1 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
6 2013-09-03 Richard Biener <rguenther@suse.de>
8 * tree-affine.c (add_elt_to_tree): Fix association issue,
9 avoid useless converts and make sure to always return a
10 properly typed result.
12 2013-09-03 Richard Biener <rguenther@suse.de>
15 * fold-const.c (negate_expr_p): Fix division case.
16 (negate_expr): Likewise.
18 2013-09-03 Richard Biener <rguenther@suse.de>
21 * tree-streamer-out.c: Include tm.h.
22 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
24 2013-09-03 Jan Hubicka <jh@suse.cz>
26 * tree-profile.c (tree_profiling): Cleanup CFG when done.
28 2013-09-03 Alan Modra <amodra@gmail.com>
30 * config.gcc (powerpc*-*-linux*): Add support for little-endian
31 multilibs to big-endian target and vice versa.
32 * config/rs6000/t-linux64: Use := assignment on all vars.
33 (MULTILIB_EXTRA_OPTS): Remove fPIC.
34 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
35 * config/rs6000/t-linux64le: New file.
36 * config/rs6000/t-linux64bele: New file.
37 * config/rs6000/t-linux64lebe: New file.
39 2013-09-02 Jan Hubicka <jh@suse.cz>
41 * ipa-inline-transform.c (inline_transform): Do not
42 optimize_inline_calls when not optimizing.
44 2013-09-02 Jan Hubicka <jh@suse.cz>
46 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
47 duplicated nodes for assembler names.
48 * symtab.c (symtab_unregister_node): Do not attempt to unlink
49 hard registers from assembler name hash.
51 2013-09-02 Jan Hubicka <jh@suse.cz>
53 * ipa-split.c (execute_split_functions): Split externally visible
54 functions called once.
56 2013-09-02 Martin Jambor <mjambor@suse.cz>
59 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
60 linked list. When finding the correct duplicate, also consider also
61 the caller in additon to its inlined_to node.
63 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
65 * config/aarch64/aarch64-simd-builtins.def
66 (dup_lane_scalar): Remove.
67 * config/aarch64/aarch64-simd.md
68 (aarch64_simd_dup): Add 'w->w' alternative.
69 (aarch64_dup_lane<mode>): Allow for VALL.
70 (aarch64_dup_lane_scalar<mode>): Remove.
71 (aarch64_dup_lane_<vswap_width_name><mode>): New.
72 (aarch64_get_lane_signed<mode>): Add w->w altenative.
73 (aarch64_get_lane_unsigned<mode>): Likewise.
74 (aarch64_get_lane<mode>): Likewise.
75 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
76 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
77 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
78 (VCON): Change container of V2SF.
79 (vswap_width_name): Likewise.
80 * config/aarch64/arm_neon.h
81 (__aarch64_vdup_lane_any): New.
82 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
83 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
84 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
86 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
89 * expr.c (emit_move_complex): Do not move complex FP values as parts if
90 the source or the destination is a single hard register.
92 2013-09-02 Richard Biener <rguenther@suse.de>
95 * tree-scalar-evolution.c (instantiate_scev_name): Allow
98 2013-09-02 Richard Biener <rguenther@suse.de>
100 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
101 arithmetic to sizetype.
103 2013-09-02 Bin Cheng <bin.cheng@arm.com>
105 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
106 Find auto-increment use both before and after candidate.
108 2013-09-02 Marek Polacek <polacek@redhat.com>
110 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
112 2013-09-01 Jan Hubicka <jh@suse.cz>
114 * Makefile.in: Add ipa-profile.o
115 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
116 * cgraph.c (struct cgraph_propagate_frequency_data,
117 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
118 ipa-profile.c; replace cgraph_ by ipa_ prefix.
119 * cgraph.h (cgraph_propagate_frequency): Remove.
120 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
121 (inline_update_callee_summaries): Update.
122 * ipa-profile.c: New file.
123 * ipa-utils.h (ipa_propagate_frequency): Declare.
124 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
125 data-streamer.h, value-prof.h
126 (symtab_remove_unreachable_nodes): Update profile.
127 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
128 account_time_size, cmp_counts, dump_histogram,
129 ipa_profile_generate_summary, ipa_profile_write_summary,
130 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
131 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
132 Move to ipa-profile.c
134 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
136 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
138 2013-09-01 Jan Hubicka <jh@suse.cz>
140 * common.opt (fdevirtualize-speculatively): New function.
141 * invoke.texi (fdevirtualize-speculatively): Document.
142 * ipa-devirt.c: Include ipa-inline.h
143 (likely_target_p): New function.
144 (ipa_devirt): New function.
145 (gate_ipa_devirt): New function.
146 (pass_data_ipa_devirt): New static var.
147 (pass_ipa_devirt): Likewise.
148 (make_pass_ipa_devirt): New function.
149 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
150 (common_handle_option): Disable devirtualization when
151 value range profiling is available.
152 * passes.def (pass_ipa_devirt): Add.
153 * timever.def (TV_IPA_DEVIRT): New timevar.
154 * tree-pass.h (make_pass_ipa_devirt):
156 2013-09-01 Iain Sandoe <iain@codesourcery.com>
158 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
159 include sanitize(undefined).
161 2013-08-31 Diego Novillo <dnovillo@google.com>
163 * Makefile.in (TREE_CORE_H): Define.
165 (GTFILES): Add tree-core.h.
166 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
168 * tree-core.h: New file.
169 Move all data structures, enum, typedefs, global
170 declarations and constants from ...
173 2013-08-31 Jan Hubicka <jh@suse.cz>
175 * bulitins.c (expand_builtin): Do not early exit for gcov
176 instrumented functions.
178 2013-08-31 Marek Polacek <polacek@redhat.com>
180 * ubsan.c: Include tm_p.h.
182 2013-08-31 Jan Hubicka <jh@suse.cz>
184 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
187 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
188 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
189 * tree-cfg.c (verify_gimple_label): ... here.
190 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
191 (ipa_merge_profiles): New function.
192 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
193 (lto_input_function_body): Likewise.
194 * ipa-utils.h (ipa_merge_profiles): Declare.
195 * lto-streamer.h (lto_input_function_body): Update prototype.
196 (emit_label_in_global_context_p): Remove.
197 * lto-symtab.c: Include ipa-utils.h
198 (lto_cgraph_replace_node): Use ipa_merge_profiles.
200 2013-08-31 Jan Hubicka <jh@suse.cz>
202 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
204 2013-08-31 Jan Hubicka <jh@suse.cz>
206 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
208 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
210 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
211 "cmp" RTX before signed_comparison_operator check to account
214 2013-08-30 Jan Hubicka <jh@suse.cz>
216 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
217 (detect_type_change_1): Rename to ...
218 (detect_type_change): ... this one; early return on non-polymorphic
220 (detect_type_change_ssa): Add comp_type parameter; update
221 use of detect_type_change.
222 (compute_complex_assign_jump_func): Add param_type parameter;
223 update use of detect_type_change_ssa.
224 (compute_complex_ancestor_jump_func): Likewise.
225 (ipa_get_callee_param_type): New function.
226 (ipa_compute_jump_functions_for_edge): Compute parameter type;
227 update calls to the jump function computation functions.
229 2013-08-30 Teresa Johnson <tejohnson@google.com>
230 Steven Bosscher <steven@gcc.gnu.org>
232 * cfgrtl.c (fixup_new_cold_bb): New routine.
233 (commit_edge_insertions): Invoke fixup_partitions.
234 (find_partition_fixes): New routine.
235 (fixup_partitions): Ditto.
236 (verify_hot_cold_block_grouping): Update comments.
237 (rtl_verify_edges): Invoke find_partition_fixes.
238 (rtl_verify_bb_pointers): Update comments.
239 (rtl_verify_bb_layout): Ditto.
240 * basic-block.h (probably_never_executed_edge_p): Declare.
241 (fixup_partitions): Ditto.
242 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
243 * bb-reorder.c (sanitize_hot_paths): New function.
244 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
246 * predict.c (probably_never_executed_edge_p): New routine.
247 * cfg.c (check_bb_profile): Add partition insanity warnings.
249 2013-08-30 Meador Inge <meadori@codesourcery.com>
251 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
253 2013-08-30 Marek Polacek <polacek@redhat.com>
255 * Makefile.in (ubsan.o): Add.
256 (c-family/c-ubsan.o): Add.
257 (builtins.o): Add ubsan.h dependency.
260 * common.opt: Add -fsanitize=undefined option.
261 (flag_sanitize): Add variable.
262 (fsanitize=): Add option. Add Driver.
263 (fsanitize=thread): Remove option.
264 (fsanitize=address): Likewise.
265 (static-libubsan): New option.
266 * doc/invoke.texi: Document the new flag and -static-libubsan.
267 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
268 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
269 * builtin-attrs.def (ATTR_COLD): Define.
270 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
271 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
272 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
273 * flag-types.h (sanitize_code): New enum.
274 * opts.c (common_handle_option): Parse command line arguments
275 of -fsanitize=. Add -fsanitize=unreachable option.
276 * varasm.c (get_variable_section): Adjust.
277 (assemble_noswitch_variable): Likewise.
278 (assemble_variable): Likewise.
279 (output_constant_def_contents): Likewise.
280 (categorize_decl_for_section): Likewise.
281 (place_block_symbol): Likewise.
282 (output_object_block): Likewise.
283 * builtins.def: Likewise.
284 * toplev.c (compile_file): Likewise.
285 (process_options): Likewise.
286 * cppbuiltin.c: Likewise.
287 * tsan.c (tsan_pass): Likewise.
288 (tsan_gate): Likewise.
289 (tsan_gate_O0): Likewise.
290 * cfgexpand.c (partition_stack_vars): Likewise.
291 (expand_stack_vars): Likewise.
292 (defer_stack_allocation): Likewise.
293 (expand_used_vars): Likewise.
294 * cfgcleanup.c (old_insns_match_p): Likewise.
295 * asan.c (asan_finish_file): Likewise.
296 (asan_instrument): Likewise.
297 (gate_asan): Likewise.
298 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
299 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
300 (asan_global_struct): Use pointer_sized_int_node instead
301 calling build_nonstandard_integer_type.
302 (initialize_sanitizer_builtins): Likewise.
303 (asan_finish_file): Likewise.
304 * gcc.c: Document %{%:function(args):X}.
305 (static_spec_functions): Add sanitize.
306 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
307 store funcval != NULL there.
308 (do_spec_1): Adjust handle_spec_function caller.
309 (handle_braces): Allow %:function(args) as condition.
310 (sanitize_spec_function): New function.
311 (ADD_STATIC_LIBUBSAN_LIBS): Define.
312 (LIBUBSAN_SPEC): Likewise.
313 (LIBUBSAN_EARLY_SPEC): Likewise.
314 (SANITIZER_SPEC): Handle libubsan.
315 (SANITIZER_EARLY_SPEC): Likewise.
316 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
317 instead of fsanitize=address.
318 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
319 instead of fsanitize=address*.
320 * builtins.c: Include ubsan.h.
321 (fold_builtin_0): Instrument __builtin_unreachable.
322 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
323 instead of flag_asan.
324 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
325 (pointer_sized_int_node): Define.
326 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
328 2013-08-30 Mike Stump <mikestump@comcast.net>
330 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
333 2013-08-29 Jan Hubicka <jh@suse.cz>
335 * cgraph.c (cgraph_function_body_availability): Handle weakref
337 * passes.def: Remove pass_fixup_cfg.
338 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
339 track when we need to remove functions.
340 (gate_ipa_inline): Execute inlining always; add comment why.
341 (pass_data_ipa_inline): Remove TODO_remove_functions.
342 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
343 do not produce summaries.
344 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
345 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
346 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
347 constant pool and vtable.
349 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
351 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
352 New arm_neon.h's internal macros to specify 64-bit constants.
353 Avoid using stdint.h's macros.
355 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
357 * recog.c (verify_changes): Verify that changes[i].old is non-zero
358 before applying REG_P.
360 2013-08-30 Jakub Jelinek <jakub@redhat.com>
362 PR tree-optimization/58277
363 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
364 after seeing too many stmts with vdef in between dombb and current
365 bb, invalidate everything.
367 2013-08-30 Richard Biener <rguenther@suse.de>
369 * fold-const.c (fold_single_bit_test): Fix overflow test.
371 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
373 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
374 and which can live in a register, always retrieve the value on entry.
375 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
376 passed by invisible reference specially.
377 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
378 (vt_add_function_parameter): Correctly deal with a parameter passed by
381 2013-08-30 Jan Hubicka <jh@suse.cz>
383 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
385 2013-08-30 Richard Biener <rguenther@suse.de>
387 PR tree-optimization/58228
388 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
389 allow invariant loads in nested loop vectorization.
391 2013-08-30 Richard Biener <rguenther@suse.de>
393 PR tree-optimization/58223
394 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
395 (has_anti_or_output_dependence): ... this and adjust to also
396 look for output dependences.
397 (mark_nodes_having_upstream_mem_writes): Adjust.
398 (rdg_flag_uses): Likewise.
400 2013-08-30 Richard Biener <rguenther@suse.de>
402 PR tree-optimization/58010
403 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
404 assert that we have a loop-closed PHI.
406 2013-08-29 Jan Hubicka <jh@suse.cz>
408 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
409 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
410 * lto-section-in.c (lto_free_function_in_decl_state): New function.
411 (lto_free_function_in_decl_state_for_node): New function.
413 2013-08-29 Xinliang David Li <davidxl@google.com>
415 * loop-unroll.c (report_unroll_peel): Minor message change.
416 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
417 Emit alignment peeling message with default -fopt-info.
418 (vect_loop_versioning): Emit loop version info message.
419 * tree-vectorizer.c (vectorize_loops): Minor message change.
420 (execute_vect_slp): Ditto.
422 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
424 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
425 of the clone from the DECL_NAME of the original function.
427 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
429 * passes.c (register_pass): Add overload.
430 * tree-pass.h (register_pass): Forward declare it. Add comment.
432 2013-08-29 Jan Hubicka <jh@suse.cz>
434 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
435 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
436 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
437 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
438 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
439 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
440 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
441 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
443 2013-08-29 Teresa Johnson <tejohnson@google.com>
445 * dumpfile.c (dump_loc): Output column number.
446 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
447 * doc/invoke.texi: Document optall -fopt-info flag.
448 * profile.c (read_profile_edge_counts): Use new dump framework.
449 (compute_branch_probabilities): Ditto.
450 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
451 when pass not in any opt group.
452 * pass_manager.h (pass_manager::get_pass_profile): New method.
453 * value-prof.c (check_counter): Use new dump framework.
454 (check_ic_target): Ditto.
455 * coverage.c (get_coverage_counts): Ditto.
456 (coverage_init): Setup new dump framework.
458 2013-08-29 Richard Biener <rguenther@suse.de>
460 PR tree-optimization/58246
461 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
462 handle the dominance check inside a basic-block.
464 2013-08-29 Richard Biener <rguenther@suse.de>
467 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
468 of default defs that appear in abnormal PHI nodes.
470 2013-08-29 Richard Biener <rguenther@suse.de>
472 PR tree-optimization/57685
473 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
474 single-use operands to avoid exponential complexity.
476 2013-08-28 Dehao Chen <dehao@google.com>
478 * ipa-inline.c (edge_badness): Fix integer underflow.
480 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
482 * gtm-builtins.def (_ITM_free): Declare leaf.
484 2013-08-28 Jakub Jelinek <jakub@redhat.com>
487 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
488 (*tls_local_dynamic_base_64_largepic): Likewise.
489 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
490 Remove predicate from call operand.
491 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
492 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
494 2013-08-28 Jeff Law <law@redhat.com>
496 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
497 checks for the number of predecessors and successors allowed.
498 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
499 which require copying a joiner block if there is a request which
500 is a subpath that requires no joiner block copying.
502 2013-08-28 Jan Hubicka <jh@suse.cz>
504 * lto-streamer-out.c (DFS_write_tree_body): Drop
505 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
506 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
507 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
508 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
509 Do not read DECL_ERROR_ISSUED.
510 (unpack_ts_decl_with_vis_value_fields): Do not read
512 (lto_input_ts_binfo_tree_pointers): Do not read
513 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
514 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
515 write DECL_ERROR_ISSUED..
516 (pack_ts_decl_with_vis_value_fields): Do not write
518 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
519 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
520 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
521 * tree.h (tree_decl_common): Update comment.
522 (DECL_ERROR_ISSUED): Remove.
524 2013-08-28 Jakub Jelinek <jakub@redhat.com>
527 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
529 2013-08-28 Jan Hubicka <jh@suse.cz>
531 * builtins.def (free): Declare leaf.
533 2013-08-27 David Malcolm <dmalcolm@redhat.com>
536 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
537 * configure: Regenerate.
539 2013-08-27 Martin Jambor <mjambor@suse.cz>
541 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
542 (ipa_ancestor_jf_data): Likewise.
543 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
544 (ipa_get_jf_pass_through_type_preserved): New function.
545 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
546 (ipa_get_jf_ancestor_type_preserved): New function.
547 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
548 (ipa_get_jf_ancestor_result): Likewise.
549 (propagate_vals_accross_pass_through): Use
550 ipa_get_jf_pass_through_result to do all the value mappings.
551 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
553 (ipa_set_jf_cst_copy): New function.
554 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
555 (ipa_set_jf_arith_pass_through): Likewise.
556 (ipa_set_ancestor_jf): Likewise.
557 (compute_complex_assign_jump_func): Set type_preserved instead of
559 (ipa_compute_jump_functions_for_edge): Likewise.
560 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
561 (update_jump_functions_after_inlining): Update type_preserved.
562 Explicitely create jump functions when combining one with pass_through.
563 (ipa_write_jump_function): Stream the type_preserved flags.
564 (ipa_read_jump_function): Likewise.
566 2013-08-27 Jakub Jelinek <jakub@redhat.com>
567 Aldy Hernandez <aldyh@redhat.com>
569 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
570 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
571 * function.h (struct function): Add has_force_vect_loops and
573 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
574 * gimple.c (gimple_build_omp_critical): Add KIND argument and
576 * gimple.def: Update CLAUSES comments.
