1 2013-08-29 Xinliang David Li <davidxl@google.com>
3 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
4 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
5 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6 Delay base decl alignment adjustment.
7 * tree-vectorizer.c (vect_destroy_datarefs): New function.
8 * tree-vectorizer.h: New data structure.
9 (set_dr_misalignment): New function.
10 (dr_misalignment): Ditto.
11 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
12 (vectorizable_load): Ditto.
13 (ensure_base_align): New function.
14 (vectorize_loops): Add dbg_cnt support.
15 (execute_vect_slp): Ditto.
16 * dbgcnt.def: New debug counter.
17 * Makefile: New dependency.
19 2013-09-03 Meador Inge <meadori@codesourcery.com>
23 2013-08-30 Meador Inge <meadori@codesourcery.com>
25 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
27 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
29 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
32 2013-09-03 Richard Biener <rguenther@suse.de>
34 * tree-affine.c (add_elt_to_tree): Fix association issue,
35 avoid useless converts and make sure to always return a
36 properly typed result.
38 2013-09-03 Richard Biener <rguenther@suse.de>
41 * fold-const.c (negate_expr_p): Fix division case.
42 (negate_expr): Likewise.
44 2013-09-03 Richard Biener <rguenther@suse.de>
47 * tree-streamer-out.c: Include tm.h.
48 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
50 2013-09-03 Jan Hubicka <jh@suse.cz>
52 * tree-profile.c (tree_profiling): Cleanup CFG when done.
54 2013-09-03 Alan Modra <amodra@gmail.com>
56 * config.gcc (powerpc*-*-linux*): Add support for little-endian
57 multilibs to big-endian target and vice versa.
58 * config/rs6000/t-linux64: Use := assignment on all vars.
59 (MULTILIB_EXTRA_OPTS): Remove fPIC.
60 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
61 * config/rs6000/t-linux64le: New file.
62 * config/rs6000/t-linux64bele: New file.
63 * config/rs6000/t-linux64lebe: New file.
65 2013-09-02 Jan Hubicka <jh@suse.cz>
67 * ipa-inline-transform.c (inline_transform): Do not
68 optimize_inline_calls when not optimizing.
70 2013-09-02 Jan Hubicka <jh@suse.cz>
72 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
73 duplicated nodes for assembler names.
74 * symtab.c (symtab_unregister_node): Do not attempt to unlink
75 hard registers from assembler name hash.
77 2013-09-02 Jan Hubicka <jh@suse.cz>
79 * ipa-split.c (execute_split_functions): Split externally visible
80 functions called once.
82 2013-09-02 Martin Jambor <mjambor@suse.cz>
85 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
86 linked list. When finding the correct duplicate, also consider also
87 the caller in additon to its inlined_to node.
89 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
91 * config/aarch64/aarch64-simd-builtins.def
92 (dup_lane_scalar): Remove.
93 * config/aarch64/aarch64-simd.md
94 (aarch64_simd_dup): Add 'w->w' alternative.
95 (aarch64_dup_lane<mode>): Allow for VALL.
96 (aarch64_dup_lane_scalar<mode>): Remove.
97 (aarch64_dup_lane_<vswap_width_name><mode>): New.
98 (aarch64_get_lane_signed<mode>): Add w->w altenative.
99 (aarch64_get_lane_unsigned<mode>): Likewise.
100 (aarch64_get_lane<mode>): Likewise.
101 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
102 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
103 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
104 (VCON): Change container of V2SF.
105 (vswap_width_name): Likewise.
106 * config/aarch64/arm_neon.h
107 (__aarch64_vdup_lane_any): New.
108 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
109 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
110 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
112 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
115 * expr.c (emit_move_complex): Do not move complex FP values as parts if
116 the source or the destination is a single hard register.
118 2013-09-02 Richard Biener <rguenther@suse.de>
121 * tree-scalar-evolution.c (instantiate_scev_name): Allow
124 2013-09-02 Richard Biener <rguenther@suse.de>
126 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
127 arithmetic to sizetype.
129 2013-09-02 Bin Cheng <bin.cheng@arm.com>
131 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
132 Find auto-increment use both before and after candidate.
134 2013-09-02 Marek Polacek <polacek@redhat.com>
136 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
138 2013-09-01 Jan Hubicka <jh@suse.cz>
140 * Makefile.in: Add ipa-profile.o
141 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
142 * cgraph.c (struct cgraph_propagate_frequency_data,
143 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
144 ipa-profile.c; replace cgraph_ by ipa_ prefix.
145 * cgraph.h (cgraph_propagate_frequency): Remove.
146 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
147 (inline_update_callee_summaries): Update.
148 * ipa-profile.c: New file.
149 * ipa-utils.h (ipa_propagate_frequency): Declare.
150 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
151 data-streamer.h, value-prof.h
152 (symtab_remove_unreachable_nodes): Update profile.
153 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
154 account_time_size, cmp_counts, dump_histogram,
155 ipa_profile_generate_summary, ipa_profile_write_summary,
156 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
157 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
158 Move to ipa-profile.c
160 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
162 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
164 2013-09-01 Jan Hubicka <jh@suse.cz>
166 * common.opt (fdevirtualize-speculatively): New function.
167 * invoke.texi (fdevirtualize-speculatively): Document.
168 * ipa-devirt.c: Include ipa-inline.h
169 (likely_target_p): New function.
170 (ipa_devirt): New function.
171 (gate_ipa_devirt): New function.
172 (pass_data_ipa_devirt): New static var.
173 (pass_ipa_devirt): Likewise.
174 (make_pass_ipa_devirt): New function.
175 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
176 (common_handle_option): Disable devirtualization when
177 value range profiling is available.
178 * passes.def (pass_ipa_devirt): Add.
179 * timever.def (TV_IPA_DEVIRT): New timevar.
180 * tree-pass.h (make_pass_ipa_devirt):
182 2013-09-01 Iain Sandoe <iain@codesourcery.com>
184 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
185 include sanitize(undefined).
187 2013-08-31 Diego Novillo <dnovillo@google.com>
189 * Makefile.in (TREE_CORE_H): Define.
191 (GTFILES): Add tree-core.h.
192 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
194 * tree-core.h: New file.
195 Move all data structures, enum, typedefs, global
196 declarations and constants from ...
199 2013-08-31 Jan Hubicka <jh@suse.cz>
201 * bulitins.c (expand_builtin): Do not early exit for gcov
202 instrumented functions.
204 2013-08-31 Marek Polacek <polacek@redhat.com>
206 * ubsan.c: Include tm_p.h.
208 2013-08-31 Jan Hubicka <jh@suse.cz>
210 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
213 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
214 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
215 * tree-cfg.c (verify_gimple_label): ... here.
216 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
217 (ipa_merge_profiles): New function.
218 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
219 (lto_input_function_body): Likewise.
220 * ipa-utils.h (ipa_merge_profiles): Declare.
221 * lto-streamer.h (lto_input_function_body): Update prototype.
222 (emit_label_in_global_context_p): Remove.
223 * lto-symtab.c: Include ipa-utils.h
224 (lto_cgraph_replace_node): Use ipa_merge_profiles.
226 2013-08-31 Jan Hubicka <jh@suse.cz>
228 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
230 2013-08-31 Jan Hubicka <jh@suse.cz>
232 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
234 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
236 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
237 "cmp" RTX before signed_comparison_operator check to account
240 2013-08-30 Jan Hubicka <jh@suse.cz>
242 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
243 (detect_type_change_1): Rename to ...
244 (detect_type_change): ... this one; early return on non-polymorphic
246 (detect_type_change_ssa): Add comp_type parameter; update
247 use of detect_type_change.
248 (compute_complex_assign_jump_func): Add param_type parameter;
249 update use of detect_type_change_ssa.
250 (compute_complex_ancestor_jump_func): Likewise.
251 (ipa_get_callee_param_type): New function.
252 (ipa_compute_jump_functions_for_edge): Compute parameter type;
253 update calls to the jump function computation functions.
255 2013-08-30 Teresa Johnson <tejohnson@google.com>
256 Steven Bosscher <steven@gcc.gnu.org>
258 * cfgrtl.c (fixup_new_cold_bb): New routine.
259 (commit_edge_insertions): Invoke fixup_partitions.
260 (find_partition_fixes): New routine.
261 (fixup_partitions): Ditto.
262 (verify_hot_cold_block_grouping): Update comments.
263 (rtl_verify_edges): Invoke find_partition_fixes.
264 (rtl_verify_bb_pointers): Update comments.
265 (rtl_verify_bb_layout): Ditto.
266 * basic-block.h (probably_never_executed_edge_p): Declare.
267 (fixup_partitions): Ditto.
268 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
269 * bb-reorder.c (sanitize_hot_paths): New function.
270 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
272 * predict.c (probably_never_executed_edge_p): New routine.
273 * cfg.c (check_bb_profile): Add partition insanity warnings.
275 2013-08-30 Meador Inge <meadori@codesourcery.com>
277 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
279 2013-08-30 Marek Polacek <polacek@redhat.com>
281 * Makefile.in (ubsan.o): Add.
282 (c-family/c-ubsan.o): Add.
283 (builtins.o): Add ubsan.h dependency.
286 * common.opt: Add -fsanitize=undefined option.
287 (flag_sanitize): Add variable.
288 (fsanitize=): Add option. Add Driver.
289 (fsanitize=thread): Remove option.
290 (fsanitize=address): Likewise.
291 (static-libubsan): New option.
292 * doc/invoke.texi: Document the new flag and -static-libubsan.
293 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
294 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
295 * builtin-attrs.def (ATTR_COLD): Define.
296 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
297 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
298 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
299 * flag-types.h (sanitize_code): New enum.
300 * opts.c (common_handle_option): Parse command line arguments
301 of -fsanitize=. Add -fsanitize=unreachable option.
302 * varasm.c (get_variable_section): Adjust.
303 (assemble_noswitch_variable): Likewise.
304 (assemble_variable): Likewise.
305 (output_constant_def_contents): Likewise.
306 (categorize_decl_for_section): Likewise.
307 (place_block_symbol): Likewise.
308 (output_object_block): Likewise.
309 * builtins.def: Likewise.
310 * toplev.c (compile_file): Likewise.
311 (process_options): Likewise.
312 * cppbuiltin.c: Likewise.
313 * tsan.c (tsan_pass): Likewise.
314 (tsan_gate): Likewise.
315 (tsan_gate_O0): Likewise.
316 * cfgexpand.c (partition_stack_vars): Likewise.
317 (expand_stack_vars): Likewise.
318 (defer_stack_allocation): Likewise.
319 (expand_used_vars): Likewise.
320 * cfgcleanup.c (old_insns_match_p): Likewise.
321 * asan.c (asan_finish_file): Likewise.
322 (asan_instrument): Likewise.
323 (gate_asan): Likewise.
324 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
325 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
326 (asan_global_struct): Use pointer_sized_int_node instead
327 calling build_nonstandard_integer_type.
328 (initialize_sanitizer_builtins): Likewise.
329 (asan_finish_file): Likewise.
330 * gcc.c: Document %{%:function(args):X}.
331 (static_spec_functions): Add sanitize.
332 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
333 store funcval != NULL there.
334 (do_spec_1): Adjust handle_spec_function caller.
335 (handle_braces): Allow %:function(args) as condition.
336 (sanitize_spec_function): New function.
337 (ADD_STATIC_LIBUBSAN_LIBS): Define.
338 (LIBUBSAN_SPEC): Likewise.
339 (LIBUBSAN_EARLY_SPEC): Likewise.
340 (SANITIZER_SPEC): Handle libubsan.
341 (SANITIZER_EARLY_SPEC): Likewise.
342 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
343 instead of fsanitize=address.
344 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
345 instead of fsanitize=address*.
346 * builtins.c: Include ubsan.h.
347 (fold_builtin_0): Instrument __builtin_unreachable.
348 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
349 instead of flag_asan.
350 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
351 (pointer_sized_int_node): Define.
352 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
354 2013-08-30 Mike Stump <mikestump@comcast.net>
356 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
359 2013-08-29 Jan Hubicka <jh@suse.cz>
361 * cgraph.c (cgraph_function_body_availability): Handle weakref
363 * passes.def: Remove pass_fixup_cfg.
364 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
365 track when we need to remove functions.
366 (gate_ipa_inline): Execute inlining always; add comment why.
367 (pass_data_ipa_inline): Remove TODO_remove_functions.
368 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
369 do not produce summaries.
370 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
371 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
372 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
373 constant pool and vtable.
375 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
377 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
378 New arm_neon.h's internal macros to specify 64-bit constants.
379 Avoid using stdint.h's macros.
381 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
383 * recog.c (verify_changes): Verify that changes[i].old is non-zero
384 before applying REG_P.
386 2013-08-30 Jakub Jelinek <jakub@redhat.com>
388 PR tree-optimization/58277
389 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
390 after seeing too many stmts with vdef in between dombb and current
391 bb, invalidate everything.
393 2013-08-30 Richard Biener <rguenther@suse.de>
395 * fold-const.c (fold_single_bit_test): Fix overflow test.
397 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
399 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
400 and which can live in a register, always retrieve the value on entry.
401 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
402 passed by invisible reference specially.
403 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
404 (vt_add_function_parameter): Correctly deal with a parameter passed by
407 2013-08-30 Jan Hubicka <jh@suse.cz>
409 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
411 2013-08-30 Richard Biener <rguenther@suse.de>
413 PR tree-optimization/58228
414 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
415 allow invariant loads in nested loop vectorization.
417 2013-08-30 Richard Biener <rguenther@suse.de>
419 PR tree-optimization/58223
420 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
421 (has_anti_or_output_dependence): ... this and adjust to also
422 look for output dependences.
423 (mark_nodes_having_upstream_mem_writes): Adjust.
424 (rdg_flag_uses): Likewise.
426 2013-08-30 Richard Biener <rguenther@suse.de>
428 PR tree-optimization/58010
429 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
430 assert that we have a loop-closed PHI.
432 2013-08-29 Jan Hubicka <jh@suse.cz>
434 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
435 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
436 * lto-section-in.c (lto_free_function_in_decl_state): New function.
437 (lto_free_function_in_decl_state_for_node): New function.
439 2013-08-29 Xinliang David Li <davidxl@google.com>
441 * loop-unroll.c (report_unroll_peel): Minor message change.
442 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
443 Emit alignment peeling message with default -fopt-info.
444 (vect_loop_versioning): Emit loop version info message.
445 * tree-vectorizer.c (vectorize_loops): Minor message change.
446 (execute_vect_slp): Ditto.
448 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
450 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
451 of the clone from the DECL_NAME of the original function.
453 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
455 * passes.c (register_pass): Add overload.
456 * tree-pass.h (register_pass): Forward declare it. Add comment.
458 2013-08-29 Jan Hubicka <jh@suse.cz>
460 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
461 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
462 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
463 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
464 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
465 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
466 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
467 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
469 2013-08-29 Teresa Johnson <tejohnson@google.com>
471 * dumpfile.c (dump_loc): Output column number.
472 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
473 * doc/invoke.texi: Document optall -fopt-info flag.
474 * profile.c (read_profile_edge_counts): Use new dump framework.
475 (compute_branch_probabilities): Ditto.
476 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
477 when pass not in any opt group.
478 * pass_manager.h (pass_manager::get_pass_profile): New method.
479 * value-prof.c (check_counter): Use new dump framework.
480 (check_ic_target): Ditto.
481 * coverage.c (get_coverage_counts): Ditto.
482 (coverage_init): Setup new dump framework.
484 2013-08-29 Richard Biener <rguenther@suse.de>
486 PR tree-optimization/58246
487 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
488 handle the dominance check inside a basic-block.
490 2013-08-29 Richard Biener <rguenther@suse.de>
493 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
494 of default defs that appear in abnormal PHI nodes.
496 2013-08-29 Richard Biener <rguenther@suse.de>
498 PR tree-optimization/57685
499 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
500 single-use operands to avoid exponential complexity.
502 2013-08-28 Dehao Chen <dehao@google.com>
504 * ipa-inline.c (edge_badness): Fix integer underflow.
506 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
508 * gtm-builtins.def (_ITM_free): Declare leaf.
510 2013-08-28 Jakub Jelinek <jakub@redhat.com>
513 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
514 (*tls_local_dynamic_base_64_largepic): Likewise.
515 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
516 Remove predicate from call operand.
517 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
518 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
520 2013-08-28 Jeff Law <law@redhat.com>
522 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
523 checks for the number of predecessors and successors allowed.
524 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
525 which require copying a joiner block if there is a request which
526 is a subpath that requires no joiner block copying.
528 2013-08-28 Jan Hubicka <jh@suse.cz>
530 * lto-streamer-out.c (DFS_write_tree_body): Drop
531 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
532 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
533 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
534 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
535 Do not read DECL_ERROR_ISSUED.
536 (unpack_ts_decl_with_vis_value_fields): Do not read
538 (lto_input_ts_binfo_tree_pointers): Do not read
539 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
540 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
541 write DECL_ERROR_ISSUED..
542 (pack_ts_decl_with_vis_value_fields): Do not write
544 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
545 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
546 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
547 * tree.h (tree_decl_common): Update comment.
548 (DECL_ERROR_ISSUED): Remove.
550 2013-08-28 Jakub Jelinek <jakub@redhat.com>
553 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
555 2013-08-28 Jan Hubicka <jh@suse.cz>
557 * builtins.def (free): Declare leaf.
559 2013-08-27 David Malcolm <dmalcolm@redhat.com>
562 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
563 * configure: Regenerate.
565 2013-08-27 Martin Jambor <mjambor@suse.cz>
567 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
568 (ipa_ancestor_jf_data): Likewise.
569 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
570 (ipa_get_jf_pass_through_type_preserved): New function.
571 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
572 (ipa_get_jf_ancestor_type_preserved): New function.
573 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
574 (ipa_get_jf_ancestor_result): Likewise.
575 (propagate_vals_accross_pass_through): Use
576 ipa_get_jf_pass_through_result to do all the value mappings.
577 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
579 (ipa_set_jf_cst_copy): New function.
580 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
581 (ipa_set_jf_arith_pass_through): Likewise.
582 (ipa_set_ancestor_jf): Likewise.
583 (compute_complex_assign_jump_func): Set type_preserved instead of
585 (ipa_compute_jump_functions_for_edge): Likewise.
586 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
587 (update_jump_functions_after_inlining): Update type_preserved.
588 Explicitely create jump functions when combining one with pass_through.
589 (ipa_write_jump_function): Stream the type_preserved flags.
590 (ipa_read_jump_function): Likewise.
592 2013-08-27 Jakub Jelinek <jakub@redhat.com>
593 Aldy Hernandez <aldyh@redhat.com>
595 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
596 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
597 * function.h (struct function): Add has_force_vect_loops and
599 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
600 * gimple.c (gimple_build_omp_critical): Add KIND argument and
602 * gimple.def: Update CLAUSES comments.
603 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
604 (gimple_build_omp_for): Add argument to prototype.
605 (gimple_omp_for_kind): New.
606 (gimple_omp_for_set_kind): New.
607 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
608 GOVD_DATA_SHARE_CLASS.
609 (enum omp_region_type): Add ORT_SIMD.
610 (gimple_add_tmp_var): Handle ORT_SIMD.
611 (gimplify_var_or_parm_decl): Same.
612 (is_gimple_stmt): Same.
613 (omp_firstprivatize_variable): Same.
614 (omp_add_variable): Only use splay_tree_insert if lookup failed.
615 (omp_notice_variable): Handle ORT_SIMD.
616 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
617 (omp_check_private): Handle ORT_SIMD.
618 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
620 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
621 Handle OMP_CLAUSE_LASTPRIVATE.
622 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
624 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
625 (gimplify_expr): Handle OMP_SIMD.
626 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
627 (expand_GOMP_SIMD_VF): New.
628 (expand_GOMP_SIMD_LAST_LANE): New.
629 * internal-fn.def (GOMP_SIMD_LANE): New.
631 (GOMP_SIMD_LAST_LANE): New.
632 * omp-low.c: Include target.h.
633 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
635 (check_omp_nesting_restrictions): Same.
637 (lower_rec_simd_input_clauses): New.
638 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
640 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
641 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
642 (expand_omp_build_assign): New.
643 (expand_omp_for_init_counts): New.
644 (expand_omp_for_init_vars): New.
645 (extract_omp_for_update_vars): New.
646 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
647 and rewrite accordingly.
648 (expand_omp_simd): New.
649 (expand_omp_for): Use expand_omp_simd.
650 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
651 (lower_omp_for): Do not lower the body.
652 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
654 * tree-flow.h (find_omp_clause): Remove prototype.
655 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
656 forcing vectorization of the loop, or if flag_tree_vectorize.
657 (gate_tree_if_conversion): Similarly.
658 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
659 gimple_build_omp_for.
660 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
661 * tree-parloops (create_parallel_loop): Pass kind argument to
662 gimple_build_omp_for.
663 * tree-pretty-print.c (dump_omp_clause): Add cases for
664 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
665 OMP_CLAUSE__SIMDUID_.
666 (dump_generic_node): Handle OMP_SIMD.
667 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
668 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
669 unroll OMP_SIMD loops here.
670 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
671 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
673 (vect_analyze_data_refs): Handle simd loops.
674 * tree-vect-loop.c (vectorizable_live_operation): Handle
676 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
677 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
678 (vectorizable_load): Same.
679 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
680 (struct simduid_to_vf): New.
681 (simduid_to_vf::hash): New.
682 (simduid_to-vf::equal): New.
683 (struct simd_array_to_simduid): New.
684 (simd_array_to_simduid::hash): New.
685 (simd_array_to_simduid::equal): New.
686 (adjust_simduid_builtins): New.
687 (struct note_simd_array_uses_struct): New.
688 (note_simd_array_uses_cb): New.
689 (note_simd_array_uses): New.
690 (vectorize_loops): Handle simd hints and adjust simd builtins
692 * tree-vectorizer.h (struct _stmt_vec_info): Add
693 simd_lane_access_p field.
694 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
695 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
696 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
697 (omp_clause_code_name): Same.
698 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
699 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
700 * tree.def (OMP_SIMD): New entry.
701 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
702 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
703 (OMP_CLAUSE_DECL): Adjust range for new clauses.
704 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
705 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
706 (OMP_CLAUSE_LINEAR_STEP): New.
707 (OMP_CLAUSE_SAFELEN_EXPR): New.
708 (OMP_CLAUSE__SIMDUID__DECL): New.
709 (find_omp_clause): New prototype.
711 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
713 * config/i386/driver-i386.c (host_detect_local_cpu): Update
714 Haswell processor detection.
716 2013-08-27 Christian Widmer <shadow@umbrox.de>
719 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
720 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
721 AVX2 capable processors.
723 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
725 * config/aarch64/arm_neon.h: Replace all inline asm implementations
726 of vget_low_* with implementations in terms of other intrinsics.
728 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
731 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
732 values to -1, 0 and 1.
734 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
736 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
737 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
738 (<optab><mode>3_insn): Remove.
739 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
740 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
741 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
742 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
743 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
744 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
745 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
746 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
747 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
748 (ror<mode>3_insn): Likewise.
749 * config/aarch64/predicates.md (aarch64_simd_register): New.
751 2013-08-27 Richard Biener <rguenther@suse.de>
753 PR tree-optimization/57521
754 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
755 one edge is non-critical.
756 (find_phi_replacement_condition): Make sure to use a non-critical
757 edge. Cleanup and remove old bug workarounds.
758 (bb_postdominates_preds): Remove.
759 (if_convertible_loop_p_1): Do not compute post-dominators.
760 (combine_blocks): Do not free post-dominators.
761 (main_tree_if_conversion): Likewise.
762 (pass_data_if_conversion): Add TODO_verify_ssa.
764 2013-08-27 DJ Delorie <dj@redhat.com>
766 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
768 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
770 * function.c (assign_parm_find_data_types): Set passed_mode and
771 nominal_mode to the TYPE_MODE of nominal_type for the built
772 pointer type in case of the struct-pass-by-reference.
774 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
776 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
777 (UINT16_TYPE): Change default to "unsigned int".
779 * config/avr/avr.opt (mfract-convert-truncate): New option.
780 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
781 is set, round negative fractional integers according to n1169
782 when converting to integer types.
784 2013-08-26 Jan Hubicka <jh@suse.cz>
786 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
787 methods can not be called indirectly when their address is not taken.
789 2013-08-26 Jan Hubicka <jh@suse.cz>
791 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
794 2013-08-26 Jan Hubicka <jh@suse.cz>
796 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
799 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
801 * reload.c (find_valid_class): Allow classes that do not include
802 FIRST_PSEUDO_REGISTER - 1.
804 2013-08-26 Jan Hubicka <jh@suse.cz>
806 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
807 fix edge count/frequency when speculation failed; fix type check
810 2013-08-26 Jan Hubicka <jh@suse.cz>
812 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
814 2013-08-26 Jan Hubicka <jh@suse.cz>
816 * ipa-inline-transform.c (inline_transform): Be ready for basic block
817 to be changed by edge redirection.
819 2013-08-26 Jan Hubicka <jh@suse.cz>
821 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
822 formating; add sanity check.
823 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
824 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
825 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
826 (resolve_noninline_speculation): Update callee keys, too.
828 2013-08-26 Jan Hubicka <jh@suse.cz>
830 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
831 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
833 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
835 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
838 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
840 * config/i386/i386.c (ix86_debug_options): Remove prototype.
841 (x86_64_elf_select_section): Ditto.
842 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
844 (ix86_pass_by_reference): Ditto.
845 (output_set_got): Ditto.
846 (ix86_unary_operator_ok): Ditto.
847 (ix86_expand_builtin): Ditto.
849 2013-08-23 Jan Hubicka <jh@suse.cz>
851 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
853 2013-08-23 Jan Hubicka <jh@suse.cz>
855 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
856 (tree_decl_with_vis): Add FINAL field.
858 2013-08-23 Jeff Law <law@redhat.com>
860 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
861 the debugging dump when the expression is fully redundant.
863 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
865 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
866 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
867 * pretty-print.c (pp_formatted_text_data): Likewise.
868 (pp_write_text_to_stream): Likewise.
869 (pp_write_text_as_dot_label_to_stream): Likewise.
870 (pp_append_r): Likewise.
871 (pp_format): Likewise.
872 (pp_flush): Likewise.
873 (pp_clear_output_area): Likewise.
874 (pp_append_text): Likewise.
875 (pp_formatted_text): Likewise.
876 (pp_remaining_character_count_for_line): Likewise.
877 (pp_newline): Likewise.
878 (pp_character): Likewise.
879 (output_buffer::~output_buffer): Define.
880 (pretty_printer::~pretty_printer): Destruct output buffer.
881 * pretty-print.h (output_buffer::~output_buffer): Declare.
882 (pretty_printer::~pretty_printer): Declare virtual.
884 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
887 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
888 HOST_WIDE_INT_M1U): New macros.
889 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
890 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
891 unsigned -1 for lshift.
892 * cse.c (cse_insn): Likewise.
893 * double-int.c (rshift_double, lshift_double): Likewise.
894 * builtins.c (fold_builtin_bitop): Likewise.
895 * combine.c (force_to_mode): Likewise.
896 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
897 * simplify-rtx.c (simplify_const_unary_operation,
898 simplify_const_binary_operation): Likewise.
899 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
900 check_va_list_escapes): Likewise.
901 * rtlanal.c (nonzero_bits1): Likewise.
902 * expmed.c (expand_smod_pow2): Likewise.
903 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
905 2013-08-23 Jan Hubicka <jh@suse.cz>
907 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
908 as having address taken.
910 2013-08-23 Jan Hubicka <jh@suse.cz>
912 * ipa-utils.h (method_class_type): Declare.
913 * ipa-devirt.c (method_class_type): Export.
915 * cgraphunit.c (analyze_functions): Do basic devirtualization;
916 do not walk base classes of anonymous types.
918 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
920 PR rtl-optimization/58220
922 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
923 handle SEQUENCE insns properly.
925 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
927 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
929 (pp_newline_and_indent): Likewise.
930 (pp_separate_with): Likewise.
931 * pretty-print.c (pp_newline_and_flush): Define.
932 (pp_newline_and_indent): Likewise.
933 (pp_separate_with): Likewise.
935 2013-08-23 Jakub Jelinek <jakub@redhat.com>
938 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
939 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
941 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
943 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
944 * gcc/config/i386/i386.md (*movti_internal): Use
945 predicate to determine if EVEX is needed.
946 (*movsi_internal): Ditto.
947 (*movdf_internal): Ditto.
948 (*movsf_internal): Ditto.
949 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
951 2013-08-23 Jakub Jelinek <jakub@redhat.com>
953 PR tree-optimization/58209
954 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
955 (find_tail_calls): Give up for pointer result types if m is non-NULL.
956 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
957 emit POINTER_PLUS_EXPR.
958 (create_tailcall_accumulator): For pointer result type accumulate in
961 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
963 * configure.ac: Add backslashes missing from the last change.
964 * configure: Regenerate.
966 2013-08-22 Jan Hubicka <jh@susue.cz>
968 * ipa.c (function_and_variable_visibility): First remember function
969 was global and then make it local.
971 2013-08-22 Julian Brown <julian@codesourcery.com>
973 * configure.ac: Add aarch64 to list of arches which use "nop" in
975 * configure: Regenerate.
977 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
979 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
980 gnu_libc_has_function.
981 * config/s390/tpf.h: Likewise.
983 2013-08-22 Jan Hubicka <jh@susue.cz>
985 * timevar.c (validate_phases): Add cast.
987 2013-08-22 Jan Hubicka <jh@susue.cz>
989 * timevar.c (validate_phases): Use size_t for memory.
990 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
992 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
994 * pretty-print.h (output_buffer::output_buffer): Declare.
995 (pretty_printer::pretty_printer): Likewise.
996 (pp_construct): Remove.
997 * pretty-print.c (output_buffer::output_buffer): Define.
998 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
999 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
1000 (print_gimple_expr): Likewise.
1001 (print_gimple_seq): Likewise.
1002 (gimple_dump_bb): Likewise.
1003 * sched-vis.c (dump_value_slim): Likewise.
1004 (dump_insn_slim): Likewise.
1005 (dump_rtl_slim): Likewise.
1006 (str_pattern_slim): Likewise.
1007 * tree-mudflap.c (mf_varname_tree): Likewise.
1008 * graph.c (print_graph_cfg): Likewise.
1009 (start_graph_dump): Likewise.
1010 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
1012 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
1013 pretty printer initialization.
1014 * coretypes.h (diagnostic_context): Remove superflous type alias
1016 (pretty_printer): Likewise. Declare directly as a class.
1017 (pretty_print_info): Remove declaration as class.
1018 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
1019 and pp_clear_output_area.
1020 (asan_add_global): Likewise.
1022 2013-08-22 Jan Hubicka <jh@suse.cz>
1024 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
1025 * ipa-utils.h (update_type_inheritance_graph): Declare.
1026 (possible_polymorphic_call_target_p): Declare.
1027 (possible_polymorphic_call_target_p): New.
1028 * ipa-devirt.c: Update toplevel comments.
1029 (cached_polymorphic_call_targets): Move up.
1030 (odr_type_d): Move ID down.
1031 (polymorphic_type_binfo_p): Update comment.
1032 (odr_hasher::remove): Likewise;
1033 (get_odr_type): Set anonymous_namespace.
1034 (dump_odr_type): Dump it.
1035 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
1036 (maybe_record_node): Record node in cached_polymorphic_call_targets.
1037 (record_binfo): Add comment.
1038 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
1040 (devirt_node_removal_hook): Do not iCE when cache is freed.
1041 (possible_polymorphic_call_target_p): New predicate.
1042 (update_type_inheritance_graph): New function.
1044 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
1045 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1046 Sergey Lega <sergey.s.lega@intel.com>
1047 Anna Tikhonova <anna.tikhonova@intel.com>
1048 Ilya Tocar <ilya.tocar@intel.com>
1049 Andrey Turetskiy <andrey.turetskiy@intel.com>
1050 Ilya Verbin <ilya.verbin@intel.com>
1051 Kirill Yukhin <kirill.yukhin@intel.com>
1052 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1054 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
1055 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
1056 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
1057 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
1058 (OPTION_MASK_ISA_AVX2_UNSET): Update.
1059 (OPTION_MASK_ISA_AVX512F_UNSET): New.
1060 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
1061 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
1062 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
1063 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
1064 OPT_mavx512pf, OPT_mavx512er cases.
1065 * config/i386/constraints.md (v): New constraint.
1066 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
1067 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
1068 (bit_AVX512CD): New.
1069 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
1070 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
1071 * config/i386/i386-c.c (ix86_target_macros_internal):
1072 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
1074 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
1075 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
1076 * config/i386/i386.c (regclass_map, dbx_register_map)
1077 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
1078 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
1079 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
1081 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
1082 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
1083 -mavx512pf options. Fix formatting.
1084 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
1085 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
1086 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
1087 -mavx512cd, -mavx512pf options.
1088 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
1089 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
1090 (ix86_preferred_output_reload_class): Replace SSE_REGS with
1092 (ix86_hard_regno_mode_ok): Support 512-bit registers.
1093 (ix86_set_reg_reg_cost): Ditto.
1094 (x86_order_regs_for_local_alloc): Ditto.
1095 (MAX_VECT_LEN): Extend to 64-byte.
1096 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
1097 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
1098 (TARGET_AVX512ER, TARGET_AVX512CD): New.
1099 (BIGGEST_ALIGNMENT): Extend to 512-bits.
1100 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
1101 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
1102 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
1103 (SSE_REG_MODE_P): Support new modes.
1104 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
1105 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
1106 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
1107 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
1108 (REG_CLASS_CONTENTS): Add new registers.
1109 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
1110 (EXT_REX_SSE_REGNO_P): New.
1111 (HI_REGISTER_NAMES): Add new registers.
1112 * config/i386/i386.md: Define constants for new registers.
1113 (mode): Add new 512-bit modes.
1114 (prefix): Support evex prefix.
1115 (isa): Support avx512f, noavx512f, fma_avx512f.
1116 (ssemodesuffix): Add new 512-bit modes.
1118 (*movxi_internal_avx512f): Ditto.
1119 (*movdi_internal): Replace constraint "x" with the new constraint "v".
1121 (*movsi_internal): Likewise.
1122 (*movdf_internal): Likewise.
1123 (*movsf_internal): Likewise.
1124 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
1125 (<code><mode>3): Likewise.
1126 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
1128 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
1129 with the new constraint "v".
1130 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
1132 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
1133 with the new constraint "v".
1134 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
1135 (<sse2>_storedqu<avxsizesuffix>): Likewise.
1136 (*<plusminus_insn><mode>3): Likewise.
1137 (<sse>_vm<plusminus_insn><mode>3): Likewise.
1138 (*mul<mode>3): Likewise.
1139 (<sse>_vmmul<mode>3): Likewise.
1140 (<sse>_div<mode>3): Likewise.
1141 (<sse>_vmdiv<mode>3): Likewise.
1142 (<sse>_sqrt<mode>2): Likewise.
1143 (<sse>_vmsqrt<mode>2): Likewise.
1144 (*<code><mode>3_finite): Likewise.
1145 (*<code><mode>3) <smaxmin>: Likewise.
1146 (<sse>_vm<code><mode>3): Likewise.
1147 (*<code><mode>3) <any_logic>: Likewise.
1148 (*fma_fmadd_<mode>): Likewise.
1149 (*fma_fmsub_<mode>): Likewise.
1150 (*fma_fnmadd_<mode>): Likewise.
1151 (*fma_fnmsub_<mode>): Likewise.
1152 (*fma_fmaddsub_<mode>): Likewise.
1153 (*fma_fmsubadd_<mode>): Likewise.
1154 (*fmai_fmadd_<mode>): Likewise.
1155 (*fmai_fmsub_<mode>): Likewise.
1156 (*fmai_fnmadd_<mode>): Likewise.
1157 (*fmai_fnmsub_<mode>): Likewise.
1158 (sse_cvtsi2ss): Likewise.
1159 (sse_cvtsi2ssq): Likewise.
1160 (sse_cvtss2si): Likewise.
1161 (sse_cvtss2si_2): Likewise.
1162 (sse_cvtss2siq): Likewise.
1163 (sse_cvtss2siq_2): Likewise.
1164 (sse_cvttss2si): Likewise.
1165 (sse_cvtss2siq_2): Likewise.
1166 (float<sseintvecmodelower><mode>2): Likewise.
1167 (sse2_cvtsd2si_2): Likewise.
1168 (sse2_cvtsd2siq_2): Likewise.
1169 (*<plusminus_insn><mode>3): Likewise.
1170 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
1171 (*<sse4_1_avx2>_mul<mode>3): Likewise.
1172 (ashr<mode>3): Likewise.
1173 (<shift_insn><mode>3): Likewise.
1174 (avx2_<code><mode>3): Likewise.
1175 (*avx2_<code><mode>3): Likewise.
1176 (*andnot<mode>3): Likewise.
1177 (*<code><mode>3) <any_logic>: Likewise.
1178 (abs<mode>2): Likewise.
1179 (avx2_permvar<mode>): Likewise.
1180 (avx2_perm<mode>_1): Likewise.
1181 (*avx_vpermilp<mode>): Likewise.
1182 (avx_vpermilvar<mode>3): Likewise.
1183 (avx2_ashrv<mode>): Likewise.
1184 (avx2_<shift_insn>v<mode>): Likewise.
1185 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
1187 * doc/rtl.texi: Document XImode.
1189 2013-08-21 Jeff Law <law@redhat.com>
1191 * tree-flow.h (register_jump_thread): Pass vector of edges
1192 instead of each important edge.
1193 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
1194 thread path into a vector and pass that to register_jump_thread.
1195 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
1196 passed in edge vector to the current 3-edge form.
1199 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1201 * dce.c (fini_dce): Call df_analyze again just in case
1202 delete_unmarked_insns removed anything.
1204 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
1206 * reload.h (struct reg_equivs): Rename to ..
1207 (struct reg_equivs_s): .. this.
1209 2013-08-20 Martin Liska <marxin.liska@gmail.com>
1211 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
1213 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1215 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
1217 2013-08-21 Jeff Law <law@redhat.com>
1219 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
1220 simplify assignments too. If the RHS collapses to a singleton
1221 range, then return the value for the range.
1223 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
1225 * config/i386/sse.md (V16): Rename to...
1227 (mov<mode>): Update iterator name.
1228 (*mov<mode>_internal): Ditto.
1229 (push<mode>1): Ditto.
1230 (movmisalign<mode>): Ditto.
1232 2013-08-20 Jan Hubicka <jh@suse.cz>
1235 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
1236 entry for direct edges.
1237 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
1239 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1241 Revert my last two changes, r201865 and r201864:
1244 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1246 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1247 instances can own GC refs.
1249 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1250 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1251 (gcc::context::gt_pch_nx): Likewise.
1252 (gcc::context::gt_pch_nx): Likewise.
1253 * ggc.h (gt_ggc_mx <T>): New.
1254 (gt_pch_nx_with_op <T>): New.
1255 (gt_pch_nx <T>): New.
1256 * passes.c (opt_pass::gt_ggc_mx): New.
1257 (opt_pass::gt_pch_nx): New.
1258 (opt_pass::gt_pch_nx_with_op): New.
1259 (pass_manager::gt_ggc_mx): New.
1260 (pass_manager::gt_pch_nx): New.
1261 (pass_manager::gt_pch_nx_with_op): New.
1262 (pass_manager::operator new): Use
1263 ggc_internal_cleared_alloc_stat rather than xcalloc.
1264 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1265 (pass_manager::gt_ggc_mx): New.
1266 (pass_manager::gt_pch_nx): New.
1267 (pass_manager::gt_pch_nx_with_op): New.
1268 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1269 (opt_pass::operator new): New.
1270 (opt_pass::gt_ggc_mx): New.
1271 (opt_pass::gt_pch_nx): New.
1272 (opt_pass::gt_pch_nx_with_op): New.
1275 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1277 * Makefile.in (GTFILES): Add context.h.
1278 * context.c (gcc::context::operator new): New.
1279 (gcc::context::gt_ggc_mx): New.
1280 (gcc::context::gt_pch_nx): New.
1281 (gcc::context::gt_pch_nx): New.
1282 * context.h (gcc::context): Add GTY((user)) marking.
1283 (gcc::context::operator new): New.
1284 (gcc::context::gt_ggc_mx): New.
1285 (gcc::context::gt_pch_nx): New.
1286 (gcc::context::gt_pch_nx): New.
1287 (g): Add GTY marking.
1288 (gt_ggc_mx (gcc::context *)): New.
1289 (gt_pch_nx (gcc::context *)): New.
1290 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1291 void *cookie)): New.
1292 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1294 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1296 * dce.c (fini_dce): Call df_analyze again just in case
1297 delete_unmarked_insns removed anything.
1299 2013-08-20 Teresa Johnson <tejohnson@google.com>
1301 PR rtl-optimizations/57451
1302 * final.c (reemit_insn_block_notes): Prevent lexical blocks
1303 from crossing split section boundaries.
1305 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1307 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
1308 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
1309 with MULTLIB_DEFAULTS.
1311 2013-08-20 Nick Clifton <nickc@redhat.com>
1313 * target.def (narrow_volatile_bitfield): Note that the default
1314 value is false, not !TARGET_STRICT_ALIGN.
1315 * doc/tm.texi: Regenerate.
1317 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
1319 Fix LIB_SPEC for systems without libpthread.
1321 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
1322 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
1324 * config/i386/linux-common.h: Likewise.
1325 * config/mips/linux-common.h: Likewise.
1327 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1329 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
1332 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1334 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1335 instances can own GC refs.
1337 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1338 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1339 (gcc::context::gt_pch_nx): Likewise.
1340 (gcc::context::gt_pch_nx): Likewise.
1341 * ggc.h (gt_ggc_mx <T>): New.
1342 (gt_pch_nx_with_op <T>): New.
1343 (gt_pch_nx <T>): New.
1344 * passes.c (opt_pass::gt_ggc_mx): New.
1345 (opt_pass::gt_pch_nx): New.
1346 (opt_pass::gt_pch_nx_with_op): New.
1347 (pass_manager::gt_ggc_mx): New.
1348 (pass_manager::gt_pch_nx): New.
1349 (pass_manager::gt_pch_nx_with_op): New.
1350 (pass_manager::operator new): Use
1351 ggc_internal_cleared_alloc_stat rather than xcalloc.
1352 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1353 (pass_manager::gt_ggc_mx): New.
1354 (pass_manager::gt_pch_nx): New.
1355 (pass_manager::gt_pch_nx_with_op): New.
1356 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1357 (opt_pass::operator new): New.
1358 (opt_pass::gt_ggc_mx): New.
1359 (opt_pass::gt_pch_nx): New.
1360 (opt_pass::gt_pch_nx_with_op): New.
1362 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1364 * Makefile.in (GTFILES): Add context.h.
1365 * context.c (gcc::context::operator new): New.
1366 (gcc::context::gt_ggc_mx): New.
1367 (gcc::context::gt_pch_nx): New.
1368 (gcc::context::gt_pch_nx): New.
1369 * context.h (gcc::context): Add GTY((user)) marking.
1370 (gcc::context::operator new): New.
1371 (gcc::context::gt_ggc_mx): New.
1372 (gcc::context::gt_pch_nx): New.
1373 (gcc::context::gt_pch_nx): New.
1374 (g): Add GTY marking.
1375 (gt_ggc_mx (gcc::context *)): New.
1376 (gt_pch_nx (gcc::context *)): New.
1377 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1378 void *cookie)): New.
1379 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1381 2013-08-20 Alan Modra <amodra@gmail.com>
1384 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
1385 (rs6000_emit_epilogue): Likewise.
1387 2013-08-19 Dehao Chen <dehao@google.com>
1389 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
1391 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
1392 Jakub Jelinek <jakub@redhat.com>
1394 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
1395 (BUILT_IN_FABSD64): Likewise.
1396 (BUILT_IN_FABSD128): Likewise.
1397 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
1398 (fold_builtin_1): Likewise.
1399 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
1400 destination and source operands.
1401 (*abstd2_fpr): Likewise.
1402 (*nabstd2_fpr): Likewise.
1404 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
1406 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
1409 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
1411 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
1414 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
1416 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
1417 * builtins.c (default_libc_has_function): New.
1418 (gnu_libc_has_function): Ditto.
1419 (no_c99_libc_has_function): Ditto.
1420 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
1421 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
1422 (fold_builtin_sincos): Likewise.
1423 (fold_builtin_cexp): Likewise.
1424 * builtins.def (DEF_C94_BUILTIN): Likewise.
1425 (DEF_C99_BUILTIN): Likewise.
1426 (DEF_C99_C90RES_BUILTIN): Likewise.
1427 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
1428 definitions to using this define.
1429 * config/darwin-protos.h (darwin_libc_has_function): New.
1430 * config/darwin.c (darwin_libc_has_function): Ditto.
1431 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
1432 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
1433 * config/darwin.h: Ditto.
1434 * config/elfos.h: Ditto.
1435 * config/freebsd.h: Ditto.
1436 * config/i386/cygming.h: Ditto.
1437 * config/i386/djgpp.h: Ditto.
1438 * config/i386/i386-interix.h: Ditto.
1439 * config/microblaze/microblaze.h: Ditto.
1440 * config/mmix/mmix.h: Ditto.
1441 * config/gnu-user.h: Ditto.
1442 * config/ia64/hpux.h: Ditto.
1443 * config/pa/pa-hpux.h: Ditto.
1444 * config/pdp11/pdp11.h: Ditto.
1445 * config/picochip/picochip.h: Ditto.
1446 * config/linux.h: Ditto.
1447 * config/netbsd.h: Ditto.
1448 * config/openbsd.h: Ditto.
1449 * config/rs6000/aix43.h: Ditto.
1450 * config/rs6000/aix51.h: Ditto.
1451 * config/rs6000/aix52.h: Ditto.
1452 * config/rs6000/aix53.h: Ditto.
1453 * config/rs6000/aix61.h: Ditto.
1454 * config/rs6000/darwin.h: Ditto.
1455 * config/rs6000/linux.h: Ditto.
1456 * config/rs6000/linux64.h: Ditto.
1457 * config/s390/tpf.h: Ditto.
1458 * config/sol2-10.h: Ditto.
1459 * config/sol2.h: Ditto.
1460 * config/vms/vms.h: Ditto.
1461 * config/vxworks.h: Ditto.
1462 * config/linux-android.c (linux_android_libc_has_function):
1463 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
1464 * config/linux-protos.h (linux_android_libc_has_function):
1466 * config/i386/i386.c (ix86_libc_has_function): New.
1467 * config/i386/i386-protos.h
1468 (ix86_libc_has_function): New declaration.
1469 * config/i386/i386.md
1470 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
1471 ("isinf<mode>2): Likewise.
1472 * convert.c (convert_to_integer): Using new target hook
1473 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
1474 TARGET_C99_FUNCTIONS.
1475 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
1476 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
1477 * coretypes.h (function_class): New enum for different
1478 classes of functions.
1479 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
1480 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
1481 (TARGET_HAS_SINCOS): Likewise.
1482 (TARGET_LIBC_HAS_FUNCTION): New.
1483 * doc/tm.texi: Regenerated.
1484 * targhooks.h (default_libc_has_function): New declaration.
1485 (no_c99_libc_has_function): Ditto.
1486 (gnu_libc_has_function): Ditto.
1487 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
1488 and TARGET_HAS_SINCOS.
1490 2013-08-18 Jan Hubicka <jh@suse.cz>
1492 * Makeifle-in (ipa-devirt.o): New.
1493 (GTFILES): Add ipa-utils.h and ipa-devirt.c
1494 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
1495 (analyze_functions): Look into possible targets of polymorphic call.
1496 * dumpfile.c (dump_files): Add type-inheritance dump.
1497 * dumpfile.h (TDI_inheritance): New.
1498 * ipa-devirt.c: New file.
1499 * ipa-utils.h (odr_type_d): Forward declare.
1500 (odr_type): New type.
1501 (build_type_inheritance_graph): Declare.
1502 (possible_polymorphic_call_targets): Declare and introduce inline
1503 variant when only edge is pased.
1504 (dump_possible_polymorphic_call_targets): Likewise.
1505 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
1506 * tree.c (type_in_anonymous_namespace_p): Break out from ...
1507 (types_same_for_odr): ... here.
1508 * tree.h (type_in_anonymous_namespace_p): Declare.
1510 2013-08-18 Jakub Jelinek <jakub@redhat.com>
1512 PR tree-optimization/58006
1513 * tree-parloops.c (take_address_of): Don't ICE if get_name
1515 (eliminate_local_variables_stmt): Remove clobber stmts.
1517 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
1519 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
1520 error message is issued for an alias to undefined symbol.
1522 2013-08-18 Jan Hubicka <jh@suse.cz>
1524 * cgraph.c (cgraph_create_indirect_edge): Discover
1525 polymorphic calls and record basic info into indirect_info.
1526 * gimple-fold.c (gimple_fold_call): When doing BINFO based
1527 devirtualization, ignore objc function calls.
1528 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
1529 call with no parm index info.
1530 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
1531 * tree.c (virtual_method_call_p): New function.
1532 * tree.h (virtual_method_call_p): Declare.
1534 2013-08-16 Jan Hubicka <jh@suse.cz>
1537 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
1539 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
1541 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
1544 2013-08-16 David Malcolm <dmalcolm@redhat.com>
1546 * gengtype.c (type_for_name): Add special-case support for
1547 locating types within the "gcc::" namespace.
1548 (open_base_files): Emit a "using namespace gcc" directive.
1550 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1553 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
1554 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
1556 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
1557 array instead of each individual operand as a separate argument.
1558 (emit_fusion_gpr_load): Likewise.
1559 (expand_fusion_gpr_load): Add new function declaration.
1561 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
1562 signature to have the operands passed as an array, instead of as
1563 separate arguments. Allow ZERO_EXTEND to be in the memory
1564 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
1565 depend on the register live/dead flags when peepholes are run.
1566 (expand_fusion_gpr_load): New function to be called from the
1567 peephole2 pass, to change the register that addis sets to be the
1569 (emit_fusion_gpr_load): Change the calling signature to have the
1570 operands passed as an array, instead of as separate arguments.
1571 Allow ZERO_EXTEND to be in the memory address, and also
1572 SIGN_EXTEND if -mpower8-fusion-sign.
1574 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
1576 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
1577 adjust the register addis loads up in the peephole2 pass. Do not
1578 depend on the register live/dead state when the peephole pass is done.
1580 2013-08-16 David Malcolm <dmalcolm@redhat.com>
1582 * gengtype.c (create_user_defined_type): Ensure that the kind
1583 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
1584 declaration is seen before the GTY((user)) marking.
1586 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
1589 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
1591 2013-08-16 Jan Hubicka <jh@suse.cz>
1593 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
1594 arugment expected_type.
1595 (gimple_fold_call): Use it.
1596 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
1597 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
1598 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
1599 (try_make_edge_direct_virtual_call): Likewise.
1600 * tree.c (obj_type_ref_class): New.
1601 * tree.h (obj_type_ref_class): Use it.
1603 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
1605 * sched-vis.c (rtl_slim_pp_initialized): Remove.
1606 (rtl_slim_pp): Likewise.
1607 (init_rtl_slim_pretty_print): Likewise.
1608 (dump_value_slim): Don't call it. Use local pretty printer.
1609 (dump_insn_slim): Likewise.
1610 (dump_rtl_slim): Likewise.
1611 (str_pattern_slim): Likewise.
1612 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
1615 2013-08-16 Jakub Jelinek <jakub@redhat.com>
1617 PR tree-optimization/58164
1618 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
1619 walk gimple_goto_dest of GIMPLE_GOTO.
1621 PR tree-optimization/58165
1622 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
1623 bi_call must be the last stmt in a bb, don't split_block, instead
1624 use fallthru edge from it and give up if there is none.
1625 Release conds vector when returning early.
1627 2013-08-14 Xinliang David Li <davidxl@google.com>
1629 * config/i386/i386.c (ix86_option_override_internal):
1630 Remove unused variable and field.
1632 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1635 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
1636 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
1637 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
1638 and Linux, correct BLKmode alignment when 128-bit alignment is
1639 required and compatibility flag is not set.
1640 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
1641 for zero-size arguments when compatibility flag is not set.
1643 2013-08-14 Jakub Jelinek <jakub@redhat.com>
1645 PR tree-optimization/58145
1646 * tree-sra.c (build_ref_for_offset): If prev_base has
1647 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
1649 2013-08-14 Xinliang David Li <davidxl@google.com>
1651 * config/i386/i386.c (ix86_option_override_internal):
1652 Fix uninitialized variable error.
1654 2013-08-14 Xinliang David Li <davidxl@google.com>
1656 * config/i386/i386.opt: Define two new options.
1657 * config/i386/x86-tune.def: Add arch selector field in macros.
1658 * config/i386/i386.h: Adjust macro definition.
1659 * config/i386/i386.c (ix86_option_override_internal):
1661 (parse_mtune_ctrl_str): New function.
1662 (set_ix86_tune_features): New function.
1663 (ix86_function_specific_restore): Call the new helper function.
1665 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
1667 PR rtl-optimization/57662
1668 * sel-sched.c (code_motion_process_successors): When the current insn
1669 is removed after the recursive traversal, break from the loop.
1670 Add comments and debug printouts.
1672 2013-08-14 Jakub Jelinek <jakub@redhat.com>
1673 Alexandre Oliva <aoliva@redhat.com>
1676 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
1677 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
1678 there also UNSPEC_PLTOFF.
1680 2013-08-14 Marek Polacek <polacek@redhat.com>
1682 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
1683 NUM_CONDITIONS bit positions.
1685 2013-08-13 Cary Coutant <ccoutant@google.com>
1687 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
1688 (attr_checksum): Hash vector contents instead of pointer.
1689 (attr_checksum_ordered): Likewise.
1691 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
1693 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
1694 when Pmode != word_mode. Add length_address attribute.
1695 (sse3_monitor_<mode>): Merge from sse3_monitor and
1696 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
1697 Pmode != word_mode. Update insn length attribute.
1698 * config/i386/i386.c (ix86_option_override_internal): Update
1699 ix86_gen_monitor selection for merged sse3_monitor insn.
1701 2013-08-13 Julian Brown <julian@codesourcery.com>
1703 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
1704 perform invalid legitimization on greater-than-word-size modes for
1707 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
1709 * ira.c (setup_class_translate_array): Use aclass instead of cl
1710 for classes not fully covered by allocno classes.
1712 2013-08-13 Jakub Jelinek <jakub@redhat.com>
1714 PR tree-optimization/57661
1715 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
1716 * tree-inline.c (tree_function_versioning): Initialize it.
1717 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
1718 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
1719 that is not being copied.
1722 * asan.c (instrument_strlen_call): Fix typo in comment.
1723 Use char * type even for the lhs of POINTER_PLUS_EXPR.
1725 2013-08-13 Steve Ellcey <sellcey@mips.com>
1727 * config/mips/mips.md (prefetch): Use lw instead of ld on
1728 loongson in 32bit mode.
1730 2013-08-13 Nick Clifton <nickc@redhat.com>
1732 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
1734 2013-08-13 Jan Hubicka <jh@suse.cz>
1736 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
1737 introduced edge; fix typo in sanity check.
1738 (cgraph_resolve_speculation): Export; improve diagnostic.
1739 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
1740 speculation at type mismatch.
1741 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
1742 (cgraph_resolve_speculation): Declare.
1743 (symtab_can_be_discarded): New function.
1744 * value-prof.c (gimple_ic_transform): Remove actual transform code.
1745 * ipa-inline-transform.c (speculation_removed): New global var.
1746 (clone_inlined_nodes): See if speculation can be removed.
1747 (inline_call): If speculations was removed, we growths may not match.
1748 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
1749 (speculation_useful_p): New function.
1750 (resolve_noninline_speculation): New function.
1751 (inline_small_functions): Resolve useless speculations.
1752 * ipa-inline.h (speculation_useful_p): Declare
1753 * ipa.c (can_replace_by_local_alias): Simplify.
1754 (ipa_profile): Produce speculative calls in non-lto, too;
1755 add simple cost model; produce local aliases.
1757 2013-08-13 David Malcolm <dmalcolm@redhat.com>
1759 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
1762 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
1764 * config/i386/i386.c (ix86_function_versions): Use error + inform.
1766 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
1768 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
1769 iterator instead of X87MODEF.
1771 2013-08-12 Perez Read <netfirewall@gmail.com>
1774 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
1775 operand 0 for intel asm alternative.
1776 (*movabs<mode>_2): Ditto for operand 1.
1778 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
1780 * config/aarch64/arm_none.h
1781 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
1783 2013-08-12 Nick Clifton <nickc@redhat.com>
1785 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
1787 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1789 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
1790 expand for QI/HImode operand to produce more effictive code for
1791 unsigned char(short) --> float(double) conversion.
1793 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
1795 * doc/invoke.texi: Mention that -ftls-model does not force the final
1798 2013-08-12 Marek Polacek <polacek@redhat.com>
1799 Marc Glisse <marc.glisse@inria.fr>
1801 PR tree-optimization/57980
1802 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
1803 when creating -1 constant.
1805 2013-08-10 Jan Hubicka <jh@suse.cz>
1807 Workaround binutils PR14342.
1808 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
1809 (gimple_init_edge_profiler): Likewise.
1810 (gimple_gen_ic_func_profiler): Likewise.
1812 2013-08-09 Jan Hubicka <jh@suse.cz>
1814 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
1816 2013-08-09 Xinliang David Li <davidxl@google.com>
1818 * config/i386/stringop.def: New file.
1819 * config/i386/stringop.opt: New file.
1820 * config/i386/i386-opts.h: Include stringopt.def.
1821 * config/i386/i386.opt: Include stringopt.opt.
1822 * config/i386/i386.c (ix86_option_override_internal):
1823 Override default size based stringop inline strategies with options.
1824 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
1827 2013-08-09 Jan Hubicka <jh@suse.cz>
1829 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
1831 2013-08-09 Jan Hubicka <jh@suse.cz>
1833 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
1835 (dump_cgraph_node): Dump profile-id.
1836 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
1837 and common_target_probability.
1838 * lto-cgraph.c (lto_output_edge): Stream common targets.
1839 (lto_output_node): Stream profile ids.
1840 (input_node): Stream profile ids.
1841 (input_edge): Stream common targets.
1842 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
1843 * ipa.c: Include value-prof.h
1844 (ipa_profile_generate_summary): Turn indirect call statement histograms
1845 into common targets.
1846 (ipa_profile): Turn common targets into speculative edges.
1848 2013-08-09 Jan Hubicka <jh@suse.cz>
1850 * cgraph.h (cgraph_node): Add profile_id.
1851 * value-prof.c (cgraph_node_map): Turn into pointer_map.
1852 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
1853 (del_node_map): Update.
1854 (find_func_by_funcdef_no): Replace by ...
1855 (find_func_by_profile_id): ... this one.
1856 (gimple_ic_transform): Do not remove useful histograms when
1857 speculation is not done; dump info when indirect call removal
1859 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
1860 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
1861 (__gcov_indirect_call_profiler_v2): .. this one.
1862 * profile.h (init_node_map): Update.
1863 * coverage.c (coverage_compute_profile_id): New function.
1864 * coverage.h (coverage_compute_profile_id): Declare.
1865 * tree-profile.c (init_ic_make_global_vars): Make
1866 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
1867 (gimple_init_edge_profiler): Update prototype of
1868 __gcov_indirect_call_profiler.
1869 (gimple_gen_ic_func_profiler): Simplify.
1870 (tree_profiling): Use init_node_map
1872 2013-08-09 Jan Hubicka <jh@suse.cz>
1874 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
1875 non-speculative refs.
1876 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
1877 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
1878 (cgraph_set_call_stmt): Likewise.
1879 (cgraph_create_edge_1): Fix release checking compilatoin;
1881 (cgraph_free_edge): Free indirect info.
1882 (cgraph_turn_edge_to_speculative): New function.
1883 (cgraph_speculative_call_info): New function.
1884 (cgraph_make_edge_direct): Return direct edge; handle speculation.
1885 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
1886 (dump_cgraph_node): Dump speculation.
1887 (verify_edge_count_and_frequency): Accept speculative edges.
1888 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
1889 (verify_cgraph_node): Handle speculation.
1890 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
1891 (cgraph_set_call_stmt): Update prototype.
1892 (cgraph_make_edge_direct): Update prototype.
1893 (cgraph_speculative_call_info): Declare.
1894 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
1895 to change; update call of ipa_find_references.
1896 * ipa-ref.c (ipa_record_reference): Fix return value; clear
1897 lto_stmt_uid and speculative flags.
1898 (ipa_dump_references): Dump speculation.
1899 (ipa_clone_references): Clone speculative flag.
1900 (ipa_clone_referring): Likewise.
1901 (ipa_clone_ref): New function.
1902 (ipa_find_reference): Look into lto_stmt_uids
1903 (ipa_clear_stmts_in_references): Do not clear speculative calls.
1904 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
1905 (ipa_find_reference): Update declaration.
1906 (ipa_clone_ref): Declare.
1907 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
1908 stream speculative flag.
1909 (lto_output_ref): Stream statements uids and speculation.
1910 (input_ref): Likewise.
1911 (input_edge): Stream speuclation.
1912 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
1913 (cgraph_set_call_stmt_including_clones): Handle speculation.
1914 * ipa-inline.c (heap_edge_removal_hook): New function.
1915 (inline_small_functions): Register it.
1916 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
1917 also initialize refs.
1918 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
1920 (try_make_edge_direct_simple_call): Likewise.
1921 (try_make_edge_direct_simple_call): Likewise.
1922 (update_indirect_edges_after_inlining): Likewise.
1923 (remove_described_reference): Look proper lto_stmt_uid.
1924 (propagate_controlled_uses): Likewise.
1925 (propagate_controlled_uses): Liekwise.
1926 * tree-inline.c (copy_bb): Copy speculative edges.
1927 (redirect_all_calls): New function.
1928 (copy_cfg_body): Do redirection after loop info is updated.
1929 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1931 2013-08-09 Jan Hubicka <jh@suse.cz>
1933 * lto-streamer-out.c (output_function): Renumber PHIs.
1934 * lto-streamer-in.c (input_function): Likewise.
1936 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1938 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1939 (get_lane_unsigned): Likewise.
1940 (dup_lane_scalar): Likewise.
1941 (get_lane): enable for VALL.
1942 * config/aarch64/aarch64-simd.md
1943 (aarch64_dup_lane_scalar<mode>): Remove.
1944 (aarch64_get_lane_signed<mode>): Likewise.
1945 (aarch64_get_lane_unsigned<mode>): Likewise.
1946 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1947 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1948 (aarch64_get_lane<mode>): Enable for all vector modes.
1949 (aarch64_get_lanedi): Remove misleading constraints.
1950 * config/aarch64/arm_neon.h
1951 (__aarch64_vget_lane_any): Define.
1952 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1953 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1954 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1955 * config/aarch64/iterators.md (VDQQH): New.
1959 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1961 * configure.ac: Add GAS check for LEON instructions on SPARC.
1962 * configure: Regenerate.
1963 * config.in: Likewise.
1964 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1966 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1967 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1969 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1970 (AS_LEON_FLAG): New macro.
1971 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1972 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1973 Deal with LEON and LEON3 for the memory model.
1974 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1975 (atomic_compare_and_swap<mode>_1): Likewise.
1976 (*atomic_compare_and_swap<mode>_1): Likewise.
1978 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1980 * config/arm/neon.md (vcond): Fix floating-point vector
1981 comparisons against 0.
1983 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1985 * lra-constraints.c (emit_spill_move): Remove assert.
1986 (process_alt_operands): Add more debugging
1987 output. Increase reject for spilling into memory. Decrease
1988 reject for reloading scratch.
1989 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1991 2013-08-08 Steve Ellcey <sellcey@mips.com>
1993 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1994 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1995 micromips incompatible. Add nan2008.
1996 (MULTILIB_DIRNAMES): Add nan2008.
1997 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1998 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1999 and micromips incompatible. Add nan2008.
2000 (MULTILIB_DIRNAMES): Add nan2008.
2001 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
2003 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
2005 PR rtl-optimization/58079
2006 * combine.c (combine_simplify_rtx): Avoid using SUBST if
2007 simplify_comparison has widened a comparison with an integer.
2009 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2011 * config/arm/neon.md (movmisalign<mode>): Disable when we
2012 don't allow unaligned accesses.
2013 (*movmisalign<mode>_neon_store): Likewise.
2014 (*movmisalign<mode>_neon_load): Likewise.
2015 (*movmisalign<mode>_neon_store): Likewise.
2016 (*movmisalign<mode>_neon_load): Likewise.
2018 2013-08-08 Jan Hubicka <jh@suse.cz>
2020 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
2021 (make_pass_rebuild_cgraph_edges): Also clear references.
2022 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
2023 * ipa-inline-transform.c (inline_transform): Remove all references
2025 * cgraphunit.c (expand_function): Remove all references after
2027 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
2028 (ipa_find_reference): Rewrite to iterator.
2029 (remove_stmt_references): Likewise.
2030 (ipa_clear_stmts_in_references): New function.
2031 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
2032 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
2034 * ipa-split.c (split_function): Remove references in split function.
2036 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2039 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
2040 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
2042 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2045 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
2046 suggested mode for the assignment isn't compatible with the
2049 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
2052 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
2054 2013-08-07 Xinliang David Li <davidxl@google.com>
2056 * config/i386/i386.opt: New option -mtune-ctrl=.
2057 * config/i386/x86-tune.def: New file.
2058 * config/i386/i386.h: include x86-tune.def.
2059 * config/i386/i386.c (ix86_option_override_internal):
2060 Parsing -mtune-ctrl= option and set tune features.
2062 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
2065 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
2066 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
2069 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2071 * rtl.h (update_alignments): Declare.
2072 * final.c (grow_label_align): New function extracted from...
2073 (shorten_branches): ...here. Call it.
2074 (update_alignments): New function.
2075 * reorg.c (sibling_labels): New variable.
2076 (get_label_before): Add SIBLING parameter. If it is non-zero, push
2077 the new label along with it onto the sibling_labels vector.
2078 (fill_simple_delay_slots): Adjust call to get_label_before.
2079 (fill_slots_from_thread): Likewise.
2080 (relax_delay_slots): Likewise.
2081 (make_return_insns): Likewise.
2082 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
2084 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2086 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
2087 document its semantics.
2088 (diagnostic_report_diagnostic): Adjust accordingly.
2090 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2092 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
2093 (sparc_option_override): ...and port to new C++ pass API.
2094 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
2096 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
2098 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
2100 2013-08-06 Caroline Tice <cmtice@google.com>
2102 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
2103 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
2104 * tree-pass.h: Add pass_vtable_verify.
2105 * varasm.c (assemble_variable): Add code to properly set the comdat
2106 section and name for the .vtable_map_vars section.
2107 (assemble_vtyv_preinit_initializer): New function.
2108 (default_sectin_type_flags): Make sure .vtable_map_vars section has
2110 * output.h: Add function decl for assemble_vtv_preinit_initializer.
2111 * vtable-verify.c: New file.
2112 * vtable-verify.h: New file.
2113 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
2114 initialiation levels.
2115 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
2116 * passes.def: Insert pass_vtable_verify.
2117 * aclocal.m4: Reorder includes.
2118 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
2119 -fvtv-counts options.
2120 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
2121 as appropriate, if -fvtable-verify=... is used.
2122 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
2123 -fvtable-verify=... is used.
2124 * Makefile.in (OBJS): Add vtable-verify.o to list.
2125 (vtable-verify.o): Add new build rule.
2126 (GTFILES): Add vtable-verify.c to list.
2127 * common.opt (fvtable-verify=): New flag.
2128 (vtv_priority): Values for fvtable-verify= flag.
2129 (fvtv-counts): New flag.
2130 (fvtv-debug): New flag.
2131 * tree.h (save_vtable_map_decl): New extern function decl.
2133 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2135 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
2136 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
2137 (pass_data_rl78_devirt): ...new pass_data instance and...
2138 (make_pass_rl78_devirt): ...new function.
2139 (rl78_asm_file_start): Port pass registration to new C++ API.
2141 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2143 * coretypes.h (rtl_opt_pass): Add.
2144 (gcc::context): Add.
2145 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
2146 (epiphany_init): Port to new C++ pass API.
2147 (epiphany_optimize_mode_switching): Likewise.
2148 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
2149 (pass_manager::get_pass_mode_switching): New.
2150 (pass_manager::get_pass_peephole2): New.
2151 * mode-switching.c (pass_mode_switching): Add clone method.
2152 * recog.c (pass_peephole2): Add clone method.
2153 (pass_split_all_insns): Add clone method.
2155 2013-08-06 David Malcolm <dmalcolm@redhat.com>
2157 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
2158 (mips_option_override): ...here, porting to new C++ API for passes.
2160 2013-08-06 Jan Hubicka <jh@suse.cz>
2162 * cgraph.c (cgraph_get_body): New function based on lto.c
2164 * cgraph.h (cgraph_get_body): Declare.
2165 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
2167 * cgraphunit.c (expand_function): Get body prior expanding.
2168 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
2169 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
2171 * passes.c (do_per_function_toporder): Get body.
2172 * tree-inline.c (expand_call_inline): Get body prior inlining it.
2173 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
2175 2013-08-06 Martin Jambor <mjambor@suse.cz>
2178 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
2179 is not re-finalized. Rename second parameter to no_collect.
2181 2013-08-06 Martin Jambor <mjambor@suse.cz>
2184 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
2185 MEM_REF has proper alignment information.
2187 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
2190 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
2192 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
2193 rtx (*) (rtx, ...) with insn_gen_fn.
2194 * genoutput.c (output_insn_data): Cast gen_? function pointers to
2195 insn_gen_fn::stored_funcptr. Add initializer braces.
2197 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2199 Rewrite how instances of passes are cloned to remove assumptions
2200 about their sizes (thus allowing pass subclasses to have
2201 additional data fields, albeit non-GC-managed ones at this point).
2203 * passes.c (make_pass_instance): Now that passes have clone
2204 methods, rewrite this function to eliminate XNEW and memcpy
2205 calls that used hardcoded sizes. Since this function no longer
2206 creates pass instances, rename it to...
2207 (add_pass_instance): ...this. Document the old way that passes were
2208 numbered and flagged, and rework this function to continue using it.
2209 (next_pass_1): Add an initial_pass argument for use by
2211 (position_pass): When adding multiple instances of a pass, use
2212 the pass's clone method, rather than relying on the XNEW/memcpy
2213 within the former make_pass_instance (now add_pass_instance).
2214 (pass_manager::pass_manager): When invoking next_pass_1, also supply
2215 the initial instance of the current pass within the pass manager.
2217 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2219 This is the automated part of the conversion of passes from C
2220 structs to C++ classes.
2222 Patch autogenerated by refactor_passes.py from
2223 https://github.com/davidmalcolm/gcc-refactoring-scripts
2224 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
2226 * asan.c (pass_asan): Convert from a global struct to a subclass of
2227 gimple_opt_pass along with...
2228 (pass_data_asan): ...new pass_data instance and...
2229 (make_pass_asan): ...new function.
2230 (pass_asan_O0): Convert from a global struct to a subclass of
2231 gimple_opt_pass along with...
2232 (pass_data_asan_O0): ...new pass_data instance and...
2233 (make_pass_asan_O0): ...new function.
2234 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
2235 subclass of rtl_opt_pass along with...
2236 (pass_data_inc_dec): ...new pass_data instance and...
2237 (make_pass_inc_dec): ...new function.
2238 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
2239 a subclass of rtl_opt_pass along with...
2240 (pass_data_reorder_blocks): ...new pass_data instance and...
2241 (make_pass_reorder_blocks): ...new function.
2242 (pass_duplicate_computed_gotos): Convert from a global struct to a
2243 subclass of rtl_opt_pass along with...
2244 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
2245 (make_pass_duplicate_computed_gotos): ...new function.
2246 (pass_partition_blocks): Convert from a global struct to a subclass of
2247 rtl_opt_pass along with...
2248 (pass_data_partition_blocks): ...new pass_data instance and...
2249 (make_pass_partition_blocks): ...new function.
2250 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
2251 struct to a subclass of rtl_opt_pass along with...
2252 (pass_data_branch_target_load_optimize1): ...new pass_data instance
2254 (make_pass_branch_target_load_optimize1): ...new function.
2255 (pass_branch_target_load_optimize2): Convert from a global struct to a
2256 subclass of rtl_opt_pass along with...
2257 (pass_data_branch_target_load_optimize2): ...new pass_data instance
2259 (make_pass_branch_target_load_optimize2): ...new function.
2260 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
2261 of rtl_opt_pass along with...
2262 (pass_data_jump): ...new pass_data instance and...
2263 (make_pass_jump): ...new function.
2264 (pass_jump2): Convert from a global struct to a subclass of
2265 rtl_opt_pass along with...
2266 (pass_data_jump2): ...new pass_data instance and...
2267 (make_pass_jump2): ...new function.
2268 * cfgexpand.c (pass_expand): Convert from a global struct to a
2269 subclass of rtl_opt_pass along with...
2270 (pass_data_expand): ...new pass_data instance and...
2271 (make_pass_expand): ...new function.
2272 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
2273 of rtl_opt_pass along with...
2274 (pass_data_free_cfg): ...new pass_data instance and...
2275 (make_pass_free_cfg): ...new function.
2276 (pass_into_cfg_layout_mode): Convert from a global struct to a
2277 subclass of rtl_opt_pass along with...
2278 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
2279 (make_pass_into_cfg_layout_mode): ...new function.
2280 (pass_outof_cfg_layout_mode): Convert from a global struct to a
2281 subclass of rtl_opt_pass along with...
2282 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
2283 (make_pass_outof_cfg_layout_mode): ...new function.
2284 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
2285 struct to a subclass of gimple_opt_pass along with...
2286 (pass_data_build_cgraph_edges): ...new pass_data instance and...
2287 (make_pass_build_cgraph_edges): ...new function.
2288 (pass_rebuild_cgraph_edges): Convert from a global struct to a
2289 subclass of gimple_opt_pass along with...
2290 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
2291 (make_pass_rebuild_cgraph_edges): ...new function.
2292 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
2293 subclass of gimple_opt_pass along with...
2294 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
2296 (make_pass_remove_cgraph_callee_edges): ...new function.
2297 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
2298 struct to a subclass of rtl_opt_pass along with...
2299 (pass_data_stack_adjustments): ...new pass_data instance and...
2300 (make_pass_stack_adjustments): ...new function.
2301 * combine.c (pass_combine): Convert from a global struct to a subclass
2302 of rtl_opt_pass along with...
2303 (pass_data_combine): ...new pass_data instance and...
2304 (make_pass_combine): ...new function.
2305 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
2306 global struct to a subclass of rtl_opt_pass along with...
2307 (pass_data_compare_elim_after_reload): ...new pass_data instance
2309 (make_pass_compare_elim_after_reload): ...new function.
2310 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
2311 of rtl_opt_pass along with...
2312 (pass_data_rtl_cprop): ...new pass_data instance and...
2313 (make_pass_rtl_cprop): ...new function.
2314 * cse.c (pass_cse): Convert from a global struct to a subclass of
2315 rtl_opt_pass along with...
2316 (pass_data_cse): ...new pass_data instance and...
2317 (make_pass_cse): ...new function.
2318 (pass_cse2): Convert from a global struct to a subclass of
2319 rtl_opt_pass along with...
2320 (pass_data_cse2): ...new pass_data instance and...
2321 (make_pass_cse2): ...new function.
2322 (pass_cse_after_global_opts): Convert from a global struct to a
2323 subclass of rtl_opt_pass along with...
2324 (pass_data_cse_after_global_opts): ...new pass_data instance and...
2325 (make_pass_cse_after_global_opts): ...new function.
2326 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
2327 of rtl_opt_pass along with...
2328 (pass_data_ud_rtl_dce): ...new pass_data instance and...
2329 (make_pass_ud_rtl_dce): ...new function.
2330 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
2331 rtl_opt_pass along with...
2332 (pass_data_fast_rtl_dce): ...new pass_data instance and...
2333 (make_pass_fast_rtl_dce): ...new function.
2334 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
2335 a subclass of rtl_opt_pass along with...
2336 (pass_data_df_initialize_opt): ...new pass_data instance and...
2337 (make_pass_df_initialize_opt): ...new function.
2338 (pass_df_initialize_no_opt): Convert from a global struct to a
2339 subclass of rtl_opt_pass along with...
2340 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
2341 (make_pass_df_initialize_no_opt): ...new function.
2342 (pass_df_finish): Convert from a global struct to a subclass of
2343 rtl_opt_pass along with...
2344 (pass_data_df_finish): ...new pass_data instance and...
2345 (make_pass_df_finish): ...new function.
2346 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
2347 rtl_opt_pass along with...
2348 (pass_data_rtl_dse1): ...new pass_data instance and...
2349 (make_pass_rtl_dse1): ...new function.
2350 (pass_rtl_dse2): Convert from a global struct to a subclass of
2351 rtl_opt_pass along with...
2352 (pass_data_rtl_dse2): ...new pass_data instance and...
2353 (make_pass_rtl_dse2): ...new function.
2354 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
2355 subclass of rtl_opt_pass along with...
2356 (pass_data_dwarf2_frame): ...new pass_data instance and...
2357 (make_pass_dwarf2_frame): ...new function.
2358 * except.c (pass_set_nothrow_function_flags): Convert from a global
2359 struct to a subclass of rtl_opt_pass along with...
2360 (pass_data_set_nothrow_function_flags): ...new pass_data instance
2362 (make_pass_set_nothrow_function_flags): ...new function.
2363 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
2364 subclass of rtl_opt_pass along with...
2365 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
2367 (make_pass_convert_to_eh_region_ranges): ...new function.
2368 * final.c (pass_compute_alignments): Convert from a global struct to a
2369 subclass of rtl_opt_pass along with...
2370 (pass_data_compute_alignments): ...new pass_data instance and...
2371 (make_pass_compute_alignments): ...new function.
2372 (pass_final): Convert from a global struct to a subclass of
2373 rtl_opt_pass along with...
2374 (pass_data_final): ...new pass_data instance and...
2375 (make_pass_final): ...new function.
2376 (pass_shorten_branches): Convert from a global struct to a subclass of
2377 rtl_opt_pass along with...
2378 (pass_data_shorten_branches): ...new pass_data instance and...
2379 (make_pass_shorten_branches): ...new function.
2380 (pass_clean_state): Convert from a global struct to a subclass of
2381 rtl_opt_pass along with...
2382 (pass_data_clean_state): ...new pass_data instance and...
2383 (make_pass_clean_state): ...new function.
2384 * function.c (pass_instantiate_virtual_regs): Convert from a global
2385 struct to a subclass of rtl_opt_pass along with...
2386 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
2387 (make_pass_instantiate_virtual_regs): ...new function.
2388 (pass_leaf_regs): Convert from a global struct to a subclass of
2389 rtl_opt_pass along with...
2390 (pass_data_leaf_regs): ...new pass_data instance and...
2391 (make_pass_leaf_regs): ...new function.
2392 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
2393 subclass of rtl_opt_pass along with...
2394 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
2396 (make_pass_thread_prologue_and_epilogue): ...new function.
2397 (pass_match_asm_constraints): Convert from a global struct to a
2398 subclass of rtl_opt_pass along with...
2399 (pass_data_match_asm_constraints): ...new pass_data instance and...
2400 (make_pass_match_asm_constraints): ...new function.
2401 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
2402 subclass of rtl_opt_pass along with...
2403 (pass_data_rtl_fwprop): ...new pass_data instance and...
2404 (make_pass_rtl_fwprop): ...new function.
2405 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
2406 rtl_opt_pass along with...
2407 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
2408 (make_pass_rtl_fwprop_addr): ...new function.
2409 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
2410 rtl_opt_pass along with...
2411 (pass_data_rtl_pre): ...new pass_data instance and...
2412 (make_pass_rtl_pre): ...new function.
2413 (pass_rtl_hoist): Convert from a global struct to a subclass of
2414 rtl_opt_pass along with...
2415 (pass_data_rtl_hoist): ...new pass_data instance and...
2416 (make_pass_rtl_hoist): ...new function.
2417 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
2418 subclass of gimple_opt_pass along with...
2419 (pass_data_lower_cf): ...new pass_data instance and...
2420 (make_pass_lower_cf): ...new function.
2421 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
2422 from a global struct to a subclass of gimple_opt_pass along with...
2423 (pass_data_strength_reduction): ...new pass_data instance and...
2424 (make_pass_strength_reduction): ...new function.
2425 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
2426 of rtl_opt_pass along with...
2427 (pass_data_rtl_ifcvt): ...new pass_data instance and...
2428 (make_pass_rtl_ifcvt): ...new function.
