[Bug bootstrap/58532] New: bootstrap failure with BOOT_CFLAGS="-g -O3"

[Bug bootstrap/58532] New: bootstrap failure with BOOT_CFLAGS="-g -O3"

[dcb314 at hotmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

            Bug ID: 58532
           Summary: bootstrap failure with BOOT_CFLAGS="-g -O3"
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com

I just tried to bootstrap gcc 4.9 trunk dated
20130925 on an AMD Phenom with BOOT_CFLAGS="-g -O3".

It failed, but on checking the difference between
-O2 and -O3, I got the following setting
of BOOT_CFLAGS to bootstrap ok.

-g -O2 -ftree-partial-pre -fipa-cp-clone -fvect-cost-model -ftree-slp-vectorize
-ftree-loop-vectorize -fgcse-after-reload -funswitch-loops
-fpredictive-commoning

It looks to me like flag -ftree-loop-distribute-patterns breaks bootstrap.

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2013-09-26
          Component|bootstrap                   |tree-optimization
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org
   Target Milestone|---                         |4.9.0
            Summary|bootstrap failure with      |[4.9 Regression] bootstrap
                   |BOOT_CFLAGS="-g -O3"        |failure with
                   |                            |BOOT_CFLAGS="-g -O3"
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
I will have a look.

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |WAITING

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Works for me.  How did you configure?  I had to enable --disable-werror at
least
and tested c and c++, thus

../configure --enable-languages=c,c++ --disable-werror

on x86_64-linux.

How exactly did it fail for you?

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[dcb314 at hotmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

--- Comment #3 from David Binderman <dcb314 at hotmail dot com> ---
(In reply to Richard Biener from comment #2)
> How exactly did it fail for you?

I tried

../src/trunk/configure --prefix=/home/dcb/gcc/results --enable-checking=yes
--enable-languages=c,c++,fortran --disable-werror

and that failed thusly

Comparing stages 2 and 3
warning: gcc/cc1-checksum.o differs
warning: gcc/cc1plus-checksum.o differs
Bootstrap comparison failure!
gcc/fortran/parse.o differs
make[2]: *** [compare] Error 1
make[2]: Leaving directory `/home/dcb/gcc/working.ok'
make[1]: *** [stage3-bubble] Error 2
make[1]: Leaving directory `/home/dcb/gcc/working.ok'
make: *** [bootstrap] Error 2

and I tried 

../src/trunk/configure --prefix=/home/dcb/gcc/results --enable-checking=yes
--enable-languages=c,c++,fortran

and that failed like this

In file included from ../../src/trunk/gcc/function.h:27:0,
                 from ../../src/trunk/gcc/df-scan.c:32:
../../src/trunk/gcc/df-scan.c: In function ‘void
df_ref_record.constprop(df_collection_rec*, rtx, basic_block, df_insn_info*,
df_ref_type, int)’:
../../src/trunk/gcc/hard-reg-set.h:156:44: error: array subscript is above
array bounds [-Werror=array-bounds]
   (!!((SET)[(BIT) / UHOST_BITS_PER_WIDE_INT] \
                                            ^

Both bootstraps on AMD Phenom Redhat Linux
>From gcc-bugs-return-430633-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Thu Sep 26 21:08:37 2013
Return-Path: <gcc-bugs-return-430633-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 29430 invoked by alias); 26 Sep 2013 21:08:37 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 29389 invoked by uid 48); 26 Sep 2013 21:08:32 -0000
From: "charles.frasch at gmail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug lto/58528] lto1: internal compiler error: in build_abbrev_table, at dwarf2out.c:7478
Date: Thu, 26 Sep 2013 21:08:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: lto
X-Bugzilla-Version: 4.8.1
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: charles.frasch at gmail dot com
X-Bugzilla-Status: UNCONFIRMED
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-58528-4-zhZA3dLikT@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-58528-4@http.gcc.gnu.org/bugzilla/>
References: <bug-58528-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-09/txt/msg01873.txt.bz2
Content-length: 353

http://gcc.gnu.org/bugzilla/show_bug.cgi?idX528

--- Comment #5 from Charles <charles.frasch at gmail dot com> ---
I think I have "reduced" this as much as possible. Need separate compilations
for each .cpp file and the six archive libraries. Will try to make a script
that makes the ars from .ii files then runs the final link where the ICE
occurs.

