_code (def_stmt) =
=3D=3D BIT_IOR_EXPR)))
> =A0 =A0 =A0 =A0{
> =A0 =A0 =A0 =A0 =A0tree op0 =3D gimple_assign_rhs1 (def_stmt);
> =A0 =A0 =A0 =A0 =A0tree op1 =3D gimple_assign_rhs2 (def_stmt);
> @@ -6747,19 +6751,95 @@ varying:
> =A0 return SSA_PROP_VARYING;
> =A0}
>
> +/* Returns operand1 of ssa-name with SSA_NAME as code, Otherwise it
> + =A0 returns NULL_TREE. =A0*/
> +static tree
> +ssa_name_get_inner_ssa_name_p (tree op)
> +{
> + =A0gimple stmt;
> +
> + =A0if (TREE_CODE (op) !=3D SSA_NAME
> + =A0 =A0 =A0|| !is_gimple_assign (SSA_NAME_DEF_STMT (op)))
> + =A0 =A0return NULL_TREE;
> + =A0stmt =3D SSA_NAME_DEF_STMT (op);
> + =A0if (gimple_assign_rhs_code (stmt) !=3D SSA_NAME)
> + =A0 =A0return NULL_TREE;
> + =A0return gimple_assign_rhs1 (stmt);
> +}

This and the following should be all a separate patch.  Please.

> +/* Returns operand of cast operation, if OP is a type-conversion. Otherw=
ise
> + =A0 return NULL_TREE. =A0*/
> +static tree
> +ssa_name_get_cast_to_p (tree op)
> +{
> + =A0gimple stmt;
> +
> + =A0if (TREE_CODE (op) !=3D SSA_NAME
> + =A0 =A0 =A0|| !is_gimple_assign (SSA_NAME_DEF_STMT (op)))
> + =A0 =A0return NULL_TREE;
> + =A0stmt =3D SSA_NAME_DEF_STMT (op);
> + =A0if (!CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (stmt)))
> + =A0 =A0return NULL_TREE;
> + =A0return gimple_assign_rhs1 (stmt);
> +}
> +
> =A0/* Simplify boolean operations if the source is known
> =A0 =A0to be already a boolean. =A0*/
> =A0static bool
> =A0simplify_truth_ops_using_ranges (gimple_stmt_iterator *gsi, gimple stm=
t)
> =A0{
> =A0 enum tree_code rhs_code =3D gimple_assign_rhs_code (stmt);
> + =A0gimple stmt2 =3D stmt;
> =A0 tree val =3D NULL;
> - =A0tree op0, op1;
> + =A0tree op0, op1, cop0, cop1;
> =A0 value_range_t *vr;
> =A0 bool sop =3D false;
> =A0 bool need_conversion;
> + =A0location_t loc =3D gimple_location (stmt);
>
> =A0 op0 =3D gimple_assign_rhs1 (stmt);
> + =A0op1 =3D NULL_TREE;
> +
> + =A0/* Handle cases with prefixed type-cast. =A0*/

What's a 'prefixed type-cast'?

Isn't most of simplify_truth(!)_ops_using_ranges obsolete now?