2429 (pass_if_after_combine): Convert from a global struct to a subclass of
2430 rtl_opt_pass along with...
2431 (pass_data_if_after_combine): ...new pass_data instance and...
2432 (make_pass_if_after_combine): ...new function.
2433 (pass_if_after_reload): Convert from a global struct to a subclass of
2434 rtl_opt_pass along with...
2435 (pass_data_if_after_reload): ...new pass_data instance and...
2436 (make_pass_if_after_reload): ...new function.
2437 * init-regs.c (pass_initialize_regs): Convert from a global struct to
2438 a subclass of rtl_opt_pass along with...
2439 (pass_data_initialize_regs): ...new pass_data instance and...
2440 (make_pass_initialize_regs): ...new function.
2441 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
2442 of ipa_opt_pass_d along with...
2443 (pass_data_ipa_cp): ...new pass_data instance and...
2444 (make_pass_ipa_cp): ...new function.
2445 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
2446 global struct to a subclass of gimple_opt_pass along with...
2447 (pass_data_inline_parameters): ...new pass_data instance and...
2448 (make_pass_inline_parameters): ...new function.
2449 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
2450 subclass of gimple_opt_pass along with...
2451 (pass_data_early_inline): ...new pass_data instance and...
2452 (make_pass_early_inline): ...new function.
2453 (pass_ipa_inline): Convert from a global struct to a subclass of
2454 ipa_opt_pass_d along with...
2455 (pass_data_ipa_inline): ...new pass_data instance and...
2456 (make_pass_ipa_inline): ...new function.
2457 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
2458 struct to a subclass of gimple_opt_pass along with...
2459 (pass_data_local_pure_const): ...new pass_data instance and...
2460 (make_pass_local_pure_const): ...new function.
2461 (pass_ipa_pure_const): Convert from a global struct to a subclass of
2462 ipa_opt_pass_d along with...
2463 (pass_data_ipa_pure_const): ...new pass_data instance and...
2464 (make_pass_ipa_pure_const): ...new function.
2465 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
2466 to a subclass of ipa_opt_pass_d along with...
2467 (pass_data_ipa_reference): ...new pass_data instance and...
2468 (make_pass_ipa_reference): ...new function.
2469 * ipa-split.c (pass_split_functions): Convert from a global struct to
2470 a subclass of gimple_opt_pass along with...
2471 (pass_data_split_functions): ...new pass_data instance and...
2472 (make_pass_split_functions): ...new function.
2473 (pass_feedback_split_functions): Convert from a global struct to a
2474 subclass of gimple_opt_pass along with...
2475 (pass_data_feedback_split_functions): ...new pass_data instance and...
2476 (make_pass_feedback_split_functions): ...new function.
2477 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
2478 global struct to a subclass of simple_ipa_opt_pass along with...
2479 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
2481 (make_pass_ipa_function_and_variable_visibility): ...new function.
2482 (pass_ipa_free_inline_summary): Convert from a global struct to a
2483 subclass of simple_ipa_opt_pass along with...
2484 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
2485 (make_pass_ipa_free_inline_summary): ...new function.
2486 (pass_ipa_whole_program_visibility): Convert from a global struct to a
2487 subclass of ipa_opt_pass_d along with...
2488 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
2490 (make_pass_ipa_whole_program_visibility): ...new function.
2491 (pass_ipa_profile): Convert from a global struct to a subclass of
2492 ipa_opt_pass_d along with...
2493 (pass_data_ipa_profile): ...new pass_data instance and...
2494 (make_pass_ipa_profile): ...new function.
2495 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
2496 ipa_opt_pass_d along with...
2497 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
2498 (make_pass_ipa_cdtor_merge): ...new function.
2499 * ira.c (pass_ira): Convert from a global struct to a subclass of
2500 rtl_opt_pass along with...
2501 (pass_data_ira): ...new pass_data instance and...
2502 (make_pass_ira): ...new function.
2503 (pass_reload): Convert from a global struct to a subclass of
2504 rtl_opt_pass along with...
2505 (pass_data_reload): ...new pass_data instance and...
2506 (make_pass_reload): ...new function.
2507 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
2508 subclass of rtl_opt_pass along with...
2509 (pass_data_cleanup_barriers): ...new pass_data instance and...
2510 (make_pass_cleanup_barriers): ...new function.
2511 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
2512 of rtl_opt_pass along with...
2513 (pass_data_loop2): ...new pass_data instance and...
2514 (make_pass_loop2): ...new function.
2515 (pass_rtl_loop_init): Convert from a global struct to a subclass of
2516 rtl_opt_pass along with...
2517 (pass_data_rtl_loop_init): ...new pass_data instance and...
2518 (make_pass_rtl_loop_init): ...new function.
2519 (pass_rtl_loop_done): Convert from a global struct to a subclass of
2520 rtl_opt_pass along with...
2521 (pass_data_rtl_loop_done): ...new pass_data instance and...
2522 (make_pass_rtl_loop_done): ...new function.
2523 (pass_rtl_move_loop_invariants): Convert from a global struct to a
2524 subclass of rtl_opt_pass along with...
2525 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
2526 (make_pass_rtl_move_loop_invariants): ...new function.
2527 (pass_rtl_unswitch): Convert from a global struct to a subclass of
2528 rtl_opt_pass along with...
2529 (pass_data_rtl_unswitch): ...new pass_data instance and...
2530 (make_pass_rtl_unswitch): ...new function.
2531 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
2532 subclass of rtl_opt_pass along with...
2533 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
2535 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2536 (pass_rtl_doloop): Convert from a global struct to a subclass of
2537 rtl_opt_pass along with...
2538 (pass_data_rtl_doloop): ...new pass_data instance and...
2539 (make_pass_rtl_doloop): ...new function.
2540 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
2541 a subclass of rtl_opt_pass along with...
2542 (pass_data_lower_subreg): ...new pass_data instance and...
2543 (make_pass_lower_subreg): ...new function.
2544 (pass_lower_subreg2): Convert from a global struct to a subclass of
2545 rtl_opt_pass along with...
2546 (pass_data_lower_subreg2): ...new pass_data instance and...
2547 (make_pass_lower_subreg2): ...new function.
2548 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
2549 struct to a subclass of ipa_opt_pass_d along with...
2550 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
2551 (make_pass_ipa_lto_gimple_out): ...new function.
2552 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
2553 of ipa_opt_pass_d along with...
2554 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
2555 (make_pass_ipa_lto_finish_out): ...new function.
2556 * mode-switching.c (pass_mode_switching): Convert from a global struct
2557 to a subclass of rtl_opt_pass along with...
2558 (pass_data_mode_switching): ...new pass_data instance and...
2559 (make_pass_mode_switching): ...new function.
2560 * modulo-sched.c (pass_sms): Convert from a global struct to a
2561 subclass of rtl_opt_pass along with...
2562 (pass_data_sms): ...new pass_data instance and...
2563 (make_pass_sms): ...new function.
2564 * omp-low.c (pass_expand_omp): Convert from a global struct to a
2565 subclass of gimple_opt_pass along with...
2566 (pass_data_expand_omp): ...new pass_data instance and...
2567 (make_pass_expand_omp): ...new function.
2568 (pass_lower_omp): Convert from a global struct to a subclass of
2569 gimple_opt_pass along with...
2570 (pass_data_lower_omp): ...new pass_data instance and...
2571 (make_pass_lower_omp): ...new function.
2572 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
2573 of gimple_opt_pass along with...
2574 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
2575 (make_pass_diagnose_omp_blocks): ...new function.
2576 * passes.c (pass_early_local_passes): Convert from a global struct to
2577 a subclass of simple_ipa_opt_pass along with...
2578 (pass_data_early_local_passes): ...new pass_data instance and...
2579 (make_pass_early_local_passes): ...new function.
2580 (pass_all_early_optimizations): Convert from a global struct to a
2581 subclass of gimple_opt_pass along with...
2582 (pass_data_all_early_optimizations): ...new pass_data instance and...
2583 (make_pass_all_early_optimizations): ...new function.
2584 (pass_all_optimizations): Convert from a global struct to a subclass
2585 of gimple_opt_pass along with...
2586 (pass_data_all_optimizations): ...new pass_data instance and...
2587 (make_pass_all_optimizations): ...new function.
2588 (pass_all_optimizations_g): Convert from a global struct to a subclass
2589 of gimple_opt_pass along with...
2590 (pass_data_all_optimizations_g): ...new pass_data instance and...
2591 (make_pass_all_optimizations_g): ...new function.
2592 (pass_rest_of_compilation): Convert from a global struct to a subclass
2593 of rtl_opt_pass along with...
2594 (pass_data_rest_of_compilation): ...new pass_data instance and...
2595 (make_pass_rest_of_compilation): ...new function.
2596 (pass_postreload): Convert from a global struct to a subclass of
2597 rtl_opt_pass along with...
2598 (pass_data_postreload): ...new pass_data instance and...
2599 (make_pass_postreload): ...new function.
2600 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
2601 subclass of rtl_opt_pass along with...
2602 (pass_data_gcse2): ...new pass_data instance and...
2603 (make_pass_gcse2): ...new function.
2604 * postreload.c (pass_postreload_cse): Convert from a global struct to
2605 a subclass of rtl_opt_pass along with...
2606 (pass_data_postreload_cse): ...new pass_data instance and...
2607 (make_pass_postreload_cse): ...new function.
2608 * predict.c (pass_profile): Convert from a global struct to a subclass
2609 of gimple_opt_pass along with...
2610 (pass_data_profile): ...new pass_data instance and...
2611 (make_pass_profile): ...new function.
2612 (pass_strip_predict_hints): Convert from a global struct to a subclass
2613 of gimple_opt_pass along with...
2614 (pass_data_strip_predict_hints): ...new pass_data instance and...
2615 (make_pass_strip_predict_hints): ...new function.
2616 * recog.c (pass_peephole2): Convert from a global struct to a subclass
2617 of rtl_opt_pass along with...
2618 (pass_data_peephole2): ...new pass_data instance and...
2619 (make_pass_peephole2): ...new function.
2620 (pass_split_all_insns): Convert from a global struct to a subclass of
2621 rtl_opt_pass along with...
2622 (pass_data_split_all_insns): ...new pass_data instance and...
2623 (make_pass_split_all_insns): ...new function.
2624 (pass_split_after_reload): Convert from a global struct to a subclass
2625 of rtl_opt_pass along with...
2626 (pass_data_split_after_reload): ...new pass_data instance and...
2627 (make_pass_split_after_reload): ...new function.
2628 (pass_split_before_regstack): Convert from a global struct to a
2629 subclass of rtl_opt_pass along with...
2630 (pass_data_split_before_regstack): ...new pass_data instance and...
2631 (make_pass_split_before_regstack): ...new function.
2632 (pass_split_before_sched2): Convert from a global struct to a subclass
2633 of rtl_opt_pass along with...
2634 (pass_data_split_before_sched2): ...new pass_data instance and...
2635 (make_pass_split_before_sched2): ...new function.
2636 (pass_split_for_shorten_branches): Convert from a global struct to a
2637 subclass of rtl_opt_pass along with...
2638 (pass_data_split_for_shorten_branches): ...new pass_data instance
2640 (make_pass_split_for_shorten_branches): ...new function.
2641 * ree.c (pass_ree): Convert from a global struct to a subclass of
2642 rtl_opt_pass along with...
2643 (pass_data_ree): ...new pass_data instance and...
2644 (make_pass_ree): ...new function.
2645 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
2646 subclass of rtl_opt_pass along with...
2647 (pass_data_stack_regs): ...new pass_data instance and...
2648 (make_pass_stack_regs): ...new function.
2649 (pass_stack_regs_run): Convert from a global struct to a subclass of
2650 rtl_opt_pass along with...
2651 (pass_data_stack_regs_run): ...new pass_data instance and...
2652 (make_pass_stack_regs_run): ...new function.
2653 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
2654 subclass of rtl_opt_pass along with...
2655 (pass_data_cprop_hardreg): ...new pass_data instance and...
2656 (make_pass_cprop_hardreg): ...new function.
2657 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
2658 subclass of rtl_opt_pass along with...
2659 (pass_data_reginfo_init): ...new pass_data instance and...
2660 (make_pass_reginfo_init): ...new function.
2661 * regmove.c (pass_regmove): Convert from a global struct to a subclass
2662 of rtl_opt_pass along with...
2663 (pass_data_regmove): ...new pass_data instance and...
2664 (make_pass_regmove): ...new function.
2665 * regrename.c (pass_regrename): Convert from a global struct to a
2666 subclass of rtl_opt_pass along with...
2667 (pass_data_regrename): ...new pass_data instance and...
2668 (make_pass_regrename): ...new function.
2669 * reorg.c (pass_delay_slots): Convert from a global struct to a
2670 subclass of rtl_opt_pass along with...
2671 (pass_data_delay_slots): ...new pass_data instance and...
2672 (make_pass_delay_slots): ...new function.
2673 (pass_machine_reorg): Convert from a global struct to a subclass of
2674 rtl_opt_pass along with...
2675 (pass_data_machine_reorg): ...new pass_data instance and...
2676 (make_pass_machine_reorg): ...new function.
2677 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
2678 of rtl_opt_pass along with...
2679 (pass_data_sched): ...new pass_data instance and...
2680 (make_pass_sched): ...new function.
2681 (pass_sched2): Convert from a global struct to a subclass of
2682 rtl_opt_pass along with...
2683 (pass_data_sched2): ...new pass_data instance and...
2684 (make_pass_sched2): ...new function.
2685 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
2686 to a subclass of rtl_opt_pass along with...
2687 (pass_data_stack_ptr_mod): ...new pass_data instance and...
2688 (make_pass_stack_ptr_mod): ...new function.
2689 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
2690 to a subclass of rtl_opt_pass along with...
2691 (pass_data_rtl_store_motion): ...new pass_data instance and...
2692 (make_pass_rtl_store_motion): ...new function.
2693 * tracer.c (pass_tracer): Convert from a global struct to a subclass
2694 of gimple_opt_pass along with...
2695 (pass_data_tracer): ...new pass_data instance and...
2696 (make_pass_tracer): ...new function.
2697 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
2698 to a subclass of gimple_opt_pass along with...
2699 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
2700 (make_pass_diagnose_tm_blocks): ...new function.
2701 (pass_lower_tm): Convert from a global struct to a subclass of
2702 gimple_opt_pass along with...
2703 (pass_data_lower_tm): ...new pass_data instance and...
2704 (make_pass_lower_tm): ...new function.
2705 (pass_tm_init): Convert from a global struct to a subclass of
2706 gimple_opt_pass along with...
2707 (pass_data_tm_init): ...new pass_data instance and...
2708 (make_pass_tm_init): ...new function.
2709 (pass_tm_mark): Convert from a global struct to a subclass of
2710 gimple_opt_pass along with...
2711 (pass_data_tm_mark): ...new pass_data instance and...
2712 (make_pass_tm_mark): ...new function.
2713 (pass_tm_edges): Convert from a global struct to a subclass of
2714 gimple_opt_pass along with...
2715 (pass_data_tm_edges): ...new pass_data instance and...
2716 (make_pass_tm_edges): ...new function.
2717 (pass_tm_memopt): Convert from a global struct to a subclass of
2718 gimple_opt_pass along with...
2719 (pass_data_tm_memopt): ...new pass_data instance and...
2720 (make_pass_tm_memopt): ...new function.
2721 (pass_ipa_tm): Convert from a global struct to a subclass of
2722 simple_ipa_opt_pass along with...
2723 (pass_data_ipa_tm): ...new pass_data instance and...
2724 (make_pass_ipa_tm): ...new function.
2725 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
2726 subclass of gimple_opt_pass along with...
2727 (pass_data_call_cdce): ...new pass_data instance and...
2728 (make_pass_call_cdce): ...new function.
2729 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
2730 subclass of gimple_opt_pass along with...
2731 (pass_data_build_cfg): ...new pass_data instance and...
2732 (make_pass_build_cfg): ...new function.
2733 (pass_split_crit_edges): Convert from a global struct to a subclass of
2734 gimple_opt_pass along with...
2735 (pass_data_split_crit_edges): ...new pass_data instance and...
2736 (make_pass_split_crit_edges): ...new function.
2737 (pass_warn_function_return): Convert from a global struct to a
2738 subclass of gimple_opt_pass along with...
2739 (pass_data_warn_function_return): ...new pass_data instance and...
2740 (make_pass_warn_function_return): ...new function.
2741 (pass_warn_function_noreturn): Convert from a global struct to a
2742 subclass of gimple_opt_pass along with...
2743 (pass_data_warn_function_noreturn): ...new pass_data instance and...
2744 (make_pass_warn_function_noreturn): ...new function.
2745 (pass_warn_unused_result): Convert from a global struct to a subclass
2746 of gimple_opt_pass along with...
2747 (pass_data_warn_unused_result): ...new pass_data instance and...
2748 (make_pass_warn_unused_result): ...new function.
2749 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
2750 a subclass of gimple_opt_pass along with...
2751 (pass_data_merge_phi): ...new pass_data instance and...
2752 (make_pass_merge_phi): ...new function.
2753 * tree-complex.c (pass_lower_complex): Convert from a global struct to
2754 a subclass of gimple_opt_pass along with...
2755 (pass_data_lower_complex): ...new pass_data instance and...
2756 (make_pass_lower_complex): ...new function.
2757 (pass_lower_complex_O0): Convert from a global struct to a subclass of
2758 gimple_opt_pass along with...
2759 (pass_data_lower_complex_O0): ...new pass_data instance and...
2760 (make_pass_lower_complex_O0): ...new function.
2761 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
2762 subclass of gimple_opt_pass along with...
2763 (pass_data_lower_eh): ...new pass_data instance and...
2764 (make_pass_lower_eh): ...new function.
2765 (pass_refactor_eh): Convert from a global struct to a subclass of
2766 gimple_opt_pass along with...
2767 (pass_data_refactor_eh): ...new pass_data instance and...
2768 (make_pass_refactor_eh): ...new function.
2769 (pass_lower_resx): Convert from a global struct to a subclass of
2770 gimple_opt_pass along with...
2771 (pass_data_lower_resx): ...new pass_data instance and...
2772 (make_pass_lower_resx): ...new function.
2773 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
2774 of gimple_opt_pass along with...
2775 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
2776 (make_pass_lower_eh_dispatch): ...new function.
2777 (pass_cleanup_eh): Convert from a global struct to a subclass of
2778 gimple_opt_pass along with...
2779 (pass_data_cleanup_eh): ...new pass_data instance and...
2780 (make_pass_cleanup_eh): ...new function.
2781 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
2782 to a subclass of simple_ipa_opt_pass along with...
2783 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
2784 (make_pass_ipa_lower_emutls): ...new function.
2785 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
2786 a subclass of gimple_opt_pass along with...
2787 (pass_data_if_conversion): ...new pass_data instance and...
2788 (make_pass_if_conversion): ...new function.
2789 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
2790 subclass of gimple_opt_pass along with...
2791 (pass_data_build_ssa): ...new pass_data instance and...
2792 (make_pass_build_ssa): ...new function.
2793 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
2794 global struct to a subclass of gimple_opt_pass along with...
2795 (pass_data_loop_distribution): ...new pass_data instance and...
2796 (make_pass_loop_distribution): ...new function.
2797 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
2798 subclass of gimple_opt_pass along with...
2799 (pass_data_mudflap_1): ...new pass_data instance and...
2800 (make_pass_mudflap_1): ...new function.
2801 (pass_mudflap_2): Convert from a global struct to a subclass of
2802 gimple_opt_pass along with...
2803 (pass_data_mudflap_2): ...new pass_data instance and...
2804 (make_pass_mudflap_2): ...new function.
2805 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
2806 subclass of gimple_opt_pass along with...
2807 (pass_data_mudflap_1): ...new pass_data instance and...
2808 (make_pass_mudflap_1): ...new function.
2809 (pass_mudflap_2): Convert from a global struct to a subclass of
2810 gimple_opt_pass along with...
2811 (pass_data_mudflap_2): ...new pass_data instance and...
2812 (make_pass_mudflap_2): ...new function.
2813 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
2814 gimple_opt_pass along with...
2815 (pass_data_nrv): ...new pass_data instance and...
2816 (make_pass_nrv): ...new function.
2817 (pass_return_slot): Convert from a global struct to a subclass of
2818 gimple_opt_pass along with...
2819 (pass_data_return_slot): ...new pass_data instance and...
2820 (make_pass_return_slot): ...new function.
2821 * tree-object-size.c (pass_object_sizes): Convert from a global struct
2822 to a subclass of gimple_opt_pass along with...
2823 (pass_data_object_sizes): ...new pass_data instance and...
2824 (make_pass_object_sizes): ...new function.
2825 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
2826 global struct to a subclass of gimple_opt_pass along with...
2827 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
2829 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2830 (pass_fixup_cfg): Convert from a global struct to a subclass of
2831 gimple_opt_pass along with...
2832 (pass_data_fixup_cfg): ...new pass_data instance and...
2833 (make_pass_fixup_cfg): ...new function.
2834 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
2835 (make_pass_mudflap_1): ...new function.
2836 (pass_mudflap_2): Replace declaration with that of...
2837 (make_pass_mudflap_2): ...new function.
2838 (pass_asan): Replace declaration with that of...
2839 (make_pass_asan): ...new function.
2840 (pass_asan_O0): Replace declaration with that of...
2841 (make_pass_asan_O0): ...new function.
2842 (pass_tsan): Replace declaration with that of...
2843 (make_pass_tsan): ...new function.
2844 (pass_tsan_O0): Replace declaration with that of...
2845 (make_pass_tsan_O0): ...new function.
2846 (pass_lower_cf): Replace declaration with that of...
2847 (make_pass_lower_cf): ...new function.
2848 (pass_refactor_eh): Replace declaration with that of...
2849 (make_pass_refactor_eh): ...new function.
2850 (pass_lower_eh): Replace declaration with that of...
2851 (make_pass_lower_eh): ...new function.
2852 (pass_lower_eh_dispatch): Replace declaration with that of...
2853 (make_pass_lower_eh_dispatch): ...new function.
2854 (pass_lower_resx): Replace declaration with that of...
2855 (make_pass_lower_resx): ...new function.
2856 (pass_build_cfg): Replace declaration with that of...
2857 (make_pass_build_cfg): ...new function.
2858 (pass_early_tree_profile): Replace declaration with that of...
2859 (make_pass_early_tree_profile): ...new function.
2860 (pass_cleanup_eh): Replace declaration with that of...
2861 (make_pass_cleanup_eh): ...new function.
2862 (pass_sra): Replace declaration with that of...
2863 (make_pass_sra): ...new function.
2864 (pass_sra_early): Replace declaration with that of...
2865 (make_pass_sra_early): ...new function.
2866 (pass_early_ipa_sra): Replace declaration with that of...
2867 (make_pass_early_ipa_sra): ...new function.
2868 (pass_tail_recursion): Replace declaration with that of...
2869 (make_pass_tail_recursion): ...new function.
2870 (pass_tail_calls): Replace declaration with that of...
2871 (make_pass_tail_calls): ...new function.
2872 (pass_tree_loop): Replace declaration with that of...
2873 (make_pass_tree_loop): ...new function.
2874 (pass_tree_loop_init): Replace declaration with that of...
2875 (make_pass_tree_loop_init): ...new function.
2876 (pass_lim): Replace declaration with that of...
2877 (make_pass_lim): ...new function.
2878 (pass_tree_unswitch): Replace declaration with that of...
2879 (make_pass_tree_unswitch): ...new function.
2880 (pass_predcom): Replace declaration with that of...
2881 (make_pass_predcom): ...new function.
2882 (pass_iv_canon): Replace declaration with that of...
2883 (make_pass_iv_canon): ...new function.
2884 (pass_scev_cprop): Replace declaration with that of...
2885 (make_pass_scev_cprop): ...new function.
2886 (pass_empty_loop): Replace declaration with that of...
2887 (make_pass_empty_loop): ...new function.
2888 (pass_record_bounds): Replace declaration with that of...
2889 (make_pass_record_bounds): ...new function.
2890 (pass_graphite): Replace declaration with that of...
2891 (make_pass_graphite): ...new function.
2892 (pass_graphite_transforms): Replace declaration with that of...
2893 (make_pass_graphite_transforms): ...new function.
2894 (pass_if_conversion): Replace declaration with that of...
2895 (make_pass_if_conversion): ...new function.
2896 (pass_loop_distribution): Replace declaration with that of...
2897 (make_pass_loop_distribution): ...new function.
2898 (pass_vectorize): Replace declaration with that of...
2899 (make_pass_vectorize): ...new function.
2900 (pass_slp_vectorize): Replace declaration with that of...
2901 (make_pass_slp_vectorize): ...new function.
2902 (pass_complete_unroll): Replace declaration with that of...
2903 (make_pass_complete_unroll): ...new function.
2904 (pass_complete_unrolli): Replace declaration with that of...
2905 (make_pass_complete_unrolli): ...new function.
2906 (pass_parallelize_loops): Replace declaration with that of...
2907 (make_pass_parallelize_loops): ...new function.
2908 (pass_loop_prefetch): Replace declaration with that of...
2909 (make_pass_loop_prefetch): ...new function.
2910 (pass_iv_optimize): Replace declaration with that of...
2911 (make_pass_iv_optimize): ...new function.
2912 (pass_tree_loop_done): Replace declaration with that of...
2913 (make_pass_tree_loop_done): ...new function.
2914 (pass_ch): Replace declaration with that of...
2915 (make_pass_ch): ...new function.
2916 (pass_ccp): Replace declaration with that of...
2917 (make_pass_ccp): ...new function.
2918 (pass_phi_only_cprop): Replace declaration with that of...
2919 (make_pass_phi_only_cprop): ...new function.
2920 (pass_build_ssa): Replace declaration with that of...
2921 (make_pass_build_ssa): ...new function.
2922 (pass_build_alias): Replace declaration with that of...
2923 (make_pass_build_alias): ...new function.
2924 (pass_build_ealias): Replace declaration with that of...
2925 (make_pass_build_ealias): ...new function.
2926 (pass_dominator): Replace declaration with that of...
2927 (make_pass_dominator): ...new function.
2928 (pass_dce): Replace declaration with that of...
2929 (make_pass_dce): ...new function.
2930 (pass_dce_loop): Replace declaration with that of...
2931 (make_pass_dce_loop): ...new function.
2932 (pass_cd_dce): Replace declaration with that of...
2933 (make_pass_cd_dce): ...new function.
2934 (pass_call_cdce): Replace declaration with that of...
2935 (make_pass_call_cdce): ...new function.
2936 (pass_merge_phi): Replace declaration with that of...
2937 (make_pass_merge_phi): ...new function.
2938 (pass_split_crit_edges): Replace declaration with that of...
2939 (make_pass_split_crit_edges): ...new function.
2940 (pass_pre): Replace declaration with that of...
2941 (make_pass_pre): ...new function.
2942 (pass_profile): Replace declaration with that of...
2943 (make_pass_profile): ...new function.
2944 (pass_strip_predict_hints): Replace declaration with that of...
2945 (make_pass_strip_predict_hints): ...new function.
2946 (pass_lower_complex_O0): Replace declaration with that of...
2947 (make_pass_lower_complex_O0): ...new function.
2948 (pass_lower_complex): Replace declaration with that of...
2949 (make_pass_lower_complex): ...new function.
2950 (pass_lower_vector): Replace declaration with that of...
2951 (make_pass_lower_vector): ...new function.
2952 (pass_lower_vector_ssa): Replace declaration with that of...
2953 (make_pass_lower_vector_ssa): ...new function.
2954 (pass_lower_omp): Replace declaration with that of...
2955 (make_pass_lower_omp): ...new function.
2956 (pass_diagnose_omp_blocks): Replace declaration with that of...
2957 (make_pass_diagnose_omp_blocks): ...new function.
2958 (pass_expand_omp): Replace declaration with that of...
2959 (make_pass_expand_omp): ...new function.
2960 (pass_expand_omp_ssa): Replace declaration with that of...
2961 (make_pass_expand_omp_ssa): ...new function.
2962 (pass_object_sizes): Replace declaration with that of...
2963 (make_pass_object_sizes): ...new function.
2964 (pass_strlen): Replace declaration with that of...
2965 (make_pass_strlen): ...new function.
2966 (pass_fold_builtins): Replace declaration with that of...
2967 (make_pass_fold_builtins): ...new function.
2968 (pass_stdarg): Replace declaration with that of...
2969 (make_pass_stdarg): ...new function.
2970 (pass_early_warn_uninitialized): Replace declaration with that of...
2971 (make_pass_early_warn_uninitialized): ...new function.
2972 (pass_late_warn_uninitialized): Replace declaration with that of...
2973 (make_pass_late_warn_uninitialized): ...new function.
2974 (pass_cse_reciprocals): Replace declaration with that of...
2975 (make_pass_cse_reciprocals): ...new function.
2976 (pass_cse_sincos): Replace declaration with that of...
2977 (make_pass_cse_sincos): ...new function.
2978 (pass_optimize_bswap): Replace declaration with that of...
2979 (make_pass_optimize_bswap): ...new function.
2980 (pass_optimize_widening_mul): Replace declaration with that of...
2981 (make_pass_optimize_widening_mul): ...new function.
2982 (pass_warn_function_return): Replace declaration with that of...
2983 (make_pass_warn_function_return): ...new function.
2984 (pass_warn_function_noreturn): Replace declaration with that of...
2985 (make_pass_warn_function_noreturn): ...new function.
2986 (pass_cselim): Replace declaration with that of...
2987 (make_pass_cselim): ...new function.
2988 (pass_phiopt): Replace declaration with that of...
2989 (make_pass_phiopt): ...new function.
2990 (pass_forwprop): Replace declaration with that of...
2991 (make_pass_forwprop): ...new function.
2992 (pass_phiprop): Replace declaration with that of...
2993 (make_pass_phiprop): ...new function.
2994 (pass_tree_ifcombine): Replace declaration with that of...
2995 (make_pass_tree_ifcombine): ...new function.
2996 (pass_dse): Replace declaration with that of...
2997 (make_pass_dse): ...new function.
2998 (pass_nrv): Replace declaration with that of...
2999 (make_pass_nrv): ...new function.
3000 (pass_rename_ssa_copies): Replace declaration with that of...
3001 (make_pass_rename_ssa_copies): ...new function.
3002 (pass_sink_code): Replace declaration with that of...
3003 (make_pass_sink_code): ...new function.
3004 (pass_fre): Replace declaration with that of...
3005 (make_pass_fre): ...new function.
3006 (pass_check_data_deps): Replace declaration with that of...
3007 (make_pass_check_data_deps): ...new function.
3008 (pass_copy_prop): Replace declaration with that of...
3009 (make_pass_copy_prop): ...new function.
3010 (pass_vrp): Replace declaration with that of...
3011 (make_pass_vrp): ...new function.
3012 (pass_uncprop): Replace declaration with that of...
3013 (make_pass_uncprop): ...new function.
3014 (pass_return_slot): Replace declaration with that of...
3015 (make_pass_return_slot): ...new function.
3016 (pass_reassoc): Replace declaration with that of...
3017 (make_pass_reassoc): ...new function.
3018 (pass_rebuild_cgraph_edges): Replace declaration with that of...
3019 (make_pass_rebuild_cgraph_edges): ...new function.
3020 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
3021 (make_pass_remove_cgraph_callee_edges): ...new function.
3022 (pass_build_cgraph_edges): Replace declaration with that of...
3023 (make_pass_build_cgraph_edges): ...new function.
3024 (pass_local_pure_const): Replace declaration with that of...
3025 (make_pass_local_pure_const): ...new function.
3026 (pass_tracer): Replace declaration with that of...
3027 (make_pass_tracer): ...new function.
3028 (pass_warn_unused_result): Replace declaration with that of...
3029 (make_pass_warn_unused_result): ...new function.
3030 (pass_diagnose_tm_blocks): Replace declaration with that of...
3031 (make_pass_diagnose_tm_blocks): ...new function.
3032 (pass_lower_tm): Replace declaration with that of...
3033 (make_pass_lower_tm): ...new function.
3034 (pass_tm_init): Replace declaration with that of...
3035 (make_pass_tm_init): ...new function.
3036 (pass_tm_mark): Replace declaration with that of...
3037 (make_pass_tm_mark): ...new function.
3038 (pass_tm_memopt): Replace declaration with that of...
3039 (make_pass_tm_memopt): ...new function.
3040 (pass_tm_edges): Replace declaration with that of...
3041 (make_pass_tm_edges): ...new function.
3042 (pass_split_functions): Replace declaration with that of...
3043 (make_pass_split_functions): ...new function.
3044 (pass_feedback_split_functions): Replace declaration with that of...
3045 (make_pass_feedback_split_functions): ...new function.
3046 (pass_strength_reduction): Replace declaration with that of...
3047 (make_pass_strength_reduction): ...new function.
3048 (pass_ipa_lower_emutls): Replace declaration with that of...
3049 (make_pass_ipa_lower_emutls): ...new function.
3050 (pass_ipa_function_and_variable_visibility): Replace declaration with
3052 (make_pass_ipa_function_and_variable_visibility): ...new function.
3053 (pass_ipa_tree_profile): Replace declaration with that of...
3054 (make_pass_ipa_tree_profile): ...new function.
3055 (pass_early_local_passes): Replace declaration with that of...
3056 (make_pass_early_local_passes): ...new function.
3057 (pass_ipa_whole_program_visibility): Replace declaration with that
3059 (make_pass_ipa_whole_program_visibility): ...new function.
3060 (pass_ipa_lto_gimple_out): Replace declaration with that of...
3061 (make_pass_ipa_lto_gimple_out): ...new function.
3062 (pass_ipa_increase_alignment): Replace declaration with that of...
3063 (make_pass_ipa_increase_alignment): ...new function.
3064 (pass_ipa_inline): Replace declaration with that of...
3065 (make_pass_ipa_inline): ...new function.
3066 (pass_ipa_free_lang_data): Replace declaration with that of...
3067 (make_pass_ipa_free_lang_data): ...new function.
3068 (pass_ipa_free_inline_summary): Replace declaration with that of...
3069 (make_pass_ipa_free_inline_summary): ...new function.
3070 (pass_ipa_cp): Replace declaration with that of...
3071 (make_pass_ipa_cp): ...new function.
3072 (pass_ipa_reference): Replace declaration with that of...
3073 (make_pass_ipa_reference): ...new function.
3074 (pass_ipa_pure_const): Replace declaration with that of...
3075 (make_pass_ipa_pure_const): ...new function.
3076 (pass_ipa_pta): Replace declaration with that of...
3077 (make_pass_ipa_pta): ...new function.
3078 (pass_ipa_lto_finish_out): Replace declaration with that of...
3079 (make_pass_ipa_lto_finish_out): ...new function.
3080 (pass_ipa_tm): Replace declaration with that of...
3081 (make_pass_ipa_tm): ...new function.
3082 (pass_ipa_profile): Replace declaration with that of...
3083 (make_pass_ipa_profile): ...new function.
3084 (pass_ipa_cdtor_merge): Replace declaration with that of...
3085 (make_pass_ipa_cdtor_merge): ...new function.
3086 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
3088 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3089 (pass_init_datastructures): Replace declaration with that of...
3090 (make_pass_init_datastructures): ...new function.
3091 (pass_fixup_cfg): Replace declaration with that of...
3092 (make_pass_fixup_cfg): ...new function.
3093 (pass_expand): Replace declaration with that of...
3094 (make_pass_expand): ...new function.
3095 (pass_instantiate_virtual_regs): Replace declaration with that of...
3096 (make_pass_instantiate_virtual_regs): ...new function.
3097 (pass_rtl_fwprop): Replace declaration with that of...
3098 (make_pass_rtl_fwprop): ...new function.
3099 (pass_rtl_fwprop_addr): Replace declaration with that of...
3100 (make_pass_rtl_fwprop_addr): ...new function.
3101 (pass_jump): Replace declaration with that of...
3102 (make_pass_jump): ...new function.
3103 (pass_jump2): Replace declaration with that of...
3104 (make_pass_jump2): ...new function.
3105 (pass_lower_subreg): Replace declaration with that of...
3106 (make_pass_lower_subreg): ...new function.
3107 (pass_cse): Replace declaration with that of...
3108 (make_pass_cse): ...new function.
3109 (pass_fast_rtl_dce): Replace declaration with that of...
3110 (make_pass_fast_rtl_dce): ...new function.
3111 (pass_ud_rtl_dce): Replace declaration with that of...
3112 (make_pass_ud_rtl_dce): ...new function.
3113 (pass_rtl_dce): Replace declaration with that of...
3114 (make_pass_rtl_dce): ...new function.
3115 (pass_rtl_dse1): Replace declaration with that of...
3116 (make_pass_rtl_dse1): ...new function.
3117 (pass_rtl_dse2): Replace declaration with that of...
3118 (make_pass_rtl_dse2): ...new function.
3119 (pass_rtl_dse3): Replace declaration with that of...
3120 (make_pass_rtl_dse3): ...new function.
3121 (pass_rtl_cprop): Replace declaration with that of...
3122 (make_pass_rtl_cprop): ...new function.
3123 (pass_rtl_pre): Replace declaration with that of...
3124 (make_pass_rtl_pre): ...new function.
3125 (pass_rtl_hoist): Replace declaration with that of...
3126 (make_pass_rtl_hoist): ...new function.
3127 (pass_rtl_store_motion): Replace declaration with that of...
3128 (make_pass_rtl_store_motion): ...new function.
3129 (pass_cse_after_global_opts): Replace declaration with that of...
3130 (make_pass_cse_after_global_opts): ...new function.
3131 (pass_rtl_ifcvt): Replace declaration with that of...
3132 (make_pass_rtl_ifcvt): ...new function.
3133 (pass_into_cfg_layout_mode): Replace declaration with that of...
3134 (make_pass_into_cfg_layout_mode): ...new function.
3135 (pass_outof_cfg_layout_mode): Replace declaration with that of...
3136 (make_pass_outof_cfg_layout_mode): ...new function.
3137 (pass_loop2): Replace declaration with that of...
3138 (make_pass_loop2): ...new function.
3139 (pass_rtl_loop_init): Replace declaration with that of...
3140 (make_pass_rtl_loop_init): ...new function.
3141 (pass_rtl_move_loop_invariants): Replace declaration with that of...
3142 (make_pass_rtl_move_loop_invariants): ...new function.
3143 (pass_rtl_unswitch): Replace declaration with that of...
3144 (make_pass_rtl_unswitch): ...new function.
3145 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
3146 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3147 (pass_rtl_doloop): Replace declaration with that of...
3148 (make_pass_rtl_doloop): ...new function.
3149 (pass_rtl_loop_done): Replace declaration with that of...
3150 (make_pass_rtl_loop_done): ...new function.
3151 (pass_web): Replace declaration with that of...
3152 (make_pass_web): ...new function.
3153 (pass_cse2): Replace declaration with that of...
3154 (make_pass_cse2): ...new function.
3155 (pass_df_initialize_opt): Replace declaration with that of...
3156 (make_pass_df_initialize_opt): ...new function.
3157 (pass_df_initialize_no_opt): Replace declaration with that of...
3158 (make_pass_df_initialize_no_opt): ...new function.
3159 (pass_reginfo_init): Replace declaration with that of...
3160 (make_pass_reginfo_init): ...new function.
3161 (pass_inc_dec): Replace declaration with that of...
3162 (make_pass_inc_dec): ...new function.
3163 (pass_stack_ptr_mod): Replace declaration with that of...
3164 (make_pass_stack_ptr_mod): ...new function.
3165 (pass_initialize_regs): Replace declaration with that of...
3166 (make_pass_initialize_regs): ...new function.
3167 (pass_combine): Replace declaration with that of...
3168 (make_pass_combine): ...new function.
3169 (pass_if_after_combine): Replace declaration with that of...
3170 (make_pass_if_after_combine): ...new function.
3171 (pass_ree): Replace declaration with that of...
3172 (make_pass_ree): ...new function.
3173 (pass_partition_blocks): Replace declaration with that of...
3174 (make_pass_partition_blocks): ...new function.
3175 (pass_match_asm_constraints): Replace declaration with that of...
3176 (make_pass_match_asm_constraints): ...new function.
3177 (pass_regmove): Replace declaration with that of...
3178 (make_pass_regmove): ...new function.
3179 (pass_split_all_insns): Replace declaration with that of...
3180 (make_pass_split_all_insns): ...new function.
3181 (pass_fast_rtl_byte_dce): Replace declaration with that of...
3182 (make_pass_fast_rtl_byte_dce): ...new function.
3183 (pass_lower_subreg2): Replace declaration with that of...
3184 (make_pass_lower_subreg2): ...new function.
3185 (pass_mode_switching): Replace declaration with that of...
3186 (make_pass_mode_switching): ...new function.
3187 (pass_sms): Replace declaration with that of...
3188 (make_pass_sms): ...new function.
3189 (pass_sched): Replace declaration with that of...
3190 (make_pass_sched): ...new function.
3191 (pass_ira): Replace declaration with that of...
3192 (make_pass_ira): ...new function.
3193 (pass_reload): Replace declaration with that of...
3194 (make_pass_reload): ...new function.
3195 (pass_clean_state): Replace declaration with that of...
3196 (make_pass_clean_state): ...new function.
3197 (pass_branch_prob): Replace declaration with that of...
3198 (make_pass_branch_prob): ...new function.
3199 (pass_value_profile_transformations): Replace declaration with that
3201 (make_pass_value_profile_transformations): ...new function.
3202 (pass_postreload_cse): Replace declaration with that of...
3203 (make_pass_postreload_cse): ...new function.
3204 (pass_gcse2): Replace declaration with that of...
3205 (make_pass_gcse2): ...new function.
3206 (pass_split_after_reload): Replace declaration with that of...
3207 (make_pass_split_after_reload): ...new function.
3208 (pass_branch_target_load_optimize1): Replace declaration with that
3210 (make_pass_branch_target_load_optimize1): ...new function.
3211 (pass_thread_prologue_and_epilogue): Replace declaration with that
3213 (make_pass_thread_prologue_and_epilogue): ...new function.
3214 (pass_stack_adjustments): Replace declaration with that of...
3215 (make_pass_stack_adjustments): ...new function.
3216 (pass_peephole2): Replace declaration with that of...
3217 (make_pass_peephole2): ...new function.
3218 (pass_if_after_reload): Replace declaration with that of...
3219 (make_pass_if_after_reload): ...new function.
3220 (pass_regrename): Replace declaration with that of...
3221 (make_pass_regrename): ...new function.
3222 (pass_cprop_hardreg): Replace declaration with that of...
3223 (make_pass_cprop_hardreg): ...new function.
3224 (pass_reorder_blocks): Replace declaration with that of...
3225 (make_pass_reorder_blocks): ...new function.
3226 (pass_branch_target_load_optimize2): Replace declaration with that
3228 (make_pass_branch_target_load_optimize2): ...new function.
3229 (pass_leaf_regs): Replace declaration with that of...
3230 (make_pass_leaf_regs): ...new function.
3231 (pass_split_before_sched2): Replace declaration with that of...
3232 (make_pass_split_before_sched2): ...new function.
3233 (pass_compare_elim_after_reload): Replace declaration with that of...
3234 (make_pass_compare_elim_after_reload): ...new function.
3235 (pass_sched2): Replace declaration with that of...
3236 (make_pass_sched2): ...new function.
3237 (pass_stack_regs): Replace declaration with that of...
3238 (make_pass_stack_regs): ...new function.
3239 (pass_stack_regs_run): Replace declaration with that of...
3240 (make_pass_stack_regs_run): ...new function.
3241 (pass_df_finish): Replace declaration with that of...
3242 (make_pass_df_finish): ...new function.
3243 (pass_compute_alignments): Replace declaration with that of...
3244 (make_pass_compute_alignments): ...new function.
3245 (pass_duplicate_computed_gotos): Replace declaration with that of...
3246 (make_pass_duplicate_computed_gotos): ...new function.
3247 (pass_variable_tracking): Replace declaration with that of...
3248 (make_pass_variable_tracking): ...new function.
3249 (pass_free_cfg): Replace declaration with that of...
3250 (make_pass_free_cfg): ...new function.
3251 (pass_machine_reorg): Replace declaration with that of...
3252 (make_pass_machine_reorg): ...new function.
3253 (pass_cleanup_barriers): Replace declaration with that of...
3254 (make_pass_cleanup_barriers): ...new function.
3255 (pass_delay_slots): Replace declaration with that of...
3256 (make_pass_delay_slots): ...new function.
3257 (pass_split_for_shorten_branches): Replace declaration with that of...
3258 (make_pass_split_for_shorten_branches): ...new function.
3259 (pass_split_before_regstack): Replace declaration with that of...
3260 (make_pass_split_before_regstack): ...new function.
3261 (pass_convert_to_eh_region_ranges): Replace declaration with that
3263 (make_pass_convert_to_eh_region_ranges): ...new function.
3264 (pass_shorten_branches): Replace declaration with that of...
3265 (make_pass_shorten_branches): ...new function.
3266 (pass_set_nothrow_function_flags): Replace declaration with that of...
3267 (make_pass_set_nothrow_function_flags): ...new function.
3268 (pass_dwarf2_frame): Replace declaration with that of...
3269 (make_pass_dwarf2_frame): ...new function.
3270 (pass_final): Replace declaration with that of...
3271 (make_pass_final): ...new function.
3272 (pass_rtl_seqabstr): Replace declaration with that of...
3273 (make_pass_rtl_seqabstr): ...new function.
3274 (pass_release_ssa_names): Replace declaration with that of...
3275 (make_pass_release_ssa_names): ...new function.
3276 (pass_early_inline): Replace declaration with that of...
3277 (make_pass_early_inline): ...new function.
3278 (pass_inline_parameters): Replace declaration with that of...
3279 (make_pass_inline_parameters): ...new function.
3280 (pass_update_address_taken): Replace declaration with that of...
3281 (make_pass_update_address_taken): ...new function.
3282 (pass_convert_switch): Replace declaration with that of...
3283 (make_pass_convert_switch): ...new function.
3284 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
3285 to a subclass of simple_ipa_opt_pass along with...
3286 (pass_data_ipa_tree_profile): ...new pass_data instance and...
3287 (make_pass_ipa_tree_profile): ...new function.
3288 * tree-sra.c (pass_sra_early): Convert from a global struct to a
3289 subclass of gimple_opt_pass along with...
3290 (pass_data_sra_early): ...new pass_data instance and...
3291 (make_pass_sra_early): ...new function.
3292 (pass_sra): Convert from a global struct to a subclass of
3293 gimple_opt_pass along with...
3294 (pass_data_sra): ...new pass_data instance and...
3295 (make_pass_sra): ...new function.
3296 (pass_early_ipa_sra): Convert from a global struct to a subclass of
3297 gimple_opt_pass along with...
3298 (pass_data_early_ipa_sra): ...new pass_data instance and...
3299 (make_pass_early_ipa_sra): ...new function.
3300 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
3301 subclass of gimple_opt_pass along with...
3302 (pass_data_ccp): ...new pass_data instance and...
3303 (make_pass_ccp): ...new function.
3304 (pass_fold_builtins): Convert from a global struct to a subclass of
3305 gimple_opt_pass along with...
3306 (pass_data_fold_builtins): ...new pass_data instance and...
3307 (make_pass_fold_builtins): ...new function.
3308 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
3309 subclass of gimple_opt_pass along with...
3310 (pass_data_copy_prop): ...new pass_data instance and...
3311 (make_pass_copy_prop): ...new function.
3312 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
3313 global struct to a subclass of gimple_opt_pass along with...
3314 (pass_data_rename_ssa_copies): ...new pass_data instance and...
3315 (make_pass_rename_ssa_copies): ...new function.
3316 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
3317 subclass of gimple_opt_pass along with...
3318 (pass_data_dce): ...new pass_data instance and...
3319 (make_pass_dce): ...new function.
3320 (pass_dce_loop): Convert from a global struct to a subclass of
3321 gimple_opt_pass along with...
3322 (pass_data_dce_loop): ...new pass_data instance and...
3323 (make_pass_dce_loop): ...new function.
3324 (pass_cd_dce): Convert from a global struct to a subclass of
3325 gimple_opt_pass along with...
3326 (pass_data_cd_dce): ...new pass_data instance and...
3327 (make_pass_cd_dce): ...new function.
3328 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
3329 subclass of gimple_opt_pass along with...
3330 (pass_data_dominator): ...new pass_data instance and...
3331 (make_pass_dominator): ...new function.
3332 (pass_phi_only_cprop): Convert from a global struct to a subclass of
3333 gimple_opt_pass along with...
3334 (pass_data_phi_only_cprop): ...new pass_data instance and...
3335 (make_pass_phi_only_cprop): ...new function.
3336 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
3337 subclass of gimple_opt_pass along with...
3338 (pass_data_dse): ...new pass_data instance and...
3339 (make_pass_dse): ...new function.
3340 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
3341 a subclass of gimple_opt_pass along with...
3342 (pass_data_forwprop): ...new pass_data instance and...
3343 (make_pass_forwprop): ...new function.
3344 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
3345 struct to a subclass of gimple_opt_pass along with...
3346 (pass_data_tree_ifcombine): ...new pass_data instance and...
3347 (make_pass_tree_ifcombine): ...new function.
3348 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
3349 subclass of gimple_opt_pass along with...
3350 (pass_data_ch): ...new pass_data instance and...
3351 (make_pass_ch): ...new function.
3352 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
3353 subclass of gimple_opt_pass along with...
3354 (pass_data_tree_loop): ...new pass_data instance and...
3355 (make_pass_tree_loop): ...new function.
3356 (pass_tree_loop_init): Convert from a global struct to a subclass of
3357 gimple_opt_pass along with...
3358 (pass_data_tree_loop_init): ...new pass_data instance and...
3359 (make_pass_tree_loop_init): ...new function.
3360 (pass_lim): Convert from a global struct to a subclass of
3361 gimple_opt_pass along with...
3362 (pass_data_lim): ...new pass_data instance and...
3363 (make_pass_lim): ...new function.
3364 (pass_tree_unswitch): Convert from a global struct to a subclass of
3365 gimple_opt_pass along with...
3366 (pass_data_tree_unswitch): ...new pass_data instance and...
3367 (make_pass_tree_unswitch): ...new function.
3368 (pass_predcom): Convert from a global struct to a subclass of
3369 gimple_opt_pass along with...
3370 (pass_data_predcom): ...new pass_data instance and...
3371 (make_pass_predcom): ...new function.
3372 (pass_vectorize): Convert from a global struct to a subclass of
3373 gimple_opt_pass along with...
3374 (pass_data_vectorize): ...new pass_data instance and...
3375 (make_pass_vectorize): ...new function.
3376 (pass_graphite): Convert from a global struct to a subclass of
3377 gimple_opt_pass along with...
3378 (pass_data_graphite): ...new pass_data instance and...
3379 (make_pass_graphite): ...new function.
3380 (pass_graphite_transforms): Convert from a global struct to a subclass
3381 of gimple_opt_pass along with...
3382 (pass_data_graphite_transforms): ...new pass_data instance and...
3383 (make_pass_graphite_transforms): ...new function.
3384 (pass_check_data_deps): Convert from a global struct to a subclass of
3385 gimple_opt_pass along with...
3386 (pass_data_check_data_deps): ...new pass_data instance and...
3387 (make_pass_check_data_deps): ...new function.
3388 (pass_iv_canon): Convert from a global struct to a subclass of
3389 gimple_opt_pass along with...
3390 (pass_data_iv_canon): ...new pass_data instance and...
3391 (make_pass_iv_canon): ...new function.
3392 (pass_scev_cprop): Convert from a global struct to a subclass of
3393 gimple_opt_pass along with...
3394 (pass_data_scev_cprop): ...new pass_data instance and...
3395 (make_pass_scev_cprop): ...new function.
3396 (pass_record_bounds): Convert from a global struct to a subclass of
3397 gimple_opt_pass along with...
3398 (pass_data_record_bounds): ...new pass_data instance and...
3399 (make_pass_record_bounds): ...new function.
3400 (pass_complete_unroll): Convert from a global struct to a subclass of
3401 gimple_opt_pass along with...
3402 (pass_data_complete_unroll): ...new pass_data instance and...
3403 (make_pass_complete_unroll): ...new function.
3404 (pass_complete_unrolli): Convert from a global struct to a subclass of
3405 gimple_opt_pass along with...
3406 (pass_data_complete_unrolli): ...new pass_data instance and...
3407 (make_pass_complete_unrolli): ...new function.
3408 (pass_parallelize_loops): Convert from a global struct to a subclass
3409 of gimple_opt_pass along with...
3410 (pass_data_parallelize_loops): ...new pass_data instance and...
3411 (make_pass_parallelize_loops): ...new function.
3412 (pass_loop_prefetch): Convert from a global struct to a subclass of
3413 gimple_opt_pass along with...
3414 (pass_data_loop_prefetch): ...new pass_data instance and...
3415 (make_pass_loop_prefetch): ...new function.
3416 (pass_iv_optimize): Convert from a global struct to a subclass of
3417 gimple_opt_pass along with...
3418 (pass_data_iv_optimize): ...new pass_data instance and...
3419 (make_pass_iv_optimize): ...new function.
3420 (pass_tree_loop_done): Convert from a global struct to a subclass of
3421 gimple_opt_pass along with...
3422 (pass_data_tree_loop_done): ...new pass_data instance and...
3423 (make_pass_tree_loop_done): ...new function.
3424 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
3425 struct to a subclass of gimple_opt_pass along with...
3426 (pass_data_cse_reciprocals): ...new pass_data instance and...
3427 (make_pass_cse_reciprocals): ...new function.
3428 (pass_cse_sincos): Convert from a global struct to a subclass of
3429 gimple_opt_pass along with...
3430 (pass_data_cse_sincos): ...new pass_data instance and...
3431 (make_pass_cse_sincos): ...new function.
3432 (pass_optimize_bswap): Convert from a global struct to a subclass of
3433 gimple_opt_pass along with...
3434 (pass_data_optimize_bswap): ...new pass_data instance and...
3435 (make_pass_optimize_bswap): ...new function.
3436 (pass_optimize_widening_mul): Convert from a global struct to a
3437 subclass of gimple_opt_pass along with...
3438 (pass_data_optimize_widening_mul): ...new pass_data instance and...
3439 (make_pass_optimize_widening_mul): ...new function.
3440 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
3441 subclass of gimple_opt_pass along with...
3442 (pass_data_phiopt): ...new pass_data instance and...
3443 (make_pass_phiopt): ...new function.
3444 (pass_cselim): Convert from a global struct to a subclass of
3445 gimple_opt_pass along with...
3446 (pass_data_cselim): ...new pass_data instance and...
3447 (make_pass_cselim): ...new function.
3448 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
3449 subclass of gimple_opt_pass along with...
3450 (pass_data_phiprop): ...new pass_data instance and...
3451 (make_pass_phiprop): ...new function.
3452 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
3453 subclass of gimple_opt_pass along with...
3454 (pass_data_pre): ...new pass_data instance and...
3455 (make_pass_pre): ...new function.
3456 (pass_fre): Convert from a global struct to a subclass of
3457 gimple_opt_pass along with...
3458 (pass_data_fre): ...new pass_data instance and...
3459 (make_pass_fre): ...new function.
3460 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
3461 subclass of gimple_opt_pass along with...
3462 (pass_data_reassoc): ...new pass_data instance and...
3463 (make_pass_reassoc): ...new function.
3464 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
3465 subclass of gimple_opt_pass along with...
3466 (pass_data_sink_code): ...new pass_data instance and...
3467 (make_pass_sink_code): ...new function.
3468 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
3469 subclass of gimple_opt_pass along with...
3470 (pass_data_strlen): ...new pass_data instance and...
3471 (make_pass_strlen): ...new function.
3472 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
3473 struct to a subclass of gimple_opt_pass along with...
3474 (pass_data_build_alias): ...new pass_data instance and...
3475 (make_pass_build_alias): ...new function.
3476 (pass_build_ealias): Convert from a global struct to a subclass of
3477 gimple_opt_pass along with...
3478 (pass_data_build_ealias): ...new pass_data instance and...
3479 (make_pass_build_ealias): ...new function.
3480 (pass_ipa_pta): Convert from a global struct to a subclass of
3481 simple_ipa_opt_pass along with...
3482 (pass_data_ipa_pta): ...new pass_data instance and...
3483 (make_pass_ipa_pta): ...new function.
3484 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
3485 subclass of gimple_opt_pass along with...
3486 (pass_data_uncprop): ...new pass_data instance and...
3487 (make_pass_uncprop): ...new function.
3488 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
3489 global struct to a subclass of gimple_opt_pass along with...
3490 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
3491 (make_pass_late_warn_uninitialized): ...new function.
3492 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
3493 to a subclass of gimple_opt_pass along with...
3494 (pass_data_init_datastructures): ...new pass_data instance and...
3495 (make_pass_init_datastructures): ...new function.
3496 (pass_early_warn_uninitialized): Convert from a global struct to a
3497 subclass of gimple_opt_pass along with...
3498 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
3499 (make_pass_early_warn_uninitialized): ...new function.
3500 (pass_update_address_taken): Convert from a global struct to a
3501 subclass of gimple_opt_pass along with...
3502 (pass_data_update_address_taken): ...new pass_data instance and...
3503 (make_pass_update_address_taken): ...new function.
3504 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
3505 struct to a subclass of gimple_opt_pass along with...
3506 (pass_data_release_ssa_names): ...new pass_data instance and...
3507 (make_pass_release_ssa_names): ...new function.
3508 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
3509 subclass of gimple_opt_pass along with...
3510 (pass_data_stdarg): ...new pass_data instance and...
3511 (make_pass_stdarg): ...new function.
3512 * tree-switch-conversion.c (pass_convert_switch): Convert from a
3513 global struct to a subclass of gimple_opt_pass along with...
3514 (pass_data_convert_switch): ...new pass_data instance and...
3515 (make_pass_convert_switch): ...new function.
3516 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
3517 to a subclass of gimple_opt_pass along with...
3518 (pass_data_tail_recursion): ...new pass_data instance and...
3519 (make_pass_tail_recursion): ...new function.
3520 (pass_tail_calls): Convert from a global struct to a subclass of
3521 gimple_opt_pass along with...
3522 (pass_data_tail_calls): ...new pass_data instance and...
3523 (make_pass_tail_calls): ...new function.
3524 * tree-vect-generic.c (pass_lower_vector): Convert from a global
3525 struct to a subclass of gimple_opt_pass along with...
3526 (pass_data_lower_vector): ...new pass_data instance and...
3527 (make_pass_lower_vector): ...new function.
3528 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
3529 gimple_opt_pass along with...
3530 (pass_data_lower_vector_ssa): ...new pass_data instance and...
3531 (make_pass_lower_vector_ssa): ...new function.
3532 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
3533 to a subclass of gimple_opt_pass along with...
3534 (pass_data_slp_vectorize): ...new pass_data instance and...
3535 (make_pass_slp_vectorize): ...new function.
3536 (pass_ipa_increase_alignment): Convert from a global struct to a
3537 subclass of simple_ipa_opt_pass along with...
3538 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
3539 (make_pass_ipa_increase_alignment): ...new function.
3540 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
3541 gimple_opt_pass along with...
3542 (pass_data_vrp): ...new pass_data instance and...
3543 (make_pass_vrp): ...new function.
3544 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
3545 subclass of simple_ipa_opt_pass along with...
3546 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
3547 (make_pass_ipa_free_lang_data): ...new function.
3548 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
3549 gimple_opt_pass along with...
3550 (pass_data_tsan): ...new pass_data instance and...
3551 (make_pass_tsan): ...new function.
3552 (pass_tsan_O0): Convert from a global struct to a subclass of
3553 gimple_opt_pass along with...
3554 (pass_data_tsan_O0): ...new pass_data instance and...
3555 (make_pass_tsan_O0): ...new function.
3556 * var-tracking.c (pass_variable_tracking): Convert from a global
3557 struct to a subclass of rtl_opt_pass along with...
3558 (pass_data_variable_tracking): ...new pass_data instance and...
3559 (make_pass_variable_tracking): ...new function.
3560 * web.c (pass_web): Convert from a global struct to a subclass of
3561 rtl_opt_pass along with...
3562 (pass_data_web): ...new pass_data instance and...
3563 (make_pass_web): ...new function.
3564 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
3565 declaration with that of...
3566 (make_pass_mode_switch_use): ...new function.
3567 (pass_resolve_sw_modes): Replace declaration with that of...
3568 (make_pass_resolve_sw_modes): ...new function.
3569 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
3570 from a global struct to a subclass of rtl_opt_pass along with...
3571 (pass_data_mode_switch_use): ...new pass_data instance and...
3572 (make_pass_mode_switch_use): ...new function.
3573 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
3574 from a global struct to a subclass of rtl_opt_pass along with...
3575 (pass_data_resolve_sw_modes): ...new pass_data instance and...
3576 (make_pass_resolve_sw_modes): ...new function.
3577 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
3578 struct to a subclass of rtl_opt_pass along with...
3579 (pass_data_insert_vzeroupper): ...new pass_data instance and...
3580 (make_pass_insert_vzeroupper): ...new function.
3581 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
3582 global struct to a subclass of rtl_opt_pass along with...
3583 (pass_data_work_around_errata): ...new pass_data instance and...
3584 (make_pass_work_around_errata): ...new function.
3585 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
3586 struct to a subclass of rtl_opt_pass along with...
3587 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
3588 (make_pass_mips_machine_reorg2): ...new function.
3590 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3592 * passes.c (pass_manager::operator new): New.
3594 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3596 Handwritten part of conversion of passes to C++ classes.
3598 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
3599 (toplev.o): Add dep on PASS_MANAGER_H.
3600 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
3601 of early local pases to reflect this moving from a global to a
3602 member of gcc::pass_manager.
3603 (cgraph_add_new_function): Likewise.
3604 * lto-cgraph.c (lto_output_node): Update for conversion of
3605 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
3606 * passes.c (opt_pass::clone): New.
3607 (opt_pass::gate): New.
3608 (opt_pass::execute): New.
3609 (opt_pass::opt_pass): New.
3610 (pass_manager::execute_early_local_passes): New.
3611 (pass_manager::execute_pass_mode_switching): new.
3612 (finish_optimization_passes): Convert to...
3613 (pass_manager::finish_optimization_passes): ...this.
3614 (finish_optimization_passes): Update for conversion of passes to
3616 (register_dump_files_1): Use has_gate since we cannot portably
3617 check a vtable entry against NULL.
3618 (dump_one_pass): Likewise.
3619 (ipa_write_summaries_2): Likewise.
3620 (ipa_write_optimization_summaries_1): Likewise.
3621 (ipa_read_summaries_1): Likewise.
3622 (ipa_read_optimization_summaries_1): Likewise.
3623 (execute_ipa_stmt_fixups): Likewise.
3624 (pass_manager::pass_manager): Rewrite pass-creation, invoking
3625 pass-creation functions rather than wiring up globals, and
3626 storing the results in fields of pass_manager generated using
3628 (pass_manager::dump_profile_report): Update for conversion of
3629 passes to C++ classes.
3630 (pass_manager::execute_ipa_summary_passes): Likewise.
3631 (execute_one_ipa_transform_pass): Likewise.
3632 (execute_one_pass): Use has_gate and has_execute since we cannot
3633 portably check a vtable entry against NULL.
3634 * pass_manager.h (pass_manager::finish_optimization_passes): New.
3635 (pass_manager): Use pass-instances.def to add fields for the
3636 various pass instances.
3637 * toplev.c (finalize): Update for move of
3638 finish_optimization_passes to a method of gcc::pass_manager.
3639 * toplev.h (finish_optimization_passes): Move to method of class
3641 * tree-pass.h (struct pass_data): New.
3642 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
3643 (opt_pass::gate): Convert to virtual function.
3644 (opt_pass::~opt_pass): New.
3645 (opt_pass::clone): New.
3646 (opt_pass::execute): Convert to virtual function.
3647 (opt_pass::opt_pass): New.
3648 (opt_pass::ctxt_): new.
3649 (gimple_opt_pass): Convert to subclass of opt_pass.
3650 (gimple_opt_pass::gimple_opt_pass): New.
3651 (rtl_opt_pass): Convert to subclass of opt_pass.
3652 (rtl_opt_pass::rtl_opt_pass): New.
3653 (ipa_opt_pass_d): Convert to subclass of opt_pass.
3654 (ipa_opt_pass_d::ipa_opt_pass_d): New.
3655 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
3656 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
3657 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
3658 invocation of pass_mode_switching to reflect this moving from a
3659 global to a member of gcc::pass_manager.
3660 (ix86_option_override): Rework how pass_insert_vzeroupper is
3661 added to the pass_manager to reflect autogenerated changes.
3662 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
3664 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
3666 PR rtl-optimization/57708
3667 * recog.c (peep2_find_free_register): Validate all regs in a
3670 2013-08-05 Jan Hubicka <jh@suse.cz>
3673 * cgraph.c (verify_cgraph_node): Accept local flags from other
3675 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
3676 (function_and_variable_visibility): Likewise.
3677 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
3679 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3681 * graph.c (init_graph_slim_pretty_print): Remove.
3682 (print_graph_cfg): Do not call it. Use local pretty printer.
3683 (start_graph_dump): Likewise.
3685 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3687 * gimple-pretty-print.c (buffer): Remove.
3688 (initialized): Likewise.
3689 (maybe_init_pretty_print): Likewise.
3690 (print_gimple_stmt): Do not call it. Use non-static local
3691 pretty_printer variable.
3692 (print_gimple_expr): Likewise.
3693 (print_gimple_seq): Likewise.
3694 (gimple_dump_bb): Likewise.
3696 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3698 * asan.c (asan_pp): Remove.
3699 (asan_pp_initialized): Likewise.
3700 (asan_pp_initialize): Likewise.
3701 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
3702 (asan_emit_stack_protection): Tidy. Use local pretty printer.
3703 (asan_add_global): Likewise.
3705 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
3707 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
3708 * diagnostic.h (diagnostic_flush_buffer): Adjust.
3709 * pretty-print.c (pp_formatted_text_data): Likewise.
3710 (pp_indent): Rename from pp_base_indent.
3711 (pp_format): Rename from pp_base_format.
3712 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
3713 (pp_format_verbatim): Rename from pp_base_format_verbatim.
3714 (pp_flush): Rename from pp_base_flush.
3715 (pp_set_line_maximum_length): Rename from
3716 pp_base_set_line_maximum_length.
3717 (pp_clear_output_area): Rename from pp_base_clear_output_area.
3718 (pp_set_prefix): Rename from pp_base_set_prefix.
3719 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
3720 (pp_emit_prefix): Rename from pp_base_emit_prefix.
3721 (pp_append_text): Rename from pp_base_append_text.
3722 (pp_formatted_text): Rename from pp_base_formatted_text.
3723 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
3724 (pp_remaining_character_count_for_line): Rename from
3725 pp_base_remaining_character_count_for_line.
3726 (pp_newline): Rename from pp_base_newline.
3727 (pp_character): Rename from pp_base_character.
3728 (pp_string): Rename from pp_base_string.
3729 (pp_maybe_space): Rename from pp_base_maybe_space.
3730 * asan.c (asan_pp_string): Adjust.
3731 (asan_emit_stack_protection): Likewise.
3732 (asan_add_global): Likewise.
3733 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
3734 * tree-mudflap.c (mf_varname_tree): Likewise.
3735 * tree-pretty-print.c (pp_tree_identifier): Rename from
3736 pp_base_tree_identifier.
3737 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
3738 Declare as function.
3740 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3742 * pretty-print.h (pp_bar_bar): New.
3743 (pp_ampersand_ampersand): Likewise.
3744 (pp_less_equal): Likewise.
3745 (pp_greater_equal): Likewise.
3746 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
3747 printer functions instead of pp_string or operators and punctuators.
3748 (dump_gimple_call): Likewise.
3749 (dump_gimple_omp_for): Likewise.
3750 (dump_gimple_transaction): Likewise.
3751 (dump_gimple_phi): Likewise.
3752 (pp_gimple_stmt_1): Likewise.
3753 * sched-vis.c (print_insn): Likewise.
3754 * tree-mudflap.c (mf_varname_tree): Likewise.
3755 * tree-pretty-print.c (dump_block_node): Likewise.
3756 (dump_generic_node): Likewise.
3758 2013-08-02 Jan Hubicka <jh@suse.cz>
3760 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
3762 * lto-streamer-out.c (tree_is_indexable): Results decls and
3763 parm decls are not indexable.
3764 (DFS_write_tree_body): Do not follow args and results.
3765 (hash_tree): Likewise.
3766 (output_functions): Rearrange so struct function is needed
3767 only when real body is output; be able to also ouptut abstract
3768 functions; output DECL_ARGUMENTS and DECL_RESULT.
3769 (lto_output): When not in WPA, ale store abstract functions.
3770 (write_symbol): Do not care about RESULT_DECL.
3771 (output_symbol_p): Handle correctly sbtract decls.
3772 * lto-streamer-in.c (input_function): Rearrange so struct
3773 function can be NULL at entry; allow streaming of
3774 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
3775 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
3776 sanity check during LTO.
3777 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
3778 RESULT_DECl and DECL_ARGUMENTS.
3779 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
3782 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3784 * pretty-print.h (pp_underscore): New.
3786 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
3787 printer functions instead of pp_character.
3788 (dump_binary_rhs): Likewise.
3789 (dump_ternary_rhs): Likewise.
3790 (dump_gimple_call_args): Likewise.
3791 (pp_points_to_solution): Likewise.
3792 (dump_gimple_call): Likewise.
3793 (dump_gimple_switch): Likewise.
3794 (dump_gimple_cond): Likewise.
3795 (dump_gimple_bind): Likewise.
3796 (dump_gimple_try): Likewise.
3797 (dump_gimple_omp_for): Likewise.
3798 (dump_gimple_omp_continue): Likewise.
3799 (dump_gimple_omp_single): Likewise.
3800 (dump_gimple_omp_sections): Likewise.
3801 (dump_gimple_omp_block): Likewise.
3802 (dump_gimple_omp_critical): Likewise.
3803 (dump_gimple_transaction): Likewise.
3804 (dump_gimple_asm): Likewise.
3805 (dump_gimple_phi): Likewise.
3806 (dump_gimple_omp_parallel): Likewise.
3807 (dump_gimple_omp_task): Likewise.
3808 (dump_gimple_omp_atomic_load): Likewise.
3809 (dump_gimple_omp_atomic_store): Likewise.
3810 (dump_gimple_mem_ops): Likewise.
3811 (pp_gimple_stmt_1): Likewise.
3812 (pp_cfg_jump): Likewise.
3813 (dump_implicit_edges): Likewise.
3814 (gimple_dump_bb_for_graph): Likewise.
3815 * graph.c (draw_cfg_node): Likewise.
3816 * langhooks.c (lhd_print_error_function): Likewise.
3817 * sched-vis.c (print_exp): Likewise.
3818 (print_value): Likewise.
3819 (print_pattern): Likewise.
3820 (print_insn): Likewise.
3821 (rtl_dump_bb_for_graph): Likewise.
3822 * tree-pretty-print.c (dump_function_declaration): Likewise.
3823 (dump_array_domain): Likewise.
3824 (dump_omp_clause): Likewise.
3825 (dump_location): Likewise.
3826 (dump_generic_node): Likewise.
3827 (print_struct_decl): Likewise.
3828 * diagnostic.c (diagnostic_show_locus): Use pp_space.
3830 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3832 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
3833 candidate table when replacing a candidate statement.
3834 (replace_rhs_if_not_dup): Likewise.
3835 (replace_one_candidate): Likewise.
3837 2013-08-02 Jan Hubicka <jh@suse.cz>
3838 Martin Liska <marxin.liska@gmail.com>
3840 * cgraphunit.c (add_new_function): Fix logic when adding from
3842 (assemble_thunk): Rename to ...
3843 (expand_thunk); .. this one; export; get it working with
3844 general functions; make produced gimple valid.
3845 * cgraph.h (expand_thunk): Declare.
3847 2013-08-02 Jan Hubicka <jh@suse.cz>
3849 * ipa-cp.c (gather_context_independent_values): Use
3850 ipa_get_param_move_cost.
3851 (get_replacement_map): Remove PARAM; move parameter folding
3853 (create_specialized_node): Update.
3854 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
3855 assert that we have gimple body; update move_cost.
3856 (count_formal_params): Assert that we have gimple body.
3857 (ipa_dump_param): New function.
3858 (ipa_alloc_node_params): Break out from ...
3859 (ipa_initialize_node_params): ... here.
3860 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
3861 (ipa_write_node_info): Stream move costs.
3862 (ipa_read_node_info): Read move costs.
3863 (ipa_update_after_lto_read): Do not recompute node params.
3864 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
3865 (ipa_get_param): Check we are not in WPA.
3866 (ipa_get_param_move_cost): New.
3867 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
3868 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
3869 parm numbers to be present.
3871 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3873 PR rtl-optimization/58048
3874 * lra-constraints.c (process_alt_operands): Don't check asm
3875 operand on register.
3877 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
3879 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
3880 the implied StoreLoad barrier for atomic operations if before.
3882 2013-08-02 Jan Hubicka <jh@suse.cz>
3883 Martin Liska <marxin.liska@gmail.com>
3885 * cgraph.c (cgraph_function_body_availability): Do not check
3887 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
3888 symtab_node_availability): Declare.
3889 * ipa.c (can_replace_by_local_alias): New.
3890 (function_and_variable_visibility): Use it.
3891 * symtab.c (symtab_for_node_and_aliases,
3892 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
3894 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3896 PR rtl-optimization/57963
3897 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
3898 (lra_constraints): Use them.
3900 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
3902 * config/arm/types.md (define_attr "type"): Add "load_acq"
3904 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
3906 (cortex_a53_store1): Likewise.
3908 2013-08-01 Jan Hubicka <jh@suse.cz>
3910 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
3911 partitions are not needed.
3913 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
3915 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
3916 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
3917 MAYBE_NON_Q_CLASS_P where appropriate.
3919 2013-08-01 Jan Hubicka <jh@suse.cz>
3921 * cgraph.h (release_function_body): Declare.
3922 * tree.c (free_lang_data_in_decl): Free, parameters and return values
3923 of unused delcarations.
3925 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3927 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3928 RTL form when subtracting a constant.
3930 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3932 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3933 Generate canonical plus rtx with negated immediate instead of minus
3935 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3937 2013-08-01 Jan Hubicka <jh@suse.cz>
3939 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3940 (cgraph_release_function_body): Likewise.
3941 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3942 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3943 to used_as_abstract_origin.
3944 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3945 symbols used as abstract origins.
3946 * cgraphunit.c (analyze_functions): Update.
3947 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3948 used_as_abstract_origin.
3949 * tree-inline.c (tree_function_versioning): Update
3950 used_as_abstract_origin; be ready for DECL_RESULT and
3951 DECL_ARGUMENTS to be NULL.
3953 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3954 for abstract functions.
3955 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3958 2013-08-01 Jan Hubicka <jh@suse.cz>
3960 * profile.c (compute_value_histograms): Fix thinko.
3962 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3964 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3965 aarch-common-protos.h to extra_headers.
3966 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3967 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3968 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3970 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3972 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3973 Include "../arm/types.md". Define "type" attribute for all patterns.
3974 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3977 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3979 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3980 to support power8 load fusion.
3981 (fusion_gpr_mem_load): Likewise.
3983 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3985 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3986 declarations for power8 load fusion.
3987 (emit_fusion_gpr_load): Likewise.
3989 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3990 tuning for power8, turn on fusion mode by default. Turn on sign
3991 extending fusion mode if normal fusion mode is on, and we are at
3993 (fusion_gpr_load_p): New function, return true if we can fuse an
3994 addis instruction with a dependent load to a GPR.
3995 (emit_fusion_gpr_load): Emit the instructions for power8 load
3998 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3999 (VSX load fusion peepholes): New peepholes to fuse together an
4000 addi instruction with a VSX load instruction.
4002 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
4003 peepholes to fuse an addis instruction with a load to a GPR base
4004 register. If we are supporting sign extending fusions, convert
4005 sign extending loads to zero extending loads and add an explicit
4008 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4010 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
4011 aarch-common-protos.h to extra_headers.
4012 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
4013 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
4014 (arm_early_store_addr_dep): Likewise.
4015 (arm_no_early_alu_shift_dep): Likewise.
4016 (arm_no_early_alu_shift_value_dep): Likewise.
4017 (arm_no_early_mul_dep): Likewise.
4018 (arm_no_early_store_addr_dep): Likewise.
4019 (arm_mac_accumulator_is_mul_result): Likewise.
4020 (arm_mac_accumulator_is_result): Likewise.
4021 * config/arm/aarch-common.c: ... here. New file.
4022 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
4024 (arm_early_store_addr_dep): Likewise.
4025 (arm_no_early_alu_shift_dep): Likewise.
4026 (arm_no_early_alu_shift_value_dep): Likewise.
4027 (arm_no_early_mul_dep): Likewise.
4028 (arm_no_early_store_addr_dep): Likewise.
4029 (arm_mac_accumulator_is_mul_result): Likewise.
4030 (arm_mac_accumulator_is_result): Likewise.
4031 * config/arm/aarch-common-protos.h: ... here. New file.
4032 * config/arm/t-arm (aarch-common.o): Define.
4034 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4036 * config/arm/arm.md: Include new file "types.md".
4037 (define_attr "type"): Move from here to ...
4038 (define_attr "mul32"): Likewise.
4039 (define_attr "mul64"): Likewise.
4040 * config/arm/types.md: ... here. New file.
4042 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
4044 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
4045 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
4047 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4049 * gen-pass-instances.awk: Fix offset of substr().
4051 2013-07-31 David Malcolm <dmalcolm@redhat.com>
4053 * Makefile.in (pass-instances.def): New.
4054 (passes.o): Replace dependency on passes.def with one on
4057 * gen-pass-instances.awk: New.
4059 * passes.c (pass_manager::pass_manager): Use pass-instances.def
4060 rather than passes.def, updating local definition of NEXT_PASS
4061 macro to add an extra NUM parameter (currently unused).
4063 2013-07-30 David Malcolm <dmalcolm@redhat.com>
4065 * Makefile.in (PASS_MANAGER_H): New.
4066 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
4067 (passes.o): Likewise.
4068 (statistics.o): Likewise.
4069 (cgraphunit.o): Likewise.
4070 (context.o): Depend on PASS_MANAGER_H.
4072 * pass_manager.h: New.
4074 * cgraphunit.c (cgraph_add_new_function): Update for moves
4075 of globals to fields of pass_manager.
4076 (analyze_function): Likewise.
4077 (expand_function): Likewise.
4078 (ipa_passes): Likewise.
4079 (compile): Likewise.
4081 * context.c (context::context): New.
4082 * context.h (context::context): New.
4083 (context::get_passes): New.
4084 (context::passes_): New.
4086 * lto-cgraph.c (input_node): Update for moves of globals to
4087 fields of pass_manager.
4089 * passes.c (all_passes): Remove, in favor of a field of the
4090 same name within the new class pass_manager.
4091 (all_small_ipa_passes): Likewise.
4092 (all_lowering_passes): Likewise.
4093 (all_regular_ipa_passes): Likewise.
4094 (all_late_ipa_passes): Likewise.
4095 (all_lto_gen_passes): Likewise.
4096 (passes_by_id): Likewise.
4097 (passes_by_id_size): Likewise.
4098 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4099 the new class pass_manager.
4100 (set_pass_for_id): Convert to...
4101 (pass_manager::set_pass_for_id): ...method.
4102 (get_pass_for_id): Convert to...
4103 (pass_manager::get_pass_for_id): ...method.
4104 (register_one_dump_file): Move body of implementation into...
4105 (pass_manager::register_one_dump_file): ...here.
4106 (register_dump_files_1): Convert to...
4107 (pass_manager::register_dump_files_1): ...method.
4108 (register_dump_files): Convert to...
4109 (pass_manager::register_dump_files): ...method.
4110 (create_pass_tab): Update for moves of globals to fields of
4112 (dump_passes): Move body of implementation into...
4113 (pass_manager::dump_passes): ...here.
4114 (register_pass): Move body of implementation into...
4115 (pass_manager::register_pass): ...here.
4116 (init_optimization_passes): Convert into...
4117 (pass_manager::pass_manager): ...constructor for new
4118 pass_manager class, and initialize the pass_lists array.
4119 (check_profile_consistency): Update for moves of globals to
4120 fields of pass_manager.
4121 (dump_profile_report): Move body of implementation into...
4122 (pass_manager::dump_profile_report): ...here.
4123 (ipa_write_summaries_1): Update for moves of pass lists from
4124 being globals to fields of pass_manager.
4125 (ipa_write_optimization_summaries): Likewise.
4126 (ipa_read_summaries): Likewise.
4127 (ipa_read_optimization_summaries): Likewise.
