Jakub Jelinek [Tue, 17 Aug 2010 18:10:20 +0000 (20:10 +0200)]
tree.h (host_integerp): Add ATTRIBUTE_PURE when not ENABLE_TREE_CHECKING.
* tree.h (host_integerp): Add ATTRIBUTE_PURE when not
ENABLE_TREE_CHECKING.
(tree_low_cst): Add inline version for !ENABLE_TREE_CHECKING
and GCC >= 4.3.
Jakub Jelinek [Tue, 17 Aug 2010 18:08:05 +0000 (20:08 +0200)]
re PR fortran/45308 ([F2003] DATE_AND_TIME does _not_ blank-pad strings)
PR fortran/45308
* intrinsics/date_and_time.c (date_and_time): Pass __{zone,time,date}_len
instead of {ZONE,TIME,DATE}_LEN as second argument to fstrcpy. Drop
asserts. Adjust comment to the F2003 wording from the F95 wording.
Jakub Jelinek [Tue, 17 Aug 2010 18:06:18 +0000 (20:06 +0200)]
re PR fortran/45304 (Functions/subroutines without arguments always treated like varargs)
PR fortran/45304
* trans-decl.c (build_library_function_decl_1): Chain on
void_list_node instead of creating a new TREE_LIST.
* trans-intrinsic.c (gfc_get_intrinsic_lib_fndecl): Likewise.
* trans-types.c (gfc_get_function_type): Likewise. Set
typelist to void_list_node for the main program.
Iain Sandoe [Tue, 17 Aug 2010 14:54:10 +0000 (14:54 +0000)]
only use _darwin10_Unwind_FindEnclosingFunction on darwin >= 10
gcc:
* unwind-dw2-fde-darwin.c (_darwin10_Unwind_FindEnclosingFunction):
Dunmmy function with NULL return unless the target is
OSX >= 10.6 (Darwin10).
libjava:
* include/posix.h: Make substitution of
_darwin10_Unwind_FindEnclosingFunction conditional on
OSX >= 10.6 (Darwin10).
Richard Guenther [Tue, 17 Aug 2010 09:55:04 +0000 (09:55 +0000)]
tree-ssa-dom.c (struct edge_info): Use a VEC for the list of conditional equivalences.
2010-08-17 Richard Guenther <rguenther@suse.de>
* tree-ssa-dom.c (struct edge_info): Use a VEC for the
list of conditional equivalences.
(free_all_edge_infos): Adjust.
(record_equivalences_from_incoming_edge): Likewise.
(record_cond): Likewise.
(build_and_record_new_cond): Likewise.
(record_conditions): Likewise.
(dom_opt_leave_block): Likewise.
PR c/40563
* c-decl.c (diagnose_uninitialized_cst_member): New function.
(finish_decl): Use it to issue a -Wc++-compat warning about
uninitialized const field in struct or union.
Daniel Kraft [Tue, 17 Aug 2010 08:20:03 +0000 (10:20 +0200)]
re PR fortran/38936 ([F03] ASSOCIATE construct / improved SELECT TYPE (a=>expr))
2010-08-17 Daniel Kraft <d@domob.eu>
PR fortran/38936
* gfortran.h (struct gfc_association_list): New member `where'.
(gfc_is_associate_pointer) New method.
* match.c (gfc_match_associate): Remember locus for each associate
name matched and do not try to set variable flag.
* parse.c (parse_associate): Use remembered locus for symbols.
* primary.c (match_variable): Instead of variable-flag check for
associate names set it for all such names used.
* symbol.c (gfc_is_associate_pointer): New method.
* resolve.c (resolve_block_construct): Don't generate assignments
to give associate-names their values.
(resolve_fl_var_and_proc): Allow associate-names to be deferred-shape.
(resolve_symbol): Set some more attributes for associate variables,
set variable flag here and check it and don't try to build an
explicitely shaped array-spec for array associate variables.
* trans-expr.c (gfc_conv_variable): Dereference in case of association
to scalar variable.
* trans-types.c (gfc_is_nodesc_array): Handle array association symbols.
