[Bug regression/51051] New: [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[Bug regression/51051] New: [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

             Bug #: 51051
           Summary: [4.7 Regression]: build fails on cris-elf building
                    libstdc++-v3
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Keywords: build, ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: regression
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: hp@gcc.gnu.org
                CC: amodra@gcc.gnu.org
              Host: x86_64-unknown-linux-gnu
            Target: cris-axis-elf


Created attachment 25763
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=25763
cc1plus -fpreprocessed new_opv.ii -melf -g -O2 -fno-implicit-templates
-fdiagnostics-show-location=once -ffunction-sections -fdata-sections
-frandom-seed=new_opv.lo -o new_opv.s

A patch in the revision range (last_known_working:first_known_failing)
181187:181189 caused the build for cris-elf to fail as follows:
libtool: compile:  /tmp/hpautotest-gcc1/cris-elf/gccobj/./gcc/xgcc
-shared-libgcc -B/tmp/hpautotest-gcc1/cris-elf/gccobj/./gcc -nostdinc++
-L/tmp/hpautotest-gcc1/cris-elf/gccobj/cris-elf/libstdc++-v3/src
-L/tmp/hpautotest-gcc1/cris-elf/gccobj/cris-elf/libstdc++-v3/src/.libs
-nostdinc -B/tmp/hpautotest-gcc1/cris-elf/gccobj/cris-elf/newlib/ -isystem
/tmp/hpautotest-gcc1/cris-elf/gccobj/cris-elf/newlib/targ-include -isystem
/tmp/hpautotest-gcc1/gcc/newlib/libc/include
-B/tmp/hpautotest-gcc1/cris-elf/gccobj/cris-elf/libgloss/cris
-L/tmp/hpautotest-gcc1/cris-elf/gccobj/cris-elf/libgloss/libnosys
-L/tmp/hpautotest-gcc1/gcc/libgloss/cris
-B/tmp/hpautotest-gcc1/cris-elf/pre/cris-elf/bin/
-B/tmp/hpautotest-gcc1/cris-elf/pre/cris-elf/lib/ -isystem
/tmp/hpautotest-gcc1/cris-elf/pre/cris-elf/include -isystem
/tmp/hpautotest-gcc1/cris-elf/pre/cris-elf/sys-include
-I/tmp/hpautotest-gcc1/gcc/libstdc++-v3/../libgcc
-I/tmp/hpautotest-gcc1/cris-elf/gccobj/cris-elf/libstdc++-v3/include/cris-elf
-I/tmp/hpautotest-gcc1/cris-elf/gccobj/cris-elf/libstdc++-v3/include
-I/tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++ -fno-implicit-templates -Wall
-Wextra -Wwrite-strings -Wcast-qual -fdiagnostics-show-location=once
-ffunction-sections -fdata-sections -frandom-seed=new_opv.lo -g -O2 -c
/tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++/new_opv.cc -o new_opv.o
/tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++/new_opv.cc: In function 'void*
operator new [](std::size_t)':
/tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++/new_opv.cc:34:1: error: missing
REG_EH_REGION note in the end of bb 2
/tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++/new_opv.cc:34:1: error: call
edges for non-call insn in bb 2
/tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++/new_opv.cc:34:1: error: in
basic block 2:
/tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++/new_opv.cc:34:1: error: flow
control insn inside a basic block
(call_insn 8 6 50 2 (parallel [
            (set (reg:SI 10 r10)
                (call (mem:QI (symbol_ref:SI ("_Znwm") [flags 0x41]
<function_decl 0x7f94a90bf300 operator new>) [0 operator new S1 A8])
                    (const_int 0 [0])))
            (clobber (reg:SI 16 srp))
        ]) /tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++/new_opv.cc:33 222
{*expanded_call_value_non_v32}
     (expr_list:REG_EH_REGION (const_int 1 [0x1])
        (nil))
    (expr_list:REG_UNUSED (use (reg:SI 10 r10))
        (nil)))
/tmp/hpautotest-gcc1/gcc/libstdc++-v3/libsupc++/new_opv.cc:34:1: internal
compiler error: in rtl_verify_flow_info_1, at cfgrtl.c:2001



Author of suspect patch in revision range CC:ed and preprocessed source
attached.

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|regression                  |rtl-optimization
   Target Milestone|---                         |4.7.0

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[rmansfield at qnx dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

Ryan Mansfield <rmansfield at qnx dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rmansfield at qnx dot com

--- Comment #1 from Ryan Mansfield <rmansfield at qnx dot com> 2011-11-09 14:11:00 UTC ---
Similar ICE happens when building libstdc++-v3 for arm-unknown-linux-gnueabi.