577 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
578 (gimple_build_omp_for): Add argument to prototype.
579 (gimple_omp_for_kind): New.
580 (gimple_omp_for_set_kind): New.
581 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
582 GOVD_DATA_SHARE_CLASS.
583 (enum omp_region_type): Add ORT_SIMD.
584 (gimple_add_tmp_var): Handle ORT_SIMD.
585 (gimplify_var_or_parm_decl): Same.
586 (is_gimple_stmt): Same.
587 (omp_firstprivatize_variable): Same.
588 (omp_add_variable): Only use splay_tree_insert if lookup failed.
589 (omp_notice_variable): Handle ORT_SIMD.
590 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
591 (omp_check_private): Handle ORT_SIMD.
592 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
594 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
595 Handle OMP_CLAUSE_LASTPRIVATE.
596 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
598 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
599 (gimplify_expr): Handle OMP_SIMD.
600 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
601 (expand_GOMP_SIMD_VF): New.
602 (expand_GOMP_SIMD_LAST_LANE): New.
603 * internal-fn.def (GOMP_SIMD_LANE): New.
605 (GOMP_SIMD_LAST_LANE): New.
606 * omp-low.c: Include target.h.
607 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
609 (check_omp_nesting_restrictions): Same.
611 (lower_rec_simd_input_clauses): New.
612 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
614 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
615 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
616 (expand_omp_build_assign): New.
617 (expand_omp_for_init_counts): New.
618 (expand_omp_for_init_vars): New.
619 (extract_omp_for_update_vars): New.
620 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
621 and rewrite accordingly.
622 (expand_omp_simd): New.
623 (expand_omp_for): Use expand_omp_simd.
624 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
625 (lower_omp_for): Do not lower the body.
626 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
628 * tree-flow.h (find_omp_clause): Remove prototype.
629 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
630 forcing vectorization of the loop, or if flag_tree_vectorize.
631 (gate_tree_if_conversion): Similarly.
632 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
633 gimple_build_omp_for.
634 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
635 * tree-parloops (create_parallel_loop): Pass kind argument to
636 gimple_build_omp_for.
637 * tree-pretty-print.c (dump_omp_clause): Add cases for
638 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
639 OMP_CLAUSE__SIMDUID_.
640 (dump_generic_node): Handle OMP_SIMD.
641 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
642 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
643 unroll OMP_SIMD loops here.
644 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
645 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
647 (vect_analyze_data_refs): Handle simd loops.
648 * tree-vect-loop.c (vectorizable_live_operation): Handle
650 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
651 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
652 (vectorizable_load): Same.
653 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
654 (struct simduid_to_vf): New.
655 (simduid_to_vf::hash): New.
656 (simduid_to-vf::equal): New.
657 (struct simd_array_to_simduid): New.
658 (simd_array_to_simduid::hash): New.
659 (simd_array_to_simduid::equal): New.
660 (adjust_simduid_builtins): New.
661 (struct note_simd_array_uses_struct): New.
662 (note_simd_array_uses_cb): New.
663 (note_simd_array_uses): New.
664 (vectorize_loops): Handle simd hints and adjust simd builtins
666 * tree-vectorizer.h (struct _stmt_vec_info): Add
667 simd_lane_access_p field.
668 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
669 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
670 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
671 (omp_clause_code_name): Same.
672 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
673 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
674 * tree.def (OMP_SIMD): New entry.
675 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
676 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
677 (OMP_CLAUSE_DECL): Adjust range for new clauses.
678 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
679 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
680 (OMP_CLAUSE_LINEAR_STEP): New.
681 (OMP_CLAUSE_SAFELEN_EXPR): New.
682 (OMP_CLAUSE__SIMDUID__DECL): New.
683 (find_omp_clause): New prototype.
685 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
687 * config/i386/driver-i386.c (host_detect_local_cpu): Update
688 Haswell processor detection.
690 2013-08-27 Christian Widmer <shadow@umbrox.de>
693 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
694 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
695 AVX2 capable processors.
697 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
699 * config/aarch64/arm_neon.h: Replace all inline asm implementations
700 of vget_low_* with implementations in terms of other intrinsics.
702 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
705 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
706 values to -1, 0 and 1.
708 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
710 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
711 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
712 (<optab><mode>3_insn): Remove.
713 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
714 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
715 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
716 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
717 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
718 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
719 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
720 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
721 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
722 (ror<mode>3_insn): Likewise.
723 * config/aarch64/predicates.md (aarch64_simd_register): New.
725 2013-08-27 Richard Biener <rguenther@suse.de>
727 PR tree-optimization/57521
728 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
729 one edge is non-critical.
730 (find_phi_replacement_condition): Make sure to use a non-critical
731 edge. Cleanup and remove old bug workarounds.
732 (bb_postdominates_preds): Remove.
733 (if_convertible_loop_p_1): Do not compute post-dominators.
734 (combine_blocks): Do not free post-dominators.
735 (main_tree_if_conversion): Likewise.
736 (pass_data_if_conversion): Add TODO_verify_ssa.
738 2013-08-27 DJ Delorie <dj@redhat.com>
740 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
742 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
744 * function.c (assign_parm_find_data_types): Set passed_mode and
745 nominal_mode to the TYPE_MODE of nominal_type for the built
746 pointer type in case of the struct-pass-by-reference.
748 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
750 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
751 (UINT16_TYPE): Change default to "unsigned int".
753 * config/avr/avr.opt (mfract-convert-truncate): New option.
754 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
755 is set, round negative fractional integers according to n1169
756 when converting to integer types.
758 2013-08-26 Jan Hubicka <jh@suse.cz>
760 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
761 methods can not be called indirectly when their address is not taken.
763 2013-08-26 Jan Hubicka <jh@suse.cz>
765 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
768 2013-08-26 Jan Hubicka <jh@suse.cz>
770 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
773 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
775 * reload.c (find_valid_class): Allow classes that do not include
776 FIRST_PSEUDO_REGISTER - 1.
778 2013-08-26 Jan Hubicka <jh@suse.cz>
780 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
781 fix edge count/frequency when speculation failed; fix type check
784 2013-08-26 Jan Hubicka <jh@suse.cz>
786 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
788 2013-08-26 Jan Hubicka <jh@suse.cz>
790 * ipa-inline-transform.c (inline_transform): Be ready for basic block
791 to be changed by edge redirection.
793 2013-08-26 Jan Hubicka <jh@suse.cz>
795 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
796 formating; add sanity check.
797 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
798 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
799 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
800 (resolve_noninline_speculation): Update callee keys, too.
802 2013-08-26 Jan Hubicka <jh@suse.cz>
804 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
805 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
807 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
809 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
812 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
814 * config/i386/i386.c (ix86_debug_options): Remove prototype.
815 (x86_64_elf_select_section): Ditto.
816 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
818 (ix86_pass_by_reference): Ditto.
819 (output_set_got): Ditto.
820 (ix86_unary_operator_ok): Ditto.
821 (ix86_expand_builtin): Ditto.
823 2013-08-23 Jan Hubicka <jh@suse.cz>
825 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
827 2013-08-23 Jan Hubicka <jh@suse.cz>
829 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
830 (tree_decl_with_vis): Add FINAL field.
832 2013-08-23 Jeff Law <law@redhat.com>
834 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
835 the debugging dump when the expression is fully redundant.
837 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
839 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
840 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
841 * pretty-print.c (pp_formatted_text_data): Likewise.
842 (pp_write_text_to_stream): Likewise.
843 (pp_write_text_as_dot_label_to_stream): Likewise.
844 (pp_append_r): Likewise.
845 (pp_format): Likewise.
846 (pp_flush): Likewise.
847 (pp_clear_output_area): Likewise.
848 (pp_append_text): Likewise.
849 (pp_formatted_text): Likewise.
850 (pp_remaining_character_count_for_line): Likewise.
851 (pp_newline): Likewise.
852 (pp_character): Likewise.
853 (output_buffer::~output_buffer): Define.
854 (pretty_printer::~pretty_printer): Destruct output buffer.
855 * pretty-print.h (output_buffer::~output_buffer): Declare.
856 (pretty_printer::~pretty_printer): Declare virtual.
858 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
861 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
862 HOST_WIDE_INT_M1U): New macros.
863 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
864 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
865 unsigned -1 for lshift.
866 * cse.c (cse_insn): Likewise.
867 * double-int.c (rshift_double, lshift_double): Likewise.
868 * builtins.c (fold_builtin_bitop): Likewise.
869 * combine.c (force_to_mode): Likewise.
870 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
871 * simplify-rtx.c (simplify_const_unary_operation,
872 simplify_const_binary_operation): Likewise.
873 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
874 check_va_list_escapes): Likewise.
875 * rtlanal.c (nonzero_bits1): Likewise.
876 * expmed.c (expand_smod_pow2): Likewise.
877 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
879 2013-08-23 Jan Hubicka <jh@suse.cz>
881 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
882 as having address taken.
884 2013-08-23 Jan Hubicka <jh@suse.cz>
886 * ipa-utils.h (method_class_type): Declare.
887 * ipa-devirt.c (method_class_type): Export.
889 * cgraphunit.c (analyze_functions): Do basic devirtualization;
890 do not walk base classes of anonymous types.
892 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
894 PR rtl-optimization/58220
896 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
897 handle SEQUENCE insns properly.
899 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
901 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
903 (pp_newline_and_indent): Likewise.
904 (pp_separate_with): Likewise.
905 * pretty-print.c (pp_newline_and_flush): Define.
906 (pp_newline_and_indent): Likewise.
907 (pp_separate_with): Likewise.
909 2013-08-23 Jakub Jelinek <jakub@redhat.com>
912 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
913 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
915 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
917 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
918 * gcc/config/i386/i386.md (*movti_internal): Use
919 predicate to determine if EVEX is needed.
920 (*movsi_internal): Ditto.
921 (*movdf_internal): Ditto.
922 (*movsf_internal): Ditto.
923 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
925 2013-08-23 Jakub Jelinek <jakub@redhat.com>
927 PR tree-optimization/58209
928 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
929 (find_tail_calls): Give up for pointer result types if m is non-NULL.
930 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
931 emit POINTER_PLUS_EXPR.
932 (create_tailcall_accumulator): For pointer result type accumulate in
935 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
937 * configure.ac: Add backslashes missing from the last change.
938 * configure: Regenerate.
940 2013-08-22 Jan Hubicka <jh@susue.cz>
942 * ipa.c (function_and_variable_visibility): First remember function
943 was global and then make it local.
945 2013-08-22 Julian Brown <julian@codesourcery.com>
947 * configure.ac: Add aarch64 to list of arches which use "nop" in
949 * configure: Regenerate.
951 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
953 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
954 gnu_libc_has_function.
955 * config/s390/tpf.h: Likewise.
957 2013-08-22 Jan Hubicka <jh@susue.cz>
959 * timevar.c (validate_phases): Add cast.
961 2013-08-22 Jan Hubicka <jh@susue.cz>
963 * timevar.c (validate_phases): Use size_t for memory.
964 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
966 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
968 * pretty-print.h (output_buffer::output_buffer): Declare.
969 (pretty_printer::pretty_printer): Likewise.
970 (pp_construct): Remove.
971 * pretty-print.c (output_buffer::output_buffer): Define.
972 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
973 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
974 (print_gimple_expr): Likewise.
975 (print_gimple_seq): Likewise.
976 (gimple_dump_bb): Likewise.
977 * sched-vis.c (dump_value_slim): Likewise.
978 (dump_insn_slim): Likewise.
979 (dump_rtl_slim): Likewise.
980 (str_pattern_slim): Likewise.
981 * tree-mudflap.c (mf_varname_tree): Likewise.
982 * graph.c (print_graph_cfg): Likewise.
983 (start_graph_dump): Likewise.
984 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
986 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
987 pretty printer initialization.
988 * coretypes.h (diagnostic_context): Remove superflous type alias
990 (pretty_printer): Likewise. Declare directly as a class.
991 (pretty_print_info): Remove declaration as class.
992 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
993 and pp_clear_output_area.
994 (asan_add_global): Likewise.
996 2013-08-22 Jan Hubicka <jh@suse.cz>
998 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
999 * ipa-utils.h (update_type_inheritance_graph): Declare.
1000 (possible_polymorphic_call_target_p): Declare.
1001 (possible_polymorphic_call_target_p): New.
1002 * ipa-devirt.c: Update toplevel comments.
1003 (cached_polymorphic_call_targets): Move up.
1004 (odr_type_d): Move ID down.
1005 (polymorphic_type_binfo_p): Update comment.
1006 (odr_hasher::remove): Likewise;
1007 (get_odr_type): Set anonymous_namespace.
1008 (dump_odr_type): Dump it.
1009 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
1010 (maybe_record_node): Record node in cached_polymorphic_call_targets.
1011 (record_binfo): Add comment.
1012 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
1014 (devirt_node_removal_hook): Do not iCE when cache is freed.
1015 (possible_polymorphic_call_target_p): New predicate.
1016 (update_type_inheritance_graph): New function.
1018 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
1019 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1020 Sergey Lega <sergey.s.lega@intel.com>
1021 Anna Tikhonova <anna.tikhonova@intel.com>
1022 Ilya Tocar <ilya.tocar@intel.com>
1023 Andrey Turetskiy <andrey.turetskiy@intel.com>
1024 Ilya Verbin <ilya.verbin@intel.com>
1025 Kirill Yukhin <kirill.yukhin@intel.com>
1026 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1028 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
1029 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
1030 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
1031 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
1032 (OPTION_MASK_ISA_AVX2_UNSET): Update.
1033 (OPTION_MASK_ISA_AVX512F_UNSET): New.
1034 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
1035 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
1036 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
1037 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
1038 OPT_mavx512pf, OPT_mavx512er cases.
1039 * config/i386/constraints.md (v): New constraint.
1040 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
1041 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
1042 (bit_AVX512CD): New.
1043 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
1044 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
1045 * config/i386/i386-c.c (ix86_target_macros_internal):
1046 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
1048 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
1049 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
1050 * config/i386/i386.c (regclass_map, dbx_register_map)
1051 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
1052 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
1053 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
1055 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
1056 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
1057 -mavx512pf options. Fix formatting.
1058 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
1059 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
1060 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
1061 -mavx512cd, -mavx512pf options.
1062 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
1063 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
1064 (ix86_preferred_output_reload_class): Replace SSE_REGS with
1066 (ix86_hard_regno_mode_ok): Support 512-bit registers.
1067 (ix86_set_reg_reg_cost): Ditto.
1068 (x86_order_regs_for_local_alloc): Ditto.
1069 (MAX_VECT_LEN): Extend to 64-byte.
1070 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
1071 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
1072 (TARGET_AVX512ER, TARGET_AVX512CD): New.
1073 (BIGGEST_ALIGNMENT): Extend to 512-bits.
1074 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
1075 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
1076 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
1077 (SSE_REG_MODE_P): Support new modes.
1078 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
1079 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
1080 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
1081 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
1082 (REG_CLASS_CONTENTS): Add new registers.
1083 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
1084 (EXT_REX_SSE_REGNO_P): New.
1085 (HI_REGISTER_NAMES): Add new registers.
1086 * config/i386/i386.md: Define constants for new registers.
1087 (mode): Add new 512-bit modes.
1088 (prefix): Support evex prefix.
1089 (isa): Support avx512f, noavx512f, fma_avx512f.
1090 (ssemodesuffix): Add new 512-bit modes.
1092 (*movxi_internal_avx512f): Ditto.
1093 (*movdi_internal): Replace constraint "x" with the new constraint "v".
1095 (*movsi_internal): Likewise.
1096 (*movdf_internal): Likewise.
1097 (*movsf_internal): Likewise.
1098 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
1099 (<code><mode>3): Likewise.
1100 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
1102 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
1103 with the new constraint "v".
1104 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
1106 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
1107 with the new constraint "v".
1108 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
1109 (<sse2>_storedqu<avxsizesuffix>): Likewise.
1110 (*<plusminus_insn><mode>3): Likewise.
1111 (<sse>_vm<plusminus_insn><mode>3): Likewise.
1112 (*mul<mode>3): Likewise.
1113 (<sse>_vmmul<mode>3): Likewise.
1114 (<sse>_div<mode>3): Likewise.
1115 (<sse>_vmdiv<mode>3): Likewise.
1116 (<sse>_sqrt<mode>2): Likewise.
1117 (<sse>_vmsqrt<mode>2): Likewise.
1118 (*<code><mode>3_finite): Likewise.
1119 (*<code><mode>3) <smaxmin>: Likewise.
1120 (<sse>_vm<code><mode>3): Likewise.
1121 (*<code><mode>3) <any_logic>: Likewise.
1122 (*fma_fmadd_<mode>): Likewise.
1123 (*fma_fmsub_<mode>): Likewise.
1124 (*fma_fnmadd_<mode>): Likewise.
1125 (*fma_fnmsub_<mode>): Likewise.
1126 (*fma_fmaddsub_<mode>): Likewise.
1127 (*fma_fmsubadd_<mode>): Likewise.
1128 (*fmai_fmadd_<mode>): Likewise.
1129 (*fmai_fmsub_<mode>): Likewise.
1130 (*fmai_fnmadd_<mode>): Likewise.
1131 (*fmai_fnmsub_<mode>): Likewise.
1132 (sse_cvtsi2ss): Likewise.
1133 (sse_cvtsi2ssq): Likewise.
1134 (sse_cvtss2si): Likewise.
1135 (sse_cvtss2si_2): Likewise.
1136 (sse_cvtss2siq): Likewise.
1137 (sse_cvtss2siq_2): Likewise.
1138 (sse_cvttss2si): Likewise.
1139 (sse_cvtss2siq_2): Likewise.
1140 (float<sseintvecmodelower><mode>2): Likewise.
1141 (sse2_cvtsd2si_2): Likewise.