> + =A0if (CONVERT_EXPR_CODE_P (rhs_code)
> + =A0 =A0 =A0&& INTEGRAL_TYPE_P (TREE_TYPE (op0))
> + =A0 =A0 =A0&& TREE_CODE (op0) =3D=3D SSA_NAME
> + =A0 =A0 =A0&& is_gimple_assign (SSA_NAME_DEF_STMT (op0))
> + =A0 =A0 =A0&& INTEGRAL_TYPE_P (TREE_TYPE (gimple_assign_lhs (stmt))))
> + =A0 =A0{
> + =A0 =A0 =A0stmt2 =3D SSA_NAME_DEF_STMT (op0);
> + =A0 =A0 =A0op0 =3D gimple_assign_rhs1 (stmt2);
> + =A0 =A0 =A0if (!INTEGRAL_TYPE_P (TREE_TYPE (op0)))
> + =A0 =A0 =A0 return false;
> + =A0 =A0 =A0rhs_code =3D gimple_assign_rhs_code (stmt2);
> + =A0 =A0 =A0if (rhs_code !=3D BIT_NOT_EXPR
> + =A0 =A0 =A0 =A0 =A0&& rhs_code !=3D TRUTH_NOT_EXPR
> + =A0 =A0 =A0 =A0 && rhs_code !=3D BIT_AND_EXPR
> + =A0 =A0 =A0 =A0 && rhs_code !=3D BIT_IOR_EXPR
> + =A0 =A0 =A0 =A0 && rhs_code !=3D BIT_XOR_EXPR
> + =A0 =A0 =A0 =A0 && rhs_code !=3D NE_EXPR && rhs_code !=3D EQ_EXPR)
> + =A0 =A0 =A0 return false;
> + =A0 =A0 =A0if (rhs_code =3D=3D BIT_AND_EXPR || rhs_code =3D=3D BIT_IOR_=
EXPR
> + =A0 =A0 =A0 =A0 || rhs_code =3D=3D BIT_XOR_EXPR
> + =A0 =A0 =A0 =A0 || rhs_code =3D=3D NE_EXPR || rhs_code =3D=3D EQ_EXPR)
> + =A0 =A0 =A0 op1 =3D gimple_assign_rhs2 (stmt2);
> + =A0 =A0 =A0if (gimple_has_location (stmt2))
> + =A0 =A0 =A0 =A0loc =3D gimple_location (stmt2);
> + =A0 =A0}
> + =A0else if (CONVERT_EXPR_CODE_P (rhs_code))
> + =A0 =A0return false;
> + =A0else if (rhs_code =3D=3D BIT_AND_EXPR || rhs_code =3D=3D BIT_IOR_EXPR
> + =A0 =A0 =A0|| rhs_code =3D=3D BIT_XOR_EXPR
> + =A0 =A0 =A0|| rhs_code =3D=3D NE_EXPR || rhs_code =3D=3D EQ_EXPR)
> + =A0 =A0op1 =3D gimple_assign_rhs2 (stmt);
> +
> + =A0/* ~X is only equivalent of !X, if type-precision is one and X has
> + =A0 =A0 an integral type. =A0*/
> + =A0if (rhs_code =3D=3D BIT_NOT_EXPR
> + =A0 =A0 =A0&& (!INTEGRAL_TYPE_P (TREE_TYPE (op0))
> + =A0 =A0 =A0 =A0 || TYPE_PRECISION (TREE_TYPE (op0)) !=3D 1))
> + =A0 =A0return false;
> +
> =A0 if (TYPE_PRECISION (TREE_TYPE (op0)) !=3D 1)
> =A0 =A0 {
> =A0 =A0 =A0 if (TREE_CODE (op0) !=3D SSA_NAME)
> @@ -6775,22 +6855,100 @@ simplify_truth_ops_using_ranges (gimple_
> =A0 =A0 =A0 =A0 return false;
> =A0 =A0 }
>
> - =A0if (rhs_code =3D=3D TRUTH_NOT_EXPR)
> + =A0if (op1 && TREE_CODE (op1) !=3D INTEGER_CST
> + =A0 =A0 =A0&& TYPE_PRECISION (TREE_TYPE (op1)) !=3D 1)
> + =A0 =A0{
> + =A0 =A0 =A0vr =3D get_value_range (op1);
> + =A0 =A0 =A0val =3D compare_range_with_value (GE_EXPR, vr, integer_zero_=
node, &sop);
> + =A0 =A0 =A0if (!val || !integer_onep (val))
> + =A0 =A0 =A0 return false;
> +
> + =A0 =A0 =A0val =3D compare_range_with_value (LE_EXPR, vr, integer_one_n=
ode, &sop);
> + =A0 =A0 =A0if (!val || !integer_onep (val))
> + =A0 =A0 =A0 return false;
> + =A0 =A0}
> +
> + =A0need_conversion =3D
> + =A0 =A0!useless_type_conversion_p (TREE_TYPE (gimple_assign_lhs (stmt)),
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 TREE_TYPE (=
op0));
> +
> + =A0/* As comparisons X !=3D 0 getting folded by prior pass to (bool) X,
> + =A0 =A0 but X =3D=3D 0 might be not folded for none boolean type of X
> + =A0 =A0 to (bool) (X ^ 1).
> + =A0 =A0 So for bitwise-binary operations we have three cases to handle:
> + =A0 =A0 a) ((bool) X) op ((bool) Y)
> + =A0 =A0 b) ((bool) X) op (Y =3D=3D 0) OR (X =3D=3D 0) op ((bool) Y)
> + =A0 =A0 c) (X =3D=3D 0) op (Y =3D=3D 0)