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

Hans-Peter Nilsson <hp at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |http://gcc.gnu.org/ml/gcc-p
                   |                            |atches/2011-11/msg01409.htm
                   |                            |l

--- Comment #2 from Hans-Peter Nilsson <hp at gcc dot gnu.org> 2011-11-10 14:18:36 UTC ---
(In reply to comment #1)
> Similar ICE happens when building libstdc++-v3 for arm-unknown-linux-gnueabi.

Can you please test the referred patch on that target?  Thanks.

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[rmansfield at qnx dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

--- Comment #3 from Ryan Mansfield <rmansfield at qnx dot com> 2011-11-10 15:01:04 UTC ---
(In reply to comment #2)
> (In reply to comment #1)
> > Similar ICE happens when building libstdc++-v3 for arm-unknown-linux-gnueabi.
> 
> Can you please test the referred patch on that target?  Thanks.

Yes, the patch fixes the ICE that occurs when building libstdc++-v3.

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[joel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

Joel Sherrill <joel at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |joel at gcc dot gnu.org

--- Comment #4 from Joel Sherrill <joel at gcc dot gnu.org> 2011-11-10 20:32:24 UTC ---
*** Bug 51087 has been marked as a duplicate of this bug. ***

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[joel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

--- Comment #5 from Joel Sherrill <joel at gcc dot gnu.org> 2011-11-10 20:33:23 UTC ---
Problem also seen on sparc-rtems and patch referenced in URL addresses the
problem.  Build proceeding now.

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

Hans-Peter Nilsson <hp at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011-11-12
     Ever Confirmed|0                           |1

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[ramana at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

Ramana Radhakrishnan <ramana at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |michael.hope at linaro dot
                   |                            |org

--- Comment #6 from Ramana Radhakrishnan <ramana at gcc dot gnu.org> 2011-11-14 14:49:36 UTC ---
*** Bug 51068 has been marked as a duplicate of this bug. ***

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[amodra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

Alan Modra <amodra at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
                 CC|amodra at gcc dot gnu.org   |
         AssignedTo|unassigned at gcc dot       |amodra at gmail dot com
                   |gnu.org                     |

--- Comment #7 from Alan Modra <amodra at gmail dot com> 2011-11-14 22:06:29 UTC ---
Patch fixing what I broke (rather than, perhaps, underlying causes as per
Bernd's patch) in http://gcc.gnu.org/ml/gcc-patches/2011-11/msg01685.html. 
I've tested that my patch fixes the cris-elf testcase, but haven't yet
bootstrapped it.

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

--- Comment #8 from Hans-Peter Nilsson <hp at gcc dot gnu.org> 2011-11-14 23:31:17 UTC ---
Author: hp
Date: Mon Nov 14 23:31:13 2011
New Revision: 181371

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=181371
Log:
    PR rtl-optimization/51051
    * cfgrtl.c (cfg_layout_can_merge_blocks_p): Return FALSE if the
    move would cause fallthrough into the exit block.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/cfgrtl.c

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[amodra at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

--- Comment #9 from Alan Modra <amodra at gcc dot gnu.org> 2011-11-15 22:10:51 UTC ---
Author: amodra
Date: Tue Nov 15 22:10:46 2011
New Revision: 181391

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=181391
Log:
    PR rtl-optimization/51051
    PR bootstrap/51086
    * function.c (thread_prologue_and_epilogue_insns): Guard
    emitting return with single_succ_p test.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/function.c

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[amodra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

Alan Modra <amodra at gmail dot com> changed:

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

--- Comment #10 from Alan Modra <amodra at gmail dot com> 2011-11-15 22:11:25 UTC ---
Fixed

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[joel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

--- Comment #11 from Joel Sherrill <joel at gcc dot gnu.org> 2011-12-07 16:56:46 UTC ---
I still have HP's patch in my local tree. Should I remove it? Or does it need
to be committed?

[Bug rtl-optimization/51051] [4.7 Regression]: build fails on cris-elf building libstdc++-v3

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51051

--- Comment #12 from Hans-Peter Nilsson <hp at gcc dot gnu.org> 2011-12-07 19:07:27 UTC ---
(In reply to comment #11)
> I still have HP's patch in my local tree.

I assume you mean Bernd's patch referenced in this PR.
(I was only doing the legwork.)

> Should I remove it? Or does it need
> to be committed?

Both are committed, see comments 8 and 9.
Please look into what makes it look different in your tree.

brgds, H-P