(gfc_sym_type): Return pointer type for association to scalar vars.
* trans-decl.c (gfc_get_symbol_decl): Defer association symbols.
(trans_associate_var): New method.
(gfc_trans_deferred_vars): Handle association symbols.
2010-08-17 Daniel Kraft <d@domob.eu>
PR fortran/38936
* gfortran.dg/associate_1.f03: Extended to test newly supported
features like association to variables.
* gfortran.dg/associate_3.f03: Removed check for illegal change
of associate-name here...
* gfortran.dg/associate_5.f03: ...and added it here.
* gfortran.dg/associate_6.f03: No longer XFAIL'ed.
* gfortran.dg/associate_7.f03: New test.
Bernd Schmidt [Mon, 16 Aug 2010 10:25:28 +0000 (10:25 +0000)]
thumb2.md (thumb2_notsi_shiftsi, [...]): Delete.
* config/arm/thumb2.md (thumb2_notsi_shiftsi,
thumbsi_notsi_shiftsi_compare0, thumb2_not_shiftsi_compare0_scratch,
thumb2_cmpsi_shiftsi, thumb2_cmpsi_shiftsi_swp, thumb2_arith_shiftsi,
thumb2_arith_shiftsi splitter, thumb2_arith_shiftsi_compare0,
thumb2_arith_shiftsi_compare0_scratch, thumb2_sub_shiftsi,
thumb2_sub_shiftsi_compare0, thumb2_sub_shiftsi_compare0_scratch,
thumb2_iorsi3): Delete.
(orsi_notsi_si): No longer a named pattern.
(orsi_not_shiftsi_si): Renamed from thumb_orsi_not_shiftsi_si.
* config/arm/predicates.md (shift_amount_operand): New.
(mult_operator): New.
* config/arm/arm.md (attr arch, attr arch_enabled, attr insn_enabled,
attr enabled): New.
(iorsi3_insn): Renamed from arm_iorsi3. Handle a new alternative if
arch matches t2.
(not_shiftsi): Renamed from arm_notsi_shiftsi. Handle Thumb2 variant.
(not_shiftsi_compare0): Likewise, renamed from
arm_notsi_shiftsi_compare0.
(not_shiftsi_compare0_scratch): Likweise, renamed from
arm_notsi_shiftsi_compare0_scratch.
(cmpsi_shiftsi): Likewise, renamed from arm_cmpsi_shiftsi.
(cmpsi_shiftsi_swp): Likewise, renamed from arm_cmpsi_shiftsi_swp.
(arith_shiftsi): Handle Thumb2 variant. Set insn_enabled attribute
so that the register alternative is disabled when the shift_operator
is MULT. Use "M" as the constraint for constants.
(arith_shiftsi splitter): Enable for TARGET_32BIT.
(arith_shiftsi_compare0): Handle Thumb2 variant. Use "M" as the
constraint for constants.
(arith_shiftsi_compare0_scratch): Likewise.
(sub_shiftsi, sub_shiftsi_compare0, sub_shiftsi_compare0_scratch):
Handle Thumb2 alternative.
Joseph Myers [Mon, 16 Aug 2010 10:17:07 +0000 (11:17 +0100)]
options.texi (NoDriverArg): Document.
* doc/options.texi (NoDriverArg): Document.
* gcc.c (cpp_unique_options): Generate -MD and -MMD instead of
-MDX and -MMDX.
* opt-functions.awk (switch_flags): Handle NoDriverArg.
* opts-common.c (decode_cmdline_option): Ignore CL_SEPARATE
marking for CL_NO_DRIVER_ARG options when in the driver.
* opts.h (CL_NO_DRIVER_ARG): Define.
(CL_PARAMS, CL_WARNING, CL_OPTIMIZATION, CL_DRIVER, CL_TARGET,
CL_COMMON): Update values.
c-family:
* c.opt (MDX): Change back to MD. Mark NoDriverArg instead of
RejectDriver.
(MMDX): Change back to MMD. Mark NoDriverArg instead of
RejectDriver.