1142 (sse2_cvtsd2siq_2): Likewise.
1143 (*<plusminus_insn><mode>3): Likewise.
1144 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
1145 (*<sse4_1_avx2>_mul<mode>3): Likewise.
1146 (ashr<mode>3): Likewise.
1147 (<shift_insn><mode>3): Likewise.
1148 (avx2_<code><mode>3): Likewise.
1149 (*avx2_<code><mode>3): Likewise.
1150 (*andnot<mode>3): Likewise.
1151 (*<code><mode>3) <any_logic>: Likewise.
1152 (abs<mode>2): Likewise.
1153 (avx2_permvar<mode>): Likewise.
1154 (avx2_perm<mode>_1): Likewise.
1155 (*avx_vpermilp<mode>): Likewise.
1156 (avx_vpermilvar<mode>3): Likewise.
1157 (avx2_ashrv<mode>): Likewise.
1158 (avx2_<shift_insn>v<mode>): Likewise.
1159 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
1161 * doc/rtl.texi: Document XImode.
1163 2013-08-21 Jeff Law <law@redhat.com>
1165 * tree-flow.h (register_jump_thread): Pass vector of edges
1166 instead of each important edge.
1167 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
1168 thread path into a vector and pass that to register_jump_thread.
1169 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
1170 passed in edge vector to the current 3-edge form.
1173 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1175 * dce.c (fini_dce): Call df_analyze again just in case
1176 delete_unmarked_insns removed anything.
1178 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
1180 * reload.h (struct reg_equivs): Rename to ..
1181 (struct reg_equivs_s): .. this.
1183 2013-08-20 Martin Liska <marxin.liska@gmail.com>
1185 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
1187 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1189 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
1191 2013-08-21 Jeff Law <law@redhat.com>
1193 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
1194 simplify assignments too. If the RHS collapses to a singleton
1195 range, then return the value for the range.
1197 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
1199 * config/i386/sse.md (V16): Rename to...
1201 (mov<mode>): Update iterator name.
1202 (*mov<mode>_internal): Ditto.
1203 (push<mode>1): Ditto.
1204 (movmisalign<mode>): Ditto.
1206 2013-08-20 Jan Hubicka <jh@suse.cz>
1209 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
1210 entry for direct edges.
1211 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
1213 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1215 Revert my last two changes, r201865 and r201864:
1218 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1220 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1221 instances can own GC refs.
1223 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1224 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1225 (gcc::context::gt_pch_nx): Likewise.
1226 (gcc::context::gt_pch_nx): Likewise.
1227 * ggc.h (gt_ggc_mx <T>): New.
1228 (gt_pch_nx_with_op <T>): New.
1229 (gt_pch_nx <T>): New.
1230 * passes.c (opt_pass::gt_ggc_mx): New.
1231 (opt_pass::gt_pch_nx): New.
1232 (opt_pass::gt_pch_nx_with_op): New.
1233 (pass_manager::gt_ggc_mx): New.
1234 (pass_manager::gt_pch_nx): New.
1235 (pass_manager::gt_pch_nx_with_op): New.
1236 (pass_manager::operator new): Use
1237 ggc_internal_cleared_alloc_stat rather than xcalloc.
1238 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1239 (pass_manager::gt_ggc_mx): New.
1240 (pass_manager::gt_pch_nx): New.
1241 (pass_manager::gt_pch_nx_with_op): New.
1242 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1243 (opt_pass::operator new): New.
1244 (opt_pass::gt_ggc_mx): New.
1245 (opt_pass::gt_pch_nx): New.
1246 (opt_pass::gt_pch_nx_with_op): New.
1249 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1251 * Makefile.in (GTFILES): Add context.h.
1252 * context.c (gcc::context::operator new): New.
1253 (gcc::context::gt_ggc_mx): New.
1254 (gcc::context::gt_pch_nx): New.
1255 (gcc::context::gt_pch_nx): New.
1256 * context.h (gcc::context): Add GTY((user)) marking.
1257 (gcc::context::operator new): New.
1258 (gcc::context::gt_ggc_mx): New.
1259 (gcc::context::gt_pch_nx): New.
1260 (gcc::context::gt_pch_nx): New.
1261 (g): Add GTY marking.
1262 (gt_ggc_mx (gcc::context *)): New.
1263 (gt_pch_nx (gcc::context *)): New.
1264 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1265 void *cookie)): New.
1266 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1268 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1270 * dce.c (fini_dce): Call df_analyze again just in case
1271 delete_unmarked_insns removed anything.
1273 2013-08-20 Teresa Johnson <tejohnson@google.com>
1275 PR rtl-optimizations/57451
1276 * final.c (reemit_insn_block_notes): Prevent lexical blocks
1277 from crossing split section boundaries.
1279 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1281 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
1282 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
1283 with MULTLIB_DEFAULTS.
1285 2013-08-20 Nick Clifton <nickc@redhat.com>
1287 * target.def (narrow_volatile_bitfield): Note that the default
1288 value is false, not !TARGET_STRICT_ALIGN.
1289 * doc/tm.texi: Regenerate.
1291 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
1293 Fix LIB_SPEC for systems without libpthread.
1295 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
1296 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
1298 * config/i386/linux-common.h: Likewise.
1299 * config/mips/linux-common.h: Likewise.
1301 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1303 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
1306 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1308 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1309 instances can own GC refs.
1311 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1312 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1313 (gcc::context::gt_pch_nx): Likewise.
1314 (gcc::context::gt_pch_nx): Likewise.
1315 * ggc.h (gt_ggc_mx <T>): New.
1316 (gt_pch_nx_with_op <T>): New.
1317 (gt_pch_nx <T>): New.
1318 * passes.c (opt_pass::gt_ggc_mx): New.
1319 (opt_pass::gt_pch_nx): New.
1320 (opt_pass::gt_pch_nx_with_op): New.
1321 (pass_manager::gt_ggc_mx): New.
1322 (pass_manager::gt_pch_nx): New.
1323 (pass_manager::gt_pch_nx_with_op): New.
1324 (pass_manager::operator new): Use
1325 ggc_internal_cleared_alloc_stat rather than xcalloc.
1326 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1327 (pass_manager::gt_ggc_mx): New.
1328 (pass_manager::gt_pch_nx): New.
1329 (pass_manager::gt_pch_nx_with_op): New.
1330 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1331 (opt_pass::operator new): New.
1332 (opt_pass::gt_ggc_mx): New.
1333 (opt_pass::gt_pch_nx): New.
1334 (opt_pass::gt_pch_nx_with_op): New.
1336 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1338 * Makefile.in (GTFILES): Add context.h.
1339 * context.c (gcc::context::operator new): New.
1340 (gcc::context::gt_ggc_mx): New.
1341 (gcc::context::gt_pch_nx): New.
1342 (gcc::context::gt_pch_nx): New.
1343 * context.h (gcc::context): Add GTY((user)) marking.
1344 (gcc::context::operator new): New.
1345 (gcc::context::gt_ggc_mx): New.
1346 (gcc::context::gt_pch_nx): New.
1347 (gcc::context::gt_pch_nx): New.
1348 (g): Add GTY marking.
1349 (gt_ggc_mx (gcc::context *)): New.
1350 (gt_pch_nx (gcc::context *)): New.
1351 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1352 void *cookie)): New.
1353 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1355 2013-08-20 Alan Modra <amodra@gmail.com>
1358 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
1359 (rs6000_emit_epilogue): Likewise.
1361 2013-08-19 Dehao Chen <dehao@google.com>
1363 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
1365 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
1366 Jakub Jelinek <jakub@redhat.com>
1368 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
1369 (BUILT_IN_FABSD64): Likewise.
1370 (BUILT_IN_FABSD128): Likewise.
1371 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
1372 (fold_builtin_1): Likewise.
1373 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
1374 destination and source operands.
1375 (*abstd2_fpr): Likewise.
1376 (*nabstd2_fpr): Likewise.
1378 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
1380 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
1383 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
1385 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
1388 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
1390 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
1391 * builtins.c (default_libc_has_function): New.
1392 (gnu_libc_has_function): Ditto.
1393 (no_c99_libc_has_function): Ditto.
1394 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
1395 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
1396 (fold_builtin_sincos): Likewise.
1397 (fold_builtin_cexp): Likewise.
1398 * builtins.def (DEF_C94_BUILTIN): Likewise.
1399 (DEF_C99_BUILTIN): Likewise.
1400 (DEF_C99_C90RES_BUILTIN): Likewise.
1401 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
1402 definitions to using this define.
1403 * config/darwin-protos.h (darwin_libc_has_function): New.
1404 * config/darwin.c (darwin_libc_has_function): Ditto.
1405 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
1406 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
1407 * config/darwin.h: Ditto.
1408 * config/elfos.h: Ditto.
1409 * config/freebsd.h: Ditto.
1410 * config/i386/cygming.h: Ditto.
1411 * config/i386/djgpp.h: Ditto.
1412 * config/i386/i386-interix.h: Ditto.
1413 * config/microblaze/microblaze.h: Ditto.
1414 * config/mmix/mmix.h: Ditto.
1415 * config/gnu-user.h: Ditto.
1416 * config/ia64/hpux.h: Ditto.
1417 * config/pa/pa-hpux.h: Ditto.
1418 * config/pdp11/pdp11.h: Ditto.
1419 * config/picochip/picochip.h: Ditto.
1420 * config/linux.h: Ditto.
1421 * config/netbsd.h: Ditto.
1422 * config/openbsd.h: Ditto.
1423 * config/rs6000/aix43.h: Ditto.
1424 * config/rs6000/aix51.h: Ditto.
1425 * config/rs6000/aix52.h: Ditto.
1426 * config/rs6000/aix53.h: Ditto.
1427 * config/rs6000/aix61.h: Ditto.
1428 * config/rs6000/darwin.h: Ditto.
1429 * config/rs6000/linux.h: Ditto.
1430 * config/rs6000/linux64.h: Ditto.
1431 * config/s390/tpf.h: Ditto.
1432 * config/sol2-10.h: Ditto.
1433 * config/sol2.h: Ditto.
1434 * config/vms/vms.h: Ditto.
1435 * config/vxworks.h: Ditto.
1436 * config/linux-android.c (linux_android_libc_has_function):
1437 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
1438 * config/linux-protos.h (linux_android_libc_has_function):
1440 * config/i386/i386.c (ix86_libc_has_function): New.
1441 * config/i386/i386-protos.h
1442 (ix86_libc_has_function): New declaration.
1443 * config/i386/i386.md
1444 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
1445 ("isinf<mode>2): Likewise.
1446 * convert.c (convert_to_integer): Using new target hook
1447 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
1448 TARGET_C99_FUNCTIONS.
1449 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
1450 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
1451 * coretypes.h (function_class): New enum for different
1452 classes of functions.
1453 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
1454 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
1455 (TARGET_HAS_SINCOS): Likewise.
1456 (TARGET_LIBC_HAS_FUNCTION): New.
1457 * doc/tm.texi: Regenerated.
1458 * targhooks.h (default_libc_has_function): New declaration.
1459 (no_c99_libc_has_function): Ditto.
1460 (gnu_libc_has_function): Ditto.
1461 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
1462 and TARGET_HAS_SINCOS.
1464 2013-08-18 Jan Hubicka <jh@suse.cz>
1466 * Makeifle-in (ipa-devirt.o): New.
1467 (GTFILES): Add ipa-utils.h and ipa-devirt.c
1468 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
1469 (analyze_functions): Look into possible targets of polymorphic call.
1470 * dumpfile.c (dump_files): Add type-inheritance dump.
1471 * dumpfile.h (TDI_inheritance): New.
1472 * ipa-devirt.c: New file.
1473 * ipa-utils.h (odr_type_d): Forward declare.
1474 (odr_type): New type.
1475 (build_type_inheritance_graph): Declare.
1476 (possible_polymorphic_call_targets): Declare and introduce inline
1477 variant when only edge is pased.
1478 (dump_possible_polymorphic_call_targets): Likewise.
1479 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
1480 * tree.c (type_in_anonymous_namespace_p): Break out from ...
1481 (types_same_for_odr): ... here.
1482 * tree.h (type_in_anonymous_namespace_p): Declare.
1484 2013-08-18 Jakub Jelinek <jakub@redhat.com>
1486 PR tree-optimization/58006
1487 * tree-parloops.c (take_address_of): Don't ICE if get_name
1489 (eliminate_local_variables_stmt): Remove clobber stmts.
1491 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
1493 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
1494 error message is issued for an alias to undefined symbol.
1496 2013-08-18 Jan Hubicka <jh@suse.cz>
1498 * cgraph.c (cgraph_create_indirect_edge): Discover
1499 polymorphic calls and record basic info into indirect_info.
1500 * gimple-fold.c (gimple_fold_call): When doing BINFO based
1501 devirtualization, ignore objc function calls.
1502 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
1503 call with no parm index info.
1504 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
1505 * tree.c (virtual_method_call_p): New function.
1506 * tree.h (virtual_method_call_p): Declare.
1508 2013-08-16 Jan Hubicka <jh@suse.cz>
1511 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
1513 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
1515 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
1518 2013-08-16 David Malcolm <dmalcolm@redhat.com>
1520 * gengtype.c (type_for_name): Add special-case support for
1521 locating types within the "gcc::" namespace.
1522 (open_base_files): Emit a "using namespace gcc" directive.
1524 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1527 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
1528 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
1530 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
1531 array instead of each individual operand as a separate argument.
1532 (emit_fusion_gpr_load): Likewise.
1533 (expand_fusion_gpr_load): Add new function declaration.
1535 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
1536 signature to have the operands passed as an array, instead of as
1537 separate arguments. Allow ZERO_EXTEND to be in the memory
1538 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
1539 depend on the register live/dead flags when peepholes are run.
1540 (expand_fusion_gpr_load): New function to be called from the
1541 peephole2 pass, to change the register that addis sets to be the
1543 (emit_fusion_gpr_load): Change the calling signature to have the
1544 operands passed as an array, instead of as separate arguments.
1545 Allow ZERO_EXTEND to be in the memory address, and also
1546 SIGN_EXTEND if -mpower8-fusion-sign.
1548 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
1550 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
1551 adjust the register addis loads up in the peephole2 pass. Do not
1552 depend on the register live/dead state when the peephole pass is done.
1554 2013-08-16 David Malcolm <dmalcolm@redhat.com>
1556 * gengtype.c (create_user_defined_type): Ensure that the kind
1557 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
1558 declaration is seen before the GTY((user)) marking.
1560 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
1563 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
1565 2013-08-16 Jan Hubicka <jh@suse.cz>
1567 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
1568 arugment expected_type.
1569 (gimple_fold_call): Use it.
1570 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
1571 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
1572 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
1573 (try_make_edge_direct_virtual_call): Likewise.
1574 * tree.c (obj_type_ref_class): New.
1575 * tree.h (obj_type_ref_class): Use it.
1577 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
1579 * sched-vis.c (rtl_slim_pp_initialized): Remove.
1580 (rtl_slim_pp): Likewise.
1581 (init_rtl_slim_pretty_print): Likewise.
1582 (dump_value_slim): Don't call it. Use local pretty printer.
1583 (dump_insn_slim): Likewise.
1584 (dump_rtl_slim): Likewise.
1585 (str_pattern_slim): Likewise.
1586 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
1589 2013-08-16 Jakub Jelinek <jakub@redhat.com>
1591 PR tree-optimization/58164
1592 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
1593 walk gimple_goto_dest of GIMPLE_GOTO.
1595 PR tree-optimization/58165
1596 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
1597 bi_call must be the last stmt in a bb, don't split_block, instead
1598 use fallthru edge from it and give up if there is none.
1599 Release conds vector when returning early.
1601 2013-08-14 Xinliang David Li <davidxl@google.com>
1603 * config/i386/i386.c (ix86_option_override_internal):
1604 Remove unused variable and field.
1606 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1609 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
1610 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
1611 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
1612 and Linux, correct BLKmode alignment when 128-bit alignment is
1613 required and compatibility flag is not set.
1614 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
1615 for zero-size arguments when compatibility flag is not set.
1617 2013-08-14 Jakub Jelinek <jakub@redhat.com>
1619 PR tree-optimization/58145
1620 * tree-sra.c (build_ref_for_offset): If prev_base has
1621 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
1623 2013-08-14 Xinliang David Li <davidxl@google.com>
1625 * config/i386/i386.c (ix86_option_override_internal):
1626 Fix uninitialized variable error.
1628 2013-08-14 Xinliang David Li <davidxl@google.com>
1630 * config/i386/i386.opt: Define two new options.
1631 * config/i386/x86-tune.def: Add arch selector field in macros.
1632 * config/i386/i386.h: Adjust macro definition.
1633 * config/i386/i386.c (ix86_option_override_internal):
1635 (parse_mtune_ctrl_str): New function.
1636 (set_ix86_tune_features): New function.
1637 (ix86_function_specific_restore): Call the new helper function.
1639 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
1641 PR rtl-optimization/57662
1642 * sel-sched.c (code_motion_process_successors): When the current insn
1643 is removed after the recursive traversal, break from the loop.
1644 Add comments and debug printouts.
1646 2013-08-14 Jakub Jelinek <jakub@redhat.com>
1647 Alexandre Oliva <aoliva@redhat.com>
1650 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
1651 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
1652 there also UNSPEC_PLTOFF.
1654 2013-08-14 Marek Polacek <polacek@redhat.com>
1656 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
1657 NUM_CONDITIONS bit positions.
1659 2013-08-13 Cary Coutant <ccoutant@google.com>
1661 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
1662 (attr_checksum): Hash vector contents instead of pointer.
1663 (attr_checksum_ordered): Likewise.
1665 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
1667 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
1668 when Pmode != word_mode. Add length_address attribute.
1669 (sse3_monitor_<mode>): Merge from sse3_monitor and
1670 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
1671 Pmode != word_mode. Update insn length attribute.
1672 * config/i386/i386.c (ix86_option_override_internal): Update
1673 ix86_gen_monitor selection for merged sse3_monitor insn.