> + =A0 =A0 The later two cases can't be handled for now, as vr tables
> + =A0 =A0 would need to be adjusted. =A0*/
> + =A0if (need_conversion
> + =A0 =A0 =A0&& (rhs_code =3D=3D BIT_XOR_EXPR
> + =A0 =A0 =A0 =A0 || rhs_code =3D=3D BIT_AND_EXPR
> + =A0 =A0 =A0 =A0 || rhs_code =3D=3D BIT_IOR_EXPR)
> + =A0 =A0 =A0&& TREE_CODE (op1) =3D=3D SSA_NAME && TREE_CODE (op0) =3D=3D=
 SSA_NAME)
> + =A0 =A0{
> + =A0 =A0 =A0cop0 =3D ssa_name_get_cast_to_p (op0);
> + =A0 =A0 =A0cop1 =3D ssa_name_get_cast_to_p (op1);
> + =A0 =A0 =A0if (!cop0 || !cop1)
> + =A0 =A0 =A0 =A0/* We would need an new statment for cases b and c, and =
we can't
> + =A0 =A0 =A0 =A0 =A0 due vr table, so bail out. =A0*/
> + =A0 =A0 =A0 =A0return false;
> +
> + =A0 =A0 =A0if (!INTEGRAL_TYPE_P (TREE_TYPE (cop0))
> + =A0 =A0 =A0 =A0 || !types_compatible_p (TREE_TYPE (cop0), TREE_TYPE (co=
p1)))
> + =A0 =A0 =A0 return false;
> + =A0 =A0 =A0need_conversion =3D
> + =A0 =A0 =A0 !useless_type_conversion_p (TREE_TYPE (gimple_assign_lhs (s=
tmt)),
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 TRE=
E_TYPE (cop0));
> + =A0 =A0 =A0if (need_conversion)
> + =A0 =A0 =A0 return false;
> + =A0 =A0 =A0op0 =3D cop0;
> + =A0 =A0 =A0op1 =3D cop1;
> +
> + =A0 =A0 =A0/* We need to re-check if value ranges for new operands
> + =A0 =A0 =A0 =A0 for 1-bit precision/range. =A0*/
> + =A0 =A0 =A0if (TYPE_PRECISION (TREE_TYPE (op0)) !=3D 1)
> + =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 if (TREE_CODE (op0) !=3D SSA_NAME)
> + =A0 =A0 =A0 =A0 =A0 return false;
> + =A0 =A0 =A0 =A0 vr =3D get_value_range (op0);
> +
> + =A0 =A0 =A0 =A0 val =3D compare_range_with_value (GE_EXPR, vr, integer_=
zero_node, &sop);
> + =A0 =A0 =A0 =A0 if (!val || !integer_onep (val))
> + =A0 =A0 =A0 =A0 =A0 return false;
> +
> + =A0 =A0 =A0 =A0 val =3D compare_range_with_value (LE_EXPR, vr, integer_=
one_node, &sop);
> + =A0 =A0 =A0 =A0 if (!val || !integer_onep (val))
> + =A0 =A0 =A0 =A0 =A0 return false;
> + =A0 =A0 =A0 }
> +
> + =A0 =A0 =A0if (op1 && TYPE_PRECISION (TREE_TYPE (op1)) !=3D 1)
> + =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 vr =3D get_value_range (op1);
> + =A0 =A0 =A0 =A0 val =3D compare_range_with_value (GE_EXPR, vr, integer_=
zero_node, &sop);
> + =A0 =A0 =A0 =A0 if (!val || !integer_onep (val))
> + =A0 =A0 =A0 =A0 =A0 return false;
> +
> + =A0 =A0 =A0 =A0 val =3D compare_range_with_value (LE_EXPR, vr, integer_=
one_node, &sop);
> + =A0 =A0 =A0 =A0 if (!val || !integer_onep (val))
> + =A0 =A0 =A0 =A0 =A0 return false;
> + =A0 =A0 =A0 }
> + =A0 =A0}
> + =A0else if (rhs_code =3D=3D TRUTH_NOT_EXPR
> + =A0 =A0 =A0 =A0 =A0|| rhs_code =3D=3D BIT_NOT_EXPR)
> =A0 =A0 {
> =A0 =A0 =A0 rhs_code =3D NE_EXPR;
> =A0 =A0 =A0 op1 =3D build_int_cst (TREE_TYPE (op0), 1);
> =A0 =A0 }
> =A0 else
> =A0 =A0 {
> - =A0 =A0 =A0op1 =3D gimple_assign_rhs2 (stmt);
> -
> =A0 =A0 =A0 /* Reduce number of cases to handle. =A0*/
> =A0 =A0 =A0 if (is_gimple_min_invariant (op1))
> =A0 =A0 =A0 =A0{
> =A0 =A0 =A0 =A0 =A0 /* Exclude anything that should have been already fol=
ded. =A0*/
> =A0 =A0 =A0 =A0 =A0if (rhs_code !=3D EQ_EXPR
> =A0 =A0 =A0 =A0 =A0 =A0 =A0&& rhs_code !=3D NE_EXPR
> - =A0 =A0 =A0 =A0 =A0 =A0 && rhs_code !=3D TRUTH_XOR_EXPR)
> + =A0 =A0 =A0 =A0 =A0 =A0 && rhs_code !=3D BIT_XOR_EXPR)
> =A0 =A0 =A0 =A0 =A0 =A0return false;
>