* c-opts.c (c_common_handle_option): Use OPT_MD and OPT_MMD
instead of OPT_MDX and OPT_MMDX.
fortran:
* lang.opt (MDX): Change back to MD. Mark NoDriverArg instead of
RejectDriver.
(MMDX): Change back to MMD. Mark NoDriverArg instead of
RejectDriver.
* cpp.c (gfc_cpp_handle_option): Use OPT_MD and OPT_MMD instead of
OPT_MDX and OPT_MMDX.
Richard Guenther [Mon, 16 Aug 2010 09:55:54 +0000 (09:55 +0000)]
tree-cfg.c (verify_types_in_gimple_reference): Verify TARGET_MEM_REF a bit.
2010-08-16 Richard Guenther <rguenther@suse.de>
* tree-cfg.c (verify_types_in_gimple_reference): Verify
TARGET_MEM_REF a bit.
* tree-ssa-address.c (addr_for_mem_ref): Adjust.
(create_mem_ref_raw): Always create TMR_OFFSET, store the
alias pointer type via it.
(copy_mem_ref_info): Adjust.
(maybe_fold_tmr): Likewise.
* tree.c (mem_ref_offset): Also handle TARGET_MEM_REF.
(reference_alias_ptr_type): Likewise.
* tree.def (TARGET_MEM_REF): Remove TMR_ORIGINAL operand,
adjust documentation of TMR_OFFSET operand.
* alias.c (get_alias_set): Do not look at TMR_ORIGINAL but
use the alias pointer type.
* expr.c (expand_expr_real_1): Do not use TMR_ORIGINAL to
initialize mem attrs but the TMR itself.
* tree-eh.c (tree_could_trap_p): Handle TARGET_MEM_REF
similar to MEM_REF.
* tree-pretty-print.c (dump_generic_node): Do not dump TMR_ORIGINAL.
* tree-ssa-loop-ivopts.c (idx_remove_ssa_names): Remove.
(unshare_and_remove_ssa_names): Likewise.
(copy_ref_info): Adjust.
* tree-ssa-pre.c (create_component_ref_by_pieces_1): Simplify
TARGET_MEM_REF case.
* tree-ssa-sccvn.c (copy_reference_ops_from_ref): Do not look
at TMR_ORIGINAL.
* tree.h (TMR_ORIGINAL): Remove.
* gimple.c (get_base_address): For TARGET_MEM_REF with a
symbol return that.
* tree-dfa.c (get_ref_base_and_extent): Handle TARGET_MEM_REF
with a symbol.
(get_addr_base_and_unit_offset): Likewise.
* tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Handle
TARGET_MEM_REFs.
(indirect_ref_may_alias_decl_p): Likewise.
(refs_may_alias_p_1): Do not bail out for TARGET_MEM_REFs.
Uros Bizjak [Sun, 15 Aug 2010 21:25:58 +0000 (23:25 +0200)]
invoke.texi (-fomit-frame-pointer): Document that starting with GCC version 4.6...
* doc/invoke.texi (-fomit-frame-pointer): Document that starting
with GCC version 4.6, the default setting (when not optimizing
for size) for 32-bit Linux x86 and 32-bit Darwin x86 targets has
been changed to -fomit-frame-pointer.
Daniel Kraft [Sun, 15 Aug 2010 19:46:21 +0000 (21:46 +0200)]
re PR fortran/38936 ([F03] ASSOCIATE construct / improved SELECT TYPE (a=>expr))
2010-08-15 Daniel Kraft <d@domob.eu>
PR fortran/38936
* gfortran.h (gfc_find_proc_namespace): New method.
* expr.c (gfc_build_intrinsic_call): No need to build symtree messing
around with namespace.
* symbol.c (gfc_find_proc_namespace): New method.
* trans-decl.c (gfc_build_qualified_array): Use it for correct
value of nest.
* primary.c (gfc_match_varspec): Handle associate-names as arrays.
* parse.c (parse_associate): Removed assignment-generation here...
* resolve.c (resolve_block_construct): ...and added it here.
