Richard Biener [Fri, 25 Oct 2013 11:51:11 +0000 (11:51 +0000)]
re PR tree-optimization/58626 (possible array wrong code bug)
2013-10-25 Richard Biener <rguenther@suse.de>
PR tree-optimization/58626
* tree-loop-distribution.c (enum rdg_dep_type): Remove
anti_dd, output_dd and input_dd.
(struct rdg_edge): Remove level and relation members.
(RDGE_LEVEL, RDGE_RELATION): Remove.
(dot_rdg_1): Adjust.
(create_rdg_edge_for_ddr): Remove.
(create_rdg_edges_for_scalar): Adjust.
(create_edge_for_control_dependence): Likewise.
(create_rdg_edges): Split into ...
(create_rdg_flow_edges): ... this
(create_rdg_cd_edges): ... and this.
(free_rdg): Adjust.
(build_rdg): Likewise, do not compute data dependences or
add edges for them.
(pg_add_dependence_edges): New function.
(pgcmp): Likewise.
(distribute_loop): First apply all non-dependence based
partition mergings. Then compute dependences between partitions
and merge and order partitions according to them.
Eric Botcazou [Fri, 25 Oct 2013 09:21:11 +0000 (09:21 +0000)]
re PR rtl-optimization/58831 (wrong code at -O2 and -O3 on x86_64-linux-gnu in 64-bit mode)
PR rtl-optimization/58831
* alias.c (init_alias_analysis): At the beginning of each iteration, set
the reg_seen[N] bit if static_reg_base_value[N] is non-null.
Nick Clifton [Fri, 25 Oct 2013 09:04:53 +0000 (09:04 +0000)]
pr57793.c: Add expected error messages for targets with small integers.
* c-c++-common/pr57793.c: Add expected error messages for
targets with small integers.
* gcc.dg/c99-stdint-1.c: Only run on 32-bit plus targets.
* gcc.dg/c99-stdint-2.c: Likewise.
* gcc.dg/cdce1.c: Likewise.
* gcc.dg/fold-overflow-1.c: Likewise.
* gcc.dg/utf-cvt.c: Likewise.
* gcc.dg/ftrapv-1.c: Only run on targets that support trapping
arithmetic.
* gcc.dg/ftrapv-2.c: Likewise.
* gcc.dg/pr30286.c: Likewise.
* gcc.dg/pr19340.c: Only run on targets that support
scheduling.
* lib/target-supports.exp (check_effective_target_trapping): New
proc. Returns true if the target supports trapping arithmetic.
re PR other/58712 (issues found by --enable-checking=valgrind)
PR ipa/58712
* cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
as argument.
(cgraph_create_edge): Use the new argument.
(cgraph_create_indirect_edge): Likewise.
Tobias Burnus [Thu, 24 Oct 2013 16:30:22 +0000 (18:30 +0200)]
re PR fortran/44646 ([F08] Implement DO CONCURRENT)
2013-10-24 Tobias Burnus <burnus@net-b.de>
PR fortran/44646
* trans-stmt.c (struct forall_info): Add do_concurrent field.
(gfc_trans_forall_1): Set it for do concurrent.
(gfc_trans_forall_loop): Mark those as annot_expr_ivdep_kind.
Marek Polacek [Thu, 24 Oct 2013 13:54:00 +0000 (13:54 +0000)]
re PR c++/58705 ([c++11] ICE with invalid initializer for _Complex variable)
PR c++/58705
cp/
* typeck2.c (check_narrowing): Don't check narrowing when the scalar
initializer is empty.
testsuite/
* g++.dg/parse/pr58705.C: New test.
configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
* configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
* Makefile.in (CXXDEPMODE): Assign and change users.
(CCDEPMODE): Delete.
* configure: Regenerate.
David Malcolm [Thu, 24 Oct 2013 02:10:10 +0000 (02:10 +0000)]
gengtype: parse base classes for some GTY-marked types
2013-10-23 David Malcolm <dmalcolm@redhat.com>
* gengtype-parse.c (require_without_advance): New.
(type): For GTY-marked types that are not GTY((user)), parse any
base classes, requiring them to be single-inheritance, and not
be templates. For non-GTY-marked types and GTY((user)),
continue to skip over any C++ inheritance specification.
* gengtype-state.c (state_writer::write_state_struct_type):
Write base class of type (if any).