4128 (execute_all_ipa_stmt_fixups): Likewise.
4130 * statistics.c (statistics_fini): Update for moves of globals to
4131 fields of pass_manager.
4133 * toplev.c (general_init): Replace call to
4134 init_optimization_passes with construction of the pass_manager
4137 * tree-pass.h (all_passes): Remove, in favor of a field of the
4138 same name within the new class pass_manager.
4139 (all_small_ipa_passes): Likewise.
4140 (all_lowering_passes): Likewise.
4141 (all_regular_ipa_passes): Likewise.
4142 (all_lto_gen_passes): Likewise.
4143 (all_late_ipa_passes): Likewise.
4144 (passes_by_id): Likewise.
4145 (passes_by_id_size): Likewise.
4146 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4147 the new class pass_manager.
4148 (get_pass_for_id): Remove.
4150 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4152 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
4155 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4157 * arm.md (mulhi3): New expand pattern.
4159 2013-07-30 Jan Hubicka <jh@suse.cz>
4160 Martin Liska <marxin.liska@gmail.com>
4162 * profile.c (compute_value_histograms): Do not ICE when
4163 there is mismatch only on some counters.
4165 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4167 PR rtl-optimization/57637
4168 * function.c (move_insn_for_shrink_wrap): Also check the
4169 GEN set of the LIVE problem for the liveness analysis
4170 if it exists, otherwise give up.
4172 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4174 PR tree-optimization/57993
4175 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
4176 replaced statement in the candidate table.
4177 (phi_add_costs): Return infinite cost when the hidden basis does
4178 not dominate all phis on which the candidate is dependent.
4179 (replace_one_candidate): Record replaced statement in the
4182 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
4184 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
4185 (ashlv2si3): New expander.
4186 (*ashlv2si3_i): New define_insn_and_split.
4187 * predicates.md (float_operation): Allow patterns with three
4190 PR rtl-optimization/58021
4191 * mode-switching.c (create_pre_exit): Always split off preceding
4192 insns if we are not at the basic block head.
4194 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
4196 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
4197 (UCLIBC_DYNAMIC_LINKER): New macro.
4198 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
4200 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
4201 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
4203 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
4204 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
4205 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
4206 for SF and DF modes. Use ieee_quad_format for TF mode.
4207 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
4208 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
4209 (mips_option_override): Handle `-mnan=legacy'.
4210 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
4211 `-mabs=2008' and `-mnan=2008'.
4212 (OPTION_DEFAULT_SPECS): Add "nan" default.
4213 (ASM_SPEC): Handle `-mnan='.
4214 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
4215 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
4216 comment accordingly.
4217 (neg<mode>2): Likewise.
4218 * config/mips/mips.opt (mabs, mnan): New options.
4219 * doc/install.texi (Configuration): Document `--with-nan=' option.
4220 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
4222 (MIPS Options): Document them.
4223 * config.gcc <mips*-*-*>: Handle `--with-nan='.
4224 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
4225 * configure: Regenerate.
4226 * config.in: Regenerate.
4228 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4230 * config/i386/i386.md (float post-reload splitters): Do not check
4231 for subregs of SSE registers.
4233 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4234 H.J. Lu <hongjiu.lu@intel.com>
4238 * config/i386/i386.md (post-reload splitter
4239 to avoid partial SSE reg dependency stalls): New pattern.
4241 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
4243 * config/s390/s390.md ("movcc"): Swap load and store instructions.
4245 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
4247 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
4248 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
4250 2013-07-26 Cary Coutant <ccoutant@google.com>
4252 * dwarf2out.c (die_checksum_ordered): Don't include template
4253 instantiations in signature.
4254 (is_template_parameter): New function.
4255 (is_template_instantiation): New function.
4256 (generate_skeleton_bottom_up): Don't include template instantiations
4258 (generate_skeleton): Likewise.
4259 (break_out_comdat_types): Move recursive call to break out nested
4261 (prune_unused_types_mark_generic_parms_dies): Call
4262 is_template_parameter.
4264 2013-07-26 Ian Bolton <ian.bolton@arm.com>
4266 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
4267 uses vector registers.
4268 * config/aarch64/iterators.md: Add attributes rtn and vas.
4270 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4271 Richard Earnshaw <richard.earnshaw@arm.com>
4273 * combine.c (simplify_comparison): Re-canonicalize operands
4275 * config/arm/arm.md (movcond_addsi): New splitter.
4277 2013-07-25 Sterling Augustine <saugustine@google.com>
4279 * dwarf2out.c (size_of_pubnames): Move code to...
4280 (include_pubname_in_output): ...here. New.
4281 (want_pubnames): Rearrange.
4282 (output_pubnames): Call include_pubname_in_output. Move assertion.
4284 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4286 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
4288 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4291 * doc/extend.texi: Remove obsolete builtins. Fix
4292 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
4294 2013-07-25 Jan Hubicka <jh@suse.cz>
4296 * cgraph.c (release_function_body): Break out from ...
4297 (cgraph_release_function_body): ... this one; also release DECL_RESULT
4299 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
4300 old_tree in the map.
4301 (create_specialized_node): Update.
4302 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
4304 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
4305 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
4306 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
4307 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
4310 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4312 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
4313 addsi3_carryin_alt2_<optab>): Correct output template.
4315 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4317 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
4318 Adjust for arm_restrict_it.
4319 Remove trailing whitespace.
4321 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
4323 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
4324 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
4326 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
4328 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
4330 PR rtl-optimization/57960
4331 * lra-constraints.c (process_alt_operands): Use the right mode
4332 when checking strict_low.
4334 2013-07-25 Jan Hubicka <jh@suse.cz>
4336 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
4337 * cgraph.c (cgraph_remove_node): Do not release function body
4338 when in cgraph streaming.
4339 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
4340 in other partitions are not considered reachable; fix handling of
4343 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4345 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
4347 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4352 * config/arm/arm.md ("*sibcall_insn): Replace use of
4353 Ss with US. Adjust output for v5 and v4t.
4354 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
4356 * config/arm/constraints.md ("Ss"): Rename to US.
4358 2013-07-25 Terry Guo <terry.guo@arm.com>
4360 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
4361 shift_add/shift_sub0/shift_sub1 RTXs.
4363 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
4364 Anton Blanchard <anton@au1.ibm.com>
4366 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
4367 (altivec_vpks<VI_char>ss): Likewise.
4368 (altivec_vpks<VI_char>us): Likewise.
4369 (altivec_vpku<VI_char>us): Likewise.
4370 (altivec_vpku<VI_char>um): Likewise.
4372 2013-07-24 David Malcolm <dmalcolm@redhat.com>
4374 Introduce context class.
4376 * Makefile.in (CONTEXT_H): New.
4377 (OBJS): Add context.o.
4378 (toplev.o): Add CONTEXT_H to dependencies.
4381 * toplev.c (general_init): Create the singleton gcc::context instance.
4387 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
4389 PR rtl-optimization/57968
4390 * mode-switching.c (create_pre_exit): Allow instructions that
4391 don't set a return register to need a non-exit mode.
4393 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4394 Anton Blanchard <anton@au1.ibm.com>
4396 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
4397 operands to vperm for little endian.
4398 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
4399 of lvsl to create the control mask for a vperm for little endian.
4401 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4402 Anton Blanchard <anton@au1.ibm.com>
4404 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
4405 two operands for little-endian.
4407 2013-07-23 Steve Ellcey <sellcey@mips.com>
4409 * config/mips/mips.c (mips_case_values_threshold): New.
4410 (TARGET_CASE_VALUES_THRESHOLD): Define.
4412 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4413 Anton Blanchard <anton@au1.ibm.com>
4415 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
4416 selection of field for vector splat in little endian mode.
4418 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4420 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
4421 expanders to rs6000.md.
4422 (ior<mode>3): Likewise.
4423 (and<mode>3): Likewise.
4424 (one_cmpl<mode>2): Likewise.
4425 (nor<mode>3): Likewise.
4426 (andc<mode>3): Likewise.
4427 (eqv<mode>3): Likewise.
4428 (nand<mode>3): Likewise.
4429 (orc<mode>3): Likewise.
4431 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
4434 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
4435 to split multi-word logical operations.
4436 (rs6000_split_logical_di): Likewise.
4437 (rs6000_split_logical): Likewise.
4439 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
4440 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
4441 and allow TImode operations in 32-bit.
4442 (vsx_and<mode>3_64bit): Likewise.
4443 (vsx_ior<mode>3_32bit): Likewise.
4444 (vsx_ior<mode>3_64bit): Likewise.
4445 (vsx_xor<mode>3_32bit): Likewise.
4446 (vsx_xor<mode>3_64bit): Likewise.
4447 (vsx_one_cmpl<mode>2_32bit): Likewise.
4448 (vsx_one_cmpl<mode>2_64bit): Likewise.
4449 (vsx_nor<mode>3_32bit): Likewise.
4450 (vsx_nor<mode>3_64bit): Likewise.
4451 (vsx_andc<mode>3_32bit): Likewise.
4452 (vsx_andc<mode>3_64bit): Likewise.
4453 (vsx_eqv<mode>3_32bit): Likewise.
4454 (vsx_eqv<mode>3_64bit): Likewise.
4455 (vsx_nand<mode>3_32bit): Likewise.
4456 (vsx_nand<mode>3_64bit): Likewise.
4457 (vsx_orc<mode>3_32bit): Likewise.
4458 (vsx_orc<mode>3_64bit): Likewise.
4460 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
4461 logical types in GPRs.
4463 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
4464 logical insns to rs6000.md, and allow TImode operations in
4466 (altivec_ior<mode>3): Likewise.
4467 (altivec_xor<mode>3): Likewise.
4468 (altivec_one_cmpl<mode>2): Likewise.
4469 (altivec_nor<mode>3): Likewise.
4470 (altivec_andc<mode>3): Likewise.
4472 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
4473 attributes for moving the 128-bit logical operations into
4475 (BOOL_REGS_OUTPUT): Likewise.
4476 (BOOL_REGS_OP1): Likewise.
4477 (BOOL_REGS_OP2): Likewise.
4478 (BOOL_REGS_UNARY): Likewise.
4479 (BOOL_REGS_AND_CR0): Likewise.
4480 (one_cmpl<mode>2): Add support for DI logical operations on
4481 32-bit, splitting the operations to 32-bit.
4485 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
4486 changes to combine the 32/64-bit code, allow logical operations on
4487 TI mode in 32-bit, and to use similar match_operator patterns like
4488 scalar mode uses. Combine the Altivec and VSX code for logical
4489 operations, and move it here.
4490 (ior<mode>3, 128-bit types): Likewise.
4491 (xor<mode>3, 128-bit types): Likewise.
4492 (one_cmpl<mode>3, 128-bit types): Likewise.
4493 (nor<mode>3, 128-bit types): Likewise.
4494 (andc<mode>3, 128-bit types): Likewise.
4495 (eqv<mode>3, 128-bit types): Likewise.
4496 (nand<mode>3, 128-bit types): Likewise.
4497 (orc<mode>3, 128-bit types): Likewise.
4498 (and<mode>3_internal): Likewise.
4499 (bool<mode>3_internal): Likewise.
4500 (boolc<mode>3_internal1): Likewise.
4501 (boolc<mode>3_internal2): Likewise.
4502 (boolcc<mode>3_internal1): Likewise.
4503 (boolcc<mode>3_internal2): Likewise.
4504 (eqv<mode>3_internal1): Likewise.
4505 (eqv<mode>3_internal2): Likewise.
4506 (one_cmpl1<mode>3_internal): Likewise.
4508 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
4510 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4511 Rename flag_stack_usage to flag_stack_usage_info.
4513 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
4515 * config/microblaze/sync.md: New file.
4516 * config/microblaze/microblaze.md: Include sync.md
4517 * config/microblaze/microblaze.c: Add print_operand 'y'.
4518 * config/microblaze/constraints.md: Add memory_contraint
4519 'Q' which is a single register.
4521 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
4523 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
4525 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
4527 * reload.c (find_reloads): Exit loop once we find this operand
4528 cannot be reloaded somehow for this alternative.
4530 * reload.c (find_reloads): Exit loop once we find a hard register.
4532 * rtlanal.c (computed_jump_p): Exit loop once we find label
4535 * i386.c (ix86_pad_returns): Exit loop after setting replace.
4537 * cfgloopmanip.c (remove_path): Exit loop after setting
4540 * gensupport.c (subst_dup): Avoid loop if code is not
4541 MATCH_DUP nor MATCH_OP_DUP.
4543 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
4545 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
4547 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4549 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
4550 true for SP_REGNUM if mode == ptr_mode.
4551 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
4552 with value R0_REGNUM + 31.
4554 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4556 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
4557 pad pointer-typed argument downward.
4559 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4561 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
4562 and __ILP32__ when the ILP32 model is in use.
4564 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4566 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
4567 (aarch64_load_symref_appropriately): In the case of
4568 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
4569 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
4570 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
4571 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
4572 if 'mode' doesn't equal to 'ptr_mode'.
4573 (aarch64_output_mi_thunk): Add an assertion on the alignment of
4574 'vcall_offset'; change to call aarch64_emit_move differently depending
4575 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
4576 to calculate the upper bound of 'vcall_offset'.
4577 (aarch64_cannot_force_const_mem): Change to also return true if
4579 (aarch64_legitimize_reload_address): In the case of large
4580 displacements, add new local variable 'xmode' and an assertion
4581 based on it; change to use 'xmode' to generate the new rtx and
4583 (aarch64_asm_trampoline_template): Change to generate the template
4584 differently depending on TARGET_ILP32 or not; change to use
4585 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
4586 (aarch64_trampoline_size): Removed.
4587 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
4588 and replace immediate literals with it. Change to use 'ptr_mode'
4589 instead of 'DImode' and call convert_memory_address if the mode
4590 of 'fnaddr' doesn't equal to 'ptr_mode'.
4591 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
4593 (aarch64_elf_asm_destructor): Likewise.
4594 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
4595 on TARGET_ILP32 instead of aarch64_trampoline_size.
4596 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
4597 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
4598 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
4599 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
4600 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
4601 (storewb_pair<GPI:mode>_<P:mode>): ... this.
4602 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
4603 depending on the value of 'mode'.
4604 (add_losym_<mode>): New.
4605 (ldr_got_small_<mode>): New, based on ldr_got_small.
4606 (ldr_got_small): Remove.
4607 (ldr_got_small_sidi): New.
4608 * config/aarch64/iterators.md (P): New.
4609 (PTR): Change to 'ptr_mode' in the condition.
4611 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4613 * config.gcc (aarch64*-*-*): Support --with-abi.
4614 (aarch64*-*-elf): Support --with-multilib-list.
4615 (aarch64*-*-linux*): Likewise.
4616 (supported_defaults): Add abi to aarch64*-*-*.
4617 * configure.ac: Mention AArch64 for --with-multilib-list.
4618 * configure: Re-generated.
4619 * config/aarch64/biarchilp32.h: New file.
4620 * config/aarch64/biarchlp64.h: New file.
4621 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
4623 (MULTILIB_DEFAULTS): Ditto.
4624 (DRIVER_SELF_SPECS): Ditto.
4625 (ASM_SPEC): Update to also substitute -mabi.
4626 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
4627 file whose name depends on -mabi= and -mbig-endian.
4628 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
4630 (POINTER_SIZE): New define.
4631 (POINTERS_EXTEND_UNSIGNED): Ditto.
4632 (enum aarch64_abi_type): New enumeration tag.
4633 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
4634 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
4635 (TARGET_ILP32): New define.
4636 * config/aarch64/aarch64.opt (mabi): New.
4638 (ilp32, lp64): New values for -mabi.
4639 * config/aarch64/t-aarch64 (comma): New define.
4640 (MULTILIB_OPTIONS): Ditto.
4641 (MULTILIB_DIRNAMES): Ditto.
4642 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
4643 * doc/invoke.texi: Document -mabi for AArch64.
4645 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
4647 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
4649 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4650 Anton Blanchard <anton@au1.ibm.com>
4652 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
4653 endianness when selecting field to splat.
4655 2013-07-22 Eric Christopher <echristo@gmail.com>
4657 * dwarf2out.c (die_odr_checksum): New function to use
4658 CHECKSUM_ macros and ULEB128 for DIE tag.
4659 (generate_type_signature): Use.
4661 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
4663 * config.gcc (sparc*-*-*): Accept leon3 processor.
4664 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
4665 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
4666 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
4667 * config/sparc/sparc.opt (enum processor_type): Add leon3.
4668 (mfix-ut699): Adjust comment.
4669 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
4670 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
4671 (CPP_CPU_SPEC): Likewise.
4672 (ASM_CPU_SPEC): Likewise.
4673 * config/sparc/sparc.c (leon3_cost): New constant.
4674 (sparc_option_override): Add leon3 support.
4675 (mem_ref): New function.
4676 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
4677 (sparc_do_work_around_errata): Look into the instruction in the delay
4678 slot and adjust accordingly. Add fix for the data cache nullify issues
4679 of the UT699. Change insertion position for the NOP.
4680 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
4681 (leon3_load): New reservation.
4682 (leon_store): Bump latency to 2.
4683 (grfpu): New automaton.
4684 (grfpu_alu): New unit.
4685 (grfpu_ds): Likewise.
4686 (leon_fp_alu): Adjust.
4687 (leon_fp_mult): Delete.
4688 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
4689 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
4690 * config/sparc/sparc.md (cpu): Add leon3.
4691 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
4693 (atomic_test_and_set): Likewise.
4696 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
4698 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
4699 default for R5900 targets.
4700 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
4701 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
4702 * config/mips/mips.c (mips_option_override): Report an error for
4703 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
4704 for -march=r5900 -mhard-float.
4706 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
4708 * df-problems.c (can_move_insns_across): Exit loop once we
4709 find a non-fixed, non-global register.
4711 * ipa-pure-const.c (propagate_nothrow): Exit loop after
4714 * omega.c (omega_eliminate_red): Break after setting red_found.
4715 (omega_problem_has_red_equations): Similarly after setting found.
4716 (omega_query_variable): Similarly after setting coupled.
4718 2013-07-22 Marek Polacek <polacek@redhat.com>
4720 * gimplify.c: Don't include gimple.h twice.
4722 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4724 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
4725 instead of TARGET_THUMB1.
4726 (Pz): New constraint.
4727 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
4729 (compare_negsi_si): Likewise.
4730 (compare_addsi2_op0): Likewise.
4731 (compare_addsi2_op1): Likewise.
4732 (addsi3_carryin_<optab>): Likewise.
4733 (addsi3_carryin_alt2_<optab>): Likewise.
4734 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
4735 for arm_restrict_it.
4736 (subsi3_carryin): Likewise.
4737 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
4738 (minmax_arithsi): Disable for arm_restrict_it.
4739 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
4740 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
4741 (satsi_<SAT:code>_shift): Likewise.
4742 (arm_shiftsi3): Add alternative for 16-bit encoding.
4743 (arm32_movhf): Disable for arm_restrict_it.
4744 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
4745 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
4747 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
4749 * config/arm/arm.md (attribute "insn"): Delete.
4750 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
4751 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
4752 (not_shiftsi): Update for attribute change.
4753 (not_shiftsi_compare0): Likewise.
4754 (not_shiftsi_compare0_scratch): Likewise.
4755 (arm_one_cmplsi2): Likewise.
4756 (thumb1_one_cmplsi2): Likewise.
4757 (notsi_compare0): Likewise.
4758 (notsi_compare0_scratch): Likewise.
4759 (thumb1_movdi_insn): Likewise.
4760 (arm_movsi_insn): Likewise.
4761 (movhi_insn_arch4): Likewise.
4762 (movhi_bytes): Likewise.
4763 (arm_movqi_insn): Likewise.
4764 (thumb1_movqi_insn): Likewise.
4765 (arm32_movhf): Likewise.
4766 (thumb1_movhf): Likewise.
4767 (arm_movsf_soft_insn): Likewise.
4768 (thumb1_movsf_insn): Likewise.
4769 (thumb_movdf_insn): Likewise.
4770 (movsicc_insn): Likewise.
4771 (movsfcc_soft_insn): Likewise.
4772 (and_scc): Likewise.
4773 (cond_move): Likewise.
4774 (if_move_not): Likewise.
4775 (if_not_move): Likewise.
4776 (if_shift_move): Likewise.
4777 (if_move_shift): Likewise.
4778 (if_shift_shift): Likewise.
4779 (if_not_arith): Likewise.
4780 (if_arith_not): Likewise.
4781 (cond_move_not): Likewise.
4782 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
4783 (neon_mov<mode>): Likewise.
4784 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
4785 (thumb2_movsi_vfp): Likewise.
4786 (movsf_vfp): Likewise.
4787 (thumb2_movsf_vfp): Likewise.
4788 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4790 (cortexa7_older_only): Likewise.
4791 (cortexa7_younger): Likewise.
4792 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
4793 (1020alu_shift_op): Likewise.
4794 (1020alu_shift_reg_op): Likewise.
4795 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
4796 (alu_shift_op): Likewise.
4797 (alu_shift_reg_op): Likewise.
4798 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
4799 (11_alu_shift_op): Likewise.
4800 (11_alu_shift_reg_op): Likewise.
4801 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
4802 (9_alu_shift_reg_op): Likewise.
4803 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4805 (cortex_a15_alu_shift): Likewise.
4806 (cortex_a15_alu_shift_reg): Likewise.
4807 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
4808 (cortex_a5_alu_shift): Likewise.
4809 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
4811 (cortex_a53_alu_shift): Likewise.
4812 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4814 (cortex_a7_alu_reg): Likewise.
4815 (cortex_a7_alu_shift): Likewise.
4816 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4817 (cortex_a8_alu_shift): Likewise.
4818 (cortex_a8_alu_shift_reg): Likewise.
4819 (cortex_a8_mov): Likewise.
4820 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4821 (cortex_a9_dp_shift): Likewise.
4822 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
4823 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
4824 (cortex_r4_mov): Likewise.
4825 (cortex_r4_alu_shift): Likewise.
4826 (cortex_r4_alu_shift_reg): Likewise.
4827 * config/arm/fa526.md (526_alu_op): Update for attribute change.
4828 (526_alu_shift_op): Likewise.
4829 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
4830 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
4831 (626te_alu_shift_op): Likewise.
4832 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
4833 (726te_alu_op): Likewise.
4834 (726te_alu_shift_op): Likewise.
4835 (726te_alu_shift_reg_op): Likewise.
4836 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
4837 (mp626_alu_shift_op): Likewise.
4838 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
4839 (pj4_alu_e1_conds): Likewise.
4840 (pj4_alu): Likewise.
4841 (pj4_alu_conds): Likewise.
4842 (pj4_shift): Likewise.
4843 (pj4_shift_conds): Likewise.
4844 (pj4_alu_shift): Likewise.
4845 (pj4_alu_shift_conds): Likewise.
4847 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4849 * config/arm/predicates.md (shiftable_operator_strict_it):
4851 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
4852 Disable cond_exec version for arm_restrict_it.
4853 (thumb2_smaxsi3): Convert to generate cond_exec.
4854 (thumb2_sminsi3): Likewise.
4855 (thumb32_umaxsi3): Likewise.
4856 (thumb2_uminsi3): Likewise.
4857 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
4858 (thumb2_neg_abssi2): Likewise.
4859 (thumb2_mov_scc): Add alternative for 16-bit encoding.
4860 (thumb2_movsicc_insn): Adjust alternatives.
4861 (thumb2_mov_negscc): Disable for arm_restrict_it.
4862 (thumb2_mov_negscc_strict_it): New pattern.
4863 (thumb2_mov_notscc_strict_it): New pattern.
4864 (thumb2_mov_notscc): Disable for arm_restrict_it.
4865 (thumb2_ior_scc): Likewise.
4866 (thumb2_ior_scc_strict_it): New pattern.
4867 (thumb2_cond_move): Adjust for arm_restrict_it.
4868 (thumb2_cond_arith): Disable for arm_restrict_it.
4869 (thumb2_cond_arith_strict_it): New pattern.
4870 (thumb2_cond_sub): Adjust for arm_restrict_it.
4871 (thumb2_movcond): Likewise.
4872 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
4873 (thumb2_zero_extendhisi2_v6): Likewise.
4874 (thumb2_zero_extendqisi2_v6): Likewise.
4875 (orsi_notsi_si): Likewise.
4876 (orsi_not_shiftsi_si): Likewise.
4878 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
4880 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
4881 instruction sequence is 1 byte shorter.
4883 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
4885 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
4886 it is not needed after split.
4888 2013-07-20 Iain Sandoe <iain@codesourcery.com>
4891 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
4892 second label for nonlocal goto receivers. Don't output pic base labels
4893 unless we're producing PIC; mark that action unreachable().
4894 (ix86_save_reg): If the function contains a nonlocal label, save the
4896 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
4897 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
4898 (update_pic_label_number_if_needed): New.
4899 (machopic_output_function_base_name): Adjust for nonlocal receiver
4901 (machopic_should_output_picbase_label): New.
4902 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
4903 (nonlocal_goto_receiver): New insn and split.
4905 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
4907 * config/aarch64/aarch64-builtins.c
4908 (aarch64_fold_builtin): Fold abs in all modes.
4909 * config/aarch64/aarch64-simd-builtins.def
4910 (abs): Enable for all modes.
4911 * config/aarch64/arm_neon.h
4912 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
4913 (vabs_f64): Add missing intrinsic.
4915 2013-07-19 Ian Bolton <ian.bolton@arm.com>
4917 * config/aarch64/arm_neon.h (vabs_s64): New function
4919 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
4922 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
4923 * config/avr/avr.md (adjust_len): Add `round'.
4924 * config/avr/avr-protos.h (avr_out_round): New prototype.
4925 (avr_out_plus): Add `out_label' argument.
4926 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4927 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4928 Handle the case where `insn' is just a pattern.
4929 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4930 (avr_out_round): New function.
4931 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4933 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4935 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4936 Add check for flag_stack_usage to handle -fstack-usage support
4938 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4940 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4941 interaction for new Power8 flags and VSX.
4943 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4946 * tree-inline.c (expand_call_inline): Emit errors during
4947 early_inlining only if optimization is not turned on.
4949 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4953 * passes.c (init_optimization_passes): Move the construction of
4954 the pass hierarchy into a new passes.def file.
4956 * Makefile.in (passes.o): Add dependency on passes.def.
4958 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4960 * passes.c (init_optimization_passes): Introduce macros for
4961 constructing the tree of passes (INSERT_PASSES_AFTER,
4962 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4963 TERMINATE_PASS_LIST).
4965 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4966 Wei Mi <wmi@google.com>
4968 PR rtl-optimization/57878
4969 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4971 (reload_pseudo_compare_func): Check nregs first for reload
4974 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4976 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4978 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4980 * read-rtl.c (validate_const_int): Once an invalid character is
4981 seen, quit the loop.
4983 * gengtype.c (write_roots): Similarly once we find the "deletable"
4984 or "if_marked" option.
4986 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4988 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4989 "xtab" and "sat". Move value "clz" from here to ...
4990 (attriubte "type"): ... here.
4991 (satsi_<SAT:code>): Delete "insn" attribute.
4992 (satsi_<SAT:code>_shift): Likewise.
4993 (arm_zero_extendqisi2addsi): Likewise.
4994 (arm_extendqisi2addsi): Likewise.
4995 (clzsi2): Update for attribute changes.
4996 (rbitsi2): Likewise.
4997 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4999 (arm_usatsihi): Likewise.
5000 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5002 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
5004 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
5005 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
5006 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
5007 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
5008 in alphabetical order.
5009 (attribute "core_cycles"): Update for attribute changes.
5010 (arm_addsi3): Likewise.
5011 (addsi3_compare0): Likewise.
5012 (addsi3_compare0_scratch): Likewise.
5013 (addsi3_compare_op1): Likewise.
5014 (addsi3_compare_op2): Likewise.
5015 (compare_addsi2_op0): Likewise.
5016 (compare_addsi2_op1): Likewise.
5017 (addsi3_carryin_shift_<optab>): Likewise.
5018 (subsi3_carryin_shift): Likewise.
5019 (rsbsi3_carryin_shift): Likewise.
5020 (arm_subsi3_insn): Likewise.
5021 (subsi3_compare0): Likewise.
5022 (subsi3_compare): Likewise.
5023 (arm_andsi3_insn): Likewise.
5024 (thumb1_andsi3_insn): Likewise.
5025 (andsi3_compare0): Likewise.
5026 (andsi3_compare0_scratch): Likewise.
5027 (zeroextractsi_compare0_scratch
5028 (andsi_not_shiftsi_si): Likewise.
5029 (iorsi3_insn): Likewise.
5030 (iorsi3_compare0): Likewise.
5031 (iorsi3_compare0_scratch): Likewise.
5032 (arm_xorsi3): Likewise.
5033 (thumb1_xorsi3_insn): Likewise.
5034 (xorsi3_compare0): Likewise.
5035 (xorsi3_compare0_scratch): Likewise.
5036 (satsi_<SAT:code>_shift): Likewise.
5038 (arm_shiftsi3): Likewise.
5039 (shiftsi3_compare0): Likewise.
5040 (not_shiftsi): Likewise.
5041 (not_shiftsi_compare0): Likewise.
5042 (not_shiftsi_compare0_scratch): Likewise.
5043 (arm_one_cmplsi2): Likewise.
5044 (thumb_one_complsi2): Likewise.
5045 (notsi_compare0): Likewise.
5046 (notsi_compare0_scratch): Likewise.
5047 (thumb1_zero_extendhisi2): Likewise.
5048 (arm_zero_extendhisi2): Likewise.
5049 (arm_zero_extendhisi2_v6): Likewise.
5050 (arm_zero_extendhisi2addsi): Likewise.
5051 (thumb1_zero_extendqisi2): Likewise.
5052 (thumb1_zero_extendqisi2_v6): Likewise.
5053 (arm_zero_extendqisi2): Likewise.
5054 (arm_zero_extendqisi2_v6): Likewise.
5055 (arm_zero_extendqisi2addsi): Likewise.
5056 (thumb1_extendhisi2): Likewise.
5057 (arm_extendhisi2): Likewise.
5058 (arm_extendhisi2_v6): Likewise.
5059 (arm_extendqisi): Likewise.
5060 (arm_extendqisi_v6): Likewise.
5061 (arm_extendqisi2addsi): Likewise.
5062 (thumb1_extendqisi2): Likewise.
5063 (thumb1_movdi_insn): Likewise.
5064 (arm_movsi_insn): Likewise.
5065 (movsi_compare0): Likewise.
5066 (movhi_insn_arch4): Likewise.
5067 (movhi_bytes): Likewise.
5068 (arm_movqi_insn): Likewise.
5069 (thumb1_movqi_insn): Likewise.
5070 (arm32_movhf): Likewise.
5071 (thumb1_movhf): Likewise.
5072 (arm_movsf_soft_insn): Likewise.
5073 (thumb1_movsf_insn): Likewise.
5074 (movdf_soft_insn): Likewise.
5075 (thumb_movdf_insn): Likewise.
5076 (arm_cmpsi_insn): Likewise.
5077 (cmpsi_shiftsi): Likewise.
5078 (cmpsi_shiftsi_swp): Likewise.
5079 (arm_cmpsi_negshiftsi_si): Likewise.
5080 (movsicc_insn): Likewise.
5081 (movsfcc_soft_insn): Likewise.
5082 (arith_shiftsi): Likewise.
5083 (arith_shiftsi_compare0
5084 (arith_shiftsi_compare0_scratch
5085 (sub_shiftsi): Likewise.
5086 (sub_shiftsi_compare0
5087 (sub_shiftsi_compare0_scratch
5088 (and_scc): Likewise.
5089 (cond_move): Likewise.
5090 (if_plus_move): Likewise.
5091 (if_move_plus): Likewise.
5092 (if_move_not): Likewise.
5093 (if_not_move): Likewise.
5094 (if_shift_move): Likewise.
5095 (if_move_shift): Likewise.
5096 (if_shift_shift): Likewise.
5097 (if_not_arith): Likewise.
5098 (if_arith_not): Likewise.
5099 (cond_move_not): Likewise.
5100 (thumb1_ashlsi3): Set type attribute.
5101 (thumb1_ashrsi3): Likewise.
5102 (thumb1_lshrsi3): Likewise.
5103 (thumb1_rotrsi3): Likewise.
5104 (shiftsi3_compare0_scratch): Likewise.
5105 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
5106 (neon_mov<mode>): Likewise.
5107 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
5109 (thumb2_movsi_insn): Likewise.
5110 (thumb2_cmpsi_neg_shiftsi): Likewise.
5111 (thumb2_extendqisi_v6): Likewise.
5112 (thumb2_zero_extendhisi2_v6): Likewise.
5113 (thumb2_zero_extendqisi2_v6): Likewise.
5114 (thumb2_shiftsi3_short): Likewise.
5115 (thumb2_addsi3_compare0_scratch): Likewise.
5116 (orsi_not_shiftsi_si): Likewise.
5117 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5118 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
5120 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
5121 (1020alu_shift_op): Likewise.
5122 (1020alu_shift_reg_op): Likewise.
5123 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
5124 (alu_shift_op): Likewise.
5125 (alu_shift_reg_op): Likewise.
5126 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
5127 (11_alu_shift_op): Likewise.
5128 (11_alu_shift_reg_op): Likewise.
5129 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
5130 (9_alu_shift_reg_op): Likewise.
5131 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5133 (cortex_a15_alu_shift): Likewise.
5134 (cortex_a15_alu_shift_reg): Likewise.
5135 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
5137 (cortex_a5_alu_shift): Likewise.
5138 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
5140 (cortex_a53_alu_shift): Likewise.
5141 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5143 (cortex_a7_alu_reg): Likewise.
5144 (cortex_a7_alu_shift): Likewise.
5145 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
5147 (cortex_a8_alu_shift): Likewise.
5148 (cortex_a8_alu_shift_reg): Likewise.
5149 (cortex_a8_mov): Likewise.
5150 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
5151 (cortex_a9_dp_shift): Likewise.
5152 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
5154 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
5156 (cortex_r4_mov): Likewise.
5157 (cortex_r4_alu_shift): Likewise.
5158 (cortex_r4_alu_shift_reg): Likewise.
5159 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
5160 (526_alu_shift_op): Likewise.
5161 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
5162 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
5163 (626te_alu_shift_op): Likewise.
5164 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
5165 (726te_alu_op): Likewise.
5166 (726te_alu_shift_op): Likewise.
5167 (726te_alu_shift_reg_op): Likewise.
5168 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
5169 (mp626_alu_shift_op): Likewise.
5170 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
5171 (pj4_alu_e1_conds): Likewise.
5172 (pj4_alu): Likewise.
5173 (pj4_alu_conds): Likewise.
5174 (pj4_shift): Likewise.
5175 (pj4_shift_conds): Likewise.
5176 (pj4_alu_shift): Likewise.
5177 (pj4_alu_shift_conds): Likewise.
5178 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5180 (cortexa7_older_only): Likewise.
5181 (cortexa7_younger): Likewise.
5183 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5185 * ipa-pure-const.c (generate_summary): Rename to...
5186 (pure_const_generate_summary): ... this.
5188 2013-07-17 Iain Sandoe <iain@codesourcery.com>
5190 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
5192 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
5195 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
5198 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5200 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
5201 enabled without -march=zEC12.
5202 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
5205 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5207 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
5208 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
5209 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
5210 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
5211 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
5212 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
5213 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
5215 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
5216 and ISA_HAS_NMADD3_NMSUB3.
5217 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
5218 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
5219 (nmsub4<mode>, nmsub3<mode>): Likewise.
5220 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
5222 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5224 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
5225 TARGET_MIPS5400 checking.
5227 2013-07-16 Jakub Jelinek <jakub@redhat.com>
5228 Peter Bergner <bergner@vnet.ibm.com>
5230 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
5231 registers in the comment.
5232 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
5233 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
5234 rather than FIRST_PSEUDO_REGISTERS.
5236 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
5238 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
5239 enable extra ISA flags with TARGET_HTM.
5241 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5243 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
5246 2013-07-15 Cong Hou <congh@google.com>
5248 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
5249 in compare function for sorting.
5251 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
5253 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
5254 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
5255 * config/rs6000/rs6000.opt: Add -mhtm option.
5256 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
5257 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
5258 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
5259 __HTM__ if the HTM instructions are available.
5260 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
5261 htm_spr_reg_operand): New define_predicates.
5262 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
5263 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
5265 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
5266 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
5267 HTM builtin functions.
5268 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
5269 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
5270 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
5271 (rs6000_builtin_mask_calculate): Likewise.
5272 (rs6000_option_override_internal): Likewise.
5273 (bdesc_htm): Add new HTM builtin support.
5274 (htm_spr_num): New function.
5275 (htm_spr_regno): Likewise.
5276 (rs6000_htm_spr_icode): Likewise.
5277 (htm_expand_builtin): Likewise.
5278 (htm_init_builtins): Likewise.
5279 (rs6000_expand_builtin): Add support for HTM builtin functions.
5280 (rs6000_init_builtins): Likewise.
5281 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
5283 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
5284 (TARGET_HTM, MASK_HTM): Define macros.
5285 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
5286 (FIXED_REGISTERS): Likewise.
5287 (CALL_USED_REGISTERS): Likewise.
5288 (CALL_REALLY_USED_REGISTERS): Likewise.
5289 (REG_ALLOC_ORDER): Likewise.
5290 (enum reg_class): Likewise.
5291 (REG_CLASS_NAMES): Likewise.
5292 (REG_CLASS_CONTENTS): Likewise.
5293 (REGISTER_NAMES): Likewise.
5294 (ADDITIONAL_REGISTER_NAMES): Likewise.
5295 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
5296 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
5297 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
5298 * config/rs6000/htm.md: New file.
5299 * config/rs6000/htmintrin.h: New file.
5300 * config/rs6000/htmxlintrin.h: New file.
5302 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
5304 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5305 Define SYMBOL_TINY_GOT, update comment.
5306 * config/aarch64/aarch64.c
5307 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
5308 (aarch64_expand_mov_immediate): Likewise.
5309 (aarch64_print_operand): Likewise.
5310 (aarch64_classify_symbol): Likewise.
5311 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
5312 (ldr_got_tiny): Define.
5314 2013-07-13 Tobias Grosser <tobias@grosser.es>
5316 PR tree-optimization/54094
5317 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
5318 scheduling dimension for the parallelism check from the polyhedral
5319 information in the AST.
5320 * graphite-dependences.c (carries_deps): Do not assume the schedule is
5323 2013-07-13 Jason Merrill <jason@redhat.com>
5325 * print-tree.c (debug_vec_tree): Use debug_raw.
5326 (debug_raw (vec<tree, va_gc> &)): New.
5327 (debug_raw (vec<tree, va_gc> *)): New.
5328 * tree.h: Declare them.
5330 2013-07-13 Bin Cheng <bin.cheng@arm.com>
5332 * ifcvt.c (ifcvt_after_combine): New static variable.
5333 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
5335 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
5336 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
5337 rest_of_handle_if_after_reload): Pass new argument for if_convert.
5339 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
5341 * config/mips/mips.c (mips_expand_call): Remove empty statement.
5343 2013-07-12 Michael Matz <matz@suse.de>
5346 * convert.c (convert_to_real): Reject non-float inner types.
5348 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
5350 * config/aarch64/aarch64-protos.h
5351 (aarch64_simd_immediate_valid_for_move): Remove.
5352 * config/aarch64/aarch64.c (simd_immediate_info): New member.
5353 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
5355 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
5357 2013-07-11 Steve Ellcey <sellcey@mips.com>
5359 * config/mips/mips.c (mips_conditional_register_usage): Do not
5360 use t[0-7] registers in MIPS16 mode when optimizing for size.
5362 2013-07-11 Sriraman Tallam <tmsriram@google.com>
5364 * config/i386/i386.c (dispatch_function_versions): Fix array
5365 indexing of function_version_info to match actual_versions.
5367 2013-07-11 Teresa Johnson <tejohnson@google.com>
5369 * vec.h (struct va_gc): Move release out-of-line.
5370 (va_gc::release): Call ggc_free on released vec.
5372 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5374 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
5375 Require GOT register as additional operand in UNSPEC.
5376 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
5377 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
5378 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
5379 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
5380 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
5381 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
5382 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
5384 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
5387 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
5388 name seen by assembler/linker rather if available.
5390 2013-07-11 Andreas Schwab <schwab@suse.de>
5392 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
5394 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
5396 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
5398 2013-07-10 Joseph Myers <joseph@codesourcery.com>
5400 * doc/tm.texi.in: Move hook documentation to ....
5401 * target.def: ... here.
5403 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
5405 * doc/tm.texi: Regenerate.
5407 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
5410 * doc/invoke.texi: Document -Wconditionally-supported.
5412 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5415 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
5418 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5421 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
5422 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
5423 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
5424 Remove duplicate devices.
5425 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
5426 * config/avr/t-multilib: Regenerate.
5427 * config/avr/avr-tables.opt: Regenerate.
5428 * doc/avr-mmcu.texi: Regenerate.
5430 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5433 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
5435 2013-07-10 Graham Stott <graham.stott@btinternet.com>
5437 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
5438 the cost of MULT when optimizing for size.
5440 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5442 * config/cr16/cr16-protos.h: Don't include target.h.
5444 2013-07-09 Joseph Myers <joseph@codesourcery.com>
5446 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
5447 adjust register size for TDmode and TFmode for VSX registers.
5449 2013-07-08 Kai Tietz <ktietz@redhat.com>
5452 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
5453 hook_bool_const_tree_true.
5455 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5457 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
5458 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
5459 * config/s390/s390.md: Define FPR*_REGNUM constants.
5460 Fix FPR2_REGNUM constant (18 -> 17).
5461 ("*trunc<BFP:mode><DFP_ALL:mode>2")
5462 ("*trunc<DFP_ALL:mode><BFP:mode>2")
5463 ("trunc<BFP:mode><DFP_ALL:mode>2")
5464 ("trunc<DFP_ALL:mode><BFP:mode>2")
5465 ("*extend<BFP:mode><DFP_ALL:mode>2")
5466 ("*extend<DFP_ALL:mode><BFP:mode>2")
5467 ("extend<BFP:mode><DFP_ALL:mode>2")
5468 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
5471 2013-07-08 Graham Stott <graham.stott@btinternet.com>
5473 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
5475 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5477 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
5478 and cfun_fpr_bit_p to cfun_fpr_save_p.
5479 (s390_frame_area, s390_register_info, s390_frame_info)
5480 (s390_emit_prologue, s390_emit_epilogue)
5481 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
5483 * config/s390/s390.h: Define *_REGNUM macros for floating point
5486 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
5488 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
5490 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
5492 PR rtl-optimization/57786
5493 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
5494 and break out of the loop when it is set to false.
5496 2013-07-08 Jakub Jelinek <jakub@redhat.com>
5499 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
5500 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
5501 (const_int 63)) 0)).
5502 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
5503 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
5504 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
5506 PR rtl-optimization/57829
5507 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
5508 mask bits outside of mode are just sign-extension from mode to HWI.
5510 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5512 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
5513 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
5514 adjust_address instead of change_address to keep info about alignment.
5515 (emit_strmov): Remove.
5516 (emit_memmov): New function.
5517 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
5518 (expand_movmem_epilogue): Likewise and return updated rtx for
5520 (expand_constant_movmem_prologue): Likewise and return updated rtx for
5521 destination and source.
5522 (decide_alignment): Refactor, handle vector_loop.
5523 (ix86_expand_movmem): Likewise.
5524 (ix86_expand_setmem): Likewise.
5525 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
5527 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
5529 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
5530 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
5532 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
5534 * config/i386/sse.md (sse_movlhps): Change alternative 3
5535 of operand 2 to "m".
5537 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
5540 * config/i386/sse.md (iptr): New mode attribute.
5541 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
5542 (<sse>_vm<plusminus_insn><mode>3): Ditto.
5543 (<sse>_vmmul<mode>3): Ditto.
5544 (<sse>_vmdiv<mode>3): Ditto.
5545 (sse_vmrcpv4sf2): Ditto.
5546 (<sse>_vmsqrt<mode>2): Ditto.
5547 (sse_vmrsqrtv4sf2): Ditto.
5548 (<sse>_vm<code><mode>3): Ditto.
5549 (avx_vmcmp<mode>3): Ditto.
5550 (<sse>_vmmaskcmp<mode>3): Ditto.
5551 (<sse>_comi): Ditto.
5552 (<sse>_ucomi): Ditto.
5553 (*xop_vmfrcz_<mode>): Ditto.
5554 (*fmai_fmadd_<mode>): Ditto.
5555 (*fmai_fmsub_<mode>): Ditto.
5556 (*fmai_fnmadd_<mode>): Ditto.
5557 (*fmai_fnmsub_<mode>): Ditto.
5558 (*fma4i_vmfmadd_<mode>): Ditto.
5559 (*fma4i_vmfmsub_<mode>): Ditto.
5560 (*fma4i_vmfnmadd_<mode>): Ditto.
5561 (*fma4i_vmfnmsub_<mode>): Ditto.
5562 (*xop_vmfrcz_<mode>): Ditto.
5563 (sse_cvtps2pi): Ditto.
5564 (sse_cvttps2pi): Ditto.
5565 (sse_cvtss2si): Ditto.
5566 (sse_cvtss2si_2): Ditto.
5567 (sse_cvtss2siq_2): Ditto.
5568 (sse_cvttss2si): Ditto.
5569 (sse_cvttss2siq): Ditto.
5570 (sse_cvtsd2si): Ditto.
5571 (sse_cvtsd2si_2): Ditto.
5572 (sse_cvtsd2siq_2): Ditto.
5573 (sse_cvttsd2si): Ditto.
5574 (sse_cvttsd2siq): Ditto.
5575 (sse_cvtsd2ss): Ditto.
5576 (sse_cvtss2sd): Ditto.
5577 (avx2_pbroadcast<mode>): Ditto.
5578 (avx2_pbroadcast<mode>_1): Ditto.
5579 (*avx_vperm_broadcast_v4sf): Ditto.
5581 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
5582 (sse_movlhps): Ditto.
5583 (sse_storehps): Ditto.
5584 (sse_loadhps): Ditto.
5585 (sse_storelps): Ditto.
5586 (sse_loadlps): Ditto.
5587 (*vec_concatv4sf): Ditto.
5588 (*vec_interleave_highv2df): Ditto.
5589 (*vec_interleave_lowv2df): Ditto.
5590 (*vec_extractv2df_1_sse): Ditto.
5591 (*vec_extractv2df_0_sse): Ditto.
5592 (sse2_storelpd): Ditto.
5593 (sse2_loadlpd): Ditto.
5594 (sse2_movsd): Ditto.
5595 (*vec_concatv4si): Ditto.
5596 (vec_concatv2di): Ditto.
5598 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
5599 for Intel asm dialect.
5600 (mmx_punpcklwd): Ditto.
5601 (mmx_punpckldq): Ditto.
5603 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
5604 for intel assembler dialect.
5606 2013-07-06 Jakub Jelinek <jakub@redhat.com>
5609 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
5610 for BUILT_IN_C{LZ,LRSB}*.
5611 * tree.h (CASE_INT_FN): Add FN##IMAX case.
5612 * tree-vrp.c (extract_range_basic): Handle
5613 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
5614 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
5615 fall thru to code calling set_value*.
5616 * builtins.c (expand_builtin): Remove *IMAX cases.
5617 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
5618 if width is bigger than 2*HWI.
5620 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
5622 PR rtl-optimization/55342
5623 * lra-int.h (lra_subreg_reload_pseudos): New.
5624 * lra.c: Add undoing optional reloads to the block diagram.
5625 (lra_subreg_reload_pseudos): New.
5626 (lra_optional_reload_pseudos): Change comments.
5627 (lra): Init and clear lra_subreg_reload_pseudos. Clear
5628 lra_optional_reload_pseudos after undo transformations.
5629 * lra-assigns.c (pseudo_prefix_title): New.
5630 (lra_setup_reg_renumber): Use it.
5631 (spill_for): Ditto. Check subreg reload pseudos too.
5632 (assign_by_spills): Consider subreg reload pseudos too.
5633 * lra-constraints.c (simplify_operand_subreg): Use
5634 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
5635 (curr_insn_transform): Recognize and do optional reloads.
5636 (undo_optional_reloads): New.
5637 (lra_undo_inheritance): Call undo_optional_reloads.
5639 2013-07-05 Thomas Quinot <quinot@adacore.com>
5641 * tree-complex.c (expand_complex_operations_1): Fix typo.
5643 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
5645 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
5646 (tune_params): New member 'const vec_costs'.
5647 * config/aarch64/aarch64.c (generic_vector_cost): New.
5648 (generic_tunings): New member 'generic_vector_cost'.
5649 (aarch64_builtin_vectorization_cost): New.
5650 (aarch64_add_stmt_cost): New.
5651 (TARGET_VECTORIZE_ADD_STMT_COST): New.
5652 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
5654 2013-07-03 Jakub Jelinek <jakub@redhat.com>
5657 * config/i386/predicates.md (vsib_address_operand): Disallow
5658 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
5660 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
5663 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
5664 expand_builtin_setjmp_receiver.
5665 (expand_label): Adjust, call expand_builtin_setjmp_receiver
5666 with NULL for the label parameter.
5667 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
5668 the frame-pointer. Adjust comments.
5669 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
5670 only if LABEL is non-NULL.
5672 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
5674 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
5675 (ARM_ABI_AAPCS64): Ditto.
5677 (ARM_DEFAULT_ABI): Ditto.
5679 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
5681 * config/aarch64/aarch64-builtins.c
5682 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
5683 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
5685 * config/aarch64/aarch64-simd.md
5686 (aarch64_ld1<VALL:mode>): New.
5687 (aarch64_st1<VALL:mode>): Likewise.
5688 * config/aarch64/arm_neon.h
5689 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
5691 2013-07-02 Sriraman Tallam <tmsriram@google.com>
5693 * config/i386/i386.c (gate_insert_vzeroupper): Check if
5695 (ix86_option_override_internal):Turn on all -mavx target flags by
5696 default as they are dependent on AVX anyway.
5698 2013-07-02 Cary Coutant <ccoutant@google.com>
5700 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
5702 (loc_checksum_ordered): Likewise.
5703 (hash_loc_operands): Remove inline keyword.
5705 2013-07-02 Jakub Jelinek <jakub@redhat.com>
5707 PR tree-optimization/57741
5708 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
5709 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
5710 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
5711 Allow REAL_CST step_exprs if flag_associative_math.
5712 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
5714 2013-07-02 Ian Bolton <ian.bolton@arm.com>
5716 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
5718 2013-07-02 Ian Bolton <ian.bolton@arm.com>
5720 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
5722 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5724 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
5726 (iorsi3_insn): Likewise.
5727 (arm_xorsi3): Likewise.
5729 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
5731 * arm.md (attribute "wtype"): Delete. Move attribute values from here
5733 (attribute "type"): ... here, and prefix with "wmmx_".
5734 (attribute "core_cycles"): Update for attribute changes.
5735 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
5736 (tbcstv4hi): Likewise.
5737 (tbcstv2si): Likewise.
5738 (iwmmxt_iordi3): Likewise.
5739 (iwmmxt_xordi3): Likewise.
5740 (iwmmxt_anddi3): Likewise.
5741 (iwmmxt_nanddi3): Likewise.
5742 (iwmmxt_arm_movdi): Likewise.
5743 (iwmmxt_movsi_insn): Likewise.
5744 (mov<mode>_internal): Likewise.
5745 (and<mode>3_iwmmxt): Likewise.
5746 (ior<mode>3_iwmmxt): Likewise.
5747 (xor<mode>3_iwmmxt): Likewise.
5748 (add<mode>3_iwmmxt): Likewise.
5749 (ssaddv8qi3): Likewise.
5750 (ssaddv4hi3): Likewise.
5751 (ssaddv2si3): Likewise.
5752 (usaddv8qi3): Likewise.
5753 (usaddv4hi3): Likewise.
5754 (usaddv2si3): Likewise.
5755 (sub<mode>3_iwmmxt): Likewise.
5756 (sssubv8qi3): Likewise.
5757 (sssubv4hi3): Likewise.
5758 (sssubv2si3): Likewise.
5759 (ussubv8qi3): Likewise.
5760 (ussubv4hi3): Likewise.
5761 (ussubv2si3): Likewise.
5762 (mulv4hi3_iwmmxt): Likewise.
5763 (smulv4hi3_highpart): Likewise.
5764 (umulv4hi3_highpart): Likewise.
5765 (iwmmxt_wmacs): Likewise.
5766 (iwmmxt_wmacsz): Likewise.
5767 (iwmmxt_wmacu): Likewise.
5768 (iwmmxt_wmacuz): Likewise.
5769 (iwmmxt_clrdi): Likewise.
5770 (iwmmxt_clrv8qi): Likewise.
5771 (iwmmxt_clr4hi): Likewise.
5772 (iwmmxt_clr2si): Likewise.
5773 (iwmmxt_uavgrndv8qi3): Likewise.
5774 (iwmmxt_uavgrndv4hi3): Likewise.
5775 (iwmmxt_uavgv8qi3): Likewise.
5776 (iwmmxt_uavgv4hi3): Likewise.
5777 (iwmmxt_tinsrb): Likewise.
5778 (iwmmxt_tinsrh): Likewise.
5779 (iwmmxt_tinsrw): Likewise.
5780 (iwmmxt_textrmub): Likewise.
5781 (iwmmxt_textrmsb): Likewise.
5782 (iwmmxt_textrmuh): Likewise.
5783 (iwmmxt_textrmsh): Likewise.
5784 (iwmmxt_textrmw): Likewise.
5785 (iwmxxt_wshufh): Likewise.
5786 (eqv8qi3): Likewise.
5787 (eqv4hi3): Likewise.
5788 (eqv2si3): Likewise.
5789 (gtuv8qi3): Likewise.
5790 (gtuv4hi3): Likewise.
5791 (gtuv2si3): Likewise.
5792 (gtv8qi3): Likewise.
5793 (gtv4hi3): Likewise.
5794 (gtv2si3): Likewise.
5795 (smax<mode>3_iwmmxt): Likewise.
5796 (umax<mode>3_iwmmxt): Likewise.
5797 (smin<mode>3_iwmmxt): Likewise.
5798 (umin<mode>3_iwmmxt): Likewise.
5799 (iwmmxt_wpackhss): Likewise.
5800 (iwmmxt_wpackwss): Likewise.
5801 (iwmmxt_wpackdss): Likewise.
5802 (iwmmxt_wpackhus): Likewise.
5803 (iwmmxt_wpackwus): Likewise.
5804 (iwmmxt_wpackdus): Likewise.
5805 (iwmmxt_wunpckihb): Likewise.
5806 (iwmmxt_wunpckihh): Likewise.
5807 (iwmmxt_wunpckihw): Likewise.
5808 (iwmmxt_wunpckilb): Likewise.
5809 (iwmmxt_wunpckilh): Likewise.
5810 (iwmmxt_wunpckilw): Likewise.
5811 (iwmmxt_wunpckehub): Likewise.
5812 (iwmmxt_wunpckehuh): Likewise.
5813 (iwmmxt_wunpckehuw): Likewise.
5814 (iwmmxt_wunpckehsb): Likewise.
5815 (iwmmxt_wunpckehsh): Likewise.
5816 (iwmmxt_wunpckehsw): Likewise.
5817 (iwmmxt_wunpckelub): Likewise.
5818 (iwmmxt_wunpckeluh): Likewise.
5819 (iwmmxt_wunpckeluw): Likewise.
5820 (iwmmxt_wunpckelsb): Likewise.
5821 (iwmmxt_wunpckelsh): Likewise.
5822 (iwmmxt_wunpckelsw): Likewise.
5823 (ror<mode>3): Likewise.
5824 (ashr<mode>3_iwmmxt): Likewise.
5825 (lshr<mode>3_iwmmxt): Likewise.
5826 (ashl<mode>3_iwmmxt): Likewise.
5827 (ror<mode>3_di): Likewise.
5828 (ashr<mode>3_di): Likewise.
5829 (lshr<mode>3_di): Likewise.
5830 (ashl<mode>3_di): Likewise.
5831 (iwmmxt_wmadds): Likewise.
5832 (iwmmxt_wmaddu): Likewise.
5833 (iwmmxt_tmia): Likewise.
5834 (iwmmxt_tmiaph): Likewise.
5835 (iwmmxt_tmiabb): Likewise.
5836 (iwmmxt_tmiatb): Likewise.
5837 (iwmmxt_tmiabt): Likewise.
5838 (iwmmxt_tmiatt): Likewise.
5839 (iwmmxt_tmovmskb): Likewise.
5840 (iwmmxt_tmovmskh): Likewise.
5841 (iwmmxt_tmovmskw): Likewise.
5842 (iwmmxt_waccb): Likewise.
5843 (iwmmxt_wacch): Likewise.
5844 (iwmmxt_waccw): Likewise.
5845 (iwmmxt_waligni): Likewise.
5846 (iwmmxt_walignr): Likewise.
5847 (iwmmxt_walignr0): Likewise.
5848 (iwmmxt_walignr1): Likewise.
5849 (iwmmxt_walignr2): Likewise.
5850 (iwmmxt_walignr3): Likewise.
5851 (iwmmxt_wsadb): Likewise.
5852 (iwmmxt_wsadh): Likewise.
5853 (iwmmxt_wsadbz): Likewise.
5854 (iwmmxt_wsadhz): Likewise.
5855 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
5856 (iwmmxt_wabsdiffb): Likewise.
5857 (iwmmxt_wabsdiffh): Likewise.
5858 (iwmmxt_wabsdiffw): Likewise.
5859 (iwmmxt_waddsubhx): Likewise
5860 (iwmmxt_wsubaddhx): Likewise.
5861 (addc<mode>3): Likewise.
5862 (iwmmxt_avg4): Likewise.
5863 (iwmmxt_avg4r): Likewise.
5864 (iwmmxt_wmaddsx): Likewise.
5865 (iwmmxt_wmaddux): Likewise.
5866 (iwmmxt_wmaddsn): Likewise.
5867 (iwmmxt_wmaddun): Likewise.
5868 (iwmmxt_wmulwsm): Likewise.
5869 (iwmmxt_wmulwum): Likewise.
5870 (iwmmxt_wmulsmr): Likewise.
5871 (iwmmxt_wmulumr): Likewise.
5872 (iwmmxt_wmulwsmr): Likewise.
5873 (iwmmxt_wmulwumr): Likewise.
5874 (iwmmxt_wmulwl): Likewise.
5875 (iwmmxt_wqmulm): Likewise.
5876 (iwmmxt_wqmulwm): Likewise.
5877 (iwmmxt_wqmulmr): Likewise.
5878 (iwmmxt_wqmulwmr): Likewise.
5879 (iwmmxt_waddbhusm): Likewise.
5880 (iwmmxt_waddbhusl): Likewise.
5881 (iwmmxt_wqmiabb): Likewise.
5882 (iwmmxt_wqmiabt): Likewise.
5883 (iwmmxt_wqmiatb): Likewise.
5884 (iwmmxt_wqmiatt): Likewise.
5885 (iwmmxt_wqmiabbn): Likewise.
5886 (iwmmxt_wqmiabtn): Likewise.
5887 (iwmmxt_wqmiatbn): Likewise.
5888 (iwmmxt_wqmiattn): Likewise.
5889 (iwmmxt_wmiabb): Likewise.
5890 (iwmmxt_wmiabt): Likewise.
5891 (iwmmxt_wmiatb): Likewise.
5892 (iwmmxt_wmiatt): Likewise.
5893 (iwmmxt_wmiabbn): Likewise.
5894 (iwmmxt_wmiabtn): Likewise.
5895 (iwmmxt_wmiatbn): Likewise.
5896 (iwmmxt_wmiattn): Likewise.
5897 (iwmmxt_wmiawbb): Likewise.
5898 (iwmmxt_wmiawbt): Likewise.
5899 (iwmmxt_wmiawtb): Likewise.
5900 (iwmmxt_wmiawtt): Likewise.
5901 (iwmmxt_wmiawbbn): Likewise.
5902 (iwmmxt_wmiawbtn): Likewise.
5903 (iwmmxt_wmiawtbn): Likewise.
5904 (iwmmxt_wmiawttn): Likewise.
5905 (iwmmxt_wmerge): Likewise.
5906 (iwmmxt_tandc<mode>3): Likewise.
5907 (iwmmxt_torc<mode>3): Likewise.
5908 (iwmmxt_torvsc<mode>3): Likewise.
5909 (iwmmxt_textrc<mode>3): Likewise.
5910 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
5911 (wmmxt_pack): Likewise.
5912 (wmmxt_mult_c1): Likewise.
5913 (wmmxt_mult_c2): Likewise.
5914 (wmmxt_alu_c1): Likewise.
5915 (wmmxt_alu_c2): Likewise.
5916 (wmmxt_alu_c3): Likewise.
5917 (wmmxt_transfer_c1): Likewise.
5918 (wmmxt_transfer_c2): Likewise.
5919 (wmmxt_transfer_c3): Likewise.
5920 (marvell_f_iwmmxt_wstr): Likewise.
5921 (marvell_f_iwmmxt_wldr): Likewise.
5923 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5925 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5927 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5930 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5931 * lra-constraints.c (need_for_split_p): Check call used hard regs
5932 living through calls.
5934 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5935 call used regs for call insn.
5937 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5940 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5941 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5943 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5945 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5947 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5949 * lra-constraints.c (need_for_split_p): Check call used hard regs
5950 living through calls.
5952 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5955 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5956 to tie with any other modes. Eliminate Altivec vector mode tests,
5957 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5958 code, to return 0 if testing MODE2 for a condition, if we've
5959 already tested MODE1 for the same condition.
5961 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5963 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5966 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5968 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5969 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5971 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5973 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5975 (aarch64_symbolic_constant_p): Remove.
5976 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5977 static. Fix line length and white space.
5978 (aarch64_symbolic_constant_p): Remove.
5979 * config/aarch64/predicates.md (aarch64_valid_symref):
5980 Use aarch64_classify_symbol_expression.
5982 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5984 * config/arm/constraints.md (Ts): New constraint.
5985 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5987 (compare_scc): Use "Ts" constraint for operand 0.
5988 (ior_scc_scc): Likewise.
5989 (and_scc_scc): Likewise.
5990 (and_scc_scc_nodom): Likewise.
5991 (ior_scc_scc_cmp): Likewise for operand 7.
5992 (and_scc_scc_cmp): Likewise.
5993 * config/arm/thumb2.md (thumb2_movsi_insn):
5994 Add alternatives for 16-bit encodings.
5995 (thumb2_movhi_insn): Likewise.
5996 (thumb2_movsicc_insn): Likewise.
5997 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5998 (thumb2_negscc): Use "Ts" constraint.
5999 Move mvn instruction outside cond_exec block.
6000 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
6001 for 16-bit encodings.
6003 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6005 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
6007 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
6008 (mulsi3subsi): Likewise.
6009 (mulsidi3adddi): Likewise.
6010 (mulsidi3_v6): Likewise.
6011 (umulsidi3_v6): Likewise.
6012 (umulsidi3adddi_v6): Likewise.
6013 (smulsi3_highpart_v6): Likewise.
6014 (umulsi3_highpart_v6): Likewise.
6015 (mulhisi3tb): Likewise.
6016 (mulhisi3bt): Likewise.
6017 (mulhisi3tt): Likewise.
6018 (maddhisi4): Likewise.
6019 (maddhisi4tb): Likewise.
6020 (maddhisi4tt): Likewise.
6021 (maddhidi4): Likewise.
6022 (maddhidi4tb): Likewise.
6023 (maddhidi4tt): Likewise.
6024 (zeroextractsi_compare0_scratch): Likewise.
6025 (insv_zero): Likewise.
6026 (insv_t2): Likewise.
6027 (anddi_notzesidi_di): Likewise.
6028 (anddi_notsesidi_di): Likewise.
6029 (andsi_notsi_si): Likewise.
6030 (iordi_zesidi_di): Likewise.
6031 (xordi_zesidi_di): Likewise.
6032 (andsi_iorsi3_notsi): Likewise.
6034 (smax_m1): Likewise.
6036 (not_shiftsi): Likewise.
6037 (unaligned_loadsi): Likewise.
6038 (unaligned_loadhis): Likewise.
6039 (unaligned_loadhiu): Likewise.
6040 (unaligned_storesi): Likewise.
6041 (unaligned_storehi): Likewise.
6042 (extv_reg): Likewise.
6043 (extzv_t2): Likewise.
6045 (udivsi3): Likewise.
6046 (arm_zero_extendhisi2addsi): Likewise.
6047 (arm_zero_extendqisi2addsi): Likewise.
6048 (compareqi_eq0): Likewise.
6049 (arm_extendhisi2_v6): Likewise.
6050 (arm_extendqisi2addsi): Likewise.
6051 (arm_movt): Likewise.
6052 (thumb2_ldrd): Likewise.
6053 (thumb2_ldrd_base): Likewise.
6054 (thumb2_ldrd_base_neg): Likewise.
6055 (thumb2_strd): Likewise.
6056 (thumb2_strd_base): Likewise.
6057 (thumb2_strd_base_neg): Likewise.
6058 (arm_negsi2): Add alternative for 16-bit encoding.
6059 (arm_one_cmplsi2): Likewise.
6061 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6063 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
6064 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
6065 (movdfcc): Likewise.
6066 * config/arm/vfp.md (*thumb2_movsf_vfp):
6067 Disable predication for arm_restrict_it.
6068 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
6069 (*thumb2_movdfcc_vfp): Likewise.
6070 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
6071 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
6072 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
6073 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
6074 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
6075 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
6076 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
6077 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
6078 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
6079 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
6080 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
6081 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
6082 Disable predication for arm_restrict_it.
6084 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
6086 * config/i386/bmiintrin.h (_bextr_u32): New.
6087 (_bextr_u64): Ditto.
6089 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
6091 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
6092 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
6093 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
6094 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
6095 * config/mips/n32-elf.h: ...this new file.
6097 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
6100 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
6101 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
6103 2013-06-27 Catherine Moore <clm@codesourcery.com>
6105 * config/mips/mips-tables.opt: Regenerate.
6106 * config/mips/mips-cpus.def: Add m14ke and m14kec.
6107 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
6108 * doc/invoke.texi: Add -m14kc.
6110 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6113 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
6114 constraints of operand 1 and 2.
6117 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
6118 to match RTL canonicalization. Swap predicates and
6119 constraints of operand 1 and 2.
6121 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
6123 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
6124 Process OP_INOUT regs for splitting too.
6126 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6128 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
6129 decl before the loop, initialize to NULL.
6130 (vectorizable_load): Initialize ptr_incr to NULL.
6132 2013-06-27 Martin Jambor <mjambor@suse.cz>
6135 * ipa-ref.h (ipa_maybe_record_reference): Declare.
6136 * ipa-ref.c (ipa_maybe_record_reference): New function.
6137 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
6138 * ipa-cp.c (create_specialized_node): Record potential references from
6140 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
6142 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6144 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
6145 parameter 'mode' of type 'enum machine_mode mode'; change to pass
6146 'mode' to force_reg.
6147 (aarch64_add_offset): Update calls to aarch64_force_temporary.
6148 (aarch64_expand_mov_immediate): Likewise.
6150 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6152 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
6153 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
6155 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6157 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
6158 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6159 (struct machine_function): Add tbegin_p.
6160 (s390_canonicalize_comparison): Fold CC mode compares to
6161 conditional jump if possible.
6162 (s390_emit_jump): Return the emitted jump.
6163 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
6164 Handle CCRAWmode compares.
6165 (s390_option_override): Default to -mhtm if available.
6166 (s390_reg_clobbered_rtx): Handle floating point regs as well.
6167 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
6168 FPRs instead of df_regs_ever_live_p.
6169 (s390_optimize_nonescaping_tx): New function.
6170 (s390_init_frame_layout): Extend clobbered_regs array to cover
6172 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
6173 (s390_expand_tbegin): New function.
6174 (enum s390_builtin): New enum definition.
6175 (code_for_builtin): New array definition.
6176 (s390_init_builtins): New function.
6177 (s390_expand_builtin): New function.
6178 (TARGET_INIT_BUILTINS): Define.
6179 (TARGET_EXPAND_BUILTIN): Define.
6180 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
6181 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
6182 (s390_alc_comparison): Likewise.
6183 * config/s390/s390-modes.def: Add CCRAWmode.
6184 * config/s390/s390.h (processor_flags): Add PF_TX.
6185 (TARGET_CPU_HTM): Define macro.
6186 (TARGET_HTM): Define macro.
6187 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
6188 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
6189 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6190 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
6191 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
6192 (TBEGIN_MASK, TBEGINC_MASK): New constants.
6193 ("*cc_to_int"): Move up.
6194 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
6195 constants other than 0.
6196 ("*ccraw_to_int"): New insn and splitter definition.
6197 ("tbegin", "tbegin_nofloat", "tbegin_retry")
6198 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
6199 ("tx_assist"): New expander.
6200 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
6201 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
6202 * config/s390/s390.opt: Add -mhtm option.
6203 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
6204 * config/s390/htmxlintrin.h: New file.
6205 * config/s390/htmintrin.h: New file.
6206 * config/s390/s390intrin.h: New file.
6207 * doc/extend.texi: Document htm builtins.
6208 * config.gcc: Add the new header files to extra_headers.
6210 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
6212 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
6213 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
6215 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
6216 Pat Haugen <pthaugen@us.ibm.com>
6217 Peter Bergner <bergner@vnet.ibm.com>
6219 * config/rs6000/power8.md: New.
6220 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
6221 setting for power8 entry.
6222 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
6223 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
6224 test for Power4/Power5 only.
6225 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
6227 (force_new_group): Adjust comment.
6228 * config/rs6000/rs6000.md: Include power8.md.
6230 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
6232 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
6233 * config/arm/arm-protos.h (arm_max_conditional_execute): New
6235 (tune_params): Update comment.
6236 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
6237 (arm_max_conditional_execute): New function.
6238 (thumb2_final_prescan_insn): Use max_insn_skipped and
6239 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
6241 2013-06-25 Jakub Jelinek <jakub@redhat.com>
6243 PR tree-optimization/57705
6244 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
6245 SSA_NAME step, provided that it is not defined inside the loop.
6246 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
6247 (get_initial_def_for_induction): Handle SSA_NAME IV step.
6249 2013-06-25 Martin Jambor <mjambor@suse.cz>
6252 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
6253 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
6255 (ipa_note_param_call): Initialize member_ptr flag.
6256 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
6257 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
6258 (ipa_write_indirect_edge_info): Stream member_ptr flag.
6259 (ipa_read_indirect_edge_info): Likewise.
6261 2013-06-25 Richard Biener <rguenther@suse.de>
6264 * passes.c (init_optimization_passes): Move pass_fold_builtins
6265 and pass_dce earlier with -Og.
6267 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
6269 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
6270 <BIT_FIELD_REF>: Remove trailing TAB.
6271 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
6274 2013-06-24 Martin Jambor <mjambor@suse.cz>
6276 PR tree-optimization/57358
6277 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
6278 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
6279 (ipa_analyze_params_uses): Generate pessimistic info when true.
6281 2013-06-24 Martin Jambor <mjambor@suse.cz>
6283 PR tree-optimization/57539
6284 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
6285 global.inlined_to of the new node to it. All callers changed.
6286 * ipa-inline-transform.c (clone_inlined_nodes): New variable
6287 inlining_into, pass it to cgraph_clone_node.
6288 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
6289 ipa_free_edge_args_substructures.
6290 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
6291 rdesc linked list. Do not assert rdesc edges have inlined caller.
6292 Assert we have found an rdesc in the rdesc list.
6294 2013-06-24 Richard Biener <rguenther@suse.de>
6296 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
6297 (pointer_set_lookup): Declare.
6298 (class pointer_map): New template class implementing a
6299 generic pointer to T map.
6300 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
6301 pointer_map<T>::contains, pointer_map<T>::insert,
6302 pointer_map<T>::traverse): New functions.
6303 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
6304 (pointer_set_lookup): New function.
6305 (pointer_set_contains): Use pointer_set_lookup.
6306 (pointer_set_insert): Likewise.
6307 (insert_aux): Remove.
6308 (struct pointer_map_t): Embed a pointer_set_t.
6309 (pointer_map_create): Adjust.
6310 (pointer_map_destroy): Likewise.
6311 (pointer_map_contains): Likewise.
6312 (pointer_map_insert): Likewise.
6313 (pointer_map_traverse): Likewise.
6314 * tree-streamer.h (struct streamer_tree_cache_d): Use a
6315 pointer_map<unsigned> instead of a pointer_map_t.
6316 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
6317 (streamer_tree_cache_lookup): Likewise.
6318 (streamer_tree_cache_create): Likewise.
6319 (streamer_tree_cache_delete): Likewise.
6320 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
6321 pointer_map<unsigned> instead of a pointer_map_t.
6322 (lto_init_tree_ref_encoder): Adjust.
6323 (lto_destroy_tree_ref_encoder): Likewise.
6324 * lto-section-out.c (lto_output_decl_index): Likewise.
6325 (lto_record_function_out_decl_state): Likewise.
6326 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
6328 2013-06-24 Richard Biener <rguenther@suse.de>
6330 PR tree-optimization/57488
6331 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
6333 2013-06-24 Alan Modra <amodra@gmail.com>
6335 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
6336 (gen_easy_altivec_constant): Likewise.
6337 * config/rs6000/predicates.md (easy_vector_constant_add_self,
6338 easy_vector_constant_msb): Likewise.
6340 2013-06-23 Jakub Jelinek <jakub@redhat.com>
6343 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
6344 add missing return true.
6346 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
6349 * config/sh/predicates.md (general_extend_operand): Invoke
6350 general_movsrc_operand for memory operands.
6351 (general_movsrc_operand): Allow reg+reg addressing, do not use
6352 general_operand for memory operands.
6354 2013-06-23 Sriraman Tallam <tmsriram@google.com>
6356 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
6357 when current target options does not apply.
6358 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
6359 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
6360 * config/i386/bmiintrin.h: Pass appropriate target
6361 attributes to header.
6362 * config/i386/mmintrin.h: Ditto.
6363 * config/i386/nmmintrin.h: Ditto.
6364 * config/i386/avx2intrin.h: Ditto.
6365 * config/i386/fxsrintrin.h: Ditto.
6366 * config/i386/tbmintrin.h: Ditto.
6367 * config/i386/xsaveintrin.h: Ditto.
6368 * config/i386/f16cintrin.h: Ditto.
6369 * config/i386/xtestintrin.h: Ditto.
6370 * config/i386/xsaveoptintrin.h: Ditto.
6371 * config/i386/bmi2intrin.h: Ditto.
6372 * config/i386/lzcntintrin.h: Ditto.
6373 * config/i386/smmintrin.h: Ditto.
6374 * config/i386/wmmintrin.h: Ditto.
6375 * config/i386/x86intrin.h: Remove all header include guards.
6376 * config/i386/prfchwintrin.h: Ditto.
6377 * config/i386/pmmintrin.h: Ditto.
6378 * config/i386/tmmintrin.h: Ditto.
6379 * config/i386/xmmintrin.h: Ditto.
6380 * config/i386/popcntintrin.h: Ditto.
6381 * config/i386/rdseedintrin.h: Ditto.
6382 * config/i386/ammintrin.h: Ditto.
6383 * config/i386/emmintrin.h: Ditto.
6384 * config/i386/immintrin.h: Remove all header include guards.
6385 * config/i386/fma4intrin.h: Ditto.
6386 * config/i386/lwpintrin.h: Ditto.
6387 * config/i386/xopintrin.h: Ditto.
6388 * config/i386/ia32intrin.h: Ditto.
6389 * config/i386/avxintrin.h: Ditto.
6390 * config/i386/rtmintrin.h: Ditto.
6391 * config/i386/fmaintrin.h: Ditto.
6392 * config/i386/mm3dnow.h: Ditto.
6394 2013-06-22 Sriraman Tallam <tmsriram@google.com>
6396 * common/config/i386/i386-common.c: Handle LZCNT.
6398 2013-06-22 Andi Kleen <ak@linux.intel.com>
6400 * doc/extend.texi: Use __atomic_store_n instead of
6401 __atomic_store in HLE example.
6403 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
6405 * config/sh/sh.c: Remove <cstdlib> workaround.
6407 2013-06-21 Andi Kleen <ak@linux.intel.com>
6409 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
6411 2013-06-21 Andi Kleen <ak@linux.intel.com>
6413 * doc/extend.texi: Document that __atomic_clear and
6414 __atomic_test_and_set should only be used with bool.
6416 2013-06-20 Jan Hubicka <jh@suse.cz>
6418 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
6420 * tree.c (decls_same_for_odr): New function.
6421 (same_for_edr): New function.
6422 (types_same_for_odr): New function.
6423 (get_binfo_at_offset): Use it.
6424 * tree.h (types_same_for_odr): Declare.
6426 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
6427 Jason Merrill <jason@redhat.com>
6429 * system.h: Include <cstdlib> as well as <stdlib.h>.
6431 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
6434 * config/i386/i386.c (construct_container): Report error if
6435 long double is used with disabled x87 float returns.
6437 2013-06-20 Jan Hubicka <jh@suse.cz>
6439 * lto-cgraph.c (input_symtab): Do not set cgraph state.
6441 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
6443 PR rtl-optimization/57425
6444 PR rtl-optimization/57569
6445 * alias.c (write_dependence_p): Remove parameters mem_mode and
6446 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
6447 Changed all callers.
6448 (canon_anti_dependence): Get comments and semantics in sync.
6449 Add parameter mem_canonicalized. Changed all callers.
6450 * rtl.h (canon_anti_dependence): Update prototype.
6452 2013-06-20 Richard Biener <rguenther@suse.de>
6454 * data-streamer-in.c (streamer_read_uhwi): Optimize single
6455 byte case, inline streamer_read_uchar and defer section
6458 2013-06-20 Richard Biener <rguenther@suse.de>
6460 PR tree-optimization/57584
6461 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
6462 SSA names into the expanded expression that take part in
6463 abnormal coalescing.
6465 2013-06-19 Sharad Singhai <singhai@google.com>
6467 * gcov.c (print_usage): Handle new option.
6468 (process_args): Ditto.
6469 (get_gcov_intermediate_filename): New function.
6470 (output_intermediate_file): New function.
6471 (output_gcov_file): New function
6472 (generate_results): Handle new option.
6473 (release_function): Relase demangled name.
6474 (read_graph_file): Handle demangled name.
6475 (output_lines): Ditto.
6476 * doc/gcov.texi: Document gcov intermediate format.
6478 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
6481 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
6482 (lra_emit_add): Use the functions. Add comment about Y as an
6485 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
6488 * collect2.c (collect_atexit): New.
6489 (collect_exit): Delete.
6490 (main): Register collect_atexit with atexit.
6491 (collect_wait): Change collect_exit to exit.
6493 * collect2.h (collect_exit): Delete.
6494 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
6496 2013-06-19 Wei Mi <wmi@google.com>
6498 PR rtl-optimization/57518
6499 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
6500 if regno is used in paradoxical subreg.
6501 (update_equiv_regs): Check pdx_subregs[regno] before
6502 set a reg to be equivalent with a mem.
6504 2013-06-19 Matthias Klose <doko@ubuntu.com>
6507 * file-find.h (find_a_file): Add a mode parameter.
6508 * file-find.c (find_a_file): Likewise.
6509 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
6510 with X_OK for the executables.
6511 * collect2.c (main): Call find_a_file with X_OK.
6513 2013-06-19 Steve Ellcey <sellcey@mips.com>
6516 * config/mips/mips.md (casesi_internal_mips16_<mode>):
6517 Use NEXT_INSN instead of next_real_insn.
6519 2013-06-19 Jan Hubicka <jh@suse.cz>
6521 * cgraph.h (const_value_known_p): Replace by ...
6522 (ctor_for_folding): .. this one.
6523 * cgraphunit.c (process_function_and_variable_attributes): Use it.
6524 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
6525 * expr.c (expand_expr_real_1): Likewise.
6526 (string_constant): Likewise.
6527 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
6528 * ipa.c (process_references): Likewise.
6529 (symtab_remove_unreachable_nodes): Likewise.
6530 * ipa-inline-analysis.c (param_change_prob): Likewise.
6531 * gimple-fold.c (canonicalize_constructor_val): Likewise.
6532 (get_base_constructor): Likwise.
6533 * varpool.c (varpool_remove_node): Likewise.
6534 (varpool_remove_initializer): LIkewise.
6535 (dump_varpool_node): LIkwise.
6536 (const_value_known_p): Rewrite to ...
6537 (ctor_for_folding): ... this one.