(resolve_variable): Handle names that are arrays but were not parsed
as such because of association.
(resolve_code): Fix BLOCK resolution.
(resolve_symbol): Generate array-spec for associate-names.
2010-08-15 Daniel Kraft <d@domob.eu>
PR fortran/38936
* gfortran.dg/associate_1.f03: Enable test for array expressions.
* gfortran.dg/associate_3.f03: Clarify comment.
* gfortran.dg/associate_5.f03: New test.
* gfortran.dg/associate_6.f03: New test.
Daniel Kraft [Sun, 15 Aug 2010 15:28:10 +0000 (17:28 +0200)]
re PR fortran/45197 ([F2008] Allow IMPURE elemental procedures)
2010-08-15 Daniel Kraft <d@domob.eu>
PR fortran/45197
* decl.c (gfc_match_prefix): Match IMPURE prefix and mark ELEMENTAL
routines not IMPURE also as PURE.
* intrinsic.c (enum klass): New class `CLASS_PURE' and renamed
`NO_CLASS' in `CLASS_IMPURE'.
(add_sym): Set symbol-attributes `pure' and `elemental' correctly.
(add_sym_0s): Renamed `NO_CLASS' in `CLASS_IMPURE'.
(add_functions): Ditto.
(add_subroutines): Ditto and mark `MOVE_ALLOC' as CLASS_PURE.
* resolve.c (gfc_pure): Do not treat ELEMENTAL as automatically PURE.
(resolve_formal_arglist): Check that arguments to ELEMENTAL procedures
are not ALLOCATABLE and have their INTENT specified.
2010-08-15 Daniel Kraft <d@domob.eu>
PR fortran/45197
* gfortran.dg/elemental_args_check_3.f90: New test.
* gfortran.dg/impure_1.f08: New test.
* gfortran.dg/impure_2.f08: New test.
* gfortran.dg/impure_3.f90: New test.
* gfortran.dg/typebound_proc_6.f03: Changed expected error message.
Ira Rosen [Sun, 15 Aug 2010 07:00:32 +0000 (07:00 +0000)]
tree-vect-data-refs.c (vect_setup_realignment): Support realignment in basic blocks.
* tree-vect-data-refs.c (vect_setup_realignment): Support realignment
in basic blocks.
(vect_supportable_dr_alignment): Check alignment for basic blocks.
* tree-vect-slp.c (vect_build_slp_tree): Allow different codes for
data references.
(vect_bb_vectorization_profitable_p): New function.
(vect_slp_analyze_bb): Call vect_bb_vectorization_profitable_p() to
check if it's worthwhile to vectorize the basic block.
* reload.h (register_move_cost, memory_move_secondary_cost,
secondary_reload_class): Adjust prototype.
* rtl.h (reg_class_subset_p): Adjust prototype.
* reload.c (secondary_reload_class): Change arguments type from
enum reg_class to reg_class_t. Change result type to reg_class_t.
* reginfo.c (register_move_cost, reg_class_subset_p): Change
arguments type from enum reg_class to reg_class_t.
(memory_move_secondary_cost): Change arguments type from
enum reg_class to reg_class_t. Change type of saved_flags to
reg_class_t.
* config/mips/mips.h (REGISTER_MOVE_COST, MEMORY_MOVE_COST):
Remove macros.
(mips_cost): Remove.
(struct mips_rtx_cost_data): Moved to mips.c.
* config/mips/mips-protos.h (mips_register_move_cost): Remove.
* config/mips/mips.c (struct mips_rtx_cost_data): Moved from mips.h.
(mips_cost): Make static.
(mips_canonicalize_move_class): Change argument type to reg_class_t.
Change result type to reg_class_t.
(mips_move_to_gpr_cost, mips_move_from_gpr_cost): Change arguments
type from enum reg_class to reg_class_t.
(mips_register_move_cost): Make static. Change arguments
type from enum reg_class to reg_class_t.
(mips_memory_move_cost): New function.