(read_state_struct_type): Read base class of type (if any).
* gengtype.c (new_structure): Add a "base_class" parameter.
(create_optional_field_): Update for new parameter to
new_structure.
(adjust_field_rtx_def): Likewise.
(adjust_field_tree_exp): Likewise.
* gengtype.h (struct type): Add "base_class" field to the s
union field.
(new_structure): Add "base" parameter.
Jason Merrill [Wed, 23 Oct 2013 18:08:56 +0000 (14:08 -0400)]
In C++11 a trivial [cd]tor might not be callable.
* class.c (user_provided_p): A function deleted on its declation
in the class is not user-provided.
(type_build_ctor_call): Also force a ctor call if we
might have a deleted or private trivial ctor.
(type_build_dtor_call): New.
(deduce_noexcept_on_destructors): Remove obsolete code.
* cp-tree.h: Declare type_build_dtor_call.
* decl.c (expand_static_init): Make sure trivial dtors are callable.
(cxx_maybe_build_cleanup): Likewise.
* except.c (build_throw): Likewise.
* init.c (build_value_init): Handle trivial but not callable ctors.
(perform_target_ctor): Make sure trivial dtor is callable.
(perform_member_init): Likewise.
(expand_cleanup_for_base): Likewise.
(build_vec_delete_1): Likewise.
(build_delete): Likewise.
(push_base_cleanups): Likewise.
(build_new_1): Avoid redundant error.
* method.c (synthesized_method_walk): Can't ever exit early in C++11.
Always process the subobject destructor.
* semantics.c (finish_compound_literal): Make sure trivial dtor is
callable.
* typeck2.c (split_nonconstant_init): Likewise.
Jakub Jelinek [Wed, 23 Oct 2013 16:19:17 +0000 (18:19 +0200)]
re PR tree-optimization/58775 (reassoc1 causes an ICE with some bool arithmetic)
PR tree-optimization/58775
PR tree-optimization/58791
* tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
(insert_stmt_after): Rewritten, don't move the stmt, but really
insert it.
(get_stmt_uid_with_default): Remove.
(build_and_add_sum): Use insert_stmt_after and
reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only
labels.
(update_range_test): Set uid on stmts added by
force_gimple_operand_gsi. Don't immediately modify statements
in inter-bb optimization, just update oe->op values.
(optimize_range_tests): Return bool whether any changed have
been made.
(update_ops): New function.
(struct inter_bb_range_test_entry): New type.
(maybe_optimize_range_tests): Perform statement changes here.
(not_dominated_by, appears_later_in_bb, get_def_stmt,
ensure_ops_are_available): Remove.
(find_insert_point): Rewritten.
(rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
return LHS of the (new resp. old) stmt. Don't call
ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
instead of last, move new stmt at the right place.
(linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
(negate_value): Likewise. Set uids.
(break_up_subtract_bb): Initialize uids.
(reassociate_bb): Adjust rewrite_expr_tree caller.
(do_reassoc): Don't call renumber_gimple_stmt_uids.
* gcc.dg/guality/pr58791-1.c: New test.
* gcc.dg/guality/pr58791-2.c: New test.
* gcc.dg/guality/pr58791-3.c: New test.
* gcc.dg/guality/pr58791-4.c: New test.
* gcc.dg/guality/pr58791-5.c: New test.
* gcc.c-torture/compile/pr58775.c: New test.
* gcc.dg/tree-ssa/reassoc-28.c: Don't scan reassoc1 dump.
Jeff Law [Wed, 23 Oct 2013 13:54:48 +0000 (07:54 -0600)]
tree-ssa-threadedge.c (thread_across_edge): Do not allow threading through joiner blocks with abnormal outgoing edges.
* tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
through joiner blocks with abnormal outgoing edges.
* tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
Add parameter JOINERS, to allow/disallow threading through joiner
blocks.
(thread_block): New. Call thread_block_1.
(mark_threaded_blocks): Remove code to filter out certain cases
of threading through joiner blocks.
(thread_through_all_blocks): Document how we can have a dangling
edge AUX field and clear it.