1675 2013-08-13 Julian Brown <julian@codesourcery.com>
1677 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
1678 perform invalid legitimization on greater-than-word-size modes for
1681 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
1683 * ira.c (setup_class_translate_array): Use aclass instead of cl
1684 for classes not fully covered by allocno classes.
1686 2013-08-13 Jakub Jelinek <jakub@redhat.com>
1688 PR tree-optimization/57661
1689 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
1690 * tree-inline.c (tree_function_versioning): Initialize it.
1691 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
1692 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
1693 that is not being copied.
1696 * asan.c (instrument_strlen_call): Fix typo in comment.
1697 Use char * type even for the lhs of POINTER_PLUS_EXPR.
1699 2013-08-13 Steve Ellcey <sellcey@mips.com>
1701 * config/mips/mips.md (prefetch): Use lw instead of ld on
1702 loongson in 32bit mode.
1704 2013-08-13 Nick Clifton <nickc@redhat.com>
1706 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
1708 2013-08-13 Jan Hubicka <jh@suse.cz>
1710 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
1711 introduced edge; fix typo in sanity check.
1712 (cgraph_resolve_speculation): Export; improve diagnostic.
1713 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
1714 speculation at type mismatch.
1715 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
1716 (cgraph_resolve_speculation): Declare.
1717 (symtab_can_be_discarded): New function.
1718 * value-prof.c (gimple_ic_transform): Remove actual transform code.
1719 * ipa-inline-transform.c (speculation_removed): New global var.
1720 (clone_inlined_nodes): See if speculation can be removed.
1721 (inline_call): If speculations was removed, we growths may not match.
1722 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
1723 (speculation_useful_p): New function.
1724 (resolve_noninline_speculation): New function.
1725 (inline_small_functions): Resolve useless speculations.
1726 * ipa-inline.h (speculation_useful_p): Declare
1727 * ipa.c (can_replace_by_local_alias): Simplify.
1728 (ipa_profile): Produce speculative calls in non-lto, too;
1729 add simple cost model; produce local aliases.
1731 2013-08-13 David Malcolm <dmalcolm@redhat.com>
1733 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
1736 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
1738 * config/i386/i386.c (ix86_function_versions): Use error + inform.
1740 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
1742 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
1743 iterator instead of X87MODEF.
1745 2013-08-12 Perez Read <netfirewall@gmail.com>
1748 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
1749 operand 0 for intel asm alternative.
1750 (*movabs<mode>_2): Ditto for operand 1.
1752 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
1754 * config/aarch64/arm_none.h
1755 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
1757 2013-08-12 Nick Clifton <nickc@redhat.com>
1759 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
1761 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1763 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
1764 expand for QI/HImode operand to produce more effictive code for
1765 unsigned char(short) --> float(double) conversion.
1767 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
1769 * doc/invoke.texi: Mention that -ftls-model does not force the final
1772 2013-08-12 Marek Polacek <polacek@redhat.com>
1773 Marc Glisse <marc.glisse@inria.fr>
1775 PR tree-optimization/57980
1776 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
1777 when creating -1 constant.
1779 2013-08-10 Jan Hubicka <jh@suse.cz>
1781 Workaround binutils PR14342.
1782 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
1783 (gimple_init_edge_profiler): Likewise.
1784 (gimple_gen_ic_func_profiler): Likewise.
1786 2013-08-09 Jan Hubicka <jh@suse.cz>
1788 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
1790 2013-08-09 Xinliang David Li <davidxl@google.com>
1792 * config/i386/stringop.def: New file.
1793 * config/i386/stringop.opt: New file.
1794 * config/i386/i386-opts.h: Include stringopt.def.
1795 * config/i386/i386.opt: Include stringopt.opt.
1796 * config/i386/i386.c (ix86_option_override_internal):
1797 Override default size based stringop inline strategies with options.
1798 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
1801 2013-08-09 Jan Hubicka <jh@suse.cz>
1803 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
1805 2013-08-09 Jan Hubicka <jh@suse.cz>
1807 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
1809 (dump_cgraph_node): Dump profile-id.
1810 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
1811 and common_target_probability.
1812 * lto-cgraph.c (lto_output_edge): Stream common targets.
1813 (lto_output_node): Stream profile ids.
1814 (input_node): Stream profile ids.
1815 (input_edge): Stream common targets.
1816 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
1817 * ipa.c: Include value-prof.h
1818 (ipa_profile_generate_summary): Turn indirect call statement histograms
1819 into common targets.
1820 (ipa_profile): Turn common targets into speculative edges.
1822 2013-08-09 Jan Hubicka <jh@suse.cz>
1824 * cgraph.h (cgraph_node): Add profile_id.
1825 * value-prof.c (cgraph_node_map): Turn into pointer_map.
1826 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
1827 (del_node_map): Update.
1828 (find_func_by_funcdef_no): Replace by ...
1829 (find_func_by_profile_id): ... this one.
1830 (gimple_ic_transform): Do not remove useful histograms when
1831 speculation is not done; dump info when indirect call removal
1833 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
1834 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
1835 (__gcov_indirect_call_profiler_v2): .. this one.
1836 * profile.h (init_node_map): Update.
1837 * coverage.c (coverage_compute_profile_id): New function.
1838 * coverage.h (coverage_compute_profile_id): Declare.
1839 * tree-profile.c (init_ic_make_global_vars): Make
1840 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
1841 (gimple_init_edge_profiler): Update prototype of
1842 __gcov_indirect_call_profiler.
1843 (gimple_gen_ic_func_profiler): Simplify.
1844 (tree_profiling): Use init_node_map
1846 2013-08-09 Jan Hubicka <jh@suse.cz>
1848 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
1849 non-speculative refs.
1850 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
1851 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
1852 (cgraph_set_call_stmt): Likewise.
1853 (cgraph_create_edge_1): Fix release checking compilatoin;
1855 (cgraph_free_edge): Free indirect info.
1856 (cgraph_turn_edge_to_speculative): New function.
1857 (cgraph_speculative_call_info): New function.
1858 (cgraph_make_edge_direct): Return direct edge; handle speculation.
1859 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
1860 (dump_cgraph_node): Dump speculation.
1861 (verify_edge_count_and_frequency): Accept speculative edges.
1862 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
1863 (verify_cgraph_node): Handle speculation.
1864 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
1865 (cgraph_set_call_stmt): Update prototype.
1866 (cgraph_make_edge_direct): Update prototype.
1867 (cgraph_speculative_call_info): Declare.
1868 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
1869 to change; update call of ipa_find_references.
1870 * ipa-ref.c (ipa_record_reference): Fix return value; clear
1871 lto_stmt_uid and speculative flags.
1872 (ipa_dump_references): Dump speculation.
1873 (ipa_clone_references): Clone speculative flag.
1874 (ipa_clone_referring): Likewise.
1875 (ipa_clone_ref): New function.
1876 (ipa_find_reference): Look into lto_stmt_uids
1877 (ipa_clear_stmts_in_references): Do not clear speculative calls.
1878 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
1879 (ipa_find_reference): Update declaration.
1880 (ipa_clone_ref): Declare.
1881 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
1882 stream speculative flag.
1883 (lto_output_ref): Stream statements uids and speculation.
1884 (input_ref): Likewise.
1885 (input_edge): Stream speuclation.
1886 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
1887 (cgraph_set_call_stmt_including_clones): Handle speculation.
1888 * ipa-inline.c (heap_edge_removal_hook): New function.
1889 (inline_small_functions): Register it.
1890 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
1891 also initialize refs.
1892 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
1894 (try_make_edge_direct_simple_call): Likewise.
1895 (try_make_edge_direct_simple_call): Likewise.
1896 (update_indirect_edges_after_inlining): Likewise.
1897 (remove_described_reference): Look proper lto_stmt_uid.
1898 (propagate_controlled_uses): Likewise.
1899 (propagate_controlled_uses): Liekwise.
1900 * tree-inline.c (copy_bb): Copy speculative edges.
1901 (redirect_all_calls): New function.
1902 (copy_cfg_body): Do redirection after loop info is updated.
1903 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1905 2013-08-09 Jan Hubicka <jh@suse.cz>
1907 * lto-streamer-out.c (output_function): Renumber PHIs.
1908 * lto-streamer-in.c (input_function): Likewise.
1910 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1912 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1913 (get_lane_unsigned): Likewise.
1914 (dup_lane_scalar): Likewise.
1915 (get_lane): enable for VALL.
1916 * config/aarch64/aarch64-simd.md
1917 (aarch64_dup_lane_scalar<mode>): Remove.
1918 (aarch64_get_lane_signed<mode>): Likewise.
1919 (aarch64_get_lane_unsigned<mode>): Likewise.
1920 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1921 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1922 (aarch64_get_lane<mode>): Enable for all vector modes.
1923 (aarch64_get_lanedi): Remove misleading constraints.
1924 * config/aarch64/arm_neon.h
1925 (__aarch64_vget_lane_any): Define.
1926 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1927 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1928 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1929 * config/aarch64/iterators.md (VDQQH): New.
1933 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1935 * configure.ac: Add GAS check for LEON instructions on SPARC.
1936 * configure: Regenerate.
1937 * config.in: Likewise.
1938 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1940 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1941 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1943 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1944 (AS_LEON_FLAG): New macro.
1945 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1946 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1947 Deal with LEON and LEON3 for the memory model.
1948 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1949 (atomic_compare_and_swap<mode>_1): Likewise.
1950 (*atomic_compare_and_swap<mode>_1): Likewise.
1952 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1954 * config/arm/neon.md (vcond): Fix floating-point vector
1955 comparisons against 0.
1957 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1959 * lra-constraints.c (emit_spill_move): Remove assert.
1960 (process_alt_operands): Add more debugging
1961 output. Increase reject for spilling into memory. Decrease
1962 reject for reloading scratch.
1963 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1965 2013-08-08 Steve Ellcey <sellcey@mips.com>
1967 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1968 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1969 micromips incompatible. Add nan2008.
1970 (MULTILIB_DIRNAMES): Add nan2008.
1971 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1972 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1973 and micromips incompatible. Add nan2008.
1974 (MULTILIB_DIRNAMES): Add nan2008.
1975 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1977 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
1979 PR rtl-optimization/58079
1980 * combine.c (combine_simplify_rtx): Avoid using SUBST if
1981 simplify_comparison has widened a comparison with an integer.
1983 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1985 * config/arm/neon.md (movmisalign<mode>): Disable when we
1986 don't allow unaligned accesses.
1987 (*movmisalign<mode>_neon_store): Likewise.
1988 (*movmisalign<mode>_neon_load): Likewise.
1989 (*movmisalign<mode>_neon_store): Likewise.
1990 (*movmisalign<mode>_neon_load): Likewise.
1992 2013-08-08 Jan Hubicka <jh@suse.cz>
1994 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
1995 (make_pass_rebuild_cgraph_edges): Also clear references.
1996 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
1997 * ipa-inline-transform.c (inline_transform): Remove all references
1999 * cgraphunit.c (expand_function): Remove all references after
2001 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
2002 (ipa_find_reference): Rewrite to iterator.
2003 (remove_stmt_references): Likewise.
2004 (ipa_clear_stmts_in_references): New function.
2005 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
2006 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
2008 * ipa-split.c (split_function): Remove references in split function.
2010 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2013 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
2014 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
2016 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2019 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
2020 suggested mode for the assignment isn't compatible with the
2023 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
2026 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
2028 2013-08-07 Xinliang David Li <davidxl@google.com>
2030 * config/i386/i386.opt: New option -mtune-ctrl=.
2031 * config/i386/x86-tune.def: New file.
2032 * config/i386/i386.h: include x86-tune.def.
2033 * config/i386/i386.c (ix86_option_override_internal):
2034 Parsing -mtune-ctrl= option and set tune features.
2036 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
2039 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
2040 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
2043 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2045 * rtl.h (update_alignments): Declare.
2046 * final.c (grow_label_align): New function extracted from...
2047 (shorten_branches): ...here. Call it.
2048 (update_alignments): New function.
2049 * reorg.c (sibling_labels): New variable.
2050 (get_label_before): Add SIBLING parameter. If it is non-zero, push
2051 the new label along with it onto the sibling_labels vector.
2052 (fill_simple_delay_slots): Adjust call to get_label_before.
2053 (fill_slots_from_thread): Likewise.
2054 (relax_delay_slots): Likewise.
2055 (make_return_insns): Likewise.
2056 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
2058 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2060 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
2061 document its semantics.
2062 (diagnostic_report_diagnostic): Adjust accordingly.
2064 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2066 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
2067 (sparc_option_override): ...and port to new C++ pass API.
2068 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
2070 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
2072 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
2074 2013-08-06 Caroline Tice <cmtice@google.com>
2076 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
2077 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
2078 * tree-pass.h: Add pass_vtable_verify.
2079 * varasm.c (assemble_variable): Add code to properly set the comdat
2080 section and name for the .vtable_map_vars section.
2081 (assemble_vtyv_preinit_initializer): New function.
2082 (default_sectin_type_flags): Make sure .vtable_map_vars section has
2084 * output.h: Add function decl for assemble_vtv_preinit_initializer.
2085 * vtable-verify.c: New file.
2086 * vtable-verify.h: New file.
2087 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
2088 initialiation levels.
2089 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
2090 * passes.def: Insert pass_vtable_verify.
2091 * aclocal.m4: Reorder includes.
2092 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
2093 -fvtv-counts options.
2094 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
2095 as appropriate, if -fvtable-verify=... is used.
2096 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
2097 -fvtable-verify=... is used.
2098 * Makefile.in (OBJS): Add vtable-verify.o to list.
2099 (vtable-verify.o): Add new build rule.
2100 (GTFILES): Add vtable-verify.c to list.
2101 * common.opt (fvtable-verify=): New flag.
2102 (vtv_priority): Values for fvtable-verify= flag.
2103 (fvtv-counts): New flag.
2104 (fvtv-debug): New flag.
2105 * tree.h (save_vtable_map_decl): New extern function decl.
2107 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2109 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
2110 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
2111 (pass_data_rl78_devirt): ...new pass_data instance and...
2112 (make_pass_rl78_devirt): ...new function.
2113 (rl78_asm_file_start): Port pass registration to new C++ API.
2115 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2117 * coretypes.h (rtl_opt_pass): Add.
2118 (gcc::context): Add.
2119 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
2120 (epiphany_init): Port to new C++ pass API.
2121 (epiphany_optimize_mode_switching): Likewise.
2122 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
2123 (pass_manager::get_pass_mode_switching): New.
2124 (pass_manager::get_pass_peephole2): New.
2125 * mode-switching.c (pass_mode_switching): Add clone method.
2126 * recog.c (pass_peephole2): Add clone method.
2127 (pass_split_all_insns): Add clone method.
2129 2013-08-06 David Malcolm <dmalcolm@redhat.com>
2131 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
2132 (mips_option_override): ...here, porting to new C++ API for passes.
2134 2013-08-06 Jan Hubicka <jh@suse.cz>
2136 * cgraph.c (cgraph_get_body): New function based on lto.c
2138 * cgraph.h (cgraph_get_body): Declare.
2139 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
2141 * cgraphunit.c (expand_function): Get body prior expanding.
2142 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
2143 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
2145 * passes.c (do_per_function_toporder): Get body.
2146 * tree-inline.c (expand_call_inline): Get body prior inlining it.
2147 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
2149 2013-08-06 Martin Jambor <mjambor@suse.cz>
2152 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
2153 is not re-finalized. Rename second parameter to no_collect.
2155 2013-08-06 Martin Jambor <mjambor@suse.cz>
2158 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
2159 MEM_REF has proper alignment information.
2161 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
2164 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
2166 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
2167 rtx (*) (rtx, ...) with insn_gen_fn.
2168 * genoutput.c (output_insn_data): Cast gen_? function pointers to
2169 insn_gen_fn::stored_funcptr. Add initializer braces.
2171 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2173 Rewrite how instances of passes are cloned to remove assumptions
2174 about their sizes (thus allowing pass subclasses to have
2175 additional data fields, albeit non-GC-managed ones at this point).
2177 * passes.c (make_pass_instance): Now that passes have clone
2178 methods, rewrite this function to eliminate XNEW and memcpy
2179 calls that used hardcoded sizes. Since this function no longer
2180 creates pass instances, rename it to...
2181 (add_pass_instance): ...this. Document the old way that passes were
2182 numbered and flagged, and rework this function to continue using it.
2183 (next_pass_1): Add an initial_pass argument for use by
2185 (position_pass): When adding multiple instances of a pass, use
2186 the pass's clone method, rather than relying on the XNEW/memcpy
2187 within the former make_pass_instance (now add_pass_instance).
2188 (pass_manager::pass_manager): When invoking next_pass_1, also supply
2189 the initial instance of the current pass within the pass manager.
2191 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2193 This is the automated part of the conversion of passes from C
2194 structs to C++ classes.
2196 Patch autogenerated by refactor_passes.py from
2197 https://github.com/davidmalcolm/gcc-refactoring-scripts
2198 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
2200 * asan.c (pass_asan): Convert from a global struct to a subclass of
2201 gimple_opt_pass along with...
2202 (pass_data_asan): ...new pass_data instance and...
2203 (make_pass_asan): ...new function.
2204 (pass_asan_O0): Convert from a global struct to a subclass of
2205 gimple_opt_pass along with...
2206 (pass_data_asan_O0): ...new pass_data instance and...
2207 (make_pass_asan_O0): ...new function.
2208 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
2209 subclass of rtl_opt_pass along with...
2210 (pass_data_inc_dec): ...new pass_data instance and...
2211 (make_pass_inc_dec): ...new function.
2212 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
2213 a subclass of rtl_opt_pass along with...
2214 (pass_data_reorder_blocks): ...new pass_data instance and...
2215 (make_pass_reorder_blocks): ...new function.
2216 (pass_duplicate_computed_gotos): Convert from a global struct to a
2217 subclass of rtl_opt_pass along with...