(TARGET_REGISTER_MOVE_COST, TARGET_MEMORY_MOVE_COST): Define.
re PR target/43358 (IRA: internal compiler error: in pool_free, at alloc-pool.c:335)
gcc/
PR rtl-optimization/43358
* ira-lives.c (process_single_reg_class_operands): Adjust the costs
of a single hard register, using simplify_subreg_regno to decide
what that register should be.
* config/mips/mips.c (CODE_FOR_loongson_pmullh): Define.
* config/mips/loongson.md (UNSPEC_LOONGSON_PMULL): Delete.
(loongson_pmull<V_suffix>): Rename to...
(mul<mode>3): ...this and use MULT instead of an UNSPEC.
Paolo Carlini [Fri, 13 Aug 2010 11:48:34 +0000 (11:48 +0000)]
ios_init.cc (ios_base::Init::~Init): Add back semicolon at the end of the _GLIBCXX_SYNCHRONIZATION_HAPPENS* macros...
2010-08-12 Paolo Carlini <paolo.carlini@oracle.com>
* src/ios_init.cc (ios_base::Init::~Init): Add back semicolon
at the end of the _GLIBCXX_SYNCHRONIZATION_HAPPENS* macros,
removed by mistake.
* include/tr1_impl/boost_sp_counted_base.h: Likewise.
* include/ext/rc_string_base.h: Likewise.
* include/bits/locale_classes.h: Likewise.
* include/bits/basic_string.h: Likewise.
* include/bits/ios_base.h: Likewise.
* include/bits/c++config: Adjust comment.
Daniel Kraft [Fri, 13 Aug 2010 07:26:05 +0000 (09:26 +0200)]
gfortran.h (array_type): New type `AS_IMPLIED_SHAPE'.
2010-08-13 Daniel Kraft <d@domob.eu>
* gfortran.h (array_type): New type `AS_IMPLIED_SHAPE'.
* array.c (gfc_match_array_spec): Match implied-shape specification and
handle AS_IMPLIED_SHAPE correctly otherwise.
* decl.c (add_init_expr_to_sym): Set upper bounds for implied-shape.
(variable_decl): Some checks for implied-shape declaration.
* resolve.c (resolve_symbol): Assert that array-spec is no longer
AS_IMPLIED_SHAPE in any case.
2010-08-13 Daniel Kraft <d@domob.eu>
* gfortran.dg/implied_shape_1.f08: New test.
* gfortran.dg/implied_shape_2.f90: New test.
* gfortran.dg/implied_shape_3.f08: New test.
* config/i386/i386.c (USE_IX86_FRAME_POINTER): Default to 0.
(override_options): If not configured with --enable-frame-pointer,
enable -fomit-frame-pointer (but not for TARGET_MACHO or when
optimizing for size), -fasynchronous-unwind-tables and
-maccumulate-outgoing-args by default.
Richard Guenther [Thu, 12 Aug 2010 10:38:05 +0000 (10:38 +0000)]
re PR tree-optimization/45232 (tree reassociation introduces undefined overflow)
2010-08-12 Richard Guenther <rguenther@suse.de>
PR tree-optimization/45232
* tree-ssa-reassoc.c (can_reassociate_p): Disable re-association
for types with undefined overflow.
(reassociate_bb): Allow re-associating of bit and min/max
operations for types with undefined overflow.
* tree-ssa-forwprop.c (associate_plusminus): New function.
(tree_ssa_forward_propagate_single_use_vars): Call it.
Sebastian Pop [Wed, 11 Aug 2010 20:31:24 +0000 (20:31 +0000)]
Fix miscompile of 416.gamess.
2010-07-29 Sebastian Pop <sebastian.pop@amd.com>
* graphite-sese-to-poly.c (rewrite_close_phi_out_of_ssa): Add back
the case removed in the previous patch, when the only phi argument
is defined in the same loop as the phi node itself. Handle it
separately from the invariant case by both propagating it outside
the region and replacing the phi node with an assign.
Sebastian Pop [Wed, 11 Aug 2010 20:31:14 +0000 (20:31 +0000)]
Fix gamess: the only constant phi nodes with one argument are is_gimple_min_invariant and SSA_NAME_IS_DEFAULT_DEF.