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |ASSIGNED

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
Ok, I reproduced it.

Bootstrap comparison failure!
gcc/dwarf2out.o differs
gcc/fortran/parse.o differs
libiberty/regex.o differs
libiberty/pic/regex.o differs

somehow GCC has miscompiled itself.

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
There is a compare-debug failure on fortran/parse.o at least, reducing that.

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
One difference happens in 057.cunrolli already, we create a preheader for a
loop depending on -g:

 ;; Function bool gfc_parse_file() (_Z14gfc_parse_filev, funcdef_no=257,
decl_ui
d=17369, symbol_order=156)

 Created preheader block for loop 4
-Created preheader block for loop 5
 ;; 10 loops found

...

@@ -18009,13 +19254,14 @@
   <bb 109>:
   st_228 = parse_spec (131);

-  # st_470 = PHI <st_228(109)>
-  # error_471 = PHI <0(109)>
+  # st_229 = PHI <st_228(109), st_232(124)>
+  # error_233 = PHI <0(109), 1(124)>
 loop:
-
-  <bb 205>:
-  # st_229 = PHI <st_470(110), st_232(207)>
-  # error_233 = PHI <error_471(110), 1(207)>
+  # DEBUG st => NULL
+  # DEBUG error => NULL
+  # DEBUG st => NULL
+  # DEBUG error => error_233
+  # DEBUG st => st_229
   st.280_230 = (int) st_229;
   switch (st.280_230) <default: <L61>, case 12: <L59>, case 33: <L60>, case
131: <L58>>

...

   <bb 124>:
   st_232 = next_statement ();
-
-  <bb 207>:
-  goto <bb 205>;
+  goto <bb 110> (loop);

(to investigate).

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
First testcase reduction ended in PR58552, re-reducing.

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
Author: rguenth
Date: Mon Sep 30 07:44:46 2013
New Revision: 203024

URL: http://gcc.gnu.org/viewcvs?rev=203024&root=gcc&view=rev
Log:
2013-09-30  Richard Biener  <rguenther@suse.de>

    PR middle-end/58532
    * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
    before looking for setjmp-like calls.

    * g++.dg/torture/pr58552.C: New testcase.

Added:
    trunk/gcc/testsuite/g++.dg/torture/pr58552.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-cfg.c

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #9 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed.

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[dcb314 at hotmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

David Binderman <dcb314 at hotmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |---

--- Comment #10 from David Binderman <dcb314 at hotmail dot com> ---
(In reply to Richard Biener from comment #9)
> Fixed.

Not sure. My second bootstrap still fails on revision 203101

../src/trunk/configure --prefix=/home/dcb/gcc/results --enable-checking=yes
--enable-languages=c,c++,fortran

produces

In file included from ../../src/trunk/gcc/function.h:27:0,
                 from ../../src/trunk/gcc/df-scan.c:32:
../../src/trunk/gcc/df-scan.c: In function ‘void
df_ref_record.constprop(df_collection_rec*, rtx, basic_block, df_insn_info*,
df_ref_type, int)’:
../../src/trunk/gcc/hard-reg-set.h:156:44: error: array subscript is above
array bounds [-Werror=array-bounds]
   (!!((SET)[(BIT) / UHOST_BITS_PER_WIDE_INT] \
                                            ^
../../src/trunk/gcc/df-scan.c:2824:18: note: in expansion of macro
‘TEST_HARD_REG_BIT’

I presume --disable-werror would avoid the problem ?
>From gcc-bugs-return-430946-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Wed Oct 02 08:29:34 2013
Return-Path: <gcc-bugs-return-430946-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 8423 invoked by alias); 2 Oct 2013 08:29:34 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 8389 invoked by uid 48); 2 Oct 2013 08:29:31 -0000
From: "manu at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c/53001] -Wfloat-conversion should be available to warn about floating point errors
Date: Wed, 02 Oct 2013 08:29:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c
X-Bugzilla-Version: unknown
X-Bugzilla-Keywords:
X-Bugzilla-Severity: enhancement
X-Bugzilla-Who: manu at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
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-53001-4-uOpOd2mvlt@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-53001-4@http.gcc.gnu.org/bugzilla/>
References: <bug-53001-4@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
X-SW-Source: 2013-10/txt/msg00090.txt.bz2
Content-length: 1371