2218 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
2219 (make_pass_duplicate_computed_gotos): ...new function.
2220 (pass_partition_blocks): Convert from a global struct to a subclass of
2221 rtl_opt_pass along with...
2222 (pass_data_partition_blocks): ...new pass_data instance and...
2223 (make_pass_partition_blocks): ...new function.
2224 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
2225 struct to a subclass of rtl_opt_pass along with...
2226 (pass_data_branch_target_load_optimize1): ...new pass_data instance
2228 (make_pass_branch_target_load_optimize1): ...new function.
2229 (pass_branch_target_load_optimize2): Convert from a global struct to a
2230 subclass of rtl_opt_pass along with...
2231 (pass_data_branch_target_load_optimize2): ...new pass_data instance
2233 (make_pass_branch_target_load_optimize2): ...new function.
2234 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
2235 of rtl_opt_pass along with...
2236 (pass_data_jump): ...new pass_data instance and...
2237 (make_pass_jump): ...new function.
2238 (pass_jump2): Convert from a global struct to a subclass of
2239 rtl_opt_pass along with...
2240 (pass_data_jump2): ...new pass_data instance and...
2241 (make_pass_jump2): ...new function.
2242 * cfgexpand.c (pass_expand): Convert from a global struct to a
2243 subclass of rtl_opt_pass along with...
2244 (pass_data_expand): ...new pass_data instance and...
2245 (make_pass_expand): ...new function.
2246 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
2247 of rtl_opt_pass along with...
2248 (pass_data_free_cfg): ...new pass_data instance and...
2249 (make_pass_free_cfg): ...new function.
2250 (pass_into_cfg_layout_mode): Convert from a global struct to a
2251 subclass of rtl_opt_pass along with...
2252 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
2253 (make_pass_into_cfg_layout_mode): ...new function.
2254 (pass_outof_cfg_layout_mode): Convert from a global struct to a
2255 subclass of rtl_opt_pass along with...
2256 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
2257 (make_pass_outof_cfg_layout_mode): ...new function.
2258 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
2259 struct to a subclass of gimple_opt_pass along with...
2260 (pass_data_build_cgraph_edges): ...new pass_data instance and...
2261 (make_pass_build_cgraph_edges): ...new function.
2262 (pass_rebuild_cgraph_edges): Convert from a global struct to a
2263 subclass of gimple_opt_pass along with...
2264 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
2265 (make_pass_rebuild_cgraph_edges): ...new function.
2266 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
2267 subclass of gimple_opt_pass along with...
2268 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
2270 (make_pass_remove_cgraph_callee_edges): ...new function.
2271 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
2272 struct to a subclass of rtl_opt_pass along with...
2273 (pass_data_stack_adjustments): ...new pass_data instance and...
2274 (make_pass_stack_adjustments): ...new function.
2275 * combine.c (pass_combine): Convert from a global struct to a subclass
2276 of rtl_opt_pass along with...
2277 (pass_data_combine): ...new pass_data instance and...
2278 (make_pass_combine): ...new function.
2279 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
2280 global struct to a subclass of rtl_opt_pass along with...
2281 (pass_data_compare_elim_after_reload): ...new pass_data instance
2283 (make_pass_compare_elim_after_reload): ...new function.
2284 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
2285 of rtl_opt_pass along with...
2286 (pass_data_rtl_cprop): ...new pass_data instance and...
2287 (make_pass_rtl_cprop): ...new function.
2288 * cse.c (pass_cse): Convert from a global struct to a subclass of
2289 rtl_opt_pass along with...
2290 (pass_data_cse): ...new pass_data instance and...
2291 (make_pass_cse): ...new function.
2292 (pass_cse2): Convert from a global struct to a subclass of
2293 rtl_opt_pass along with...
2294 (pass_data_cse2): ...new pass_data instance and...
2295 (make_pass_cse2): ...new function.
2296 (pass_cse_after_global_opts): Convert from a global struct to a
2297 subclass of rtl_opt_pass along with...
2298 (pass_data_cse_after_global_opts): ...new pass_data instance and...
2299 (make_pass_cse_after_global_opts): ...new function.
2300 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
2301 of rtl_opt_pass along with...
2302 (pass_data_ud_rtl_dce): ...new pass_data instance and...
2303 (make_pass_ud_rtl_dce): ...new function.
2304 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
2305 rtl_opt_pass along with...
2306 (pass_data_fast_rtl_dce): ...new pass_data instance and...
2307 (make_pass_fast_rtl_dce): ...new function.
2308 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
2309 a subclass of rtl_opt_pass along with...
2310 (pass_data_df_initialize_opt): ...new pass_data instance and...
2311 (make_pass_df_initialize_opt): ...new function.
2312 (pass_df_initialize_no_opt): Convert from a global struct to a
2313 subclass of rtl_opt_pass along with...
2314 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
2315 (make_pass_df_initialize_no_opt): ...new function.
2316 (pass_df_finish): Convert from a global struct to a subclass of
2317 rtl_opt_pass along with...
2318 (pass_data_df_finish): ...new pass_data instance and...
2319 (make_pass_df_finish): ...new function.
2320 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
2321 rtl_opt_pass along with...
2322 (pass_data_rtl_dse1): ...new pass_data instance and...
2323 (make_pass_rtl_dse1): ...new function.
2324 (pass_rtl_dse2): Convert from a global struct to a subclass of
2325 rtl_opt_pass along with...
2326 (pass_data_rtl_dse2): ...new pass_data instance and...
2327 (make_pass_rtl_dse2): ...new function.
2328 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
2329 subclass of rtl_opt_pass along with...
2330 (pass_data_dwarf2_frame): ...new pass_data instance and...
2331 (make_pass_dwarf2_frame): ...new function.
2332 * except.c (pass_set_nothrow_function_flags): Convert from a global
2333 struct to a subclass of rtl_opt_pass along with...
2334 (pass_data_set_nothrow_function_flags): ...new pass_data instance
2336 (make_pass_set_nothrow_function_flags): ...new function.
2337 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
2338 subclass of rtl_opt_pass along with...
2339 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
2341 (make_pass_convert_to_eh_region_ranges): ...new function.
2342 * final.c (pass_compute_alignments): Convert from a global struct to a
2343 subclass of rtl_opt_pass along with...
2344 (pass_data_compute_alignments): ...new pass_data instance and...
2345 (make_pass_compute_alignments): ...new function.
2346 (pass_final): Convert from a global struct to a subclass of
2347 rtl_opt_pass along with...
2348 (pass_data_final): ...new pass_data instance and...
2349 (make_pass_final): ...new function.
2350 (pass_shorten_branches): Convert from a global struct to a subclass of
2351 rtl_opt_pass along with...
2352 (pass_data_shorten_branches): ...new pass_data instance and...
2353 (make_pass_shorten_branches): ...new function.
2354 (pass_clean_state): Convert from a global struct to a subclass of
2355 rtl_opt_pass along with...
2356 (pass_data_clean_state): ...new pass_data instance and...
2357 (make_pass_clean_state): ...new function.
2358 * function.c (pass_instantiate_virtual_regs): Convert from a global
2359 struct to a subclass of rtl_opt_pass along with...
2360 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
2361 (make_pass_instantiate_virtual_regs): ...new function.
2362 (pass_leaf_regs): Convert from a global struct to a subclass of
2363 rtl_opt_pass along with...
2364 (pass_data_leaf_regs): ...new pass_data instance and...
2365 (make_pass_leaf_regs): ...new function.
2366 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
2367 subclass of rtl_opt_pass along with...
2368 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
2370 (make_pass_thread_prologue_and_epilogue): ...new function.
2371 (pass_match_asm_constraints): Convert from a global struct to a
2372 subclass of rtl_opt_pass along with...
2373 (pass_data_match_asm_constraints): ...new pass_data instance and...
2374 (make_pass_match_asm_constraints): ...new function.
2375 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
2376 subclass of rtl_opt_pass along with...
2377 (pass_data_rtl_fwprop): ...new pass_data instance and...
2378 (make_pass_rtl_fwprop): ...new function.
2379 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
2380 rtl_opt_pass along with...
2381 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
2382 (make_pass_rtl_fwprop_addr): ...new function.
2383 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
2384 rtl_opt_pass along with...
2385 (pass_data_rtl_pre): ...new pass_data instance and...
2386 (make_pass_rtl_pre): ...new function.
2387 (pass_rtl_hoist): Convert from a global struct to a subclass of
2388 rtl_opt_pass along with...
2389 (pass_data_rtl_hoist): ...new pass_data instance and...
2390 (make_pass_rtl_hoist): ...new function.
2391 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
2392 subclass of gimple_opt_pass along with...
2393 (pass_data_lower_cf): ...new pass_data instance and...
2394 (make_pass_lower_cf): ...new function.
2395 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
2396 from a global struct to a subclass of gimple_opt_pass along with...
2397 (pass_data_strength_reduction): ...new pass_data instance and...
2398 (make_pass_strength_reduction): ...new function.
2399 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
2400 of rtl_opt_pass along with...
2401 (pass_data_rtl_ifcvt): ...new pass_data instance and...
2402 (make_pass_rtl_ifcvt): ...new function.
2403 (pass_if_after_combine): Convert from a global struct to a subclass of
2404 rtl_opt_pass along with...
2405 (pass_data_if_after_combine): ...new pass_data instance and...
2406 (make_pass_if_after_combine): ...new function.
2407 (pass_if_after_reload): Convert from a global struct to a subclass of
2408 rtl_opt_pass along with...
2409 (pass_data_if_after_reload): ...new pass_data instance and...
2410 (make_pass_if_after_reload): ...new function.
2411 * init-regs.c (pass_initialize_regs): Convert from a global struct to
2412 a subclass of rtl_opt_pass along with...
2413 (pass_data_initialize_regs): ...new pass_data instance and...
2414 (make_pass_initialize_regs): ...new function.
2415 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
2416 of ipa_opt_pass_d along with...
2417 (pass_data_ipa_cp): ...new pass_data instance and...
2418 (make_pass_ipa_cp): ...new function.
2419 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
2420 global struct to a subclass of gimple_opt_pass along with...
2421 (pass_data_inline_parameters): ...new pass_data instance and...
2422 (make_pass_inline_parameters): ...new function.
2423 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
2424 subclass of gimple_opt_pass along with...
2425 (pass_data_early_inline): ...new pass_data instance and...
2426 (make_pass_early_inline): ...new function.
2427 (pass_ipa_inline): Convert from a global struct to a subclass of
2428 ipa_opt_pass_d along with...
2429 (pass_data_ipa_inline): ...new pass_data instance and...
2430 (make_pass_ipa_inline): ...new function.
2431 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
2432 struct to a subclass of gimple_opt_pass along with...
2433 (pass_data_local_pure_const): ...new pass_data instance and...
2434 (make_pass_local_pure_const): ...new function.
2435 (pass_ipa_pure_const): Convert from a global struct to a subclass of
2436 ipa_opt_pass_d along with...
2437 (pass_data_ipa_pure_const): ...new pass_data instance and...
2438 (make_pass_ipa_pure_const): ...new function.
2439 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
2440 to a subclass of ipa_opt_pass_d along with...
2441 (pass_data_ipa_reference): ...new pass_data instance and...
2442 (make_pass_ipa_reference): ...new function.
2443 * ipa-split.c (pass_split_functions): Convert from a global struct to
2444 a subclass of gimple_opt_pass along with...
2445 (pass_data_split_functions): ...new pass_data instance and...
2446 (make_pass_split_functions): ...new function.
2447 (pass_feedback_split_functions): Convert from a global struct to a
2448 subclass of gimple_opt_pass along with...
2449 (pass_data_feedback_split_functions): ...new pass_data instance and...
2450 (make_pass_feedback_split_functions): ...new function.
2451 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
2452 global struct to a subclass of simple_ipa_opt_pass along with...
2453 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
2455 (make_pass_ipa_function_and_variable_visibility): ...new function.
2456 (pass_ipa_free_inline_summary): Convert from a global struct to a
2457 subclass of simple_ipa_opt_pass along with...
2458 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
2459 (make_pass_ipa_free_inline_summary): ...new function.
2460 (pass_ipa_whole_program_visibility): Convert from a global struct to a
2461 subclass of ipa_opt_pass_d along with...
2462 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
2464 (make_pass_ipa_whole_program_visibility): ...new function.
2465 (pass_ipa_profile): Convert from a global struct to a subclass of
2466 ipa_opt_pass_d along with...
2467 (pass_data_ipa_profile): ...new pass_data instance and...
2468 (make_pass_ipa_profile): ...new function.
2469 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
2470 ipa_opt_pass_d along with...
2471 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
2472 (make_pass_ipa_cdtor_merge): ...new function.
2473 * ira.c (pass_ira): Convert from a global struct to a subclass of
2474 rtl_opt_pass along with...
2475 (pass_data_ira): ...new pass_data instance and...
2476 (make_pass_ira): ...new function.
2477 (pass_reload): Convert from a global struct to a subclass of
2478 rtl_opt_pass along with...
2479 (pass_data_reload): ...new pass_data instance and...
2480 (make_pass_reload): ...new function.
2481 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
2482 subclass of rtl_opt_pass along with...
2483 (pass_data_cleanup_barriers): ...new pass_data instance and...
2484 (make_pass_cleanup_barriers): ...new function.
2485 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
2486 of rtl_opt_pass along with...
2487 (pass_data_loop2): ...new pass_data instance and...
2488 (make_pass_loop2): ...new function.
2489 (pass_rtl_loop_init): Convert from a global struct to a subclass of
2490 rtl_opt_pass along with...
2491 (pass_data_rtl_loop_init): ...new pass_data instance and...
2492 (make_pass_rtl_loop_init): ...new function.
2493 (pass_rtl_loop_done): Convert from a global struct to a subclass of
2494 rtl_opt_pass along with...
2495 (pass_data_rtl_loop_done): ...new pass_data instance and...
2496 (make_pass_rtl_loop_done): ...new function.
2497 (pass_rtl_move_loop_invariants): Convert from a global struct to a
2498 subclass of rtl_opt_pass along with...
2499 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
2500 (make_pass_rtl_move_loop_invariants): ...new function.
2501 (pass_rtl_unswitch): Convert from a global struct to a subclass of
2502 rtl_opt_pass along with...
2503 (pass_data_rtl_unswitch): ...new pass_data instance and...
2504 (make_pass_rtl_unswitch): ...new function.
2505 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
2506 subclass of rtl_opt_pass along with...
2507 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
2509 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2510 (pass_rtl_doloop): Convert from a global struct to a subclass of
2511 rtl_opt_pass along with...
2512 (pass_data_rtl_doloop): ...new pass_data instance and...
2513 (make_pass_rtl_doloop): ...new function.
2514 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
2515 a subclass of rtl_opt_pass along with...
2516 (pass_data_lower_subreg): ...new pass_data instance and...
2517 (make_pass_lower_subreg): ...new function.
2518 (pass_lower_subreg2): Convert from a global struct to a subclass of
2519 rtl_opt_pass along with...
2520 (pass_data_lower_subreg2): ...new pass_data instance and...
2521 (make_pass_lower_subreg2): ...new function.
2522 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
2523 struct to a subclass of ipa_opt_pass_d along with...
2524 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
2525 (make_pass_ipa_lto_gimple_out): ...new function.
2526 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
2527 of ipa_opt_pass_d along with...
2528 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
2529 (make_pass_ipa_lto_finish_out): ...new function.
2530 * mode-switching.c (pass_mode_switching): Convert from a global struct
2531 to a subclass of rtl_opt_pass along with...
2532 (pass_data_mode_switching): ...new pass_data instance and...
2533 (make_pass_mode_switching): ...new function.
2534 * modulo-sched.c (pass_sms): Convert from a global struct to a
2535 subclass of rtl_opt_pass along with...
2536 (pass_data_sms): ...new pass_data instance and...
2537 (make_pass_sms): ...new function.
2538 * omp-low.c (pass_expand_omp): Convert from a global struct to a
2539 subclass of gimple_opt_pass along with...
2540 (pass_data_expand_omp): ...new pass_data instance and...
2541 (make_pass_expand_omp): ...new function.
2542 (pass_lower_omp): Convert from a global struct to a subclass of
2543 gimple_opt_pass along with...
2544 (pass_data_lower_omp): ...new pass_data instance and...
2545 (make_pass_lower_omp): ...new function.
2546 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
2547 of gimple_opt_pass along with...
2548 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
2549 (make_pass_diagnose_omp_blocks): ...new function.
2550 * passes.c (pass_early_local_passes): Convert from a global struct to
2551 a subclass of simple_ipa_opt_pass along with...
2552 (pass_data_early_local_passes): ...new pass_data instance and...
2553 (make_pass_early_local_passes): ...new function.
2554 (pass_all_early_optimizations): Convert from a global struct to a
2555 subclass of gimple_opt_pass along with...
2556 (pass_data_all_early_optimizations): ...new pass_data instance and...
2557 (make_pass_all_early_optimizations): ...new function.
2558 (pass_all_optimizations): Convert from a global struct to a subclass
2559 of gimple_opt_pass along with...
2560 (pass_data_all_optimizations): ...new pass_data instance and...
2561 (make_pass_all_optimizations): ...new function.
2562 (pass_all_optimizations_g): Convert from a global struct to a subclass
2563 of gimple_opt_pass along with...
2564 (pass_data_all_optimizations_g): ...new pass_data instance and...
2565 (make_pass_all_optimizations_g): ...new function.
2566 (pass_rest_of_compilation): Convert from a global struct to a subclass
2567 of rtl_opt_pass along with...
2568 (pass_data_rest_of_compilation): ...new pass_data instance and...
2569 (make_pass_rest_of_compilation): ...new function.
2570 (pass_postreload): Convert from a global struct to a subclass of
2571 rtl_opt_pass along with...
2572 (pass_data_postreload): ...new pass_data instance and...
2573 (make_pass_postreload): ...new function.
2574 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
2575 subclass of rtl_opt_pass along with...
2576 (pass_data_gcse2): ...new pass_data instance and...