6539 2013-06-19 Jakub Jelinek <jakub@redhat.com>
6542 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
6543 PERSONALITY in $PATH derived prefixes.
6545 2013-06-19 Jeff Law <law@redhat.com>
6547 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
6550 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
6551 (simplify_bitwise_binary): Use it to simpify certain binary ops on
6554 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
6556 * config/arm/vfp.md: Move VFP instruction classification documentation
6558 * config/arm/arm.md: ... here. Update instruction classification
6561 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
6563 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
6564 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
6565 pattern. Use more efficient sequences on ARMv5 and Thumb2.
6567 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
6570 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6571 with NEXT_INSN. Use tablejump_p to check for jump table data
6574 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
6577 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
6578 that now in C++ the value is correct per the C++ standards.
6580 2013-06-19 Richard Biener <rguenther@suse.de>
6582 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
6585 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6588 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6591 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6592 with next_active_insn.
6594 2013-06-18 Sriraman Tallam <tmsriram@google.com>
6596 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
6597 functions are inlined during failures to flag an error.
6598 * tree-inline.c (expand_call_inline): Allow the error to be flagged
6599 in early inline pass.
6601 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
6603 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
6606 2013-06-18 Julian Brown <julian@codesourcery.com>
6608 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
6609 Permit virtual register pre-reload if !strict.
6610 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
6612 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
6614 * config/arm/neon.md (movmisalign<mode>): Use
6615 neon_perm_struct_or_reg_operand instead of
6616 neon_struct_or_register_operand.
6617 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
6618 neon_permissive_struct_operand instead of neon_struct_operand.
6619 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
6620 neon_vector_mem_operand.
6621 * config/arm/predicates.md (neon_struct_operand): Adjust call to
6622 neon_vector_mem_operand.
6623 (neon_permissive_struct_operand): New.
6624 (neon_struct_or_register_operand): Rename to...
6625 (neon_perm_struct_or_reg_operand): This. Adjust call to
6626 neon_vector_mem_operand.
6628 2013-06-18 Richard Biener <rguenther@suse.de>
6630 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
6631 * lto-streamer.h: Include pointer-set.h.
6632 (struct lto_decl_slot): Remove.
6633 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
6634 Remove next_index entry.
6635 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
6636 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
6637 (lto_init_tree_ref_encoder): Adjust.
6638 (lto_destroy_tree_ref_encoder): Likewise.
6639 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
6640 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
6641 (lto_output_decl_index): Adjust.
6642 (lto_new_out_decl_state): Likewise.
6643 (lto_record_function_out_decl_state): Likewise.
6644 * lto-streamer-out.c (copy_function): Likewise.
6646 2013-06-18 Richard Biener <rguenther@suse.de>
6648 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
6649 * cgraphunit.c: Include cfgloop.h.
6650 (init_lowered_empty_function): Initialize the loop tree.
6651 (assemble_thunk): Insert new BBs into loops.
6653 2013-06-18 Richard Biener <rguenther@suse.de>
6655 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
6656 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
6657 the map from cache entry to cache index optional.
6658 (streamer_tree_cache_replace_tree): Adjust accordingly.
6659 (streamer_tree_cache_append): Likewise.
6660 (streamer_tree_cache_delete): Likewise.
6661 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
6662 streamer cache map from cache entry to cache index.
6663 * lto-streamer-out.c (create_output_block): Adjust.
6665 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
6667 * config/arm/arm.md (attribute "insn"): Move multiplication and
6668 division attributes to...
6669 (attribute "type"): ... here. Remove mult.
6670 (attribute "mul32"): New attribute.
6671 (attribute "mul64"): Add umaal.
6672 (*arm_mulsi3): Update attributes.
6673 (*arm_mulsi3_v6): Likewise.
6674 (*thumb_mulsi3): Likewise.
6675 (*thumb_mulsi3_v6): Likewise.
6676 (*mulsi3_compare0): Likewise.
6677 (*mulsi3_compare0_v6): Likewise.
6678 (*mulsi_compare0_scratch): Likewise.
6679 (*mulsi_compare0_scratch_v6): Likewise.
6680 (*mulsi3addsi): Likewise.
6681 (*mulsi3addsi_v6): Likewise.
6682 (*mulsi3addsi_compare0): Likewise.
6683 (*mulsi3addsi_compare0_v6): Likewise.
6684 (*mulsi3addsi_compare0_scratch): Likewise.
6685 (*mulsi3addsi_compare0_scratch_v6): Likewise.
6686 (*mulsi3subsi): Likewise.
6687 (*mulsidi3adddi): Likewise.
6688 (*mulsi3addsi_v6): Likewise.
6689 (*mulsidi3adddi_v6): Likewise.
6690 (*mulsidi3_nov6): Likewise.
6691 (*mulsidi3_v6): Likewise.
6692 (*umulsidi3_nov6): Likewise.
6693 (*umulsidi3_v6): Likewise.
6694 (*umulsidi3adddi): Likewise.
6695 (*umulsidi3adddi_v6): Likewise.
6696 (*smulsi3_highpart_nov6): Likewise.
6697 (*smulsi3_highpart_v6): Likewise.
6698 (*umulsi3_highpart_nov6): Likewise.
6699 (*umulsi3_highpart_v6): Likewise.
6700 (mulhisi3): Likewise.
6701 (*mulhisi3tb): Likewise.
6702 (*mulhisi3bt): Likewise.
6703 (*mulhisi3tt): Likewise.
6704 (maddhisi4): Likewise.
6705 (*maddhisi4tb): Likewise.
6706 (*maddhisi4tt): Likewise.
6707 (maddhidi4): Likewise.
6708 (*maddhidi4tb): Likewise.
6709 (*maddhidi4tt): Likewise.
6711 (udivsi3): Likewise.
6712 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
6713 (thumb2_mulsi_short_compare0): Likewise.
6714 (thumb2_mulsi_short_compare0_scratch): Likewise.
6715 * config/arm/arm1020e.md (1020mult1): Update attribute change.
6716 (1020mult2): Likewise.
6717 (1020mult3): Likewise.
6718 (1020mult4): Likewise.
6719 (1020mult5): Likewise.
6720 (1020mult6): Likewise.
6721 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
6723 (cortex_a15_mult64): Likewise.
6724 (cortex_a15_sdiv): Likewise.
6725 (cortex_a15_udiv): Likewise.
6726 * config/arm/arm1026ejs.md (mult1): Update attribute change.
6732 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
6733 (pj4_ir_div): Likewise.
6734 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
6735 (11_mult2): Likewise.
6736 (11_mult3): Likewise.
6737 (11_mult4): Likewise.
6738 (11_mult5): Likewise.
6739 (11_mult6): Likewise.
6740 (11_mult7): Likewise.
6741 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
6742 (cortex_a8_mla): Likewise.
6743 (cortex_a8_mull): Likewise.
6744 (cortex_a8_smulwy): Likewise.
6745 (cortex_a8_smlald): Likewise.
6746 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
6747 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
6748 (cortex_r4_mul_3): Likewise.
6749 (cortex_r4_mla_4): Likewise.
6750 (cortex_r4_mla_3): Likewise.
6751 (cortex_r4_smlald): Likewise.
6752 (cortex_r4_mull): Likewise.
6753 (cortex_r4_sdiv): Likewise.
6754 (cortex_r4_udiv): Likewise.
6755 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
6756 (cortex_a7_idiv): Likewise.
6757 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
6758 (9_mult2): Likewise.
6759 (9_mult3): Likewise.
6760 (9_mult4): Likewise.
6761 (9_mult5): Likewise.
6762 (9_mult6): Likewise.
6763 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
6764 (cortex_a53_sdiv): Likewise.
6765 (cortex_a53_udiv): Likewise.
6766 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
6767 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
6768 (mp626_mult2): Likewise.
6769 (mp626_mult3): Likewise.
6770 (mp626_mult4): Likewise.
6771 * config/arm/fa526.md (526_mult1): Update attribute change.
6772 (526_mult2): Likewise.
6773 * config/arm/arm-generic.md (mult): Update attribute change.
6774 (mult_ldsched_strongarm): Likewise.
6775 (mult_ldsched): Likewise.
6776 (multi_cycle): Likewise.
6777 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
6778 * config/arm/fa606te.md (606te_mult1): Update attribute change.
6779 (606te_mult2): Likewise.
6780 (606te_mult3): Likewise.
6781 (606te_mult4): Likewise.
6782 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
6783 (cortex_a9_mac16): Likewise.
6784 (cortex_a9_multiply): Likewise.
6785 (cortex_a9_mac): Likewise.
6786 (cortex_a9_multiply_long): Likewise.
6787 * config/arm/fa626te.md (626te_mult1): Update attribute change.
6788 (626te_mult2): Likewise.
6789 (626te_mult3): Likewise.
6790 (626te_mult4): Likewise.
6792 2013-06-18 Richard Biener <rguenther@suse.de>
6795 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
6797 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6800 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6801 with next_active_insn.
6803 2013-06-18 Alan Modra <amodra@gmail.com>
6805 * config/rs6000/rs6000.h (enum data_align): New.
6806 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
6807 (DATA_ABI_ALIGNMENT): Define.
6808 (CONSTANT_ALIGNMENT): Correct comment.
6809 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
6810 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
6812 2013-06-17 David Malcolm <dmalcolm@redhat.com>
6814 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
6815 ATTRIBUTE_UNUSED marking.
6817 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
6819 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
6820 alternative and update.
6821 (aarch64_dup_lanedi): Delete.
6822 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
6823 * config/aarch64/aarch64-simd-builtins.def: Update.
6825 2013-06-17 Richard Biener <rguenther@suse.de>
6827 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
6828 (lto_input_scc): Declare.
6829 (lto_input_tree_1): Likewise.
6830 (struct lto_stats_d): Add num_tree_bodies_output and
6831 num_pickle_refs_output.
6832 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
6833 (lto_read_tree_1): Split out from ...
6834 (lto_read_tree): ... this.
6835 (lto_input_scc): New function.
6836 (lto_input_tree_1): Split out from ...
6837 (lto_input_tree): ... this. Handle LTO_tree_scc.
6838 (lto_data_in_create): Create the streamer cache without hashes.
6839 * lto-streamer-out.c (create_output_block): Create the streamer
6840 cache with hashes when not doing WPA.
6841 (lto_write_tree_1): Split out from ...
6842 (lto_write_tree): ... this.
6843 (get_symbol_initial_value): New function.
6844 (lto_output_tree_1): Split out from ...
6845 (lto_output_tree): ... this. Write trees as series of SCCs
6846 using a DFS walk via DFS_write_tree.
6847 (struct sccs, struct scc_entry): New types.
6848 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
6849 (DFS_write_tree_body): New function.
6850 (DFS_write_tree): Likewise.
6851 (hash_tree): Likewise.
6852 (scc_entry_compare): Likewise.
6853 (hash_scc): Likewise.
6854 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
6855 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
6856 TREE_CHAIN as regular reference.
6857 (streamer_read_integer_cst): Remove.
6858 (streamer_get_pickled_tree): Adjust.
6859 * tree-streamer-out.c (streamer_write_chain): Disable streaming
6860 of DECL_EXTERNALs in BLOCK_VARS for now.
6861 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
6863 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
6864 Add hash value argument and record that if hashes are recorded
6866 (streamer_tree_cache_insert_1): Adjust.
6867 (streamer_tree_cache_insert): Likewise.
6868 (streamer_tree_cache_insert_at): Rename to ...
6869 (streamer_tree_cache_replace_tree): ... this and adjust.
6870 (streamer_tree_cache_append): Adjust.
6871 (record_common_node): Likewise.
6872 (streamer_tree_cache_create): Add argument whether to
6873 record hash values together with trees.
6874 (streamer_tree_cache_delete): Adjust.
6875 * tree-streamer.h (struct streamer_tree_cache_d): Add
6877 (streamer_read_integer_cst): Remove.
6878 (streamer_tree_cache_insert): Adjust.
6879 (streamer_tree_cache_append): Likewise.
6880 (streamer_tree_cache_insert_at): Rename to ...
6881 (streamer_tree_cache_replace_tree): ... this and adjust.
6882 (streamer_tree_cache_create): Add argument whether to record hashes.
6883 (streamer_tree_cache_get): Rename to ...
6884 (streamer_tree_cache_get_tree): ... this.
6885 (streamer_tree_cache_get_hash): New function.
6886 * tree.c (cache_integer_cst): New function.
6887 * tree.h (cache_integer_cst): Declare.
6888 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
6889 * lto-symtab.c (lto_varpool_replace_node): Only release
6890 DECL_INITIAL of non-prevailing decls.
6891 * varpool.c (varpool_remove_initializer): Do not release
6892 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
6894 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6896 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
6897 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
6898 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
6899 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
6900 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
6901 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
6902 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
6903 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
6904 instead of TARGET_64BIT.
6905 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
6906 Require ISA_HAS_<D>DIV.
6908 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
6910 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
6911 (mips*-*-linux*): Move default with_llsc setting to where other
6913 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
6915 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
6916 Likewise. Remove default with_tune setting. Move default float
6917 setting to its own block. Handle with_llsc in the same block as above.
6919 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
6921 PR rtl-optimization/57425
6922 PR rtl-optimization/57569
6923 * alias.c (write_dependence_p): Add new parameters mem_mode,
6924 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6925 Changed all callers.
6926 (canon_anti_dependence): New function.
6927 * cse.c (check_dependence): Use canon_anti_dependence.
6928 * cselib.c (cselib_invalidate_mem): Likewise.
6929 * rtl.h (canon_anti_dependence): Declare.
6931 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6933 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6934 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6935 ".set mips3" for 64-bit targets.
6937 2013-06-15 Dehao Chen <dehao@google.com>
6939 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6940 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6941 (gimple_check_call_args): Likewise.
6942 * value-prof.c (check_ic_target): Likewise.
6943 * ipa-inline.c (early_inliner): Likewise.
6944 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6945 * cgraph.c (cgraph_create_edge_1): Likewise.
6946 (cgraph_make_edge_direct): Likewise.
6948 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6951 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6952 rs6000_output_move_128bit to handle emitting quad memory
6953 operations. Set attribute length to 8 bytes.
6955 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6957 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6959 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6960 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6961 (aarch64_<su>mlsl<mode>): Likewise.
6963 2013-06-14 Mike Stump <mikestump@comcast.net>
6965 * Makefile.in (TARGET_H): Add insn-codes.h.
6967 2013-06-14 Alan Modra <amodra@gmail.com>
6971 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6972 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6973 bitfield expansion when EXPAND_MEMORY.
6974 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6976 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6978 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6979 test for clearing quad memory on 32-bit later.
6981 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6983 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6984 (fold_negate_expr): Likewise.
6985 (fold_real_zero_addition_p): Handle vectors.
6986 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6988 2013-06-14 Alan Modra <amodra@gmail.com>
6990 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6992 2013-06-13 Jan Hubicka <jh@suse.cz>
6994 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6995 Local comdats are not externally visible.
6996 * symtab.c (dump_symtab_base): Dump externally visible.
6997 (verify_symtab_base): Verify back links in the symtab hash.
6999 2013-06-13 Bin Cheng <bin.cheng@arm.com>
7001 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
7002 CONVERT_EXPR as equal nodes.
7004 2013-06-13 Bin Cheng <bin.cheng@arm.com>
7006 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
7008 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
7010 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
7011 Generalize to complex and vector.
7012 * tree.c (build_all_ones_cst): New function.
7013 * tree.h (build_all_ones_cst): Declare it.
7015 2013-06-13 Alan Modra <amodra@gmail.com>
7017 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
7018 * config/rs6000/rs6000.md (signbittf2): New insn.
7019 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
7020 (abstf2_internal, cmptf_internal2): Likewise.
7021 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
7023 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7024 Pat Haugen <pthaugen@us.ibm.com>
7025 Peter Bergner <bergner@vnet.ibm.com>
7027 * config/rs6000/rs6000.c (emit_load_locked): Add support for
7028 power8 byte, half-word, and quad-word atomic instructions.
7029 (emit_store_conditional): Likewise.
7030 (rs6000_expand_atomic_compare_and_swap): Likewise.
7031 (rs6000_expand_atomic_op): Likewise.
7033 * config/rs6000/sync.md (larx): Add new modes for power8.
7035 (AINT): New mode iterator to include TImode as well as normal
7036 integer modes on power8.
7037 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
7038 that VSX registers are not considered. Use AINT mode iterator
7039 instead of INT1 to allow inclusion of quad word atomic operations
7041 (load_locked<mode>): Likewise.
7042 (store_conditional<mode>): Likewise.
7043 (atomic_compare_and_swap<mode>): Likewise.
7044 (atomic_exchange<mode>): Likewise.
7045 (atomic_nand<mode>): Likewise.
7046 (atomic_fetch_<fetchop_name><mode>): Likewise.
7047 (atomic_nand_fetch<mode>): Likewise.
7048 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
7050 (ATOMIC): On power8, add QImode, HImode modes.
7051 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
7052 modes that promote to SImode.
7053 (load_lockedti): Convert TImode arguments to PTImode, so that we
7054 get a guaranteed even/odd register pair.
7055 (load_lockedpti): Likewise.
7056 (store_conditionalti): Likewise.
7057 (store_conditionalpti): Likewise.
7059 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
7060 atomic load/store instructions.
7063 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
7065 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
7067 (insn_count): New attribute, with most cases extracted from...
7068 (length): ...here. Redefine most cases in terms of insn_count.
7069 (single_insn): Delete.
7070 (can_delay): Use insn_count to check for single instructions.
7071 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
7072 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
7073 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
7074 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
7075 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
7076 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
7077 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
7078 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
7079 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
7080 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
7081 rather than "length".
7082 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
7083 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
7084 Use "insn_count" rather than "length".
7085 * config/mips/mips-dsp.md
7086 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
7087 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
7090 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
7092 PR tree-optimization/57361
7093 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
7095 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
7097 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
7099 (aarch64_simd_combine<mode>): New instruction expansion.
7100 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
7102 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
7103 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
7105 2013-06-12 Jan Hubicka <jh@suse.cz>
7107 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
7108 decl has when in streaming stage.
7109 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
7110 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
7112 2013-06-12 Roland Stigge <stigge@antcom.de>
7115 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
7117 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7119 PR tree-optimization/57537
7120 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
7121 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
7123 2013-06-12 Richard Biener <rguenther@suse.de>
7125 * data-streamer.h (streamer_write_char_stream): CSE
7126 obs->current_pointer.
7127 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
7128 streamer_write_char_stream manually and optimize the resulting loop.
7129 (streamer_write_hwi_stream): Likewise.
7131 2013-06-12 Jan Hubicka <jh@suse.cz>
7133 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
7134 * cgraph.h (varpool_create_empty_node): Declare.
7135 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
7137 * symtab.c (symtab_unregister_node): Be lax about missin entries
7139 (symtab_get_node): Update comment.
7140 * varpool.c (varpool_create_empty_node): Break out from ...
7141 (varpool_node_for_decl): ... here.
7142 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
7144 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
7146 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
7147 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
7148 part. Use straight-line flow at the end.
7149 <COMPONENT_REF>: Remove superfluous else.
7150 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
7152 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7155 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
7156 target hook even for !TREE_PUBLIC decls. If no resolution info
7157 is available, return false for common and external decls.
7159 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7161 * config/rl78/constraints.md (U): New constraint.
7162 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
7165 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7168 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
7169 to allow returning address to AT_PLATFORM name.
7171 2013-06-11 Jan Hubicka <jh@suse.cz>
7173 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
7174 * cgraph.h (symtab_node_base): Add weakref flag.
7175 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
7176 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
7177 (output_weakrefs): Use weakref flag.
7178 * fold-const.c (simple_operand_p): Handle WEAK.
7179 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
7180 * ipa.c (varpool_externally_visible_p): Drop weakref.
7181 (function_and_variable_visibility): Update comment; fix weakref
7182 sanity checks; do not clear DECL_WEAK on them.
7183 * lto-cgraph.c (lto_output_node): update.
7184 (lto_output_varpool_node): Update.
7185 (input_overwrite_node): Update.
7186 (input_node): Update.
7187 (input_varpool_node): Update.
7188 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
7189 (lto_symtab_merge_symbols): Add sanity check.
7190 (lto_symtab_prevailing_decl): Do not special case weakrefs.
7191 * passes.c (rest_of_decl_compilation): Set static flag, too.
7192 * symtab.c (dump_symtab_base): Dump weakref.
7193 (verify_symtab_base): Sanity check weakrefs.
7194 (symtab_make_decl_local): Remove duplicated code.
7195 (symtab_alias_ultimate_target): Simplify.
7196 * varpool.c (varpool_create_variable_alias): Set weakref flag.
7198 2013-06-11 DJ Delorie <dj@redhat.com>
7200 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
7201 (rl78_unwind_word_mode): New.
7203 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7205 * final.c (debug_prefix_maps): Make static.
7207 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7209 * function.c (initial_trampoline): Remove stray copy.
7211 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
7213 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
7215 2013-06-11 Martin Jambor <mjambor@suse.cz>
7217 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
7218 within bounds at the beginning of the function.
7220 2013-06-11 Alan Modra <amodra@gmail.com>
7222 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
7224 (get_named_section): Don't NULL !DECL_P decl.
7226 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
7228 * doc/invoke.texi (core-avx2): Document.
7230 (atom): Updated with MOVBE.
7232 2013-06-11 Richard Biener <rguenther@suse.de>
7234 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
7236 2013-06-11 Anton Blanchard <anton@samba.org>
7238 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
7239 correct shift value in little-endian mode.
7241 2013-06-11 Jakub Jelinek <jakub@redhat.com>
7244 * varasm.c (get_variable_align): Move #endif to the right place.
7246 2013-06-10 Cary Coutant <ccoutant@google.com>
7248 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
7249 for hash so that hash table traversal order is deterministic.
7251 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
7252 Pat Haugen <pthaugen@us.ibm.com>
7253 Peter Bergner <bergner@vnet.ibm.com>
7255 * config/rs6000/vector.md (GPR move splitter): Do not split moves
7256 of vectors in GPRS if they are direct moves or quad word load or
7259 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
7261 (direct_move_p): Likewise.
7262 (quad_load_store_p): Likewise.
7264 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
7265 classes into bins based on the physical register type.
7266 (reg_class_to_reg_type): Likewise.
7267 (IS_STD_REG_TYPE): Likewise.
7268 (IS_FP_VECT_REG_TYPE): Likewise.
7269 (reload_fpr_gpr): Arrays to determine what insn to use if we can
7270 use direct move instructions.
7271 (reload_gpr_vsx): Likewise.
7272 (reload_vsx_gpr): Likewise.
7273 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
7274 information that is a simplification of register classes. Also
7275 precalculate direct move reload helpers.
7276 (direct_move_p): New function to return true if the operation can
7277 be done as a direct move instruciton.
7278 (quad_load_store_p): New function to return true if the operation
7279 is a quad memory operation.
7280 (rs6000_legitimize_address): If quad memory, only allow register
7281 indirect for TImode addresses.
7282 (rs6000_legitimate_address_p): Likewise.
7283 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
7284 (rs6000_reload_register_type): Likewise.
7285 (register_to_reg_type): Return register type.
7286 (rs6000_secondary_reload_simple_move): New helper function for
7287 secondary reload and secondary memory needed to identify anything
7288 that is a simple move, and does not need reloading.
7289 (rs6000_secondary_reload_direct_move): New helper function for
7290 secondary reload to identify cases that can be done with several
7291 instructions via the direct move instructions.
7292 (rs6000_secondary_reload_move): New helper function for secondary
7293 reload to identify moves between register types that can be done.
7294 (rs6000_secondary_reload): Add support for quad memory operations
7295 and for direct move.
7296 (rs6000_secondary_memory_needed): Likewise.
7297 (rs6000_debug_secondary_memory_needed): Change argument names.
7298 (rs6000_output_move_128bit): New function to return the move to
7299 use for 128-bit moves, including knowing about the various
7300 limitations of quad memory operations.
7302 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
7303 memory operations. call rs6000_output_move_128bit for the actual
7304 instruciton(s) to generate.
7305 (vsx_movti_64bit): Likewise.
7307 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
7308 (UNSPEC_P8V_MTVSRWZ): Likewise.
7309 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
7310 (UNSPEC_P8V_MTVSRD): Likewise.
7311 (UNSPEC_P8V_XXPERMDI): Likewise.
7312 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
7313 (UNSPEC_FUSION_GPR): Likewise.
7314 (FMOVE128_GPR): New iterator for direct move.
7315 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
7318 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
7319 loads and direct move instructions.
7320 (zero_extendsidi2_lfiwzx): Likewise.
7321 (extendsidi2_lfiwax): Likewise.
7322 (extendsidi2_nocell): Likewise.
7323 (floatsi<mode>2_lfiwax): Likewise.
7325 (floatunssi<mode>2_lfiwzx): Likewise.
7327 (fix_trunc<mode>_stfiwx): Likewise.
7328 (fixuns_trunc<mode>_stfiwx): Likewise.
7329 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
7330 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
7331 (parity<mode>2_cmpb): Set length/type attr.
7332 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
7333 for 'mr.' to fast_compare.
7334 (bpermd_<mode>): Change type attr to popcnt.
7335 (p8_fmrgow_<mode>): New insns for power8 direct move support.
7336 (p8_mtvsrwz_1): Likewise.
7337 (p8_mtvsrwz_2): Likewise.
7338 (reload_fpr_from_gpr<mode>): Likewise.
7339 (p8_mtvsrd_1): Likewise.
7340 (p8_mtvsrd_2): Likewise.
7341 (p8_xxpermdi_<mode>): Likewise.
7342 (reload_vsx_from_gpr<mode>): Likewise.
7343 (reload_vsx_from_gprsf): Likewise.
7344 (p8_mfvsrd_3_<mode>): LIkewise.
7345 (reload_gpr_from_vsx<mode>): Likewise.
7346 (reload_gpr_from_vsxsf): Likewise.
7347 (p8_mfvsrd_4_disf): Likewise.
7348 (multi-word GPR splits): Do not split direct moves or quad memory
7351 2013-06-10 David Malcolm <dmalcolm@redhat.com>
7353 * tree-into-ssa.c (interesting_blocks): Make static.
7355 2013-06-10 Jakub Jelinek <jakub@redhat.com>
7358 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
7359 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
7360 Use DATA_ABI_ALIGNMENT for that case instead if defined.
7361 (get_variable_align): New function.
7362 (get_variable_section, emit_bss, emit_common,
7363 assemble_variable_contents, place_block_symbol): Use
7364 get_variable_align instead of DECL_ALIGN.
7365 (assemble_noswitch_variable): Add align argument, use it
7366 instead of DECL_ALIGN.
7367 (assemble_variable): Adjust caller. Use get_variable_align
7368 instead of DECL_ALIGN.
7369 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
7371 (DATA_ABI_ALIGNMENT): Define.
7372 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
7373 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
7374 opt is false, only return the psABI mandated alignment increase.
7375 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
7376 (DATA_ABI_ALIGNMENT): ... this.
7377 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
7378 (DATA_ABI_ALIGNMENT): ... this.
7379 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
7380 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
7381 (DATA_ABI_ALIGNMENT): ... this.
7382 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
7383 * doc/tm.texi: Regenerated.
7385 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
7387 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
7388 cmp_code to construct REG_EQUAL note.
7390 2013-06-09 Jakub Jelinek <jakub@redhat.com>
7393 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
7394 that operands[2] doesn't overlap with operands[0].
7396 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
7397 Jan Hubicka <jh@suse.cz>
7399 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
7400 hack to mark symbols as used.
7402 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
7404 PR rtl-optimization/57559
7405 * lra-constraints.c (process_alt_operands): Don't discourage
7406 memory with known offset for offsetable memory constraint.
7407 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
7409 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
7411 * varasm.c (struct oc_local_state): Reorder fields.
7412 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
7413 and adjust accordingly.
7414 (output_constructor): Reorder initialization code and adjust call to
7415 output_constructor_bitfield.
7417 2013-06-07 Jan Hubicka <jh@suse.cz>
7419 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
7421 2013-06-07 David Malcolm <dmalcolm@redhat.com>
7423 * tree-object-size.c (unknown): Make const.
7425 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7427 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
7428 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
7429 for last alternative in the cpu_facility attribute.
7431 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7434 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
7435 (xordi3): Change operand 2 constraint to arm_xordi_operand.
7436 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
7437 * config/arm/constraints.md (Dg): New constraint.
7438 * config/arm/neon.md (xordi3_neon): Remove.
7439 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
7440 * config/arm/predicates.md (arm_xordi_operand): New predicate.
7442 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7444 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
7445 Clean up alternatives.
7447 2013-06-07 Alan Modra <amodra@gmail.com>
7449 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
7452 2013-06-07 Alan Modra <amodra@gmail.com>
7454 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
7456 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7458 * config/arm/constraints.md (Df): New constraint.
7459 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
7460 Correct length attribute for last two alternatives.
7462 2013-06-07 Alan Modra <amodra@gmail.com>
7464 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7465 override user -mfp-in-toc.
7466 (offsettable_ok_by_alignment): Consider just the current access
7467 rather than the whole object, unless BLKmode. Handle
7468 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
7469 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
7470 for -mcmodel=medium.
7471 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
7472 override user -mfp-in-toc or -msum-in-toc. Default to
7473 -mno-fp-in-toc for -mcmodel=medium.
7475 2013-06-06 DJ Delorie <dj@redhat.com>
7477 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
7478 TARGET_VALID_POINTER_MODE.
7480 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
7481 Pat Haugen <pthaugen@us.ibm.com>
7482 Peter Bergner <bergner@vnet.ibm.com>
7484 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7485 Document new power8 builtins.
7487 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
7488 condition code register, to allow 128-bit logical operations to be
7489 done in the VSX or GPR registers.
7490 (nor<mode>3): Use the canonical form for nor.
7491 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
7492 vclz*, and vpopcnt* vector instructions.
7493 (nand<mode>3): Likewise.
7494 (orc<mode>3): Likewise.
7495 (clz<mode>2): LIkewise.
7496 (popcount<mode>2): Likewise.
7498 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
7499 that only the GPRs are recognized.
7501 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7502 support for new power8 builtins.
7504 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
7506 (xscvdpspn): Likewise.
7512 (vpopcnt): Likewise.
7513 (vpopcntb): Likewise.
7514 (vpopcnth): Likewise.
7515 (vpopcntw): Likewise.
7516 (vpopcntd): Likewise.
7521 (eqv_v16qi3): Likewise.
7522 (eqv_v8hi3): Likewise.
7523 (eqv_v4si3): Likewise.
7524 (eqv_v2di3): Likewise.
7525 (eqv_v4sf3): Likewise.
7526 (eqv_v2df3): Likewise.
7528 (nand_v16qi3): Likewise.
7529 (nand_v8hi3): Likewise.
7530 (nand_v4si3): Likewise.
7531 (nand_v2di3): Likewise.
7532 (nand_v4sf3): Likewise.
7533 (nand_v2df3): Likewise.
7535 (orc_v16qi3): Likewise.
7536 (orc_v8hi3): Likewise.
7537 (orc_v4si3): Likewise.
7538 (orc_v2di3): Likewise.
7539 (orc_v4sf3): Likewise.
7540 (orc_v2df3): Likewise.
7542 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
7543 allow power8 quad mode in 64-bit.
7544 (rs6000_builtin_vectorized_function): Add support to vectorize
7545 ISA 2.07 count leading zeros, population count builtins.
7546 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
7547 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
7548 (builtin_function_type): Add vgbbd builtin function which takes an
7550 (altivec_expand_vec_perm_const): Add support for new power8 merge
7553 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
7554 that does not include TImdoe for use with 32-bit.
7555 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
7557 (UNSPEC_VSX_CVDPSPN): Likewise.
7558 (vsx_xscvdpspn): Likewise.
7559 (vsx_xscvspdpn): Likewise.
7560 (vsx_xscvdpspn_scalar): Likewise.
7561 (vsx_xscvspdpn_directmove): Likewise.
7562 (vsx_and<mode>3): Split logical operations into 32-bit and
7563 64-bit. Add support to do logical operations on TImode as well as
7564 VSX vector types. Allow logical operations to be done in either
7565 VSX registers or in general purpose registers in 64-bit mode. Add
7566 splitters if GPRs were used. For AND, add clobber of CCmode to
7567 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
7569 (vsx_and<mode>3_32bit): Likewise.
7570 (vsx_and<mode>3_64bit): Likewise.
7571 (vsx_ior<mode>3): Likewise.
7572 (vsx_ior<mode>3_32bit): Likewise.
7573 (vsx_ior<mode>3_64bit): Likewise.
7574 (vsx_xor<mode>3): Likewise.
7575 (vsx_xor<mode>3_32bit): Likewise.
7576 (vsx_xor<mode>3_64bit): Likewise.
7577 (vsx_one_cmpl<mode>2): Likewise.
7578 (vsx_one_cmpl<mode>2_32bit): Likewise.
7579 (vsx_one_cmpl<mode>2_64bit): Likewise.
7580 (vsx_nor<mode>3): Likewise.
7581 (vsx_nor<mode>3_32bit): Likewise.
7582 (vsx_nor<mode>3_64bit): Likewise.
7583 (vsx_andc<mode>3): Likewise.
7584 (vsx_andc<mode>3_32bit): Likewise.
7585 (vsx_andc<mode>3_64bit): Likewise.
7586 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
7587 and xxlorc instructions.
7588 (vsx_eqv<mode>3_64bit): Likewise.
7589 (vsx_nand<mode>3_32bit): Likewise.
7590 (vsx_nand<mode>3_64bit): Likewise.
7591 (vsx_orc<mode>3_32bit): Likewise.
7592 (vsx_orc<mode>3_64bit): Likewise.
7594 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
7596 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
7598 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
7599 (p8_vmrgow): Likewise.
7600 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
7601 GPRs to be split under VSX.
7602 (p8v_clz<mode>2): Add power8 count leading zero support.
7603 (p8v_popcount<mode>2): Add power8 population count support.
7604 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
7607 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
7610 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
7612 (vec_nand): Likewise.
7613 (vec_vclz): Likewise.
7614 (vec_vclzb): Likewise.
7615 (vec_vclzd): Likewise.
7616 (vec_vclzh): Likewise.
7617 (vec_vclzw): Likewise.
7618 (vec_vgbbd): Likewise.
7619 (vec_vmrgew): Likewise.
7620 (vec_vmrgow): Likewise.
7621 (vec_vpopcnt): Likewise.
7622 (vec_vpopcntb): Likewise.
7623 (vec_vpopcntd): Likewise.
7624 (vec_vpopcnth): Likewise.
7625 (vec_vpopcntw): Likewise.
7627 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7629 PR rtl-optimization/57468
7630 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
7633 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7635 PR rtl-optimization/57459
7636 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
7637 type when setting live regs.
7639 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7641 * config/s390/s390.opt (mlra): New option.
7642 * config/s390/s390.c (s390_decompose_address): Check displacement
7643 for all registers for LRA.
7644 (s390_secondary_reload): Don't used secondary reloads for LRA.
7645 (s390_lra_p): New function.
7646 (TARGET_LRA_P): Define.
7647 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
7648 of attribute cpu_facility to zarch for the last alternative.
7649 (*cmpmem_short): Ditto.
7651 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
7653 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
7654 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
7655 (arm_expand_prologue): Likewise.
7657 2013-06-06 Teresa Johnson <tejohnson@google.com>
7660 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
7661 as this is now done by redirect_edge_and_branch_force.
7662 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
7663 barriers, and fix interaction with splitting.
7664 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
7665 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
7666 reflect changes made in the routine.
7667 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
7668 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
7669 since this is called in cfglayout mode, and replace partition fixup
7670 with assert as that is now done by force_nonfallthru_and_redirect.
7671 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
7672 already be marked with region crossing note.
7673 (insert_section_boundary_note): Make non-static, gate on flag
7674 has_bb_partition, rewrite to also check for multiple partitions.
7675 (rest_of_handle_reorder_blocks): Remove call to
7676 insert_section_boundary_note, now done later during free_cfg.
7677 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
7678 * bb-reorder.h (insert_section_boundary_note): Declare.
7679 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
7680 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
7681 invoke insert_section_boundary_note.
7682 (try_redirect_by_replacing_jump): Remove unnecessary
7683 check for region crossing note.
7684 (fixup_partition_crossing): New function.
7685 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
7686 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
7687 in non-cfglayout mode.
7688 (force_nonfallthru_and_redirect): Fixup partition boundaries,
7689 remove old code that tried to do this. Emit barrier correctly
7690 when we are in cfglayout mode.
7691 (last_bb_in_partition): New function.
7692 (rtl_split_edge): Correctly fixup partition boundaries.
7693 (commit_one_edge_insertion): Remove old code that tried to
7694 fixup region crossing edge since this is now handled in
7695 split_block, and set up insertion point correctly since
7696 block may now end in a jump.
7697 (verify_hot_cold_block_grouping): Guard against checking when not in
7698 linearized RTL mode.
7699 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
7701 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
7702 rtl_verify_flow_info, so not called in cfglayout mode.
7703 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
7704 (fixup_reorder_chain): Remove old code that attempted to fixup region
7705 crossing note as this is now handled in force_nonfallthru_and_redirect.
7706 (duplicate_insn_chain): Don't duplicate switch section notes.
7707 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
7709 * basic-block.h (emit_barrier_after_bb): Declare.
7711 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7713 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
7714 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
7715 arm_usatsihi): Adjust alternatives for arm_restrict_it.
7717 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7719 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
7721 * config/arm/ldmstm.md: Regenerate.
7723 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7725 * config/arm/sync.md (atomic_loaddi_1):
7726 Disable predication for arm_restrict_it.
7727 (arm_load_exclusive<mode>): Likewise.
7728 (arm_load_exclusivesi): Likewise.
7729 (arm_load_exclusivedi): Likewise.
7730 (arm_load_acquire_exclusive<mode>): Likewise.
7731 (arm_load_acquire_exclusivesi): Likewise.
7732 (arm_load_acquire_exclusivedi): Likewise.
7733 (arm_store_exclusive<mode>): Likewise.
7734 (arm_store_exclusive<mode>): Likewise.
7735 (arm_store_release_exclusivedi): Likewise.
7736 (arm_store_release_exclusive<mode>): Likewise.
7738 2013-06-06 Richard Biener <rguenther@suse.de>
7740 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
7742 (lto_tag_is_tree_code_p): Adjust.
7743 (lto_tag_is_gimple_code_p): Likewise.
7744 (lto_gimple_code_to_tag): Likewise.
7745 (lto_tag_to_gimple_code): Likewise.
7746 (lto_tree_code_to_tag): Likewise.
7747 (lto_tag_to_tree_code): Likewise.
7748 * data-streamer.h (streamer_write_hwi_in_range): Use
7749 uhwi streaming to stream the normalized range.
7750 (streamer_read_hwi_in_range): Likewise.
7752 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7754 * config/arm/arm.md (enabled_for_depr_it): New attribute.
7755 (predicable_short_it): Likewise.
7756 (predicated): Likewise.
7757 (enabled): Handle above.
7758 (define_cond_exec): Set predicated attribute to yes.
7760 2013-06-05 Mike Stump <mikestump@comcast.net>
7762 * gdbinit.in (__FUNCTION__): Add.
7764 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
7766 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
7767 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
7769 2013-06-05 Jan Hubicka <jh@suse.cz>
7771 * varasm.c (mark_decl_referenced): Revert the removal until targets
7774 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
7776 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
7777 instead of mark_decl_referenced.
7779 2013-06-05 Jan Hubicka <jh@suse.cz>
7781 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
7782 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
7783 and symtab_used_from_object_file_p.
7784 (cgraph_make_node_local_1): Clear forced_by_abi.
7785 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
7786 * cgraph.h (symtab_node_base): Add forced_by_abi.
7787 (decide_is_variable_needed): Remove.
7788 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
7789 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
7790 (decide_is_symbol_needed): ... this one; handle symbols in general;
7791 always analyze virtuals; honnor forced_by_abi.
7792 (cgraph_finalize_function): Update.
7793 (varpool_finalize_decl): Update.
7794 (symbol_defined_and_needed): Remove.
7795 (analyze_functions): Update.
7796 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
7797 output_refs, input_overwrite_node): Handle forced_by_abi.
7798 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
7799 (address_taken_from_non_vtable_p): ... this one.
7800 (comdat_can_be_unshared_p_1): New function.
7801 (cgraph_comdat_can_be_unshared_p): Rename to ...
7802 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
7803 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
7804 (function_and_variable_visibility): Clear forced_by_abi as needed.
7805 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
7806 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
7807 * symtab.c (dump_symtab_base): Dump forced_by_abi.
7808 * varpool.c (decide_is_variable_needed): Remove.
7810 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7812 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
7813 (arm_option_override): Override arm_restrict_it where appropriate.
7814 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
7815 * config/arm/arm.opt (mrestrict-it): New command-line option.
7816 * doc/invoke.texi: Document -mrestrict-it.
7818 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7820 * tsan.c (tsan_atomic_table): Make const.
7822 2013-06-05 Richard Biener <rguenther@suse.de>
7824 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
7825 index associated with the tree we are supposed to replace.
7826 * tree-streamer-out.c (pack_ts_base_value_fields): Output
7827 TREE_ASM_WRITTEN as zero for everything but SSA names.
7829 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7831 * tree-ssa-structalias.c (call_stmt_vars): Make static.
7833 2013-06-04 Jan Hubicka <jh@suse.cz>
7835 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
7836 (input_node, input_varpool_node): Handle correctly external same
7838 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
7839 nodes at ltrans stage.
7841 2013-06-04 Jan Hubicka <jh@suse.cz>
7843 * ipa-inline.c (update_caller_keys): Fix availability test.
7844 (update_callee_keys): Likewise.
7845 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
7846 to follow ELF standard.
7848 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
7850 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
7851 (mips64r5900el-*-elf*): New configurations.
7852 * config/mips/mips-cpus.def (r5900): New processor.
7853 * config/mips/mips-tables.opt: Regenerate.
7854 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
7855 (mips_issue_rate): Handle PROCESSOR_R5900.
7856 (mips_reorg_process_insns): Force reorder mode for the R5900.
7857 * config/mips/mips.h (TARGET_MIPS5900): Define.
7858 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
7860 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
7862 * config/mips/mips.md (processor): Add r5900.
7863 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
7865 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7867 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
7868 into function to generate MOVI instruction.
7869 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
7870 (aarch64_preferred_simd_mode): Turn into wrapper.
7871 (aarch64_output_scalar_simd_mov_immediate): New function.
7872 * config/aarch64/aarch64-protos.h: Add prototype for above.
7874 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7876 * config/aarch64/aarch64.c (simd_immediate_info): Remove
7877 element_char member.
7878 (sizetochar): Return signed char.
7879 (aarch64_simd_valid_immediate): Remove elchar and other
7880 unnecessary variables.
7881 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
7882 Calculate element_char as required.
7883 * config/aarch64/aarch64-protos.h: Update and move prototype
7884 for aarch64_output_simd_mov_immediate.
7885 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
7888 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7890 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
7891 information completed by aarch64_simd_valid_immediate.
7892 (aarch64_legitimate_constant_p): Update arguments.
7893 (aarch64_simd_valid_immediate): Work with struct rather than many
7895 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
7896 (aarch64_simd_make_constant): Update arguments.
7897 (aarch64_output_simd_mov_immediate): Work with struct rather than
7898 many pointers. Output immediate directly rather than as operand.
7899 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7901 * config/aarch64/constraints.md (Dn): Update arguments.
7903 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7905 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
7907 (aarch64_simd_immediate_valid_for_move): Remove.
7908 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
7909 (aarch64_simd_make_constant): Update call.
7910 (aarch64_output_simd_mov_immediate): Update call.
7911 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7913 * config/aarch64/constraints.md (Dn): Update call.
7915 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7917 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
7918 return type to bool for prototype.
7919 (aarch64_legitimate_constant_p): Check for true instead of not -1.
7920 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
7921 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
7923 2013-06-04 Catherine Moore <clm@codesourcery.com>
7925 * config/mips/mips.opt (meva): New.
7926 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7927 (ASM_SPEC): Handle -meva.
7928 * doc/invoke.texi (meva): Document.
7930 2013-06-04 Alan Modra <amodra@gmail.com>
7932 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7935 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7937 * rtl.def: Add extra fourth optional field to define_cond_exec.
7938 * gensupport.c (process_one_cond_exec): Process attributes from
7940 * doc/md.texi: Document fourth field in define_cond_exec.
7942 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7944 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7945 out the processing order as in store_bit_field_1.
7947 2013-06-04 Jan Hubicka <jh@suse.cz>
7950 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7951 non-VAR_DECL node if it does not exist yet.
7953 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7955 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7956 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7957 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7958 target_cpu_default setting.
7960 2013-06-03 Teresa Johnson <tejohnson@google.com>
7962 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7963 default to -fopt-info=optimized instead of all.
7964 * doc/invoke.texi: Ditto.
7965 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7966 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7967 (execute_vect_slp): Emit BB vectorization success under
7968 MSG_OPTIMIZED_LOCATIONS.
7969 * tree-vect-slp.c (vect_slp_transform_bb): Change
7970 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7971 * tree-vect-loop.c (vect_transform_loop): Ditto.
7973 2013-06-03 Jason Merrill <jason@redhat.com>
7976 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7977 Use TARGET_EXPR for C++.
7979 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7981 PR rtl-optimization/57268
7982 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7983 if DEBUG_INSN_P (insn).
7986 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7988 PR rtl-optimization/57268
7989 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7990 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7992 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7994 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7995 (ix86_avoid_lea_for_addr): Likewise.
7996 (exact_dependency_1): Likewise.
7997 (ix86_adjust_cost): Likewise.
7998 (swap_top_of_ready_list): Fix formatting and !reload_completed check
8000 (do_reorder_for_imul): Fix typo, formatting and
8001 !reload_completed check removed.
8002 (ix86_sched_reorder): Fix typo and formatting.
8003 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
8006 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
8008 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
8010 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
8012 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
8013 <STRING_CST>: Likewise.
8014 <VECTOR_CST>: Likewise.
8016 2013-06-01 Janus Weil <janus@gcc.gnu.org>
8017 Mikael Morin <mikael@gcc.gnu.org>
8019 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
8020 * config.in: Regenerated.
8021 * configure: Regenerated.
8023 2013-06-01 Jan Hubicka <jh@suse.cz>
8026 * cgraphunit.c (compile): When weakref is not supported,
8027 set up transparent aliases before final output pass.
8028 * varasm.c (assemble_alias): Do not try to do it here.
8030 2013-06-01 Jan Hubicka <jh@suse.cz>
8033 * passes.c (for_per_function): Skip unanalyzed functions.
8035 2013-06-01 Jan Hubicka <jh@suse.cz>
8037 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
8038 (lto_symtab_merge_symbols_1): ... this one.
8039 (lto_symtab_merge_cgraph_nodes): Rename to ...
8040 (lto_symtab_merge_symbols): ... this one; simplify.
8041 * cgraph.c (same_body_aliases_done): Rename to ...
8042 (cpp_implicit_aliases_done): ... this one.
8043 (cgraph_create_function_alias): Update.
8044 (cgraph_same_body_alias): Update.
8045 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
8046 (verify_edge_corresponds_to_fndecl): Simplify.
8047 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
8048 (cgraph_node): Remove same_body_alias.
8049 (varpool_node): Remove alias_of and extra_name_alias.
8050 (same_body_aliases_done): Rename to ..
8051 (cpp_implicit_aliases_done): ... this one.
8052 (symtab_alias_ultimate_target): Add default parameter.
8053 (symtab_resolve_alias): New function.
8054 (fixup_same_cpp_alias_visibility): Declare.
8055 (cgraph_function_node): Add default parameter.
8056 (cgraph_node_asm_name): Likewise.
8057 (cgraph_function_or_thunk_node): Add default parameter; do
8058 not ICE when it is NULL.
8059 (varpool_variable_node): Likewise.
8060 * tree-emutls.c (create_emultls_var): Update.
8061 (ipa_lower_emutls): Update.
8062 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
8063 (cgraph_reset_node): Reset alias info.
8064 (cgraph_finalize_function): Update.
8065 (fixup_same_cpp_alias_visibility): Move to symtab.c.
8066 (analyze_function): Simplify.
8067 (cgraph_process_same_body_aliases): Simplify.
8068 (analyze_functions): Fixup same body aliases.
8069 (handle_alias_pairs): Simplify.
8070 (assemble_thunk): Update.
8071 (assemble_thunks_and_aliases): Update.
8072 (output_weakrefs): Rewrite.
8073 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
8074 (lto_output_varpool_node): Likewise.
8075 (compute_ltrans_boundary): Remve assert.
8076 (get_alias_symbol): New functoin.
8077 (input_node): Rewrite alias handling.
8078 (input_varpool_node): Likewise.
8079 * ipa-pure-const.c (propagate_pure_const): Fix formating.
8080 * ipa.c (process_references): Handle weakrefs correctly.
8081 (symtab_remove_unreachable_nodes): Likewise.
8082 * trans-mem.c (get_cg_data): Update.
8083 (ipa_tm_create_version_alias): Update.
8084 (ipa_tm_execute): Update.
8085 * symtab.c (dump_symtab_base): Dump aliases.
8086 (verify_symtab_base): Verify aliases.
8087 (symtab_node_availability): New function.
8088 (symtab_alias_ultimate_target): Simplify.
8089 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
8090 handle all the fixup cases.
8091 (symtab_resolve_alias): New function.
8092 * passes.c (ipa_write_summaries): Handle weakrefs.
8093 * varpool.c (varpool_analyze_node): Simplify.
8094 (assemble_aliases): Update.
8095 (varpool_create_variable_alias): Simplify.
8096 (varpool_extra_name_alias): Simplify.
8097 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
8098 (lto_symtab_merge_symbols): ... this one.
8100 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
8103 PR rtl-optimization/57268
8104 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8105 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8107 2013-06-01 Tobias Burnus <burnus@net-b.de>
8110 2013-05-31 Tobias Burnus <burnus@net-b.de>
8113 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
8116 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8118 PR rtl-optimization/57268
8119 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8120 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8122 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8124 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
8125 unordered comparison operators when -fno-trapping-math is in effect
8127 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
8128 and implement unordered comparison operators properly on the e500.
8130 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8132 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
8133 for constant scalar integers.
8134 (simplify_relational_operation_1): Likewise.
8136 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
8138 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
8139 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
8142 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8143 Igor Zamyatin <igor.zamyatin@intel.com>
8145 Silvermont (SLM) architecture performance tuning.
8146 * config/i386/i386.h (enum ix86_tune_indices): Add
8147 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
8148 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
8150 * config/i386/i386.c (initial_ix86_tune_features)
8151 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
8152 (ix86_lea_outperforms): Handle Silvermont tuning.
8153 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
8155 (ix86_use_lea_for_mov): Likewise.
8156 (ix86_avoid_lea_for_addr): Likewise.
8157 (ix86_lea_for_add_ok): Likewise.
8158 (exact_dependency_1): New function.
8159 (exact_store_load_dependency): Likewise.
8160 (ix86_adjust_cost): Handle Silvermont tuning.
8161 (do_reoder_for_imul): Likewise.
8162 (swap_top_of_ready_list): New function.
8163 (ix86_sched_reorder): Changed to handle Silvermont tuning.
8165 * config/i386/i386.md (peepholes that split memory operand in fp
8168 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8170 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
8171 Remove un-necessary braces.
8173 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8175 * config/aarch64/aarch64.c (aarch64_classify_symbol):
8176 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
8178 2013-05-31 Tobias Burnus <burnus@net-b.de>
8181 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
8183 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8186 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
8187 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
8188 * config/arm/neon.md (iordi3_neon): Remove.
8189 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
8190 * config/arm/predicates.md (imm_for_neon_logic_operand):
8191 Move to earlier in the file.
8192 (neon_logic_op2): Likewise.
8193 (arm_iordi_operand_neon): New predicate.
8195 2013-05-31 Richard Biener <rguenther@suse.de>
8197 PR tree-optimization/57478
8198 PR tree-optimization/57453
8199 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
8202 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8204 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
8205 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
8207 2013-05-30 Tobias Burnus <burnus@net-b.de>
8208 Thomas Koenig <tkoenig@gcc.gnu.org>
8211 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
8212 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
8214 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
8216 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
8218 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
8220 * target.def (register_usage_leveling_p): New hook.
8221 * targhooks.c (default_register_usage_leveling_p): New.
8222 * targhooks.h (default_register_usage_leveling_p): New prototype.
8223 * lra-assigns.c (register_usage_leveling_p): Use the hook.
8224 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
8225 * doc/tm.texi: Update.
8226 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
8228 2013-05-30 Ian Bolton <ian.bolton@arm.com>
8230 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
8231 (*insv_reg<mode>): New define_insn.
8233 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
8235 PR rtl-optimization/57439
8236 * postreload.c (move2add_valid_value_p): Check that we have
8237 a zero subreg_regno_offset when accessing the register in
8240 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8241 Igor Zamyatin <igor.zamyatin@intel.com>
8243 Silvermont (SLM) architecture pipeline model, tuning and
8245 * config.gcc: Add slm config options and target.
8247 * config/i386/slm.md: New.
8249 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
8251 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
8253 (ix86_target_macros_internal): Likewise.
8255 * gcc/config/i386/i386.c (slm_cost): New cost.
8256 (m_SLM): New macro flag.
8257 (initial_ix86_tune_features): Set m_SLM.
8258 (x86_accumulate_outgoing_args): Likewise.
8259 (x86_arch_always_fancy_math_387): Likewise.
8260 (processor_target_table): Add slm cost.
8261 (cpu_names): Add slm cpu name.
8262 (x86_option_override_internal): Set SLM ISA.
8263 (ix86_issue_rate): New case PROCESSOR_SLM.
8264 (ia32_multipass_dfa_lookahead): Likewise.
8265 (fold_builtin_cpu): Add slm.
8267 * config/i386/i386.h (TARGET_SLM): New target macro.
8268 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
8269 (processor_type): Add PROCESSOR_SLM.
8271 * config/i386/i386.md (cpu): Add new value "slm".
8272 (slm.md): Include slm.md.
8274 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
8275 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8277 * config/arm/arm-protos.h: Add and update function protos.
8278 * config/arm/arm.c (use_simple_return_p): New added.
8279 (thumb2_expand_return): Check simple_return flag.
8280 * config/arm/arm.md: Add simple_return and conditional simple_return.
8281 * config/arm/iterators.md: Add iterator for return and simple_return.
8283 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8285 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
8286 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
8287 (arm_emit_vfp_multi_reg_pop): Likewise.
8288 (thumb2_emit_ldrd_pop): Likewise.
8289 (arm_expand_epilogue): Add misc REG_CFA notes.
8290 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
8292 2013-05-29 Lawrence Crowl <crowl@google.com>
8294 * config/arm/t-arm: Update for below.
8296 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
8297 Change type to hash_table. Update dependent calls and types.
8299 * config/i386/t-cygming: Update for below.
8301 * config/i386/t-interix: Update for below.
8303 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
8304 Change type to hash_table. Update dependent calls and types.
8305 (i386_find_on_wrapper_list::wrappers): Likewise.
8307 * config/ia64/t-ia64: Update for below.
8309 * config/ia64/ia64.c (bundle_state_table):
8310 Change type to hash_table. Update dependent calls and types.
8312 * config/mips/mips.c (mips_reorg_process_insns::htab):
8313 Change type to hash_table. Update dependent calls and types.
8315 * config/sol2.c (solaris_comdat_htab):
8316 Change type to hash_table. Update dependent calls and types.
8318 * config/t-sol2: Update for above.
8320 2013-05-29 Teresa Johnson <tejohnson@google.com>
8322 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
8323 functions are not yet marked as defined.
8325 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
8326 Pat Haugen <pthaugen@us.ibm.com>
8327 Peter Bergner <bergner@vnet.ibm.com>
8329 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
8333 (vrotl<mode>3): Likewise.
8334 (vashl<mode>3): Likewise.
8335 (vlshr<mode>3): Likewise.
8336 (vashr<mode>3): Likewise.
8338 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8339 support for power8 V2DI builtins.
8341 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
8342 power8 V2DI builtins.
8343 (vupkhsw): Likewise.
8344 (vupklsw): Likewise.
8345 (vaddudm): Likewise.
8350 (vpkudum): Likewise.
8351 (vpksdss): Likewise.
8352 (vpkudus): Likewise.
8353 (vpksdus): Likewise.
8358 (vsubudm): Likewise.
8359 (vcmpequd): Likewise.
8360 (vcmpgtsd): Likewise.
8361 (vcmpgtud): Likewise.
8362 (vcmpequd_p): Likewise.
8363 (vcmpgtsd_p): Likewise.
8364 (vcmpgtud_p): Likewise.
8365 (vupkhsw): Likewise.
8366 (vupklsw): Likewise.
8367 (vaddudm): Likewise.
8372 (vpksdss): Likewise.
8373 (vpksdus): Likewise.
8374 (vpkudum): Likewise.
8375 (vpkudus): Likewise.
8380 (vsubudm): Likewise.
8382 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
8383 support for power8 V2DI instructions.
8385 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
8386 power8 V2DI instructions. Combine pack and unpack insns to use an
8387 iterator for each mode. Check whether a particular mode supports
8388 Altivec instructions instead of just checking TARGET_ALTIVEC.
8389 (UNSPEC_VPKUWUM): Likewise.
8390 (UNSPEC_VPKSHSS): Likewise.
8391 (UNSPEC_VPKSWSS): Likewise.
8392 (UNSPEC_VPKUHUS): Likewise.
8393 (UNSPEC_VPKSHUS): Likewise.
8394 (UNSPEC_VPKUWUS): Likewise.
8395 (UNSPEC_VPKSWUS): Likewise.
8396 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
8397 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
8398 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
8399 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
8400 (UNSPEC_VUPKHSB): Likewise.
8401 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
8402 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
8403 (UNSPEC_VUPKHSH): Likewise.
8404 (UNSPEC_VUPKLSB): Likewise.
8405 (UNSPEC_VUPKLSH): Likewise.
8407 (VI_char): Likewise.
8408 (VI_scalar): Likewise.
8409 (VI_unit): Likewise.
8411 (VP_small): Likewise.
8412 (VP_small_lc): Likewise.
8413 (VU_char): Likewise.
8414 (add<mode>3): Likewise.
8415 (altivec_vaddcuw): Likewise.
8416 (altivec_vaddu<VI_char>s): Likewise.
8417 (altivec_vadds<VI_char>s): Likewise.
8418 (sub<mode>3): Likewise.
8419 (altivec_vsubcuw): Likewise.
8420 (altivec_vsubu<VI_char>s): Likewise.
8421 (altivec_vsubs<VI_char>s): Likewise.
8422 (altivec_vavgs<VI_char>): Likewise.
8423 (altivec_vcmpbfp): Likewise.
8424 (altivec_eq<mode>): Likewise.
8425 (altivec_gt<mode>): Likewise.
8426 (altivec_gtu<mode>): Likewise.
8427 (umax<mode>3): Likewise.
8428 (smax<mode>3): Likewise.
8429 (umin<mode>3): Likewise.
8430 (smin<mode>3): Likewise.
8431 (altivec_vpkuhum): Likewise.
8432 (altivec_vpkuwum): Likewise.
8433 (altivec_vpkshss): Likewise.
8434 (altivec_vpkswss): Likewise.
8435 (altivec_vpkuhus): Likewise.
8436 (altivec_vpkshus): Likewise.
8437 (altivec_vpkuwus): Likewise.
8438 (altivec_vpkswus): Likewise.
8439 (altivec_vpks<VI_char>ss): Likewise.
8440 (altivec_vpks<VI_char>us): Likewise.
8441 (altivec_vpku<VI_char>us): Likewise.
8442 (altivec_vpku<VI_char>um): Likewise.
8443 (altivec_vrl<VI_char>): Likewise.
8444 (altivec_vsl<VI_char>): Likewise.
8445 (altivec_vsr<VI_char>): Likewise.
8446 (altivec_vsra<VI_char>): Likewise.
8447 (altivec_vsldoi_<mode>): Likewise.
8448 (altivec_vupkhsb): Likewise.
8449 (altivec_vupkhs<VU_char>): Likewise.
8450 (altivec_vupkls<VU_char>): Likewise.
8451 (altivec_vupkhsh): Likewise.
8452 (altivec_vupklsb): Likewise.
8453 (altivec_vupklsh): Likewise.
8454 (altivec_vcmpequ<VI_char>_p): Likewise.
8455 (altivec_vcmpgts<VI_char>_p): Likewise.
8456 (altivec_vcmpgtu<VI_char>_p): Likewise.
8457 (abs<mode>2): Likewise.
8458 (vec_unpacks_hi_v16qi): Likewise.
8459 (vec_unpacks_hi_v8hi): Likewise.
8460 (vec_unpacks_lo_v16qi): Likewise.
8461 (vec_unpacks_hi_<VP_small_lc>): Likewise.
8462 (vec_unpacks_lo_v8hi): Likewise.
8463 (vec_unpacks_lo_<VP_small_lc>): Likewise.
8464 (vec_pack_trunc_v8h): Likewise.
8465 (vec_pack_trunc_v4si): Likewise.
8466 (vec_pack_trunc_<mode>): Likewise.
8468 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
8470 (vec_vmaxsd): Likewise.
8471 (vec_vmaxud): Likewise.
8472 (vec_vminsd): Likewise.
8473 (vec_vminud): Likewise.
8474 (vec_vpksdss): Likewise.
8475 (vec_vpksdus): Likewise.
8476 (vec_vpkudum): Likewise.
8477 (vec_vpkudus): Likewise.
8478 (vec_vrld): Likewise.
8479 (vec_vsld): Likewise.
8480 (vec_vsrad): Likewise.
8481 (vec_vsrd): Likewise.
8482 (vec_vsubudm): Likewise.
8483 (vec_vupkhsw): Likewise.
8484 (vec_vupklsw): Likewise.
8486 2013-05-29 Jan Hubicka <jh@suse.cz>
8488 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
8489 flags; reorder rest of fields in more consistent way.
8490 (varpool_node): Remove analyzed, finalized and alias.
8491 (cgraph_ndoe): Likewise.
8492 (symtab_alias_ultimate_target): New function.
8493 (cgraph_function_node): Move offline.
8494 (cgraph_reset_node): Declare.
8495 (cgraph_comdat_can_be_unshared_p): Remove.
8496 (varpool_remove_initializer): Declare.
8497 (varpool_first_defined_variable, varpool_next_defined_variable
8498 cgraph_first_defined_function, cgraph_next_defined_function): Update.
8499 (cgraph_function_with_gimple_body_p): Update.
8500 (varpool_all_refs_explicit_p): Update.
8501 (symtab_alias_target): New function.
8502 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
8503 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
8504 (cgraph_function_or_thunk_node): Simplify using
8505 symtab_alias_ultimate_target.
8506 (varpool_variable_node): Likewise.
8507 * cgraph.c (cgraph_create_function_alias): Update.
8508 (cgraph_add_thunk): Update.
8509 (cgraph_remove_node): Update.
8510 (dump_cgraph_node): Do not dump removed flags.
8511 (cgraph_function_body_availability): Update.
8512 (cgraph_propagate_frequency): Update.
8513 (verify_cgraph_node): Check sanity of local flag.
8514 (cgraph_function_node): Move here from cgraph.h; revamp for
8515 cgraph_function_or_thunk_node.
8516 * lto-symtab.c (lto_varpool_replace_node): Update.
8517 (lto_symtab_resolve_can_prevail_p): Update.
8518 (lto_symtab_merge_cgraph_nodes): Update.
8519 * ipa-cp.c (determine_versionability, initialize_node_lattices,
8520 propagate_constants_accross_call, devirtualization_time_bonus,
8521 ipcp_propagate_stage): Update.
8522 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
8523 * ipa-inline-transform.c (clone_inlined_nodes,
8524 preserve_function_body_p): Update.
8525 * ipa-reference.c (propagate): Update.
8526 (write_node_summary_p): Update.
8527 * toplev.c (wrapup_global_declaration_2): Update.
8528 * cgraphunit.c (cgraph_analyze_function): Rename to ...
8529 (analyze_function) ... this one.
8530 (cgraph_process_new_functions): Update.
8531 (cgraph_reset_node): Export.
8532 (cgraph_finalize_function): Update.
8533 (cgraph_add_new_function): Update.
8534 (process_function_and_variable_attributes): Update.
8535 (varpool_finalize_decl): Update.
8536 (symbol_finalized): Remove.
8537 (symbol_finalized_and_needed): Rename to ...
8538 (symbol_defined_and_needed): ... update.
8539 (cgraph_analyze_functions): Update.
8540 (handle_alias_pairs): Update.
8541 (mark_functions_to_output): Update.
8542 (assemble_thunk): Update.
8543 (output_in_order): Update.
8544 (output_weakrefs): Update.
8545 (finalize_compilation_unit): Update.
8546 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
8547 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
8548 input_node, input_varpool_node): Update.
8549 * dbxout.c (dbxout_expand_expr): Update.
8550 * cgraphclones.c (cgraph_clone_node): Update.
8551 (cgraph_copy_node_for_versioning): Update.
8552 (cgraph_materialize_clone): Update.
8553 (cgraph_materialize_all_clones): Update.
8554 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
8555 propagate_pure_const, propagate_nothrow): Update.
8556 * lto-streamer-out.c (lto_output, write_symbol): Update.
8557 * ipa-utils.c (ipa_reverse_postorder): Update.
8558 * ipa-inline.c (can_inline_edge_p): Update.
8559 (update_caller_keys, ipa_inline): Update.
8560 * dwarf2out.c (reference_to_unused,
8561 premark_types_used_by_global_vars_helper): Update.
8562 * tree-eh.c (tree_could_trap_p): Update.
8563 * ipa-split.c (consider_split, execute_split_functions): Update.
8564 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
8565 has_addr_references_p): Update; move ahead in file for better
8567 (process_references): Simplify.
8568 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
8570 (cgraph_comdat_can_be_unshared_p): Make static.
8571 (cgraph_externally_visible_p): Update.
8572 (varpool_externally_visible_p): Update.
8573 (function_and_variable_visibility): Update.
8574 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
8575 ipa_tm_mark_force_output_node): Update.
8576 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
8577 estimate_edge_devirt_benefit, inline_generate_summary,
8578 inline_write_summary): Update.
8579 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
8580 * ipa-prop.c (ipa_compute_jump_functions): Update.
8581 (ipa_print_node_params, ipa_prop_read_section,
8582 ipa_update_after_lto_read, read_replacements_section): Update.
8583 * varasm.c (mark_decl_referenced): Update.
8584 (assemble_alias, dump_tm_clone_pairs): Update.
8585 * tree-inline.c (copy_bb): Update.
8586 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
8588 * symtab.c (dump_symtab_base): Print new flags.
8589 (verify_symtab_base): Verify new flags.
8590 (symtab_alias_ultimate_target): New function.
8591 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
8592 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
8594 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
8596 * i386.c (ix86_get_function_versions_dispatcher,
8597 ix86_generate_version_dispatcher_body): Update.
8598 (fold_builtin_cpu): Use varpool_add_new_variable.
8599 * varpool.c (varpool_remove_initializer): Break out from ...
8600 (varpool_remove_node): ... this one.
8601 (dump_varpool_node, varpool_node_for_asm,
8602 cgraph_variable_initializer_availability, varpool_analyze_node,
8603 varpool_assemble_decl, varpool_remove_unreferenced_decls,
8604 varpool_finalize_named_section_flags, varpool_create_variable_alias):
8607 2013-05-29 Jan Hubicka <jh@suse.cz>
8609 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
8611 2013-05-29 Easwaran Raman <eraman@google.com>
8613 PR tree-optimization/57442
8614 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
8615 when control exits the main loop.
8617 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
8619 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
8621 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
8622 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
8623 * rx/t-rx: Add rx100 under multi library matches option for nofpu
8626 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8628 PR tree-optimization/57441
8629 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
8630 Don't limit size of incr_vec to number of candidates.
8632 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
8634 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
8635 and mips16 directories.
8636 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
8637 (MULTILIB_DIRNAMES): Ditto.
8638 (MULTILIB_EXCEPTIONS): Add new exceptions.
8639 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
8640 (MULTILIB_DIRNAMES): Ditto.
8641 (MULTILIB_EXCEPTIONS): Add new exceptions.
8643 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8644 Marcus Shawcroft <marcus.shawcroft@arm.com>
8646 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
8647 SYMBOL_TINY_ABSOLUTE.
8648 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
8649 SYMBOL_TINY_ABSOLUTE.
8650 (aarch64_expand_mov_immediate): Likewise.
8651 (aarch64_classify_symbol): Likewise.
8652 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
8653 Permit SYMBOL_TINY_ABSOLUTE.
8654 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
8656 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8657 Marcus Shawcroft <marcus.shawcroft@arm.com>
8659 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
8660 Refactor if/switch. Replace gcc_assert with if.
8662 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8664 * config/i386/i386.c (initial_ix86_tune_features): Enable
8665 FP Reassociation for AMD bdver1 and bdver2.
8667 2013-05-29 Martin Jambor <mjambor@suse.cz>
8669 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
8670 and IMAGPART_EXPR do not occur within other handled_components.
8672 2013-05-29 Richard Biener <rguenther@suse.de>
8674 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
8675 access on whether the use is in the BB we currently try to
8677 (vect_bb_vectorization_profitable_p): Pass the BB we currently
8678 vectorize to vect_bb_slp_scalar_cost.
8680 2013-05-29 Richard Biener <rguenther@suse.de>
8682 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
8683 computing scalar cost offsetted by stmts that are kept live
8685 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
8686 for computation of scalar cost.
8688 2013-05-28 Steve Ellcey <sellcey@mips.com>
8690 * config/mips/mips-cpus.def (mips32r2): Change processor type.
8692 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
8694 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
8695 array notation built-in reduction functions.
8696 * doc/passes.texi (Passes): Added documentation about changes done
8698 * doc/invoke.texi (C Dialect Options): Added documentation about
8699 the -fcilkplus flag.
8700 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
8701 (BUILTINS_DEF): Depend on cilkplus.def.
8702 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
8703 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
8704 * cilkplus.def: New file.
8706 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
8708 PR rtl-optimization/57439
8709 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
8711 2013-05-28 Easwaran Raman <eraman@google.com>
8713 PR tree-optimization/57337
8714 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
8715 (find_insert_point): Correctly identify the insertion point
8716 when two statements with the same UID is compared.
8718 2013-05-28 Richard Biener <rguenther@suse.de>
8720 PR tree-optimization/56787
8721 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
8722 from the list of data references.
8723 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
8725 (vect_analyze_loop_operations): Likewise.
8726 (vect_transform_loop): Remove clobbers.
8728 2013-05-28 Martin Jambor <mjambor@suse.cz>
8730 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
8731 and REALPART_EXPRs have scalar type.
8733 2013-05-28 Richard Biener <rguenther@suse.de>
8735 PR tree-optimization/57411
8736 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
8738 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
8739 virtual operand propagation.
8741 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
8743 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
8744 destination register for bmasksi_vis.
8745 (vector_init_bshuffle): Likewise.
8746 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
8748 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
8750 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
8751 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
8752 mode if the instruction isn't available in the original mode.
8753 * config/sparc/sparc.opt (mfix-ut699): New option.
8754 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
8755 (divdf3): Turn into expander.
8756 (divdf3_nofix): New insn.
8757 (divdf3_fix): Likewise.
8758 (divsf3): Disable if -mfix-ut699.
8759 (sqrtdf2): Turn into expander.
8760 (sqrtdf2_nofix): New insn.
8761 (sqrtdf2_fix): Likewise.
8762 (sqrtsf2): Disable if -mfix-ut699.
8764 2013-05-27 Richard Biener <rguenther@suse.de>
8767 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
8768 block for the new loop.
8770 2013-05-27 Richard Biener <rguenther@suse.de>
8772 PR tree-optimization/57343
8773 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
8774 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
8775 (number_of_iterations_cond): Do not build the folded tree.
8777 2013-05-27 Richard Biener <rguenther@suse.de>
8781 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8782 OEP_CONSTANT_ADDRESS_OF retained.
8784 PR tree-optimization/57417
8785 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
8787 (set_ssa_val_to): Compare addresses using
8788 get_addr_base_and_unit_offset.
8790 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
8792 PR rtl-optimization/56833
8793 * postreload.c (move2add_record_mode): New function.
8794 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
8795 (move2add_use_add2_insn): Use move2add_record_sym_value.
8796 (move2add_use_add3_insn): Likewise.
8797 (reload_cse_move2add): Use move2add_valid_value_p and
8798 move2add_record_mode. Invalidate call-clobbered and REG_INC
8799 affected regs by setting reg_mode to VOIDmode.
8800 (move2add_note_store): Don't pretend the inside of a SUBREG is
8801 the actual destination. Invalidate single/leading registers by
8802 setting reg_mode to VOIDmode.
8803 Use move2add_record_sym_value, move2add_valid_value_p and
8804 move2add_record_mode.
8806 2013-05-27 Richard Biener <rguenther@suse.de>
8808 PR tree-optimization/57396
8809 * tree-affine.c (double_int_constant_multiple_p): Properly
8810 return false for val == 0 and div != 0.
8812 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8814 * config/mips/mips.h: Use #elif in preprocessor conditions.
8816 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8819 * config/mips/constraints.md (kl): New constraint.
8820 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
8821 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
8822 constraint for operand 0. Split after CSE for MIPS16. Emit a move
8824 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
8826 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8829 * config/mips/mips.c (mips_can_inline_p): New function.
8830 (TARGET_CAN_INLINE_P): Define.
8832 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
8834 * sched-int.h (ds_t, dw_t): Make unsigned int.
8835 Fix documentation that describes how all the ds_t bits are used.
8836 Reserve the last bit for delayed-branch scheduling.
8837 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
8838 (BITS_PER_DEP_WEAK): Fix definition and documentation.
8839 (gen_dep_weak_1): Remove prototype.
8840 * sched-deps.c (get_dep_weak_1): Make static.
8841 * target.def (speculate_insn, needs_block_p, gen_spec_check,
8842 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
8843 * doc/tm.texi: Regenerate.
8844 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
8846 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
8849 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
8851 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
8852 Sandra Loosemore <sandra@codesourcery.com>
8854 * config.gcc (powerpc-*): Allow native for with-cpu.
8856 2013-05-24 Jeff Law <law@redhat.com>
8858 PR tree-optimization/57124
8859 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
8860 conversion feeding a condition if the range has an overflow
8861 if -fstrict-overflow. Add warnings for when we do make the
8864 2013-05-24 Dehao Chen <dehao@google.com>
8866 * tree-cfg.c (locus_discrim_map): Fix the typo.
8867 (locus_discrim_hasher): Likewise.
8868 (locus_discrim_hasher::hash): Likewise.
8869 (locus_discrim_hasher::equal): Likewise.
8871 2013-05-24 Martin Jambor <mjambor@suse.cz>
8873 PR tree-optimization/57294
8874 * cgraph.h (ipa_record_stmt_references): Declare.
8875 * cgraphbuild.c (ipa_record_stmt_references): New function.
8876 (build_cgraph_edges): Use ipa_record_stmt_references.
8877 (rebuild_cgraph_edges): Likewise.
8878 (cgraph_rebuild_references): Likewise.
8879 * ipa-prop.c (ipa_modify_call_arguments): Discard references
8880 associated with the old statement and build references from the
8881 newly built statements.
8882 * ipa-ref.c (ipa_remove_stmt_references): New function.
8883 * ipa-ref.h (ipa_remove_stmt_references): Declare.
8885 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
8887 * lra-constraints.c (emit_spill_move): Use smaller mode for
8889 (check_and_process_move): Consider mem-reg moves for secondary
8891 (curr_insn_transform): Don't lose insns emitted before for
8892 secondary memory moves.
8893 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
8894 reg set up in the current insn.
8896 2013-05-24 Dehao Chen <dehao@google.com>
8898 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
8900 (locus_descrim_hasher::equal): Likewise.
8901 (build_gimple_cfg): New discriminator assignment algorithm.
8902 (make_edges): Likewise.
8903 (next_discriminator_for_locus): Likewise.
8904 (same_line_p): Likewise.
8905 (assign_discriminators): Likewise.
8906 (make_cond_expr_edges): Likewise.
8907 (make_gimple_switch_edges): Likewise.
8908 (make_goto_expr_edges): Likewise.
8909 (make_gimple_asm_edges): Likewise.
8911 2013-05-24 Ian Bolton <ian.bolton@arm.com>
8913 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
8914 X format specifier to only display bottom 16 bits.
8915 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
8916 immediate to match for operand 2, since it will be masked.
8918 2013-05-24 Richard Biener <rguenther@suse.de>
8920 PR tree-optimization/57287
8921 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
8922 all SSA names that occur in abnormal PHIs.
8924 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8927 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8928 that index is not negative.
8930 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8932 PR rtl-optimization/55177
8933 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8934 (simplify_byte_swapping_operation): New.
8935 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8936 (simplify_relational_operation_1): Deal with BSWAP.
8938 2013-05-23 Richard Henderson <rth@redhat.com>
8941 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8942 (ix86_reorg): Call it.
8944 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8947 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8948 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8949 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8951 2013-05-23 Christian Bruel <christian.bruel@st.com>
8954 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8956 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8957 Marcus Shawcroft <marcus.shawcroft@arm.com>
8959 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8960 * config/aarch64/constraints.md (Usa): Remove.
8961 * doc/md.texi (AArch64 Usa): Remove.
8963 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8964 Marcus Shawcroft <marcus.shawcroft@arm.com>
8966 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8967 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8968 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8969 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8971 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8973 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8974 instruction (AdvSIMD).
8975 * config/aarch64/aarch64-builtins.c
8976 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8977 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8979 2013-05-23 Martin Jambor <mjambor@suse.cz>
8982 * tree.h (contains_bitfld_component_ref_p): Declare.
8983 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8984 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8986 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8987 not access a bit-field. Assert all final offsets are byte-aligned.
8989 2013-05-23 Richard Biener <rguenther@suse.de>
8991 PR tree-optimization/57380
8992 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8993 least one invariant or re-used load.
8994 * passes.c (init_optimization_passes): Move pass_phiprop before
8997 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8999 * config/aarch64/aarch64-simd.md
9000 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
9002 2013-05-23 Richard Biener <rguenther@suse.de>
9005 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
9006 OEP_CONSTANT_ADDRESS_OF retained.
9008 2013-05-23 Jakub Jelinek <jakub@redhat.com>
9011 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
9012 don't lower unit. Handle unit not being always BITS_PER_WORD.
9014 2013-05-23 Richard Biener <rguenther@suse.de>
9016 PR rtl-optimization/57341
9017 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
9018 instead of true_dependence.
9020 2013-05-22 David Malcolm <dmalcolm@redhat.com>
9022 * bb-reorder.c (branch_threshold): Make const.
9023 (exec_threshold): Ditto.
9025 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9026 Pat Haugen <pthaugen@us.ibm.com>
9027 Peter Bergner <bergner@vnet.ibm.com>
9029 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
9030 documentation for the power8 crypto builtins.
9032 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
9034 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
9035 macros for defining power8 builtin functions.
9036 (BU_P8V_AV_2): Likewise.
9037 (BU_P8V_AV_P): Likewise.
9038 (BU_P8V_VSX_1): Likewise.
9039 (BU_P8V_OVERLOAD_1): Likewise.
9040 (BU_P8V_OVERLOAD_2): Likewise.
9041 (BU_CRYPTO_1): Likewise.
9042 (BU_CRYPTO_2): Likewise.
9043 (BU_CRYPTO_3): Likewise.
9044 (BU_CRYPTO_OVERLOAD_1): Likewise.
9045 (BU_CRYPTO_OVERLOAD_2): Likewise.
9046 (XSCVSPDP): Fix typo, point to the correct instruction.
9047 (VCIPHER): Add power8 crypto builtins.
9048 (VCIPHERLAST): Likewise.
9049 (VNCIPHER): Likewise.
9050 (VNCIPHERLAST): Likewise.
9051 (VPMSUMB): Likewise.
9052 (VPMSUMH): Likewise.
9053 (VPMSUMW): Likewise.
9054 (VPERMXOR_V2DI): Likewise.
9055 (VPERMXOR_V4SI: Likewise.
9056 (VPERMXOR_V8HI: Likewise.
9057 (VPERMXOR_V16QI: Likewise.
9058 (VSHASIGMAW): Likewise.
9059 (VSHASIGMAD): Likewise.
9061 (VPERMXOR): Likewise.
9062 (VSHASIGMA): Likewise.
9064 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9065 __CRYPTO__ if the crypto instructions are available.
9066 (altivec_overloaded_builtins): Add support for overloaded power8
9069 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
9070 support for power8 crypto builtins.
9071 (builtin_function_type): Likewise.
9072 (altivec_init_builtins): Add support for builtins that take vector
9073 long long (V2DI) arguments.
