[Bug middle-end/53106] New: [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[Bug middle-end/53106] New: [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

             Bug #: 53106
           Summary: [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to
                    build
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: hjl.tools@gmail.com
                CC: areg.melikadamyan@gmail.com, hubicka@gcc.gnu.org


On Linux/x86-64 and Linux/ia32, revision 186700:

http://gcc.gnu.org/ml/gcc-cvs/2012-04/msg00652.html

caused 2 benchmarks in SPEC CPU 2006 failed to build.

483.xalancbmk on Linux/x86-64:

g++ -c -o UnionDatatypeValidator.o -DSPEC_CPU -DNDEBUG  -DAPP_NO_THREADS
-DXALAN_INMEM_MSG_LOADER -I. -Ixercesc -Ixercesc/dom -Ixercesc/dom/impl
-Ixercesc/sax -Ixercesc/util/MsgLoaders/InMemory
-Ixercesc/util/Transcoders/Iconv -Ixalanc/include -DPROJ_XMLPARSER
-DPROJ_XMLUTIL -DPROJ_PARSERS -DPROJ_SAX4C -DPROJ_SAX2 -DPROJ_DOM
-DPROJ_VALIDATORS -DXML_USE_NATIVE_TRANSCODER -DXML_USE_INMEM_MESSAGELOADER -O3
-funroll-loops -ffast-math    -DSPEC_CPU_LP64  -DSPEC_CPU_LINUX    
UnionDatatypeValidator.cpp
...
UnionDatatypeValidator.cpp:605:1: error: caller edge frequency 3624 does not
match BB frequency 334
 XERCES_CPP_NAMESPACE_END
 ^
_ZN11xercesc_2_59XMLString6equalsEPKtS2_.part.2/2066 (static bool
xercesc_2_5::XMLString::_ZN11xercesc_2_59XMLString6equalsEPKtS2_.part.2(const
XMLCh*)) @0x2aaaaba46888
  Type: function
  Visibility: prevailing_def_ironly artificial
  References: 
  Referring: 
  Function static bool
xercesc_2_5::XMLString::_ZN11xercesc_2_59XMLString6equalsEPKtS2_.part.2(const
XMLCh*)/2066 is inline copy in static bool xercesc_2_5::XMLString::equals(const
XMLCh*, const XMLCh*)/179
  Availability: local
  Function flags: analyzed body local finalized
  Called by: _ZN11xercesc_2_59XMLString6equalsEPKtS2_/179 (3.62 per call)
(inlined) 
  Calls: 
UnionDatatypeValidator.cpp:605:1: internal compiler error: verify_cgraph_node
failed
 XERCES_CPP_NAMESPACE_END
 ^
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> 2012-04-24 21:50:20 UTC ---
On Linux/ia32, 447.dealII failed to build:

g++ -m32 -c -o fe_raviart_thomas.o -DSPEC_CPU -DNDEBUG  -Iinclude
-DBOOST_DISABLE_THREADS -Ddeal_II_dimension=3 -O3 -funroll-loops -msse2
-mfpmath=sse -ffast-math           fe_raviart_thomas.cc
fe_raviart_thomas.cc:2004:51: error: caller edge frequency 3943 does not match
BB frequency 390
 template class FE_RaviartThomas<deal_II_dimension>;
                                                   ^
_ZNSt14_Bit_referenceaSEb.part.9/2392 (std::_Bit_reference&
std::_Bit_reference::_ZNSt14_Bit_referenceaSEb.part.9()) @0x2aaaae94e4e0
  Type: function
  Visibility: prevailing_def_ironly visibility_specified artificial
  previous sharing asm name: 3263
  References: 
  Referring: 
  Function std::_Bit_reference&
std::_Bit_reference::_ZNSt14_Bit_referenceaSEb.part.9()/2392 is inline copy in
std::_Bit_reference& std::_Bit_reference::operator=(bool)/273
  Availability: local
  Function flags: analyzed body local finalized
  Called by: _ZNSt14_Bit_referenceaSEb/273 (3.94 per call) (inlined) 
  Calls: 
fe_raviart_thomas.cc:2004:51: internal compiler error: verify_cgraph_node
failed
 template class FE_RaviartThomas<deal_II_dimension>;
                                                   ^
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

--- Comment #2 from H.J. Lu <hjl.tools at gmail dot com> 2012-04-24 23:12:19 UTC ---
[hjl@gnu-35 delta]$ /export/gnu/import/git/gcc-test-spec/usr/bin/gcc -mx32 -O3
-S testcase.cc
testcase.cc:31:35: error: caller edge frequency 3943 does not match BB
frequency 390
 template class FE_RaviartThomas<3>;
                                   ^
_ZNSt14_Bit_referenceaSEb.part.2/256 (std::_Bit_reference&
std::_Bit_reference::_ZNSt14_Bit_referenceaSEb.part.2()) @0x7ffff112ed68
  Type: function
  Visibility: prevailing_def_ironly visibility_specified artificial
  References: 
  Referring: 
  Function std::_Bit_reference&
std::_Bit_reference::_ZNSt14_Bit_referenceaSEb.part.2()/256 is inline copy in
std::_Bit_reference& std::_Bit_reference::operator=(bool)/26
  Availability: local
  Function flags: analyzed body local finalized
  Called by: _ZNSt14_Bit_referenceaSEb/26 (3.94 per call) (inlined) 
  Calls: 
testcase.cc:31:35: internal compiler error: verify_cgraph_node failed
 template class FE_RaviartThomas<3>;
                                   ^
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
[hjl@gnu-35 delta]$ /export/gnu/import/git/gcc-test-spec/usr/bin/gcc -m32 -O3
-S testcase.cc
testcase.cc:31:35: error: caller edge frequency 3943 does not match BB
frequency 390
 template class FE_RaviartThomas<3>;
                                   ^
_ZNSt14_Bit_referenceaSEb.part.2/256 (std::_Bit_reference&
std::_Bit_reference::_ZNSt14_Bit_referenceaSEb.part.2()) @0x7ffff112ed68
  Type: function
  Visibility: prevailing_def_ironly visibility_specified artificial
  References: 
  Referring: 
  Function std::_Bit_reference&
std::_Bit_reference::_ZNSt14_Bit_referenceaSEb.part.2()/256 is inline copy in
std::_Bit_reference& std::_Bit_reference::operator=(bool)/26
  Availability: local
  Function flags: analyzed body local finalized
  Called by: _ZNSt14_Bit_referenceaSEb/26 (3.94 per call) (inlined) 
  Calls: 
testcase.cc:31:35: internal compiler error: verify_cgraph_node failed
 template class FE_RaviartThomas<3>;
                                   ^
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
[hjl@gnu-35 delta]$