2577 (make_pass_gcse2): ...new function.
2578 * postreload.c (pass_postreload_cse): Convert from a global struct to
2579 a subclass of rtl_opt_pass along with...
2580 (pass_data_postreload_cse): ...new pass_data instance and...
2581 (make_pass_postreload_cse): ...new function.
2582 * predict.c (pass_profile): Convert from a global struct to a subclass
2583 of gimple_opt_pass along with...
2584 (pass_data_profile): ...new pass_data instance and...
2585 (make_pass_profile): ...new function.
2586 (pass_strip_predict_hints): Convert from a global struct to a subclass
2587 of gimple_opt_pass along with...
2588 (pass_data_strip_predict_hints): ...new pass_data instance and...
2589 (make_pass_strip_predict_hints): ...new function.
2590 * recog.c (pass_peephole2): Convert from a global struct to a subclass
2591 of rtl_opt_pass along with...
2592 (pass_data_peephole2): ...new pass_data instance and...
2593 (make_pass_peephole2): ...new function.
2594 (pass_split_all_insns): Convert from a global struct to a subclass of
2595 rtl_opt_pass along with...
2596 (pass_data_split_all_insns): ...new pass_data instance and...
2597 (make_pass_split_all_insns): ...new function.
2598 (pass_split_after_reload): Convert from a global struct to a subclass
2599 of rtl_opt_pass along with...
2600 (pass_data_split_after_reload): ...new pass_data instance and...
2601 (make_pass_split_after_reload): ...new function.
2602 (pass_split_before_regstack): Convert from a global struct to a
2603 subclass of rtl_opt_pass along with...
2604 (pass_data_split_before_regstack): ...new pass_data instance and...
2605 (make_pass_split_before_regstack): ...new function.
2606 (pass_split_before_sched2): Convert from a global struct to a subclass
2607 of rtl_opt_pass along with...
2608 (pass_data_split_before_sched2): ...new pass_data instance and...
2609 (make_pass_split_before_sched2): ...new function.
2610 (pass_split_for_shorten_branches): Convert from a global struct to a
2611 subclass of rtl_opt_pass along with...
2612 (pass_data_split_for_shorten_branches): ...new pass_data instance
2614 (make_pass_split_for_shorten_branches): ...new function.
2615 * ree.c (pass_ree): Convert from a global struct to a subclass of
2616 rtl_opt_pass along with...
2617 (pass_data_ree): ...new pass_data instance and...
2618 (make_pass_ree): ...new function.
2619 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
2620 subclass of rtl_opt_pass along with...
2621 (pass_data_stack_regs): ...new pass_data instance and...
2622 (make_pass_stack_regs): ...new function.
2623 (pass_stack_regs_run): Convert from a global struct to a subclass of
2624 rtl_opt_pass along with...
2625 (pass_data_stack_regs_run): ...new pass_data instance and...
2626 (make_pass_stack_regs_run): ...new function.
2627 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
2628 subclass of rtl_opt_pass along with...
2629 (pass_data_cprop_hardreg): ...new pass_data instance and...
2630 (make_pass_cprop_hardreg): ...new function.
2631 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
2632 subclass of rtl_opt_pass along with...
2633 (pass_data_reginfo_init): ...new pass_data instance and...
2634 (make_pass_reginfo_init): ...new function.
2635 * regmove.c (pass_regmove): Convert from a global struct to a subclass
2636 of rtl_opt_pass along with...
2637 (pass_data_regmove): ...new pass_data instance and...
2638 (make_pass_regmove): ...new function.
2639 * regrename.c (pass_regrename): Convert from a global struct to a
2640 subclass of rtl_opt_pass along with...
2641 (pass_data_regrename): ...new pass_data instance and...
2642 (make_pass_regrename): ...new function.
2643 * reorg.c (pass_delay_slots): Convert from a global struct to a
2644 subclass of rtl_opt_pass along with...
2645 (pass_data_delay_slots): ...new pass_data instance and...
2646 (make_pass_delay_slots): ...new function.
2647 (pass_machine_reorg): Convert from a global struct to a subclass of
2648 rtl_opt_pass along with...
2649 (pass_data_machine_reorg): ...new pass_data instance and...
2650 (make_pass_machine_reorg): ...new function.
2651 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
2652 of rtl_opt_pass along with...
2653 (pass_data_sched): ...new pass_data instance and...
2654 (make_pass_sched): ...new function.
2655 (pass_sched2): Convert from a global struct to a subclass of
2656 rtl_opt_pass along with...
2657 (pass_data_sched2): ...new pass_data instance and...
2658 (make_pass_sched2): ...new function.
2659 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
2660 to a subclass of rtl_opt_pass along with...
2661 (pass_data_stack_ptr_mod): ...new pass_data instance and...
2662 (make_pass_stack_ptr_mod): ...new function.
2663 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
2664 to a subclass of rtl_opt_pass along with...
2665 (pass_data_rtl_store_motion): ...new pass_data instance and...
2666 (make_pass_rtl_store_motion): ...new function.
2667 * tracer.c (pass_tracer): Convert from a global struct to a subclass
2668 of gimple_opt_pass along with...
2669 (pass_data_tracer): ...new pass_data instance and...
2670 (make_pass_tracer): ...new function.
2671 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
2672 to a subclass of gimple_opt_pass along with...
2673 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
2674 (make_pass_diagnose_tm_blocks): ...new function.
2675 (pass_lower_tm): Convert from a global struct to a subclass of
2676 gimple_opt_pass along with...
2677 (pass_data_lower_tm): ...new pass_data instance and...
2678 (make_pass_lower_tm): ...new function.
2679 (pass_tm_init): Convert from a global struct to a subclass of
2680 gimple_opt_pass along with...
2681 (pass_data_tm_init): ...new pass_data instance and...
2682 (make_pass_tm_init): ...new function.
2683 (pass_tm_mark): Convert from a global struct to a subclass of
2684 gimple_opt_pass along with...
2685 (pass_data_tm_mark): ...new pass_data instance and...
2686 (make_pass_tm_mark): ...new function.
2687 (pass_tm_edges): Convert from a global struct to a subclass of
2688 gimple_opt_pass along with...
2689 (pass_data_tm_edges): ...new pass_data instance and...
2690 (make_pass_tm_edges): ...new function.
2691 (pass_tm_memopt): Convert from a global struct to a subclass of
2692 gimple_opt_pass along with...
2693 (pass_data_tm_memopt): ...new pass_data instance and...
2694 (make_pass_tm_memopt): ...new function.
2695 (pass_ipa_tm): Convert from a global struct to a subclass of
2696 simple_ipa_opt_pass along with...
2697 (pass_data_ipa_tm): ...new pass_data instance and...
2698 (make_pass_ipa_tm): ...new function.
2699 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
2700 subclass of gimple_opt_pass along with...
2701 (pass_data_call_cdce): ...new pass_data instance and...
2702 (make_pass_call_cdce): ...new function.
2703 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
2704 subclass of gimple_opt_pass along with...
2705 (pass_data_build_cfg): ...new pass_data instance and...
2706 (make_pass_build_cfg): ...new function.
2707 (pass_split_crit_edges): Convert from a global struct to a subclass of
2708 gimple_opt_pass along with...
2709 (pass_data_split_crit_edges): ...new pass_data instance and...
2710 (make_pass_split_crit_edges): ...new function.
2711 (pass_warn_function_return): Convert from a global struct to a
2712 subclass of gimple_opt_pass along with...
2713 (pass_data_warn_function_return): ...new pass_data instance and...
2714 (make_pass_warn_function_return): ...new function.
2715 (pass_warn_function_noreturn): Convert from a global struct to a
2716 subclass of gimple_opt_pass along with...
2717 (pass_data_warn_function_noreturn): ...new pass_data instance and...
2718 (make_pass_warn_function_noreturn): ...new function.
2719 (pass_warn_unused_result): Convert from a global struct to a subclass
2720 of gimple_opt_pass along with...
2721 (pass_data_warn_unused_result): ...new pass_data instance and...
2722 (make_pass_warn_unused_result): ...new function.
2723 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
2724 a subclass of gimple_opt_pass along with...
2725 (pass_data_merge_phi): ...new pass_data instance and...
2726 (make_pass_merge_phi): ...new function.
2727 * tree-complex.c (pass_lower_complex): Convert from a global struct to
2728 a subclass of gimple_opt_pass along with...
2729 (pass_data_lower_complex): ...new pass_data instance and...
2730 (make_pass_lower_complex): ...new function.
2731 (pass_lower_complex_O0): Convert from a global struct to a subclass of
2732 gimple_opt_pass along with...
2733 (pass_data_lower_complex_O0): ...new pass_data instance and...
2734 (make_pass_lower_complex_O0): ...new function.
2735 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
2736 subclass of gimple_opt_pass along with...
2737 (pass_data_lower_eh): ...new pass_data instance and...
2738 (make_pass_lower_eh): ...new function.
2739 (pass_refactor_eh): Convert from a global struct to a subclass of
2740 gimple_opt_pass along with...
2741 (pass_data_refactor_eh): ...new pass_data instance and...
2742 (make_pass_refactor_eh): ...new function.
2743 (pass_lower_resx): Convert from a global struct to a subclass of
2744 gimple_opt_pass along with...
2745 (pass_data_lower_resx): ...new pass_data instance and...
2746 (make_pass_lower_resx): ...new function.
2747 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
2748 of gimple_opt_pass along with...
2749 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
2750 (make_pass_lower_eh_dispatch): ...new function.
2751 (pass_cleanup_eh): Convert from a global struct to a subclass of
2752 gimple_opt_pass along with...
2753 (pass_data_cleanup_eh): ...new pass_data instance and...
2754 (make_pass_cleanup_eh): ...new function.
2755 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
2756 to a subclass of simple_ipa_opt_pass along with...
2757 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
2758 (make_pass_ipa_lower_emutls): ...new function.
2759 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
2760 a subclass of gimple_opt_pass along with...
2761 (pass_data_if_conversion): ...new pass_data instance and...
2762 (make_pass_if_conversion): ...new function.
2763 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
2764 subclass of gimple_opt_pass along with...
2765 (pass_data_build_ssa): ...new pass_data instance and...
2766 (make_pass_build_ssa): ...new function.
2767 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
2768 global struct to a subclass of gimple_opt_pass along with...
2769 (pass_data_loop_distribution): ...new pass_data instance and...
2770 (make_pass_loop_distribution): ...new function.
2771 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
2772 subclass of gimple_opt_pass along with...
2773 (pass_data_mudflap_1): ...new pass_data instance and...
2774 (make_pass_mudflap_1): ...new function.
2775 (pass_mudflap_2): Convert from a global struct to a subclass of
2776 gimple_opt_pass along with...
2777 (pass_data_mudflap_2): ...new pass_data instance and...
2778 (make_pass_mudflap_2): ...new function.
2779 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
2780 subclass of gimple_opt_pass along with...
2781 (pass_data_mudflap_1): ...new pass_data instance and...
2782 (make_pass_mudflap_1): ...new function.
2783 (pass_mudflap_2): Convert from a global struct to a subclass of
2784 gimple_opt_pass along with...
2785 (pass_data_mudflap_2): ...new pass_data instance and...
2786 (make_pass_mudflap_2): ...new function.
2787 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
2788 gimple_opt_pass along with...
2789 (pass_data_nrv): ...new pass_data instance and...
2790 (make_pass_nrv): ...new function.
2791 (pass_return_slot): Convert from a global struct to a subclass of
2792 gimple_opt_pass along with...
2793 (pass_data_return_slot): ...new pass_data instance and...
2794 (make_pass_return_slot): ...new function.
2795 * tree-object-size.c (pass_object_sizes): Convert from a global struct
2796 to a subclass of gimple_opt_pass along with...
2797 (pass_data_object_sizes): ...new pass_data instance and...
2798 (make_pass_object_sizes): ...new function.
2799 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
2800 global struct to a subclass of gimple_opt_pass along with...
2801 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
2803 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2804 (pass_fixup_cfg): Convert from a global struct to a subclass of
2805 gimple_opt_pass along with...
2806 (pass_data_fixup_cfg): ...new pass_data instance and...
2807 (make_pass_fixup_cfg): ...new function.
2808 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
2809 (make_pass_mudflap_1): ...new function.
2810 (pass_mudflap_2): Replace declaration with that of...
2811 (make_pass_mudflap_2): ...new function.
2812 (pass_asan): Replace declaration with that of...
2813 (make_pass_asan): ...new function.
2814 (pass_asan_O0): Replace declaration with that of...
2815 (make_pass_asan_O0): ...new function.
2816 (pass_tsan): Replace declaration with that of...
2817 (make_pass_tsan): ...new function.
2818 (pass_tsan_O0): Replace declaration with that of...
2819 (make_pass_tsan_O0): ...new function.
2820 (pass_lower_cf): Replace declaration with that of...
2821 (make_pass_lower_cf): ...new function.
2822 (pass_refactor_eh): Replace declaration with that of...
2823 (make_pass_refactor_eh): ...new function.
2824 (pass_lower_eh): Replace declaration with that of...
2825 (make_pass_lower_eh): ...new function.
2826 (pass_lower_eh_dispatch): Replace declaration with that of...
2827 (make_pass_lower_eh_dispatch): ...new function.
2828 (pass_lower_resx): Replace declaration with that of...
2829 (make_pass_lower_resx): ...new function.
2830 (pass_build_cfg): Replace declaration with that of...
2831 (make_pass_build_cfg): ...new function.
2832 (pass_early_tree_profile): Replace declaration with that of...
2833 (make_pass_early_tree_profile): ...new function.
2834 (pass_cleanup_eh): Replace declaration with that of...
2835 (make_pass_cleanup_eh): ...new function.
2836 (pass_sra): Replace declaration with that of...
2837 (make_pass_sra): ...new function.
2838 (pass_sra_early): Replace declaration with that of...
2839 (make_pass_sra_early): ...new function.
2840 (pass_early_ipa_sra): Replace declaration with that of...
2841 (make_pass_early_ipa_sra): ...new function.
2842 (pass_tail_recursion): Replace declaration with that of...
2843 (make_pass_tail_recursion): ...new function.
2844 (pass_tail_calls): Replace declaration with that of...
2845 (make_pass_tail_calls): ...new function.
2846 (pass_tree_loop): Replace declaration with that of...
2847 (make_pass_tree_loop): ...new function.
2848 (pass_tree_loop_init): Replace declaration with that of...
2849 (make_pass_tree_loop_init): ...new function.
2850 (pass_lim): Replace declaration with that of...
2851 (make_pass_lim): ...new function.
2852 (pass_tree_unswitch): Replace declaration with that of...
2853 (make_pass_tree_unswitch): ...new function.
2854 (pass_predcom): Replace declaration with that of...
2855 (make_pass_predcom): ...new function.
2856 (pass_iv_canon): Replace declaration with that of...
2857 (make_pass_iv_canon): ...new function.
2858 (pass_scev_cprop): Replace declaration with that of...
2859 (make_pass_scev_cprop): ...new function.
2860 (pass_empty_loop): Replace declaration with that of...
2861 (make_pass_empty_loop): ...new function.
2862 (pass_record_bounds): Replace declaration with that of...
2863 (make_pass_record_bounds): ...new function.
2864 (pass_graphite): Replace declaration with that of...
2865 (make_pass_graphite): ...new function.
2866 (pass_graphite_transforms): Replace declaration with that of...
2867 (make_pass_graphite_transforms): ...new function.
2868 (pass_if_conversion): Replace declaration with that of...
2869 (make_pass_if_conversion): ...new function.
2870 (pass_loop_distribution): Replace declaration with that of...
2871 (make_pass_loop_distribution): ...new function.
2872 (pass_vectorize): Replace declaration with that of...
2873 (make_pass_vectorize): ...new function.
2874 (pass_slp_vectorize): Replace declaration with that of...
2875 (make_pass_slp_vectorize): ...new function.
2876 (pass_complete_unroll): Replace declaration with that of...
2877 (make_pass_complete_unroll): ...new function.
2878 (pass_complete_unrolli): Replace declaration with that of...
2879 (make_pass_complete_unrolli): ...new function.
2880 (pass_parallelize_loops): Replace declaration with that of...
2881 (make_pass_parallelize_loops): ...new function.
2882 (pass_loop_prefetch): Replace declaration with that of...
2883 (make_pass_loop_prefetch): ...new function.
2884 (pass_iv_optimize): Replace declaration with that of...
2885 (make_pass_iv_optimize): ...new function.
2886 (pass_tree_loop_done): Replace declaration with that of...
2887 (make_pass_tree_loop_done): ...new function.
2888 (pass_ch): Replace declaration with that of...
2889 (make_pass_ch): ...new function.
2890 (pass_ccp): Replace declaration with that of...
2891 (make_pass_ccp): ...new function.
2892 (pass_phi_only_cprop): Replace declaration with that of...
2893 (make_pass_phi_only_cprop): ...new function.
2894 (pass_build_ssa): Replace declaration with that of...
2895 (make_pass_build_ssa): ...new function.
2896 (pass_build_alias): Replace declaration with that of...
2897 (make_pass_build_alias): ...new function.
2898 (pass_build_ealias): Replace declaration with that of...
2899 (make_pass_build_ealias): ...new function.
2900 (pass_dominator): Replace declaration with that of...
2901 (make_pass_dominator): ...new function.
2902 (pass_dce): Replace declaration with that of...
2903 (make_pass_dce): ...new function.
2904 (pass_dce_loop): Replace declaration with that of...
2905 (make_pass_dce_loop): ...new function.
2906 (pass_cd_dce): Replace declaration with that of...
2907 (make_pass_cd_dce): ...new function.
2908 (pass_call_cdce): Replace declaration with that of...
2909 (make_pass_call_cdce): ...new function.
2910 (pass_merge_phi): Replace declaration with that of...
2911 (make_pass_merge_phi): ...new function.
2912 (pass_split_crit_edges): Replace declaration with that of...
2913 (make_pass_split_crit_edges): ...new function.