> =A0 =A0 =A0 =A0 =A0if (!integer_zerop (op1)
> @@ -6810,18 +6968,6 @@ simplify_truth_ops_using_ranges (gimple_
> =A0 =A0 =A0 =A0 =A0/* Punt on A =3D=3D B as there is no BIT_XNOR_EXPR. =
=A0*/
> =A0 =A0 =A0 =A0 =A0if (rhs_code =3D=3D EQ_EXPR)
> =A0 =A0 =A0 =A0 =A0 =A0return false;
> -
> - =A0 =A0 =A0 =A0 if (TYPE_PRECISION (TREE_TYPE (op1)) !=3D 1)
> - =A0 =A0 =A0 =A0 =A0 {
> - =A0 =A0 =A0 =A0 =A0 =A0 vr =3D get_value_range (op1);
> - =A0 =A0 =A0 =A0 =A0 =A0 val =3D compare_range_with_value (GE_EXPR, vr, =
integer_zero_node, &sop);
> - =A0 =A0 =A0 =A0 =A0 =A0 if (!val || !integer_onep (val))
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 return false;
> -
> - =A0 =A0 =A0 =A0 =A0 =A0 val =3D compare_range_with_value (LE_EXPR, vr, =
integer_one_node, &sop);
> - =A0 =A0 =A0 =A0 =A0 =A0 if (!val || !integer_onep (val))
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 return false;
> - =A0 =A0 =A0 =A0 =A0 }
> =A0 =A0 =A0 =A0}
> =A0 =A0 }
>
> @@ -6834,11 +6980,8 @@ simplify_truth_ops_using_ranges (gimple_
> =A0 =A0 =A0 else
> =A0 =A0 =A0 =A0location =3D gimple_location (stmt);
>
> - =A0 =A0 =A0if (rhs_code =3D=3D TRUTH_AND_EXPR || rhs_code =3D=3D TRUTH_=
OR_EXPR)
> - =A0 =A0 =A0 =A0warning_at (location, OPT_Wstrict_overflow,
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 _("assuming signed overflow does no=
t occur when "
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 "simplifying && or || to & or |=
"));
> - =A0 =A0 =A0else
> + =A0 =A0 =A0if (rhs_code !=3D BIT_AND_EXPR && rhs_code !=3D BIT_IOR_EXPR
> + =A0 =A0 =A0 =A0 && rhs_code !=3D BIT_XOR_EXPR)
> =A0 =A0 =A0 =A0 warning_at (location, OPT_Wstrict_overflow,
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0_("assuming signed overflow does n=
ot occur when "
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0"simplifying =3D=3D, !=3D or !=
 to identity or ^"));
> @@ -6856,19 +6999,17 @@ simplify_truth_ops_using_ranges (gimple_
>
> =A0 switch (rhs_code)
> =A0 =A0 {
> - =A0 =A0case TRUTH_AND_EXPR:
> - =A0 =A0 =A0rhs_code =3D BIT_AND_EXPR;
> - =A0 =A0 =A0break;
> - =A0 =A0case TRUTH_OR_EXPR:
> - =A0 =A0 =A0rhs_code =3D BIT_IOR_EXPR;
> + =A0 =A0case BIT_AND_EXPR:
> + =A0 =A0case BIT_IOR_EXPR:
> =A0 =A0 =A0 break;
> - =A0 =A0case TRUTH_XOR_EXPR:
> + =A0 =A0case BIT_XOR_EXPR:
> =A0 =A0 case NE_EXPR:
> =A0 =A0 =A0 if (integer_zerop (op1))
> =A0 =A0 =A0 =A0{
> =A0 =A0 =A0 =A0 =A0gimple_assign_set_rhs_with_ops (gsi,
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0need_conversion ? NOP_EXPR : SSA_NAME,
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0op0, NULL);
> + =A0 =A0 =A0 =A0 gimple_set_location (stmt, loc);
> =A0 =A0 =A0 =A0 =A0update_stmt (gsi_stmt (*gsi));
> =A0 =A0 =A0 =A0 =A0return true;
> =A0 =A0 =A0 =A0}
> @@ -6879,10 +7020,20 @@ simplify_truth_ops_using_ranges (gimple_
> =A0 =A0 =A0 gcc_unreachable ();
> =A0 =A0 }
>
> + =A0/* We can't insert here new expression as otherwise
> + =A0 =A0 tracked vr tables getting out of bounds. =A0*/
> =A0 if (need_conversion)
> =A0 =A0 return false;
>
> + =A0/* Reduce here SSA_NAME -> SSA_NAME. =A0*/
> + =A0while ((cop0 =3D ssa_name_get_inner_ssa_name_p (op0)) !=3D NULL_TREE)
> + =A0 =A0op0 =3D cop0;
> +
> + =A0while ((cop1 =3D ssa_name_get_inner_ssa_name_p (op1)) !=3D NULL_TREE)
> + =A0 =A0op1 =3D cop1;
> +

