[Bug tree-optimization/47283] New: [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[Bug tree-optimization/47283] New: [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[doko at ubuntu dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

           Summary: [4.6 regression] ICE in refs_may_alias_p_1, at
                    tree-ssa-alias.c
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: doko@ubuntu.com


this was the only ICE seen in a partial distribution rebuilds on ix86 and
x86_64

trunk 20101113, on i686-linux-gnu, works with -Os, -O1, fails with -O2, -O3

$ g++-4.6 -c kcategorizedview.ii -g -O3
../../kdeui/itemviews/kcategorizedview.cpp: In member function 'virtual void
KCategorizedView::mouseMoveEvent(QMouseEvent*)':
../../kdeui/itemviews/kcategorizedview.cpp:961:75: warning: 'virtual void
KCategoryDrawerV2::mouseButtonMoved(const QModelIndex&, QMouseEvent*)' is
deprecated (declared at ../../kdeui/itemviews/kcategorydrawer.h:127)
[-Wdeprecated-declarations]
../../kdeui/itemviews/kcategorizedview.cpp: In member function 'virtual void
KCategorizedView::mousePressEvent(QMouseEvent*)':
../../kdeui/itemviews/kcategorizedview.cpp:999:77: warning: 'virtual void
KCategoryDrawerV2::mouseButtonPressed(const QModelIndex&, QMouseEvent*)' is
deprecated (declared at ../../kdeui/itemviews/kcategorydrawer.h:123)
[-Wdeprecated-declarations]
../../kdeui/itemviews/kcategorizedview.cpp: In member function 'virtual void
KCategorizedView::mouseReleaseEvent(QMouseEvent*)':
../../kdeui/itemviews/kcategorizedview.cpp:1030:78: warning: 'virtual void
KCategoryDrawerV2::mouseButtonReleased(const QModelIndex&, QMouseEvent*)' is
deprecated (declared at ../../kdeui/itemviews/kcategorydrawer.h:125)
[-Wdeprecated-declarations]
In file included from ../../kdeui/itemviews/kcategorizedview.cpp:1470:0:
./kcategorizedview.moc: In member function 'virtual int
KCategorizedView::qt_metacall(QMetaObject::Call, int, void**)':
./kcategorizedview.moc:87:162: warning: 'virtual void
KCategorizedView::rowsInsertedArtifficial(const QModelIndex&, int, int)' is
deprecated (declared at ../../kdeui/itemviews/kcategorizedview.cpp:1433)
[-Wdeprecated-declarations]
./kcategorizedview.moc:88:150: warning: 'virtual void
KCategorizedView::rowsRemoved(const QModelIndex&, int, int)' is deprecated
(declared at ../../kdeui/itemviews/kcategorizedview.cpp:1444)
[-Wdeprecated-declarations]
../../kdeui/itemviews/kcategorizedview.cpp: In member function 'virtual void
KCategorizedView::updateGeometries()':
../../kdeui/itemviews/kcategorizedview.cpp:1378:1: internal compiler error: in
refs_may_alias_p_1, at tree-ssa-alias.c:1081
Please submit a full bug report,
with preprocessed source if appropriate.

[Bug tree-optimization/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[doko at ubuntu dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #1 from Matthias Klose <doko at ubuntu dot com> 2011-01-13 22:11:25 UTC ---
Created attachment 22962
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22962
preprocessed source

[Bug tree-optimization/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.01.13 23:02:03
                 CC|                            |jamborm at gcc dot gnu.org
   Target Milestone|---                         |4.6.0
     Ever Confirmed|0                           |1

--- Comment #2 from H.J. Lu <hjl.tools at gmail dot com> 2011-01-13 23:02:03 UTC ---
It is caused by revision 164135:

http://gcc.gnu.org/ml/gcc-cvs/2010-09/msg00427.html

[Bug tree-optimization/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #3 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-01-14 11:03:21 UTC ---
Doesn't reproduce for me on x86_64 with -m32.

[Bug tree-optimization/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #4 from Martin Jambor <jamborm at gcc dot gnu.org> 2011-01-14 12:03:51 UTC ---
OK, I can reproduce this and will have a look.  The first interesting
observation is that -g is necessary to trigger the error.

[Bug tree-optimization/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #5 from Martin Jambor <jamborm at gcc dot gnu.org> 2011-01-17 18:12:53 UTC ---
THe ICE happens because refs_may_alias_p_1 gets an ao_ref initialized
from a MEM_REF of an ADDR_EXPR of a component_ref.
get_ref_base_and_extent then removes the MEM_REF but the base it
returns is a COMPONENT_REF which then confuses refs_may_alias_p_1.