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53001

--- Comment #21 from Manuel López-Ibáñez <manu at gcc dot gnu.org> ---
(In reply to Joshua Cogliati from comment #20)
> Created attachment 30937 [details]
> Patch to add -Wfloat-conversion option against trunk
> 
> Added one more changed needed to get it to compile (which only occured if I
> did an rm -Rf in the build directry, the make bootstrap didn't catch the
> libcpp warning found.)  
> 
> The only thing I have left is to possibly modify the test scripts (to use
> -Wfloat-conversion instead of -Wconversion for the relevant tests) and
> recompile against the current trunk.

I would humbly suggest that you first recompile against current trunk, then
submit to gcc-patches and CC Dodji and Jason Merrill (you can find the emails
in MAINTAINERS) asking for comments before spending time and effort on updating
the test cases. The maintainers may ask you to change the name, or they might
not like that it triggers so often in GCC (so they may ask you to not enable it
with -Wextra), or they might simply not like the patch at all. 

Please mention that it is your first patch to GCC and your copyright assignment
is already in order. People are reluctant to review patches unless the papers
are in order since there is a high chance that they will never be committed.
>From gcc-bugs-return-430947-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Wed Oct 02 08:31:13 2013
Return-Path: <gcc-bugs-return-430947-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 9820 invoked by alias); 2 Oct 2013 08:31:13 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 9774 invoked by uid 48); 2 Oct 2013 08:31:10 -0000
From: "paolo.carlini at oracle dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/24208] C++ front-end can still be sped up
Date: Wed, 02 Oct 2013 08:31:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.0.0
X-Bugzilla-Keywords: compile-time-hog
X-Bugzilla-Severity: normal
X-Bugzilla-Who: paolo.carlini at oracle dot com
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: cc
Message-ID: <bug-24208-4-DKz8TkyPif@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-24208-4@http.gcc.gnu.org/bugzilla/>
References: <bug-24208-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-10/txt/msg00091.txt.bz2
Content-length: 455

http://gcc.gnu.org/bugzilla/show_bug.cgi?id$208

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tsaunders at mozilla dot com

--- Comment #8 from Paolo Carlini <paolo.carlini at oracle dot com> ---
*** Bug 58575 has been marked as a duplicate of this bug. ***

[Bug tree-optimization/58532] [4.9 Regression] bootstrap failure with BOOT_CFLAGS="-g -O3"

[rguenther at suse dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532

--- Comment #11 from rguenther at suse dot de <rguenther at suse dot de> ---
On Wed, 2 Oct 2013, dcb314 at hotmail dot com wrote:

> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58532
> 
> David Binderman <dcb314 at hotmail dot com> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>              Status|RESOLVED                    |REOPENED
>          Resolution|FIXED                       |---
> 
> --- Comment #10 from David Binderman <dcb314 at hotmail dot com> ---
> (In reply to Richard Biener from comment #9)
> > Fixed.
> 
> Not sure. My second bootstrap still fails on revision 203101
> 
> ../src/trunk/configure --prefix=/home/dcb/gcc/results --enable-checking=yes
> --enable-languages=c,c++,fortran
> 
> produces
> 
> In file included from ../../src/trunk/gcc/function.h:27:0,
>                  from ../../src/trunk/gcc/df-scan.c:32:
> ../../src/trunk/gcc/df-scan.c: In function ?void
> df_ref_record.constprop(df_collection_rec*, rtx, basic_block, df_insn_info*,
> df_ref_type, int)?:
> ../../src/trunk/gcc/hard-reg-set.h:156:44: error: array subscript is above
> array bounds [-Werror=array-bounds]
>    (!!((SET)[(BIT) / UHOST_BITS_PER_WIDE_INT] \
>                                             ^
> ../../src/trunk/gcc/df-scan.c:2824:18: note: in expansion of macro
> ?TEST_HARD_REG_BIT?
> 
> I presume --disable-werror would avoid the problem ?

Yes, it's known that non-standard flags cause spurious warnings so
--disable-werror is recommended if you change flags.

Richard.