* Makefile.am (write_entries_to_file_split): New variable.
(write_entries_to_file): Write out that many entries at once,
for speed.
* Makefile.in: Regenerate.
Richard Guenther [Sat, 15 Mar 2008 18:22:26 +0000 (18:22 +0000)]
tree-ssa-ccp.c (ccp_fold): Also read from constant values and fold constant aggregate refs.
2008-03-15 Richard Guenther <rguenther@suse.de>
* tree-ssa-ccp.c (ccp_fold): Also read from constant values
and fold constant aggregate refs.
(fold_const_aggregate_ref): Handle string constants
and constructors in ARRAY_REFs. Handle INDIRECT_REF.
(evaluate_stmt): Simplify now that ccp_fold folds constant
aggregate refs.
* tree-ssa-ccp.c (get_symbol_constant_value): Export.
(fold_const_aggregate_ref): Likewise.
(get_value): Return NULL if we don't have any values.
(ccp_finalize): Set const_val to NULL after freeing it.
* tree-flow.h (get_symbol_constant_value): Declare.
(fold_const_aggregate_ref): Likewise.
* tree-ssa-sccvn.c (try_to_simplify): Use them.
Richard Guenther [Sat, 15 Mar 2008 14:27:55 +0000 (14:27 +0000)]
re PR middle-end/35593 (spurious warning "array subscript is below array bounds" with void* function argument plus -O2)
2008-03-15 Richard Guenther <rguenther@suse.de>
PR middle-end/35593
* tree-ssa-ccp.c (maybe_fold_offset_to_array_ref): Make sure
to not produce negative array indices if not allowed. Add
parameter to indicate that.
(maybe_fold_offset_to_component_ref): Allow negative array
indices only for the first member of a structure.
(maybe_fold_offset_to_reference): Allow negative array indices.
(maybe_fold_stmt_addition): Likewise.
Bjoern Haase [Sat, 15 Mar 2008 12:23:45 +0000 (12:23 +0000)]
avr.c (avr_arch_types): Add avr6 entry.
* gcc/gcc/config/avr/avr.c (avr_arch_types): Add avr6 entry.
(avr_arch): Add ARCH_AVR6.
(avr_mcu_types): Add 'atmega2560' and 'atmega2561' entry.
(initial_elimination_offset): Initialize and use 'avr_pc_size'
instead of fixed value 2.
(print_operand_address): Use gs() asm specifier instead of pm().
(avr_assemble_integer): (Ditto.).
(avr_output_addr_vec_elt): (Ditto.).
(print_operand): Handle "!" code.
* gcc/gcc/config/avr/avr.h (TARGET_CPU_CPP_BUILTINS): Add
__AVR_3_BYTE_PC__, __AVR_2_BYTE_PC__ and __AVR_HAVE_JMP_CALL__.
(AVR_HAVE_EIJMP_EICALL): Define.
(AVR_3_BYTE_PC): Redefine.
(AVR_2_BYTE_PC): (Ditto.).
(PRINT_OPERAND_PUNCT_VALID_P): Add '!' code.
(LINK_SPEC): Add atmega2560 and atmega2561.
(CRT_BINUTILS_SPEC): Add atmega2560 (crtm2560.o) and atmega2561
(crtm2561.o).
* gcc/gcc/config/avr/avr.md (call_insn): Use eicall instead of icall
for 3 byte PC devices.
(call_value_insn): (Ditto.).
(*tablejump_enh): Use eijmp instead of ijmp for 3 byte PC devices.
(indirect_jump): Use only for for 2 byte PC devices.
(*tablejump): (Ditto.).
(*indirect_jump_avr6): Add insn.
(*tablejump_rjmp): Don't use for 3 byte PC devices.
* gcc/gcc/config/avr/libgcc.S (__prologue_saves__): Use eijmp
instead of ijmp for 3 byte PC devices.
(__tablejump2__): (Ditto.).
* gcc/gcc/config/avr/t-avr (MULITLIB_OPTIONS): Add avr6 architecture.
(MULITLIB_DIRNAMES): (Ditto.).
(MULTILIB_MATCHES): Add atmega2560 and atmega2561 to list.
Uros Bizjak [Sat, 15 Mar 2008 11:17:05 +0000 (12:17 +0100)]
mmx.md ("sse2_umulv1siv2di3"): Rename from "sse2_umulsidi3".
* config/i386/mmx.md ("sse2_umulv1siv2di3"): Rename from
"sse2_umulsidi3". Use V1DI mode for operand 0.
("mmx_psadbw"): Use V1DI mode for operand 0.
* config/i386/i386-modes.def (V1SI): New vector mode.
* config/i386/i386.c (struct builtin_description)
[IX86_BUILTIN_PMULUDQ]: Use CODE_FOR_sse2_umulv1siv1di3.