9075 * config/rs6000/crypto.md: New file, define power8 crypto
9078 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9079 Pat Haugen <pthaugen@us.ibm.com>
9080 Peter Bergner <bergner@vnet.ibm.com>
9082 * doc/invoke.texi (Option Summary): Add power8 options.
9083 (RS/6000 and PowerPC Options): Likewise.
9085 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
9086 constraints.md instead of rs6000.h. Reorder w* constraints. Add
9087 wm, wn, wr documentation.
9089 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
9090 registers if direct move instructions are enabled.
9091 (wn): New constraint for no registers.
9092 (wq): New constraint for quad word even GPR registers.
9093 (wr): New constraint if 64-bit instructions are enabled.
9094 (wv): New constraint if power8 vector instructions are enabled.
9095 (wQ): New constraint for quad word memory locations.
9097 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
9098 constraint for 0..15 for crypto instructions.
9099 (gpc_reg_operand): If VSX allow registers in VSX registers as well
9100 as GPR and floating point registers.
9101 (int_reg_operand): New predicate to match only GPR registers.
9102 (base_reg_operand): New predicate to match base registers.
9103 (quad_int_reg_operand): New predicate to match even GPR registers
9104 for quad memory operations.
9105 (vsx_reg_or_cint_operand): New predicate to allow vector logical
9106 operations in both GPR and VSX registers.
9107 (quad_memory_operand): New predicate for quad memory operations.
9108 (reg_or_indexed_operand): New predicate for direct move support.
9110 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
9111 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
9112 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
9113 (POWERPC_MASKS): Add power8 options.
9114 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
9117 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9118 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
9120 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
9121 (-mpower8-fusion): New power8 options.
9122 (-mpower8-fusion-sign): Likewise.
9123 (-mpower8-vector): Likewise.
9124 (-mcrypto): Likewise.
9125 (-mdirect-move): Likewise.
9126 (-mquad-memory): Likewise.
9128 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
9130 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
9132 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
9133 (rs6000_debug_vector_unit): Add p8_vector.
9134 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
9135 definitions. Also print fusion state.
9136 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
9137 (rs6000_builtin_mask_calculate): Add power8 builtin support.
9138 (rs6000_option_override_internal): Add support for power8.
9139 (rs6000_common_init_builtins): Add debugging for skipped builtins
9141 (rs6000_adjust_cost): Add power8 support.
9142 (rs6000_issue_rate): Likewise.
9143 (insn_must_be_first_in_group): Likewise.
9144 (insn_must_be_last_in_group): Likewise.
9145 (force_new_group): Likewise.
9146 (rs6000_register_move_cost): Likewise.
9147 (rs6000_opt_masks): Likewise.
9149 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
9150 power8 capable assembler, default to power7 options.
9151 (TARGET_DIRECT_MOVE): Likewise.
9152 (TARGET_CRYPTO): Likewise.
9153 (TARGET_P8_VECTOR): Likewise.
9154 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
9155 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
9156 (VECTOR_MEM_P8_VECTOR_P): Likewise.
9157 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
9158 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
9159 (TARGET_XSCVDPSPN): Likewise.
9160 (TARGET_XSCVSPDPN): Likewsie.
9161 (TARGET_SYNC_HI_QI): Likewise.
9162 (TARGET_SYNC_TI): Likewise.
9163 (MASK_CRYPTO): Likewise.
9164 (MASK_DIRECT_MOVE): Likewise.
9165 (MASK_P8_FUSION): Likewise.
9166 (MASK_P8_VECTOR): Likewise.
9167 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
9168 temporary used by some of the direct move instructions to get two FP
9169 temporary registers does not force creation of a stack frame.
9170 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
9171 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
9172 that any VSX registers are tieable, even if they are also an
9173 Altivec vector mode.
9174 (r6000_reg_class_enum): Add wm, wr, wv constraints.
9175 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
9176 (RS6000_BTM_CRYPTO): Likewise.
9177 (RS6000_BTM_COMMON): Likewise.
9179 * config/rs6000/rs6000.md (cpu attribute): Add power8.
9180 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
9181 (enum rs6000_vector): Add power8 vector support.
9183 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9187 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
9188 (any_sibcall_could_use_r3): this and handle indirect calls.
9189 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
9191 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9193 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
9195 2013-05-22 Richard Biener <rguenther@suse.de>
9198 * profile.c (branch_prob): Do not split blocks that are
9199 abnormally receiving from ECF_RETURNS_TWICE functions.
9201 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9203 * recog.c (offsettable_address_addr_space_p): Fix calculation of
9204 address mode. Move pointer mode initialization to the same place.
9206 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9208 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
9209 while it has any effect.
9211 2013-05-21 Easwaran Raman <eraman@google.com>
9213 PR tree-optimization/57322
9214 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
9215 UID of the statement added to the BB to be 1.
9217 2013-05-21 Jakub Jelinek <jakub@redhat.com>
9219 PR tree-optimization/57331
9220 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
9221 of conversion from pointer type to integral type with integer.
9223 2013-05-21 Martin Jambor <mjambor@suse.cz>
9226 * ipa-prop.c (ipa_read_node_info): Process param_used and
9227 controlled_uses in the same order as when writing.
9229 2013-05-21 Magnus Granberg <baldrick@free.fr>
9232 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
9234 2013-05-21 Richard Biener <rguenther@suse.de>
9236 PR tree-optimization/57318
9237 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
9238 estimate stmts with side-effects as likely eliminated.
9240 2013-05-21 Richard Biener <rguenther@suse.de>
9242 PR tree-optimization/57330
9243 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
9244 preserve the call stmts fntype.
9246 2013-05-21 Richard Biener <rguenther@suse.de>
9248 PR tree-optimization/57303
9249 * tree-ssa-sink.c (statement_sink_location): Improve killing
9250 stmt detection and properly handle self-assignments.
9252 2013-05-21 Christian Bruel <christian.bruel@st.com>
9254 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
9255 spanning registers. LEAF_REG_REMAP is supported only for contiguous
9256 registers. Set register size out of the PARALLEL loop.
9258 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
9261 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
9262 (*fmasf4, *fmasf4_media): New insns.
9264 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9266 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
9267 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
9268 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
9269 (mips_idiv_insns): Update the comments to say that the returned
9270 instruction counts are in units of BASE_INSN_LENGTH.
9271 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
9272 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
9273 using 2 rather than 4 as the length of indirect MIPS16 and
9274 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
9275 length of a NOP. Don't divide MIPS16 lengths by 2.
9276 (mips16_split_long_branches): Assume a branch is long if the
9277 length is greater than 4 rather than 8.
9278 * config/mips/mips.md (length): Give MIPS16 lengths directly,
9279 rather than multiplying them by 2. Multiply instruction counts
9280 by BASE_INSN_LENGTH rather than 4.
9281 (*jump_mips16, tls_get_tp_mips16_<mode>)
9282 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
9284 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9286 * config/mips/mips.md (extended_mips16): Remove branch case.
9287 (length): Remove duplicated extended_mips16 test.
9289 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9291 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
9293 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
9295 * recog.h (Recog_data): Rename to...
9296 (recog_data_d): ...this.
9297 (recog_data): Update accordingly.
9298 * recog.c (recog_data): Likewise.
9299 * reload.c (save_recog_data): Likewise.
9300 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
9301 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
9303 2013-05-17 Julian Brown <julian@codesourcery.com>
9305 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
9306 found in a REG_EQUAL note, invalidate it.
9308 2013-05-17 Easwaran Raman <eraman@google.com>
9310 * tree-ssa-reassoc.c (find_insert_point): New function.
9311 (insert_stmt_after): Likewise.
9312 (get_def_stmt): Likewise.
9313 (ensure_ops_are_available): Likewise.
9314 (not_dominated_by): Likewise.
9315 (rewrite_expr_tree): Do not move statements beyond what is
9316 necessary. Remove call to swap_ops_for_binary_stmt...
9317 (reassociate_bb): ... and move it here.
9318 (build_and_add_sum): Assign UIDs for new statements.
9319 (linearize_expr): Likewise.
9320 (do_reassoc): Renumber gimple statement UIDs.
9322 2013-05-17 Jan Hubicka <jh@suse.cz>
9324 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
9326 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
9327 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
9329 (output_weakrefs): Update.
9331 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
9332 Martin Jambor <mjambor@suse.cz>
9335 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
9336 value that corresponds to the given aggval is found in values vector.
9338 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
9340 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
9341 sse, sse2, sse3, ssse3 and sse4a flags to options.
9343 2013-05-17 David Malcolm <dmalcolm@redhat.com>
9345 * gengtype-state.c: (s_expr_writer): New class, to handle
9346 prettifying of output layout of s-expressions.
9347 (state_writer): New class, to write out gtype.state.
9348 (state_written_type_count): Move this variable into member data of
9350 (s_expr_writer::s_expr_writer): New code: constructor for new class
9351 (state_writer::state_writer(): ditto
9352 (s_expr_writer::write_new_line): New function
9353 (s_expr_writer::write_any_indent): ditto
9354 (s_expr_writer::begin_s_expr): ditto
9355 (s_expr_writer::end_s_expr): ditto
9356 (write_state_fileloc): convert to method of state_writer...
9357 (state_writer:: write_state_fileloc): ...and use methods of
9358 s_expr_writer to write indentation into the gtype.state output file
9359 to visually represent the hierarchical structure of the list
9361 (write_state_fields): ditto, renaming to...
9362 (state_writer::write_state_fields)
9363 (write_state_a_string): ditto, renaming to...
9364 (state_writer::write_state_a_string)
9365 (write_state_string_option): ditto, renaming to...
9366 (state_writer::write_state_string_option)
9367 (write_state_type_option): ditto, renaming to...
9368 (state_writer::write_state_type_option)
9369 (write_state_nested_option): ditto, renaming to...
9370 (state_writer::write_state_nested_option)
9371 (write_state_option): ditto, renaming to...
9372 (state_writer::write_state_option)
9373 (write_state_options): ditto, renaming to...
9374 (state_writer::write_state_options)
9375 (write_state_lang_bitmap): ditto, renaming to...
9376 (state_writer::write_state_lang_bitmap)
9377 (write_state_version): ditto, renaming to...
9378 (state_writer::write_state_version)
9379 (write_state_scalar_type): ditto, renaming to...
9380 (state_writer::write_state_scalar_type)
9381 (write_state_string_type): ditto, renaming to...
9382 (state_writer::write_state_string_type)
9383 (write_state_undefined_type): ditto, renaming to...
9384 (state_writer::write_state_undefined_type)
9385 (write_state_struct_union_type): ditto, renaming to...
9386 (state_writer::write_state_struct_union_type)
9387 (write_state_struct_type): ditto, renaming to...
9388 (state_writer::write_state_struct_type)
9389 (write_state_user_struct_type): ditto, renaming to...
9390 (state_writer::write_state_user_struct_type)
9391 (write_state_lang_struct_type): ditto, renaming to...
9392 (state_writer::write_state_lang_struct_type)
9393 (write_state_param_struct_type): ditto, renaming to...
9394 (state_writer::write_state_param_struct_type)
9395 (write_state_pointer_type): ditto, renaming to...
9396 (state_writer::write_state_pointer_type)
9397 (write_state_array_type): ditto, renaming to...
9398 (state_writer::write_state_array_type)
9399 (write_state_gc_used): ditto, renaming to...
9400 (state_writer::write_state_gc_used)
9401 (write_state_common_type_content): ditto, renaming to...
9402 (state_writer::write_state_common_type_content)
9403 (write_state_type): ditto, renaming to...
9404 (state_writer::write_state_type)
9405 (write_state_pair_list): ditto, renaming to...
9406 (state_writer::write_state_pair_list)
9407 (write_state_pair): ditto, renaming to...
9408 (state_writer::write_state_pair)
9409 (write_state_typedefs): ditto, renaming to...
9410 (state_writer::write_state_typedefs)
9411 (write_state_structures): ditto, renaming to...
9412 (state_writer::write_state_structures)
9413 (write_state_param_structs): ditto, renaming to...
9414 (state_writer::write_state_param_structs)
9415 (write_state_variables): ditto, renaming to...
9416 (state_writer::write_state_variables)
9417 (write_state_srcdir): ditto, renaming to...
9418 (state_writer::write_state_srcdir)
9419 (write_state_files_list): ditto, renaming to...
9420 (state_writer::write_state_files_list)
9421 (write_state_languages): ditto, renaming to...
9422 (state_writer::write_state_languages)
9423 (write_state): create a state_writer instance and use it when
9424 writing out the state file
9426 2013-05-17 Mike Stump <mikestump@comcast.net>
9428 PR rtl-optimization/57304
9429 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
9430 accessing DF_REF_REAL_LOC.
9432 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9434 PR rtl-optimization/57281
9435 PR rtl-optimization/57300
9436 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
9437 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
9438 what the other splitter did if the registers are dead.
9440 2013-05-17 Richard Biener <rguenther@suse.de>
9442 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
9445 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9447 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
9450 2013-05-17 Marek Polacek <polacek@redhat.com>
9452 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
9453 length when doing non-zero store of storing '\0' to '\0'.
9455 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9457 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
9458 vect_external_def oprnd1 with loop_vinfo, try to emit
9459 optional cast, negation and and stmts on the loop preheader
9460 edge instead of into the pattern def seq.
9462 PR tree-optimization/57051
9463 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
9464 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
9466 2013-05-16 Nick Clifton <nickc@redhat.com>
9468 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
9469 (rl78_is_naked_func): New function.
9470 (rl78_expand_prologue): Skip prologue generation for naked functions.
9471 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
9472 * doc/extend.texi (naked): Add RL78 to the list of processors
9473 that supports this attribute.
9475 2013-05-16 Jeff Law <law@redhat.com>
9477 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
9479 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
9481 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
9482 cache parameters using detect_caches_amd also for CYRIX,
9483 NSC and TM2 signatures.
9485 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
9486 Dzianis Kahanovich <mahatma@eu.by>
9490 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9491 VIA/Centaur processors and determine their cache parameters
9492 using detect_caches_amd.
9494 2013-05-16 Teresa Johnson <tejohnson@google.com>
9496 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
9497 (rtl_verify_edges): New function.
9498 (rtl_verify_bb_insns): Ditto.
9499 (rtl_verify_bb_pointers): Ditto.
9500 (rtl_verify_bb_insn_chain): Ditto.
9501 (rtl_verify_fallthru): Ditto.
9502 (rtl_verify_bb_layout): Ditto.
9503 (rtl_verify_flow_info_1): Outline checks into new functions.
9504 (rtl_verify_flow_info): Ditto.
9506 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
9508 * cfghooks.c (copy_bbs): Add update_dominance argument.
9509 * cfghooks.h (copy_bbs): Update prototype.
9510 * tree-cfg.c (gimple_duplicate_sese_region):
9511 Add update_dominance argument.
9512 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
9513 * tree-ssa-loop-ch.c (copy_loop_headers): Update
9514 gimple_duplicate_sese_region call.
9515 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
9516 Update copy_bbs call.
9517 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
9518 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
9520 2013-05-16 Jakub Jelinek <jakub@redhat.com>
9522 * tree-vectorizer.h (NUM_PATTERNS): Increment.
9523 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
9524 vect_recog_rotate_pattern.
9525 (vect_recog_rotate_pattern): New function.
9527 2013-05-16 Jason Merrill <jason@redhat.com>
9529 * Makefile.in (LLINKER): New variable.
9530 (mostlyclean): Remove link mutex.
9531 * configure.ac: Handle --enable-link-mutex.
9532 * lock-and-run.sh: New script.
9534 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9537 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
9540 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9542 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
9544 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
9546 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
9547 * config/arm/arm.c (next_consecutive_mem): New function.
9548 (gen_movmem_ldrd_strd): Likewise.
9549 * config/arm/arm.md (movmemqi): Update condition and code.
9550 (unaligned_loaddi, unaligned_storedi): New patterns.
9552 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9554 * config.gcc: Obsolete *-*-solaris2.9*.
9555 * doc/install.texi (Specific, *-*-solaris2*): Document it.
9557 2013-05-16 Richard Biener <rguenther@suse.de>
9559 * passes.c (init_optimization_passes): Move pass_parallelize_loops
9560 earlier, after GRAPHITE transforms and IV canonicalization.
9562 2013-05-16 Jakub Jelinek <jakub@redhat.com>
9564 * omp-low.c (extract_omp_for_data): For collapsed loops,
9565 if at least one of the loops is known at compile time to
9566 iterate zero times, set count to 0.
9567 (expand_omp_regimplify_p): New function.
9568 (expand_omp_for_generic): For collapsed loops, if at least
9569 one of the loops isn't known to iterate at least once,
9570 add runtime check with setting count to 0.
9571 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
9572 For unsigned types if it isn't known at compile time that
9573 the loop will iterate at least once, add runtime check to bypass
9574 the whole loop if initial condition isn't true.
9576 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
9578 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
9580 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
9583 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
9584 transformations to avoid an infinite loop.
9586 2013-05-16 Marek Polacek <polacek@redhat.com>
9588 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
9590 2013-05-15 Leif Ekblad <leif@rdos.net>
9592 * config/i386/i386.c (ix86_decompose_address): Use
9593 DEFAULT_TLS_SEG_REG to access TLS segment register.
9594 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
9595 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
9596 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
9598 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
9601 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
9602 sibling calls to functions that would normally be lazily bound,
9603 unless $gp is call-clobbered.
9605 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
9607 * config/i386/i386.c (ix86_option_override_internal): Update
9608 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
9609 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
9610 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
9611 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
9613 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
9615 2013-05-15 Andreas Schwab <schwab@suse.de>
9617 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
9618 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
9621 2013-05-15 Teresa Johnson <tejohnson@google.com>
9623 * loop-unroll.c (report_unroll_peel): Check decision before
9624 emitting unroll/peel message.
9626 2013-05-15 Teresa Johnson <tejohnson@google.com>
9628 * function.h (has_bb_partition): New rtl_data flag.
9629 (bb_reorder_complete): Ditto.
9630 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
9631 instead of flag_reorder_blocks_and_partition.
9632 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
9633 with some enhancements.
9634 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
9635 * bb-reorder.c (connect_traces): Check for has_bb_partition
9636 instead of flag_reorder_blocks_and_partition.
9637 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
9638 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
9639 verify_hot_cold_block_grouping.
9640 (partition_hot_cold_basic_blocks): Set has_bb_partition.
9642 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9645 * config/arm/predicates.md (call_insn_operand): New predicate.
9646 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
9647 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
9648 if insn is not a tail call.
9649 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
9651 * config/arm/arm.h (enum reg_class): New caller save register class.
9652 (REG_CLASS_NAMES): Likewise.
9653 (REG_CLASS_CONTENTS): Likewise.
9654 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
9657 2013-05-15 Richard Biener <rguenther@suse.de>
9659 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
9660 of MSG_OPTIMIZED_LOCATIONS.
9661 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
9662 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
9664 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
9665 of MSG_OPTIMIZED_LOCATIONS.
9666 (execute_vect_slp): Likewise.
9667 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
9668 (vect_create_cond_for_alias_checks): Likewise.
9669 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
9670 (vect_recog_widen_mult_pattern): Likewise.
9671 (vect_recog_widen_sum_pattern): Likewise.
9672 (vect_recog_over_widening_pattern): Likewise.
9673 (vect_recog_widen_shift_pattern): Likewise.
9674 (vect_recog_vector_vector_shift_pattern): Likewise.
9675 (vect_recog_divmod_pattern): Likewise.
9676 (vect_recog_mixed_size_cond_pattern): Likewise.
9677 (vect_recog_bool_pattern): Likewise.
9678 (vect_pattern_recog_1): Likewise.
9680 2013-05-15 Martin Jambor <mjambor@suse.cz>
9682 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
9683 non-functions to builtin_unreachable.
9684 * ipa-inline-transform.c (inline_call): Do not assert estimates were
9685 correct when new direct edges were discovered.
9687 2013-05-15 Martin Jambor <mjambor@suse.cz>
9689 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
9690 header, print symbol order instead of node uid, print more information
9691 about indirect edge targets.
9692 (ipa_make_edge_direct_to_target): Print symbol order instead of node
9694 (ipa_make_edge_direct_to_target): Likewise.
9695 (remove_described_reference): Likewise.
9696 (propagate_controlled_uses): Likewise.
9697 (ipa_print_node_params): Also print symbol order.
9698 (ipcp_transform_function): Print symbol order instead of node uids.
9699 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
9700 (cgraph_get_create_real_symbol_node): Likewise.
9701 * ipa-cp.c (print_lattice): Likewise.
9702 (print_all_lattices): Likewise.
9703 (determine_versionability): Likewise.
9704 (initialize_node_lattices): Likewise.
9705 (estimate_local_effects): Likewise.
9706 (update_profiling_info): Likewise.
9707 (create_specialized_node): Likewise.
9708 (perhaps_add_new_callers): Likewise.
9709 (decide_about_value): Likewise.
9710 (decide_whether_version_node): Likewise.
9711 (identify_dead_nodes): Likewise.
9712 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
9713 (dump_inline_summary): Likewise.
9714 (estimate_node_size_and_time): Likewise.
9715 (inline_analyze_function): Likewise.
9716 * ipa-inline.c (report_inline_failed_reason): Likewise.
9717 (want_early_inline_function_p): Likewise.
9718 (edge_badness): Likewise.
9719 (update_edge_key): Likewise.
9720 (inline_small_functions): Likewise. Add dumping of order to two other
9722 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
9723 instead of node uids.
9724 (propagate_pure_const): Likewise.
9725 (propagate_pure_const): Likewise.
9726 * ipa-utils.c (dump_cgraph_node_set): Likewise.
9727 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
9728 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
9730 * tree-pretty-print.c (dump_function_header): Likewise.
9731 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
9732 Print symbol order instead of node uids.
9734 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9736 * config/s390/s390.c (s390_register_move_cost): Don't impose the
9737 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
9739 2013-05-15 Richard Biener <rguenther@suse.de>
9741 PR tree-optimization/57275
9742 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
9743 return value for fail to do runtime alias checks for gather loads.
9745 2013-05-15 Jan Hubicka <jh@suse.cz>
9749 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
9750 weakrefs are not external.
9751 (lto_symtab_merge_decls): Fix thinko when dealing with
9752 non-lto_symtab decls.
9753 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
9754 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
9755 * varpool.c (dump_varpool_node): Dump more flags.
9757 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9759 * config/i386/i386.c (processor_alias_table): Add instruction
9760 FSGSBASE for AMD bdver3 architecture.
9762 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9764 * tree.c (warn_deprecated_use): Print file:line using locus color.
9765 * diagnostic.c (diagnostic_report_current_module): Print file:line
9766 and file:line:column using locus color.
9768 2013-05-14 Mike Stump <mikestump@comcast.net>
9770 * gdbinit.in: Add __null.
9772 2013-05-14 Mike Stump <mikestump@comcast.net>
9774 * recog.h: Rename struct recog_data to Recog_data.
9775 * recog.c: Likewise.
9776 * reload.c (can_reload_into): Likewise.
9777 * config/picochip/picochip.c: Likewise.
9779 2013-05-14 Mike Stump <mikestump@comcast.net>
9781 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
9783 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
9785 * resource.h (struct resources): Remove unch_memory member.
9786 (CLEAR_RESOURCE): Don't clear unch_memory.
9787 * resource.c (mark_referenced_resources): Don't set it.
9788 (mark_set_resources): Likewise.
9789 (mark_target_live_regs): Don't clear it.
9790 (init_resource_info): Likewise.
9791 * reorg.c (resource_conflicts_p): Don't compare it.
9792 (redundant_insn): Don't set it.
9794 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
9796 * emit-rtl.c (next_label): Remove unused function.
9797 (skip_consecutive_labels, link_cc0_insns): Move to ...
9798 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
9799 only place where these functions are used, and make them static.
9801 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9803 * fold-const.c (fold_negate_expr): Handle vectors.
9804 (fold_truth_not_expr): Make it static.
9805 (fold_invert_truthvalue): New static function.
9806 (invert_truthvalue_loc): Handle vectors. Do not call
9807 fold_truth_not_expr directly.
9808 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
9809 <TRUTH_NOT_EXPR>: Do not cast to boolean.
9810 (fold_comparison): Handle vector constants.
9811 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
9812 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
9813 * tree.h (fold_truth_not_expr): Remove declaration.
9815 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
9817 * config/aarch64/aarch64-simd.md
9818 (aarch64_vcond_internal<mode>): Rename to...
9819 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
9820 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
9821 float modes. Clarify all iterator modes.
9822 (vcond<mode><mode>): Use new name for vcond expanders.
9823 (vcond<v_cmp_result><mode>): Likewise.
9824 (vcondu<mode><mode>: Likewise.
9825 * config/aarch64/iterators.md (VDQF_COND): New.
9827 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9830 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
9831 variable for the shift amount. Check that we shift by non-negative
9834 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
9837 * config/arm/arm.h (EPILOGUE_USES): Only return true
9838 for LR_REGNUM after epilogue_completed.
9840 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
9842 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
9845 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9848 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
9849 and Solaris 11+/x86 with gld.
9850 * configure: Regenerate.
9852 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9854 * expmed.c (expand_shift_1): Canonicalize rotates by
9855 constant bitsize / 2 to bitsize - 1.
9856 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
9857 case ROTATERT>: Likewise.
9860 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9862 * config/i386/i386.md (rotateinv): New code attr.
9863 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9864 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9865 roll $31, %eax, etc.
9867 2013-05-14 Richard Biener <rguenther@suse.de>
9870 * tree-eh.c (sink_clobbers): Give up for successors with
9871 multiple predecessors and no virtual uses.
9873 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
9875 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
9876 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
9878 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9881 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
9882 the case when both op0 and op1 have VOIDmode.
9884 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9886 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
9887 in multiply-accumulate mode.
9889 2013-05-13 Guozhi Wei <carrot@google.com>
9891 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
9893 2013-05-13 Kai Tietz <ktietz@redhat.com>
9896 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
9897 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
9898 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
9899 * config/i386/i386.c (ix86_option_override_internal): Likewise.
9900 (ix86_expand_prologue): Likewise.
9901 (ix86_expand_split_stack_prologue): Likewise.
9902 (legitimate_pic_address_disp_p): Likewise.
9903 (legitimize_pic_address): Likewise.
9904 (legitimize_tls_address): Likewise.
9905 (legitimize_pe_coff_symbol): Likewise.
9906 (output_pic_addr_const): Likewise.
9907 (construct_plt_address): Likewise.
9908 (ix86_expand_call): Likewise.
9909 (x86_output_mi_thunk): Likewise.
9910 (x86_function_profiler): Likewise.
9912 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
9914 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
9915 similar switch cases.
9916 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
9917 (aarch64_simd_mov_to_<mode>low): Delete.
9918 (aarch64_simd_mov_to_<mode>high): Delete.
9919 (move_lo_quad_<mode>): Add w<-r alternative.
9920 (aarch64_simd_move_hi_quad_<mode>): Likewise.
9921 (aarch64_simd_mov_from_*): Update type attribute.
9922 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
9925 2013-05-13 Jan Hubicka <jh@suse.cz>
9927 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9928 * config/i386/i386.c (ix86_compute_frame_layout,
9929 ix86_expand_epilogue, emit_i387_cw_initialization,
9930 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9931 ix86_local_alignment): Fix use of size/speed predicates.
9933 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9935 PR tree-optimization/45216
9936 PR tree-optimization/57157
9937 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9938 the (-Y) & (B - 1) variant if OP is |.
9939 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9940 return shifted. Use (-op1) & (prec - 1) as other_amount
9941 instead of prec - op1.
9943 2013-05-13 Martin Jambor <mjambor@suse.cz>
9946 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9947 (ipa_constant_data): New type.
9948 (ipa_jump_func): Use ipa_constant_data to hold information about
9949 constant jump functions.
9950 (ipa_get_jf_constant): Adjust to jump function type changes.
9951 (ipa_get_jf_constant_rdesc): New function.
9952 (ipa_param_descriptor): New field controlled_uses.
9953 (ipa_get_controlled_uses): New function.
9954 (ipa_set_controlled_uses): Likewise.
9955 * ipa-ref.h (ipa_find_reference): Declare.
9956 * ipa-prop.c (ipa_cst_ref_desc): New type.
9957 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9959 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9960 New parameter cs. Adjust all callers.
9961 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9962 (remove_described_reference): New function.
9963 (jfunc_rdesc_usable): Likewise.
9964 (try_make_edge_direct_simple_call): Decrement controlled use count,
9965 attempt to remove reference if it hits zero.
9966 (combine_controlled_uses_counters): New function.
9967 (propagate_controlled_uses): Likewise.
9968 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9969 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9970 (ipa_print_node_params): Print described use counter.
9971 (ipa_write_jump_function): Adjust to jump function type changes.
9972 (ipa_read_jump_function): New parameter CS, pass it to
9973 ipa_set_jf_constant. Adjust caller.
9974 (ipa_write_node_info): Stream controlled use count
9975 (ipa_read_node_info): Likewise.
9976 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9978 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9979 count. Remove cloning-added reference if it reaches zero.
9980 * ipa-ref.c (ipa_find_reference): New function.
9982 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9984 * config/i386/i386.c (processor_target_table): Modified default
9985 alignment values for AMD BD and BT architectures.
9987 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9989 * tree-vect-generic.c (uniform_vector_p): Move ...
9990 * tree.c (uniform_vector_p): ... here.
9991 * tree.h (uniform_vector_p): Declare it.
9992 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9995 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9997 PR tree-optimization/57230
9998 * tree-ssa-strlen.c (handle_char_store): Record length for
9999 array store from STRING_CST.
10001 PR tree-optimization/57230
10002 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
10005 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
10007 * config/epiphany/epiphany.c (epiphany_init): Check size of
10008 NUM_MODES_FOR_MODE_SWITCHING.
10009 (epiphany_expand_prologue):
10010 Remove CONFIG_REGNUM initial value handling code.
10011 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
10012 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
10013 (emit_set_fp_mode, epiphany_mode_after): Likewise.
10014 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
10015 Don't return 1 for FP_MODE_NONE.
10016 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
10017 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
10018 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
10019 * config/epiphany/epiphany.md (save_config): New pattern.
10021 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
10023 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
10025 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
10027 * config/i386/i386.md (memory): Handle sseishft1.
10028 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
10029 (*vec_extractv2di_1): Ditto.
10031 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
10033 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
10036 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
10038 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
10039 Add mthumb/march=armv7-a multilib.
10040 Add mthumb/march=armv7-r multilib.
10041 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
10043 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
10045 * config/v850/t-rtems: Add more multilibs.
10047 2013-05-10 Richard Biener <rguenther@suse.de>
10049 PR tree-optimization/57214
10050 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
10051 not propagate from SSA names that occur in abnormal PHI nodes.
10053 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
10055 * stor-layout.c (element_precision): New function.
10056 * machmode.h (element_precision): Declare it.
10057 * tree.c (build_minus_one_cst): New function.
10058 (element_precision): Likewise.
10059 * tree.h (build_minus_one_cst): Declare new function.
10060 (element_precision): Likewise.
10061 * fold-const.c (operand_equal_p): Use element_precision.
10062 (fold_binary_loc): Handle vector types.
10063 * convert.c (convert_to_integer): Use element_precision.
10064 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
10067 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
10069 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
10070 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
10071 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
10072 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10073 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
10074 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
10075 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
10076 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10077 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
10078 (Uuw8): New constraints.
10079 (Usb4): Move into alphabetical order.
10080 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
10081 (sd8_operand, ub8_operand, uw8_operand): New predicates.
10082 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
10083 previously unnamed patterns.
10084 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
10085 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
10086 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
10087 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
10088 of set_attr_alternative/if_then_else. Use extended_mips16 instead
10089 of specific lengths.
10091 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10093 * config/i386/i386.md (rotateinv): New code attr.
10094 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10095 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10096 roll $31, %eax, etc.
10098 PR tree-optimization/45216
10099 PR tree-optimization/57157
10100 * tree-ssa-forwprop.c (simplify_rotate): New function.
10101 (ssa_forward_propagate_and_combine): Call it.
10103 2013-05-10 Richard Biener <rguenther@suse.de>
10105 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
10106 disable peeling when we version for aliasing.
10107 (vector_alignment_reachable_p): Honor explicit user alignment.
10108 (vect_supportable_dr_alignment): Likewise.
10109 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
10110 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
10111 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
10112 then peeling to arrange for the cost-model check to come first.
10114 2013-05-10 Alan Modra <amodra@gmail.com>
10116 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
10117 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
10118 * configure: Regenerate.
10120 2013-05-10 Alan Modra <amodra@gmail.com>
10123 * varasm.c (default_elf_select_section): Move !DECL_P check..
10124 (get_named_section): ..to here before calling get_section_name.
10126 (default_section_type_flags): Add DECL_P check.
10127 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
10128 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
10130 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
10132 * config/epiphany/epiphany.c (epiphany_expand_prologue):
10133 When using gen_stack_adjust_str with a register offset, add a
10134 REG_FRAME_RELATED_EXPR note.
10136 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
10138 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
10139 (*vec_extractv4si_zext_mem): Ditto.
10140 (*vec_extractv2di): Add 0->x and x->x alternatives.
10141 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
10142 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
10144 2013-05-09 Jason Merrill <jason@redhat.com>
10146 N3639 C++1y VLA support
10147 * gimplify.c (gimplify_vla_decl): Don't touch an existing
10150 * tree.c (build_constructor_va): New.
10151 * tree.h: Declare it.
10153 2013-05-09 Martin Jambor <mjambor@suse.cz>
10156 * gimple-fold.c (canonicalize_constructor_val): Call
10157 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
10159 2013-05-09 Jan Hubicka <jh@suse.cz>
10160 Richard Biener <rguenther@suse.de>
10163 * symtab.c (symtab_make_decl_local): Do not add private names.
10165 2013-05-09 Jan Hubicka <jh@suse.cz>
10168 * symtab.c (insert_to_assembler_name_hash): Handle clones.
10169 (unlink_from_assembler_name_hash): Likewise.
10170 (symtab_prevail_in_asm_name_hash, symtab_register_node,
10171 symtab_unregister_node, symtab_initialize_asm_name_hash,
10172 change_decl_assembler_name): Update.
10174 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
10176 * config/aarch64/aarch64.md: New movtf split.
10177 (*movtf_aarch64): Update.
10178 (aarch64_movdi_tilow): Handle TF modes and rename to
10179 aarch64_movdi_<mode>low.
10180 (aarch64_movdi_tihigh): Handle TF modes and rename to
10181 aarch64_movdi_<mode>high
10182 (aarch64_movtihigh_di): Handle TF modes and rename to
10183 aarch64_mov<mode>high_di
10184 (aarch64_movtilow_di): Handle TF modes and rename to
10185 aarch64_mov<mode>low_di
10186 (aarch64_movtilow_tilow): Remove spurious whitespace.
10187 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
10189 (aarch64_print_operand): Update.
10191 2013-05-09 Alan Modra <amodra@gmail.com>
10193 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
10195 * configure: Regenerate.
10197 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10199 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
10200 splitter preparation statements.
10201 * config/i386/sse.md (*vec_extract* splitters): Ditto.
10202 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
10205 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10207 * gimple-ssa-strength-reduction.c (count_candidates): Change
10208 return value to int.
10209 (analyze_candidates_and_replace): Change type of length to int.
10211 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10213 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
10214 (*vec_extract<mode>): Use VI12_128 mode iterator.
10215 (*vec_extract<mode>_mem): Ditto.
10216 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
10219 2013-05-08 Diego Novillo <dnovillo@google.com>
10224 2012-08-17 Diego Novillo <dnovillo@google.com>
10227 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
10228 * config.in: Regenerate.
10229 * configure: Regenerate.
10230 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
10232 2013-05-08 Jan Hubicka <jh@suse.cz>
10235 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
10236 * cgraph.h (symtab_node_base): Add unique_name.
10237 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
10238 input_overwrite_node, input_varpool_node): Stream unique_name.
10239 * cgraphclones.c (cgraph_create_virtual_clone,
10240 cgraph_function_versioning): Set unique_name.
10241 * ipa.c (function_and_variable_visibility): Set unique_name.
10243 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10245 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
10246 (alloc_cand_and_find_basis): Restrict conditional candidate
10247 processing to CAND_MULTs.
10249 2013-05-08 Jan Hubicka <jh@suse.cz>
10252 lto-symtab.c (lto_symtab_symbol_p): New function.
10253 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
10254 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
10255 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
10256 Skip static symbols.
10258 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
10260 PR tree-optimization/57200
10261 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
10262 Only call inform if the preceding warning_at returns true.
10264 2013-05-07 Han Shen <shenhan@google.com>
10266 * cfgexpand.c (record_or_union_type_has_array_p): New function.
10267 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
10268 * common.opt (fstack-protector-strong): New option.
10269 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
10270 * doc/invoke.texi (Optimization Options): Document
10271 "-fstack-protector-strong".
10272 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
10274 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10276 * config/mips/mips.c (mips_machine_reorg2): Return 0.
10278 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
10280 * ira.c (update_equiv_regs): Add insn having equiv memory even if
10281 it is not lhs of the insn.
10282 (setup_reg_equiv): Remove insn having equiv memory which it is not
10284 * lra-constraints.c (process_address): Try to improve generation
10285 code for address base + disp.
10286 (lra_constraints): Make correct the code for checking insn setting
10287 up backward equivalence. Remove insn only if it is in the init
10289 * lra-eliminations.c (update_reg_eliminate): Change return value.
10290 (lra_eliminate): Use the result.
10292 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
10294 * config/i386/sse.md (ssescalarnummask): New mode attribute.
10295 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
10296 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
10297 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
10298 register target operands.
10299 (*vec_extractv8hi_sse2): New pattern.
10300 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
10301 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
10302 (*vec_extract<mode>_mem): New insn and split pattern.
10304 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
10306 * config/arm/arm.c (arm_asan_shadow_offset): New function.
10307 (TARGET_ASAN_SHADOW_OFFSET): Define.
10308 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
10309 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
10311 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10313 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
10314 (incr_vec_index): Return -1 if increment not found.
10315 (create_add_on_incoming_edge): Assert if increment not found.
10316 (record_increment): Limit number of increments recorded.
10317 (all_phi_incrs_profitable): Return false if an increment not found.
10318 (replace_profitable_candidates): Don't process increments that were
10320 (analyze_candidates_and_replace): Limit size of incr_vec.
10322 2013-05-07 Richard Biener <rguenther@suse.de>
10324 * calls.c (special_function_p): setjmp-like functions are leaf.
10325 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
10326 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
10328 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
10330 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
10331 (aarch64_simd_mov<mode>): New expander.
10332 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
10333 (aarch64_simd_mov_to_<mode>high): Likewise.
10334 (aarch64_simd_mov_from_<mode>low): Likewise.
10335 (aarch64_simd_mov_from_<mode>high): Likewise.
10336 (aarch64_dup_lane<mode>): Update.
10337 (aarch64_dup_lanedi): New instruction pattern.
10338 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
10339 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
10341 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10343 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
10344 (replace_mult_candidate): Remove unnecessary argument; remove
10345 unnecessary parameter from call to introduce_cast_before_cand.
10346 (replace_unconditional_candidate): Remove unnecessary parameter
10347 from call to replace_mult_candidate.
10348 (replace_conditional_candidate): Likewise.
10349 (insert_initializers): Use make_temp_ssa_name.
10350 (introduce_cast_before_cand): Remove unnecessary argument; use
10351 make_temp_ssa_name.
10352 (replace_one_candidate): Remove unnecessary argument; remove
10353 unnecessary parameter from calls to introduce_cast_before_cand.
10354 (replace_profitable_candidates): Remove unnecessary parameters
10355 from calls to replace_one_candidate.
10357 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10359 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
10360 phi def as possibly hiding a basis for a CAND_ADD whose operands
10361 have been commuted in the analysis.
10362 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
10364 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10366 * config/aarch64/aarch64.md
10367 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
10368 shift value between 0-4.
10370 2013-05-07 Richard Biener <rguenther@suse.de>
10372 * double-int.h (rshift): New overload.
10373 * double-int.c (rshift): New function.
10374 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
10375 (create_reference_ops_from_ref): Remove.
10376 (vn_reference_insert): Use shared ops for constructing the
10377 reference and copy it.
10379 2013-05-07 Richard Biener <rguenther@suse.de>
10381 PR middle-end/57190
10382 * tree-eh.c (sink_clobbers): Properly propagate
10383 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
10385 2013-05-07 Jakub Jelinek <jakub@redhat.com>
10387 PR tree-optimization/57149
10388 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
10389 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
10390 collect_phi_def_edges, execute_late_warn_uninitialized): Use
10391 uninit_undefined_value_p instead of ssa_undefined_value_p.
10394 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
10395 for modifier == EXPAND_INITIALIZER.
10397 2013-05-07 Anton Blanchard <anton@samba.org>
10399 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
10400 for powerpc64 little endian.
10401 * configure: Regenerate.
10403 2013-05-06 Graham Stott <grahams@btinternet.com>
10405 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
10406 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
10407 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
10408 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
10410 2013-05-06 Graham Stott <grahams@btinternet.com>
10412 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
10413 codes which allow non-lvalues.
10415 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
10417 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
10418 components are all 1s.
10419 (integer_minus_onep): New function.
10420 * tree.h (integer_minus_onep): Declare it.
10421 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
10422 integer_minus_onep instead of integer_all_onesp.
10424 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10427 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
10428 variations of these patterns.
10430 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
10432 * config/i386/i386.md (isa): Add x64_sse4 member.
10433 (enabled): Handle x64_sse4.
10434 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
10435 instruction for 64bit SSE4_1 targets. Update insn attributes.
10436 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
10437 instruction for SSE4_1 targets. Update insn attributes.
10438 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
10439 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
10440 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
10442 (*vec_extractv4si): Rename from *sse4_1_pextrd.
10443 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
10444 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
10446 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10449 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
10451 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
10453 * final.c (do_assembler_dialects): Don't handle curly braces and
10454 vertical bar escaped by % as dialect delimiters.
10455 (output_asm_insn): Print curly braces and vertical bar if escaped
10456 by % and ASSEMBLER_DIALECT defined.
10457 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
10458 * doc/tm.texi: Regenerated.
10460 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10462 * config/mips/mips.c: Include tree-pass.h.
10463 (mips_reorg): Split in pre- and post-dbr_schedule parts.
10464 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
10465 (pass_mips_machine_reorg2): New machine specific pass.
10466 (insert_pass_mips_machine_reorg2): New pass plugin definition.
10467 (mips_option_override): Register the new pass.
10468 * rtl.h (cleanup_barriers): Remove prototype.
10469 (dbr_schedule): Likewise.
10470 * jump.c (cleanup_barriers): Make static.
10471 * reorg.c (dbr_schedule): Likewise.
10473 2013-05-06 Richard Biener <rguenther@suse.de>
10475 PR tree-optimization/57185
10476 * tree-parloops.c (add_field_for_reduction): Handle anonymous
10477 SSA names properly.
10479 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
10482 * config/i386/i386.c (add_parameter_dependencies): Add dependence
10483 between "first_arg" and "insn", not "last" and "insn".
10485 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10487 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
10488 (find_candidates_in_block): Re-enable slsr_process_phi.
10489 (create_phi_basis): Fix double counting of candidate adjustment.
10491 2013-05-06 Richard Biener <rguenther@suse.de>
10493 PR middle-end/57147
10494 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
10495 the edge is also fallthru, preserve it and just clear the
10497 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
10498 also complex, preserve that and just clear the fallthru flag.
10499 * tree-inline.c (update_ssa_across_abnormal_edges): Also
10500 update virtual operands.
10502 2013-05-06 Alan Modra <amodra@gmail.com>
10504 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
10505 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
10506 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
10507 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
10508 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
10509 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
10511 2013-05-06 Alan Modra <amodra@gmail.com>
10513 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
10514 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
10515 (DEFAULT_ASM_ENDIAN): Define.
10516 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
10517 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
10518 Update -K PIC clause from sysv4.h.
10519 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
10520 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
10522 2013-05-06 Alan Modra <amodra@gmail.com>
10524 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
10525 twice for little-endian.
10526 (ashrdi3_no_power, ashrdi3): Support little-endian.
10528 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10531 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
10532 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
10534 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
10535 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
10536 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
10539 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
10540 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
10542 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
10543 * config/pa/pa.opt: Make mbig-switch a no-op.
10544 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
10545 (CASE_VECTOR_MODE): Always return SImode.
10546 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
10547 for the !TARGET_BIG_SWITCH case.
10548 * config/pa/pa-linux.h: Likewise.
10549 * config/pa/pa-openbsd.h: Likewise.
10550 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
10551 * config/pa/pa.md (short_jump): Remove define_insn.
10552 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
10553 (casesi0): Remove define_insn.
10554 (type): Remove btable_branch.
10555 (pa_combine_type): Likewise.
10556 (in_nullified_branch_delay): Likewise.
10557 (in_call_delay): Likewise.
10558 (define_delay): Likewise.
10559 (define_insn_reservation "Z3"): Likewise.
10560 (define_insn_reservation "Z4"): Likewise.
10561 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
10562 (pa_adjust_insn_length): Remove adjustment for btable branches.
10563 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
10566 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
10568 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
10569 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
10570 Add m->r,x alternatives.
10571 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
10572 splitters using SWI48x mode iterator.
10573 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
10574 TARGET_64BIT. Add m->x alternative.
10575 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
10576 Add o->x alternative. Enable for TARGET_SSE.
10577 (sse_storeq): Remove expander.
10578 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
10579 with memory input operand.
10580 (*vec_extractv2di_1 splitter): New.
10581 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
10582 * config/i386/i386.md (ssevecmodelower): New mode attribute.
10584 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
10586 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
10587 (INT_LOWPART): Delete.
10588 (extract_MB): Adjust.
10589 (extract_ME): Adjust.
10590 (print_operand): Adjust.
10592 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
10594 * config/rs6000/predicates.md (reg_or_add_cint_operand,
10595 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
10596 (reg_or_logical_cint_operand, easy_fp_constant,
10597 logical_const_operand): Delete "CONST_DOUBLE" case.
10598 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
10599 "HOST_BITS_PER_WIDE_INT == 64" test.
10600 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
10601 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
10602 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
10603 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
10605 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
10606 CONST_DOUBLE DImode/VOIDmode case.
10607 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
10608 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
10609 CONST_DOUBLE VOIDmode case.
10610 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
10611 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
10612 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
10613 Delete CONST_DOUBLE case.
10614 (splitters for mov FMOVE64 const_double): Delete
10615 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
10616 "HOST_BITS_PER_WIDE_INT >= 64" test.
10617 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
10619 (mov DI const_double): Delete.
10621 2013-05-04 Jakub Jelinek <jakub@redhat.com>
10623 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
10624 on op shows all bits zero in mode of a lowpart subreg, return zero.
10626 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
10629 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
10630 to save TFmode registers and DImode to save TImode registers for
10631 caller save operations.
10632 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
10633 mark being partially clobbered since they only use the first
10636 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
10637 and TDmode only use the upper 64-bits of each VSX register.
10639 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10641 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
10642 (find_candidates_in_block): Disable slsr_process_phi.
10644 2013-05-03 Guozhi Wei <carrot@google.com>
10646 * coverage.c (coverage_obj_init): Move the construction of gcov
10648 (build_init_ctor): ... here.
10650 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10652 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
10653 (slsr_cand_d): Redefine def_phi.
10654 (stride_status, phi_adjust_status, count_phis_status): New enums.
10655 (find_phi_def): New.
10656 (find_basis_for_base_expr): New.
10657 (find_basis_for_candidate): Handle hidden bases.
10658 (alloc_cand_and_find_basis): Handle phi candidates.
10659 (slsr_process_phi): New.
10660 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
10661 (create_mul_imm_cand): Likewise.
10662 (create_add_ssa_cand): Exclude phi base candidates.
10663 (create_add_imm_cand): Likewise.
10664 (slsr_process_cast): Likewise.
10665 (slsr_process_copy): Likewise.
10666 (find_candidates_in_block): Handle phi candidates.
10667 (dump_candidate): Likewise.
10668 (unconditional_cands): Delete.
10669 (unconditional_cands_with_known_stride_p): Delete.
10670 (phi_dependent_cand_p): New.
10671 (cand_increment): Handle phi-dependent candidates.
10672 (replace_dependent): Delete.
10673 (replace_mult_candidate): New.
10674 (replace_unconditional_candidate): New.
10675 (incr_vec_index): Move to avoid forward reference.
10676 (create_add_on_incoming_edge): New.
10677 (create_phi_basis): New.
10678 (replace_dependents): Delete.
10679 (replace_conditional_candidate): New.
10680 (phi_add_costs): New.
10681 (replace_uncond_cands_and_profitable_phis): New.
10682 (record_increment): Handle phi adjustments.
10683 (record_phi_increments): New.
10684 (record_increments): Handle phi adjustments.
10685 (phi_incr_cost): New.
10686 (lowest_cost_path): Handle phis.
10687 (total_savings): Likewise.
10688 (analyze_increments): Likewise.
10689 (ncd_with_phi): New.
10690 (ncd_of_cand_and_phis): New.
10691 (nearest_common_dominator_for_cands): Handle phi increments.
10692 (all_phi_incrs_profitable): New.
10693 (replace_profitable_candidates): Handle phi-dependent candidates.
10694 (analyze_candidates_and_replace): Likewise.
10696 2013-05-03 Teresa Johnson <tejohnson@google.com>
10699 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
10700 do not exceed REG_BR_PROB_BASE.
10702 2013-05-03 Jeff Law <law@redhat.com>
10704 PR tree-optimization/57144
10705 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
10706 operand of the condition will bit into the new type when eliminating
10707 a cast feeding a condition.
10709 2013-05-03 Jakub Jelinek <jakub@redhat.com>
10711 PR rtl-optimization/57130
10712 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
10713 of COMPARE as in_code to the recursive call if needed.
10715 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
10717 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
10718 (enabled): Handle new members.
10719 * config/i386/sse.md (*vec_concatv2si): Merge from
10720 *vec_concatv2si_sse2 and vec_concatv2si_sse.
10721 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
10723 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
10725 PR tree-optimization/57027
10726 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
10727 for fnms opportunity, check we got the prerequisite kind
10728 of tree / gimple before using accessor functions.
10730 2013-05-03 Richard Biener <rguenther@suse.de>
10732 * double-int.h (lshift): New overload without precision
10733 and arith argument.
10734 (operator *=, operator +=, operator -=): Move ...
10735 * double-int.c (operator *=, operator +=, operator -=): ... here
10736 and implement more efficiently.
10737 (mul_double_with_sign): Remove.
10738 (lshift_double): Adjust to take unsinged shift argument, push
10739 dispatching code to callers.
10740 (mul_double_wide_with_sign): Add early out for callers that
10741 are not interested in high parts or overflow.
10742 (lshift): New function.
10743 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
10744 dispatch code here.
10745 (lrotate, rrotate): Use logical shifts.
10746 * expr.c (get_inner_reference): Use lshift.
10747 * fixed-value.c (do_fixed_divide): Likewise.
10748 * tree-dfa.c (get_ref_base_and_extent): Likewise.
10749 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
10750 (indirect_refs_may_alias_p): Likewise.
10751 (stmt_kills_ref_p_1): Likewise.
10753 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
10755 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
10757 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
10759 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
10760 scalar form of FABD instruction.
10762 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
10764 * lra-constraints.c (process_alt_operands): Add checking alt
10765 number to choose the best alternative.
10767 2013-05-02 Richard Biener <rguenther@suse.de>
10769 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
10770 bitmap and its handling.
10771 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
10773 2013-05-02 Richard Biener <rguenther@suse.de>
10775 PR middle-end/57140
10776 * tree-inline.c (copy_loops): Properly handle removed loops.
10777 (copy_cfg_body): Mark destination loops for fixup if source
10778 loops needed fixup.
10780 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
10783 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
10784 generating simple_return for naked functions.
10786 2013-05-02 Martin Jambor <mjambor@suse.cz>
10788 PR middle-end/56988
10789 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
10790 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
10792 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
10793 ipa_agg_replacement_value structures.
10794 (known_aggs_to_agg_replacement_list): Likewise.
10795 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
10796 (read_agg_replacement_chain): Likewise.
10797 (ipcp_transform_function): Also check that by_ref flags match.
10799 2013-05-02 Richard Biener <rguenther@suse.de>
10801 * graphds.h (struct graph): Add obstack member.
10802 * graphds.c (new_graph): Initialize obstack and allocate
10804 (add_edge): Allocate edge from the obstack.
10805 (free_graph): Free the obstack instead of all edges and vertices.
10807 2013-05-02 Teresa Johnson <tejohnson@google.com>
10809 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
10811 * cfg.c (update_bb_profile_for_threading): Ditto.
10812 * tree-inline.c (copy_bb): Ditto.
10813 (copy_edges_for_bb): Ditto.
10814 (initialize_cfun): Ditto.
10815 (copy_cfg_body): Ditto.
10816 (expand_call_inline): Ditto.
10817 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
10818 (estimate_node_size_and_time): Ditto.
10819 (inline_merge_summary): Ditto.
10820 * cgraphclones.c (cgraph_clone_edge): Ditto.
10821 (cgraph_clone_node): Ditto.
10822 * sched-rgn.c (compute_dom_prob_ps): Ditto.
10823 (compute_trg_info): Ditto.
10825 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10827 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
10828 S reg when fp attribute set.
10829 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
10831 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10833 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
10835 (*and_one_cmplsi3_compare0_uxtw): Likewise.
10836 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
10837 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
10839 2013-05-02 Richard Biener <rguenther@suse.de>
10841 * tree-scalar-evolution.c (scev_info_hasher): Remove.
10842 (struct instantiate_cache_entry): New type.
10843 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
10844 (struct instantiate_cache_type): New type.
10845 (set_instantiated_value, get_instantiated_value): Remove.
10846 (get_instantiated_value_entry): New function.
10847 (instantiate_scev_name): Use the new cache and adjust.
10848 (instantiate_scev_poly): Adjust.
10849 (instantiate_scev_binary): Likewise.
10850 (instantiate_array_ref): Likewise.
10851 (instantiate_scev_convert): Likewise.
10852 (instantiate_scev_not): Likewise.
10853 (instantiate_scev_3): Likewise.
10854 (instantiate_scev_2): Likewise.
10855 (instantiate_scev_r): Likewise.
10856 (instantiate_scev): Likewise.
10857 (resolve_mixers): Likewise.
10859 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
10862 * lra-constraints.c (best_small_class_operands_num): Remove.
10863 (process_alt_operands): Remove small_class_operands_num. Take
10864 small classes operands into losers and only if the operand is not
10865 matched. Modify debugging output.
10866 (curr_insn_transform): Remove best_small_class_operands_num.
10869 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10871 * config/aarch64/aarch64-builtins.c
10872 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
10873 * config/aarch64/aarch64-simd-builtins.def
10874 (reduc_splus_): Add new modes.
10875 (reduc_uplus_): New.
10876 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
10877 (reduc_uplus_v4sf): Likewise.
10878 (reduc_splus_v4sf): Likewise.
10879 (aarch64_addv<mode>): Likewise.
10880 (reduc_uplus_<mode>): Likewise.
10881 (reduc_splus_<mode>): Likewise.
10882 (aarch64_addvv2di): Likewise.
10883 (reduc_uplus_v2di): Likewise.
10884 (reduc_splus_v2di): Likewise.
10885 (aarch64_addvv2si): Likewise.
10886 (reduc_uplus_v2si): Likewise.
10887 (reduc_splus_v2si): Likewise.
10888 (reduc_<sur>plus_<mode>): New.
10889 (reduc_<sur>plus_v2di): Likewise.
10890 (reduc_<sur>plus_v2si): Likewise.
10891 (reduc_<sur>plus_v4sf): Likewise.
10892 (aarch64_addpv4sf): Likewise.
10893 * config/aarch64/arm_neon.h
10894 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
10895 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
10896 add UNSPEC_SADDV, UNSPEC_UADDV.
10898 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
10900 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10902 * config/aarch64/arm_neon.h
10903 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
10905 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10907 * config/aarch64/aarch64-builtins
10908 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
10910 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10912 * config/aarch64/aarch64-simd-builtins.def
10913 (reduc_smax_): New.
10914 (reduc_smin_): Likewise.
10915 (reduc_umax_): Likewise.
10916 (reduc_umin_): Likewise.
10917 (reduc_smax_nan_): Likewise.
10918 (reduc_smin_nan_): Likewise.
10921 (smax): Update for V2SF, V4SF and V2DF modes.
10924 (smin_nan): Likewise.
10925 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10926 (<su><maxmin><mode>3): ...This, refactor.
10927 (s<maxmin><mode>3): New.
10928 (<maxmin_uns><mode>3): Likewise.
10929 (reduc_<maxmin_uns>_<mode>): Refactor.
10930 (reduc_<maxmin_uns>_v4sf): Likewise.
10931 (reduc_<maxmin_uns>_v2si): Likewise.
10932 (aarch64_<fmaxmin><mode>: Remove.
10933 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10935 (vmin<q>_f<32,64>): Likewise.
10936 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10938 (su): Add mappings for smax, smin, umax, umin.
10940 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10941 (FMAXMIN): Rename as...
10942 (FMAXMIN_UNS): ...This.
10944 (fmaxminv): Likewise.
10945 (fmaxmin): Likewise.
10947 (maxmin_uns_op): Likewise.
10949 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10951 * config/aarch64/arm_neon.h
10952 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10953 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10954 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10956 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10958 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10959 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10961 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10963 * config/aarch64/aarch64-simd.md
10964 (vcond<mode>_internal): Handle special cases for constant masks.
10965 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10966 (vcondu<mode><mode>): Likewise.
10967 (vcond<v_cmp_result><mode>): New.
10969 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10971 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10972 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10973 * config/aarch64/aarch64-simd-builtins.def
10974 (cmeq): Update to BUILTIN_VALLDI.
10979 * config/aarch64/arm_neon.h
10980 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10981 to builtins or C as appropriate.
10983 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10985 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10987 (cmhi): Rename to...
10989 * config/aarch64/aarch64-simd.md
10990 (simd_mode): Add SF.
10991 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10992 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10993 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10994 (cstore<mode>_neg): ...This.
10995 * config/aarch64/iterators.md
10997 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10998 (COMPARISONS): New.
10999 (UCOMPARISONS): Likewise.
11000 (optab): Add missing comparisons.
11006 (VCMP_S): Likewise.
11007 (VCMP_U): Likewise.
11008 (V_cmp_result): Add DF, SF modes.
11009 (v_cmp_result): Likewise.
11011 (vmtype): Likewise.
11012 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
11014 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
11016 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
11017 define_insn to define_insn_and_split.
11018 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
11019 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
11020 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
11021 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
11022 (thumb2_negscc): Likewise.
11024 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11026 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
11028 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11030 * config/arm/thumb2.md: Remove trailing whitespaces.
11032 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11034 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
11035 Use gen_int_mode rather than GEN_INT.
11037 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
11039 * value-prof.c (stream_in_histogram_value): Remove the strayed
11042 2013-04-30 Richard Biener <rguenther@suse.de>
11044 PR middle-end/57122
11045 * cfghooks.c (split_edge): Properly check for the loop latch edge.
11047 2013-04-30 Richard Biener <rguenther@suse.de>
11049 PR middle-end/57107
11050 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
11052 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
11054 PR rtl-optimization/56957
11055 PR rtl-optimization/57105
11056 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
11057 variable. Use just INSN_UID for determining whether an insn
11058 should be only disconnected from the insn stream.
11059 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
11061 2013-04-30 Jakub Jelinek <jakub@redhat.com>
11063 PR tree-optimization/57104
11064 * tsan.c (instrument_expr): Don't instrument accesses to
11065 DECL_HARD_REGISTER VAR_DECLs.
11067 2013-04-30 Richard Biener <rguenther@suse.de>
11069 * function.h (loops_for_fn): New inline function.
11070 (set_loops_for_fn): Likewise.
11071 * cfgloop.h (place_new_loop): Add struct function parameter.
11072 (get_loop): Likewise.
11073 (get_loops): Likewise.
11074 (number_of_loops): Likewise.
11075 (fel_next): Adjust.
11076 (fel_init): Likewise.
11077 * cfg.c (get_loop_copy): Adjust.
11078 * cfgloop.c (flow_loops_dump): Likewise.
11079 (record_loop_exits): Likewise.
11080 (verify_loop_structure): Likewise.
11081 * cfgloopanal.c (mark_irreducible_loops): Likewise.
11082 (estimate_reg_pressure_cost): Likewise.
11083 (mark_loop_exit_edges): Likewise.
11084 * cfgloopmanip.c (place_new_loop): Likewise.
11085 (add_loop): Likewise.
11086 (duplicate_loop): Likewise.
11087 * graph.c (draw_cfg_nodes): Likewise.
11088 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
11089 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
11090 (extract_affine_chrec): Likewise.
11091 (build_scop_iteration_domain): Likewise.
11092 * graphite.c (graphite_initialize): Likewise.
11093 * ira-build.c (create_loop_tree_nodes): Likewise.
11094 (more_one_region_p): Likewise.
11095 (rebuild_regno_allocno_maps): Likewise.
11096 (mark_loops_for_removal): Likewise.
11097 (mark_all_loops_for_removal): Likewise.
11098 (remove_unnecessary_regions): Likewise.
11099 (ira_build): Likewise.
11100 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
11101 * loop-init.c (fix_loop_structure): Likewise.
11102 (gate_rtl_move_loop_invariants): Likewise.
11103 (gate_rtl_unswitch): Likewise.
11104 (gate_rtl_unroll_and_peel_loops): Likewise.
11105 (rtl_doloop): Likewise.
11106 * lto-streamer-in.c (input_cfg): Likewise.
11107 * lto-streamer-out.c (output_cfg): Likewise.
11108 * modulo-sched.c (sms_schedule): Likewise.
11109 * predict.c (tree_estimate_probability): Likewise.
11110 (tree_estimate_probability_driver): Likewise.
11111 (estimate_loops): Likewise.
11112 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
11113 (move_sese_region_to_fn): Likewise.
11114 (debug_loop_num): Likewise.
11115 * tree-chrec.c (chrec_evaluate): Likewise.
11116 (hide_evolution_in_other_loops_than_loop): Likewise.
11117 (chrec_component_in_loop_num): Likewise.
11118 (reset_evolution_in_loop): Likewise.
11119 (evolution_function_is_invariant_rec_p): Likewise.
11120 * tree-if-conv.c (main_tree_if_conversion): Likewise.
11121 * tree-inline.c (copy_loops): Likewise.
11122 (copy_cfg_body): Likewise.
11123 (tree_function_versioning): Likewise.
11124 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
11125 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
11127 (add_to_evolution_1): Likewise.
11128 (scev_const_prop): Likewise.
11129 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
11130 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
11131 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
11132 (tree_ssa_lim_initialize): Likewise.
11133 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
11134 (verify_loop_closed_ssa): Likewise.
11135 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
11136 (tree_ssa_loop_im): Likewise.
11137 (tree_ssa_loop_unswitch): Likewise.
11138 (tree_vectorize): Likewise.
11139 (check_data_deps): Likewise.
11140 (tree_ssa_loop_ivcanon): Likewise.
11141 (tree_ssa_loop_bounds): Likewise.
11142 (tree_complete_unroll): Likewise.
11143 (tree_complete_unroll_inner): Likewise.
11144 (tree_parallelize_loops): Likewise.
11145 (tree_ssa_loop_prefetch): Likewise.
11146 (tree_ssa_loop_ivopts): Likewise.
11147 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
11148 * tree-vectorizer.c (vectorize_loops): Likewise.
11150 2013-04-29 Mike Frysinger <vapier@gentoo.org>
11152 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
11153 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
11154 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
11155 with EABI_LINK_SPEC.
11157 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11160 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
11163 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
11166 * lra-constraints.c (process_alt_operands): Discourage a bit more
11167 using memory for pseudos. Print cost dump for alternatives.
11168 Modify cost values for conflicts with early clobbers.
11169 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
11171 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11174 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
11176 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11178 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
11179 from/to S register.
11180 (movdi_aarch64): Support LDR/STR from/to D register.
11182 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11184 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
11185 or higher by default.
11187 2013-04-29 Richard Biener <rguenther@suse.de>
11189 PR middle-end/57075
11190 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
11191 even if not adding abnormal edges for calls that can make
11194 2013-04-29 Richard Biener <rguenther@suse.de>
11196 PR middle-end/57103
11197 * tree-cfg.c (move_stmt_op): Fix condition under which to update
11199 (move_stmt_r): Remove redundant checking.
11201 2013-04-29 Teresa Johnson <tejohnson@google.com>
11204 * basic-block.h (apply_scale): New function.
11205 (apply_probability): Use apply_scale.
11206 * gimple-streamer-in.c (input_bb): Ditto.
11207 * lto-streamer-in.c (input_cfg): Ditto.
11208 * lto-cgraph.c (merge_profile_summaries): Ditto.
11209 * tree-optimize.c (execute_fixup_cfg): Ditto.
11210 * tree-inline.c (copy_bb): Update comment to use apply_scale.
11211 (copy_edges_for_bb): Ditto.
11212 (copy_cfg_body): Ditto.
11214 2013-04-29 Tom de Vries <tom@codesourcery.com>
11216 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
11217 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
11218 (tail_merge_optimize): Handle current_loops == NULL.
11220 2013-04-26 Jeff Law <law@redhat.com>
11222 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
11223 (simplify_cond_using_ranges): Generalize code to simplify
11224 COND_EXPRs where one argument is a constant and the other
11225 is an SSA_NAME created by an integral type conversion.
11227 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11229 * config/arm/arm.md (store_minmaxsi): Use only when
11230 optimize_insn_for_size_p.
11232 2013-04-29 Christian Bruel <christian.bruel@st.com>
11235 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
11237 2013-04-29 Richard Biener <rguenther@suse.de>
11239 PR middle-end/57089
11240 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
11241 loop tree make sure to schedule a fixup for the child as well.
11242 (expand_omp_for_generic): Properly add loops.
11243 (expand_omp_for_static_nochunk): Likewise.
11244 (expand_omp_for_static_chunk): Likewise.
11245 (expand_omp_for): For the degenerate case fixup loops.
11246 (expand_omp_sections): Fix default bb placement in loops.
11247 (expand_omp_atomic_pipeline): Properly add loops.
11249 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11251 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
11253 2013-04-29 Tom de Vries <tom@codesourcery.com>
11255 * tree-ssa-tail-merge.c: Update header comment.
11257 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11259 * config/aarch64/arm_neon.h
11260 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
11261 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
11262 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
11263 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
11264 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
11265 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
11266 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
11267 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
11269 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11271 * config/aarch64/aarch64-simd.md
11272 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
11273 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
11274 fix_trunc, fixuns_trunc.
11275 (ftrunc<VDQF:mode>2): New.
11276 * config/aarch64/iterators.md (optab): Add fix, fixuns.
11277 (fix_trunc_optab): New.
11279 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11281 * config/aarch64/aarch64-builtins.c
11282 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
11283 iceilf, lround, iroundf.
11285 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11288 * config/i386/i386.h (enum ix86_tune_indices)
11289 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
11290 New, split from X86_TUNE_INTER_UNIT_MOVES.
11291 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
11292 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
11293 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
11294 (TARGET_INTER_UNIT_MOVES): Remove.
11295 * config/i386/i386.c (initial_ix86_tune_features): Update.
11296 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
11297 (ix86_expand_convert_uns_didf_sse): Use
11298 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11299 (ix86_expand_vector_init_one_nonzero): Ditto.
11300 (ix86_expand_vector_init_interleave): Ditto.
11301 (inline_secondary_memory_needed): Return true for moves from SSE class
11302 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
11303 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
11304 * config/i386/constraints.md (Yi, Ym): Depend on
11305 TARGET_INTER_UNIT_MOVES_TO_VEC.
11306 (Yj, Yn): New constraints.
11307 * config/i386/i386.md (*movdi_internal): Change constraints of
11308 operand 1 from Yi to Yj and from Ym to Yn.
11309 (*movsi_internal): Ditto.
11310 (*movdf_internal): Ditto.
11311 (*movsf_internal): Ditto.
11312 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
11313 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11314 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
11315 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
11316 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
11317 * config/i386/sse.md (movdi_to_sse): Ditto.
11318 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
11319 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
11320 TARGET_INTER_UNIT_MOVES.
11321 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
11322 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
11323 instead of TARGET_INTER_UNIT_MOVES.
11324 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
11325 operand 1 from Yi to Yj and from Ym to Yn.
11327 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11329 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
11330 (float_truncate_hi_): Likewise.
11331 (float_extend_lo_): Likewise.
11332 (float_truncate_lo_): Likewise.
11333 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
11334 (aarch64_float_extend_lo_v2df): Likewise.
11335 (vec_unpacks_hi_v4sf): Likewise.
11336 (aarch64_float_truncate_lo_v2sf): Likewise.
11337 (aarch64_float_truncate_hi_v4sf): Likewise.
11338 (vec_pack_trunc_v2df): Likewise.
11339 (vec_pack_trunc_df): Likewise.
11341 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11343 * config/aarch64/aarch64-builtins.c
11344 (aarch64_fold_builtin): Fold float conversions.
11345 * config/aarch64/aarch64-simd-builtins.def
11346 (floatv2si, floatv4si, floatv2di): New.
11347 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
11348 * config/aarch64/aarch64-simd.md
11349 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
11350 * config/aarch64/iterators.md (FLOATUORS): New.
11351 (optab): Add float, floatuns.
11352 (su_optab): Likewise.
11354 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11356 * config/aarch64/aarch64-builtins.c
11357 (aarch64_builtin_vectorized_function): Use new names for
11359 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
11360 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
11361 (fcvtzu): Split as...
11362 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
11363 (fcvtas): Split as...
11364 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
11365 (fcvtau): Split as...
11366 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
11367 (fcvtps): Split as...
11368 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
11369 (fcvtpu): Split as...
11370 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
11371 (fcvtms): Split as...
11372 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
11373 (fcvtmu): Split as...
11374 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
11375 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
11376 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
11377 (lfrintnusf, lfrintnudf): Likewise.
11378 * config/aarch64/aarch64-simd.md
11379 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
11381 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
11382 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
11383 (fcvt_pattern): Likewise.
11385 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11387 * config/aarch64/aarch64-simd.md
11388 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
11389 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
11391 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11393 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
11394 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
11395 (vrnd<a,m,n,p>_f32): Implement using builtins.
11396 (vrnd<i,x><q>_f<32, 64>): New.
11398 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11400 * config/aarch64/aarch64-builtins.c
11401 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
11402 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
11403 (frintz): Rename to...
11405 (frintp): Rename to...
11407 (frintm): Rename to...
11409 (frinti): Rename to...
11410 (nearbyint): ...this.
11411 (frintx): Rename to...
11413 (frinta): Rename to...
11415 * config/aarch64/aarch64-simd.md
11416 (aarch64_frint<frint_suffix><mode>): Delete.
11417 (<frint_pattern><mode>2): Convert to insn.
11418 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
11419 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
11420 (frint_pattern): Likewise.
11421 (frint_suffix): Likewise.
11423 2013-04-29 Richard Biener <rguenther@suse.de>
11425 PR tree-optimization/57081
11426 * loop-init.c: Include tree-flow.h.
11427 (loop_optimizer_finalize): Free number of iteration estimates.
11428 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
11430 2013-04-29 Jakub Jelinek <jakub@redhat.com>
11432 PR tree-optimization/57083
11433 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
11434 non-singleton shift count range, zero extend low_bound for uns case.
11436 * config/i386/predicates.md (general_vector_operand): New predicate.
11437 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
11438 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
11439 if they aren't nonimmediate operands. If their original values
11440 satisfy const_vector_equal_evenodd_p, don't shift them.
11441 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
11442 predicates. For the SSE4.1 case force operands[{1,2}] into registers
11443 if not nonimmediate_operand.
11444 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
11445 instead of register_operand.
11446 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
11448 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
11450 * stor-layout.c (finalize_size_functions): Allocate a structure and
11451 reset cfun before dumping the functions.
11453 2013-04-27 Jakub Jelinek <jakub@redhat.com>
11455 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
11458 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
11459 use xop_pmacsdqh if uns_p.
11460 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
11461 the immediate rotate count.
11463 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
11465 * rtl.h (struct rtx_def): Add comment for field jump.
11466 (LRA_SUBREG_P): New macro.
11467 * recog.c (register_operand): Check LRA_SUBREG_P.
11468 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11470 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11471 (lra_final_code_change): Skip subreg change for operators.
11472 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11473 if there are no operand changes.
11474 * lra-constraints.c (curr_insn_set): New.
11475 (match_reload): Set LRA_SUBREG_P.
11476 (emit_spill_move): Ditto.
11477 (check_and_process_move): Use curr_insn_set. Process only single
11478 set insns. Don't initialize sec_mem_p and change_p.
11479 (simplify_operand_subreg): Use LRA_SUBREG_P.
11480 (reg_in_class_p): New function.
11481 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11482 of #ifdef. Add code to remove cycling.
11483 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11484 non-null disp. Reload inner instead of disp when base and index
11485 are null. Try to put lo_sum into register.
11486 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11487 (check_and_process_move): Move code for move cost check to
11488 simple_move_p. Remove equiv_substitution.
11489 (simple_move_p): New function.
11490 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11491 curr_insn_set. Call check_and_process_move only for single set
11492 insns. Use the new function. Move call of check_and_process_move
11493 after operand equiv substitution and address process.
11495 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11498 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
11499 with nonlocal goto receivers or returns twice calls, ignore
11500 unininitialized values from abnormal edges to nl goto receiver
11501 or returns twice call.
11503 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11505 PR tree-optimization/57051
11506 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
11507 and VEC_RSHIFT_EXPR if shift count is a multiple of element
11510 2013-04-26 Richard Biener <rguenther@suse.de>
11512 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
11513 (expand_omp_taskreg): Likewise. Mark loops for fixup.
11514 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
11515 (fixup_loop_arrays_after_move): New function.
11516 (move_sese_region_to_fn): Properly outline the loop tree parts
11517 of the SESE region.
11519 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
11521 * config/i386/i386.md (type, unit): Fix long lines.
11523 2013-04-26 Richard Biener <rguenther@suse.de>
11525 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
11526 (lto-streamer-out.o): Likewise.
11527 * cfgloop.c (init_loops_structure): Export, add struct function
11528 argument and adjust.
11529 (flow_loops_find): Adjust.
11530 * cfgloop.h (enum loop_estimation): Add EST_LAST.
11531 (init_loops_structure): Declare.
11532 * lto-streamer-in.c: Include cfgloop.h.
11533 (input_cfg): Input the loop tree.
11534 * lto-streamer-out.c: Include cfgloop.h.
11535 (output_cfg): Output the loop tree.
11536 (output_struct_function_base): Do not drop PROP_loops.
11538 2013-03-26 Richard Biener <rguenther@suse.de>
11540 * tree-cfg.c (execute_build_cfg): Build the loop tree.
11541 (pass_build_cfg): Provide PROP_loops.
11542 (move_sese_region_to_fn): Remove loops that are outlined into fn
11544 * tree-inline.c: Include cfgloop.h.
11545 (initialize_cfun): Do not drop PROP_loops.
11546 (copy_loops): New function.
11547 (copy_cfg_body): Copy loop structure.
11548 (tree_function_versioning): Initialize destination loop tree.
11549 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
11550 (pass_parallelize_loops): Do IL verification.
11551 * loop-init.c (loop_optimizer_init): Fixup loops if required.
11552 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
11553 the CFG make sure we fixup loops as well.
11554 * tree-ssa-tail-merge.c: Include cfgloop.h.
11555 (replace_block_by): When merging loop latches mark loops for fixup.
11556 * lto-streamer-out.c (output_struct_function_base): Drop
11557 PROP_loops for now.
11558 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
11559 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
11560 * ipa-split.c: Include cfgloop.h.
11561 (split_function): Add the new return block to the loop tree root.
11562 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
11563 whether we have removed the forwarder block.
11564 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
11565 * cfgloop.h (place_new_loop): Declare.
11566 * cfgloopmanip.c (place_new_loop): Export.
11567 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
11568 (tree-switch-conversion.o): Likewise.
11569 (tree-complex.o): Likewise.
11570 (tree-inline.o): Likewise.
11571 (tree-ssa-tailmerge.o): Likewise.
11572 (ipa-split.o): Likewise.
11573 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
11574 (tree-ssa-copy.o): Likewise.
11575 * tree-switch-conversion.c: Include cfgloop.h
11576 (process_switch): If we emit a bit-test cascade, schedule loops
11578 * tree-complex.c: Include cfgloop.h.
11579 (expand_complex_div_wide): Properly add new basic-blocks to loops.
11580 * asan.c: Include cfgloop.h.
11581 (create_cond_insert_point): Properly add new basic-blocks to
11582 loops, schedule loop fixup.
11583 * cfgloop.c (verify_loop_structure): Check that looks are not
11585 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
11587 (expand_omp_for_generic): Likewise.
11588 (expand_omp_sections): Likewise.
11589 (expand_omp_atomic_pipeline): Schedule loops for fixup.
11590 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
11591 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
11592 is initialized, not when loops are present.
11593 * tree-parloops.c (parallelize_loops): Remove checking here.
11594 * passes.c (init_optimization_passes): Schedule a copy-propagation
11595 pass before complete unrolling of inner loops.
11597 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11599 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
11600 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
11601 (colorize_init): Add argument to _WIN32 version.
11602 * toplev.c: Include diagnostic-color.h.
11603 (process_options): Default to -fdiagnostics-color=auto if
11604 GCC_COLORS env var is in the environment.
11605 * common.opt (fdiagnostics-color=): Add Var and Init.
11606 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
11607 env var is in the environment, the default is auto rather than never.
11609 * diagnostic.h (file_name_as_prefix): Add context argument.
11610 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
11611 the string as locus.
11612 * langhooks.c (lhd_print_error_function): Adjust caller.
11614 2013-04-25 Lawrence Crowl <crowl@google.com>
11616 * var-tracking.c (shared_hash_def::htab):
11617 Change type to hash_table. Update dependent calls and types.
11619 2013-04-25 Lawrence Crowl <crowl@google.com>
11621 * Makefile.in: Update as needed below.
11623 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
11624 Move declaration to after the type's method definitons.
11626 * attribs.c (htab_t scoped_attributes::attribute_hash):
11627 Change type to hash_table. Update dependent calls and types.
11629 * bitmap.c (htab_t bitmap_desc_hash):
11630 Change type to hash_table. Update dependent calls and types.
11632 * cselib.c (htab_t cselib_hash_table):
11633 Change type to hash_table. Update dependent calls and types.
11635 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
11636 (hash_string_slot_node): Move implementation into lto-streamer.h
11637 struct string_slot_hasher.
11638 (eq_string_slot_node): Likewise.
11640 * data-streamer-out.c: Update output_block::string_hash_table
11641 dependent calls and types.
11643 * dwarf2cfi.c (htab_t trace_index):
11644 Change type to hash_table. Update dependent calls and types.
11646 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
11647 Change type to hash_table. Update dependent calls and types.
11648 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
11649 (htab_t optimize_external_refs::map): Likewise.
11650 (htab_t output_comp_unit::extern_map): Likewise.
11651 (htab_t output_comdat_type_unit::extern_map): Likewise.
11652 (htab_t output_macinfo::macinfo_htab): Likewise.
11653 (htab_t optimize_location_lists::htab): Likewise.
11654 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
11656 * except.c (htab_t ehspec_hash_type):
11657 Change type to hash_table. Update dependent calls and types.
11658 (assign_filter_values::ttypes): Likewise.
11659 (assign_filter_values::ehspec): Likewise.
11660 (sjlj_assign_call_site_values::ar_hash): Likewise.
11661 (convert_to_eh_region_ranges::ar_hash): Likewise.
11663 * gcse.c (htab_t pre_ldst_table):
11664 Change type to hash_table. Update dependent calls and types.
11666 * ggc-common.c (htab_t saving_htab):
11667 Change type to hash_table. Update dependent calls and types.
11668 (htab_t loc_hash): Likewise.
11669 (htab_t ptr_hash): Likewise.
11670 (call_count): Rename ggc_call_count.
11671 (call_alloc): Rename ggc_call_alloc.
11672 (loc_descriptor): Rename make_loc_descriptor.
11673 (add_statistics): Rename ggc_add_statistics.
11675 * ggc-common.c (saving_htab):
11676 Change type to hash_table. Update dependent calls and types.
11678 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
11679 (push_gimplify_context): Likewise.
11680 (pop_gimplify_context): Likewise.
11681 (struct gimple_temp_hash_elt): Added.
11682 (struct gimplify_hasher): Likewise.
11683 (struct gimplify_ctx.temp_htab):
11684 Change type to hash_table. Update dependent calls and types.
11686 * gimple-fold.c: Include gimplify-ctx.h.