2914 (pass_pre): Replace declaration with that of...
2915 (make_pass_pre): ...new function.
2916 (pass_profile): Replace declaration with that of...
2917 (make_pass_profile): ...new function.
2918 (pass_strip_predict_hints): Replace declaration with that of...
2919 (make_pass_strip_predict_hints): ...new function.
2920 (pass_lower_complex_O0): Replace declaration with that of...
2921 (make_pass_lower_complex_O0): ...new function.
2922 (pass_lower_complex): Replace declaration with that of...
2923 (make_pass_lower_complex): ...new function.
2924 (pass_lower_vector): Replace declaration with that of...
2925 (make_pass_lower_vector): ...new function.
2926 (pass_lower_vector_ssa): Replace declaration with that of...
2927 (make_pass_lower_vector_ssa): ...new function.
2928 (pass_lower_omp): Replace declaration with that of...
2929 (make_pass_lower_omp): ...new function.
2930 (pass_diagnose_omp_blocks): Replace declaration with that of...
2931 (make_pass_diagnose_omp_blocks): ...new function.
2932 (pass_expand_omp): Replace declaration with that of...
2933 (make_pass_expand_omp): ...new function.
2934 (pass_expand_omp_ssa): Replace declaration with that of...
2935 (make_pass_expand_omp_ssa): ...new function.
2936 (pass_object_sizes): Replace declaration with that of...
2937 (make_pass_object_sizes): ...new function.
2938 (pass_strlen): Replace declaration with that of...
2939 (make_pass_strlen): ...new function.
2940 (pass_fold_builtins): Replace declaration with that of...
2941 (make_pass_fold_builtins): ...new function.
2942 (pass_stdarg): Replace declaration with that of...
2943 (make_pass_stdarg): ...new function.
2944 (pass_early_warn_uninitialized): Replace declaration with that of...
2945 (make_pass_early_warn_uninitialized): ...new function.
2946 (pass_late_warn_uninitialized): Replace declaration with that of...
2947 (make_pass_late_warn_uninitialized): ...new function.
2948 (pass_cse_reciprocals): Replace declaration with that of...
2949 (make_pass_cse_reciprocals): ...new function.
2950 (pass_cse_sincos): Replace declaration with that of...
2951 (make_pass_cse_sincos): ...new function.
2952 (pass_optimize_bswap): Replace declaration with that of...
2953 (make_pass_optimize_bswap): ...new function.
2954 (pass_optimize_widening_mul): Replace declaration with that of...
2955 (make_pass_optimize_widening_mul): ...new function.
2956 (pass_warn_function_return): Replace declaration with that of...
2957 (make_pass_warn_function_return): ...new function.
2958 (pass_warn_function_noreturn): Replace declaration with that of...
2959 (make_pass_warn_function_noreturn): ...new function.
2960 (pass_cselim): Replace declaration with that of...
2961 (make_pass_cselim): ...new function.
2962 (pass_phiopt): Replace declaration with that of...
2963 (make_pass_phiopt): ...new function.
2964 (pass_forwprop): Replace declaration with that of...
2965 (make_pass_forwprop): ...new function.
2966 (pass_phiprop): Replace declaration with that of...
2967 (make_pass_phiprop): ...new function.
2968 (pass_tree_ifcombine): Replace declaration with that of...
2969 (make_pass_tree_ifcombine): ...new function.
2970 (pass_dse): Replace declaration with that of...
2971 (make_pass_dse): ...new function.
2972 (pass_nrv): Replace declaration with that of...
2973 (make_pass_nrv): ...new function.
2974 (pass_rename_ssa_copies): Replace declaration with that of...
2975 (make_pass_rename_ssa_copies): ...new function.
2976 (pass_sink_code): Replace declaration with that of...
2977 (make_pass_sink_code): ...new function.
2978 (pass_fre): Replace declaration with that of...
2979 (make_pass_fre): ...new function.
2980 (pass_check_data_deps): Replace declaration with that of...
2981 (make_pass_check_data_deps): ...new function.
2982 (pass_copy_prop): Replace declaration with that of...
2983 (make_pass_copy_prop): ...new function.
2984 (pass_vrp): Replace declaration with that of...
2985 (make_pass_vrp): ...new function.
2986 (pass_uncprop): Replace declaration with that of...
2987 (make_pass_uncprop): ...new function.
2988 (pass_return_slot): Replace declaration with that of...
2989 (make_pass_return_slot): ...new function.
2990 (pass_reassoc): Replace declaration with that of...
2991 (make_pass_reassoc): ...new function.
2992 (pass_rebuild_cgraph_edges): Replace declaration with that of...
2993 (make_pass_rebuild_cgraph_edges): ...new function.
2994 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
2995 (make_pass_remove_cgraph_callee_edges): ...new function.
2996 (pass_build_cgraph_edges): Replace declaration with that of...
2997 (make_pass_build_cgraph_edges): ...new function.
2998 (pass_local_pure_const): Replace declaration with that of...
2999 (make_pass_local_pure_const): ...new function.
3000 (pass_tracer): Replace declaration with that of...
3001 (make_pass_tracer): ...new function.
3002 (pass_warn_unused_result): Replace declaration with that of...
3003 (make_pass_warn_unused_result): ...new function.
3004 (pass_diagnose_tm_blocks): Replace declaration with that of...
3005 (make_pass_diagnose_tm_blocks): ...new function.
3006 (pass_lower_tm): Replace declaration with that of...
3007 (make_pass_lower_tm): ...new function.
3008 (pass_tm_init): Replace declaration with that of...
3009 (make_pass_tm_init): ...new function.
3010 (pass_tm_mark): Replace declaration with that of...
3011 (make_pass_tm_mark): ...new function.
3012 (pass_tm_memopt): Replace declaration with that of...
3013 (make_pass_tm_memopt): ...new function.
3014 (pass_tm_edges): Replace declaration with that of...
3015 (make_pass_tm_edges): ...new function.
3016 (pass_split_functions): Replace declaration with that of...
3017 (make_pass_split_functions): ...new function.
3018 (pass_feedback_split_functions): Replace declaration with that of...
3019 (make_pass_feedback_split_functions): ...new function.
3020 (pass_strength_reduction): Replace declaration with that of...
3021 (make_pass_strength_reduction): ...new function.
3022 (pass_ipa_lower_emutls): Replace declaration with that of...
3023 (make_pass_ipa_lower_emutls): ...new function.
3024 (pass_ipa_function_and_variable_visibility): Replace declaration with
3026 (make_pass_ipa_function_and_variable_visibility): ...new function.
3027 (pass_ipa_tree_profile): Replace declaration with that of...
3028 (make_pass_ipa_tree_profile): ...new function.
3029 (pass_early_local_passes): Replace declaration with that of...
3030 (make_pass_early_local_passes): ...new function.
3031 (pass_ipa_whole_program_visibility): Replace declaration with that
3033 (make_pass_ipa_whole_program_visibility): ...new function.
3034 (pass_ipa_lto_gimple_out): Replace declaration with that of...
3035 (make_pass_ipa_lto_gimple_out): ...new function.
3036 (pass_ipa_increase_alignment): Replace declaration with that of...
3037 (make_pass_ipa_increase_alignment): ...new function.
3038 (pass_ipa_inline): Replace declaration with that of...
3039 (make_pass_ipa_inline): ...new function.
3040 (pass_ipa_free_lang_data): Replace declaration with that of...
3041 (make_pass_ipa_free_lang_data): ...new function.
3042 (pass_ipa_free_inline_summary): Replace declaration with that of...
3043 (make_pass_ipa_free_inline_summary): ...new function.
3044 (pass_ipa_cp): Replace declaration with that of...
3045 (make_pass_ipa_cp): ...new function.
3046 (pass_ipa_reference): Replace declaration with that of...
3047 (make_pass_ipa_reference): ...new function.
3048 (pass_ipa_pure_const): Replace declaration with that of...
3049 (make_pass_ipa_pure_const): ...new function.
3050 (pass_ipa_pta): Replace declaration with that of...
3051 (make_pass_ipa_pta): ...new function.
3052 (pass_ipa_lto_finish_out): Replace declaration with that of...
3053 (make_pass_ipa_lto_finish_out): ...new function.
3054 (pass_ipa_tm): Replace declaration with that of...
3055 (make_pass_ipa_tm): ...new function.
3056 (pass_ipa_profile): Replace declaration with that of...
3057 (make_pass_ipa_profile): ...new function.
3058 (pass_ipa_cdtor_merge): Replace declaration with that of...
3059 (make_pass_ipa_cdtor_merge): ...new function.
3060 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
3062 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3063 (pass_init_datastructures): Replace declaration with that of...
3064 (make_pass_init_datastructures): ...new function.
3065 (pass_fixup_cfg): Replace declaration with that of...
3066 (make_pass_fixup_cfg): ...new function.
3067 (pass_expand): Replace declaration with that of...
3068 (make_pass_expand): ...new function.
3069 (pass_instantiate_virtual_regs): Replace declaration with that of...
3070 (make_pass_instantiate_virtual_regs): ...new function.
3071 (pass_rtl_fwprop): Replace declaration with that of...
3072 (make_pass_rtl_fwprop): ...new function.
3073 (pass_rtl_fwprop_addr): Replace declaration with that of...
3074 (make_pass_rtl_fwprop_addr): ...new function.
3075 (pass_jump): Replace declaration with that of...
3076 (make_pass_jump): ...new function.
3077 (pass_jump2): Replace declaration with that of...
3078 (make_pass_jump2): ...new function.
3079 (pass_lower_subreg): Replace declaration with that of...
3080 (make_pass_lower_subreg): ...new function.
3081 (pass_cse): Replace declaration with that of...
3082 (make_pass_cse): ...new function.
3083 (pass_fast_rtl_dce): Replace declaration with that of...
3084 (make_pass_fast_rtl_dce): ...new function.
3085 (pass_ud_rtl_dce): Replace declaration with that of...
3086 (make_pass_ud_rtl_dce): ...new function.
3087 (pass_rtl_dce): Replace declaration with that of...
3088 (make_pass_rtl_dce): ...new function.
3089 (pass_rtl_dse1): Replace declaration with that of...
3090 (make_pass_rtl_dse1): ...new function.
3091 (pass_rtl_dse2): Replace declaration with that of...
3092 (make_pass_rtl_dse2): ...new function.
3093 (pass_rtl_dse3): Replace declaration with that of...
3094 (make_pass_rtl_dse3): ...new function.
3095 (pass_rtl_cprop): Replace declaration with that of...
3096 (make_pass_rtl_cprop): ...new function.
3097 (pass_rtl_pre): Replace declaration with that of...
3098 (make_pass_rtl_pre): ...new function.
3099 (pass_rtl_hoist): Replace declaration with that of...
3100 (make_pass_rtl_hoist): ...new function.
3101 (pass_rtl_store_motion): Replace declaration with that of...
3102 (make_pass_rtl_store_motion): ...new function.
3103 (pass_cse_after_global_opts): Replace declaration with that of...
3104 (make_pass_cse_after_global_opts): ...new function.
3105 (pass_rtl_ifcvt): Replace declaration with that of...
3106 (make_pass_rtl_ifcvt): ...new function.
3107 (pass_into_cfg_layout_mode): Replace declaration with that of...
3108 (make_pass_into_cfg_layout_mode): ...new function.
3109 (pass_outof_cfg_layout_mode): Replace declaration with that of...
3110 (make_pass_outof_cfg_layout_mode): ...new function.
3111 (pass_loop2): Replace declaration with that of...
3112 (make_pass_loop2): ...new function.
3113 (pass_rtl_loop_init): Replace declaration with that of...
3114 (make_pass_rtl_loop_init): ...new function.
3115 (pass_rtl_move_loop_invariants): Replace declaration with that of...
3116 (make_pass_rtl_move_loop_invariants): ...new function.
3117 (pass_rtl_unswitch): Replace declaration with that of...
3118 (make_pass_rtl_unswitch): ...new function.
3119 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
3120 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3121 (pass_rtl_doloop): Replace declaration with that of...
3122 (make_pass_rtl_doloop): ...new function.
3123 (pass_rtl_loop_done): Replace declaration with that of...
3124 (make_pass_rtl_loop_done): ...new function.
3125 (pass_web): Replace declaration with that of...
3126 (make_pass_web): ...new function.
3127 (pass_cse2): Replace declaration with that of...
3128 (make_pass_cse2): ...new function.
3129 (pass_df_initialize_opt): Replace declaration with that of...
3130 (make_pass_df_initialize_opt): ...new function.
3131 (pass_df_initialize_no_opt): Replace declaration with that of...
3132 (make_pass_df_initialize_no_opt): ...new function.
3133 (pass_reginfo_init): Replace declaration with that of...
3134 (make_pass_reginfo_init): ...new function.
3135 (pass_inc_dec): Replace declaration with that of...
3136 (make_pass_inc_dec): ...new function.
3137 (pass_stack_ptr_mod): Replace declaration with that of...
3138 (make_pass_stack_ptr_mod): ...new function.
3139 (pass_initialize_regs): Replace declaration with that of...
3140 (make_pass_initialize_regs): ...new function.
3141 (pass_combine): Replace declaration with that of...
3142 (make_pass_combine): ...new function.
3143 (pass_if_after_combine): Replace declaration with that of...
3144 (make_pass_if_after_combine): ...new function.
3145 (pass_ree): Replace declaration with that of...
3146 (make_pass_ree): ...new function.
3147 (pass_partition_blocks): Replace declaration with that of...
3148 (make_pass_partition_blocks): ...new function.
3149 (pass_match_asm_constraints): Replace declaration with that of...
3150 (make_pass_match_asm_constraints): ...new function.
3151 (pass_regmove): Replace declaration with that of...
3152 (make_pass_regmove): ...new function.
3153 (pass_split_all_insns): Replace declaration with that of...
3154 (make_pass_split_all_insns): ...new function.
3155 (pass_fast_rtl_byte_dce): Replace declaration with that of...
3156 (make_pass_fast_rtl_byte_dce): ...new function.
3157 (pass_lower_subreg2): Replace declaration with that of...
3158 (make_pass_lower_subreg2): ...new function.
3159 (pass_mode_switching): Replace declaration with that of...
3160 (make_pass_mode_switching): ...new function.
3161 (pass_sms): Replace declaration with that of...
3162 (make_pass_sms): ...new function.
3163 (pass_sched): Replace declaration with that of...
3164 (make_pass_sched): ...new function.
3165 (pass_ira): Replace declaration with that of...
3166 (make_pass_ira): ...new function.
3167 (pass_reload): Replace declaration with that of...
3168 (make_pass_reload): ...new function.
3169 (pass_clean_state): Replace declaration with that of...
3170 (make_pass_clean_state): ...new function.
3171 (pass_branch_prob): Replace declaration with that of...
3172 (make_pass_branch_prob): ...new function.
3173 (pass_value_profile_transformations): Replace declaration with that
3175 (make_pass_value_profile_transformations): ...new function.
3176 (pass_postreload_cse): Replace declaration with that of...
3177 (make_pass_postreload_cse): ...new function.
3178 (pass_gcse2): Replace declaration with that of...
3179 (make_pass_gcse2): ...new function.
3180 (pass_split_after_reload): Replace declaration with that of...
3181 (make_pass_split_after_reload): ...new function.
3182 (pass_branch_target_load_optimize1): Replace declaration with that
3184 (make_pass_branch_target_load_optimize1): ...new function.
3185 (pass_thread_prologue_and_epilogue): Replace declaration with that
3187 (make_pass_thread_prologue_and_epilogue): ...new function.
3188 (pass_stack_adjustments): Replace declaration with that of...
3189 (make_pass_stack_adjustments): ...new function.
3190 (pass_peephole2): Replace declaration with that of...
3191 (make_pass_peephole2): ...new function.
3192 (pass_if_after_reload): Replace declaration with that of...
3193 (make_pass_if_after_reload): ...new function.
3194 (pass_regrename): Replace declaration with that of...
3195 (make_pass_regrename): ...new function.
3196 (pass_cprop_hardreg): Replace declaration with that of...
3197 (make_pass_cprop_hardreg): ...new function.
3198 (pass_reorder_blocks): Replace declaration with that of...
3199 (make_pass_reorder_blocks): ...new function.
3200 (pass_branch_target_load_optimize2): Replace declaration with that
3202 (make_pass_branch_target_load_optimize2): ...new function.
3203 (pass_leaf_regs): Replace declaration with that of...
3204 (make_pass_leaf_regs): ...new function.
3205 (pass_split_before_sched2): Replace declaration with that of...
3206 (make_pass_split_before_sched2): ...new function.
3207 (pass_compare_elim_after_reload): Replace declaration with that of...
3208 (make_pass_compare_elim_after_reload): ...new function.
3209 (pass_sched2): Replace declaration with that of...
3210 (make_pass_sched2): ...new function.
3211 (pass_stack_regs): Replace declaration with that of...
3212 (make_pass_stack_regs): ...new function.
3213 (pass_stack_regs_run): Replace declaration with that of...
3214 (make_pass_stack_regs_run): ...new function.
3215 (pass_df_finish): Replace declaration with that of...
3216 (make_pass_df_finish): ...new function.
3217 (pass_compute_alignments): Replace declaration with that of...
3218 (make_pass_compute_alignments): ...new function.
3219 (pass_duplicate_computed_gotos): Replace declaration with that of...
3220 (make_pass_duplicate_computed_gotos): ...new function.
3221 (pass_variable_tracking): Replace declaration with that of...
3222 (make_pass_variable_tracking): ...new function.
3223 (pass_free_cfg): Replace declaration with that of...
3224 (make_pass_free_cfg): ...new function.
3225 (pass_machine_reorg): Replace declaration with that of...
3226 (make_pass_machine_reorg): ...new function.
3227 (pass_cleanup_barriers): Replace declaration with that of...
3228 (make_pass_cleanup_barriers): ...new function.
3229 (pass_delay_slots): Replace declaration with that of...