(v1di_ftype_v8qi_v8qi): Rename from di_ftype_v8qi_v8qi.
(v1di_ftype_v2si_v2si): Rename from di_ftype_v2si_v2si.
(ix86_init_mmx_sse_builtins) [__builtin_ia32_psadbw]: Use
v1di_ftype_v8qi_v8qi type.
[__builtin_ia32_pmuludq]: Use v1di_ftype_v2si_v2si type.
Ralf Wildenhues [Sat, 15 Mar 2008 10:22:36 +0000 (10:22 +0000)]
gcc/ada/
* gnat_rm.texi (Implementation Defined Characteristics)
(Wide_Text_IO, Wide_Wide_Text_IO): Add @var annotations where
appropriate.
* gnat_ugn.texi (Wide Character Encodings, Switches for gnatbind)
(Switches for gnatchop, Installing a library): Likewise.
Richard Guenther [Fri, 14 Mar 2008 17:05:48 +0000 (17:05 +0000)]
re PR tree-optimization/34172 (Missed store ccp optimization)
2008-03-14 Richard Guenther <rguenther@suse.de>
PR tree-optimization/34172
* tree-flow.h (refs_may_alias_p): Declare.
(get_single_def_stmt): Likewise.
(get_single_def_stmt_from_phi): Likewise.
(get_single_def_stmt_with_phi): Likewise.
* tree-dfa.c (refs_may_alias_p): New function.
(get_single_def_stmt): Likewise.
(get_single_def_stmt_from_phi): Likewise.
(get_single_def_stmt_with_phi): Likewise.
* tree-ssa-sccvn.c (get_def_ref_stmt_vuses): New function.
(vn_reference_lookup_1): New helper function.
(vn_reference_lookup): Walk the virtual use-def chain to
continue searching for a match if the def does not alias the
reference we are looking for.
Richard Guenther [Fri, 14 Mar 2008 14:52:07 +0000 (14:52 +0000)]
re PR tree-optimization/34043 (Missed optimization causing extra loads and stores when using x86_64 builtin function together with aggregate types.)
2008-03-14 Richard Guenther <rguenther@suse.de>
PR tree-optimization/34043
PR tree-optimization/33989
* tree-ssa-pre.c (execute_pre): Allow SCCVN to do insertion
when doing FRE.
(bitmap_find_leader): Use extra argument to verify dominance
relationship inside a basic-block.
(can_PRE_operation): Add VIEW_CONVERT_EXPR.
(find_leader_in_sets): Adjust.
(create_component_ref_by_pieces): Take extra argument for
dominance check, handle lookup failures.
(find_or_generate_expression): Likewise.
(create_expression_by_pieces): Likewise.
(insert_into_preds_of_block): Adjust.
(create_value_expr_from): If asked for, verify all operands
are in the blocks AVAIL_OUT set.
(make_values_for_stmt): Check for SSA_NAMEs that are life
over an abnormal edge.
(compute_avail): Remove such check.
(do_SCCVN_insertion): New function.
(eliminate): If we do not find a leader suitable for replacement
insert a replacement expression from SCCVN if available.
* tree-ssa-sccvn.h (run_scc_vn): Update prototype.
(struct vn_ssa_aux): Add needs_insertion flag.
* tree-ssa-sccvn.c (may_insert): New global flag.
(copy_reference_ops_from_ref): Value-number union member access
based on its size, not type and member if insertion is allowed.
(visit_reference_op_load): For a weak match from union type
punning lookup a view-converted value and insert a SSA_NAME
for that value if that is not found.
(visit_use): Make dumps shorter. Do not disallow value numbering
SSA_NAMEs that are life over an abnormal edge to constants.
(free_scc_vn): Release inserted SSA_NAMEs.
(run_scc_vn): New flag to specify whether insertion is allowed.
Process SSA_NAMEs in forward order.
* tree-ssa-loop-im.c (for_each_index): Handle invariant
ADDR_EXPRs inside VIEW_CONVERT_EXPR.
* fold-const.c (fold_unary): Fold VIEW_CONVERT_EXPRs from/to
pointer type to/from integral types that do not change the
precision to regular conversions.
Jerry DeLisle [Fri, 14 Mar 2008 14:17:41 +0000 (14:17 +0000)]
re PR testsuite/35406 (gfortran.dg/ldist-1.f90 and gcc.dg/tree-ssa/ldist-4.c don't work)
PR testsuite/35406
* gcc.dg/tree-ssa/ldist-4.c: Update expected scan match
count from 0 to 1 for lp64 targets.
* gfortran.dg/ldist-1.f90: Update test to check
for 5 loops distributed instead of 4 for lp64 targets.