11688 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
11689 Change type to hash_table. Update dependent calls and types.
11690 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
11691 avoid potential global name collision.
11693 * gimplify.c: Include gimplify-ctx.h.
11694 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
11695 (htab_t gimplify_ctx::temp_htab):
11696 Update dependent calls and types for new type hash_table.
11697 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
11698 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
11700 * gimplify-ctx.h: New.
11701 (struct gimple_temp_hash_elt): Move from gimplify.c.
11702 (class gimplify_hasher): New.
11703 (struct gimplify_ctx): Move from gimple.h.
11704 (htab_t gimplify_ctx::temp_htab):
11705 Change type to hash_table. Update dependent calls and types.
11707 * graphite-clast-to-gimple.c: Include graphite-htab.h.
11708 (htab_t ivs_params::newivs_index):
11709 Change type to hash_table. Update dependent calls and types.
11710 (htab_t ivs_params::params_index): Likewise.
11711 (htab_t print_generated_program::params_index): Likewise.
11712 (htab_t gloog::newivs_index): Likewise.
11713 (htab_t gloog::params_index): Likewise.
11715 * graphite.c: Include graphite-htab.h.
11716 4htab_t graphite_transform_loops::bb_pbb_mapping):
11717 Change type to hash_table. Update dependent calls and types.
11719 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
11720 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
11721 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
11723 * graphite-dependences.c: Include graphite-htab.h.
11724 (loop_is_parallel_p): Change hash table type of parameter.
11726 * graphite-htab.h: New.
11727 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
11728 (extern find_pbb_via_hash): Move from graphite-poly.h.
11729 (extern loop_is_parallel_p): Move from graphite-poly.h.
11730 (extern get_loop_body_pbbs): Move from graphite-poly.h.
11732 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
11733 (extern loop_is_parallel_p): Move to graphite-htab.h.
11734 (extern get_loop_body_pbbs): Move to graphite-htab.h.
11736 * haifa-sched.c (htab_t delay_htab):
11737 Change type to hash_table. Update dependent calls and types.
11738 (htab_t delay_htab_i2): Likewise.
11740 * ira-color.c (htab_t allocno_hard_regs_htab):
11741 Change type to hash_table. Update dependent calls and types.
11743 * ira-costs.c (htab_t cost_classes_htab):
11744 Change type to hash_table. Update dependent calls and types.
11746 * loop-invariant.c (htab_t merge_identical_invariants::eq):
11747 Change type to hash_table. Update dependent calls and types.
11749 * loop-iv.c (htab_t bivs):
11750 Change type to hash_table. Update dependent calls and types.
11752 * loop-unroll.c (htab_t opt_info::insns_to_split):
11753 Change type to hash_table. Update dependent calls and types.
11754 (htab_t opt_info::insns_with_var_to_expand): Likewise.
11756 * lto-streamer.h (struct string_slot): Move from data-streamer.h
11757 (struct string_slot_hasher): New.
11758 (htab_t output_block::string_hash_table):
11759 Change type to hash_table. Update dependent calls and types.
11761 * lto-streamer-in.c (freeing_string_slot_hasher): New.
11762 (htab_t file_name_hash_table):
11763 Change type to hash_table. Update dependent calls and types.
11765 * lto-streamer-out.c: Update output_block::string_hash_table dependent
11768 * lto-streamer.c (htab_t tree_htab):
11769 Change type to hash_table. Update dependent calls and types.
11771 * omp-low.c: Include gimplify-ctx.h.
11773 * passes.c (htab_t name_to_pass_map):
11774 Change type to hash_table. Update dependent calls and types.
11775 (pass_traverse): Rename to passes_pass_traverse.
11777 * plugin.c (htab_t event_tab):
11778 Change type to hash_table. Update dependent calls and types.
11780 * postreload-gcse.c (htab_t expr_table):
11781 Change type to hash_table. Update dependent calls and types.
11782 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
11784 * sese.c (debug_rename_map_1): Make extern.
11785 (htab_t copy_bb_and_scalar_dependences::rename_map):
11786 Change type to hash_table. Update dependent calls and types.
11788 * sese.h (extern debug_rename_map): Move to .c file.
11790 * store-motion.c (htab_t store_motion_mems_table):
11791 Change type to hash_table. Update dependent calls and types.
11793 * trans-mem.c (htab_t tm_new_mem_hash):
11794 Change type to hash_table. Update dependent calls and types.
11796 * tree-browser.c (htab_t TB_up_ht):
11797 Change type to hash_table. Update dependent calls and types.
11799 * tree-cfg.c (htab_t discriminator_per_locus):
11800 Change type to hash_table. Update dependent calls and types.
11802 * tree-complex.c: Include tree-hasher.h
11803 (htab_t complex_variable_components):
11804 Change type to hash_table. Update dependent calls and types.
11806 * tree-eh.c (htab_t finally_tree):
11807 Change type to hash_table. Update dependent calls and types.
11809 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
11810 struct int_tree_hasher.
11811 (extern int_tree_map_eq): Likewise.
11812 (uid_decl_map_hash): Removed.
11813 (extern decl_tree_map_eq): Likewise.
11815 * tree-hasher.h: New.
11816 (struct int_tree_hasher): New.
11817 (typedef int_tree_htab_type): New.
11819 * tree-inline.c: Include gimplify-ctx.h.
11821 * tree-mudflap.c: Include gimplify-ctx.h.
11823 * tree-parloops.c: Include tree-hasher.h.
11824 (htab_t eliminate_local_variables_stmt::decl_address):
11825 Change type to hash_table. Update dependent calls and types.
11826 (htab_t separate_decls_in_region::decl_copies): Likewise.
11828 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
11829 Change type to hash_table. Update dependent calls and types.
11831 * tree-sra.c (candidates):
11832 Change type to hash_table. Update dependent calls and types.
11834 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
11836 (int_tree_map_hash): Likewise.
11838 * tree-ssa-dom.c (htab_t avail_exprs):
11839 Change type to hash_table. Update dependent calls and types.
11841 * tree-ssa-live.c (var_map_base_init::tree_to_index):
11842 Change type to hash_table. Update dependent calls and types.
11844 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
11845 Change type to hash_table. Update dependent calls and types.
11847 * tree-ssa-phiopt.c (seen_ssa_names):
11848 Change type to hash_table. Update dependent calls and types.
11850 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
11851 Change type to hash_table. Update dependent calls and types.
11853 * tree-ssa-uncprop.c (equiv):
11854 Change type to hash_table. Update dependent calls and types.
11856 2013-04-25 Jakub Jelinek <jakub@redhat.com>
11858 PR rtl-optimization/57003
11859 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
11860 call note_stores with kill_clobbered_value callback again after
11861 killing regs_invalidated_by_call.
11863 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11865 * config/aarch64/aarch64-simd.md
11866 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
11867 (aarch64_simd_bsl<mode>): Likewise.
11868 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
11870 2013-04-25 Marek Polacek <polacek@redhat.com>
11872 PR tree-optimization/57066
11873 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
11875 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
11877 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
11879 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11881 * config/aarch64/aarch64-builtins.c
11882 (aarch64_fold_builtin): New.
11883 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
11884 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
11885 * config/aarch64/aarch64-simd-builtins.def (abs): New.
11886 * config/aarch64/arm_neon.h
11887 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
11889 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11890 Tejas Belagod <tejas.belagod@arm.com>
11892 * config/aarch64/aarch64-builtins.c
11893 (aarch64_gimple_fold_builtin): New.
11894 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
11895 * config/aarch64/aarch64-simd-builtins.def (addv): New.
11896 * config/aarch64/aarch64-simd.md (addpv4sf): New.
11897 (addvv4sf): Update.
11898 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
11900 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11902 * config/aarch64/aarch64.md
11903 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
11905 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11907 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
11908 (*ngcsi_uxtw): New pattern.
11910 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11911 Julian Brown <julian@codesourcery.com>
11913 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
11914 (TB_DREG): Add T_V4HF.
11915 (v4hf_UP): New macro.
11916 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11917 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11918 Handle initialisation of V4HF. Adjust initialisation of reinterpret
11920 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11921 (arm_vector_mode_supported_p): Handle V4HF.
11922 (arm_mangle_map): Handle V4HFmode.
11923 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11924 * config/arm/arm_neon_builtins.def: Add entries for
11925 vcvtv4hfv4sf, vcvtv4sfv4hf.
11926 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11927 (neon_vcvtv4hfv4sf): Likewise.
11928 * config/arm/neon-gen.ml: Handle half-precision floating point
11930 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11931 * config/arm/arm_neon.h: Regenerate.
11932 * config/arm/neon.ml (type elts): Add F16.
11933 (type vectype): Add T_float16x4, T_floatHF.
11934 (type vecmode): Add V4HF.
11935 (type features): Add Requires_FP_bit feature.
11936 (elt_width): Handle F16.
11937 (elt_class): Likewise.
11938 (elt_of_class_width): Likewise.
11939 (mode_of_elt): Refactor.
11940 (type_for_elt): Handle F16, fix error messages.
11941 (vectype_size): Handle T_float16x4.
11942 (vcvt_sh): New function.
11943 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11944 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11945 (string_of_mode): Handle V4HF.
11946 * doc/arm-neon-intrinsics.texi: Regenerate.
11948 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11950 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11951 format specifier in 'X' case.
11953 2013-04-25 Alan Modra <amodra@gmail.com>
11956 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11957 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11958 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11959 Repeat for many other rotate/shift and mask patterns using subregs.
11960 Name lshiftrt insns.
11961 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11962 on WORDS_BIG_ENDIAN.
11964 2013-04-25 Alan Modra <amodra@gmail.com>
11966 * config.gcc: Support little-endian powerpc-linux targets.
11967 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11968 (LINK_OS_LINUX_SPEC): Define.
11969 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11970 Preserve MASK_LITTLE_ENDIAN.
11971 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11972 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11973 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11974 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11975 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11976 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11977 Correct fp word order for little-endian. Don't shift toc entries
11978 smaller than a word for little-endian.
11979 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11980 (bswapdi2 splits): Correct low-part subreg for little-endian.
11981 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11982 low/high where such is correct only for be.
11983 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11984 little-endian for -mcall-aixdesc.
11986 2013-04-25 Alan Modra <amodra@gmail.com>
11988 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11989 replace_equiv_address_nv.
11991 2013-04-25 Alan Modra <amodra@gmail.com>
11993 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11995 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11998 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11999 * rtl.h (struct rtx_def): ...
12001 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12003 PR rtl-optimizations/57046
12004 * lra-constraints (split_reg): Set up lra_risky_transformations_p
12005 for multi-reg splits.
12007 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
12009 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
12011 2013-04-24 Sterling Augustine <saugustine@google.com>
12013 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
12014 (comp_dir_string, debug_str_dwo_section): New.
12015 (DEBUG_STR_DWO_SECTION): Rename to ...
12016 (DEBUG_DWO_STR_SECTION): ... this.
12017 (DEBUG_NORM_STR_SECTION): Delete.
12018 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
12019 (DEBUG_STR_DWO_SECTION_FLAGS): New.
12020 (find_AT_string): Move most logic to ...
12021 (find_AT_string_in_table): ... here. New.
12022 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
12023 add_skeleton_AT_string. Delete logic.
12024 (output_skeleton_debug_sections): Remove call to
12025 add_top_level_skeleton_die_attrs.
12026 (add_comp_dir_attribute): Move logic to comp_dir_string.
12027 (dwarf2out_init): Initialize debug_str_dwo_section.
12028 (output_indirect_string): Call find_string_form.
12029 (output_indirect_strings): Rewrite.
12030 (prune_unused_types): Empty skeleton_debug_str_hash.
12031 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
12032 (dwarf2out_finish): Call output_indirect_strings.
12034 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12036 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
12038 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12040 * rtl.h (struct rtx_def): Add comment for field jump.
12041 (LRA_SUBREG_P): New macro.
12042 * recog.c (register_operand): Check LRA_SUBREG_P.
12043 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12045 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12046 (lra_final_code_change): Skip subreg change for operators.
12047 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12048 if there are no operand changes.
12049 * lra-constraints.c (curr_insn_set): New.
12050 (match_reload): Set LRA_SUBREG_P.
12051 (emit_spill_move): Ditto.
12052 (check_and_process_move): Use curr_insn_set. Process only single
12053 set insns. Don't initialize sec_mem_p and change_p.
12054 (simplify_operand_subreg): Use LRA_SUBREG_P.
12055 (reg_in_class_p): New function.
12056 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12057 of #ifdef. Add code to remove cycling.
12058 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12059 non-null disp. Reload inner instead of disp when base and index
12060 are null. Try to put lo_sum into register.
12061 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12062 (check_and_process_move): Move code for move cost check to
12063 simple_move_p. Remove equiv_substitution.
12064 (simple_move_p): New function.
12065 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12066 curr_insn_set. Call check_and_process_move only for single set
12067 insns. Use the new function. Move call of check_and_process_move
12068 after operand equiv substitution and address process.
12070 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
12072 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
12073 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
12074 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
12076 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12078 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
12080 2013-04-24 Marek Polacek <polacek@redhat.com>
12082 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
12083 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
12084 (select_loops_exit_conditions): Likewise.
12085 (number_of_iterations_for_all_loops): Likewise.
12086 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
12087 (scev_analysis): Likewise.
12089 2013-04-02 Catherine Moore <clm@codesourcery.com>
12090 Chao-ying Fu <fu@mips.com>
12092 * config/mips/micromips.md (jraddiusp): New pattern.
12093 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
12094 instruction if possible.
12096 2013-04-24 Alan Modra <amodra@gmail.com>
12098 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
12100 2013-04-24 Julian Brown <julian@codesourcery.com>
12101 Chung-Lin Tang <cltang@codesourcery.com>
12103 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
12104 dependency behavior in enumeration type DIE generation. Add TODO note
12105 to comments about future DW_FORM_sdata/udata re-work of related code.
12107 2013-04-23 Lawrence Crowl <crowl@google.com>
12109 * Makefile.in: Update as needed below.
12111 * hash-table.h (class hash_table):
12112 Correct many methods with parameter types compare_type to the correct
12113 value_type. (Correct code was unlikely to notice the change.)
12114 (hash_table::elements_with_deleted) New.
12115 (class hashtable::iterator): New.
12116 (hashtable::begin()): New.
12117 (hashtable::end()): New.
12118 (FOR_EACH_HASH_TABLE_ELEMENT): New.
12120 * statistics.c (statistics_hashes):
12121 Change type to hash_table. Update dependent calls and types.
12123 * tree-into-ssa.c (var_infos):
12124 Change type to hash_table. Update dependent calls and types.
12126 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
12127 Change type to hash_table. Update dependent calls and types.
12129 * tree-ssa-loop-im.c (struct mem_ref.refs):
12130 Change type to hash_table. Update dependent calls and types.
12132 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
12133 Change type to hash_table. Update dependent calls and types.
12135 * tree-ssa-sccvn.c (vn_tables_s::nary):
12136 Change type to hash_table. Update dependent calls and types.
12137 (vn_tables_s::phis): Likewise.
12138 (vn_tables_s::references): Likewise.
12140 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
12141 (vn_reference_eq): Update parameter and return types.
12143 * tree-ssa-structalias.c (pointer_equiv_class_table):
12144 Change type to hash_table. Update dependent calls and types.
12145 (location_equiv_class_table): Likewise.
12147 * tree-vect-data-refs.c: Consequential changes for making
12148 peeling a hash_table.
12150 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
12151 (destroy_loop_vec_info): Dependent hash_table update.
12153 * tree-vectorizer.h (peeling_htab):
12154 Change type to hash_table. Update dependent calls and types.
12156 2013-04-23 Shiva Chen <shiva0217@gmail.com>
12158 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
12159 to check the register content is equal or not.
12160 * lra-constraints.c (match_reload): Use lra_assign_reg_val
12161 to assign register content record.
12162 * lra-eliminations.c (update_reg_eliminate): Use
12163 lra_update_reg_val_offset to update register content offset.
12164 * lra-int.h (struct lra_reg): Add offset member.
12165 (lra_reg_val_equal_p): New static inline function.
12166 (lra_update_reg_val_offset): New static inline function.
12167 (lra_assign_reg_val): New static inline function.
12168 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
12169 to assign register content record.
12170 (initialize_lra_reg_info_element): Initial offset to zero.
12172 2013-04-23 Catherine Moore <clm@codesourcery.com>
12174 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
12175 operands. Record compression.
12177 2013-04-23 Xinliang David Li <davidxl@google.com>
12179 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
12181 2013-04-23 Richard Biener <rguenther@suse.de>
12183 PR middle-end/57036
12184 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
12185 parameter, only add abnormal goto edges from the copied body
12186 if the call could perform abnormal gotos.
12187 (copy_cfg_body): Adjust.
12189 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
12191 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
12193 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
12195 * coretypes.h (gimple_stmt_iterator): Add struct to make
12198 2013-04-23 Richard Biener <rguenther@suse.de>
12200 PR tree-optimization/57026
12201 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
12202 from SSA names occuring in abnormal PHI nodes.
12204 2013-04-22 Andi Kleen <ak@linux.intel.com>
12206 * lto/lto.c (print_lto_report_1): Fix LTO report names.
12208 2013-04-22 Andi Kleen <ak@linux.intel.com>
12210 * lto/lto.c (print_lto_report_1): Declare early.
12211 (read_cgraph_and_symbols): Call print_lto_report_1 early.
12213 2013-04-22 Andi Kleen <ak@linux.intel.com>
12215 * common.opt (-flto-report-wpa): Add.
12216 * doc/invoke.texi (-flto-report-wpa): Add.
12217 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
12220 2013-04-22 Xinliang David Li <davidxl@google.com>
12222 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
12223 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
12224 * Makefile.in: New dependency
12226 David Daney <ddaney.cavm@gmail.com>
12228 * configure.ac (gcc_cv_as_micromips_support): Use the
12229 --fatal-warnings option.
12230 * configure: Regenerate.
12232 2013-04-22 Marek Polacek <polacek@redhat.com>
12235 * tsan.c (instrument_expr): Don't instrument expression
12236 in case its size is zero.
12238 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
12242 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
12244 * config/alpha/alpha.c (TARGET_LRA_P): New define.
12246 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12248 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
12249 (gimple_stmt_iterator): New typedef.
12250 * gimple.h (gimple_stmt_iterator): Rename to...
12251 (gimple_stmt_iterator_d): ... This.
12252 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
12253 trees be valid for GIMPLE and GENERIC.
12254 (TARGET_GIMPLE_FOLD_BUILTIN): New.
12255 * gimple-fold.c (gimple_fold_call): Call target hook
12256 gimple_fold_builtin.
12257 * hooks.c (hook_bool_gsiptr_false): New.
12258 * hooks.h (hook_bool_gsiptr_false): New.
12259 * target.def (fold_stmt): New.
12260 * doc/tm.texi: Regenerate.
12262 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
12265 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
12266 a set sp if no stack realignment.
12268 2013-04-22 Nick Clifton <nickc@redhat.com>
12270 * config.gcc (tilegx-linux): Extend extra_objs rather than
12272 (tilepro-linux): Likewise.
12274 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12276 * config/aarch64/aarch64-builtins.c
12278 (CF0, CF1, CF2, CF3, CF4, CF10): New.
12279 (VAR<1-12>): Add MAP parameter.
12280 (BUILTIN_*): Likewise.
12281 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
12282 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
12283 (aarch64_ushl_n<mode>): Likewise.
12284 (aarch64_sshr_n<mode>): Likewise.
12285 (aarch64_ushr_n<mode>): Likewise.
12286 (aarch64_<maxmin><mode>): Likewise.
12287 (aarch64_sqrt<mode>): Likewise.
12288 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
12289 (vshr<q>_n_*): Likewise.
12291 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12293 * config/aarch64/aarch64-builtins.c
12294 (aarch64_simd_builtin_type_mode): Handle SF types.
12296 (BUILTIN_GPF): Define.
12297 (aarch64_init_simd_builtins): Handle SF types.
12298 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
12299 (frecps): Likewise.
12300 (frecpx): Likewise.
12301 * config/aarch64/aarch64-simd.md
12302 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
12303 (aarch64_frecpe<mode>): New.
12304 (aarch64_frecps<mode>): Likewise.
12305 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
12306 (v8type): Add frecp<esx>.
12307 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
12308 (aarch64_frecps<mode>): Likewise.
12309 * config/aarch64/iterators.md (FRECP): New.
12310 (frecp_suffix): Likewise.
12311 * config/aarch64/arm_neon.h
12312 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
12314 2013-04-22 Christian Bruel <christian.bruel@st.com>
12317 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
12318 (REG_CLASS_NAMES): Idem.
12319 (REG_CLASS_CONTENTS): Idem.
12320 (REGCLASS_HAS_FP_REG): Idem.
12321 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
12322 (sh_conditional_register_usage): Idem.
12324 2013-04-21 Jeff Law <law@redhat.com>
12326 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
12327 (ssa_forward_propagate_and_combine): Use it.
12329 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12331 * lra.c: Update the flow chart diagram.
12333 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12335 PR rtl-optimization/56847
12336 * lra-constraints.c (process_alt_operands): Discourage alternative
12337 with non-matche doffsettable memory constraint fro memory with
12340 2013-04-19 Richard Biener <rguenther@suse.de>
12342 PR tree-optimization/56982
12343 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
12345 * gimplify.c (gimplify_call_expr): Notice special calls.
12346 (gimplify_modify_expr): Likewise.
12347 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
12348 abnormal control flow receivers.
12349 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
12350 in the same way as cfun->has_nonlocal_labels.
12351 (gimple_purge_dead_abnormal_call_edges): Likewise.
12352 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
12353 receivers start a basic-block.
12355 2013-04-19 Richard Biener <rguenther@suse.de>
12357 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
12359 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
12360 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
12361 (SLP_TREE_LOAD_PERMUTATION): Add.
12362 (vect_transform_slp_perm_load): Adjust prototype.
12363 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
12364 (vect_free_slp_instance): Likewise.
12365 (vect_create_new_slp_node): Likewise.
12366 (vect_supported_slp_permutation_p): Remove.
12367 (vect_slp_rearrange_stmts): Adjust.
12368 (vect_supported_load_permutation_p): Likewise. Inline
12369 vect_supported_slp_permutation_p here.
12370 (vect_analyze_slp_instance): Compute load permutations per
12371 slp node instead of per instance.
12372 (vect_get_slp_defs): Adjust.
12373 (vect_transform_slp_perm_load): Likewise.
12374 (vect_schedule_slp_instance): Remove redundant code.
12375 (vect_schedule_slp): Remove hack for PR56270, add it ...
12376 * tree-vect-stmts.c (vectorizable_load): ... here, do not
12377 CSE loads for SLP. Adjust.
12379 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
12381 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
12382 spelling in two comments.
12384 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
12387 * config/arm/arm.c (load_multiple_sequence): Require SP
12388 as base register for loads if SP is in the register list.
12390 2013-04-19 Martin Jambor <mjambor@suse.cz>
12392 PR tree-optimization/56718
12393 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
12394 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
12395 and made public. Adjusted all callers.
12396 (ipa_intraprocedural_devirtualization): New function.
12397 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
12398 (ipa_intraprocedural_devirtualization): Likewise.
12399 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
12401 2013-04-19 Richard Biener <rguenther@suse.de>
12403 PR tree-optimization/57000
12404 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
12406 2013-04-19 Terry Guo <terry.guo@arm.com>
12408 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
12410 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
12411 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
12412 (cortex_m4_fmacs): Use new reservations.
12413 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
12415 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
12417 PR rtl-optimization/56999
12418 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
12420 (lra_coalesce): Remove split_origin_bitmap and related code.
12421 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
12422 ranges if necessary.
12424 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
12426 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
12428 (ix86_expand_call): Remove clobbered_registers array and use
12429 x86_64_ms_sysv_extra_clobbered_registers instead.
12430 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
12432 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
12434 * config/i386/i386.md (*call_rex64_ms_sysv): Use
12435 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
12436 (*call_value_rex64_ms_sysv): Ditto.
12438 2013-04-18 Cary Coutant <ccoutant@google.com>
12440 * dwarf2out.c (output_pubnames): Check die_perennial_p of
12441 parent instead of die_mark.
12443 2013-04-18 Diego Novillo <dnovillo@google.com>
12445 * gimple.c (create_gimple_tmp): New.
12446 (get_expr_type): New.
12447 (build_assign): New.
12448 (build_type_cast): New.
12449 * gimple.h (enum ssa_mode): Define.
12450 (gimple_seq_set_location): New.
12451 * asan.c (build_check_stmt): Change some gimple_build_* calls
12452 to use build_assign and build_type_cast.
12454 2013-04-18 Richard Biener <rguenther@suse.de>
12456 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
12457 handle negative step. Remove redundant checks.
12458 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
12459 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
12460 for negative step and grouped loads fail to vectorize.
12462 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
12464 * emit-rtl.c (reset_insn_used_flags): New function.
12465 (reset_all_used_flags): Use it.
12466 (verify_insn_sharing): New function.
12467 (verify_rtl_sharing): Fix verification for SEQUENCEs.
12469 2013-04-18 Jakub Jelinek <jakub@redhat.com>
12471 PR tree-optimization/56984
12472 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
12473 and (x >> M) >= N don't register any assertion if N << M is the
12476 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
12478 * lower-subreg.c (resolve_simple_move): If called self-recursive,
12479 do not delete_insn insns that have not yet been emitted, only
12480 unlink them with remove_insn.
12481 * df-scan.c (df_insn_delete): Revert r197492.
12483 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
12485 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
12486 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
12488 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
12490 * config/arm/arm.md (movsicc_insn): Convert define_insn into
12491 define_insn_and_split.
12492 (and_scc,ior_scc,negscc): Likewise.
12493 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
12495 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
12497 * config/arm/arm.c (use_return_insn): Return 0 for targets that
12498 can benefit from using a sequence of LDRD instructions in epilogue
12499 instead of a single LDM instruction.
12501 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
12504 * doc/extend.texi: Fix typo.
12506 2013-04-17 Richard Biener <rguenther@suse.de>
12508 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
12509 (vect_build_slp_tree): ... here.
12510 (vect_build_slp_tree_1): Compute which stmts of the SLP group
12511 match. Remove special-casing of mismatched complex loads.
12512 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
12513 re-try the match with swapped commutative operands.
12514 (vect_supported_load_permutation_p): Remove special-casing of
12515 mismatched complex loads.
12516 (vect_analyze_slp_instance): Adjust.
12518 2013-04-17 Richard Biener <rguenther@suse.de>
12520 PR rtl-optimization/56921
12521 * cfgloop.h (struct loop): Add simple_loop_desc member.
12522 (struct niter_desc): Mark with GTY(()).
12523 (simple_loop_desc): Do not use aux field but simple_loop_desc.
12524 * loop-iv.c (get_simple_loop_desc): Likewise.
12525 (free_simple_loop_desc): Likewise.
12528 2013-04-16 Richard Biener <rguenther@suse.de>
12530 PR rtl-optimization/56921
12531 * loop-init.c (pass_rtl_move_loop_invariants): Add
12532 TODO_do_not_ggc_collect to todo_flags_finish.
12533 (pass_rtl_unswitch): Same.
12534 (pass_rtl_unroll_and_peel_loops): Same.
12535 (pass_rtl_doloop): Same.
12537 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
12539 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
12540 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
12541 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
12543 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
12544 * tree-streamer.c (record_common_node): Adjust reference in comment.
12546 2013-04-17 Terry Guo <terry.guo@arm.com>
12548 * config/arm/cortex-m4.md: Add a new bypass.
12550 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12552 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
12554 (*subs_<optab><mode>_multp2): New pattern.
12555 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
12556 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
12558 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12560 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
12561 (*subs_mul_imm_<mode>): New pattern.
12563 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
12566 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
12567 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
12568 (vsx_movti_32bit): Change j->wa to O->wa.
12570 2013-04-16 Richard Biener <rguenther@suse.de>
12572 PR rtl-optimization/56921
12573 * loop-init.c (pass_rtl_move_loop_invariants): Add
12574 TODO_do_not_ggc_collect to todo_flags_finish.
12575 (pass_rtl_unswitch): Same.
12576 (pass_rtl_unroll_and_peel_loops): Same.
12577 (pass_rtl_doloop): Same.
12579 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
12581 * config/arm/arm.c (emit_multi_reg_push): New declaration
12582 for an existing function.
12583 (arm_emit_strd_push): New function.
12584 (arm_expand_prologue): Used here.
12585 (arm_emit_ldrd_pop): New function.
12586 (arm_expand_epilogue): Used here.
12587 (arm_get_frame_offsets): Update condition.
12588 (arm_emit_multi_reg_pop): Add a special case for load of a single
12589 register with writeback.
12591 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
12593 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
12596 2013-04-16 Richard Biener <rguenther@suse.de>
12598 PR tree-optimization/56756
12599 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
12600 (first_mem_ref_loc): New.
12601 (execute_sm): Place the load temporarily before a previous
12602 access instead of in the latch edge to ensure its SSA dependencies
12603 are defined at points dominating the load.
12605 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
12607 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
12608 correct fix by moving header and footer insn to the footer of
12609 the merged basic block. Clear BB_END of the merged-away block.
12611 PR middle-end/43631
12612 * emit-rtl.c (make_note_raw): New function.
12613 (link_insn_into_chain): New static inline function.
12614 (add_insn): Use it.
12615 (add_insn_before, add_insn_after): Factor insn chain linking code...
12616 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
12617 using link_insn_into_chain.
12618 (note_outside_basic_block_p): New helper function for emit_note_after
12619 and emit_note_before.
12620 (emit_note_after): Use nobb variant of add_insn_after if the note
12621 should not be contained in a basic block.
12622 (emit_note_before): Use nobb variant of add_insn_before if the note
12623 should not be contained in a basic block.
12624 (emit_note_copy): Use make_note_raw.
12625 (emit_note): Likewise.
12626 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
12627 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
12628 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
12629 the moved barrier the tail of the basic block it follows.
12630 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
12632 2013-04-15 Jakub Jelinek <jakub@redhat.com>
12634 PR tree-optimization/56962
12635 * gimple-ssa-strength-reduction.c (record_increment): Only set
12636 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
12637 either rhs1 or rhs2 is equal to c->base_expr.
12639 2013-04-15 Richard Biener <rguenther@suse.de>
12641 PR tree-optimization/56933
12642 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
12644 (GROUP_READ_WRITE_DEPENDENCE): Remove.
12645 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
12646 * tree-vect-data-refs.c (vect_analyze_group_access): Move
12647 dependence check ...
12648 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
12650 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
12651 GROUP_READ_WRITE_DEPENDENCE.
12653 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12655 * emit-rtl.c (reset_all_used_flags): New function.
12656 (verify_rtl_sharing): Call reset_all_used_flags before and after
12657 performing the checks.
12659 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12661 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
12662 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
12663 * config/arm/constraints.md (De): New constraint.
12664 * config/arm/neon.md (anddi3_neon): Delete.
12665 (neon_vand<mode>): Expand to standard anddi3 pattern.
12666 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
12667 Move earlier in the file.
12668 (neon_inv_logic_op2): Likewise.
12669 (arm_anddi_operand_neon): New predicate.
12671 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12673 * configure.ac (gcc_cv_ld_as_needed): Set
12674 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
12675 Use -z ignore, -z record on *-*-solaris2*.
12676 (HAVE_LD_AS_NEEDED): Update comment.
12677 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
12678 * configure: Regenerate.
12679 * config.in: Regenerate.
12680 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
12681 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
12682 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
12683 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
12684 equivalents. Fix markup.
12685 * doc/tm.texi: Regenerate.
12687 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
12689 * config/i386/i386.opt: New option mstack-protector-guard=.
12690 * config/i386/i386-opts.h: Add enum stack_protector_guard.
12691 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
12692 TARGET_SSP_TLS_GUARD.
12693 * config/i386/i386.c (ix86_option_override_internal): Set
12694 ix86_stack_protector_guard.
12695 * config/i386/i386.md (stack_protect_set): Enable for
12696 TARGET_SSP_TLS_GUARD only.
12697 (stack_protect_set_<mode>): Ditto.
12698 (stack_protect_test): Ditto.
12699 (stack_protect_test_<mode>): Ditto.
12700 * doc/invoke.texi (i386 Option): Document.
12702 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
12705 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
12706 (S_MODES): Set H_MODE bit.
12707 (SF_MODES): Set only S_MODE and SF_MODE bits.
12708 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
12709 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
12710 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
12711 <MODE_FLOAT>: Likewise.
12713 2013-04-15 Joey Ye <joey.ye@arm.com>
12715 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
12717 2013-04-15 Joey Ye <joey.ye@arm.com>
12719 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
12721 (thumb_far_jump_used_p): Count instruction size and set
12724 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
12726 * reorg.c (fill_simple_delay_slots): Reindent block of code.
12727 * resource.c (mark_target_live_regs): Reformat conditional block.
12729 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
12731 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
12732 notes, they are emitted only just before final.
12733 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
12735 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
12737 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
12738 * cfgrtl.c (delete_insn): Call it here instead.
12739 * lra-spills.c (lra_final_code_change): Use delete_insn.
12740 * haifa-sched.c (sched_remove_insn): Likewise.
12741 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
12742 returning to the nop pool.
12743 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
12744 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
12746 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
12748 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
12749 * doc/tm.texi: Regenerated.
12751 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
12753 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
12756 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
12758 * df-core.c (df_find_def): Compare register numbers.
12759 (df_find_use): Likewise.
12761 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
12764 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
12765 lra_in_progress for return.
12767 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
12769 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
12770 define_insn into define_insn_and_split and emit movsicc patterns.
12772 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
12774 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
12776 2013-04-12 Richard Biener <rguenther@suse.de>
12778 * tree-pass.h (TODO_do_not_ggc_collect): New.
12779 * passes.c (execute_one_ipa_transform_pass): Honor
12780 TODO_do_not_ggc_collect.
12781 (execute_one_pass): Likewise.
12784 2013-04-10 Richard Biener <rguenther@suse.de>
12786 * passes.c (init_optimization_passes): Remove reload pass.
12787 * ira.c (do_reload): Merge into ...
12789 (rest_of_handle_reload): Remove.
12790 (pass_reload): Likewise.
12791 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12792 of reload for vzeroupper pass placement.
12794 2013-04-12 Jakub Jelinek <jakub@redhat.com>
12796 PR tree-optimization/56918
12797 PR tree-optimization/56920
12798 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
12799 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
12800 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
12801 use wide_mul_with_sign method.
12803 2013-04-12 Richard Biener <rguenther@suse.de>
12805 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
12806 not be considered a gimple constant.
12808 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
12810 * fold-const.c (const_binop): Handle vector shifts by a scalar.
12811 (fold_binary_loc): Call const_binop also for mixed vector-scalar
12814 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
12815 Jakub Jelinek <jakub@redhat.com>
12817 * opts.c: Include diagnostic-color.h.
12818 (common_handle_option): Handle OPT_fdiagnostics_color_.
12819 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
12820 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
12821 (diagnostic-color.o): New.
12822 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
12823 (diagnostic_color_rule): New enum.
12824 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
12825 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
12826 the location string.
12827 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
12828 either NULL, or color kind.
12829 * diagnostic-color.c: New file.
12830 * diagnostic-color.h: New file.
12831 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
12833 * doc/invoke.texi (-fdiagnostics-color): Document.
12834 * pretty-print.h (pp_show_color): Define.
12835 (struct pretty_print_info): Add show_color field.
12836 * diagnostic.c: Include diagnostic-color.h.
12837 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
12838 macros. Colorize error:, warning: etc. strings and also the location
12840 (diagnostic_show_locus): Colorize the caret line.
12841 * pretty-print.c: Include diagnostic-color.h.
12842 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
12843 inside of %< %> quotes or quoted through q format modifier.
12845 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12847 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
12849 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12851 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
12852 code in CC_NZ mode.
12853 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
12856 2013-04-11 Marek Polacek <polacek@redhat.com>
12858 PR tree-optimization/48184
12859 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
12861 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
12863 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
12864 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
12865 (skip_simple_arithmetic): Tidy up.
12866 * tree.h (skip_simple_constant_arithmetic): Declare.
12868 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12870 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
12872 2013-04-11 Richard Biener <rguenther@suse.de>
12874 * tree-vect-loop.c (get_initial_def_for_induction): Properly
12875 generate vector constants.
12877 2013-04-11 Richard Biener <rguenther@suse.de>
12879 PR tree-optimization/56878
12880 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
12881 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
12883 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
12884 Prefer to align the DR with the most invariant base address.
12886 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12888 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
12891 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
12893 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
12894 floating-point vector comparisons against 0.
12896 2013-04-11 Jakub Jelinek <jakub@redhat.com>
12898 PR tree-optimization/56899
12899 * fold-const.c (extract_muldiv_1): Apply distributive law
12900 only if TYPE_OVERFLOW_WRAPS (ctype).
12902 2013-04-11 Bin Cheng <bin.cheng@arm.com>
12905 * ira-costs.c (scan_one_insn): Check whether the source rtx of
12906 loading has side effect.
12908 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
12910 * config/sparc/sparc.c: Include tree-pass.h.
12911 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
12912 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
12913 head of file. Change return type. Split off gate function.
12914 (sparc_gate_work_around_errata): New function.
12915 (pass_work_around_errata): New pass definition.
12916 (insert_pass_work_around_errata) New pass insert definition to
12917 insert pass_work_around_errata just after delayed-branch scheduling.
12918 (sparc_option_override): Insert the pass.
12919 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
12921 2013-04-10 David S. Miller <davem@davemloft.net>
12923 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12924 or -mcpu=hypersparc.
12926 * target.def (cstore_mode): New hook.
12927 * target.h: Include insn-codes.h
12928 * targhooks.c: Likewise.
12929 (default_cstore_mode): New function.
12930 * targhooks.h: Declare it.
12931 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12932 * doc/tm.texi: Rebuild.
12933 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12934 target hook, rather than inspecting the insn_data.
12935 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12936 (TARGET_CSTORE_MODE): Redefine.
12937 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12939 * config/sparc/predicates.md (cstore_result_operand): New special
12941 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12942 Use it for operand 0.
12943 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12944 (*snesi_special): Likewise.
12945 (*snesi_zero): Likewise.
12946 (*seqsi_zero): Likewise.
12947 (*sltu_insn): Likewise.
12948 (*sgeu_insn): Likewise.
12949 (*seqdi_special): Make operand 0 and comparison operation be of
12951 (*snedi_special): Likewise.
12952 (*snedi_special_vis3): Likewise.
12953 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12954 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12955 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12956 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12957 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12958 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12959 (*sltu_extend_sp64): Likewise.
12960 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12961 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12962 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12963 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12964 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12966 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12968 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12969 (aarch64_start_file): Use the new function.
12971 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12972 Jason Merrill <jason@redhat.com>
12974 * common.opt: Add -gdwarf.
12975 * opts.c (common_handle_option): Handle it.
12976 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12978 2013-04-10 Richard Biener <rguenther@suse.de>
12980 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12981 (execute_one_ipa_transform_pass): But unconditionally here.
12982 (execute_one_pass): And here.
12983 (init_optimization_passes): Remove reload pass.
12984 * tree-pass.h (TODO_ggc_collect): Remove.
12985 (pass_reload): Likewise.
12986 * ira.c (do_reload): Merge into ...
12988 (rest_of_handle_reload): Remove.
12989 (pass_reload): Likewise.
12990 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12991 of reload for vzeroupper pass placement.
12992 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12993 and todo_flags_finish of all passes.
12995 2013-04-10 Richard Biener <rguenther@suse.de>
12997 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12998 first_const_oprnd field, rename first_def_type to first_op_type.
12999 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
13000 (vect_get_and_check_slp_defs): Always use the type of the
13001 operand. Allow mixed vect_external_def, vect_constant_def types.
13002 (vect_get_constant_vectors): Handle mixed vect_external_def,
13003 vect_constant_def types.
13005 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
13007 PR tree-optimization/55524
13008 * tree-ssa-math-opts.c
13009 (convert_mult_to_fma): Don't use an fms construct
13010 when we don't have an fms operation, but fnma, and it looks
13011 likely that we'll be able to use the latter.
13013 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
13015 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
13017 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
13018 inline fail caused by overwritable functions.
13020 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
13022 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
13023 unnecessary bits in the constant power of two case.
13025 2013-04-10 Richard Biener <rguenther@suse.de>
13027 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
13028 broken code swapping operands.
13029 (vect_build_slp_tree): Do not compute load permutations here.
13030 (vect_analyze_slp_instance): Compute load permutations here,
13031 after building the SLP tree.
13033 2013-04-09 Christian Bruel <christian.bruel@st.com>
13035 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
13036 of next/prev_real_insn.
13038 2013-04-09 Jan Hubicka <jh@suse.cz>
13040 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13041 Drop aliased parameter.
13042 (function_and_variable_visibility): Do not handle alias pairs.
13043 * cgraph.c (varpool_externally_visible_p): Update prototype.
13044 * varpool.c (varpool_add_new_variable): Update.
13046 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13048 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
13050 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
13052 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
13054 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
13056 2013-04-09 Marek Polacek <polacek@redhat.com>
13058 PR tree-optimization/48762
13059 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
13061 2013-04-09 Richard Biener <rguenther@suse.de>
13063 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
13065 (vect_build_slp_tree): Likewise.
13066 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
13067 calculating the cost of a SLP instance.
13068 (vect_analyze_slp_instance): Use it from here, after building
13071 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13073 PR middle-end/56883
13074 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
13075 expand_omp_for_static_chunk): Use simple_p = true in
13076 force_gimple_operand_gsi calls when assigning to addressable decls.
13078 2013-04-09 Jeff Law <law@redhat.com>
13080 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
13081 when the boolean was created by converting a wider object which
13082 had a boolean range.
13084 2013-04-09 Richard Biener <rguenther@suse.de>
13086 * tree-vectorizer.h (slp_void_p): Remove.
13087 (slp_tree): Typedef before _slp_tree declaration.
13088 (struct _slp_tree): Use a vector of slp_tree as children.
13089 (vect_get_place_in_interleaving_chain): Remove.
13090 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
13092 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
13094 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
13095 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
13096 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
13097 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
13098 Use slp_node instead of slp_void_p and adjust.
13100 2013-04-09 Richard Biener <rguenther@suse.de>
13102 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
13103 work that is not necessary.
13105 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13107 PR tree-optimization/56854
13108 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
13109 forward into clobber stmts if it would change MEM_REF lhs into
13112 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
13114 * tree.c (type_hash_lookup, type_hash_add): Make static.
13115 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
13117 2013-04-09 Richard Biener <rguenther@suse.de>
13119 * tree.h (unsave_expr_now): Remove.
13120 * tree-inline.c (mark_local_for_remap_r): Remove.
13121 (unsave_expr_1): Likewise.
13122 (unsave_r): Likewise.
13123 (unsave_expr_now): Likewise.
13124 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
13125 (propagate_tree_value): Likewise.
13127 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
13129 * doc/rtl.texi (sequence): Rewrite documentation to match the
13130 current use of SEQUENCE rtl objects.
13131 * rtl.def (SEQUENCE): Likewise.
13133 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
13134 Update documentation.
13135 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
13136 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
13138 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
13140 2013-04-08 Teresa Johnson <tejohnson@google.com>
13142 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
13143 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
13145 (estimate_edge_size_and_time): Add comment to suggest using rounding
13147 (estimate_node_size_and_time): Ditto.
13148 (remap_edge_change_prob): Use helper rounding divide methods.
13149 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
13150 (gimple_mod_pow2_value_transform): Ditto.
13151 (gimple_mod_subtract_transform): Ditto.
13152 (gimple_ic_transform): Ditto.
13153 (gimple_stringops_transform): Ditto.
13154 * stmt.c (conditional_probability): Ditto.
13155 (emit_case_dispatch_table): Ditto.
13156 * lto-cgraph.c (merge_profile_summaries): Ditto.
13157 * tree-optimize.c (execute_fixup_cfg): Ditto.
13158 * cfgcleanup.c (try_forward_edges): Ditto.
13159 * cfgloopmanip.c (scale_loop_profile): Ditto.
13161 (duplicate_loop_to_header_edge): Ditto.
13162 (lv_adjust_loop_entry_edge): Ditto.
13163 * tree-vect-loop.c (vect_transform_loop): Ditto.
13164 * profile.c (compute_branch_probabilities): Ditto.
13165 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
13166 * lto-streamer-in.c (input_cfg): Ditto.
13167 * gimple-streamer-in.c (input_bb): Ditto.
13168 * ipa-cp.c (update_profiling_info): Ditto.
13169 (update_specialized_profile): Ditto.
13170 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
13171 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
13173 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13174 (compute_trg_info): Ditto.
13175 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
13176 (purge_dead_edges): Ditto.
13177 * loop-unswitch.c (unswitch_loop): Ditto.
13178 * cgraphclones.c (cgraph_clone_edge): Ditto.
13179 (cgraph_clone_node): Ditto.
13180 * tree-inline.c (copy_bb): Ditto.
13181 (copy_edges_for_bb): Ditto.
13182 (initialize_cfun): Ditto.
13183 (copy_cfg_body): Ditto.
13184 (expand_call_inline): Ditto.
13186 2013-04-08 Kai Tietz <ktietz@redhat.com>
13188 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
13189 TARGET_CYGWIN64 by TARGET_64BIT.
13191 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13193 * config/epiphany/epiphany.md (GPR_1): New constant.
13194 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
13195 * config/epiphany/epiphany.c (gen_compare_reg):
13196 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
13197 is already in place.
13198 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
13199 Don't require being called during rtl expansion; If y operlaps r0,
13201 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
13202 (epiphany_expand_epilogue): Likewise.
13204 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
13205 Don't use CC_FPmode for ORDERED / UNORDERED.
13206 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
13208 * config/epiphany/constraints.md (CnL): New constraint.
13209 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
13210 * config/epiphany/predicates.md (add_operand): Allow 1024.
13212 * config/epiphany/epiphany.md (logical_op): New code iterator.
13213 (op_mnc): New code attribute.
13214 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
13215 (mov_f+1, mov_f+2): New peephole2 patterns.
13217 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
13218 (cstoresi4): Also allow re-use of zero result when doing a NE
13219 comparison to a non-zero operand.
13220 Use (clobber (scratch)) for first insn if the gpr output is not needed.
13222 * config/epiphany/epiphany.md (<insn_opname>v2si3):
13223 Use gen_addsi3_i / gen_subsi3_i.
13225 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13229 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
13230 contain anything but clobbers, at most one __builtin_stack_restore,
13231 optionally debug stmts and final resx, and if it has at least one
13232 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
13233 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
13234 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
13235 which isn't defaut definition, remove them.
13236 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
13237 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
13238 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
13239 with MEM_REF LHS with SSA_NAME address.
13241 2013-04-08 Jeff Law <law@redhat.com>
13243 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
13245 2013-04-08 Richard Biener <rguenther@suse.de>
13247 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
13249 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
13250 determined vector type.
13251 (vect_analyze_data_refs): Likewise.
13252 (vect_get_new_vect_var): Adjust.
13253 (vect_create_destination_var): Preserve SSA name versions.
13254 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
13255 not dump anything here.
13257 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13259 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
13260 Add member lr_slot_known.
13261 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
13263 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
13264 Remove code that sets lr_slot_offset according to what a previous
13265 version of epiphany_emit_save_restore used to do.
13266 (epiphany_emit_save_restore): When doing an lr save or restore,
13267 set/verify lr_slot_known and lr_slot_offset.
13269 2013-04-08 Xinyu Qi <xyqi@marvell.com>
13272 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
13275 2013-04-08 Richard Biener <rguenther@suse.de>
13277 * alias.c (find_base_term): Fix thinko in previous change.
13279 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13281 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
13282 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
13283 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
13284 if possible to compute val.
13285 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
13286 For QImode integers don't require anything about precision. Use
13287 const_with_all_bytes_same to find out if the constant doesn't have
13288 repeated bytes in it.
13290 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13292 * config/s390/s390.c (s390_expand_insv): Only accept insertions
13295 2013-04-08 Marek Polacek <polacek@redhat.com>
13297 PR rtl-optimization/48182
13298 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
13301 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13304 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
13305 nuses, make sure we have a label.
13307 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13310 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
13311 (rs6000_emit_swdiv_low_precision): Remove.
13312 (rs6000_emit_swdiv): Rewrite to handle between one and four
13313 iterations of Newton-Raphson generally; modify required number of
13314 iterations for some cases.
13315 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
13317 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
13319 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
13320 set-but-unused variable.
13322 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
13323 basic blocks of released function bodies garbage-collectable.
13325 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
13326 (struct rtl_opt_pass): Add TODO_df_finish.
13328 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
13330 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13332 * config/arm/constraints.md (q): New constraint.
13333 * config/arm/ldrdstrd.md: New file.
13334 * config/arm/arm.md (ldrdstrd.md) New include.
13335 (arm_movdi): Use "q" instead of "r" constraint
13336 for double-word memory access.
13337 (movdf_soft_insn): Likewise.
13338 * config/arm/vfp.md (movdi_vfp): Likewise.
13339 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
13340 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
13341 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
13342 (mem_ok_for_ldrd_strd): Likewise.
13343 (output_move_double): Update assertion.
13345 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13347 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
13349 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13351 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
13352 define_insn_and_split.
13353 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
13355 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13357 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
13358 define_insn_and_split.
13359 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
13360 (shiftsi3_compare): New pattern.
13361 (rrx): New pattern.
13362 * config/arm/unspecs.md (UNSPEC_RRX): New.
13364 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13366 * config/arm/arm.md (negdi_extendsidi): New pattern.
13367 (negdi_zero_extendsidi): Likewise.
13369 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13371 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
13372 define_insn_and_split.
13373 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
13374 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
13376 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13378 * config/arm/arm.md (arm_subdi3): Convert define_insn into
13379 define_insn_and_split.
13380 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
13381 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
13383 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13385 * config/arm/arm.md (subsi3_carryin): New pattern.
13386 (subsi3_carryin_const): Likewise.
13387 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
13388 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
13390 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13392 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
13394 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13396 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
13397 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
13399 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13401 * config/arm/arm.c (arm_expand_builtin): Change fcode
13402 type to unsigned int.
13404 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13406 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
13408 2013-04-04 Ian Lance Taylor <iant@google.com>
13410 * doc/standards.texi (Standards): The Go frontend supports the Go 1
13413 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
13415 PR middle-end/56729
13416 * df-scan.c (df_insn_delete): Disable failing assert.
13418 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13420 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
13421 New function prototype.
13422 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
13423 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
13424 (arm_builtin_vectorized_function): New function.
13426 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13428 * config/arm/arm_neon_builtins.def: New file.
13429 * config/arm/arm.c (neon_builtin_data): Move contents to
13430 arm_neon_builtins.def.
13431 (enum arm_builtins): Include neon builtin definitions.
13432 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
13433 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
13435 2013-04-04 Marek Polacek <polacek@redhat.com>
13437 PR tree-optimization/48186
13438 * predict.c (maybe_hot_frequency_p): Return false if
13439 HOT_BB_FREQUENCY_FRACTION is 0.
13440 (cgraph_maybe_hot_edge_p): Likewise.
13442 2013-04-04 Richard Biener <rguenther@suse.de>
13444 PR tree-optimization/56826
13445 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
13448 2013-04-04 Richard Biener <rguenther@suse.de>
13450 PR tree-optimization/56213
13451 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
13452 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
13454 2013-04-04 Richard Biener <rguenther@suse.de>
13456 PR tree-optimization/56837
13457 * tree-loop-distribution.c (classify_partition): For non-zero
13458 values require that the value has the same precision as its
13459 mode to be useful as memset value.
13461 2013-04-03 Nick Clifton <nickc@redhat.com>
13463 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
13464 (fmssf4): Use fmsf.s on E3V5 architectures.
13465 (fnmasf4): Use fnmaf.s on E3V5 architectures.
13466 (fnmssf4): Use fnmsf.s on E3V5 architectures.
13468 2013-04-03 Jeff Law <law@redhat.com>
13470 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
13471 (lra-eliminations.o): Likewise.
13473 2013-04-03 Teresa Johnson <tejohnson@google.com>
13475 * gcov-io.c (compute_working_sets): Moved most of body of old
13476 compute_working_sets here from profile.c.
13477 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
13478 (gcov_working_set_t): Moved typedef here from basic-block.h
13479 (compute_working_set): Declare.
13480 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
13481 (get_working_sets): Renamed from compute_working_set,
13482 replace most of body with call to new compute_working_sets.
13483 (get_exec_counts): Replace call to compute_working_sets
13484 to get_working_sets.
13485 * profile.h (get_working_sets): Renamed from compute_working_set.
13486 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
13487 to get_working_sets.
13488 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
13489 * gcov-dump.c (dump_working_sets): New function.
13491 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
13493 * hwint.c (sext_hwi, zext_hwi): New functions.
13494 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
13495 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
13496 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
13497 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
13498 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
13499 (sext_hwi, zext_hwi): New functions.
13501 2013-04-03 Jeff Law <law@redhat.com>
13503 PR tree-optimization/56799
13504 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
13505 back test for widening conversion erroneously dropped in prior change.
13507 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13510 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
13511 instead of next_real_insn.
13513 2013-04-03 Marek Polacek <polacek@redhat.com>
13516 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
13518 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13521 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
13523 (thumb1_output_casesi): Likewise.
13524 (thumb2_output_casesi): Likewise.
13526 2013-04-03 Richard Biener <rguenther@suse.de>
13528 PR tree-optimization/56817
13529 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
13531 (tree_unroll_loops_completely_1): ... new function to manually
13532 walk the loop tree, properly defering outer loops of unrolled
13533 loops to later iterations.
13535 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13537 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
13538 (vectorizable_load): Likewise.
13539 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
13540 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
13542 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13544 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
13547 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13549 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
13551 2013-04-03 Bin Cheng <bin.cheng@arm.com>
13553 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
13555 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13557 PR tree-optimization/56790
13558 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
13561 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13563 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
13565 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
13566 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
13569 2013-04-03 Jakub Jelinek <jakub@redhat.com>
13572 * tree.h (force_folding_builtin_constant_p): New decl.
13573 * builtins.c (force_folding_builtin_constant_p): New variable.
13574 (fold_builtin_constant_p): Fold immediately also if
13575 force_folding_builtin_constant_p.
13577 2013-04-03 Richard Biener <rguenther@suse.de>
13579 PR tree-optimization/56812
13580 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
13581 DRs of the same interleaving chain are independent.
13583 2013-04-02 Jason Merrill <jason@redhat.com>
13585 * gdbinit.in (pbb): Use debug fn.
13587 2013-04-02 Lawrence Crowl <crowl@google.com>
13589 * sese.h (struct ivtype_map_elt_s): Remove unused.
13590 (extern debug_ivtype_map): Remove unused.
13591 (extern eq_ivtype_map_elts): Remove unused.
13592 * sese.c (debug_ivtype_map): Removed unused.
13593 (debug_ivtype_map_1): Removed unused.
13594 (debug_ivtype_elt): Remove unused.
13595 (eq_ivtype_map_elts): Remove unused.
13598 2013-04-02 Kai Tietz <ktietz@redhat.com>
13601 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
13602 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
13603 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
13605 (legitimize_pe_coff_symbol): Likewise.
13606 (is_imported_p): New helper-function.
13607 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
13608 for Windows x64 targets.
13609 (ix86_expand_prologue): Optimize for pe-coff targets.
13610 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
13611 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
13612 medium/large code-model.
13613 (legitimize_pic_address): Likewise.
13614 (legitimize_tls_address): Likewise.
13615 (ix86_expand_call): Likewise.
13616 (x86_output_mi_thunk): Likewise.
13617 (get_dllimport_decl): Add new beimport argument.
13618 (construct_plt_address): Don't assert for x64 pe-coff targets.
13619 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
13621 (SYMBOL_FLAG_STUBVAR): New macro.
13622 (SYMBOL_REF_STUBVAR_P): Likewise.
13623 * config/i386/winnt.c (stub_list): New structure.
13624 (stub_head): New local variable.
13625 (i386_pe_record_stub): New function.
13626 (i386_pe_file_end): Emit refptr-stubs.
13628 2013-04-02 Jakub Jelinek <jakub@redhat.com>
13630 PR rtl-optimization/56745
13631 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
13632 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
13635 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
13636 and both of them are MEM_REFs, just compare first argument for
13637 equality and attempt to deal even with differing offsets.
13640 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
13641 of gimple_clobber_p to be MEM_REF.
13642 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
13643 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
13644 after gimplification.
13645 * asan.c (get_mem_ref_of_assignment): Don't instrument
13646 gimple_clobber_p stmts.
13647 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
13648 gimple_clobber_p stmt if they have MEM_REF lhs and
13649 are dead because of another gimple_clobber_p stmt.
13650 * tree-ssa-live.c (clear_unused_block_pointer): Treat
13651 gimple_clobber_p stmts like debug stmts.
13652 (remove_unused_locals): Remove clobbers with MEM_REF lhs
13653 that refer to unused VAR_DECLs or uninitialized values.
13654 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
13655 gimple_clobber_p stmts if they refer to removed parameters.
13656 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
13659 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
13661 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
13662 using SWI48 mode attribute.
13664 2013-04-02 Wei Mi <wmi@google.com>
13666 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
13667 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
13668 *<rotate_insn><mode>3_mask in i386.md.
13670 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
13672 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
13674 2013-04-02 Richard Biener <rguenther@suse.de>
13676 PR tree-optimization/56778
13677 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
13678 Runtime alias tests are not supported for gather loads.
13679 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
13680 stmts referenced from SSA operands before updating SSA form.
13682 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
13683 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13685 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
13686 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
13687 * config/arm/cortex-a53.md: New file.
13688 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
13689 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
13690 * config/arm/arm.c (arm_issue_rate): Likewise.
13691 * config/arm/arm-tune.md: Regenerate
13692 * config/arm/arm-tables.opt: Regenerate.
13693 * config/arm/arm-cores.def: Add cortex-a53.
13695 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
13697 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
13700 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
13702 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
13703 scalar load/store operations using B/H registers.
13704 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
13706 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
13708 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
13710 * config/aarch64/aarch64.c
13711 (aarch64_simd_scalar_immediate_valid_for_move): New.
13712 * config/aarch64/aarch64-protos.h
13713 (aarch64_simd_scalar_immediate_valid_for_move): New.
13714 * config/aarch64/constraints.md (Dh, Dq): New.
13715 * config/aarch64/iterators.md (hq): New.
13717 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
13719 * reorg.c (get_branch_condition): Deal with conditional returns.
13720 (fill_simple_delay_slots): Remove dead code dealing with jumps.
13722 2013-04-01 Wei Mi <wmi@google.com>
13724 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
13725 Truncate operand 2 using %b asm operand modifier.
13726 (*<shift_insn><mode>3_mask): Ditto.
13727 (*<rotate_insn><mode>3_mask): Ditto.
13729 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
13731 PR middle-end/56798
13732 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
13734 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
13736 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
13738 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
13740 2013-03-30 Lawrence Crowl <crowl@google.com>
13742 * dse.c (clear_alias_sets): Remove never set.
13743 (disqualified_clear_alias_sets): Remove never set.
13744 (clear_alias_mode_pool): Remove never set.
13745 (dse_step0): Remove condition that is never true.
13746 (canon_address): Remove condition that is never true.
13747 (dse_step7): Remove condition that is never true.
13748 (rest_of_handle_dse): Remove condition that is never true.
13749 (rest_of_handle_dse::did_global): Remove never read from above.
13750 (dse_step2_spill): Remove never called from above.
13751 (dse_step5_spill): Remove never called from above.
13753 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
13755 * doc/md.texi (Standard Names) <casesi>: Update documentation for
13756 JUMP_TABLE_DATA changes.
13757 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
13758 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
13759 (Insns) <jump_table_data>: New entry.
13760 * doc/tm.texi: Regenerate.
13762 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
13764 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
13765 for table jump at the end of a basic block using tablejump_p.
13766 * targhooks.c (default_invalid_within_doloop): Likewise.
13767 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
13768 target hook implementation that is identical to the default hook.
13769 (rs6000_invalid_within_doloop): Remove.
13771 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
13772 unused variable from tablejump_p call.
13774 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
13775 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
13776 (INSN_DELETED_P): Likewise.
13777 (emit_jump_table_data): New prototype.
13778 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
13779 after 4th as unused.
13780 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
13781 * sched-vis.c (print_insn): Likewise.
13782 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
13783 insn for compatibility with back ends that use next_active_insn to
13784 identify jump table data.
13785 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
13786 (remove_insn): Likewise.
13787 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
13789 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
13790 (emit_jump_table_data): New function.
13792 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
13793 basic block, a JUMP_TABLE_DATA never is.
13794 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
13795 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
13796 off from code handling real insns.
13797 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
13798 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
13799 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
13800 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
13801 is not a NONDEBUG_INSN_P.
13802 * ira-costs.c (scan_one_insn): Likewise.
13803 * jump.c (mark_all_labels): Likewise.
13804 (mark_jump_label_1): Likewise.
13805 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13806 * lra.c (get_insn_freq): Expect all insns reaching here to be in
13808 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
13809 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
13810 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
13811 JUMP_TABLE_DATA_P insns.
13812 (calculate_elim_costs_all_insns): Likewise.
13813 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
13814 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
13815 (delete_output_reload): Code style fixups.
13816 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
13817 insn flags on this non-insn.
13818 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
13819 as scheduling barriers, for pre-change compatibility.
13820 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
13821 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
13823 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
13824 redundant JUMP_TABLE_DATA_P test.
13825 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
13826 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13827 (frv_for_each_packet): Likewise.
13828 * config/i386/i386.c (min_insn_size): Likewise.
13829 (ix86_avoid_jump_mispredicts): Likewise.
13830 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13831 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
13832 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13833 (mips16_insn_length): Robustify.
13834 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
13835 (mips16_split_long_branches): Likewise.
13836 * config/pa/pa.c (pa_combine_instructions): Likewise.
13837 * config/rs6000/rs6000.c (get_next_active_insn): Treat
13838 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
13839 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
13840 as contributing to pool range lengths.
13841 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
13842 Remove redundant JUMP_TABLE_DATA_P test.
13843 (sh_loop_align): Likewise.
13844 (split_branches): Likewise.
13845 (sh_insn_length_adjustment): Likewise.
13846 * config/spu/spu.c (get_branch_target): Likewise.
13848 2013-03-29 Jan Hubicka <jh@suse.cz>
13850 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
13851 gcov streaming; stream hot bb threshold to ltrans.
13852 * predict.c (get_hot_bb_threshold): Break out from ....
13853 (maybe_hot_count_p): ... here.
13854 (set_hot_bb_threshold): New function.
13855 * lto-section-in.c (lto_section_name): Add profile.
13856 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
13857 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
13858 and data-streamer.h
13859 (histogram_entry): New structure.
13860 (histogram, histogram_pool): New global vars.
13861 (histogram_hash): New structure.
13862 (histogram_hash::hash): New method.
13863 (histogram_hash::equal): Likewise.
13864 (account_time_size): New function.
13865 (cmp_counts): New function.
13866 (dump_histogram): New function.
13867 (ipa_profile_generate_summary): New function.
13868 (ipa_profile_write_summary): New function.
13869 (ipa_profile_read_summary): New function.
13870 (ipa_profile): Decide on threshold.
13871 (pass_ipa_profile): Add ipa_profile_write_summary and
13872 ipa_profile_read_summary.
13873 * Makefile.in (ipa.o): Update dependencies.
13874 * lto-streamer.h (LTO_section_ipa_profile): New section.
13876 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
13878 * tree.h (VAR_P): New.
13880 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
13883 * doc/invoke.texi ([-fwhole-program]): Fix typo.
13885 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
13887 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
13888 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13889 (control_flow_insn_p): Likewise.
13890 * cfgrtl.c (duplicate_insn_chain): Likewise.
13891 * final.c (get_attr_length_1): Likewise.
13892 (shorten_branches): Likewise.
13893 (final_scan_insn): Likewise.
13894 * function.c (instantiate_virtual_regs): Likewise.
13895 * gcse.c (insert_insn_end_basic_block): Likewise.
13896 * ira-costs.c (scan_one_insn): Likewise.
13897 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13898 * lra.c (check_rtl): Likewise.
13899 * reload1.c (elimination_costs_in_insn): Likewise.
13900 * reorg.c (follow_jumps): Likewise.
13902 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
13903 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13904 (thumb_far_jump_used_p): Likewise.
13905 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
13906 (workaround_speculation): Likewise.
13907 (add_sched_insns_for_speculation): Likewise.
13908 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
13909 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13910 (frv_for_each_packet): Likewise.
13911 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
13912 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
13913 (final_emit_insn_group_barriers): Likewise.
13914 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13915 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13916 (mips16_insn_length): Likewise.
13917 * config/pa/pa.c (pa_reorg): Likewise.
13918 (pa_combine_instructions): Likewise.
13919 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
13920 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
13921 (sh_reorg): Likewise.
13922 (split_branches): Likewise.
13923 * config/spu/spu.c (get_branch_target): Likewise.
13925 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13928 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13930 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13931 Fix declaration name.
13933 2013-03-28 Lawrence Crowl <crowl@google.com>
13935 * graphds.h (struct graph.indicies): Remove unused.
13936 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13937 (SCOP_ORIGINAL_PDDRS): Remove unused.
13938 * sese.h (extern insert_loop_close_phis): Removed unused.
13939 (extern insert_guard_phis): Removed unused.
13940 (extern ivtype_map_elt_info): Removed unused.
13941 (new_ivtype_map_elt): Removed unused.
13942 * sese.c (ivtype_map_elt_info): Removed unused.
13944 2013-03-28 Lawrence Crowl <crowl@google.com>
13946 * Makefile.in: Add several missing include dependences.
13948 (test-dump.o): New. This object is not added to any executable,
13949 but is present for ad-hoc testing.
13951 (debug (const bitmap_head_def &)): New.
13952 (debug (const bitmap_head_def *)): New.
13954 (extern debug (const bitmap_head_def &)): New.
13955 (extern debug (const bitmap_head_def *)): New.
13957 (debug (edge_def &)): New.
13958 (debug (edge_def *)): New.
13960 (debug (basic_block_def &)): New.
13961 (debug (basic_block_def *)): New.
13963 (dump_node (const_tree, int, FILE *)): Correct source file.
13965 (debug (die_struct &)): New.
13966 (debug (die_struct *)): New.
13968 (extern debug (die_struct &)): New.
13969 (extern debug (die_struct *)): New.
13970 * gimple-pretty-print.c
13971 (debug (gimple_statement_d &)): New.
13972 (debug (gimple_statement_d *)): New.
13973 * gimple-pretty-print.h
13974 (extern debug (gimple_statement_d &)): New.
13975 (extern debug (gimple_statement_d *)): New.
13977 (debug (ira_allocno_copy &)): New.
13978 (debug (ira_allocno_copy *)): New.
13979 (debug (ira_allocno &)): New.
13980 (debug (ira_allocno *)): New.
13982 (extern debug (ira_allocno_copy &)): New.
13983 (extern debug (ira_allocno_copy *)): New.
13984 (extern debug (ira_allocno &)): New.
13985 (extern debug (ira_allocno *)): New.
13987 (debug (live_range &)): New.
13988 (debug (live_range *)): New.
13990 (debug (lra_live_range &)): New.
13991 (debug (lra_live_range *)): New.
13993 (debug (lra_live_range &)): New.
13994 (debug (lra_live_range *)): New.
13996 (debug (omega_pb_d &)): New.
13997 (debug (omega_pb_d *)): New.
13999 (extern debug (omega_pb_d &)): New.
14000 (extern debug (omega_pb_d *)): New.
14002 (debug (const rtx_def &)): New.
14003 (debug (const rtx_def *)): New.
14005 (debug_tree (tree): Move within file.
14006 (debug_raw (const tree_node &)): New.
14007 (debug_raw (const tree_node *)): New.
14008 (dump_tree_via_hooks (const tree_node *, int)): New.
14009 (debug (const tree_node &)): New.
14010 (debug (const tree_node *)): New.
14011 (debug_verbose (const tree_node &)): New.
14012 (debug_verbose (const tree_node *)): New.
14013 (debug_head (const tree_node &)): New.
14014 (debug_head (const tree_node *)): New.
14015 (debug_body (const tree_node &)): New.
14016 (debug_body (const tree_node *)): New.
14017 (debug_vec_tree (tree): Move and reimplement in terms of dump.
14018 (debug (vec<tree, va_gc> &)): New.
14019 (debug (vec<tree, va_gc> *)): New.
14021 (extern debug (const rtx_def &)): New.
14022 (extern debug (const rtx_def *)): New.
14024 (debug_raw (simple_bitmap_def &)): New.
14025 (debug_raw (simple_bitmap_def *)): New.
14026 (debug (simple_bitmap_def &)): New.
14027 (debug (simple_bitmap_def *)): New.
14029 (extern debug (simple_bitmap_def &)): New.
14030 (extern debug (simple_bitmap_def *)): New.
14031 (extern debug_raw (simple_bitmap_def &)): New.
14032 (extern debug_raw (simple_bitmap_def *)): New.
14034 (debug (vinsn_def &)): New.
14035 (debug (vinsn_def *)): New.
14036 (debug_verbose (vinsn_def &)): New.
14037 (debug_verbose (vinsn_def *)): New.
14038 (debug (expr_def &)): New.
14039 (debug (expr_def *)): New.
14040 (debug_verbose (expr_def &)): New.
14041 (debug_verbose (expr_def *)): New.
14042 (debug (vec<rtx> &)): New.
14043 (debug (vec<rtx> *)): New.
14045 (extern debug (vinsn_def &)): New.
14046 (extern debug (vinsn_def *)): New.
14047 (extern debug_verbose (vinsn_def &)): New.
14048 (extern debug_verbose (vinsn_def *)): New.
14049 (extern debug (expr_def &)): New.
14050 (extern debug (expr_def *)): New.
14051 (extern debug_verbose (expr_def &)): New.
14052 (extern debug_verbose (expr_def *)): New.
14053 (extern debug (vec<rtx> &)): New.
14054 (extern debug (vec<rtx> *)): New.
14056 (_list_iter_cond_expr): Make inline instead of static.
14058 (debug (sreal &)): New.
14059 (debug (sreal *)): New.
14061 (extern debug (sreal &)): New.
14062 (extern debug (sreal *)): New.
14064 (extern debug_raw (const tree_node &)): New.
14065 (extern debug_raw (const tree_node *)): New.
14066 (extern debug (const tree_node &)): New.
14067 (extern debug (const tree_node *)): New.
14068 (extern debug_verbose (const tree_node &)): New.
14069 (extern debug_verbose (const tree_node *)): New.
14070 (extern debug_head (const tree_node &)): New.
14071 (extern debug_head (const tree_node *)): New.
14072 (extern debug_body (const tree_node &)): New.
14073 (extern debug_body (const tree_node *)): New.
14074 (extern debug (vec<tree, va_gc> &)): New.
14075 (extern debug (vec<tree, va_gc> *)): New.
14077 (debug (struct loop &)): New.
14078 (debug (struct loop *)): New.
14079 (debug_verbose (struct loop &)): New.
14080 (debug_verbose (struct loop *)): New.
14081 * tree-dump.c: Add header dependence.
14083 (extern debug (struct loop &)): New.
14084 (extern debug (struct loop *)): New.
14085 (extern debug_verbose (struct loop &)): New.
14086 (extern debug_verbose (struct loop *)): New.
14088 (debug (data_reference &)): New.
14089 (debug (data_reference *)): New.
14090 (debug (vec<data_reference_p> &)): New.
14091 (debug (vec<data_reference_p> *)): New.
14092 (debug (vec<ddr_p> &)): New.
14093 (debug (vec<ddr_p> *)): New.
14095 (extern debug (data_reference &)): New.
14096 (extern debug (data_reference *)): New.
14097 (extern debug (vec<data_reference_p> &)): New.
14098 (extern debug (vec<data_reference_p> *)): New.
14099 (extern debug (vec<ddr_p> &)): New.
14100 (extern debug (vec<ddr_p> *)): New.
14102 (debug (pt_solution &)): New.
14103 (debug (pt_solution *)): New.
14105 (extern debug (pt_solution &)): New.
14106 (extern debug (pt_solution *)): New.
14108 (debug (_var_map &)): New.
14109 (debug (_var_map *)): New.
14110 (debug (tree_live_info_d &)): New.
14111 (debug (tree_live_info_d *)): New.
14113 (extern debug (_var_map &)): New.
14114 (extern debug (_var_map *)): New.
14115 (extern debug (tree_live_info_d &)): New.
14116 (extern debug (tree_live_info_d *)): New.
14118 2013-03-28 Jan Hubicka <jh@suse.cz>
14120 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
14122 2013-03-28 Ian Bolton <ian.bolton@arm.com>
14124 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
14125 record only when desired or required.
14127 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
14129 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
14130 *vec_extractv2di_1_rex64. Use x64 isa attribute.
14132 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14134 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
14135 (*andsi3_compare0_uxtw): New pattern.
14136 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
14137 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
14139 2013-03-28 Jan Hubicka <jh@suse.cz>
14141 * data-streamer-in.c (streamer_read_gcov_count): New function.
14142 * gimple-streamer-out.c: Include value-prof.h.
14143 (output_gimple_stmt): Output histogram.
14144 (output_bb): Use streamer_write_gcov_count.
14145 * value-prof.c: Include data-streamer.h
14146 (dump_histogram_value): Add HIST_TYPE_MAX.
14147 (stream_out_histogram_value): New function.
14148 (stream_in_histogram_value): New function.
14149 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
14150 (stream_out_histogram_value, stream_in_histogram_value): Declare.
14151 * data-streamer-out.c (streamer_write_gcov_count): New function.
14152 (streamer_write_gcov_count_stream): New function.
14153 * lto-cgraph.c (lto_output_edge): Update counter streaming.
14154 (lto_output_node): Likewise.
14155 (input_node, input_edge): Likewise.
14156 * lto-streamer-out.c (output_cfg): Update streaming.
14157 * lto-streamer-in.c (input_cfg): Likewise.
14158 * data-streamer.h (streamer_write_gcov_count,
14159 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
14160 * gimple-streamer-in.c: Include value-prof.h
14161 (input_gimple_stmt): Input histograms.
14162 (input_bb): Update profile streaming.
14164 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14166 * genmodes.c (emit_max_int): New function.
14167 (emit_insn_modes_h): Added call to emit_max_function.
14168 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
14170 * machmode.def: Fixed comment.
14172 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14174 * combine.c (try_combine): Removed useless assert.
14175 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
14177 2013-03-28 Marek Polacek <polacek@redhat.com>
14178 Richard Biener <rguenther@suse.de>
14180 PR tree-optimization/56695
14181 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
14182 build signed result of a vector comparison.
14183 * tree-cfg.c (verify_gimple_comparison): Check that a result
14184 of a vector comparison has signed type.
14186 2013-03-28 Richard Biener <rguenther@suse.de>
14188 PR tree-optimization/37021
14189 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
14190 do not restrict gaps between groups.
14191 * tree-vect-stmts.c (vectorizable_load): Properly account for
14192 a gap between groups.
14194 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
14196 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
14197 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
14200 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
14202 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
14203 * doc/extend.texi (Named Address Spaces): Ditto.
14204 (Variable Attributes): Ditto.
14206 2013-03-27 Kai Tietz <ktietz@redhat.com>
14208 * config.build: Add support for cygwin x64 target.
14209 * config.gcc: Likewise.
14210 * config.host: Likewise.
14211 * configure.ac: Likewise
14212 * configure: Regenerated.
14214 2013-03-27 Kai Tietz <ktietz@redhat.com>
14216 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
14217 * config/i386/t-cygwin-w64: New file.
14218 * config/i386/cygwin-w64.h: New file.
14219 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
14220 and add support for x64-cygwin target.
14221 (CPP_SPEC): Likewise.
14222 (CXX_WRAP_SPEC_LIST): Undefine before define.
14223 (LIBGCJ_SONAME): Use 15 as version.
14225 2013-03-27 Richard Biener <rguenther@suse.de>
14227 PR tree-optimization/56716
14228 * tree-ssa-structalias.c (perform_var_substitution): Adjust
14229 dumping for ref nodes.
14231 2013-03-27 Martin Jambor <mjambor@suse.cz>
14233 PR tree-optimization/55334
14234 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
14235 restricted pointers to arrays.
14237 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
14239 * Makefile.in (.SUFFIXES): Add .cc.
14240 (.c.o): Apply same recipe for implicit rule .cc.o.
14242 2013-03-27 Richard Biener <rguenther@suse.de>
14244 PR tree-optimization/37021
14245 * tree-vect-data-refs.c (vect_check_strided_load): Allow
14246 REALPART/IMAGPART_EXPRs around the supported refs.
14247 * tree-ssa-structalias.c (find_func_aliases): Assume that
14248 floating-point values are not used to transfer pointers.
14250 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
14252 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
14253 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
14254 * gcc/doc/tm.texi: Regenerate.
14255 * gcc/targhooks.h (default_has_ifunc_p): New.
14256 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
14257 * gcc/config/linux-protos.h: New file.
14258 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
14259 this hook for linux which disables support of indirect functions in
14261 * gcc/config/linux-android.c: New file.
14262 * gcc/config/t-linux-android.c: Ditto.
14263 * gcc/config.gcc: Added new object file linux-android.o.
14264 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
14265 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
14266 * gcc/varasm.c (do_assemble_alias): Likewise.
14267 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
14268 doesn't support indirect functions.
14269 * configure: Regenerate.
14271 2013-03-27 Bin Cheng <bin.cheng@arm.com>
14274 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
14275 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
14278 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14280 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
14282 2013-03-27 Terry Guo <terry.guo@arm.com>
14284 * config/arm/arm-cores.def: Added core cortex-r7.
14285 * config/arm/arm-tune.md: Regenerated.
14286 * config/arm/arm-tables.opt: Regenerated.
14287 * doc/invoke.texi: Added entry for core cortex-r7.
14289 2013-03-27 Walter Lee <walt@tilera.com>
14291 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
14292 double-decrement of next_scratch_regno.
14294 2013-03-27 Walter Lee <walt@tilera.com>
14296 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
14298 (insn_v1mulus): Ditto.
14299 (insn_v2muls): Ditto.
14301 2013-03-27 Walter Lee <walt@tilera.com>
14303 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
14304 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
14306 2013-03-27 Walter Lee <walt@tilera.com>
14308 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
14309 (*sibcall_value): Ditto.
14311 2013-03-27 Walter Lee <walt@tilera.com>
14313 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
14314 (insn_mnz_v8qi): ... this ...
14315 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
14317 (insn_v<n>mnz): Replaced by ...
14318 (insn_v1mnz): ... this ...
14319 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
14321 (insn_mz_<mode>): Replaced by ...
14322 (insn_mz_v8qi): ... this ...
14323 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
14325 (insn_v<n>mz): Replaced by ...
14326 (insn_v1mz): ... this ...
14327 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
14330 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
14332 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
14334 2013-03-26 Roland McGrath <mcgrathr@google.com>
14336 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
14337 than fprintf with a non-constant, non-format string.
14339 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14341 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
14342 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
14343 operand 0 predicate.
14344 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
14345 attribute. Use general_x64nomem_operand as operand 1 predicate.
14346 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
14347 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14348 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
14349 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14350 (mov<mode>_insv_1): Remove expander. Merge insn with
14351 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
14352 Use general_x64nomem_operand as operand 1 predicate.
14353 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
14354 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
14355 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14356 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
14357 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
14358 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
14359 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14360 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
14361 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
14362 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
14363 (general_x64nomem_operand): Ditto.
14365 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
14367 * config/rtems.opt: Add -pthread option.
14369 2013-03-26 Richard Biener <rguenther@suse.de>
14371 * alias.c (find_base_term): Avoid redundant and not used recursion.
14372 (base_alias_check): Get the initial base term from the caller.
14373 (true_dependence_1): Compute and pass base terms to base_alias_check.
14374 (write_dependence_p): Likewise.
14375 (may_alias_p): Likewise.
14377 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
14379 * config/aarch64/aarch64.c (aarch64_classify_address): Support
14380 PC-relative load in SI modes and above only.
14382 2013-03-26 Xinyu Qi <xyqi@marvell.com>
14384 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
14385 * config/arm/iwmmxt.md (WCGR0): Update.
14386 (WCGR1, WCGR2, WCGR3): Likewise.
14388 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14390 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
14391 Use x64 and nox64 isa attributes.
14393 2013-03-26 Richard Biener <rguenther@suse.de>
14395 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
14396 alignment computations and rely on get_object_alignment_1
14397 for the !TYPE_P case.
14398 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
14400 2013-03-26 Walter Lee <walt@tilera.com>
14402 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
14403 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
14405 2013-03-25 Jeff Law <law@redhat.com>
14407 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
14408 check for INTEGRAL_TYPE_P that was missing due to checking in
14409 wrong version of prior patch.