??

> =A0 gimple_assign_set_rhs_with_ops (gsi, rhs_code, op0, op1);
> + =A0gimple_set_location (stmt, loc);
> =A0 update_stmt (gsi_stmt (*gsi));
> =A0 return true;

Well, at least previously the function was readable and now it looks
like spaghetti.

> =A0}
> @@ -7417,10 +7568,8 @@ simplify_stmt_using_ranges (gimple_stmt_
> =A0 =A0 =A0 =A0{
> =A0 =A0 =A0 =A0case EQ_EXPR:
> =A0 =A0 =A0 =A0case NE_EXPR:
> + =A0 =A0 =A0 case BIT_NOT_EXPR:
> =A0 =A0 =A0 =A0case TRUTH_NOT_EXPR:
> - =A0 =A0 =A0 case TRUTH_AND_EXPR:
> - =A0 =A0 =A0 case TRUTH_OR_EXPR:
> - =A0 =A0 =A0 =A0case TRUTH_XOR_EXPR:
> =A0 =A0 =A0 =A0 =A0 /* Transform EQ_EXPR, NE_EXPR, TRUTH_NOT_EXPR into BI=
T_XOR_EXPR
> =A0 =A0 =A0 =A0 =A0 =A0 or identity if the RHS is zero or one, and the LH=
S are known
> =A0 =A0 =A0 =A0 =A0 =A0 to be boolean values. =A0Transform all TRUTH_*_EX=
PR into
> @@ -7452,13 +7601,21 @@ simplify_stmt_using_ranges (gimple_stmt_
> =A0 =A0 =A0 =A0 =A0 =A0 if all the bits being cleared are already cleared=
 or
> =A0 =A0 =A0 =A0 =A0 =A0 all the bits being set are already set. =A0*/
> =A0 =A0 =A0 =A0 =A0if (INTEGRAL_TYPE_P (TREE_TYPE (rhs1)))
> - =A0 =A0 =A0 =A0 =A0 return simplify_bit_ops_using_ranges (gsi, stmt);
> + =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 if (simplify_truth_ops_using_ranges (gsi, stmt))
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return true;
> + =A0 =A0 =A0 =A0 =A0 =A0 return simplify_bit_ops_using_ranges (gsi, stmt=
);
> + =A0 =A0 =A0 =A0 =A0 }
> =A0 =A0 =A0 =A0 =A0break;
>
> =A0 =A0 =A0 =A0CASE_CONVERT:
> =A0 =A0 =A0 =A0 =A0if (TREE_CODE (rhs1) =3D=3D SSA_NAME
> =A0 =A0 =A0 =A0 =A0 =A0 =A0&& INTEGRAL_TYPE_P (TREE_TYPE (rhs1)))
> - =A0 =A0 =A0 =A0 =A0 return simplify_conversion_using_ranges (stmt);
> + =A0 =A0 =A0 =A0 =A0 {
> + =A0 =A0 =A0 =A0 =A0 =A0 if (simplify_truth_ops_using_ranges (gsi, stmt))
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return true;
> + =A0 =A0 =A0 =A0 =A0 =A0 return simplify_conversion_using_ranges (stmt);
> + =A0 =A0 =A0 =A0 =A0 }
> =A0 =A0 =A0 =A0 =A0break;
>
> =A0 =A0 =A0 =A0default:
>