Uros Bizjak [Thu, 13 Mar 2008 09:49:43 +0000 (10:49 +0100)]
re PR c++/35553 (-fkeep-inline-functions and -O errors out in SSE headers)
PR target/35553
* config/i386/i386.h (TARGET_CPU_CPP_BUILTINS): Define
__SSE_USE_INLINED_FUNC__ when using -flag-keep-inline-functions
together with optimization.
* config/i386/xmmintrin.h: Use __SSE_USE_INLINED_FUNC__ instead of
__OPTIMIZE__ to choose between inlined intrinsic SSE function having
immediate arguments and its equivalent macro definition.
* config/i386/bmintrin.h: Ditto.
* config/i386/smmintrin.h: Ditto.
* config/i386/tmmintrin.h: Ditto.
* config/i386/mmintrin-common.h: Ditto.
* config/i386/ammintrin.h: Ditto.
* config/i386/emmintrin.h: Ditto.
testsuite/ChangeLog:
PR target/35553
* g++.dg/other/i386-3.C: New test.
Danny Smith [Thu, 13 Mar 2008 07:06:30 +0000 (07:06 +0000)]
re PR target/35054 (No documentation of #pragma push_macro("macro_name"))
PR 35054
* doc/extend.texi (Structure-Packing Pragmas): Replace "Win32"
with the phrase "Microsoft Windows compilers".
(Push/Pop Macro Pragmas): New subsection. Document
#pragma push_macro and pragma pop_macro.
Paul Brook [Wed, 12 Mar 2008 23:19:41 +0000 (23:19 +0000)]
thumb2.md: Extend peephole to cover 3-arg subs.
2008-03-12 Paul Brook <paul@codesourcery.com>
gcc/
* config/arm/thumb2.md: Extend peephole to cover 3-arg subs.
(thumb2_alusi3_short): Exclude PLUS and MINUS.
(thumb2_addsi_shortim): Rename ...
(thumb2_addsi_short): ... to this. Allow register operands.
(thumb2_subsi_short): New pattern.
(thumb2_one_cmplsi2_short,
thumb2_negsi2_short): New patterns and peepholes.
Richard Guenther [Wed, 12 Mar 2008 14:25:48 +0000 (14:25 +0000)]
revert: [multiple changes]
2008-03-12 Richard Guenther <rguenther@suse.de>
PR c++/35469
Revert:
2008-02-04 Richard Guenther <rguenther@suse.de>
PR java/35035
* decl.c (record_builtin_java_type): Make jboolean a
integer type again where its mode doesn't match that of bool.
2008-01-25 Richard Guenther <rguenther@suse.de>
PR c++/33887
* decl.c (record_builtin_java_type): Make __java_boolean
a variant of bool.
* typeck.c (structural_comptypes): Move TYPE_FOR_JAVA check
after TYPE_MAIN_VARIANT check.
Jakub Jelinek [Wed, 12 Mar 2008 09:55:48 +0000 (10:55 +0100)]
re PR middle-end/35549 (Invalid use of copy-in/out for shared vars in nested parallels)
PR middle-end/35549
* omp-low.c (maybe_lookup_decl): Constify first argument.
(use_pointer_for_field): Change last argument from bool to
omp_context *. Disallow shared copy-in/out in nested
parallel if decl is shared in outer parallel too.
(build_outer_var_ref, scan_sharing_clauses,
lower_rec_input_clauses, lower_copyprivate_clauses,
lower_send_clauses, lower_send_shared_vars): Adjust callers.
pa.h (TARGET_LONG_PIC_SDIFF_CALL): Conditionalize define on TARGET_HPUX.
* pa.h (TARGET_LONG_PIC_SDIFF_CALL): Conditionalize define on
TARGET_HPUX. Revise comment.
(TARGET_LONG_PIC_PCREL_CALL): Revise comment.
* pa.c (output_call): Update for revised TARGET_LONG_PIC_SDIFF_CALL.
Use sr4 variant of `be' instruction when not generating PIC code.
(attr_length_call): Adjust for above change.
Uros Bizjak [Tue, 11 Mar 2008 19:18:48 +0000 (20:18 +0100)]
re PR target/35540 (Segmentation fault with __builtin_parity() and -O1)
PR target/35540
* config/i386/i386.md (paritysi2, paritydi2): Use register_operand
constraint for operand 1.
(paritysi2_cmp): Use register_operand constraint for operand 2.
Use earlyclobber modifier for operand 1. Remove support for
memory operands.
(paritydi2_cmp): Use register_operand constraint for operand 3.
Use earlyclobber modifier for operand 1. Remove support for
memory operands.
testsuite/ChangeLog:
PR target/35540
* gcc.target/i386/pr35540.c: New test.
Paul Brook [Tue, 11 Mar 2008 17:43:22 +0000 (17:43 +0000)]
arm.c (use_return_insn): Check TARGET_APCS_FRAME.