14411 2013-03-25 Walter Lee <walt@tilera.com>
14413 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
14414 TILEGX_INSN_SHUFFLEBYTES1.
14415 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
14417 (tilegx_builtins): Ditto.
14418 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
14420 2013-03-25 Walter Lee <walt@tilera.com>
14422 * config/tilegx/tilegx.md (floatsisf2): New pattern.
14423 (floatunssisf2): New pattern.
14424 (floatsidf2): New pattern.
14425 (floatunssidf2): New pattern.
14427 2013-03-25 Walter Lee <walt@tilera.com>
14429 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
14430 tests for constraint J, K, N, P.
14432 2013-03-25 Walter Lee <walt@tilera.com>
14434 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
14435 Use indirect/pcrel encoding.
14436 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
14439 2013-03-25 Steve Ellcey <sellcey@mips.com>
14441 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
14442 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
14443 * config/mips/mips.c (mips_option_override): Set IMADD default.
14444 * config/mips/mips.h (PTF_AVOID_IMADD): New.
14445 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
14446 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
14447 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
14448 * doc/invoke.texi (-mimadd/-mno-imadd): New.
14450 2013-03-25 Jeff Law <law@redhat.com>
14452 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
14453 slightly to avoid creating and folding useless trees. Simplify
14454 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
14456 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
14458 * config/i386/i386.md (*zero_extendsidi2): Merge with
14459 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
14460 * config/i386/predicates.md (x86_64_zext_operand): Rename from
14461 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
14462 targets. Clarify comment.
14464 2013-03-25 Martin Jambor <mjambor@suse.cz>
14466 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
14467 pass-through jump functions differently.
14468 (ipa_read_jump_function): Likewise. Also use setter functions to set
14471 2013-03-25 Martin Jambor <mjambor@suse.cz>
14473 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
14474 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
14476 (ipa_get_indirect_edge_target): New function.
14477 (devirtualization_time_bonus): New parameter known_aggs, pass it to
14478 ipa_get_indirect_edge_target. Update all callers.
14479 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
14480 ipa_get_indirect_edge_target_1 instead of calling
14481 ipa_get_indirect_edge_target.
14482 (create_specialized_node): Pass aggvlas to
14483 ipcp_discover_new_direct_edges.
14485 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14487 * config/arm/arm.md (f_sels, f_seld): New types.
14488 (*cmov<mode>): New pattern.
14489 * config/arm/predicates.md (arm_vsel_comparison_operator): New
14492 2013-03-25 Kai Tietz <ktietz@redhat.com>
14494 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
14495 POSIX-printf for mingw-hosted builds.
14497 2013-03-25 Richard Biener <rguenther@suse.de>
14499 PR middle-end/56694
14500 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
14501 must-not-throw stmt location.
14503 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14505 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
14506 Emit load-acquire versions when acq is true.
14507 (arm_emit_store_exclusive): Add rel parameter.
14508 Emit store-release versions when rel is true.
14509 (arm_split_compare_and_swap): Use acquire-release instructions
14511 of barriers when appropriate.
14512 (arm_split_atomic_op): Likewise.
14513 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
14514 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
14515 (VUNSPEC_SLX): Likewise.
14516 (VUNSPEC_LDA): Likewise.
14517 (VUNSPEC_STL): Likewise.
14518 * config/arm/sync.md (atomic_load<mode>): New pattern.
14519 (atomic_store<mode>): Likewise.
14520 (arm_load_acquire_exclusive<mode>): Likewise.
14521 (arm_load_acquire_exclusivesi): Likewise.
14522 (arm_load_acquire_exclusivedi): Likewise.
14523 (arm_store_release_exclusive<mode>): Likewise.
14525 2013-03-25 Catherine Moore <clm@codesourcery.com>
14527 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
14528 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
14529 * config/mip/predicates.md (lwsp_swsp_operand,
14530 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
14531 sb16_operand, db4_operand, db7_operand, ib3_operand,
14532 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
14533 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
14534 andi16_operand): New predicates.
14535 * config/mips/mips.md (compression): New attribute.
14536 (enabled): New attribute.
14537 (length): Consider compression in computing length.
14538 (shift_compression): New code attribute.
14539 (*add<mode>3): New operands. Record compression.
14540 (sub<mode>3): Likewise.
14541 (one_cmpl<mode>2): Likewise.
14542 (*and<mode>3): Likewise.
14543 (*ior<mode>3): Likewise.
14544 (unnamed pattern for xor): Likewise.
14545 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
14546 (*<optab><mode>3): Likewise.
14547 (*mov<mode>_internal: Likewise.
14548 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
14549 (mips_unsigned_immediate_p): New.
14550 (umips_lwsp_swsp_address_p): New.
14551 (m16_based_address_p): New.
14552 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
14553 (mips_unsigned_immediate_p): New prototype.
14554 (lwsp_swsp_address_p): New prototype.
14555 (m16_based_address_p): New prototype.
14556 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
14557 (mips_signed_immediate_p): New function.
14558 (m16_based_address_p): New function.
14559 (lwsp_swsp_address_p): New function.
14560 (mips_print_operand_punctuation): Recognize short delay slot insns
14561 for microMIPS.add<mode>3"
14563 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14566 * config/arm/iterators.md (v_cmp_result): New mode attribute.
14567 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
14569 2013-03-25 Richard Biener <rguenther@suse.de>
14571 PR tree-optimization/56689
14572 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
14575 2013-03-25 Richard Biener <rguenther@suse.de>
14577 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
14579 (memory_references): Likewise.
14580 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
14581 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
14582 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
14583 (gather_mem_refs_in_loops): Fold into ...
14584 (analyze_memory_references): ... this. Move initialization
14585 to tree_ssa_lim_initialize.
14586 (fill_always_executed_in): Rename to ...
14587 (fill_always_executed_in_1): ... this.
14588 (fill_always_executed_in): Move contains_call computation to
14589 this new function from ...
14590 (tree_ssa_lim_initialize): ... here.
14591 (tree_ssa_lim): Call fill_always_executed_in.
14593 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
14595 * postreload.c (reload_combine): Fix code detecting returns.
14597 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
14599 * function.c (emit_use_return_register_into_block): On cc0 targets,
14600 do not emit the sequence between cc0 setter and user.
14602 2013-03-25 Kai Tietz <ktietz@redhat.com>
14604 * config/i386/predicates.md (local_symbolic_operand): Interpret
14605 dll-imported symbols as none-local.
14607 2013-03-25 Richard Biener <rguenther@suse.de>
14609 * tree-ssa-loop-im.c (struct depend): Remove.
14610 (struct lim_aux_data): Make depends a vec of gimples.
14611 (free_lim_aux_data): Adjust.
14612 (add_dependency): Likewise.
14613 (set_level): Likewise.
14615 2013-03-25 Richard Biener <rguenther@suse.de>
14617 PR middle-end/56434
14618 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
14619 the pointer returned by calls with ECF_MALLOC set.
14621 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
14623 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
14625 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
14627 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
14628 using MMXMODE mode iterator.
14629 (*move<mode>_internal): Merge with *movv2sf_internal and
14630 *movv2sf_internal_rex64 using MMXMODE mode iterator.
14632 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
14634 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
14635 (record_last_mem_set_info): Likewise.
14637 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
14638 of XNEWVEC followed by memset.
14639 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
14641 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
14643 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
14644 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
14645 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
14646 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
14647 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
14648 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
14649 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
14650 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
14651 BARRIER_P instead of GET_CODE.
14653 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14655 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
14656 inaccuracy in the probing code.
14658 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
14659 (ctrapdi4): Likewise.
14661 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14663 * calls.c (expand_call): Add missing guard to code handling return
14664 of non-BLKmode structures in MSB.
14665 * function.c (expand_function_end): Likewise.
14667 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14669 * combine.c (try_combine): Adjust comment. Do not add the set of
14670 insn #0 if the destination indirectly is set or dies in insn #2.
14671 Tidy up code to distribute a new note.
14673 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
14675 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
14676 also for alternatives 16 and 17.
14678 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
14680 * config/i386/sse.md (*mov<mode>_internal): Merge with
14681 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
14682 Emit insn template depending on type attribute. Use
14683 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14684 movd instead of movq mnemonic for interunit moves. Rewrite mode
14685 attribute calculation. Remove unit attribute calculation.
14686 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
14687 Set prefix_data16 attribute for DImode ssemov types.
14688 Use Ym instead of y for SSE-MMX conversion alternatives.
14689 Reorder operand constraints.
14691 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
14693 * df.h (df_insn_delete): Adjust prototype.
14694 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
14695 and let it decide whether mark the basic block dirty.
14696 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
14697 * df-scan.c (df_insn_info_delete): New helper function, split
14698 off from df_insn_delete.
14699 (df_scan_free_bb_info): Use it.
14700 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
14702 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
14703 that the insn is actually an insn and it has a non-NULL basic block.
14704 Do not mark basic block dirty if only deleting a DEBUG_INSN.
14706 2013-03-22 Richard Biener <rguenther@suse.de>
14708 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
14710 (mem_ref_alloc): Do not allocate them.
14711 (refs_independent_p): Do not query or maintain a cache.
14713 2013-03-22 Richard Biener <rguenther@suse.de>
14715 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
14716 (gather_mem_refs_in_loops): Do not compute it.
14717 (analyze_memory_references): Do not allocate it.
14718 (tree_ssa_lim_finalize): Do not free it.
14719 (for_all_locs_in_loop): Do not query all_refs_in_loop.
14721 2013-03-22 Richard Biener <rguenther@suse.de>
14723 * is-a.h (as_a): Use gcc_checking_assert.
14725 2013-03-22 Ian Bolton <ian.bolton@arm.com>
14727 * config/aarch64/aarch64.c (aarch64_print_operand): New
14728 format specifier for printing a constant in hex.
14729 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
14730 format specifier for printing second operand.
14732 2013-03-22 Richard Biener <rguenther@suse.de>
14734 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
14736 (gather_mem_refs_in_loops): Perform store accumulation here.
14737 (create_vop_ref_mapping_loop): Remove.
14738 (create_vop_ref_mapping): Likewise.
14739 (analyze_memory_references): Initialize refs_stored_in_loop.
14740 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
14741 (record_indep_loop): Remove.
14742 (record_dep_loop): New function.
14743 (ref_indep_loop_p_1): Adjust to only walk over references
14744 in the loop, not its subloops.
14745 (ref_indep_loop_p): Rename to ...
14746 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
14747 maintaining a more fine-grained cache.
14748 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
14749 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
14751 2013-03-22 Richard Biener <rguenther@suse.de>
14753 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
14754 (struct mem_ref): Make accesses_in_loop a vec of a vec of
14755 aggregate mem_ref_loc.
14756 (free_mem_ref_locs): Inline into ...
14757 (memref_free): ... this and adjust.
14758 (mem_ref_alloc): Adjust.
14759 (mem_ref_locs_alloc): Remove.
14760 (record_mem_ref_loc): Adjust.
14761 (get_all_locs_in_loop): Rewrite into ...
14762 (for_all_locs_in_loop): ... this iterator.
14763 (rewrite_mem_ref_loc): New functor.
14764 (rewrite_mem_refs): Use for_all_locs_in_loop.
14765 (sm_set_flag_if_changed): New functor.
14766 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
14767 (ref_always_accessed): New functor.
14768 (ref_always_accessed_p): Use for_all_locs_in_loop.
14770 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14772 * tree-pass.h (PROP_gimple_lvec): New.
14773 * passes.c (dump_properties): Handle PROP_gimple_lvec.
14774 (init_optimization_passes): Move pass_lower_vector.
14775 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
14777 (pass_lower_vector): Provide PROP_gimple_lvec.
14778 (pass_lower_vector_ssa): Likewise.
14779 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
14781 2013-03-21 Mark Wielaard <mjw@redhat.com>
14783 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
14785 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14787 * config/i386/i386.md (*movdi_internal): Disparage slightly
14788 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
14789 conversion alternatives.
14791 2013-03-21 Jakub Jelinek <jakub@redhat.com>
14793 PR middle-end/48087
14794 * diagnostic.def (DK_WERROR): New kind.
14795 * diagnostic.h (werrorcount): Define.
14796 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
14797 promoted to DK_ERROR, increment DK_WERROR counter instead of
14799 * toplev.c (toplev_main): Call print_ignored_options even if
14800 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
14801 even if just werrorcount is non-zero.
14804 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
14806 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
14807 (string_cst_pool_decl): New function.
14808 (optimize_one_addr_into_implicit_ptr): New function.
14809 (resolve_addr_in_expr): Optimize DWARF location expression
14810 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
14811 which doesn't live in memory, but has DW_AT_location or
14812 DW_AT_const_value, or refers to a string literal, into
14813 DW_OP_GNU_implicit_pointer.
14814 (optimize_location_into_implicit_ptr): New function.
14815 (resolve_addr): If removing DW_AT_location of a variable because
14816 it was DW_OP_addr of address of the variable, but the variable doesn't
14817 live in memory, try to emit const value attribute for the initializer.
14819 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14821 * tree.h (VECTOR_TYPE_P): New macro.
14822 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
14823 TYPE_MODE): Use it.
14824 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
14825 VEC_COND_EXPR cannot be lvalues.
14826 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
14828 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14830 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
14831 Restrict the transformation to equal modes.
14833 2013-03-21 Richard Biener <rguenther@suse.de>
14835 PR tree-optimization/39326
14836 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
14837 (MEM_ANALYZABLE): Adjust.
14838 (record_mem_ref_loc): Move bitmap ops ...
14839 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
14840 unanalyzable refs, do not record locations for it.
14841 (analyze_memory_references): Allocate ref zero as shared
14843 (refs_independent_p): Do not test for unanalyzed mems here.
14844 (ref_indep_loop_p_1): Special-case disambiguation against
14845 the unanalyzed ref.
14846 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
14848 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
14850 * config/arm/arm-protos.h (tune_params): Add
14851 prefer_neon_for_64bits field.
14852 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
14853 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
14854 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
14855 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
14856 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
14857 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
14858 (arm_option_override): Handle -mneon-for-64bits new option.
14859 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
14860 (prefer_neon_for_64bits): Declare new variable.
14861 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
14862 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
14863 (arch_enabled): Handle new arch types. Remove support for onlya8
14865 (one_cmpldi2): Use new arch names.
14866 (zero_extend<mode>di2, extend<mode>di2): Ditto.
14867 * config/arm/arm.opt (mneon-for-64bits): Add option.
14868 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
14869 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
14870 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
14872 * doc/invoke.texi (-mneon-for-64bits): Document.
14874 2013-03-21 Richard Biener <rguenther@suse.de>
14876 PR tree-optimization/39326
14877 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
14878 (sort_bbs_in_loop_postorder_cmp): New function.
14879 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
14881 2013-03-21 Richard Biener <rguenther@suse.de>
14883 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
14884 (vect_insert_into_interleaving_chain): Likewise.
14885 (vect_drs_dependent_in_basic_block): Inline ...
14886 (vect_slp_analyze_data_ref_dependence): ... here. New function,
14888 (vect_analyze_data_ref_dependence): ... here. Simplify.
14889 (vect_check_interleaving): Simplify.
14890 (vect_analyze_data_ref_dependences): Likewise. Split out ...
14891 (vect_slp_analyze_data_ref_dependences): ... this new function.
14892 (dr_group_sort_cmp): New function.
14893 (vect_analyze_data_ref_accesses): Compute data-reference groups
14894 here instead of in vect_analyze_data_ref_dependence. Use
14895 a more efficient algorithm.
14896 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
14897 vect_slp_analyze_data_ref_dependences. Call
14898 vect_analyze_data_ref_accesses earlier.
14899 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
14900 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
14901 (vect_slp_analyze_data_ref_dependences): New prototype.
14903 2013-03-21 Richard Biener <rguenther@suse.de>
14905 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
14906 ref is stored in the loop.
14907 (find_refs_for_sm): Walk only over all stores.
14908 (store_motion_loop): Allocate from lim_bitmap_obstack.
14909 (store_motion): Likewise.
14911 2013-03-21 Richard Biener <rguenther@suse.de>
14913 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
14914 Update virtual SSA form.
14916 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14918 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
14919 * configure: Regenerate.
14920 * config.in: Regenerate.
14921 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
14922 if !HAVE_LD_EH_FRAME_CIEV3.
14924 2013-03-21 Richard Biener <rguenther@suse.de>
14926 * tree-cfg.c (verify_expr_no_block): New function.
14927 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14928 nor DECL_VALUE_EXPR have locations with associated blocks.
14929 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14930 (clear_unused_block_pointer): Remove code dealing with
14931 blocks in DECL_DEBUG_EXPR locations.
14933 2013-03-21 Richard Biener <rguenther@suse.de>
14935 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14936 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14937 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14938 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14939 instead of DECL_DEBUG_EXPR_IS_FROM.
14940 * gimplify.c (gimplify_modify_expr): Likewise.
14941 * tree-cfg.c (verify_expr_location_1): Likewise.
14942 * tree-complex.c (create_one_component_var): Likewise.
14943 * tree-sra.c (create_access_replacement): Likewise.
14944 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14945 (clear_unused_block_pointer): Likewise.
14946 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14947 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14948 * var-tracking.c (var_debug_decl): Likewise.
14949 (track_expr_p): Likewise.
14950 * tree-inline.c (add_local_variables): Likewise. Set
14951 DECL_HAS_DEBUG_EXPR_P after copying it.
14952 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14953 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14955 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14958 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14959 * configure: Regenerate.
14960 * config.in: Regenerate.
14961 * config/i386/i386.md (*movdf_internal): Use
14962 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14963 movd instead of movq mnemonic for interunit moves.
14964 (*movdi_internal): Ditto.
14966 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14968 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14969 (abd<mode>_3): New pattern.
14970 (aba<mode>_3): New pattern.
14971 (fabd<mode>_3): New pattern.
14973 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14975 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14976 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14977 occurrence of REGISTER_PREFIX as its empty string.
14979 2013-03-20 Jeff Law <law@redhat.com>
14981 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14982 addititional equivalences for equality comparisons between an SSA_NAME
14983 and a constant where the SSA_NAME was set from a widening conversion.
14985 2013-03-20 Walter Lee <walt@tilera.com>
14987 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14989 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14991 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14992 depending on type attribute.
14993 (*movti_internal): Ditto.
14994 (*movtf_internal): Ditto.
14995 (*movxf_internal): Ditto.
14996 (*movdf_internal): Ditto.
14997 (*movsf_internal): Ditto.
14999 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15001 * config/i386/i386.md (*movti_internal): Set prefix attribute to
15002 maybe_vex for sselog1 and ssemov types.
15003 (*movdi_internal): Reorder operand constraints.
15004 (*movsi_internal): Ditto. Set prefix attribute to
15005 maybe_vex for sselog1 and ssemov types.
15006 (*movtf_internal): Set prefix attribute to maybe_vex
15007 for sselog1 and ssemov types.
15008 (*movdf_internal): Ditto. Set prefix_data16 attribute for
15009 DImode ssemov types. Reorder operand constraints.
15010 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
15011 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
15012 attribute for SImode ssemov types. Reorder operand constraints.
15014 2013-03-20 Martin Jambor <mjambor@suse.cz>
15016 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
15017 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
15019 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
15021 * config/rs6000/predicates.md (indexed_address, update_address_mem
15022 update_indexed_address_mem): New predicates.
15023 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
15024 attribute for load/store instructions.
15025 * config/rs6000/dfp.md (movsd_store): Likewise.
15026 (movsd_load): Likewise.
15027 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
15028 (unnamed HI->DI extend define_insn): Likewise.
15029 (unnamed SI->DI extend define_insn): Likewise.
15030 (unnamed QI->SI extend define_insn): Likewise.
15031 (unnamed QI->HI extend define_insn): Likewise.
15032 (unnamed HI->SI extend define_insn): Likewise.
15033 (unnamed HI->SI extend define_insn): Likewise.
15034 (extendsfdf2_fpr): Likewise.
15035 (movsi_internal1): Likewise.
15036 (movsi_internal1_single): Likewise.
15037 (movhi_internal): Likewise.
15038 (movqi_internal): Likewise.
15039 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
15040 attribute for load/store instructions.
15041 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
15043 (mov<mode>_softfloat): Likewise.
15044 (mov<mode>_hardfloat32): Likewise.
15045 (mov<mode>_hardfloat64): Likewise.
15046 (mov<mode>_softfloat64): Likewise.
15047 (movdi_internal32): Likewise.
15048 (movdi_internal64): Likewise.
15049 (probe_stack_<mode>): Likewise.
15051 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
15053 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
15054 floating point, and decimal floating point to reload iterator.
15056 * config/rs6000/constraints.md (wl constraint): New constraints to
15057 return FLOAT_REGS if certain options are used to reduce the number
15058 of separate patterns that exist in the file.
15059 (wx constraint): Likewise.
15060 (wz constraint): Likewise.
15062 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15063 -mdebug=reg, print wg, wl, wx, and wz constraints.
15064 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15065 Initialize the reload functions for 64-bit binary/decimal floating
15067 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15068 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15069 create the buffer on the stack to overcome not having a 32-bit
15071 (rs6000_emit_move): Likewise.
15072 (rs6000_secondary_memory_needed_rtx): Likewise.
15073 (rs6000_alloc_sdmode_stack_slot): Likewise.
15074 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15075 via xxlxor, just like DFmode 0.0.
15077 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
15078 define as 1 if we are running on a power7 or newer.
15079 (enum r6000_reg_class_enum): Add new constraints.
15081 * config/rs6000/dfp.md (movsd): Delete, combine with binary
15082 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15083 with other moves by using conditional constraits (wg). Use LFIWZX
15084 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
15085 (movsd splitter): Likewise.
15086 (movsd_hardfloat): Likewise.
15087 (movsd_softfloat): Likewise.
15089 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
15090 binary and decimal floating point moves.
15091 (fmove_ok): New attributes to combine binary and decimal floating
15092 point moves, and to combine power6x (mfpgpr) moves along normal
15094 (real_value_to_target): Likewise.
15095 (f32_lr): Likewise.
15096 (f32_lm): Likewise.
15097 (f32_li): Likewise.
15098 (f32_sr): Likewise.
15099 (f32_sm): Likewise.
15100 (f32_si): Likewise.
15101 (movsf): Combine binary and decimal floating point moves. Combine
15102 power6x (mfpgpr) moves with other moves by using conditional
15103 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
15104 (mov<mode> for SFmode/SDmode); Likewise.
15105 (SFmode/SDmode splitters): Likewise.
15106 (movsf_hardfloat): Likewise.
15107 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
15108 (movsf_softfloat): Likewise.
15109 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
15111 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
15112 wx and wz constraints.
15114 * config/rs6000/constraints.md (wg constraint): New constraint to
15115 return FLOAT_REGS if -mmfpgpr (power6x) was used.
15117 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
15120 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15121 -mdebug=reg, print wg, wl, wx, and wz constraints.
15122 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15123 Initialize the reload functions for 64-bit binary/decimal floating
15125 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15126 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15127 create the buffer on the stack to overcome not having a 32-bit
15129 (rs6000_emit_move): Likewise.
15130 (rs6000_secondary_memory_needed_rtx): Likewise.
15131 (rs6000_alloc_sdmode_stack_slot): Likewise.
15132 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15133 via xxlxor, just like DFmode 0.0.
15135 * config/rs6000/dfp.md (movdd): Delete, combine with binary
15136 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15137 with other moves by using conditional constraits (wg). Use LFIWZX
15138 and STFIWX for loading SDmode on power7.
15139 (movdd splitters): Likewise.
15140 (movdd_hardfloat32): Likewise.
15141 (movdd_softfloat32): Likewise.
15142 (movdd_hardfloat64_mfpgpr): Likewise.
15143 (movdd_hardfloat64): Likewise.
15144 (movdd_softfloat64): Likewise.
15146 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
15147 64-bit binary and decimal floating point moves.
15148 (FMOVE64X): Likewise.
15149 (movdf): Combine 64-bit binary and decimal floating point moves.
15150 Combine power6x (mfpgpr) moves with other moves by using
15151 conditional constraits (wg).
15152 (mov<mode> for DFmode/DDmode): Likewise.
15153 (DFmode/DDmode splitters): Likewise.
15154 (movdf_hardfloat32): Likewise.
15155 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
15156 (movdf_softfloat32): Likewise.
15157 (movdf_hardfloat64_mfpgpr): Likewise.
15158 (movdf_hardfloat64): Likewise.
15159 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
15160 (movdf_softfloat64): Likewise.
15161 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
15162 (reload_<mode>_load): Move to later in the file so they aren't in
15163 the middle of the floating point move insns.
15164 (reload_<mode>_store): Likewise.
15166 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
15169 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
15170 constraint if -mdebug=reg.
15171 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
15172 Enable using dd reload support if needed.
15174 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
15175 binary and decimal floating point moves in rs6000.md.
15176 (movtd_internal): Likewise.
15178 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
15179 decimal floating point moves.
15181 (movtf_internal): Likewise.
15182 (mov<mode>_internal, TDmode/TFmode): Likewise.
15183 (movtf_softfloat): Likewise.
15184 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
15186 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
15187 movdi_internal64, using wg constraint for move direct operations.
15188 (movdi_internal64): Likewise.
15190 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
15191 MODES_TIEABLE_P for selected modes. Print the numerical value of
15192 the various virtual registers. Use GPR/FPR first/last values,
15193 instead of hard coding the register numbers. Print which modes
15194 have reload functions registered.
15195 (rs6000_option_override_internal): If -mdebug=reg, trace the options
15196 settings before/after setting cpu, target and subtarget settings.
15197 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
15198 and for secondary reload failures in rs6000_secondary_reload_inner.
15199 (rs6000_secondary_reload_fail): Likewise.
15200 (rs6000_secondary_reload_inner): Likewise.
15202 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
15203 macros for first/last GPR and FPR registers.
15204 (LAST_GPR_REGNO): Likewise.
15205 (FIRST_FPR_REGNO): Likewise.
15206 (LAST_FPR_REGNO): Likewise.
15208 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
15209 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
15210 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
15211 (vcond<mode><mode>): Likewise.
15212 (vcondu<mode><mode>): Likewise.
15213 (vector_gtu<mode>): Likewise.
15214 (vector_gte<mode>): Likewise.
15215 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
15216 to prevent the compiler from converting DImode operations to TImode.
15217 (ior<mode>3): Likewise.
15218 (and<mode>3): Likewise.
15219 (one_cmpl<mode>2): Likewise.
15220 (nor<mode>3): Likewise.
15221 (andc<mode>3): Likewise.
15223 * config/rs6000/constraints.md (wt constraint): New constraint
15224 that returns VSX_REGS if TImode is allowed in VSX registers.
15226 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
15227 constant under VSX.
15229 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
15230 similar to TImode, but it is restricted to being in the GPRs.
15232 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
15233 TImode to occupy a single VSX register.
15235 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
15236 -mvsx-timode for power7/power8.
15237 (power7 cpu): Likewise.
15238 (power8 cpu): Likewise.
15240 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
15241 sure that TFmode/TDmode take up two registers if they are ever
15242 allowed in the upper VSX registers.
15243 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
15245 (rs6000_init_hard_regno_mode_ok): Likewise.
15246 (rs6000_debug_reg_global): Add debugging for PTImode and wt
15247 constraint. Print if LRA is turned on.
15248 (rs6000_option_override_internal): Give an error if -mvsx-timode
15249 and VSX is not enabled.
15250 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
15251 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
15252 to reg+offset addressing. Use PTImode when checking offset
15253 addresses for validity.
15254 (reg_offset_addressing_ok_p): Likewise.
15255 (rs6000_legitimate_offset_address_p): Likewise.
15256 (rs6000_legitimize_address): Likewise.
15257 (rs6000_legitimize_reload_address): Likewise.
15258 (rs6000_legitimate_address_p): Likewise.
15259 (rs6000_eliminate_indexed_memrefs): Likewise.
15260 (rs6000_emit_move): Likewise.
15261 (rs6000_secondary_reload): Likewise.
15262 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
15263 reloads to fpr registers to continue to use reg+offset addressing,
15264 but 64-bit reloads to altivec registers need reg+reg addressing.
15265 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
15266 it. Treat LO_SUM like a PLUS operation.
15267 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
15268 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
15269 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
15270 registers to share a register with a smaller sized type, since VSX
15271 puts scalars in the upper 64-bits.
15272 (print_operand): Add support for PTImode.
15273 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
15274 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
15275 registers, but don't have arithmetic support.
15276 (rs6000_memory_move_cost): Add test for VSX.
15277 (rs6000_opt_masks): Add -mvsx-timode.
15279 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
15282 (VSr): Use wt constraint for TImode.
15283 (VSv): Drop TImode support.
15284 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
15285 (vsx_movti_64bit): Likewise.
15286 (vsx_movti_32bit): Likewise.
15287 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
15288 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
15289 one '?' on the appropriate output constraint. Do not allow TImode
15290 logical operations on 32-bit systems.
15291 (vsx_ior<mode>3): Likewise.
15292 (vsx_xor<mode>3): Likewise.
15293 (vsx_one_cmpl<mode>2): Likewise.
15294 (vsx_nor<mode>3): Likewise.
15295 (vsx_andc<mode>3): Likewise.
15296 (vsx_concat_<mode>): Likewise.
15297 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
15299 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
15300 OPTION_MASK_VSX_TIMODE.
15301 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
15302 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
15304 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
15305 (TI2 iterator): New iterator for TImode, PTImode.
15306 (wd mode attribute): Add values for vector types.
15307 (movti_string): Replace TI move operations with operations for TImode
15308 and PTImode. Add support for TImode being allowed in VSX registers.
15309 (mov<mode>_string, TImode/PTImode): Likewise.
15310 (movti_ppc64): Likewise.
15311 (mov<mode>_ppc64, TImode/PTImode): Likewise.
15312 (TI mode splitters): Likewise.
15314 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
15317 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
15319 PR tree-optimization/56355
15320 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
15321 Also handle integers with undefined overflow.
15323 2013-03-20 Catherine Moore <clm@codesourcery.com>
15324 Maciej W. Rozycki <macro@codesourcery.com>
15325 Tom de Vries <tom@codesourcery.com>
15326 Nathan Sidwell <nathan@codesourcery.com>
15327 Iain Sandoe <iain@codesourcery.com>
15328 Nathan Froyd <froydnj@codesourcery.com>
15329 Chao-ying Fu <fu@mips.com>
15331 * doc/extend.texi: (micromips, nomicromips, nocompression):
15332 Document new function attributes.
15333 * doc/invoke.texi (minterlink-compressed, mmicromips,
15334 m14k, m14ke, m14kec): Document new options.
15335 (minterlink-mips16): Update documentation.
15336 * doc/md.texi (ZC, ZD): Document new constraints.
15337 * configure.ac (gcc_cv_as_micromips): Check if linker
15338 supports the .set micromips directive.
15339 * configure: Regenerate.
15340 * config.in: Regenerate.
15341 * config/mips/mips-tables.opt: Regenerate.
15342 * config/mips/micromips.md: New file.
15343 * constraints.md (ZC, ZD): New constraints.
15344 * config/mips/predicates.md (movep_src_register): New predicate.
15345 (movep_src_operand): New predicate.
15346 (non_volatile_mem_operand): New predicate.
15347 * config/mips/mips.md (multimem): New type.
15348 (length): Differentiate between 17-bit and 18-bit branch offsets.
15349 (MOVEP1, MOVEP2): New mode iterator.
15350 (mov_<load>l): Use ZC constraint.
15351 (mov_<load>r): Likewise.
15352 (mov_<store>l): Likewise.
15353 (mov_<store>r): Likewise.
15354 (*branch_equality<mode>_inverted): Add microMIPS support.
15355 (*branch_equality<mode>): Likewise.
15356 (*jump_absolute): Likewise.
15357 (indirect_jump_<mode>): Likewise.
15358 (tablejump_<mode>): Likewise.
15359 (<optab>_internal): Likewise.
15360 (sibcall_internal): Likewise.
15361 (sibcall_value_internal): Likewise.
15362 (prefetch): Use constraint ZD.
15363 * config/mips/mips.opt (minterlink-compressed): New option.
15364 (minterlink-mips16): Now an alias for minterlink-compressed.
15365 (mmicromips): New option.
15366 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15367 (compare_and_swap_12): Likewise.
15368 (sync_add<mode>): Likewise.
15369 (sync_<optab>_12): Likewise.
15370 (sync_old_<optab>_12): Likewise.
15371 (sync_new_<optab>_12): Likewise.
15372 (sync_nand_12): Likewise.
15373 (sync_old_nand_12): Likewise.
15374 (sync_new_nand_12): Likewise.
15375 (sync_sub<mode>): Likewise.
15376 (sync_old_add<mode>): Likewise.
15377 (sync_old_sub<mode>): Likewise.
15378 (sync_new_add<mode>): Likewise.
15379 (sync_new_sub<mode>): Likewise.
15380 (sync_<optab><mode>): Likewise.
15381 (sync_old_<optab><mode>): Likewise.
15382 (sync_new_<optab><mode>): Likewise.
15383 (sync_nand<mode>): Likewise.
15384 (sync_old_nand<mode>): Likewise.
15385 (sync_new_nand<mode>): Likewise.
15386 (sync_lock_test_and_set<mode>): Likewise.
15387 (test_and_set_12): Likewise.
15388 (atomic_compare_and_swap<mode>): Likewise.
15389 (atomic_exchange<mode>_llsc): Likewise.
15390 (atomic_fetch_add<mode>_llsc): Likewise.
15391 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15392 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15393 (umips_save_restore_pattern_p): Likewise.
15394 (umips_load_store_pair_p): Likewise.
15395 (umips_output_load_store_pair): Likewise.
15396 (umips_movep_target_p): Likewise.
15397 (umips_12bit_offset_address_p): Likewise.
15398 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15399 (mips_base_mips16): Rename this...
15400 (mips_base_compression_flags): ...to this. Update all uses.
15401 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15402 (mips_mips16_decl_p): Delete.
15403 (mips_nomips16_decl_p): Delete.
15404 (mips_get_compress_on_flags): New function.
15405 (mips_get_compress_off_flags): New function.
15406 (mips_get_compress_mode): New function.
15407 (mips_get_compress_on_name): New function.
15408 (mips_get_compress_off_name): New function.
15409 (mips_insert_attributes): Support multiple compression types.
15410 (mips_merge_decl_attributes): Likewise.
15411 (umips_12bit_offset_address_p): New function.
15412 (mips_start_function_definition): Emit .set micromips directive.
15413 (mips_call_may_need_jalx_p): New function.
15414 (mips_function_ok_for_sibcall): Add microMIPS support.
15415 (mips_print_operand_punctuation): Support short delay slots and
15417 (umips_swm_mask, umips_swm_encoding): New.
15418 (umips_build_save_restore): New function.
15419 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15420 (was_mips16_p): Remove.
15421 (old_compression_mode): New.
15422 (mips_set_compression_mode): New function.
15423 (mips_set_current_function): Add microMIPS support.
15424 (mips_option_override): Likewise.
15425 (umips_save_restore_pattern_p): New function.
15426 (umips_output_save_restore): New function.
15427 (umips_load_store_pair_p_1): New function.
15428 (umips_load_store_pair_p): New function.
15429 (umips_output_load_store_pair_1): New function.
15430 (umips_output_load_store_pair): New function.
15431 (umips_movep_target_p) New function.
15432 (mips_prepare_pch_save): Add microMIPS support.
15433 * config/mips/mips.h (TARGET_COMPRESSION): New.
15434 (TARGET_CPU_CPP_BUILTINS): Update macro
15435 to use new compression flags and to support microMIPS.
15436 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15437 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15438 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15439 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15440 (ASM_SPEC): Support mmicromips and mno-micromips.
15441 (M16STORE_REG_P): New macro.
15442 (MIPS_CALL): Support TARGET_MICROMIPS.
15443 (MICROMIPS_J): New macro.
15444 (mips_base_mips16): Rename this...
15445 (mips_base_compression_flags): ...to this.
15446 (UMIPS_12BIT_OFFSET_P): New macro.
15447 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15448 (MULTILIB_DIRNAMES): Likewise.
15449 2013-03-20 Richard Biener <rguenther@suse.de>
15451 PR tree-optimization/56661
15452 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
15453 the result does not have to be distinct.
15455 2013-03-20 Richard Biener <rguenther@suse.de>
15457 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
15460 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15461 Steven Bosscher <steven@gcc.gnu.org>
15463 PR rtl-optimization/56605
15464 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
15466 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15469 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
15470 that require movd instead of movq.
15472 2013-03-20 Richard Biener <rguenther@suse.de>
15474 * tree-ssa-structalias.c (struct variable_info): Add pointer
15475 to the first field of an aggregate with sub-vars. Make
15476 this and the pointer to the next subfield its ID.
15477 (vi_next): New function.
15478 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
15479 storedanything_id, integer_id): Increment by one.
15480 (new_var_info, get_call_vi, lookup_call_clobber_vi,
15481 get_call_clobber_vi): Adjust.
15482 (solution_set_expand): Simplify and speedup.
15483 (solution_set_add): Inline into ...
15484 (set_union_with_increment): ... this. Adjust accordingly.
15485 (do_sd_constraint): Likewise.
15486 (do_ds_constraint): Likewise.
15487 (do_complex_constraint): Simplify.
15488 (build_pred_graph): Adjust.
15489 (solve_graph): Likewise. Simplify and speedup.
15490 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
15491 get_constraint_for_component_ref, get_constraint_for_1,
15492 first_vi_for_offset, first_or_preceding_vi_for_offset,
15493 create_function_info_for, create_variable_info_for_1,
15494 create_variable_info_for, intra_create_variable_infos): Adjust.
15495 (init_base_vars): Push NULL for ID zero.
15496 (compute_points_to_sets): Adjust.
15498 2013-03-20 Richard Biener <rguenther@suse.de>
15500 * cfgloop.c (verify_loop_structure): Streamline and avoid
15501 ICEing on corrupt loop tree.
15502 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
15505 2013-03-20 Richard Biener <rguenther@suse.de>
15507 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
15508 check whether an SSA update is needed.
15510 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
15512 * config/mips/constraints.md (T): Rename to...
15516 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
15517 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
15519 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15521 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
15522 (*subsi3_carryin_uxtw): Likewise.
15524 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15526 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
15527 (*rorsi3_insn_uxtw): Likewise.
15529 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15531 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
15532 (*extrsi5_insn_uxtw): Likewise.
15534 2013-03-19 Richard Biener <rguenther@suse.de>
15536 PR tree-optimization/56273
15537 * passes.c (init_optimization_passes): Move second VRP after DOM.
15539 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
15541 * config/i386/i386.md (*movti_internal): Merge from
15542 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
15543 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
15544 nox64 isa attributes.
15546 2013-03-18 Richard Biener <rguenther@suse.de>
15548 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
15550 (merge_node_constraints): Likewise.
15551 (build_succ_graph): Likewise.
15552 (valid_graph_edge): Inline into single caller.
15553 (unify_nodes): Likewise. Use bitmap_set_bit return value
15555 (scc_visit): Fix formatting and variable use.
15556 (do_sd_constraint): Use gcc_checking_assert.
15557 (do_ds_constraint): Likewise.
15558 (do_complex_constraint): Likewise.
15559 (condense_visit): Likewise. Cleanup.
15560 (dump_pred_graph): New function.
15561 (perform_var_substitution): Dump the pred-graph before
15562 variable substitution.
15563 (find_equivalent_node): Use gcc_checking_assert.
15564 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
15566 2013-03-18 Richard Biener <rguenther@suse.de>
15568 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
15569 Remove cond_expr_stmt_list argument and do not gimplify the
15571 (vect_loop_versioning): Adjust.
15572 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
15573 Cleanup to use less temporaries.
15574 (vect_create_data_ref_ptr): Cleanup.
15576 2013-03-18 Jakub Jelinek <jakub@redhat.com>
15578 PR tree-optimization/56635
15579 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
15580 require types_compatible_p types.
15582 2013-03-18 Nick Clifton <nickc@redhat.com>
15584 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
15585 spurious backslash.
15587 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
15588 Add missing line to comment describing function.
15590 2013-03-18 Richard Biener <rguenther@suse.de>
15592 PR tree-optimization/56210
15593 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
15594 Handle string / character search functions.
15595 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
15597 2013-03-18 Richard Biener <rguenther@suse.de>
15599 PR middle-end/56483
15600 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
15601 and implement properly.
15602 * gimple.h (gimple_cond_single_var_p): Remove.
15604 2013-03-18 Richard Biener <rguenther@suse.de>
15606 * tree-data-ref.h (find_data_references_in_loop): Declare.
15607 * tree-data-ref.c (get_references_in_stmt): Use a stack
15608 vector pre-allocated in the callers.
15609 (find_data_references_in_stmt): Adjust.
15610 (graphite_find_data_references_in_stmt): Likewise.
15611 (create_rdg_vertices): Likewise.
15612 (find_data_references_in_loop): Export.
15613 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
15614 Compute dependences here...
15615 (vect_analyze_data_refs): ...not here. When we encounter
15616 a non-vectorizable data reference in basic-block vectorization
15617 truncate the data reference vector. Do not bother to
15618 fixup data-dependence information for gather loads.
15619 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
15620 of data references, as reported.
15622 2013-03-18 Richard Biener <rguenther@suse.de>
15624 PR tree-optimization/3713
15625 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
15626 has_constants and expr.
15627 (stmt_has_constants): Properly valueize SSA names when deciding
15628 whether the stmt has constants.
15630 2013-03-18 Richard Biener <rguenther@suse.de>
15632 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
15633 whole function when there is nothing to do.
15634 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
15635 * tree-vectorizer.c (vectorize_loops): Update virtual and
15636 loop-closed SSA once.
15637 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
15639 2013-03-18 Richard Biener <rguenther@suse.de>
15641 PR middle-end/56113
15642 * domwalk.c (bb_postorder): New global static.
15643 (cmp_bb_postorder): New function.
15644 (walk_dominator_tree): Replace scheme imposing an order for
15645 visiting dominator sons by one sorting them at the time they
15646 are pushed on the stack.
15648 2013-03-18 Richard Biener <rguenther@suse.de>
15650 PR tree-optimization/39326
15651 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
15652 (struct mem_ref): Replace mem member with ao_ref typed member.
15653 (MEM_ANALYZABLE): Adjust.
15654 (memref_eq): Likewise.
15655 (mem_ref_alloc): Likewise.
15656 (gather_mem_refs_stmt): Likewise.
15657 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
15658 (execute_sm_if_changed_flag_set): Adjust.
15659 (execute_sm): Likewise.
15660 (ref_always_accessed_p): Likewise.
15661 (refs_independent_p): Likewise.
15662 (can_sm_ref_p): Likewise.
15664 2013-03-18 Jakub Jelinek <jakub@redhat.com>
15667 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
15668 return 1 even for !unsignedp.
15670 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15672 * config/i386/i386.md (isa): Add x64 and nox64.
15673 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
15674 (*pushtf): Enable *roF alternative for x64 isa only.
15675 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
15676 mode attribute of integer alternatives to DImode for TARGET_64BIT.
15677 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
15678 (*movtf_internal): Merge from *movtf_internal_rex64 and
15679 *movtf_internal_sse. Use x64 and nox64 isa attributes.
15680 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
15681 nox64 isa attributes.
15682 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
15683 nox64 isa attributes.
15684 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
15686 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15688 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15690 2013-03-17 Jakub Jelinek <jakub@redhat.com>
15693 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
15694 class names. Remove trailing comma after "ALL_REGS".
15696 2013-03-16 Jan Hubicka <jh@suse.cz>
15698 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
15699 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
15700 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
15701 of cgraph_get_create_node.
15702 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
15704 2013-03-16 Jason Merrill <jason@redhat.com>
15707 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
15708 with DW_AT_default_value.
15710 2013-03-16 Jakub Jelinek <jakub@redhat.com>
15712 * BASE-VER: Set to 4.9.0.
15714 2013-03-14 Andi Kleen <ak@linux.intel.com>
15717 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
15718 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
15719 Document _x* TSX intrinsics.
15721 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
15722 David Holsgrove <david.holsgrove@xilinx.com>
15724 * configure.ac: Add MicroBlaze TLS support detection.
15725 * configure: Regenerate.
15726 * config/microblaze/microblaze-protos.h
15727 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
15728 symbol_mentioned_p, label_mentioned_p): Add prototypes.
15729 * config/microblaze/microblaze.c (microblaze_address_type): Add
15730 ADDRESS_TLS and tls_reloc address types.
15731 (microblaze_address_info): Add tls_reloc.
15732 (TARGET_HAVE_TLS): Define.
15733 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
15734 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
15735 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
15736 load_tls_operand, microblaze_call_tls_get_addr,
15737 microblaze_legitimize_tls_address): New functions.
15738 (microblaze_classify_unspec): Handle UNSPEC_TLS.
15739 (get_base_reg): Use microblaze_tls_symbol_p.
15740 (microblaze_classify_address): Handle TLS.
15741 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
15742 label_mentioned_p and microblaze_tls_referenced_p.
15743 (microblaze_legitimize_address): Handle TLS.
15744 (microblaze_address_insns): Handle ADDRESS_TLS.
15745 (pic_address_needs_scratch): Handle TLS.
15746 (print_operand_address): Handle TLS.
15747 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
15748 (microblaze_expand_move): Handle TLS.
15749 (microblaze_legitimate_constant_p): Check
15750 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
15751 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
15752 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
15753 (PIC_OFFSET_TABLE_REGNUM): Set.
15754 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
15755 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
15756 (addsi3, movsi_internal2, movdf_internal): Update constraints
15757 * config/microblaze/predicates.md (arith_plus_operand): Define
15758 (move_operand): Redefine as move_src_operand,
15759 check microblaze_tls_referenced_p.
15761 2013-03-14 Ian Bolton <ian.bolton@arm.com>
15763 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
15764 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
15766 2013-03-14 Ian Bolton <ian.bolton@arm.com>
15768 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
15771 2013-03-14 Jakub Jelinek <jakub@redhat.com>
15773 PR tree-optimization/53265
15774 * common.opt (Waggressive-loop-optimizations): New option.
15775 * tree-ssa-loop-niter.c: Include tree-pass.h.
15776 (do_warn_aggressive_loop_optimizations): New function.
15777 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
15778 if number_of_latch_executions returned constant.
15779 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
15780 early. If number_of_latch_executions returned constant, set
15781 nb_iterations_upper_bound back to it.
15782 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
15784 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
15785 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
15787 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
15788 (MULTILIB_OSDIRNAMES): Set.
15789 * genmultilib: If defaultosdirname doesn't start with :: , set
15790 defaultosdirname2 instead, clear it and emit two . multilib_raw
15791 entries instead of just one.
15793 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
15795 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15796 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
15797 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15798 (SUBTARGET_OVERRIDE_OPTIONS): New.
15800 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
15803 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
15804 (musermode): Convert to Var(TARGET_USERMODE).
15805 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
15806 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
15807 * config/sh/sh.c (sh_option_override): Use
15808 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
15809 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
15811 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
15813 (udivsi3_i4_single, divsi3_i4_single): Use
15814 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
15816 2013-03-13 Dave Korn <dave.korn.cygwin@....>
15818 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
15821 2013-03-13 Richard Biener <rguenther@suse.de>
15823 PR tree-optimization/56608
15824 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
15825 calls when vectorizing basic-blocks.
15827 2013-03-13 Jakub Jelinek <jakub@redhat.com>
15830 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
15833 2013-03-12 Jakub Jelinek <jakub@redhat.com>
15835 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
15837 2013-03-11 Jan Hubicka <jh@suse.cz>
15840 * lto-streamer-out.c (output_symbol_p): Skip references from
15841 constructors of external variables.
15843 2013-03-11 Jan Hubicka <jh@suse.cz>
15845 PR middle-end/56571
15846 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
15848 * emit-rtl.c (verify_rtx_sharing): Likewise.
15849 (copy_insn_1): Likewise.
15850 * rtl.c (copy_rtx): Likewise.
15852 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
15855 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
15856 output_operand_lossage message.
15858 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
15861 * arm.c (shift_op): Validate RTL pattern on the fly.
15862 (arm_print_operand, case 'S'): Don't use shift_operator to validate
15865 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15868 * config/pa/pa.md (call_value): Check for calls to powf and direct to
15869 new call patterns that clobber %fr12.
15870 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
15871 split and postreload patterns.
15872 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
15873 registers %fr12 and %fr12R as call used.
15875 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
15877 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
15878 (canon_address, record_store, replace_read, check_mem_read_rtx,
15879 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
15880 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
15881 rest_of_handle_dse): Likewise.
15883 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
15885 PR middle-end/56524
15886 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
15888 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
15889 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
15890 (save_optabs_if_changed): Replace with...
15891 (init_tree_optimization_optabs): ...this.
15892 * optabs.c (save_optabs_if_changed): Rename to...
15893 (init_tree_optimization_optabs): ...this. Take the optimization node
15894 as argument. Do nothing if the base optabs are already correct.
15895 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
15896 to recompute optabs.
15897 * function.h (function): Remove optabs field.
15898 * function.c (invoke_set_current_function_hook): Call
15899 init_tree_optimization_optabs. Use the result to initialize
15902 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
15904 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
15905 if GTMA_HAS_NO_INSTRUMENTATION.
15906 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
15907 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
15908 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
15909 * gimple-pretty-print.c (dump_gimple_transaction): Handle
15910 GTMA_HAS_NO_INSTRUMENTATION.
15912 2013-03-08 Jakub Jelinek <jakub@redhat.com>
15914 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
15917 2013-03-08 Marek Polacek <polacek@redhat.com>
15918 Jakub Jelinek <jakub@redhat.com>
15920 PR tree-optimization/56478
15921 * predict.c (is_comparison_with_loop_invariant_p): Change the
15922 type of loop_step to tree.
15923 (predict_loops): Adjust.
15924 (predict_iv_comparison): Perform the computations on double_ints.
15926 2013-03-08 Richard Biener <rguenther@suse.de>
15928 PR tree-optimization/56570
15929 * tree-cfg.c (verify_expr_location_1): Verify locations for
15931 * tree-sra.c (create_access_replacement): Strip locations
15932 from DECL_DEBUG_EXPRs.
15934 2013-03-08 Richard Biener <rguenther@suse.de>
15936 * tree-inline.c (expand_call_inline): Do not associate
15937 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15938 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15940 2013-03-08 Richard Biener <rguenther@suse.de>
15942 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15943 or block changes with -Og. Fix for location / block encoding
15944 changes and PHI arguments with locations.
15946 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15948 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15950 (struct output_info): Likewise.
15951 (register_overhead): Remove bad gcc_assert.
15952 (bitmap_find_bit): If there is only a single bitmap element, do not
15953 count a miss as a search.
15954 (print_statistics): Update for counter type changes.
15955 (dump_bitmap_statistics): Likewise. Print headers such that they
15956 are properly lined up with the printed counters.
15958 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15960 PR tree-optimization/56559
15961 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15962 check that it has only a single use.
15964 2013-03-07 Richard Biener <rguenther@suse.de>
15966 * doc/invoke.texi (fwhole-program): Discourage use in combination
15969 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15971 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15973 PR tree-optimization/56539
15974 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15975 instead of GSI_CONTINUE_LINKING as last argument to
15976 force_gimple_operand_gsi. Adjust function comment.
15978 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15981 PR middle-end/56548
15982 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15983 promoted mode, convert the result back to the original mode.
15985 2013-03-06 Richard Biener <rguenther@suse.de>
15987 PR middle-end/56294
15988 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15989 (insert_updated_phi_nodes_compare_uids): New function.
15990 (update_ssa): Sort symbols_to_rename after UID before
15991 traversing it to insert PHI nodes.
15993 2013-03-06 Richard Biener <rguenther@suse.de>
15995 PR middle-end/50494
15996 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15997 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
16000 2013-02-13 Richard Biener <rguenther@suse.de>
16003 * varasm.c (output_constant_def_1): Get the decl representing
16004 the constant as argument.
16005 (output_constant_def): Wrap output_constant_def_1.
16006 (make_decl_rtl): Use output_constant_def_1 with the decl
16007 representing the constant.
16008 (build_constant_desc): Optionally re-use a decl already
16009 representing the constant.
16010 (tree_output_constant_def): Adjust.
16012 2013-03-06 Joey Ye <joey.ye@arm.com>
16015 * gcc.c (convert_white_space): New function.
16016 (main): Handles white space in function name.
16018 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
16021 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
16022 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
16023 to SH_DIV_CALL_TABLE for TARGET_SH2.
16024 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
16026 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
16027 call-table options.
16029 2013-03-05 Sterling Augustine <saugustine@google.com>
16030 Cary Coutant <ccoutant@google.com>
16033 * dwarf2out.c (resolve_addr): Don't call
16034 remove_loc_list_addr_table_entries a second time for the same
16037 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16040 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
16041 (avoid_complex_debug_insns): New function.
16042 (expand_debug_locations): Call it.
16044 PR rtl-optimization/56484
16045 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
16046 lifetimes of hard registers on small register class machines.
16048 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16050 * config/microblaze/microblaze-protos.h: Rename
16051 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
16052 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
16054 (microblaze_fast_interrupt_function_p): New function.
16055 (microblaze_is_interrupt_handler): Rename to
16056 microblaze_is_interrupt_variant and add fast_interrupt check.
16057 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
16058 (save_restore_insns): Likewise.
16059 (compute_frame_size): Likewise.
16060 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
16061 (microblaze_globalize_label): Likewise.
16062 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
16063 * config/microblaze/microblaze.md: Use wrapper
16064 microblaze_is_interrupt_variant.
16066 2013-03-05 Kai Tietz <ktietz@redhat.com>
16068 * sdbout.c (sdbout_one_type): Switch to current function's section
16069 supporting cold/hot.
16071 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16073 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
16076 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16078 PR middle-end/56461
16079 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
16080 if VALGRIND_GET_VBITS is defined, temporarily make object
16081 memory all defined, and restore previous valgrind addressability
16082 and definability afterwards. Free this_object at the end.
16084 PR middle-end/56461
16085 * lra.c (lra): Call lra_clear_live_ranges if live_p,
16086 right before calling lra_create_live_ranges, also call it
16087 when clearing live_p. Only call lra_clear_live_ranges
16088 at the end if live_p.
16090 PR middle-end/56461
16091 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
16093 2013-03-05 Richard Biener <rguenther@suse.de>
16095 PR tree-optimization/56521
16096 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
16099 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
16102 * except.h (remove_unreachable_eh_regions): New prototype.
16103 * except.c (remove_eh_handler_splicer): New function, split out
16104 of remove_eh_handler.
16105 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
16106 warning about running it on many EH regions one at a time.
16107 (remove_unreachable_eh_regions_worker): New function, walk the
16108 EH tree in depth-first order and remove non-marked regions.
16109 (remove_unreachable_eh_regions): New function.
16110 * tree-eh.c (mark_reachable_handlers): New function, split out
16111 from remove_unreachable_handlers.
16112 (remove_unreachable_handlers): Use mark_reachable_handlers and
16113 remove_unreachable_eh_regions.
16114 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
16115 and remove_unreachable_eh_regions.
16117 2013-03-05 Richard Biener <rguenther@suse.de>
16119 PR middle-end/56525
16120 * loop-init.c (fix_loop_structure): Remove loops in two stages,
16121 not freeing them until the end.
16123 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16125 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
16127 2013-03-05 Richard Biener <rguenther@suse.de>
16129 PR tree-optimization/56270
16130 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
16131 of loads after scheduling an SLP instance.
16133 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16135 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
16137 (check_gcc_parallelize): Run guality.exp as a separate job from
16138 vect.exp with unsorted.exp and $(dg_target_exps) separately from
16139 struct-layout-1.exp with stackalign.exp.
16141 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
16143 PR middle-end/56461
16144 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
16145 load_index sbitmap even if some bit in it isn't set.
16147 PR middle-end/56461
16148 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
16149 (discover_iteration_bound_by_body_walk): Change queues to
16150 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
16151 spelling in comment. Call safe_push on queues[bound_index] directly.
16152 Release queues[queue_index] in every iteration unconditionally.
16153 Release bounds vector.
16155 PR middle-end/56461
16156 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
16157 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
16158 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
16161 2013-03-05 Richard Biener <rguenther@suse.de>
16164 * tree-inline.c (remap_blocks_to_null): New function.
16165 (expand_call_inline): When expanding a call stmt without
16166 an associated block inline remap all callee blocks to NULL.
16168 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16170 PR rtl-optimization/56494
16171 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
16172 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
16173 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
16175 PR middle-end/56461
16176 * sel-sched-ir.c (free_sched_pools): Release
16177 succs_info_pool.stack[succs_info_pool.max_top] vectors too
16178 if succs_info_pool.max_top isn't -1.
16181 * opts.c (opts_obstack, opts_concat): Moved to...
16182 * opts-common.c (opts_obstack, opts_concat): ... here.
16184 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16186 PR middle-end/56461
16187 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
16189 2013-03-04 Martin Jambor <mjambor@suse.cz>
16191 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
16192 all appropriate places.
16194 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
16196 PR tree-optimization/56424
16197 * ipa-split.c (split_function): Do not set the RSO flag if result is
16198 not by reference and its type is a register type.
16200 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
16202 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
16203 (microblaze_legitimate_pic_operand): Likewise
16204 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
16205 new function microblaze_legitimate_pic_operand
16206 * config/microblaze/microblaze-protos.h
16207 (microblaze_legitimate_pic_operand): Declare.
16209 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16211 * config/microblaze/predicates.md (call_insn_simple_operand):
16212 New predicate for supported rtx code types.
16213 * config/microblaze/microblaze.md (call_internal1): Use
16214 call_insn_simple_operand predicate.
16216 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16218 PR middle-end/56461
16219 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
16220 partitions.ordered_remove.
16222 PR middle-end/56461
16223 * tree-vect-stmts.c (vectorizable_conversion): Don't call
16224 vec_oprnds0.create (1) for modifier == NONE.
16226 PR middle-end/56461
16227 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
16228 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
16229 vec_oprnds1 right before pushing anything to it for
16232 PR middle-end/56461
16233 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
16234 set nbbs to 0 instead of having separate code path.
16235 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
16236 instead of false as last argument if returning NULL.
16238 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
16240 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
16241 the attribute is now called "target" instead of "option".
16242 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
16243 * doc/tm.texi.in (Target Attributes): Likewise document the correct
16244 attribute/pragma name for TARGET_OPTION_VALID_P and
16245 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
16246 * doc/tm.texi: Regenerated.
16248 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
16250 * config/microblaze/microblaze.c:
16251 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
16252 * config/microblaze/microblaze.h: Add -mxl-reorder to
16254 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
16255 instructions emitted if TARGET_REORDER.
16256 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
16257 or 0 for -m/-mno case, but initialises as 2 to detect default use case
16260 2013-03-01 Xinliang David Li <davidxl@google.com>
16262 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
16265 2013-03-01 Jakub Jelinek <jakub@redhat.com>
16267 PR middle-end/56461
16268 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
16269 vector even when returning true. Fix up function comment formatting.
16271 PR middle-end/56461
16272 * ira-build.c (ira_loop_nodes_count): New variable.
16273 (create_loop_tree_nodes): Initialize it.
16274 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
16276 PR middle-end/56461
16277 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
16278 method on dr_chain and result_chain.
16279 * tree-vect-stmts.c (vectorizable_store): Only call
16280 result_chain.create if j == 0.
16282 PR middle-end/56461
16283 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
16284 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
16285 before overwriting it.
16287 2013-03-01 Tobias Burnus <burnus@net-b.de>
16289 * doc/extended.texi (C Extensions): Change order in @menu
16291 (Other MIPS Built-in Functions): Move last MIPS entry before
16292 "picoChip Built-in Functions".
16293 (SH Built-in Functions): Move after RX Built-in Functions.
16294 * doc/gcc.texi (Introduction): Change order in @menu
16296 * doc/md.texi (Constraints): Ditto.
16297 * gty.texi (Type Information): Ditto.
16298 (User-provided marking routines for template types): Make
16300 * doc/invoke.texi (AArch64 Options): Move before
16301 "Adapteva Epiphany Options".
16303 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
16304 Jakub Jelinek <jakub@redhat.com>
16307 * asan.c (gate_asan): Lookup no_sanitize_address instead of
16308 no_address_safety_analysis attribute.
16309 * doc/extend.texi (no_address_safety_attribute): Rename to
16310 no_sanitize_address attribute, mention no_address_safety_analysis
16311 attribute as deprecated alias.
16313 2013-02-28 Jakub Jelinek <jakub@redhat.com>
16315 PR middle-end/56461
16316 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
16317 type to vec<vec<tree> > *.
16318 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
16319 to be vec<tree> instead of vec<tree> *, set vec_defs
16320 to vNULL and call vec_defs.create (number_of_vects), adjust other
16322 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
16323 vectorizable_condition): Adjust vect_get_slp_defs callers.
16325 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16327 * config/aarch64/aarch64.c
16328 (aarch64_float_const_representable): Remove unused variable.
16330 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16332 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
16334 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16336 * config/aarch64/aarch64-builtins.c
16337 (aarch64_init_simd_builtins): Make static.
16339 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16341 * config/aarch64/aarch64.c
16342 (aarch64_simd_make_constant): Make static.
16344 2013-02-28 Martin Jambor <mjambor@suse.cz>
16346 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
16347 with no initialization to the RHS of debug statements.
16349 2013-02-28 Martin Jambor <mjambor@suse.cz>
16351 PR tree-optimization/56294
16352 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
16354 (get_access_replacement): Do not call create_access_replacement.
16355 Assert a replacement exists.
16356 (get_repl_default_def_ssa_name): Create the replacement declaration
16359 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16361 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
16362 final_end_function.
16364 2013-02-28 Marek Polacek <polacek@redhat.com>
16366 PR rtl-optimization/56466
16367 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
16368 if we're changing a loop.
16369 (peel_loops_completely): Likewise.
16371 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
16374 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
16376 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16379 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
16380 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
16381 INTX_FTYPE_FX, FX_FTYPE_INTX.
16382 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
16384 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16386 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
16387 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
16388 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
16389 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
16390 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
16391 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
16392 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
16393 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
16394 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
16395 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
16396 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
16397 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
16398 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
16399 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
16400 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
16401 (avrxmega6): Increase max flash segments from 5 to 6.
16402 * config/avr/t-multilib: Regenerate.
16403 * config/avr/avr-tables.opt: Regenerate.
16404 * doc/avr-mmcu.texi: Regenerate.
16406 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16408 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
16409 (avr_device_to_arch): Rename to avr_device_to_ld.
16410 (avr_device_to_as): New prototype.
16411 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
16412 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
16413 * config/avr/driver-avr.c (avr_device_to_as): New.
16414 (avr_device_to_arch): Rename to avr_device_to_ld.
16416 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16418 PR middle-end/56461
16419 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
16420 method on dr_chain and result_chain.
16422 PR middle-end/56461
16423 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
16424 pointer_set_destroy on not_executed_last_iteration.
16426 PR middle-end/56461
16427 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
16429 PR middle-end/56461
16430 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
16431 FOR_EACH_DEFINED_FUNCTION when freeing state.
16433 PR middle-end/56461
16434 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
16436 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
16439 PR middle-end/56461
16440 * ipa-cp.c (decide_whether_version_node): Call vec_free on
16441 known_aggs[i].items and release known_aggs vector.
16443 PR middle-end/56461
16444 * ipa-reference.c (propagate): Free node_info even for alias nodes.
16446 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16448 * config/microblaze/microblaze.c (microblaze_emit_compare):
16449 Use xor for EQ/NE comparisions.
16450 * config/microblaze/microblaze.md (cstoresf4): Add constraints
16451 (cbranchsf4): Adjust operator to comparison_operator.
16453 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16455 PR middle-end/56461
16456 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
16458 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
16459 vec_safe_push, always update *slot.
16460 (redirect_edge_var_map_clear): Use vec_free.
16461 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
16462 (free_var_map_entry): Use vec_free.
16463 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
16464 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
16466 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
16468 PR middle-end/45472
16469 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
16470 when the may_trap_p bit of the exprs being merged differs.
16471 Reorder tests for speculativeness in the logical and operator.
16473 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16475 * incpath.c (add_standard_paths): Use reconcat instead of concat
16476 where appropriate and avoid leaking memory.
16478 * opts.h: Include obstack.h.
16479 (opts_concat): New prototype.
16480 (opts_obstack): New declaration.
16481 * opts.c (opts_concat): New function.
16482 (opts_obstack): New variable.
16483 (init_options_struct): Call gcc_init_obstack on opts_obstack.
16484 (finish_options): Use opts_concat instead of concat
16485 and XOBNEWVEC instead of XNEWVEC.
16486 * opts-common.c (generate_canonical_option, decode_cmdline_option,
16487 generate_option): Likewise.
16488 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
16489 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
16492 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
16493 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
16495 2013-02-26 Jakub Jelinek <jakub@redhat.com>
16497 PR middle-end/56461
16498 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
16500 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
16502 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
16503 (arm_block_move_unaligned_straight): Likewise.
16504 (arm_adjust_block_mem): Likewise.
16506 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
16509 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
16510 temp, cond and label.
16511 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
16514 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
16515 * config/c6x/c6x.h (dbx_register_map): Update declaration.
16518 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
16519 of prologue/epilogue functions.
16522 * config/tilegx/tilegx.c (tilegx_expand_prologue):
16523 Remove unused variable cfa_offset.
16524 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
16527 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
16528 type promotion to unsigned.
16531 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
16532 for HOST_WIDE_INT of 32 bit / same size as int.
16533 (arm_block_move_unaligned_straight): Likewise.
16534 (arm_adjust_block_mem): Likewise.
16537 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
16540 2013-02-26 Marek Polacek <polacek@redhat.com>
16542 PR tree-optimization/56426
16543 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
16545 2013-02-26 Richard Biener <rguenther@suse.de>
16548 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
16549 unused variable loops.
16551 2013-02-26 Jakub Jelinek <jakub@redhat.com>
16553 PR tree-optimization/56448
16554 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
16555 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
16556 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
16557 later operands of the references, or even first operand for
16558 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
16560 PR tree-optimization/56443
16561 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
16562 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
16563 to type_for_mode langhook.
16565 2013-02-25 Matt Turner <mattst88@gmail.com>
16567 * doc/invoke.texi: Document r4700.
16569 2013-02-25 Richard Biener <rguenther@suse.de>
16571 PR tree-optimization/56175
16572 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
16574 (simplify_bitwise_binary): ... here. Also guard the conversion
16575 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
16577 2013-02-25 Catherine Moore <clm@codesourcery.com>
16580 2013-02-24 Catherine Moore <clm@codesourcery.com>
16581 Maciej W. Rozycki <macro@codesourcery.com>
16582 Tom de Vries <tom@codesourcery.com>
16583 Nathan Sidwell <nathan@codesourcery.com>
16584 Iain Sandoe <iain@codesourcery.com>
16585 Nathan Froyd <froydnj@codesourcery.com>
16586 Chao-ying Fu <fu@mips.com>
16588 * doc/extend.texi: (micromips, nomicromips, nocompression):
16589 Document new function attributes.
16590 * doc/invoke.texi (minterlink-compressed, mmicromips,
16591 m14k, m14ke, m14kec): Document new options.
16592 (minterlink-mips16): Update documentation.
16593 * doc/md.texi (ZC, ZD): Document new constraints.
16594 * configure.ac (gcc_cv_as_micromips): Check if linker
16595 supports the .set micromips directive.
16596 * configure: Regenerate.
16597 * config.in: Regenerate.
16598 * config/mips/mips-tables.opt: Regenerate.
16599 * config/mips/micromips.md: New file.
16600 * constraints.md (ZC, AD): New constraints.
16601 * config/mips/predicates.md (movep_src_register): New predicate.
16602 (movep_src_operand): New predicate.
16603 (non_volatile_mem_operand): New predicate.
16604 * config/mips/mips.md (multimem): New type.
16605 (length): Differentiate between 17-bit and 18-bit branch offsets.
16606 (MOVEP1, MOVEP2): New mode iterator.
16607 (mov_<load>l): Use ZC constraint.
16608 (mov_<load>r): Likewise.
16609 (mov_<store>l): Likewise.
16610 (mov_<store>r): Likewise.
16611 (*branch_equality<mode>_inverted): Add microMIPS support.
16612 (*branch_equality<mode>): Likewise.
16613 (*jump_absolute): Likewise.
16614 (indirect_jump_<mode>): Likewise.
16615 (tablejump_<mode>): Likewise.
16616 (<optab>_internal): Likewise.
16617 (sibcall_internal): Likewise.
16618 (sibcall_value_internal): Likewise.
16619 (prefetch): Use constraint ZD.
16620 * config/mips/mips.opt (minterlink-compressed): New option.
16621 (minterlink-mips16): Now an alias for minterlink-compressed.
16622 (mmicromips): New option.
16623 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16624 (compare_and_swap_12): Likewise.
16625 (sync_add<mode>): Likewise.
16626 (sync_<optab>_12): Likewise.
16627 (sync_old_<optab>_12): Likewise.
16628 (sync_new_<optab>_12): Likewise.
16629 (sync_nand_12): Likewise.
16630 (sync_old_nand_12): Likewise.
16631 (sync_new_nand_12): Likewise.
16632 (sync_sub<mode>): Likewise.
16633 (sync_old_add<mode>): Likewise.
16634 (sync_old_sub<mode>): Likewise.
16635 (sync_new_add<mode>): Likewise.
16636 (sync_new_sub<mode>): Likewise.
16637 (sync_<optab><mode>): Likewise.
16638 (sync_old_<optab><mode>): Likewise.
16639 (sync_new_<optab><mode>): Likewise.
16640 (sync_nand<mode>): Likewise.
16641 (sync_old_nand<mode>): Likewise.
16642 (sync_new_nand<mode>): Likewise.
16643 (sync_lock_test_and_set<mode>): Likewise.
16644 (test_and_set_12): Likewise.
16645 (atomic_compare_and_swap<mode>): Likewise.
16646 (atomic_exchange<mode>_llsc): Likewise.
16647 (atomic_fetch_add<mode>_llsc): Likewise.
16648 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16649 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16650 (umips_save_restore_pattern_p): Likewise.
16651 (umips_load_store_pair_p): Likewise.
16652 (umips_output_load_store_pair): Likewise.
16653 (umips_movep_target_p): Likewise.
16654 (umips_12bit_offset_address_p): Likewise.
16655 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16656 (mips_base_mips16): Rename this...
16657 (mips_base_compression_flags): ...to this. Update all uses.
16658 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16659 (mips_mips16_decl_p): Delete.
16660 (mips_nomips16_decl_p): Delete.
16661 (mips_get_compress_on_flags): New function.
16662 (mips_get_compress_off_flags): New function.
16663 (mips_get_compress_mode): New function.
16664 (mips_get_compress_on_name): New function.
16665 (mips_get_compress_off_name): New function.
16666 (mips_insert_attributes): Support multiple compression types.
16667 (mips_merge_decl_attributes): Likewise.
16668 (umips_12bit_offset_address_p): New function.
16669 (mips_start_function_definition): Emit .set micromips directive.
16670 (mips_call_may_need_jalx_p): New function.
16671 (mips_function_ok_for_sibcall): Add microMIPS support.
16672 (mips_print_operand_punctuation): Support short delay slots and
16674 (umips_swm_mask, umips_swm_encoding): New.
16675 (umips_build_save_restore): New function.
16676 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16677 (was_mips16_p): Remove.
16678 (old_compression_mode): New.
16679 (mips_set_compression_mode): New function.
16680 (mips_set_current_function): Add microMIPS support.
16681 (mips_option_override): Likewise.
16682 (umips_save_restore_pattern_p): New function.
16683 (umips_output_save_restore): New function.
16684 (umips_load_store_pair_p_1): New function.
16685 (umips_load_store_pair_p): New function.
16686 (umips_output_load_store_pair_1): New function.
16687 (umips_output_load_store_pair): New function.
16688 (umips_movep_target_p) New function.
16689 (mips_prepare_pch_save): Add microMIPS support.
16690 * config/mips/mips.h (TARGET_COMPRESSION): New.
16691 (TARGET_CPU_CPP_BUILTINS): Update macro
16692 to use new compression flags and to support microMIPS.
16693 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16694 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16695 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16696 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16697 (ASM_SPEC): Support mmicromips and mno-micromips.
16698 (M16STORE_REG_P): New macro.
16699 (MIPS_CALL): Support TARGET_MICROMIPS.
16700 (MICROMIPS_J): New macro.
16701 (mips_base_mips16): Rename this...
16702 (mips_base_compression_flags): ...to this.
16703 (UMIPS_12BIT_OFFSET_P): New macro.
16704 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16705 (MULTILIB_DIRNAMES): Likewise.
16707 2013-02-25 Tom de Vries <tom@codesourcery.com>
16709 PR rtl-optimization/56131
16710 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
16711 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
16712 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
16714 2013-02-25 Tobias Burnus <burnus@net-b.de>
16716 * doc/invoke.texi (-fsanitize=): Move from optimization
16717 to debugging options.
16719 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
16721 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
16723 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
16724 Alexander Monakov <amonakov@ispras.ru>
16726 PR middle-end/56077
16727 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
16728 flush pending lists also on non-jumps. Adjust comment.
16730 2013-02-24 Catherine Moore <clm@codesourcery.com>
16731 Maciej W. Rozycki <macro@codesourcery.com>
16732 Tom de Vries <tom@codesourcery.com>
16733 Nathan Sidwell <nathan@codesourcery.com>
16734 Iain Sandoe <iain@codesourcery.com>
16735 Nathan Froyd <froydnj@codesourcery.com>
16736 Chao-ying Fu <fu@mips.com>
16738 * doc/extend.texi: (micromips, nomicromips, nocompression):
16739 Document new function attributes.
16740 * doc/invoke.texi (minterlink-compressed, mmicromips,
16741 m14k, m14ke, m14kec): Document new options.
16742 (minterlink-mips16): Update documentation.
16743 * doc/md.texi (ZC, ZD): Document new constraints.
16744 * configure.ac (gcc_cv_as_micromips): Check if linker
16745 supports the .set micromips directive.
16746 * configure: Regenerate.
16747 * config.in: Regenerate.
16748 * config/mips/mips-tables.opt: Regenerate.
16749 * config/mips/micromips.md: New file.
16750 * constraints.md (ZC, AD): New constraints.
16751 * config/mips/predicates.md (movep_src_register): New predicate.
16752 (movep_src_operand): New predicate.
16753 (non_volatile_mem_operand): New predicate.
16754 * config/mips/mips.md (multimem): New type.
16755 (length): Differentiate between 17-bit and 18-bit branch offsets.
16756 (MOVEP1, MOVEP2): New mode iterator.
16757 (mov_<load>l): Use ZC constraint.
16758 (mov_<load>r): Likewise.
16759 (mov_<store>l): Likewise.
16760 (mov_<store>r): Likewise.
16761 (*branch_equality<mode>_inverted): Add microMIPS support.
16762 (*branch_equality<mode>): Likewise.
16763 (*jump_absolute): Likewise.
16764 (indirect_jump_<mode>): Likewise.
16765 (tablejump_<mode>): Likewise.
16766 (<optab>_internal): Likewise.
16767 (sibcall_internal): Likewise.
16768 (sibcall_value_internal): Likewise.
16769 (prefetch): Use constraint ZD.
16770 * config/mips/mips.opt (minterlink-compressed): New option.
16771 (minterlink-mips16): Now an alias for minterlink-compressed.
16772 (mmicromips): New option.
16773 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16774 (compare_and_swap_12): Likewise.
16775 (sync_add<mode>): Likewise.
16776 (sync_<optab>_12): Likewise.
16777 (sync_old_<optab>_12): Likewise.
16778 (sync_new_<optab>_12): Likewise.
16779 (sync_nand_12): Likewise.
16780 (sync_old_nand_12): Likewise.
16781 (sync_new_nand_12): Likewise.
16782 (sync_sub<mode>): Likewise.
16783 (sync_old_add<mode>): Likewise.
16784 (sync_old_sub<mode>): Likewise.
16785 (sync_new_add<mode>): Likewise.
16786 (sync_new_sub<mode>): Likewise.
16787 (sync_<optab><mode>): Likewise.
16788 (sync_old_<optab><mode>): Likewise.
16789 (sync_new_<optab><mode>): Likewise.
16790 (sync_nand<mode>): Likewise.
16791 (sync_old_nand<mode>): Likewise.
16792 (sync_new_nand<mode>): Likewise.
16793 (sync_lock_test_and_set<mode>): Likewise.
16794 (test_and_set_12): Likewise.
16795 (atomic_compare_and_swap<mode>): Likewise.
16796 (atomic_exchange<mode>_llsc): Likewise.
16797 (atomic_fetch_add<mode>_llsc): Likewise.
16798 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16799 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16800 (umips_save_restore_pattern_p): Likewise.
16801 (umips_load_store_pair_p): Likewise.
16802 (umips_output_load_store_pair): Likewise.
16803 (umips_movep_target_p): Likewise.
16804 (umips_12bit_offset_address_p): Likewise.
16805 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16806 (mips_base_mips16): Rename this...
16807 (mips_base_compression_flags): ...to this. Update all uses.
16808 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16809 (mips_mips16_decl_p): Delete.
16810 (mips_nomips16_decl_p): Delete.
16811 (mips_get_compress_on_flags): New function.
16812 (mips_get_compress_off_flags): New function.
16813 (mips_get_compress_mode): New function.
16814 (mips_get_compress_on_name): New function.
16815 (mips_get_compress_off_name): New function.
16816 (mips_insert_attributes): Support multiple compression types.
16817 (mips_merge_decl_attributes): Likewise.
16818 (umips_12bit_offset_address_p): New function.
16819 (mips_start_function_definition): Emit .set micromips directive.
16820 (mips_call_may_need_jalx_p): New function.
16821 (mips_function_ok_for_sibcall): Add microMIPS support.
16822 (mips_print_operand_punctuation): Support short delay slots and
16824 (umips_swm_mask, umips_swm_encoding): New.
16825 (umips_build_save_restore): New function.
16826 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16827 (was_mips16_p): Remove.
16828 (old_compression_mode): New.
16829 (mips_set_compression_mode): New function.
16830 (mips_set_current_function): Add microMIPS support.
16831 (mips_option_override): Likewise.
16832 (umips_save_restore_pattern_p): New function.
16833 (umips_output_save_restore): New function.
16834 (umips_load_store_pair_p_1): New function.
16835 (umips_load_store_pair_p): New function.
16836 (umips_output_load_store_pair_1): New function.
16837 (umips_output_load_store_pair): New function.
16838 (umips_movep_target_p) New function.
16839 (mips_prepare_pch_save): Add microMIPS support.
16840 * config/mips/mips.h (TARGET_COMPRESSION): New.
16841 (TARGET_CPU_CPP_BUILTINS): Update macro
16842 to use new compression flags and to support microMIPS.
16843 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16844 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16845 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16846 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16847 (ASM_SPEC): Support mmicromips and mno-micromips.
16848 (M16STORE_REG_P): New macro.
16849 (MIPS_CALL): Support TARGET_MICROMIPS.
16850 (MICROMIPS_J): New macro.
16851 (mips_base_mips16): Rename this...
16852 (mips_base_compression_flags): ...to this.
16853 (UMIPS_12BIT_OFFSET_P): New macro.
16854 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16855 (MULTILIB_DIRNAMES): Likewise.
16857 2013-02-24 Jakub Jelinek <jakub@redhat.com>
16860 * target-globals.c (save_target_globals): For init_reg_sets and
16861 target_reinit remporarily set this_fn_optabs to this_target_optabs.
16863 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
16865 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
16866 * config/aarch64/t-aarch64
16867 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
16869 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
16871 PR inline-asm/56148
16872 * lra-constraints.c (process_alt_operands): Reload operand
16873 conflicting with earlier clobber only if no more other conflicting
16876 2013-02-22 Jakub Jelinek <jakub@redhat.com>
16879 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
16880 if not linking a shared library.
16882 2013-02-22 Seth LaForge <sethml@google.com>
16884 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
16886 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
16888 * config/arm/arm.md (split for extendsidi): Update condition.
16889 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
16890 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
16891 (qhs_zextenddi_cstr): Likewise.
16893 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16895 PR middle-end/56420
16896 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
16897 avoid signed wrapping.
16898 (expand_mult): Handle properly multiplication by
16899 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
16900 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
16901 in the compiler if coeff is HOST_WIDE_INT_MIN.
16902 (expand_divmod): Don't make ext_op1 static, change it's type to
16903 uhwi. Avoid undefined behavior in -INTVAL (op1).
16905 PR rtl-optimization/50339
16906 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
16908 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
16909 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
16910 into splitting_ashiftrt field.