Paolo Carlini [Wed, 23 Oct 2013 11:48:26 +0000 (11:48 +0000)]
re PR libstdc++/58815 (Casting/Conversion operator for std::decimal not supported)
2013-10-23 Paolo Carlini <paolo.carlini@oracle.com>
PR libstdc++/58815
* include/decimal/decimal (decimal32::operator long long(),
decimal64::operator long long(), decimal128::operator long long()):
Add in c++11 mode per n3407.
* testsuite/decimal/pr58815.cc: New.
Jakub Jelinek [Wed, 23 Oct 2013 06:32:23 +0000 (08:32 +0200)]
gimple-pretty-print.c (dump_ssaname_info): Always print "# " before the info, not after it.
* gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
the info, not after it.
(gump_gimple_phi): Add COMMENT argument, if true, print "# " after
dump_ssaname_info call.
(pp_gimple_stmt_1): Adjust caller.
(dump_phi_nodes): Likewise. Don't print "# " here.
Jan Hubicka [Tue, 22 Oct 2013 19:15:02 +0000 (21:15 +0200)]
i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
* i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
* x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
* i386.c (expand_small_movmem_or_setmem): New function.
(expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New function
(alg_usable_p): Add support for value ranges; cleanup.
(ix86_expand_set_or_movmem): Add support for misaligned moves.
* doc/invoke.texi: Document -ggnu-pubnames.
* common.opt: Add new option -ggnu-pubnames and modify -gpubnames
logic.
* dwarf2out.c: Include gdb/gdb-index.h.
(DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
debug_generate_pub_sections.
(is_java, output_pubtables, output_pubname): New functions.
(include_pubname_in_output): Handle debug_generate_pub_sections at
level 2.
(size_of_pubnames): Use new local space_for_flags based on
debug_generate_pub_sections.
(output_pubnames): Unify pubnames and pubtypes output logic.
Genericize comments. Call output_pubname.
(dwarf2out_finish): Move logic to output_pubnames and call it.
Bill Schmidt [Tue, 22 Oct 2013 17:31:17 +0000 (17:31 +0000)]
rs6000.c (altivec_expand_vec_perm_const): Reverse meaning of merge-high and merge-low masks for little endian...
gcc:
2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
* config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
meaning of merge-high and merge-low masks for little endian; avoid
use of vector-pack masks for little endian for mismatched modes.
gcc/testsuite:
2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
* gcc.target/powerpc/altivec-perm-1.c: Move the two vector pack
tests into...
* gcc.target/powerpc/altivec-perm-3.c: ...this new test, which is
restricted to big-endian targets.
Paul Thomas [Tue, 22 Oct 2013 04:40:57 +0000 (04:40 +0000)]
PR fortran 57893
2013-10-22 Paul Thomas <pault@gcc.gnu.org>
PR fortran 57893
* class.c : Include target-memory.h.
(gfc_find_intrinsic_vtab) Build a minimal expression so that
gfc_element_size can be used to obtain the storage size, rather
that the kind value.
2013-10-22 Paul Thomas <pault@gcc.gnu.org>
PR fortran 57893
* gfortran.dg/unlimited_polymorphic_13.f90 : New test.
Diego Novillo [Mon, 21 Oct 2013 19:36:37 +0000 (15:36 -0400)]
Re-factor inclusion of tree.h.
This moves tree.h out of every header. This exposes dependencies of
tree.h in files that should probably not need it after tree and gimple
are separated.
After this change, no header should include tree.h directly. It should
only be included by a .c file. Unfortunately, I did not find an
automatic way of forcing this.
Tested on x86_64 with all languages enabled and using
contrib/config-list.mk.
Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
and get_inner_reference returning different pmode for non-volatile
bit-field members dependent on flag_strict_volatile_bitfields.
* stor-layout.c (layout_decl): Remove special handling of
flag_strict_volatile_bitfields.
* expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
Jeff Law [Mon, 21 Oct 2013 15:25:09 +0000 (09:25 -0600)]
tree-ssa-threadedge.c (thread_through_normal_block): New argument VISITED.
* tree-ssa-threadedge.c (thread_through_normal_block): New argument VISITED.
Remove VISISTED as a local variable. When we have a threadable jump, verify
the destination of the jump has not been visised.
(thread_across_edge): Allocate VISITED bitmap once at function scope and
use it throughout. Make sure to set appropriate bits in VISITED for E (start
of jump thread path).
* tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading through
a joiner if any edge on the path has a recorded jump thread.