2008-03-11 Paul Brook <paul@codesourcery.com>
Vladimir Prus <vladimir@codesourcery.com>
* config/arm/arm.c (use_return_insn): Check TARGET_APCS_FRAME.
(arm_compute_save_reg0_reg12_mask): Always
check if register 11 must be saved. Always safe hard frame pointer
when frame_pointer_needeed.
(arm_compute_save_reg_mask): Save IP and PC
only with apcs frames.
(arm_output_epilogue): Adjust Thumb2 codepath to
be also invoked and work for ARM non-apcs frames.
(arm_expand_prologue): Don't bother saving IP
for non-apcs frame, since it's not clobbered by
prologue code. Implement non-apcs frame
layout.
re PR libfortran/32812 (random_seed and date_and_time)
PR libfortran/32812
* intrinsics/random.c (scramble_seed, unscramble_seed): New
functions.
(random_seed_i4): Scramble the seed the user gives us before
storing it, and unscramble it when we return it back later.
Andrew Pinski [Tue, 11 Mar 2008 09:36:51 +0000 (09:36 +0000)]
re PR tree-optimization/31358 (IV-OPTs produces a weird MEM_REF)
2008-03-11 Andrew Pinski <andrew_pinski@playstation.sony.com>
Richard Guenther <rguenther@suse.de>
PR tree-optimization/31358
* tree-ssa-loop-manip.c (create_iv): Call force_gimple_operand for
the step with a NULL_TREE.
* tree-ssa-loop-ivopts.c (find_bivs): Convert the step
to sizetype if type is a pointer type.
(add_candidate_1): Don't convert the base and step to
the generic type if the orginal type is a pointer type.
(add_iv_value_candidates): Use sizetype for the step
if type is a pointer type.
(cand_value_at): Likewise.
* tree-ssa-address.c (add_to_parts): Use POINTER_PLUS_EXPR
for pointer types.
* tree-affine.c (tree_to_aff_combination <POINTER_PLUS_EXPR>):
Don't convert the tem affine to the type.
(add_elt_to_tree): Use sizetype for the step if a pointer.
Use POINTER_PLUS_EXPR for pointers.
(aff_combination_to_tree): Use sizetype for the step if a
pointer.
Co-Authored-By: Richard Guenther <rguenther@suse.de>
From-SVN: r133102
Jim Meyering [Mon, 10 Mar 2008 22:08:34 +0000 (22:08 +0000)]
Don't leak upon failed realloc.
libjava
2008-03-10 Jim Meyering <meyering@redhat.com>
Don't leak upon failed realloc.
* gnu/classpath/natSystemProperties.cc
(SystemProperties::insertSystemProperties):
libjava/classpath
2008-03-10 Jim Meyering <meyering@redhat.com>
Don't leak upon failed realloc.
* native/jni/classpath/jcl.c (JCL_realloc): Upon failed realloc,
free the original buffer before throwing the exception.
Janis Johnson [Mon, 10 Mar 2008 21:33:09 +0000 (21:33 +0000)]
patch_tester.sh (initialization): Initialize svnpatch and stop.
2008-03-10 Janis Johnson <janis187@us.ibm.com>
* patch_tester.sh (initialization): Initialize svnpatch and stop.
(usage): Add -svnpath and -stop.
(makedir): New.
(argument handling): Process -stop and -svnpath.
(setup code): Use makedir, error out if initial svn checkout fails.
(update): Use svnpath. Invoke contrib/gcc_update.
(apply_patch): Require that patch was created at top level. Use eval
with option variables. Don't use bootstrap target for make. Verify
that some tests were run.
(bootntest_patched): Use snvpath.
(main loop): For -stop, exit when there are no more patches to test.
Richard Guenther [Mon, 10 Mar 2008 17:14:45 +0000 (17:14 +0000)]
re PR tree-optimization/34677 (PREs insert_fake_stores is mostly useless)
2008-03-10 Richard Guenther <rguenther@suse.de>
PR tree-optimization/34677
* tree-ssa-pre.c (modify_expr_node_pool): Remove.
(poolify_tree): Likewise.
(modify_expr_template): Likewise.
(poolify_modify_stmt): Likewise.
(insert_fake_stores): Handle all component-ref style stores
in addition to INDIRECT_REF. Also handle complex types.
Do not poolify the inserted load.
(realify_fake_stores): Do not rebuild the tree but only
make it a SSA_NAME copy.
(init_pre): Remove initialzation of modify_expr_template.
Do not allocate modify_expr_node_pool.
(fini_pre): Do not free modify_expr_node_pool.
* gcc.dg/tree-ssa/loadpre23.c: New testcase.
* gcc.dg/tree-ssa/loadpre24.c: Likewise.
* gcc.dg/tree-ssa/loadpre25.c: Likewise.
projects / thirdparty / gcc.git / log