The reference is fetched from MEM_EXPR (mem) in ao_ref_from_mem.
However, it should be noted that mem is

  (gdb) call debug_rtx (mem)
  (mem:SI (debug_implicit_ptr:SI D.58673) [0 D.58673.wd+0 S4 A32])

so perhaps some of the callers should ignore debug instructions or
debug-whatever-the-above-is?  I don't know which one that would be,
the backtrace is (I added a few asserts to alias.c to produce it so
the line numbers are a bit off):

#0  fancy_abort (file=0x8aa5c84 "/home/mjambor/gcc/mine/gcc/alias.c", line=276, 
    function=0x8aa62a2 "ao_ref_from_mem") at
/home/mjambor/gcc/mine/gcc/diagnostic.c:893
#1  0x082e4432 in ao_ref_from_mem (ref=0xbfffe828, mem=0xb5941bac) at
/home/mjambor/gcc/mine/gcc/alias.c:276
#2  0x082e4bc8 in rtx_refs_may_alias_p (x=0xb5caf7c8, mem=<value optimized
out>, tbaa_p=0 '\000')
    at /home/mjambor/gcc/mine/gcc/alias.c:380
#3  0x082e50a3 in write_dependence_p (mem=0xb5941bac, x=0xb5caf7c8,
writep=<value optimized out>)
    at /home/mjambor/gcc/mine/gcc/alias.c:2598
#4  0x08a003c9 in sched_analyze_1 (deps=<value optimized out>, x=<value
optimized out>, insn=<value optimized out>)
    at /home/mjambor/gcc/mine/gcc/sched-deps.c:2321
#5  0x08a042ac in sched_analyze_insn (deps=<value optimized out>, x=0xb5da6054,
insn=0xb633e678)
    at /home/mjambor/gcc/mine/gcc/sched-deps.c:2656
#6  0x08a05df3 in deps_analyze_insn (deps=0xbfffeae0, insn=0xb633e678)
    at /home/mjambor/gcc/mine/gcc/sched-deps.c:3259
#7  0x08a065b3 in sched_analyze (deps=0xbfffeae0, head=0xb633e678,
tail=0xb623cc80)
    at /home/mjambor/gcc/mine/gcc/sched-deps.c:3407
#8  0x08573c3e in compute_block_dependences (bb=<value optimized out>)
    at /home/mjambor/gcc/mine/gcc/sched-rgn.c:2725
#9  sched_rgn_compute_dependencies (bb=<value optimized out>) at
/home/mjambor/gcc/mine/gcc/sched-rgn.c:3162
#10 0x08576484 in schedule_region (rgn=<value optimized out>) at
/home/mjambor/gcc/mine/gcc/sched-rgn.c:2937
#11 schedule_insns (rgn=<value optimized out>) at
/home/mjambor/gcc/mine/gcc/sched-rgn.c:3321
#12 0x08576a8e in rest_of_handle_sched2 () at
/home/mjambor/gcc/mine/gcc/sched-rgn.c:3542
#13 0x085193b9 in execute_one_pass (pass=0x8cfaa60) at
/home/mjambor/gcc/mine/gcc/passes.c:1561

[Bug tree-optimization/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
                 CC|                            |jakub at gcc dot gnu.org

[Bug tree-optimization/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
         AssignedTo|unassigned at gcc dot       |jakub at gcc dot gnu.org
                   |gnu.org                     |

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-01-19 21:29:07 UTC ---
Mine.

[Bug tree-optimization/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-01-19 23:13:37 UTC ---
Created attachment 23037
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23037
pr47283.C

Reduced testcase.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-01-20 09:20:31 UTC ---
Created attachment 23041
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23041
gcc46-pr47283.patch

Untested fix.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-01-20 16:40:39 UTC ---
Author: jakub
Date: Thu Jan 20 16:40:36 2011
New Revision: 169057

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=169057
Log:
    PR debug/47283
    * cfgexpand.c (expand_debug_expr): Instead of generating
    (mem (debug_implicit_ptr)) for MEM_REFs use COMPONENT_REF
    etc. handling.

    * g++.dg/debug/pr47283.C: New test.

Added:
    trunk/gcc/testsuite/g++.dg/debug/pr47283.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/cfgexpand.c
    trunk/gcc/testsuite/ChangeLog

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-01-20 22:53:06 UTC ---
Fixed.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Uros Bizjak <ubizjak at gmail dot com> changed:

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

--- Comment #11 from Uros Bizjak <ubizjak at gmail dot com> 2011-02-15 19:21:13 UTC ---
As seen in [1], still fails on ia64-hp-hpux11.23.

The bug can be triggered with a crosscompiler to the above target:

~/gcc-build-xxx/gcc/xgcc -v
Using built-in specs.
COLLECT_GCC=/home/uros/gcc-build-xxx/gcc/xgcc
Target: ia64-hp-hpux11.23
Configured with: ../gcc-svn/trunk/configure --target=ia64-hp-hpux11.23
--enable-languages=c,c++
Thread model: posix
gcc version 4.6.0 20110215 (experimental) [trunk revision 170189] (GCC) 

~/gcc-build-xxx/gcc/cc1plus -O2 -gdwarf-2 -quiet -milp32 pr47283.C 
pr47283.C: In member function ‘virtual void E::f10()’:
pr47283.C:58:1: internal compiler error: in refs_may_alias_p_1, at
tree-ssa-alias.c:1083
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

[1] http://gcc.gnu.org/ml/gcc-testresults/2011-02/msg01661.html

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #12 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-16 08:55:41 UTC ---
Created attachment 23361
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23361
gcc46-pr47283.patch