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> 2012-04-24 23:13:19 UTC ---
Testcase

[hjl@gnu-35 delta]$ cat testcase.cc
#include <vector>
template <int dim>
class FiniteElement
{
  public:
    FiniteElement ( const std::vector<std::vector<bool> > &nonzero_components);
};
template <int dim>
class FE_RaviartThomas : public FiniteElement<dim>
{
  public:
    FE_RaviartThomas (const unsigned int p);
    static std::vector<bool>
    get_ria_vector (const unsigned int degree);
};
template <int dim>
FE_RaviartThomas<dim>::FE_RaviartThomas (const unsigned int degree)
  :
      FiniteElement<dim> ( std::vector<std::vector<bool> >(1))
{
}
template <int dim>
std::vector<bool>
FE_RaviartThomas<dim>::get_ria_vector (const unsigned int degree)
{
  std::vector<bool> ret_val(degree,false);
  for (unsigned int i = 0 ; i < degree ; ++i)
    ret_val[i] = true;
  return ret_val;
}
template class FE_RaviartThomas<3>;
[hjl@gnu-35 delta]$

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.8.0

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

Jan Hubicka <hubicka at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2012-04-25
         AssignedTo|unassigned at gcc dot       |hubicka at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #4 from Jan Hubicka <hubicka at gcc dot gnu.org> 2012-04-25 14:34:15 UTC ---
Thanks for the testcase!

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |UNCONFIRMED
     Ever Confirmed|1                           |0

--- Comment #5 from H.J. Lu <hjl.tools at gmail dot com> 2012-04-25 14:38:47 UTC ---
On Linux/x86-64, with

-O3 -funroll-loops -ffast-math -fwhole-program -flto=jobserver
-fuse-linker-plugin

403.gcc, 483.xalancbmk, 447.dealII, 453.povray, 454.calculix and
465.tonto fail with similar error message.

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

--- Comment #6 from Jan Hubicka <hubicka at gcc dot gnu.org> 2012-04-25 16:11:43 UTC ---
This is previously latent bug in frequency verification.  We check that
frequencies of edges match frequencies of basic block. This check is disabled
when function is inline, because then the frequencies of edges are scaled to
the same base as in the outer function.

Now what happens is that the function gets cloned, then the original version
gets fully inlined and finally becomes unreachable.  We however need to keep
function body around and thus remove_unreachable_nodes turns the inline clone
back to offline re-enabling the check that finally fires.

I am not sure how we got around this problem previously.  I am doiuble
checking.

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

--- Comment #7 from Jan Hubicka <hubicka at gcc dot gnu.org> 2012-04-25 16:21:03 UTC ---
Actually we make the node unanalyzed in this case.  There is one misupdated
place.  I am testing the following patch.
Index: ipa.c
===================================================================
--- ipa.c       (revision 186827)
+++ ipa.c       (working copy)
@@ -262,7 +262,8 @@ cgraph_remove_unreachable_nodes (bool be
                  for (next = cgraph (node->symbol.same_comdat_group);
                       next != node;
                       next = cgraph (next->symbol.same_comdat_group))
-                   if (!pointer_set_insert (reachable, next))
+                   if (!next->global.inlined_to
+                       && !pointer_set_insert (reachable, next))
                      enqueue_cgraph_node (next, &first, reachable);
                }
            }
@@ -276,7 +277,7 @@ cgraph_remove_unreachable_nodes (bool be
            {
              bool noninline = node->clone_of->symbol.decl !=
node->symbol.decl;
              node = node->clone_of;
-             if (noninline && !pointer_set_insert (reachable, node) &&
!node->symbol.aux)
+             if (noninline && !pointer_set_contains (reachable, node) &&
!node->symbol.aux)
                {
                  enqueue_cgraph_node (node, &first, reachable);
                  break;

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

--- Comment #8 from William J. Schmidt <wschmidt at gcc dot gnu.org> 2012-05-03 15:16:39 UTC ---
FWIW, I verified that Honza's proposed patch fixes the build problems for
483.xlancbmk and 32-bit 447.dealII on powerpc-linux.  Any ETA for getting this
committed?

Thanks,
Bill

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

--- Comment #9 from Jan Hubicka <hubicka at gcc dot gnu.org> 2012-05-03 19:18:56 UTC ---
Author: hubicka
Date: Thu May  3 19:18:51 2012
New Revision: 187114

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=187114
Log:

    PR middle-end/53106
    * ipa.c (cgraph_remove_unreachable_nodes): Fix handling of clones.

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

[Bug middle-end/53106] [4.8 Regression] Benchmarks in SPEC CPU 2006 failed to build

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53106

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

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

--- Comment #10 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-09-07 11:27:41 UTC ---
Fixed I presume.