On 18.05.2015 12:08, Vasily Averin wrote:
could you please explain the problem in more details?
is there a special reason why you use a different gcc Version for the
kernels you provide
cat /proc/version
Linux version 2.6.32-042stab106.4 (root@kbuild-rh6-x64) (gcc version
4.4.6 20120305 (Red Hat 4.4.6-4) (GCC) ) #1 SMP Fri Mar 27 15:19:28 MSK 2015
that either CentOS6 or RHEL6 ship with:
gcc -v
gcc-Version 4.4.7 20120313 (Red Hat 4.4.7-11) (GCC)
problem is difference between gcc 4.4.6 20120305 and gcc 4.4.7-11:
# rpm -q --changelog gcc | less
* Wed Aug 27 2014 Marek Polacek <pola...@redhat.com> 4.4.7-11
- support passing non-POD through ... (#1087806)
* Wed Aug 06 2014 Jakub Jelinek <ja...@redhat.com> 4.4.7-10
- backport two further OpenMP 4.0 libgomp tasking fixes (#1099549)
- fix scheduler wrong-code with DEBUG_INSNs containing volatile ASM_OPERANDS
(#1127118, PR rtl-optimization/61801)
* Fri Jul 04 2014 Jakub Jelinek <ja...@redhat.com> 4.4.7-9
- make sure conversion between aggregates with different
sizes is never considered useless (#1113878)
- fix gfortran ICE on old-style initialization of derived type
components (#1113793)
* Mon Jun 23 2014 Marek Polacek <pola...@redhat.com> 4.4.7-8
- fix up last change
* Thu Jun 19 2014 Marek Polacek <pola...@redhat.com> 4.4.7-7
- limit warning about passing objects of non-POD type (#1087806)
* Wed May 21 2014 Jakub Jelinek <ja...@redhat.com> 4.4.7-6
- backport OpenMP 4.0 support to libgomp (library only; #1099549,
PR libgomp/58691)
* Tue Apr 15 2014 Marek Polacek <pola...@redhat.com> 4.4.7-5
- fix std::uncaught_exception (#1085442, PR c++/59224)
- don't nuke gcov files if any is missing (#1008798)
- don't ICE with missing template-header (#858351, PR c++/54653)
- don't ICE when creating an epilog for reduction (#875472)
- declare type_info (#1061435, PR libstdc++/56468)
- check elements for error_mark_node (#1027003, PR c++/43028)
* Thu Jul 18 2013 Jakub Jelinek <ja...@redhat.com> 4.4.7-4
- fix diagnostics reporting with digraphs (#906234)
- handle %{, %| and %} in inline asm on multiple asm dialect
targets as {, | and } rather than asm dialect separators (#908025)
- fix ICE in build_zero_init_1 with va_list data member in a class
(#921758, PR c++/56403)
- ignore out of range registers in the unwinder (#959564, PR target/49146)
- fix ICE in fold_stmt on debug stmts (#967003)
* Thu Oct 18 2012 Jakub Jelinek <ja...@redhat.com> 4.4.7-3
- don't emit srak instruction on s390{,x} if not compiling for
z196 or later CPU (#867878)
* Tue Oct 09 2012 Jakub Jelinek <ja...@redhat.com> 4.4.7-2
- update from gcc-4_4-branch
- GCC 4.4.7 release
- fix up -freorder-blocks-and-partition EH handling
(#821901, PR middle-end/45566)
- add -f{,no-}strict-enums option, make -fno-strict-enums the default
for C++
(#826882, PR c++/43680)
- lazily declare destructor when needed (#750545)
- fix graphite ICE with VIEW_CONVERT_EXPR (#820281)
- fix power6 ICE with bswap (#819100, PR target/53199)
- handle multilib options in gnatmake (#801144)
- fix up std::__uninitialized_copy (#808590)
- fix ICE in cp_tree_equal (#831832, PR c++/54858)
- use F_SETLKW locking for *.gcda not just in libgcov.a, but also
in the compiler when reading it (PR gcov-profile/54487)
* Mon Mar 05 2012 Jakub Jelinek <ja...@redhat.com> 4.4.6-4
===========================================================================
--
Best regards,
Gena
_______________________________________________
Users mailing list
Users@openvz.org
https://lists.openvz.org/mailman/listinfo/users