Untested patch for this.  Can anyone test this on ia64-hpux -milp32 and also on
spu-elf (or m32c)?  The latter two because of the address space fixes.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P1                          |P4
                 CC|                            |sje at gcc dot gnu.org,
                   |                            |uweigand at gcc dot gnu.org

--- Comment #13 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-16 08:57:35 UTC ---
That said, downgrading to P4 as none of the Pmode != ptr_mode or special
address space targets are primary or secondary targets.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #14 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-16 13:50:05 UTC ---
FYI I've bootstrapped/regtested #c12 patch on x86_64-linux and i686-linux
(though, of course, it doesn't make much difference there, given that
POINTERS_EXTEND_UNSIGNED is not defined there nor it has special address
spaces).

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[sje at cup dot hp.com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Steve Ellcey <sje at cup dot hp.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sje at cup dot hp.com

--- Comment #15 from Steve Ellcey <sje at cup dot hp.com> 2011-02-17 15:54:04 UTC ---
I tested the patch from comment #12 on ia64-hp-hpux11.23 (in 32 and 64 bit
modes) and it looks good.  There were no regressions and the test
g++.dg/debug/pr47283.C now passes.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[uweigand at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #16 from Ulrich Weigand <uweigand at gcc dot gnu.org> 2011-02-17 23:24:45 UTC ---
I tested the patch from comment #12 on spu-elf with no regressions.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #17 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-28 15:47:20 UTC ---
Author: jakub
Date: Mon Feb 28 15:47:18 2011
New Revision: 170566

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=170566
Log:
    PR debug/47283
    * cfgexpand.c (convert_debug_memory_address): Add AS parameter.
    Use target address_mode and pointer_mode hooks instead of hardcoded
    Pmode and ptr_mode.  Handle some simple cases of extending if
    POINTERS_EXTEND_UNSIGNED < 0.
    (expand_debug_expr) <case MEM_REF, INDIRECT_REF, TARGET_MEM_REF>:
    Call convert_debug_memory_address.
    (expand_debug_expr) <case ADDR_EXPR>: Pass as to
    convert_debug_memory_address.

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

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #18 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-28 16:19:10 UTC ---
Fixed.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Uros Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P4                          |P3
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #19 from Uros Bizjak <ubizjak at gmail dot com> 2011-03-01 19:29:19 UTC ---
(In reply to comment #18)
> Fixed.

Unfortunately, the same failure now happens on ia64-suse-linux-gnu [1]

Again, the bug can be triggered with a crosscompiler:

~/gcc-build-xxx/gcc/xgcc -v
Using built-in specs.
COLLECT_GCC=/home/uros/gcc-build-xxx/gcc/xgcc
Target: ia64-suse-linux-gnu
Configured with: ../gcc-svn/trunk/configure --target=ia64-suse-linux-gnu
--enable-languages=c,c++
Thread model: posix
gcc version 4.6.0 20110301 (experimental) [trunk revision 170594] (GCC) 

~/gcc-build-xxx/gcc/cc1plus -O2 -gdwarf-2 -quiet pr47283.C 
pr47283.C: In member function ‘virtual void E::f10()’:
pr47283.C:58:1: internal compiler error: in refs_may_alias_p_1, at
tree-ssa-alias.c:1083
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

I have reopend the PR with P3, since this is now a failure on secondary target.

[1] http://gcc.gnu.org/ml/gcc-testresults/2011-03/msg00057.html

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[sje at cup dot hp.com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #20 from Steve Ellcey <sje at cup dot hp.com> 2011-03-01 21:50:38 UTC ---
It is still working for me on ia64-hp-hpux11.23 in 32 and 64 bit
modes but it is failing on my ia64-debian-linux-gnu system.  I failed to notice
that earlier.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #21 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-03-02 10:12:10 UTC ---
Created attachment 23511
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23511
gcc46-pr47283.patch

Untested fix.

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #22 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-03-03 12:10:44 UTC ---
Author: rguenth
Date: Thu Mar  3 12:10:40 2011
New Revision: 170650

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=170650
Log:
2011-03-03  Richard Guenther  <rguenther@suse.de>

    PR middle-end/47283
    * tree-ssa-alias.c (ptr_deref_may_alias_decl_p): Make code
    match comment.
    (refs_may_alias_p_1): For release branches return true if
    we are confused by our input.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/tree-ssa-alias.c

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

--- Comment #23 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-03-03 16:06:38 UTC ---
Author: jakub
Date: Thu Mar  3 16:06:33 2011
New Revision: 170654

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=170654
Log:
    PR debug/47283
    * cfgexpand.c (expand_debug_expr) <case MEM_REF>: If MEM_REF
    first operand is not is_gimple_mem_ref_addr, try to fold it.
    If the operand still isn't is_gimple_mem_ref_addr, clear
    MEM_EXPR on op0.

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

[Bug debug/47283] [4.6 regression] ICE in refs_may_alias_p_1, at tree-ssa-alias.c

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47283

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #24 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-03-03 16:14:50 UTC ---
Fixed.