2010-07-28 Sebastian Pop <sebastian.pop@amd.com>
* graphite-sese-to-poly.c (rewrite_close_phi_out_of_ssa): The only
constant phi nodes with one argument are is_gimple_min_invariant
and SSA_NAME_IS_DEFAULT_DEF.
Remove cloog_initialize and cloog_finalize when using official CLooG releases.
This requires to initialize the Parma Polyhedra Library by graphite,
as newer CLooG versions are not bound to a PPL backend.
2010-07-27 Andreas Simbuerger <simbuerg@fim.uni-passau.de>
* graphite.c (graphite_initialize): Do not initialize
CLooG and initialize the Parma Polyhedra Library
manually when using CLOOG_ORG.
(graphite_finalize): Do not finalize CLooG and finalize
the Parma Polyhedra Library manually when using CLOOG_ORG.
* graphite-cloog-compat.h (cloog_initialize): Hide function
when using CLOOG_ORG.
(cloog_finalize): Same.
Adapt to new CLAST structure within upstream CLooG version and retain compatibility to CLooG Legacy.
2010-07-27 Andreas Simbuerger <simbuerg@fim.uni-passau.de>
* graphite-clast-to-gimple.c (clast_name_to_gcc): Parameter
type of NAME now depends on used CLooG version.
(clast_to_gcc_expression): Replace expr_* with clast_expr_*.
(gcc_type_for_clast_expr): Same.
(print_clast_stmt): Replace pprint with clast_pprint.
* graphite-cloog-compat.h: Provide compatibility macros for
CLooG Legacy.
(clast_name_p): New.
(clast_expr_term): New.
(clast_expr_red): New.
(clast_expr_bin): New.
(clast_pprint): New.
Sebastian Pop [Wed, 11 Aug 2010 20:30:09 +0000 (20:30 +0000)]
Call scev_reset_htab when Graphite out-of-SSA has changed something.
2010-07-27 Sebastian Pop <sebastian.pop@amd.com>
* graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps): Returns
a bool.
(rewrite_commutative_reductions_out_of_ssa_close_phi): Same.
(rewrite_commutative_reductions_out_of_ssa_loop): Same.
(rewrite_cross_bb_scalar_deps_out_of_ssa): Call scev_reset_htab
when something has been changed.
(rewrite_commutative_reductions_out_of_ssa): Same.
Sebastian Pop [Wed, 11 Aug 2010 20:29:59 +0000 (20:29 +0000)]
Fix tonto: call propagate_expr_outside_region.
2010-07-26 Sebastian Pop <sebastian.pop@amd.com>
* graphite-sese-to-poly.c (propagate_expr_outside_region): Also
handle the case when def is in the sese region.
(rewrite_close_phi_out_of_ssa): Call propagate_expr_outside_region
for invariant expressions.
(rewrite_cross_bb_phi_deps): Removed.
(rewrite_cross_bb_scalar_deps): Also handle GIMPLE_PHI nodes.
(rewrite_cross_bb_scalar_deps_out_of_ssa): Do not call
rewrite_cross_bb_phi_deps.
Sebastian Pop [Wed, 11 Aug 2010 20:28:54 +0000 (20:28 +0000)]
Do not translate out-of-SSA close phi nodes that can be analyzed with scev.
2010-07-20 Sebastian Pop <sebastian.pop@amd.com>
* graphite-sese-to-poly.c (rewrite_close_phi_out_of_ssa): Add an
extra parameter for the region. Call scev_analyzable_p.
(rewrite_reductions_out_of_ssa): Update call to
rewrite_close_phi_out_of_ssa.
(rewrite_cross_bb_phi_deps): Same.
(rewrite_commutative_reductions_out_of_ssa_loop): Add an extra
parameter for the region. Call scev_analyzable_p.
(rewrite_commutative_reductions_out_of_ssa): Update call to
rewrite_commutative_reductions_out_of_ssa_loop.