16911 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
16913 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
16916 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
16918 PR middle-end/56108
16919 * trans-mem.c (execute_tm_mark): Do not expand transactions that
16920 are sure to go irrevocable.
16922 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
16924 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16925 scalars are valid operands.
16927 2013-02-21 Martin Jambor <mjambor@suse.cz>
16929 PR tree-optimization/56310
16930 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16931 only matching indices and non-negative final offsets.
16932 (intersect_aggregates_with_edge): Pass src_idx to
16933 agg_replacements_to_vector. Pass src_idx insstead of index to
16934 intersect_with_agg_replacements.
16936 2013-02-21 Martin Jambor <mjambor@suse.cz>
16938 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16939 instead of hard-wired defaults.
16941 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16943 * doc/invoke.texi (MIPS Options): Update documentation of the
16944 floating-point multiply-accumulate instruction restrictions.
16946 2013-02-21 Kostya Serebryany <kcc@google.com>
16948 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16949 asan_shadow_offset on x86_64 linux.
16951 2013-02-21 Richard Biener <rguenther@suse.de>
16953 PR tree-optimization/56415
16955 2013-02-11 Richard Biener <rguenther@suse.de>
16957 PR tree-optimization/56273
16958 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16961 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16964 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16967 PR inline-asm/56405
16968 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16969 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16971 2013-02-20 Jan Hubicka <jh@suse.cz>
16973 PR tree-optimization/56265
16974 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16975 when target is referenced for first time.
16977 2013-02-20 Richard Biener <rguenther@suse.de>
16979 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16980 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16981 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16982 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16983 not return anything.
16984 (rename_ssa_copies): Do not remove unused locals.
16985 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16986 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16987 * passes.c (execute_function_todo): Do not schedule unused locals
16988 removal if cleanup_tree_cfg did something.
16989 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16990 about the number of removed locals.
16992 2013-02-20 Richard Biener <rguenther@suse.de>
16994 PR tree-optimization/56398
16995 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16997 2013-02-20 Martin Jambor <mjambor@suse.cz>
16999 PR tree-optimization/55334
17000 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
17001 restricted pointers to arrays.
17003 2013-02-20 Richard Biener <rguenther@suse.de>
17004 Jakub Jelinek <jakub@redhat.com>
17006 PR tree-optimization/56396
17007 * tree-ssa-ccp.c (n_const_val): New static variable.
17008 (get_value): Return NULL for SSA names we don't have a lattice
17010 (ccp_initialize): Initialize n_const_val.
17011 * tree-ssa-copy.c (n_copy_of): New static variable.
17012 (init_copy_prop): Initialize n_copy_of.
17013 (get_value): Return NULL_TREE for SSA names we don't have a
17016 2013-02-20 Martin Jambor <mjambor@suse.cz>
17018 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
17020 2013-02-20 Richard Biener <rguenther@suse.de>
17022 * genpreds.c (write_lookup_constraint): Do not compare first
17023 letter of the constraint again.
17025 2013-02-20 Richard Biener <rguenther@suse.de>
17027 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
17029 (get_use_iv_cost): Terminate hashtable walk when coming across
17032 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
17034 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
17035 reassociation for avx2 targets.
17037 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17039 * config/microblaze/microblaze.c: microblaze_has_clz = 0
17040 Add version check for v8.10.a to enable microblaze_has_clz
17041 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
17042 version and TARGET_PATTERN_COMPARE check
17043 * config/microblaze/microblaze.md: New clzsi2 instruction
17045 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17047 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
17048 function before branching.
17050 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17052 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
17054 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
17056 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17058 PR middle-end/55889
17059 * sel-sched.c: Include ira.h.
17060 (implicit_clobber_conflict_p): New function.
17061 (moveup_expr): Use it.
17062 * Makefile.in (sel-sched.o): Depend on ira.h.
17064 2013-02-19 Richard Biener <rguenther@suse.de>
17066 PR tree-optimization/56384
17067 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
17068 (vn_hash_type): Split out from ...
17069 (vn_hash_constant_with_type): ... here.
17070 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
17071 (vn_phi_eq): Compare types from vn_phi_s structure.
17072 (vn_phi_lookup): Populate vn_phi_s type.
17073 (vn_phi_insert): Likewise.
17075 2013-02-19 Jakub Jelinek <jakub@redhat.com>
17077 PR tree-optimization/56350
17078 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
17079 if haven't found reduction or nested cycle operand, rather than
17080 asserting we must find it.
17082 PR tree-optimization/56381
17083 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
17086 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
17087 Jakub Jelinek <jakub@redhat.com>
17090 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
17091 (swap_optab_enable): Same.
17092 (init_all_optabs): Use argument instead of global.
17093 * tree.h (struct tree_optimization_option): New field target_optabs.
17094 * expr.h (init_all_optabs): Add argument to prototype.
17095 (TREE_OPTIMIZATION_OPTABS): New.
17096 (save_optabs_if_changed): Protoize.
17097 * optabs.h: Declare this_fn_optabs.
17098 * optabs.c (save_optabs_if_changed): New.
17099 Declare this_fn_optabs.
17100 (init_optabs): Add argument to init_all_optabs() call.
17101 * function.c (invoke_set_current_function_hook): Handle per
17103 * function.h (struct function): New field optabs.
17104 * config/mips/mips.c (mips_set_mips16_mode): Handle when
17105 optimization_current_node has changed.
17106 * target-globals.h (save_target_globals_default_opts): Protoize.
17107 * target-globals.c (save_target_globals_default_opts): New.
17109 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17112 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
17113 registers %fr12 and %fr12R as call used.
17116 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
17117 and HImode, require all displacements to be an integer multiple of
17119 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
17120 only allow QImode and HImode when reload is in progress and strict is
17121 true. Likewise for symbolic addresses. Use base14_operand to check
17122 displacements in REG+BASE addresses.
17124 2013-02-18 Richard Biener <rguenther@suse.de>
17126 PR tree-optimization/56366
17127 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17128 handle sign-conversion of outer-loop initial induction value.
17130 2013-02-18 Richard Biener <rguenther@suse.de>
17132 PR middle-end/56349
17133 * cfghooks.c (merge_blocks): If we merge a latch into another
17134 block adjust references to it.
17135 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
17136 (verify_loop_structure): Verify that a recorded latch is in fact
17139 2013-02-18 Richard Biener <rguenther@suse.de>
17141 PR tree-optimization/56321
17142 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
17143 order SSA name release and virtual operand unlinking.
17145 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17147 * config/microblaze/microblaze.md (save_stack_block): Define.
17148 (restore_stack_block): Likewise.
17150 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17152 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
17153 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
17154 * config/microblaze/microblaze.c (microblaze_option_override):
17155 Bail out early for PIC modes when target does not support PIC.
17157 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17159 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
17160 Replace with a microblaze version.
17161 (microblaze_trampoline_init): Adapt for microblaze.
17162 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
17165 2013-02-16 Jakub Jelinek <jakub@redhat.com>
17166 Dodji Seketeli <dodji@redhat.com>
17169 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
17170 (instrument_mem_region_access): Do not forget to always put
17171 instrumentation of the of 'base' and 'base + len' in a "if (len !=
17172 0) statement, even for cases where either 'base' or 'base + len'
17173 are not instrumented -- because they have been previously
17174 instrumented. Simplify the logic by putting all the statements
17175 instrument 'base + len' inside a sequence, and then insert that
17176 sequence right before the current insertion point. Then, to
17177 instrument 'base + len', just get an iterator on that statement.
17178 And do not forget to update the pointer to iterator the function
17179 received as argument.
17181 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
17183 PR rtl-optimization/56348
17184 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
17186 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
17188 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
17189 (clean_graph_dump_file): Pass base to start_graph_dump.
17191 2013-02-14 Richard Henderson <rth@redhat.com>
17194 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
17196 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
17198 * collect2-aix.h: Define F_LOADONLY.
17200 2013-02-14 Richard Biener <rguenther@suse.de>
17203 * varasm.c (output_constant_def_1): Get the decl representing
17204 the constant as argument.
17205 (output_constant_def): Wrap output_constant_def_1.
17206 (make_decl_rtl): Use output_constant_def_1 with the decl
17207 representing the constant.
17208 (build_constant_desc): Optionally re-use a decl already
17209 representing the constant.
17210 (tree_output_constant_def): Adjust.
17212 2013-02-14 Dodji Seketeli <dodji@redhat.com>
17215 * asan.c (instrument_builtin_call): Really put the length of the
17216 second source argument into src1_len.
17218 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17220 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
17221 argument. If it is false, don't create edge from then_bb to
17223 (insert_if_then_before_iter): Pass true to it.
17224 (build_check_stmt): Pass false to it.
17225 (transform_statements): Flush hash table only on extended basic
17226 block boundaries, rather than at the beginning of every bb.
17227 Don't flush hash table on nonfreeing_call_p calls.
17228 * tree-flow.h (nonfreeing_call_p): New prototype.
17229 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
17231 2013-02-13 David S. Miller <davem@davemloft.net>
17233 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
17235 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
17238 * ira.c (max_regno_before_ira): Move from ...
17240 (fix_reg_equiv_init): Use max_regno_before_ira instead of
17243 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17245 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
17247 2013-02-13 Richard Biener <rguenther@suse.de>
17250 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
17251 globals in MEM_REFs.
17253 2013-02-13 Richard Biener <rguenther@suse.de>
17255 * loop-init.c (loop_optimizer_init): Clear loop state when
17256 re-initializing preserved loops.
17257 * loop-unswitch.c (unswitch_single_loop): Return whether
17258 we unswitched the loop. Do not verify loop state here.
17259 (unswitch_loops): When we unswitched a loop discover new loops.
17261 2013-02-13 Kostya Serebryany <kcc@google.com>
17263 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
17265 * sanitizer.def: Rename __asan_init to __asan_init_v1.
17267 2013-02-12 Dodji Seketeli <dodji@redhat.com>
17269 Avoid instrumenting duplicated memory access in the same basic block
17270 * Makefile.in (asan.o): Add new dependency on hash-table.h
17271 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
17272 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
17273 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
17274 (free_mem_ref_resources, has_mem_ref_been_instrumented)
17275 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
17276 (get_mem_ref_of_assignment): New functions.
17277 (get_mem_refs_of_builtin_call): Extract from
17278 instrument_builtin_call and tweak a little bit to make it fit with
17280 (instrument_builtin_call): Use the new
17281 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
17282 of is_gimple_builtin_call.
17283 (instrument_derefs, instrument_mem_region_access): Insert the
17284 instrumented memory reference into the hash table.
17285 (maybe_instrument_assignment): Renamed instrument_assignment into
17286 this, and change it to advance the iterator when instrumentation
17287 actually happened and return true in that case. This makes it
17288 homogeneous with maybe_instrument_assignment, and thus give a
17289 chance to callers to be more 'regular'.
17290 (transform_statements): Clear the memory reference hash table
17291 whenever we enter a new BB, when we cross a function call, or when
17292 we are done transforming statements. Use
17293 maybe_instrument_assignment instead of instrumentation. No more
17294 need to special case maybe_instrument_assignment and advance the
17295 iterator after calling it; it's now handled just like
17296 maybe_instrument_call. Update comment.
17298 2013-02-13 Richard Biener <rguenther@suse.de>
17300 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
17301 Fix loop discovery code.
17303 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
17305 PR inline-asm/56148
17306 * lra-constraints.c (process_alt_operands): Match early clobber
17307 operand with itself. Check conflicts with earlyclobber only if
17308 the operand is not reloaded. Prefer to reload conflicting operand
17309 if earlyclobber and matching operands are the same.
17311 2013-02-12 Richard Biener <rguenther@suse.de>
17314 * lto-streamer-out.c (write_symbol): Do not output symbols
17315 for hard register variables.
17317 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
17320 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
17321 (umulsidi3_insn, mulsidi3_insn): New insns.
17323 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
17325 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
17326 (struct tune_params): Add vec_costs field.
17327 * config/arm/arm.c (arm_builtin_vectorization_cost)
17328 (arm_add_stmt_cost): New functions.
17329 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
17330 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
17331 (arm_default_vec_cost): New struct of type cpu_vec_costs.
17332 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
17333 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
17334 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
17335 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
17337 2013-02-12 Richard Biener <rguenther@suse.de>
17340 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
17341 decls again if possible.
17343 2013-02-12 Richard Biener <rguenther@suse.de>
17345 PR middle-end/56288
17346 * tree-ssa.c (verify_ssa_name): Fix check, move
17347 SSA_NAME_IN_FREE_LIST check up.
17349 2013-02-12 Jakub Jelinek <jakub@redhat.com>
17350 Steven Bosscher <steven@gcc.gnu.org>
17352 PR rtl-optimization/56151
17353 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
17354 equal to op0 or op1, and last_insn pattern is CODE operation
17355 with MEM dest and one of the operands matches that MEM.
17357 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
17359 * doc/extend.texi: Document Function Multiversioning and "default"
17360 parameter string to target attribute.
17361 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
17362 target attribute parameter is "default".
17363 (ix86_compare_version_priority): Remove checks for target attribute.
17364 (ix86_mangle_function_version_assembler_name): Change error to sorry.
17365 Remove check for target attribute equal to NULL. Add assert.
17366 (ix86_generate_version_dispatcher_body): Change error to sorry.
17368 2013-02-11 Iain Sandoe <iain@codesourcery.com>
17369 Jack Howarth <howarth@bromo.med.uc.edu>
17370 Patrick Marlier <patrick.marlier@gmail.com>
17373 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
17374 define ENDFILE_SPEC as TM_DESTRUCTOR.
17375 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
17377 2013-02-11 Alexander Potapenko <glider@google.com>
17378 Jack Howarth <howarth@bromo.med.uc.edu>
17379 Jakub Jelinek <jakub@redhat.com>
17382 * config/darwin.c (cdtor_record): Rename ctor_record.
17383 (sort_cdtor_records): Rename sort_ctor_records.
17384 (finalize_dtors): New routine to sort destructors by
17385 priority before use in assemble_integer.
17386 (machopic_asm_out_destructor): Use finalize_dtors if needed.
17388 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
17390 PR rtl-optimization/56275
17391 * simplify-rtx.c (avoid_constant_pool_reference): Check that
17392 offset is non-negative and less than cmode size before
17393 calling simplify_subreg.
17395 2013-02-11 Richard Biener <rguenther@suse.de>
17397 PR tree-optimization/56264
17398 * cfgloop.h (fix_loop_structure): Adjust prototype.
17399 * loop-init.c (fix_loop_structure): Return the number of
17400 newly discovered loops.
17401 * tree-cfgcleanup.c (repair_loop_structures): When new loops
17402 are discovered, do a full loop-closed SSA rewrite.
17404 2013-02-11 Richard Biener <rguenther@suse.de>
17406 PR tree-optimization/56273
17407 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
17409 (check_array_ref): Fix missing newline in dumps.
17410 (search_for_addr_array): Likewise.
17412 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
17414 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
17416 2013-02-09 Jakub Jelinek <jakub@redhat.com>
17419 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
17421 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
17423 PR rtl-optimization/56246
17424 * lra-constraints.c (simplify_operand_subreg): Try to reuse
17426 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
17427 constraints are satisfied.
17429 2013-02-08 Jeff Law <law@redhat.com>
17432 * emit-rtl.c (reg_is_parm_p): New function.
17433 * regs.h (reg_is_parm_p): New prototype.
17434 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
17435 callee-clobbered registers.
17437 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
17440 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
17441 If there is no implicit builtin declaration, just return NULL.
17443 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17445 * config/i386/sse.md (FMAMODEM): New mode iterator.
17446 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
17447 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
17449 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17451 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
17452 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
17453 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
17455 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17457 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
17458 (microblaze*-*-elf): Likewise.
17459 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
17461 * config/microblaze/microblaze-c.c: Add builtin defines for
17462 _LITTLE_ENDIAN and _BIG_ENDIAN.
17463 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
17464 add to TARGET_DEFAULT flags.
17465 Expand ASM_SPEC and LINK_SPEC.
17466 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
17467 * config/microblaze/microblaze.md: Update extendsidi2 and
17468 movdi_internal instructions to use low-order / high-order reg
17470 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
17471 options and inversemask / mask of LITTLE_ENDIAN.
17472 * config/microblaze/t-microblaze: Expand multilib options to
17473 include mlittle-endian (le) and update exceptions patterns.
17475 2013-02-08 Jakub Jelinek <jakub@redhat.com>
17477 PR rtl-optimization/56195
17478 * lra-constraints.c (get_reload_reg): Don't reuse regs
17479 if they have smaller mode than requested, if they have
17480 wider mode than requested, try to return a SUBREG.
17482 PR tree-optimization/56250
17483 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
17484 if type is unsigned and code isn't MULT_EXPR.
17486 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
17488 PR tree-optimization/56064
17489 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
17490 bits according to mode.
17491 * fixed-value.h (fixed_from_double_int)
17492 (const_fixed_from_double_int): Adjust comments.
17494 2013-02-08 Richard Biener <rguenther@suse.de>
17497 * lto-streamer.h (struct data_in): Remove current_file, current_line
17498 and current_col members.
17499 * lto-streamer-out.c (lto_output_location): Stream changed bits
17500 en-block for efficiency.
17501 * lto-streamer-in.c (clear_line_info): Remove.
17502 (lto_input_location): Cache current file, line and column
17503 globally via local statics. Read changed bits en-block.
17504 (input_function): Do not call clear_line_info.
17505 (lto_read_body): Likewise.
17506 (lto_input_toplevel_asms): Likewise.
17508 2013-02-08 Michael Matz <matz@suse.de>
17510 PR tree-optimization/52448
17511 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
17512 (nt_call_phase): New static.
17513 (add_or_mark_expr): Only mark accesses with newer phase than any
17515 (nonfreeing_call_p): New.
17516 (nt_init_block): Update nt_call_phase, mark blocks as visited.
17517 (nt_fini_block): Keep blocks marked as visited.
17518 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
17520 2013-02-08 Richard Biener <rguenther@suse.de>
17522 * ira.c (ira): Free broken dominator information.
17524 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17526 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
17528 2013-02-08 Marek Polacek <polacek@redhat.com>
17530 * cfgloop.c (verify_loop_structure): Add more checking of headers.
17532 2013-02-08 Richard Biener <rguenther@suse.de>
17534 PR middle-end/56181
17535 * cfgloop.h (flow_loops_find): Adjust.
17536 (bb_loop_header_p): Declare.
17537 * cfgloop.c (bb_loop_header_p): New function split out from ...
17538 (flow_loops_find): ... here. Adjust function signature,
17539 support incremental loop structure update.
17540 (verify_loop_structure): Cleanup. Verify a loop is a loop.
17541 * cfgloopmanip.c (fix_loop_structure): Move ...
17542 * loop-init.c (fix_loop_structure): ... here.
17543 (apply_loop_flags): Split out from ...
17544 (loop_optimizer_init): ... here.
17545 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
17546 in incremental mode, only remove dead loops here.
17548 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
17551 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
17552 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
17553 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
17554 (*round<mode>3.libgcc): New insns for fixed-modes.
17555 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
17556 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
17557 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
17558 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
17559 implementations. Define to __builtin_avr_absFX,
17560 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
17561 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
17562 __builtin_avr_countlsFX, respectively.
17563 * config/avr/avr-c.c (target.h): Include it.
17564 (enum avr_builtin_id): New enum.
17565 (avr_resolve_overloaded_builtin): New static function.
17566 (avr_register_target_pragmas): Use it to set
17567 targetm.resolve_overloaded_builtin.
17568 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
17569 tree nodes used by DEF_BUILTIN.
17570 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
17571 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
17572 <AVR_BUILTIN_xxBITS>: Same.
17574 2013-02-08 Richard Biener <rguenther@suse.de>
17576 * cfgloop.c (verify_loop_structure): Properly handle
17577 a loop exiting to another loop header.
17578 * ira-int.h (ira_loops): Remove.
17579 * ira.c (ira_loops): Remove.
17580 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
17581 (do_reload): Use loop_optimizer_finalize.
17582 * ira-build.c (create_loop_tree_nodes): Use get_loops and
17583 number_of_loops to access the loop tree.
17584 (more_one_region_p): Likewise.
17585 (finish_loop_tree_nodes): Likewise.
17586 (rebuild_regno_allocno_maps): Likewise.
17587 (mark_loops_for_removal): Likewise.
17588 (mark_all_loops_for_removal): Likewise.
17589 (remove_unnecessary_regions): Likewise.
17590 (ira_build): Likewise.
17591 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
17593 2013-02-08 Richard Biener <rguenther@suse.de>
17595 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
17596 * ipa-pure-const.c (analyze_function): Avoid calling
17597 mark_irreducible_loops twice.
17598 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
17600 2013-02-07 David S. Miller <davem@davemloft.net>
17602 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
17604 * var-tracking.c (vt_add_function_parameter): Test the presence of
17605 HAVE_window_save properly and do not remap argument registers when
17606 we have a leaf function.
17608 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
17611 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
17613 * config/i386/i386.c (ix86_print_operand): Ditto.
17615 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
17617 * lra-constraints.c (process_alt_operands): Fix recently added comment.
17619 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
17621 PR rtl-optimization/56225
17622 * lra-constraints.c (process_alt_operands): Check that reload hard
17623 reg can hold value for strict_low_part.
17625 2013-02-07 Jakub Jelinek <jakub@redhat.com>
17628 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
17629 dwarf2out_end_function.
17630 (in_first_function_p, maybe_at_text_label_p,
17631 first_loclabel_num_not_at_text_label): New variables.
17632 (dwarf2out_var_location): In the first function find out
17633 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
17634 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
17637 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
17639 PR rtl-optimization/56178
17640 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
17641 SUBREG of a register. Tidy up related block of code.
17642 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
17643 note if the source is a register or a SUBREG of a register.
17645 2013-02-07 Jakub Jelinek <jakub@redhat.com>
17648 * config/rs6000/rs6000.md (ptrm): New mode attr.
17649 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
17650 call_value_indirect_aix<pttrsize>,
17651 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
17654 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17656 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
17657 if -bnortl. Convert to strcmp and strncmp.
17659 2013-02-07 Alan Modra <amodra@gmail.com>
17662 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
17663 addresses won't wrap when offsetting.
17664 (rs6000_secondary_reload): Provide secondary reloads needed for
17665 wrapping LO_SUM addresses.
17667 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
17669 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
17670 MACH, just __MACH__.
17672 2013-02-06 Richard Biener <rguenther@suse.de>
17674 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
17675 instead of calling fix_loop_structure.
17677 2013-02-06 Jakub Jelinek <jakub@redhat.com>
17679 PR middle-end/56217
17680 * omp-low.c (use_pointer_for_field): Return false if
17681 lower_send_shared_vars doesn't generate any copy-out code.
17683 2013-02-06 Tom de Vries <tom@codesourcery.com>
17685 PR rtl-optimization/56131
17686 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
17687 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
17688 of the label is NULL. Add comment.
17690 2013-02-05 Jakub Jelinek <jakub@redhat.com>
17692 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
17695 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
17696 (STATIC_LIBTSAN_LIBS): Likewise.
17697 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
17698 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
17699 is defined, don't add anything else beyond that.
17700 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
17701 (LINK_COMMAND_SPEC): Use them.
17703 PR tree-optimization/56205
17704 * tree-stdarg.c (check_all_va_list_escapes): Return true if
17705 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
17706 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
17708 2013-02-05 Richard Biener <rguenther@suse.de>
17710 PR tree-optimization/53342
17711 PR tree-optimization/53185
17712 * tree-vectorizer.h (vect_check_strided_load): Remove.
17713 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
17714 not disallow peeling for vectorized strided loads.
17715 (vect_check_strided_load): Make static and simplify.
17716 (vect_analyze_data_refs): Adjust.
17717 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
17718 correctly when vectorizing strided loads.
17720 2013-02-05 Richard Biener <rguenther@suse.de>
17722 * doc/install.texi: Refer to ISL, not PPL.
17724 2013-02-05 Jan Hubicka <jh@suse.cz>
17726 PR tree-optimization/55789
17727 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
17729 2013-02-05 Jan Hubicka <jh@suse.cz>
17731 PR tree-optimization/55789
17732 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
17733 the dead call anyway.
17735 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
17738 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
17740 2013-02-04 Alexander Potapenko <glider@google.com>
17741 Jack Howarth <howarth@bromo.med.uc.edu>
17742 Jakub Jelinek <jakub@redhat.com>
17745 * config/darwin.c (sort_ctor_records): Stabilized qsort
17746 on constructor priority by using original position.
17747 (finalize_ctors): New routine to sort constructors by
17748 priority before use in assemble_integer.
17749 (machopic_asm_out_constructor): Use finalize_ctors if needed.
17751 2013-02-04 Jakub Jelinek <jakub@redhat.com>
17754 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
17755 about visibility on artificial decls.
17756 * config/sol2.c (solaris_assemble_visibility): Likewise.
17758 2013-02-04 Kai Tietz <ktietz@redhat.com>
17761 * config/i386/i386.c (function_value_ms_64): Add additional valtype
17762 argument and improve checking of return-argument types for 16-byte
17764 (ix86_function_value_1): Add additional valtype argument on call
17765 of function_value_64.
17766 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
17767 handling infunction_value_64 function.
17769 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
17771 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
17773 2013-02-04 Richard Biener <rguenther@suse.de>
17775 PR tree-optimization/56188
17776 * tree-ssa-structalias.c (label_visit): Consider case with
17777 initially non-empty points-to set.
17778 (perform_var_substitution): Dump node mapping and clean up.
17780 2013-02-04 Richard Guenther <rguenther@suse.de>
17783 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
17784 node prevail as last resort.
17785 (lto_symtab_merge_decls): Remove guard on LTRANS here.
17786 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
17788 2013-02-04 Richard Biener <rguenther@suse.de>
17790 PR tree-optimization/56113
17791 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
17793 (equiv_class_lookup_or_add): ... this.
17794 (label_visit): Adjust and fix error in previous patch.
17795 (perform_var_substitution): Adjust.
17797 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
17799 * config/sh/divtab.c: Fix formatting and comments throughout the file.
17800 * config/sh/sh4-300.md: Likewise.
17801 * config/sh/sh4a.md: Likewise.
17802 * config/sh/constraints.md: Likewise.
17803 * config/sh/sh.md: Likewise.
17804 * config/sh/netbsd-elf.h: Likewise.
17805 * config/sh/predicates.md: Likewise.
17806 * config/sh/sh-protos.h: Likewise.
17807 * config/sh/ushmedia.h: Likewise.
17808 * config/sh/linux.h: Likewise.
17809 * config/sh/sh.c: Likewise.
17810 * config/sh/superh.h: Likewise.
17811 * config/sh/elf.h: Likewise.
17812 * config/sh/sh4.md: Likewise.
17813 * config/sh/sh.h: Likewise.
17815 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17817 * config/pa/constraints.md: Adjust unused letters. Change "T"
17818 constraint to match_test floating_point_store_memory_operand().
17819 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
17820 (base14_operand): New.
17821 (floating_point_store_memory_operand): New.
17822 (integer_store_memory_operand): Revise to use base14_operand and
17823 reg_plus_base_memory_operand.
17824 (move_dest_operand): Allow symbolic_memory_operands.
17825 (symbolic_memory_operand): Check for LO_SOM.
17826 (symbolic_operand): Change default case to break.
17827 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
17828 CONST_DOUBLE values to be reloaded by putting them into memory when
17829 the destination is a floating point register.
17830 (movdf): Remove code to handle CONST_DOUBLE.
17832 (reload_indf_r1): New.
17833 (reload_insf_r1): New.
17834 Consistently use "Q" and "T" constraints with integer and floating
17835 point move instructions, respectively.
17836 (movdi): Remove FAIL.
17837 Change predicate for source operand unamed DImode move from
17838 general_operand to move_src_operand.
17839 (umulsidi3): Change predicate for destination operand to
17841 Likewise for similar unamed patterns.
17842 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
17843 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
17844 (hppa_legitimize_address): Simplify mask calculation.
17845 (pa_emit_move_sequence): Revised handling of secondary reloads from
17846 REG+D addresses for floating point loads and stores. Directly handle
17847 loading CONST0_RTX (mode) to a floating point register.
17848 (pa_secondary_reload): Handle reloading DF and SFmode constant values
17849 to floating point registers. Don't restrict secondary reloads to
17850 floating point registers to integer modes. Revise some comments and
17852 (TARGET_LEGITIMATE_ADDRESS_P): Define.
17853 (pa_legitimate_address_p): New.
17854 (pa_legitimize_reload_address): New.
17855 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
17856 (STRICT_REG_OK_FOR_BASE_P): New.
17857 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
17858 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
17860 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
17861 Andrew Dixie <andrewd@gentrack.com>
17863 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
17866 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
17868 * expmed.c (extract_bit_field_1): Pass the full width of the
17869 structure to get_best_reg_extraction_insn.
17871 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17874 * configure.ac (use_cxa_atexit): Add AIX.
17875 * configure: Regenerate.
17877 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
17879 2013-02-01 Jakub Jelinek <jakub@redhat.com>
17882 * final.c (need_profile_function): New variable.
17883 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
17884 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
17885 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
17886 notes, targetm.asm_out.function_prologue doesn't emit anything,
17887 HAVE_prologue and profiler should be emitted before prologue,
17888 set need_profile_function instead of emitting it.
17889 (final_scan_insn): If need_profile_function, emit
17890 profile_function on the first NOTE_INSN_BASIC_BLOCK or
17891 NOTE_INSN_FUNCTION_BEG note.
17893 2013-02-01 Richard Henderson <rth@redhat.com>
17895 * config/rs6000/rs6000.md (smulditi3): New.
17898 * config/alpha/alpha.md (umulditi3): New.
17900 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17902 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
17903 (ASM_OUTPUT_ALIGNED_LOCAL): New.
17905 2013-02-01 Richard Biener <rguenther@suse.de>
17907 PR tree-optimization/56113
17908 * tree-ssa-structalias.c (label_visit): Reduce work for
17909 single-predecessor nodes.
17911 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
17913 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
17914 range isn't testing for zero.
17916 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
17918 PR middle-end/56113
17919 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
17921 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
17922 Nick Clifton <nickc@redhat.com>
17924 * config/v850/constraints.md (Q): Define as a memory constraint.
17925 * config/v850/predicates.md (label_ref_operand): New predicate.
17926 (e3v5_shift_operand): New predicate.
17927 (ior_operator): New predicate.
17928 * config/v850/t-v850: Add e3v5 multilib.
17929 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17930 (v850_gen_movdi): Prototype.
17931 * config/v850/v850.c: Add support for e3v5 architecture.
17932 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17934 (construct_save_jarl): Add e3v5 long JARL support.
17935 (v850_adjust_insn_length): New function. Adjust length of call
17936 insns when using e3v5 instructions.
17937 (v850_gen_movdi): New function: Generate instructions to move a
17939 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17940 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17941 (TARGET_USE_FPU): Enable for e3v5.
17942 (CONST_OK_FOR_W): New macro.
17943 (ADJUST_INSN_LENGTH): Define.
17944 * config/v850/v850.md (UNSPEC_LOOP): Define.
17945 (attr cpu): Add v850e3v5.
17946 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17947 (movdi): New pattern.
17948 (movdi_internal): New pattern.
17949 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17950 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17951 (cstoresf4): Likewise.
17952 (cstoredf4): Likewise.
17953 (insv): New pattern.
17954 (rotlso3_a): New pattern.
17955 (rotlsi3_b): New pattern
17956 (rotlsi3_v850e3v5): New pattern.
17957 (doloop_begin): New pattern.
17958 (fix_loop_counter): New pattern.
17959 (doloop_end): New pattern.
17960 (branch_normal): Add e3v5 long branch support.
17961 (branch_invert): Likewise.
17962 (branch_z_normal): Likewise.
17963 (branch_z_invert): Likewise.
17964 (branch_nz_normal): Likewise.
17965 (branch_nz_invert): Likewise.
17966 (call_internal_short): Add e3v5 register-indirect JARL support.
17967 (call_internal_long): Likewise.
17968 (call_value_internal_short): Likewise.
17969 (call_value_internal_long): Likewise.
17970 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17971 (mloop): New option.
17972 * config.gcc: Add support for configuring v840e3v5 target.
17973 * doc/invoke.texi: Document new v850 specific command line options.
17975 2013-01-31 Paul Koning <ni1d@arrl.net>
17979 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17980 children if parent is a class.
17981 (prune_unused_types_prune): Don't add DW_AT_declaration.
17983 2013-01-31 Richard Biener <rguenther@suse.de>
17985 PR tree-optimization/56157
17986 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17987 match up operand with SLP child.
17989 2013-01-31 Jason Merrill <jason@redhat.com>
17992 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17993 parameters the first time.
17994 (gen_scheduled_generic_parms_dies): Check completeness here.
17996 2013-01-31 Richard Biener <rguenther@suse.de>
17998 PR middle-end/53073
17999 * common.opt (faggressive-loop-optimizations): New flag,
18000 enabled by default.
18001 * doc/invoke.texi (faggressive-loop-optimizations): Document.
18002 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
18003 infer_loop_bounds_from_undefined by it.
18005 2013-01-31 Richard Biener <rguenther@suse.de>
18007 PR tree-optimization/56150
18008 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
18009 visit virtual operands.
18010 (find_uses_to_rename_bb): Likewise.
18012 2013-01-31 Richard Biener <rguenther@suse.de>
18014 PR tree-optimization/56150
18015 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
18016 mixed store non-store stmts.
18018 2013-01-30 Jakub Jelinek <jakub@redhat.com>
18021 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
18022 LIBASAN_EARLY_SPEC is defined.
18023 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
18024 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
18026 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
18029 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
18030 invalid args instead of ICEing on it.
18031 (ix86_valid_target_attribute_tree): Return error_mark_node if
18032 ix86_valid_target_attribute_inner_p failed.
18033 (ix86_valid_target_attribute_p): Return false only if
18034 ix86_valid_target_attribute_tree returned error_mark_node. Allow
18035 target("default") attribute.
18036 (sorted_attr_string): Change argument from const char * to tree,
18037 merge in all target attribute arguments rather than just one.
18038 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
18039 instead of free. Avoid using strcat.
18040 (ix86_mangle_function_version_assembler_name): Mangle
18041 target("default") as if no target attribute is present. Adjust
18042 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
18043 instead of xmalloc and XDELETEVEC instead of free.
18044 (ix86_function_versions): Don't return true if one of the decls
18045 doesn't have target attribute. If they don't and one of the decls
18046 is DECL_FUNCTION_VERSIONED, report an error. Adjust
18047 sorted_attr_string caller. Use XDELETEVEC instead of free.
18048 (ix86_supports_function_versions): Remove.
18049 (make_name): Fix up formatting.
18050 (make_dispatcher_decl): Remove resolver_name and its initialization.
18051 Avoid leaking memory.
18052 (is_function_default_version): Return true if there is
18053 target("default") attribute rather than no target attribute at all.
18054 (make_resolver_func): Avoid leaking memory.
18055 (ix86_generate_version_dispatcher_body): Likewise.
18056 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
18057 * target.def (supports_function_versions): Remove.
18058 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
18059 * doc/tm.texi: Regenerated.
18061 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
18063 PR rtl-optimization/56144
18064 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
18065 for values with side effects.
18067 2013-01-30 Richard Biener <rguenther@suse.de>
18069 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
18070 (sparseset_pop): Likewise.
18071 * cfganal.c (compute_idf): Likewise. Increase work-stack size
18072 to be able to use quick_push in the worker loop.
18074 2013-01-30 Marek Polacek <polacek@redhat.com>
18076 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
18078 2013-01-30 Richard Biener <rguenther@suse.de>
18081 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
18083 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
18085 PR tree-optimization/56064
18086 * fixed-value.c (fixed_from_double_int): New function.
18087 * fixed-value.h (fixed_from_double_int): New prototype.
18088 (const_fixed_from_double_int): New static inline function.
18089 * fold-const.c (native_interpret_fixed): New static function.
18090 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
18091 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
18092 (native_encode_fixed): New static function.
18093 (native_encode_expr) <FIXED_CST>: Use it.
18094 (native_interpret_int): Move double_int worker code to...
18095 * double-int.c (double_int::from_buffer): ...this new static method.
18096 * double-int.h (double_int::from_buffer): Prototype it.
18098 2013-01-30 Richard Biener <rguenther@suse.de>
18100 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
18101 New pointer-map and obstack.
18102 (init_alias_vars): Allocate pointer-map and obstack.
18103 (delete_points_to_sets): Free them.
18104 (find_what_var_points_to): Cache result.
18105 (find_what_p_points_to): Adjust for changed interface of
18106 find_what_var_points_to.
18107 (compute_points_to_sets): Likewise.
18108 (ipa_pta_execute): Likewise.
18110 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18112 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
18113 * configure: Regenerate.
18114 * config.in: Regenerate.
18115 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
18116 #nobits/#progbits if supported.
18118 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
18121 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
18122 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
18123 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
18125 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18127 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
18128 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
18130 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18132 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
18133 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
18135 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18137 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
18139 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
18140 * config/arm/cortex-a7.md: New bypasses using
18141 arm_mac_accumulator_is_result.
18143 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18145 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
18146 (cortex_a7_neon_mla): Likewise.
18147 (cortex_a7_fpfmad): New reservation.
18148 (cortex_a7_fpmacs): Use ffmas and update required units.
18149 (cortex_a7_fpmuld): Update required units and latency.
18150 (cortex_a7_fpmacd): Likewise.
18151 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
18152 (cortex_a7_neon). Likewise.
18153 (bypass) Update participating units.
18155 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18157 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
18158 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
18160 * config/arm/vfp11.md (vfp_farith): Use ffmas.
18161 (vfp_fmul): Use ffmad.
18162 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
18163 (cortex_r4_fmacd): Use ffmad.
18164 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
18165 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
18166 (cortex_a9_fmacd): Use ffmad.
18167 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
18168 (cortex_a8_vfp_macd): Use ffmad.
18169 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
18170 (cortex_a5_fpmacd): Use ffmad.
18171 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
18172 (cortex_a15_vfp_macd): Use ffmad.
18173 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
18175 2013-01-29 Jason Merrill <jason@redhat.com>
18178 * varasm.c (default_assemble_visibility): Don't warn about
18179 visibility on artificial decls.
18181 2013-01-29 Richard Biener <rguenther@suse.de>
18183 PR tree-optimization/56113
18184 * tree-ssa-structalias.c (equiv_class_lookup): Also return
18186 (label_visit): Free duplicate bitmaps and record the leader instead.
18187 (perform_var_substitution): Adjust.
18189 2013-01-29 Richard Biener <rguenther@suse.de>
18191 PR tree-optimization/55270
18192 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
18193 the CFG, schedule loops for fixup.
18195 2013-01-29 Nick Clifton <nickc@redhat.com>
18197 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
18200 2013-01-28 Leif Ekblad <leif@rdos.net>
18202 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
18203 * config/i386/i386.h (TARGET_RDOS): New macro.
18204 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
18205 * config/i386/i386.c (ix86_option_override_internal): For 64bit
18206 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
18207 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
18208 DEFAULT_LARGE_SECTION_THRESHOLD.
18209 * config/i386/i386.md (R14_REG, R15_REG): New constants.
18210 * config/i386/rdos.h: New file.
18211 * config/i386/rdos64.h: New file.
18213 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
18216 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
18219 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18221 PR rtl-optimization/56117
18222 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
18223 call cselib_lookup_from_insn on the MEM before calling
18224 add_insn_mem_dependence.
18226 2013-01-28 Richard Biener <rguenther@suse.de>
18228 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
18229 to a stmt that didn't have one.
18230 (copy_phis_for_bb): Likewise for PHI arguments.
18231 (copy_debug_stmt): Likewise for debug stmts.
18233 2013-01-28 Richard Biener <rguenther@suse.de>
18235 PR tree-optimization/56034
18236 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
18237 (partition_builtin_p): Adjust.
18238 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
18239 it is the last partition.
18240 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
18241 up the vertex for the definition.
18242 (classify_partition): Classify whether a partition is a
18243 PKIND_REDUCTION, thus has uses outside of the loop.
18244 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
18245 Merge all PKIND_REDUCTION partitions into the last partition.
18246 (tree_loop_distribution): Seed partitions from reductions as well.
18248 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18250 PR tree-optimization/56125
18251 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
18252 pow(x,c) into sqrt(x) * powi(x, n/2) or
18253 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
18254 optimizing for size.
18255 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
18256 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
18259 PR tree-optimization/56094
18260 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
18261 to UNKNOWN_LOCATION while gimplifying expr.
18263 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
18266 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
18267 operand 0 in movabs insn template for -masm=intel asm alternative.
18268 (*movabs<mode>_2): Ditto for operand 1.
18270 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
18273 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
18276 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18278 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
18281 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18283 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
18284 Undef to avoid warning.
18286 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
18288 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
18289 * configure: Regenerate.
18291 2013-01-25 Jakub Jelinek <jakub@redhat.com>
18293 PR tree-optimization/56098
18294 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
18295 for stmts with volatile ops.
18296 (cond_store_replacement): Don't optimize if assign has volatile ops.
18297 (cond_if_else_store_replacement_1): Don't optimize if either
18298 then_assign or else_assign have volatile ops.
18299 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
18302 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18304 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
18306 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18308 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
18309 missing ':' in asm example.
18311 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
18313 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
18314 entries into lane and laneq entries.
18315 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
18316 Remove AdvSIMD scalar modes.
18317 (aarch64_sq<r>dmulh_laneq<mode>): New.
18318 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
18320 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
18321 builtin implementations to relfect changes in RTL in aarch64-simd.md.
18322 * config/aarch64/iterators.md (VCOND): New.
18325 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18328 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
18329 Add NULL LIBNAME argument to existing definitions.
18330 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
18331 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
18332 * config/avr/avr.c (DEF_BUILTIN): Same.
18333 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
18334 (avr_expand_builtin): Expand to a vanilla call if a libgcc
18335 implementation is available (DECL_ASSEMBLER_NAME is set).
18336 (avr_fold_absfx): New static function.
18337 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
18338 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
18339 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
18340 AVR_BUILTIN_ABSLLK.
18341 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
18342 (abshk, absk, abslk, absllk): Provide as static inline functions.
18344 2013-01-25 Marek Polacek <polacek@redhat.com>
18346 PR tree-optimization/56035
18347 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
18349 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
18351 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
18352 (*movtf_internal_rex64): Add (!o,C) alternative
18353 (*movxf_internal_rex64): Ditto.
18354 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
18356 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
18358 * doc/invoke.texi: fix typo.
18359 * doc/objc.texi: fix typo.
18361 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
18363 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
18364 for the first two alternatives.
18366 2013-01-24 Diego Novillo <dnovillo@google.com>
18368 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
18369 (ggc-zone.o): Remove.
18370 * configure.ac: Remove option --with-gc.
18371 * configure: Re-generate.
18372 * doc/install.texi: Remove documentation for --with-gc.
18373 * gengtype.c (write_enum_defn): Remove. Update all users.
18374 (write_Types_process_field): Remove generation of gt_e_* argument.
18375 (output_type_enum): Remove. Update all users.
18376 (write_enum_defn): Remove. Update all users.
18377 (enum alloc_zone): Remove. Update all users.
18378 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
18379 * ggc-common.c (ggc_splay_alloc): Remove first argument.
18380 Update all callers.
18381 (struct ptr_data): Remove field TYPE. Update all users.
18382 (gt_pch_note_object): Remove argument TYPE. Update all users.
18383 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
18385 * ggc-none.c (ggc_alloc_typed_stat): Remove.
18386 (struct alloc_zone): Remove.
18387 (ggc_internal_alloc_zone_stat): Remove.
18388 (ggc_internal_cleared_alloc_zone_stat): Remove.
18389 * ggc-page.c (ggc_alloc_typed_stat): Remove.
18390 (ggc_pch_count_object): Remove last argument. Update all users.
18391 (ggc_pch_alloc_object): Remove last argument. Update all users.
18392 (struct alloc_zone): Remove.
18393 * ggc-zone.c: Remove.
18394 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
18395 (struct alloc_zone): Remove.
18396 (ggc_alloc_typed_stat): Remove.
18397 (ggc_alloc_typed): Remove.
18398 (ggc_splay_alloc): Remove first argument.
18399 (rtl_zone): Remove. Update all users.
18400 (tree_zone): Remove. Update all users.
18401 (tree_id_zone): Remove. Update all users.
18402 (ggc_internal_zone_alloc_stat): Remove. Update all users.
18403 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
18404 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
18405 * tree-ssanames.c: Remove references to zone allocator in comments.
18407 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
18409 * config/avr/avr.c (avr_out_fract): Make register numbers that
18410 might be outside of source operand signed.
18412 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
18414 * config/i386/constraints.md (Yf): New constraint.
18415 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
18416 of f constraint to conditionaly disable x87 register preferences.
18417 (*movdf_internal): Ditto.
18418 (*movsf_internal): Ditto.
18420 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
18422 PR inline-asm/55934
18423 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
18424 that have operands with impossible constraints.
18425 Add a FIXME for a speed-up opportunity.
18426 * lra-constraints.c (process_alt_operands): Verify that a class
18427 selected from constraints on asms is valid for the operand mode.
18428 (curr_insn_transform): Remove incorrect comment.
18430 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
18432 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
18433 TOC operand is a valid symbol ref in the constant pool.
18435 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18437 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
18439 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
18442 * config/avr/stdfix.h: New file.
18443 * t-avr (stdfix-gcc.h): New rule to build it.
18444 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
18446 2013-01-23 Kostya Serebryany <kcc@google.com>
18448 * config/darwin.h: remove dependency on
18449 CoreFoundation (asan on Mac OS).
18451 2013-01-23 Jakub Jelinek <jakub@redhat.com>
18454 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
18455 instead of cmpdi_operand for first comparison operand.
18456 Don't assert that comparison operands aren't both constants.
18458 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
18460 * doc/install.texi (Downloading the Source): Update references to
18461 downloading separate components.
18463 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
18465 * doc/extend.texi (__int128): Improve grammar.
18467 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
18470 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
18471 alternative to (o,r).
18472 (*movdi_internal_rex64): Remove (!o,n) alternative.
18473 (DImode immediate->memory splitter): Remove.
18474 (DImode immediate->memory peephole2): Remove.
18475 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
18476 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
18477 alternative to (!o,*r).
18478 (*movtf_internal_sse): New pattern.
18479 (*movxf_internal_rex64): New pattern.
18480 (*movxf_internal): Disable for TARGET_64BIT.
18481 (*movdf_internal_rex64): Remove (!o,F) alternative.
18483 2013-01-22 Jakub Jelinek <jakub@redhat.com>
18485 PR middle-end/56074
18486 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
18487 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
18488 * tree-vect-loop-manip.c (find_loop_location): Also ignore
18489 stmt locations where LOCATION_LOCUS of the stmt location is
18490 UNKNOWN_LOCATION or BUILTINS_LOCATION.
18493 * config/i386/i386.md (UNSPEC_STOS): New.
18494 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
18495 *strsetqi_1): Add UNSPEC_STOS.
18497 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
18500 * doc/invoke.texi: Remove left over -Wsynth example.
18502 2013-01-21 Jakub Jelinek <jakub@redhat.com>
18504 PR tree-optimization/56051
18505 * fold-const.c (fold_binary_loc): Don't fold
18506 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
18507 a narrowing conversion, or widening conversion from signed
18510 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
18512 PR rtl-optimization/56023
18513 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
18514 dependent on debug instruction.
18516 2013-01-21 Martin Jambor <mjambor@suse.cz>
18518 PR middle-end/56022
18519 * function.c (allocate_struct_function): Call
18520 invoke_set_current_function_hook earlier.
18522 2013-01-21 Jakub Jelinek <jakub@redhat.com>
18524 * reload1.c (init_reload): Only initialize reload_obstack
18525 during the first call.
18527 2013-01-21 Marek Polacek <polacek@redhat.com>
18529 * cfgloop.c (verify_loop_structure): Fix up grammar.
18531 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
18533 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
18534 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
18536 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18539 * config/arm/marvell-pj4.md: Update copyright year.
18540 Fix up use of alu to alu_reg and simple_alu_imm.
18542 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
18544 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
18546 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
18549 * lra-constraints.c (curr_insn_transform): Don't reuse original
18550 insn for secondary memory move when memory mode should be different.
18552 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18554 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
18555 atomic_storedi_1): New patterns.
18557 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18559 btver2 pipeline descriptions.
18560 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
18562 * config/i386/i386.md (btver2_decode): New type attributes.
18563 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
18565 * config/i386/btver2.md: New file describing btver2 pipelines.
18567 2013-01-19 Andrew Pinski <apinski@cavium.com>
18569 PR tree-optimization/52631
18570 * tree-ssa-sccvn (visit_use): Before looking up the original
18571 statement, try looking up the simplified expression.
18573 2013-01-19 Anthony Green <green@moxielogic.com>
18575 * config/moxie/moxie.c (moxie_expand_prologue): Set
18576 current_function_static_stack_size.
18578 2013-01-18 Jakub Jelinek <jakub@redhat.com>
18580 PR tree-optimization/56029
18581 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
18582 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
18584 2013-01-18 Sharad Singhai <singhai@google.com>
18586 PR tree-optimization/55995
18587 * dumpfile.c (dump_loc): Print location only if available.
18588 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
18590 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
18593 * lra-constraints.c (curr_insn_transform): Reuse original insn for
18594 secondary memory move.
18595 (inherit_reload_reg): Use rclass instead of cl for
18596 check_secondary_memory_needed_p.
18598 2013-01-18 Jakub Jelinek <jakub@redhat.com>
18600 PR middle-end/56015
18601 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
18602 the case where writing real complex part of target modifies op1.
18604 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
18606 * config/aarch64/aarch64-simd.md
18607 (aarch64_vcond_internal<mode>): Handle unordered cases.
18608 * config/aarch64/iterators.md (v_cmp_result): New.
18610 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
18611 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18613 * config/arm/marvell-pj4.md: New file.
18614 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
18615 * config/arm/arm.md (generic_sched): Add marvell_pj4.
18616 (generic_vfp): Likewise.
18617 * config/arm/arm-cores.def: Add marvell-pj4.
18618 * config/arm/arm-tune.md: Regenerate.
18619 * config/arm/arm-tables.opt: Regenerate.
18620 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
18621 * doc/invoke.texi: Document marvell-pj4.
18623 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
18625 * config/aarch64/arm_neon.h: Map scalar types to standard types.
18627 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18632 * var-tracking.c (negative_power_of_two_p): New.
18633 (global_get_addr_cache, local_get_addr_cache): New.
18634 (get_addr_from_global_cache, get_addr_from_local_cache): New.
18635 (vt_canonicalize_addr): Rewrite using the above. Adjust the
18637 (vt_stack_offset_p): Remove.
18638 (vt_canon_true_dep): Always canonicalize loc's address.
18639 (clobber_overlapping_mems): Make sure we have a MEM.
18640 (local_get_addr_clear_given_value): New.
18641 (val_reset): Clear local cached entries.
18642 (compute_bb_dataflow): Create and release the local cache.
18643 Disable duplicate MEMs clobbering.
18644 (emit_notes_in_bb): Clobber MEMs likewise.
18645 (vt_emit_notes): Create and release the local cache.
18646 (vt_initialize, vt_finalize): Create and release the global
18647 cache, respectively.
18648 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
18650 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18652 PR libmudflap/53359
18653 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
18654 not found in the symtab.
18656 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18659 PR rtl-optimization/55547
18660 PR rtl-optimization/53827
18663 * alias.c (offset_overlap_p): New, factored out of...
18664 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
18665 the conservative special case for symbolic constants. Don't
18666 adjust zero sizes on alignment.
18668 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
18670 PR rtl-optimization/52573
18671 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
18672 REG_UNUSED for the same register.
18674 2013-01-17 Richard Biener <rguenther@suse.de>
18675 Marek Polacek <polacek@redhat.com>
18677 PR rtl-optimization/55833
18678 * loop-unswitch.c (unswitch_loops): Move loop verification...
18679 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
18680 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
18681 Set it to true when we're removing a loop from hierarchy tree in
18682 an irreducible region.
18683 (fix_bb_placements): Adjust caller.
18684 (fix_loop_placements): Likewise.
18686 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
18688 * config/avr/builtins.def (DEF_BUILTIN): Factor out
18689 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
18690 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
18691 Remove ID. Adjust comments.
18692 * config/avr/avr-c.c (avr_builtin_name): Remove.
18693 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
18694 * config/avr/avr.c (avr_tolower): New static function.
18695 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
18696 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
18697 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
18700 2013-01-17 Jan Hubicka <jh@suse.cz>
18702 PR tree-optimization/55273
18703 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
18705 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
18708 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
18709 store through atomic_store<mode>_1.
18710 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
18712 2013-01-17 Martin Jambor <mjambor@suse.cz>
18714 PR tree-optimizations/55264
18715 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
18716 for virtual methods.
18717 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
18718 virtual methods before inlining is over.
18719 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
18721 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
18724 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
18726 PR rtl-optimization/56005
18727 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
18728 pending reads for prefetch.
18730 2013-01-16 Ian Bolton <ian.bolton@arm.com>
18732 * config/aarch64/aarch64.md
18733 (*cstoresi_neg_uxtw): New pattern.
18734 (*cmovsi_insn_uxtw): New pattern.
18735 (*<optab>si3_uxtw): New pattern.
18736 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
18737 (*<optab>si3_insn_uxtw): New pattern.
18738 (*bswapsi2_uxtw): New pattern.
18740 2013-01-16 Richard Biener <rguenther@suse.de>
18742 * tree-inline.c (tree_function_versioning): Remove set but
18743 never used variable.
18745 2013-01-16 Richard Biener <rguenther@suse.de>
18747 PR tree-optimization/55964
18748 * tree-flow.h (rename_variables_in_loop): Remove.
18749 (rename_variables_in_bb): Likewise.
18750 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
18751 (copy_loop_before): Adjust and delete update-ssa status.
18752 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
18753 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
18754 (rename_variables_in_loop): Remove.
18755 (slpeel_update_phis_for_duplicate_loop): Likewise.
18756 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
18757 use available cfg machinery instead of duplicating it.
18758 Update PHI nodes and perform poor-mans SSA update here.
18759 (slpeel_tree_peel_loop_to_edge): Adjust.
18761 2013-01-16 Richard Biener <rguenther@suse.de>
18763 PR tree-optimization/54767
18764 PR tree-optimization/53465
18765 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
18766 (vrp_visit_phi_node): For PHI arguments coming via backedges
18767 drop all symbolical range information.
18768 (execute_vrp): Compute backedges.
18770 2013-01-16 Richard Biener <rguenther@suse.de>
18772 * doc/install.texi: Update CLooG and ISL requirements to
18775 2013-01-16 Christian Bruel <christian.bruel@st.com>
18778 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
18779 (broken_move): Handle UNSPECV_SP_SWITCH_B.
18780 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
18782 2013-01-16 DJ Delorie <dj@redhat.com>
18784 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
18785 (UNSPECV_SP_SWITCH_E): New.
18786 (sp_switch_1): Change to an unspec.
18787 (sp_switch_2): Change to an unspec. Don't use post-inc when we
18790 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
18792 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
18793 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
18794 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
18795 (expand_mem_thread_fence): Ditto.
18796 (expand_mem_signal_fence): Ditto.
18797 (expand_atomic_load): Ditto.
18798 (expand_atomic_store): Ditto.
18800 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
18802 PR rtl-optimization/55547
18803 PR rtl-optimization/53827
18806 * alias.c (memrefs_conflict_p): Set sizes to negative after
18809 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18812 * function.c (thread_prologue_and_epilogue_insns): Always
18813 add crtl->drap_reg to set_up_by_prologue.set, even if
18814 stack_realign_drap is false.
18816 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18818 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
18819 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
18820 *call): Fix indention.
18822 2013-01-15 Tom de Vries <tom@codesourcery.com>
18825 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
18828 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
18830 PR rtl-optimization/55153
18831 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
18833 2013-01-15 Martin Jambor <mjambor@suse.cz>
18835 PR tree-optimization/55920
18836 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
18837 accesses as grp_to_be_debug_replaced.
18839 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18841 PR tree-optimization/55920
18842 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
18843 there is non-useless type conversion needed from debug rhs to lhs,
18844 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
18846 2013-01-15 Joseph Myers <joseph@codesourcery.com>
18847 Mikael Pettersson <mikpe@it.uu.se>
18850 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
18852 (ASM_OUTPUT_CASE_LABEL): Remove.
18853 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
18854 * final.c (shorten_branches): Update alignment of labels before
18855 jump tables if CASE_VECTOR_SHORTEN_MODE.
18857 2013-01-15 Richard Biener <rguenther@suse.de>
18860 * system.h: Do not include gmp.h for building host tools.
18862 2013-01-15 Richard Biener <rguenther@suse.de>
18864 PR middle-end/55882
18865 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
18866 account for bitpos when computing alignment.
18868 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
18870 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
18871 (ix86_target_macros_internal): Likewise.
18873 * config/i386/i386.c (m_CORE2I7): Removed.
18874 (m_CORE_HASWELL): New macro.
18875 (m_CORE_ALL): Likewise.
18876 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
18877 (initial_ix86_arch_features): Likewise.
18878 (processor_target_table): Initializations for Core avx2.
18879 (cpu_names): New names "core-avx2".
18880 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
18881 PROCESSOR_CORE_HASWELL.
18882 (ix86_issue_rate): New case.
18883 (ia32_multipass_dfa_lookahead): Likewise.
18884 (ix86_sched_init_global): Likewise.
18886 * config/i386/i386.h (TARGET_HASWELL): New macro.
18887 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
18888 (processor_type): New PROCESSOR_HASWELL.
18890 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18892 PR tree-optimization/55955
18893 * tree-vect-loop.c (vectorizable_reduction): Give up early on
18894 *SHIFT_EXPR and *ROTATE_EXPR codes.
18896 PR tree-optimization/48766
18897 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
18898 -ftrapv disable -fwrapv.
18900 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18903 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
18904 etc. to 1 and not to __flash.
18905 Use LL suffix for __INT24_MAX__ with -mint8.
18906 Use ULL suffix for __UINT24_MAX__ with -mint8.
18908 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18910 * config/avr/avr-arch.h
18911 (struct base_arch_s): Use typedef avr_arch_t instead.
18912 (struct arch_info_s): Use typedef avr_arch_info_t instead.
18913 (struct mcu_type_s): Use typedef avr_mcu_t instead.
18914 * config/avr/avr.c: Same.
18915 * config/avr/avr-devices.c: Same.
18916 * config/avr/driver-avr.c: Same.
18917 * config/avr/gen-avr-mmcu-texi.c: Same.
18918 * config/avr/avr-mcus.def: Adjust comment.
18920 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
18922 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
18923 * config/aarch64/iterators.md (VALLDI): New.
18925 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18926 Andi Kleen <ak@linux.intel.com>
18929 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18930 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18933 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18935 * config/avr/avr-stdint.h: Remove trailing blanks.
18936 * config/avr/avr-log.h: Same.
18937 * config/avr/avr-arch.h: Same.
18938 * config/avr/avr-devices.c: Same.
18939 * config/avr/avr-dimode.md: Same.
18940 * config/avr/predicates.md: Same.
18941 * config/avr/avr-c.c: Same. And fix typo.
18943 * config/avr/avr-protos.h: Same. And:
18944 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18945 (init_cumulative_args): Rename to avr_init_cumulative_args.
18946 (expand_prologue): Rename to avr_expand_prologue.
18947 (expand_epilogue): Rename to avr_expand_epilogue.
18948 (adjust_insn_length): Rename to avr_adjust_insn_length.
18949 (notice_update_cc): Rename to avr_notice_update_cc.
18950 (final_prescan_insn): Rename to avr_final_prescan_insn.
18951 * config/avr/avr.c: Same.
18952 * config/avr/avr.h: Same.
18953 * config/avr/avr.md: Remove trailing blanks.
18954 (prologue): Use avr_expand_prologue.
18955 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18957 2013-01-14 Richard Biener <rguenther@suse.de>
18959 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18960 verify_location, collect_subblocks): New functions.
18961 (verify_gimple_in_cfg): Verify that locations only reference
18962 BLOCKs in the functions BLOCK tree.
18964 2013-01-14 Richard Biener <rguenther@suse.de>
18966 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18968 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18970 (insert_out_of_ssa_copy_on_edge): Likewise.
18971 (rewrite_close_phi_out_of_ssa): Likewise.
18972 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18974 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18975 propagated constants.
18976 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18979 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18981 * config/avr/avr-modes.def: Add GPL copyright notice.
18983 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18985 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18986 MEMMODEL_MASK to determine memory model.
18987 (atomic_store<mode>): Ditto from operands[2].
18988 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18990 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18993 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18994 (fold_gimple_assign): Don't call unshare_expr here.
18995 (fold_ctor_reference): Call unshare_expr.
18997 2013-01-13 Terry Guo <terry.guo@arm.com>
18999 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
19000 * doc/fragments.texi: Document MULTILIB_REUSE.
19001 * gcc.c (multilib_reuse): New internal spec.
19002 (set_multilib_dir): Also search multilib from multilib_reuse.
19003 * genmultilib (tmpmultilib3): Refactor code.
19004 (tmpmultilib4): Ditto.
19005 (multilib_reuse): New multilib argument.
19007 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
19009 * Makefile.in: Update copyright.
19011 2013-01-12 Tom de Vries <tom@codesourcery.com>
19013 PR middle-end/55890
19014 * calls.c (expand_call): Check if arg_nr is valid.
19016 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19018 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
19019 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
19020 documentation. Add missing '__' in front of
19021 __builtin_ia32_packssdw256.
19023 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19026 * config/s390/s390.c (s390_preferred_reload_class): Do not return
19027 NO_REGS for larl operands.
19028 (s390_reload_larl_operand): Use s390_load_address instead of
19031 2013-01-11 Richard Biener <rguenther@suse.de>
19033 * tree-cfg.c (verify_node_sharing_1): Split out from ...
19034 (verify_node_sharing): ... here.
19035 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
19037 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
19039 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
19040 Substitute TREECHECKING.
19041 * configure: Regenerate.
19042 * Makefile.in (TREECHECKING): New.
19044 2013-01-11 Richard Guenther <rguenther@suse.de>
19046 PR tree-optimization/44061
19047 * tree-vrp.c (extract_range_basic): Compute zero as
19048 value-range for __builtin_constant_p of function parameters.
19050 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
19052 Update copyright years.
19054 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
19056 PR rtl-optimization/55672
19057 * lra-eliminations.c (mark_not_eliminable): Permit addition with
19058 const to be eliminable.
19060 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
19062 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
19063 * configure: Regenerate.
19065 2013-01-10 Richard Biener <rguenther@suse.de>
19067 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
19069 2013-01-10 Richard Biener <rguenther@suse.de>
19072 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
19073 locations for virtual PHI arguments.
19074 (rewrite_update_phi_arguments): Likewise.
19076 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
19078 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
19081 2013-01-10 Jakub Jelinek <jakub@redhat.com>
19083 PR tree-optimization/55921
19084 * tree-complex.c (expand_complex_asm): New function.
19085 (expand_complex_operations_1): Call it for GIMPLE_ASM.
19087 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19090 * config/s390/s390.c (s390_symref_operand_p)
19091 (s390_loadrelative_operand_p): Merge the two functions.
19092 (s390_check_qrst_address, print_operand_address): Add parameters
19093 to s390_loadrelative_operand_p invokation.
19094 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
19095 (s390_reload_larl_operand, s390_secondary_reload): Use
19096 s390_loadrelative_operand_p instead of s390_symref_operand_p.
19097 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
19099 2013-01-09 Mike Stump <mikestump@comcast.net>
19101 * dse.c (record_store): Remove unnecessary assert.
19103 2013-01-09 Jan Hubicka <jh@suse.cz>
19105 PR tree-optimization/55569
19106 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
19107 * cfgloop.h (scale_loop_profile): Likewise.
19109 2013-01-09 Jan Hubicka <jh@suse.cz>
19112 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
19114 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
19116 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
19118 PR middle-end/55114
19119 * expr.h (maybe_emit_group_store): Declare.
19120 * expr.c (maybe_emit_group_store): New function.
19121 * builtins.c (expand_builtin_int_roundingfn): Call it.
19122 (expand_builtin_int_roundingfn_2): Likewise.
19124 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
19126 PR rtl-optimization/55829
19127 * lra-constraints.c (match_reload): Add code for absent output.
19128 (curr_insn_transform): Add code for reloads of matched inputs
19131 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
19133 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
19134 attribute of movddup insn to DF.
19135 (*vec_interleave_lowv2df): Ditto.
19136 (vec_dupv2df): Ditto.
19138 2013-01-09 Jan Hubicka <jh@suse.cz>
19140 PR tree-optimiation/55875
19141 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
19142 EVERY_ITERATION parameter.
19143 (number_of_iterations_exit): Check if exit is executed every iteration.
19144 (idx_infer_loop_bounds): Similarly here.
19145 (n_of_executions_at_most): Simplify
19146 to only test for cases where statement is dominated by the
19147 particular bound; handle correctly the "postdominance" test.
19148 (scev_probably_wraps_p): Use max loop iterations info
19149 as a global bound first.
19151 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
19152 Nick Clifton <nickc@redhat.com>
19154 * config/v850/v850.md (cbranchsf4): New pattern.
19155 (cstoresf4): New pattern.
19156 (cbranchdf4): New pattern.
19157 (cstoredf4): New pattern.
19158 (movsicc): Disallow floating point comparisons.
19159 (cmpsf_le_insn): Fix order of operators.
19160 (cmpsf_lt_insn): Likewise.
19161 (cmpsf_eq_insn): Likewise.
19162 (cmpdf_le_insn): Likewise.
19163 (cmpdf_lt_insn): Likewise.
19164 (cmpdf_eq_insn): Likewise.
19165 (cmpsf_ge_insn): Use LE comparison.
19166 (cmpdf_ge_insn): Likewise.
19167 (cmpsf_gt_insn): Use LT comparison.
19168 (cmpdf_gt_insn): Likewise.
19169 (cmpsf_ne_insn): Delete pattern.
19170 (cmpdf_ne_insn): Delete pattern.
19171 * config/v850/v850.c (v850_gen_float_compare): Use
19172 gen_cmpdf_eq_insn for NE comparison.
19173 (v850_float_z_comparison_operator)
19174 (v850_float_nz_comparison_operator): Move from here ...
19175 * config/v850/predicates.md: ... to here. Move GT and GE
19176 comparisons into v850_float_z_comparison_operator.
19177 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
19179 (v850_float_nz_comparison_operator): Likewise.
19181 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19183 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
19184 with calls to gen_insvsi/gen_insvdi.
19186 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19188 * config/i386/i386.c (initial_ix86_tune_features): Set up
19189 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
19191 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
19192 Jakub Jelinek <jakub@redhat.com>
19194 PR tree-optimization/48189
19195 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
19196 If nitercst is 0, don't predict the exit edge.
19198 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19200 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
19201 in asm_fprintf with reg_names.
19202 (aarch64_print_operand_address): Likewise.
19203 (aarch64_return_addr): Likewise.
19204 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
19206 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19208 * config/pa/pa.h (VAL_U6_BITS_P): Define.
19209 (INT_U6_BITS): Likewise.
19210 * config/pa/predicates.md (uint6_operand): New predicate.
19211 (shift5_operand, shift6_operand): Likewise.
19212 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
19214 (lshrdi3): Use shift6_operand.
19215 (shrpsi4, shrpdi4): New insn patterns.
19216 (extzv): Delete expander.
19217 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
19218 predicates in unamed zero extract patterns. Tighten common constraint.
19219 (extv): Delete expander.
19220 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
19221 predicates in unamed sign extract patterns. Tighten common constraint.
19222 (insv): Delete expander.
19223 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
19224 predicates in unamed insert patterns. Tighten common constraint.
19225 Change uint32_operand predicate to uint6_operand predicate in unamed
19226 DImode pattern to insert constant values of type 1...1xxxx.
19228 2013-01-04 Jan Hubicka <jh@suse.cz>
19230 PR tree-optimization/55823
19231 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
19234 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19235 Uros Bizjak <ubizjak@gmail.com>
19237 PR rtl-optimization/55845
19238 * df-problems.c (can_move_insns_across): Stop scanning at
19239 volatile_insn_p source instruction or give up if
19240 across_from .. across_to range contains any volatile_insn_p
19243 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19245 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
19246 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
19248 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
19249 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
19251 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19254 * asan.c (asan_clear_shadow): New function.
19255 (asan_emit_stack_protection): Use it.
19257 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19259 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
19260 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
19261 with tab instead of space.
19263 2013-01-08 Nick Clifton <nickc@redhat.com>
19265 * config/rl78/rl78.c (rl78_expand_prologue): Always select
19266 register bank 0 at the start of an interrupt handler.
19267 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
19270 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19272 * config/aarch64/aarch64-simd.md
19273 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
19274 (aarch64_simd_bsl): Likewise.
19275 (aarch64_vcond_internal<mode>): Likewise.
19276 (vcond<mode><mode>): Likewise.
19277 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
19278 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
19280 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19282 * config/aarch64/aarch64-builtins.c
19283 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
19285 2013-01-08 Martin Jambor <mjambor@suse.cz>
19288 * tree-sra.c (analyze_access_subtree): Return true also after
19289 potentially creating a debug-only replacement.
19291 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19293 PR middle-end/55890
19294 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
19296 PR tree-optimization/54120
19297 * tree-vrp.c (range_fits_type_p): Don't allow
19298 src_precision < precision from signed vr to unsigned_p
19299 if vr->min or vr->max is negative.
19300 (simplify_float_conversion_using_ranges): Test can_float_p
19301 against CODE_FOR_nothing.
19303 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19304 Richard Biener <rguenther@suse.de>
19306 PR middle-end/55851
19307 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
19308 types instead of just INTEGER_TYPE types.
19310 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
19312 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
19315 2013-01-07 Steve Ellcey <sellcey@mips.com>
19318 * config/mips/mips.opt: Change mad to mmad to match documentation.
19320 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19323 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
19324 .progmemx.data now.
19326 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19329 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
19330 (avr_addrspace_t): Add .section_name field.
19331 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
19333 (avr_addrspace): Same. Initialize .section_name. Remove last
19334 NULL entry. Put __memx into .progmemx.data.
19335 (progmem_section_prefix): Remove.
19336 (avr_asm_init_sections): No need to initialize progmem_section.
19337 (avr_asm_named_section): Use avr_addrspace[].section_name to get
19338 section name prefix.
19339 (avr_asm_select_section): Ditto. And use get_unnamed_section to
19340 retrieve the progmem section.
19341 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
19342 boundary to run over avr_addrspace[].
19343 (avr_register_target_pragmas): Ditto.
19345 2013-01-06 Jakub Jelinek <jakub@redhat.com>
19347 * varasm.c (output_constant_def_contents): For asan_protect_global
19348 protected strings, adjust DECL_ALIGN if needed, before testing for
19350 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
19351 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
19353 (output_object_block): For asan protected decls, emit asan padding
19354 after their contents.
19355 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
19356 (asan_finish_file): Test it here instead.
19358 2013-01-07 Nick Clifton <nickc@redhat.com>
19359 Matthias Klose <doko@debian.org>
19360 Doug Kwan <dougkwan@google.com>
19361 H.J. Lu <hongjiu.lu@intel.com>
19364 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
19366 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
19368 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
19370 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
19372 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
19374 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19377 * doc/install.texi (Cross-Compiler-Specific Options): Document
19380 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
19382 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
19383 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
19384 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
19385 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
19386 vqmovun_high_s64): Fix source operand number and update copyright.
19388 2013-01-07 Richard Biener <rguenther@suse.de>
19390 PR middle-end/55890
19391 * gimple.h (gimple_call_builtin_p): New overload.
19392 * gimple.c (validate_call): New function.
19393 (gimple_call_builtin_p): Likewise.
19394 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19395 Use gimple_call_builtin_p.
19396 (find_func_clobbers): Likewise.
19397 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
19398 (strlen_optimize_stmt): Likewise.
19400 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
19402 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
19403 (vld1q_dup_*): Likewise.
19404 (vld1_*): Likewise.
19405 (vld1q_*): Likewise.
19406 (vld1_lane_*): Likewise.
19407 (vld1q_lane_*): Likewise.
19409 2013-01-07 Richard Biener <rguenther@suse.de>
19411 * lto-streamer.h (LTO_minor_version): Bump to 2.
19413 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
19415 * config/aarch64/aarch64-protos.h
19416 (aarch64_const_double_zero_rtx_p): Rename to...
19417 (aarch64_float_const_zero_rtx_p): ...this.
19418 (aarch64_float_const_representable_p): New.
19419 (aarch64_output_simd_mov_immediate): Likewise.
19420 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
19421 move immediate case.
19422 * config/aarch64/aarch64.c
19423 (aarch64_const_double_zero_rtx_p): Rename to...
19424 (aarch64_float_const_zero_rtx_p): ...this.
19425 (aarch64_print_operand): Allow printing of new constants.
19426 (aarch64_valid_floating_const): New.
19427 (aarch64_legitimate_constant_p): Check for valid floating-point
19429 (aarch64_simd_valid_immediate): Likewise.
19430 (aarch64_vect_float_const_representable_p): New.
19431 (aarch64_float_const_representable_p): Likewise.
19432 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
19433 (aarch64_output_simd_mov_immediate): New.
19434 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
19435 (*movdf_aarch64): Likewise.
19436 * config/aarch64/constraints.md (Ufc): New.
19437 (Y): call aarch64_float_const_zero_rtx.
19438 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
19440 2013-01-07 Richard Biener <rguenther@suse.de>
19442 PR tree-optimization/55888
19443 PR tree-optimization/55862
19444 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
19445 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
19446 not if it is contained therein.
19448 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19450 * config/avr/t-avr: Typo.
19452 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19455 * config/avr/t-avr: Don't automatically rebuild
19456 $(srcdir)/config/avr/t-multilib
19457 $(srcdir)/config/avr/avr-tables.opt
19458 $(srcdir)/doc/avr-mmcu.texi
19459 (avr-mcus): New phony target to build them on request.
19460 (s-avr-mlib, s-avr-mmcu-texi): Remove.
19461 * avr/avr-mcus.def: Adjust comments.
19463 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
19465 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
19467 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
19469 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
19471 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
19473 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
19475 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
19477 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
19478 to generate profiling.
19479 * config/rs6000/aix64.h (LIB_SPEC): Same.
19481 2013-01-04 Andrew Pinski <apinski@cavium.com>
19483 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
19485 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
19487 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
19489 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
19491 (ix86_expand_move): Ditto.
19492 (ix86_zero_extend_to_Pmode): Ditto.
19493 (ix86_expand_call): Ditto.
19494 (ix86_expand_special_args_builtin): Ditto.
19495 (ix86_expand_builtin): Ditto.
19497 2013-01-04 Richard Biener <rguenther@suse.de>
19499 PR tree-optimization/55862
19500 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
19501 translating them through PHI nodes.
19503 2013-01-04 Martin Jambor <mjambor@suse.cz>
19505 PR tree-optimization/55755
19506 * tree-sra.c (sra_modify_assign): Do not check that an access has no
19507 children when trying to avoid producing a VIEW_CONVERT_EXPR.
19509 2013-01-04 Marek Polacek <polacek@redhat.com>
19511 PR middle-end/55859
19512 * opts.c (default_options_optimization): Clarify error message.
19514 2013-01-04 Richard Biener <rguenther@suse.de>
19516 PR middle-end/55863
19517 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
19520 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19523 * config/pa/pa.md (movsi): Revert previous change.
19524 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
19527 2013-01-03 Richard Henderson <rth@redhat.com>
19529 * config/i386/i386.c (ix86_expand_move): Always assign to op1
19530 after eliminating TLS symbols.
19532 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
19535 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
19536 * graphite-poly.c (debug_gmp_value): Likewise.
19538 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
19541 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
19542 selected code model, define __code_mode_small__, __code_model_medium__,
19543 __code_model_large__, __code_model_32__ or __code_model_kernel__.
19544 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
19545 xchg temporary register with %k. Declare temporary register as
19547 [__x86_64__]: For medium and large code models, preserve %rbx register.
19549 2013-01-03 Richard Biener <rguenther@suse.de>
19551 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
19552 (dump_subscript): Adjust.
19553 (finalize_ddr_dependent): Do not dump redundant info.
19554 (analyze_siv_subscript): Adjust.
19555 (subscript_dependence_tester): Likewise.
19556 (compute_affine_dependence): Likewise.
19558 2013-01-03 Richard Biener <rguenther@suse.de>
19561 2013-01-03 Richard Biener <rguenther@suse.de>
19563 PR tree-optimization/55857
19564 * tree-vect-stmts.c (vectorizable_load): Do not setup
19565 re-alignment for invariant loads.
19567 2013-01-02 Richard Biener <rguenther@suse.de>
19569 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
19570 invariant load do not generate a vector load from the scalar location.
19572 2013-01-03 Richard Biener <rguenther@suse.de>
19574 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
19575 for not vectorizing.
19576 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
19577 not build INDIRECT_REFs, call get_name once only.
19578 (vect_create_data_ref_ptr): Likewise. Dump base object kind
19579 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
19581 2013-01-03 Richard Biener <rguenther@suse.de>
19583 PR tree-optimization/55857
19584 * tree-vect-stmts.c (vectorizable_load): Do not setup
19585 re-alignment for invariant loads.
19587 2013-01-03 Richard Biener <rguenther@suse.de>
19590 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
19591 prefer a built-in decl.
19593 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19595 * gcc.c (process_command): Update copyright notice dates.
19596 * gcov.c (print_version): Likewise.
19597 * gcov-dump.c (print_version): Likewise.
19599 PR rtl-optimization/55838
19600 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
19601 iv0.step, iv1.step and step.
19603 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19604 Marc Glisse <marc.glisse@inria.fr>
19606 PR tree-optimization/55832
19607 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
19608 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
19609 integer_{one,zero}_node.
19611 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19614 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
19615 * var-tracking.c (reverse_op): Don't add reverse ops to
19616 VALUEs that have already
19617 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
19619 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
19621 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
19623 2013-01-02 Teresa Johnson <tejohnson@google.com>
19625 * dumpfile.c (dump_loc): Print filename with location.
19626 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
19627 new location_t parameter to emit complete unroll message with
19628 new dump framework.
19629 (canonicalize_loop_induction_variables): Compute loops location
19630 and pass to try_unroll_loop_completely.
19631 * loop-unroll.c (report_unroll_peel): New function.
19632 (peel_loops_completely): Use new dump format with location
19633 for main dumpfile message, and invoke report_unroll_peel on success.
19634 (decide_unrolling_and_peeling): Ditto.
19635 (decide_peel_once_rolling): Remove old dumpfile message subsumed
19636 by report_unroll_peel.
19637 (decide_peel_completely): Ditto.
19638 (decide_unroll_constant_iterations): Ditto.
19639 (decide_unroll_runtime_iterations): Ditto.
19640 (decide_peel_simple): Ditto.
19641 (decide_unroll_stupid): Ditto.
19642 * cfgloop.c (get_loop_location): New function.
19643 * cfgloop.h (get_loop_location): Declare.
19645 2013-01-02 Sriraman Tallam <tmsriram@google.com>
19647 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
19650 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19652 PR middle-end/55198
19653 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
19654 BLKmode objects when EXPAND_MEMORY is specified.
19656 2013-01-02 Sriraman Tallam <tmsriram@google.com>
19658 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
19660 (fold_builtin_cpu): Do not share cpu model decls across statements.
19662 2013-01-02 Jason Merrill <jason@redhat.com>
19665 * tree.c (build_array_type_1): Revert earlier change.
19667 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
19669 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
19671 * config/aarch64/aarch64-tune.md: Re-generate.
19673 2013-01-02 Richard Biener <rguenther@suse.de>
19675 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
19676 invariant load do not generate a vector load from the scalar location.
19678 2013-01-02 Richard Biener <rguenther@suse.de>
19681 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
19682 * configure: Regenerate.
19684 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
19686 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
19687 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
19688 (expand_builtin_int_roundingfn_2): Keep the original target around
19689 for the fallback case.
19691 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
19693 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
19694 to be clear for sign changes.
19696 2013-01-01 Jan Hubicka <jh@suse.cz>
19698 * ipa-inline-analysis.c: Fix formatting.
19700 2013-01-01 Jakub Jelinek <jakub@redhat.com>
19702 PR tree-optimization/55831
19703 * tree-vect-loop.c (get_initial_def_for_induction): Use
19704 gsi_after_labels instead of gsi_start_bb.
19706 Copyright (C) 2013 Free Software Foundation, Inc.
19708 Copying and distribution of this file, with or without modification,
19709 are permitted in any medium without royalty provided the copyright
19710 notice and this notice are preserved.