3230 (make_pass_delay_slots): ...new function.
3231 (pass_split_for_shorten_branches): Replace declaration with that of...
3232 (make_pass_split_for_shorten_branches): ...new function.
3233 (pass_split_before_regstack): Replace declaration with that of...
3234 (make_pass_split_before_regstack): ...new function.
3235 (pass_convert_to_eh_region_ranges): Replace declaration with that
3237 (make_pass_convert_to_eh_region_ranges): ...new function.
3238 (pass_shorten_branches): Replace declaration with that of...
3239 (make_pass_shorten_branches): ...new function.
3240 (pass_set_nothrow_function_flags): Replace declaration with that of...
3241 (make_pass_set_nothrow_function_flags): ...new function.
3242 (pass_dwarf2_frame): Replace declaration with that of...
3243 (make_pass_dwarf2_frame): ...new function.
3244 (pass_final): Replace declaration with that of...
3245 (make_pass_final): ...new function.
3246 (pass_rtl_seqabstr): Replace declaration with that of...
3247 (make_pass_rtl_seqabstr): ...new function.
3248 (pass_release_ssa_names): Replace declaration with that of...
3249 (make_pass_release_ssa_names): ...new function.
3250 (pass_early_inline): Replace declaration with that of...
3251 (make_pass_early_inline): ...new function.
3252 (pass_inline_parameters): Replace declaration with that of...
3253 (make_pass_inline_parameters): ...new function.
3254 (pass_update_address_taken): Replace declaration with that of...
3255 (make_pass_update_address_taken): ...new function.
3256 (pass_convert_switch): Replace declaration with that of...
3257 (make_pass_convert_switch): ...new function.
3258 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
3259 to a subclass of simple_ipa_opt_pass along with...
3260 (pass_data_ipa_tree_profile): ...new pass_data instance and...
3261 (make_pass_ipa_tree_profile): ...new function.
3262 * tree-sra.c (pass_sra_early): Convert from a global struct to a
3263 subclass of gimple_opt_pass along with...
3264 (pass_data_sra_early): ...new pass_data instance and...
3265 (make_pass_sra_early): ...new function.
3266 (pass_sra): Convert from a global struct to a subclass of
3267 gimple_opt_pass along with...
3268 (pass_data_sra): ...new pass_data instance and...
3269 (make_pass_sra): ...new function.
3270 (pass_early_ipa_sra): Convert from a global struct to a subclass of
3271 gimple_opt_pass along with...
3272 (pass_data_early_ipa_sra): ...new pass_data instance and...
3273 (make_pass_early_ipa_sra): ...new function.
3274 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
3275 subclass of gimple_opt_pass along with...
3276 (pass_data_ccp): ...new pass_data instance and...
3277 (make_pass_ccp): ...new function.
3278 (pass_fold_builtins): Convert from a global struct to a subclass of
3279 gimple_opt_pass along with...
3280 (pass_data_fold_builtins): ...new pass_data instance and...
3281 (make_pass_fold_builtins): ...new function.
3282 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
3283 subclass of gimple_opt_pass along with...
3284 (pass_data_copy_prop): ...new pass_data instance and...
3285 (make_pass_copy_prop): ...new function.
3286 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
3287 global struct to a subclass of gimple_opt_pass along with...
3288 (pass_data_rename_ssa_copies): ...new pass_data instance and...
3289 (make_pass_rename_ssa_copies): ...new function.
3290 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
3291 subclass of gimple_opt_pass along with...
3292 (pass_data_dce): ...new pass_data instance and...
3293 (make_pass_dce): ...new function.
3294 (pass_dce_loop): Convert from a global struct to a subclass of
3295 gimple_opt_pass along with...
3296 (pass_data_dce_loop): ...new pass_data instance and...
3297 (make_pass_dce_loop): ...new function.
3298 (pass_cd_dce): Convert from a global struct to a subclass of
3299 gimple_opt_pass along with...
3300 (pass_data_cd_dce): ...new pass_data instance and...
3301 (make_pass_cd_dce): ...new function.
3302 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
3303 subclass of gimple_opt_pass along with...
3304 (pass_data_dominator): ...new pass_data instance and...
3305 (make_pass_dominator): ...new function.
3306 (pass_phi_only_cprop): Convert from a global struct to a subclass of
3307 gimple_opt_pass along with...
3308 (pass_data_phi_only_cprop): ...new pass_data instance and...
3309 (make_pass_phi_only_cprop): ...new function.
3310 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
3311 subclass of gimple_opt_pass along with...
3312 (pass_data_dse): ...new pass_data instance and...
3313 (make_pass_dse): ...new function.
3314 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
3315 a subclass of gimple_opt_pass along with...
3316 (pass_data_forwprop): ...new pass_data instance and...
3317 (make_pass_forwprop): ...new function.
3318 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
3319 struct to a subclass of gimple_opt_pass along with...
3320 (pass_data_tree_ifcombine): ...new pass_data instance and...
3321 (make_pass_tree_ifcombine): ...new function.
3322 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
3323 subclass of gimple_opt_pass along with...
3324 (pass_data_ch): ...new pass_data instance and...
3325 (make_pass_ch): ...new function.
3326 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
3327 subclass of gimple_opt_pass along with...
3328 (pass_data_tree_loop): ...new pass_data instance and...
3329 (make_pass_tree_loop): ...new function.
3330 (pass_tree_loop_init): Convert from a global struct to a subclass of
3331 gimple_opt_pass along with...
3332 (pass_data_tree_loop_init): ...new pass_data instance and...
3333 (make_pass_tree_loop_init): ...new function.
3334 (pass_lim): Convert from a global struct to a subclass of
3335 gimple_opt_pass along with...
3336 (pass_data_lim): ...new pass_data instance and...
3337 (make_pass_lim): ...new function.
3338 (pass_tree_unswitch): Convert from a global struct to a subclass of
3339 gimple_opt_pass along with...
3340 (pass_data_tree_unswitch): ...new pass_data instance and...
3341 (make_pass_tree_unswitch): ...new function.
3342 (pass_predcom): Convert from a global struct to a subclass of
3343 gimple_opt_pass along with...
3344 (pass_data_predcom): ...new pass_data instance and...
3345 (make_pass_predcom): ...new function.
3346 (pass_vectorize): Convert from a global struct to a subclass of
3347 gimple_opt_pass along with...
3348 (pass_data_vectorize): ...new pass_data instance and...
3349 (make_pass_vectorize): ...new function.
3350 (pass_graphite): Convert from a global struct to a subclass of
3351 gimple_opt_pass along with...
3352 (pass_data_graphite): ...new pass_data instance and...
3353 (make_pass_graphite): ...new function.
3354 (pass_graphite_transforms): Convert from a global struct to a subclass
3355 of gimple_opt_pass along with...
3356 (pass_data_graphite_transforms): ...new pass_data instance and...
3357 (make_pass_graphite_transforms): ...new function.
3358 (pass_check_data_deps): Convert from a global struct to a subclass of
3359 gimple_opt_pass along with...
3360 (pass_data_check_data_deps): ...new pass_data instance and...
3361 (make_pass_check_data_deps): ...new function.
3362 (pass_iv_canon): Convert from a global struct to a subclass of
3363 gimple_opt_pass along with...
3364 (pass_data_iv_canon): ...new pass_data instance and...
3365 (make_pass_iv_canon): ...new function.
3366 (pass_scev_cprop): Convert from a global struct to a subclass of
3367 gimple_opt_pass along with...
3368 (pass_data_scev_cprop): ...new pass_data instance and...
3369 (make_pass_scev_cprop): ...new function.
3370 (pass_record_bounds): Convert from a global struct to a subclass of
3371 gimple_opt_pass along with...
3372 (pass_data_record_bounds): ...new pass_data instance and...
3373 (make_pass_record_bounds): ...new function.
3374 (pass_complete_unroll): Convert from a global struct to a subclass of
3375 gimple_opt_pass along with...
3376 (pass_data_complete_unroll): ...new pass_data instance and...
3377 (make_pass_complete_unroll): ...new function.
3378 (pass_complete_unrolli): Convert from a global struct to a subclass of
3379 gimple_opt_pass along with...
3380 (pass_data_complete_unrolli): ...new pass_data instance and...
3381 (make_pass_complete_unrolli): ...new function.
3382 (pass_parallelize_loops): Convert from a global struct to a subclass
3383 of gimple_opt_pass along with...
3384 (pass_data_parallelize_loops): ...new pass_data instance and...
3385 (make_pass_parallelize_loops): ...new function.
3386 (pass_loop_prefetch): Convert from a global struct to a subclass of
3387 gimple_opt_pass along with...
3388 (pass_data_loop_prefetch): ...new pass_data instance and...
3389 (make_pass_loop_prefetch): ...new function.
3390 (pass_iv_optimize): Convert from a global struct to a subclass of
3391 gimple_opt_pass along with...
3392 (pass_data_iv_optimize): ...new pass_data instance and...
3393 (make_pass_iv_optimize): ...new function.
3394 (pass_tree_loop_done): Convert from a global struct to a subclass of
3395 gimple_opt_pass along with...
3396 (pass_data_tree_loop_done): ...new pass_data instance and...
3397 (make_pass_tree_loop_done): ...new function.
3398 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
3399 struct to a subclass of gimple_opt_pass along with...
3400 (pass_data_cse_reciprocals): ...new pass_data instance and...
3401 (make_pass_cse_reciprocals): ...new function.
3402 (pass_cse_sincos): Convert from a global struct to a subclass of
3403 gimple_opt_pass along with...
3404 (pass_data_cse_sincos): ...new pass_data instance and...
3405 (make_pass_cse_sincos): ...new function.
3406 (pass_optimize_bswap): Convert from a global struct to a subclass of
3407 gimple_opt_pass along with...
3408 (pass_data_optimize_bswap): ...new pass_data instance and...
3409 (make_pass_optimize_bswap): ...new function.
3410 (pass_optimize_widening_mul): Convert from a global struct to a
3411 subclass of gimple_opt_pass along with...
3412 (pass_data_optimize_widening_mul): ...new pass_data instance and...
3413 (make_pass_optimize_widening_mul): ...new function.
3414 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
3415 subclass of gimple_opt_pass along with...
3416 (pass_data_phiopt): ...new pass_data instance and...
3417 (make_pass_phiopt): ...new function.
3418 (pass_cselim): Convert from a global struct to a subclass of
3419 gimple_opt_pass along with...
3420 (pass_data_cselim): ...new pass_data instance and...
3421 (make_pass_cselim): ...new function.
3422 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
3423 subclass of gimple_opt_pass along with...
3424 (pass_data_phiprop): ...new pass_data instance and...
3425 (make_pass_phiprop): ...new function.
3426 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
3427 subclass of gimple_opt_pass along with...
3428 (pass_data_pre): ...new pass_data instance and...
3429 (make_pass_pre): ...new function.
3430 (pass_fre): Convert from a global struct to a subclass of
3431 gimple_opt_pass along with...
3432 (pass_data_fre): ...new pass_data instance and...
3433 (make_pass_fre): ...new function.
3434 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
3435 subclass of gimple_opt_pass along with...
3436 (pass_data_reassoc): ...new pass_data instance and...
3437 (make_pass_reassoc): ...new function.
3438 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
3439 subclass of gimple_opt_pass along with...
3440 (pass_data_sink_code): ...new pass_data instance and...
3441 (make_pass_sink_code): ...new function.
3442 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
3443 subclass of gimple_opt_pass along with...
3444 (pass_data_strlen): ...new pass_data instance and...
3445 (make_pass_strlen): ...new function.
3446 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
3447 struct to a subclass of gimple_opt_pass along with...
3448 (pass_data_build_alias): ...new pass_data instance and...
3449 (make_pass_build_alias): ...new function.
3450 (pass_build_ealias): Convert from a global struct to a subclass of
3451 gimple_opt_pass along with...
3452 (pass_data_build_ealias): ...new pass_data instance and...
3453 (make_pass_build_ealias): ...new function.
3454 (pass_ipa_pta): Convert from a global struct to a subclass of
3455 simple_ipa_opt_pass along with...
3456 (pass_data_ipa_pta): ...new pass_data instance and...
3457 (make_pass_ipa_pta): ...new function.
3458 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
3459 subclass of gimple_opt_pass along with...
3460 (pass_data_uncprop): ...new pass_data instance and...
3461 (make_pass_uncprop): ...new function.
3462 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
3463 global struct to a subclass of gimple_opt_pass along with...
3464 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
3465 (make_pass_late_warn_uninitialized): ...new function.
3466 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
3467 to a subclass of gimple_opt_pass along with...
3468 (pass_data_init_datastructures): ...new pass_data instance and...
3469 (make_pass_init_datastructures): ...new function.
3470 (pass_early_warn_uninitialized): Convert from a global struct to a
3471 subclass of gimple_opt_pass along with...
3472 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
3473 (make_pass_early_warn_uninitialized): ...new function.
3474 (pass_update_address_taken): Convert from a global struct to a
3475 subclass of gimple_opt_pass along with...
3476 (pass_data_update_address_taken): ...new pass_data instance and...
3477 (make_pass_update_address_taken): ...new function.
3478 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
3479 struct to a subclass of gimple_opt_pass along with...
3480 (pass_data_release_ssa_names): ...new pass_data instance and...
3481 (make_pass_release_ssa_names): ...new function.
3482 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
3483 subclass of gimple_opt_pass along with...
3484 (pass_data_stdarg): ...new pass_data instance and...
3485 (make_pass_stdarg): ...new function.
3486 * tree-switch-conversion.c (pass_convert_switch): Convert from a
3487 global struct to a subclass of gimple_opt_pass along with...
3488 (pass_data_convert_switch): ...new pass_data instance and...
3489 (make_pass_convert_switch): ...new function.
3490 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
3491 to a subclass of gimple_opt_pass along with...
3492 (pass_data_tail_recursion): ...new pass_data instance and...
3493 (make_pass_tail_recursion): ...new function.
3494 (pass_tail_calls): Convert from a global struct to a subclass of
3495 gimple_opt_pass along with...
3496 (pass_data_tail_calls): ...new pass_data instance and...
3497 (make_pass_tail_calls): ...new function.
3498 * tree-vect-generic.c (pass_lower_vector): Convert from a global
3499 struct to a subclass of gimple_opt_pass along with...
3500 (pass_data_lower_vector): ...new pass_data instance and...
3501 (make_pass_lower_vector): ...new function.
3502 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
3503 gimple_opt_pass along with...
3504 (pass_data_lower_vector_ssa): ...new pass_data instance and...
3505 (make_pass_lower_vector_ssa): ...new function.
3506 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
3507 to a subclass of gimple_opt_pass along with...
3508 (pass_data_slp_vectorize): ...new pass_data instance and...
3509 (make_pass_slp_vectorize): ...new function.
3510 (pass_ipa_increase_alignment): Convert from a global struct to a
3511 subclass of simple_ipa_opt_pass along with...
3512 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
3513 (make_pass_ipa_increase_alignment): ...new function.
3514 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
3515 gimple_opt_pass along with...
3516 (pass_data_vrp): ...new pass_data instance and...
3517 (make_pass_vrp): ...new function.
3518 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
3519 subclass of simple_ipa_opt_pass along with...
3520 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
3521 (make_pass_ipa_free_lang_data): ...new function.
3522 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
3523 gimple_opt_pass along with...
3524 (pass_data_tsan): ...new pass_data instance and...
3525 (make_pass_tsan): ...new function.
3526 (pass_tsan_O0): Convert from a global struct to a subclass of
3527 gimple_opt_pass along with...
3528 (pass_data_tsan_O0): ...new pass_data instance and...
3529 (make_pass_tsan_O0): ...new function.
3530 * var-tracking.c (pass_variable_tracking): Convert from a global
3531 struct to a subclass of rtl_opt_pass along with...
3532 (pass_data_variable_tracking): ...new pass_data instance and...
3533 (make_pass_variable_tracking): ...new function.
3534 * web.c (pass_web): Convert from a global struct to a subclass of
3535 rtl_opt_pass along with...
3536 (pass_data_web): ...new pass_data instance and...
3537 (make_pass_web): ...new function.
3538 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
3539 declaration with that of...
3540 (make_pass_mode_switch_use): ...new function.
3541 (pass_resolve_sw_modes): Replace declaration with that of...
3542 (make_pass_resolve_sw_modes): ...new function.
3543 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
3544 from a global struct to a subclass of rtl_opt_pass along with...
3545 (pass_data_mode_switch_use): ...new pass_data instance and...
3546 (make_pass_mode_switch_use): ...new function.
3547 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
3548 from a global struct to a subclass of rtl_opt_pass along with...
3549 (pass_data_resolve_sw_modes): ...new pass_data instance and...
3550 (make_pass_resolve_sw_modes): ...new function.
3551 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
3552 struct to a subclass of rtl_opt_pass along with...
3553 (pass_data_insert_vzeroupper): ...new pass_data instance and...
3554 (make_pass_insert_vzeroupper): ...new function.
3555 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
3556 global struct to a subclass of rtl_opt_pass along with...
3557 (pass_data_work_around_errata): ...new pass_data instance and...
3558 (make_pass_work_around_errata): ...new function.
3559 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
3560 struct to a subclass of rtl_opt_pass along with...
3561 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
3562 (make_pass_mips_machine_reorg2): ...new function.
3564 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3566 * passes.c (pass_manager::operator new): New.
3568 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3570 Handwritten part of conversion of passes to C++ classes.
3572 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
3573 (toplev.o): Add dep on PASS_MANAGER_H.
3574 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
3575 of early local pases to reflect this moving from a global to a
3576 member of gcc::pass_manager.
3577 (cgraph_add_new_function): Likewise.
3578 * lto-cgraph.c (lto_output_node): Update for conversion of
3579 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
3580 * passes.c (opt_pass::clone): New.
3581 (opt_pass::gate): New.
3582 (opt_pass::execute): New.
3583 (opt_pass::opt_pass): New.
3584 (pass_manager::execute_early_local_passes): New.
3585 (pass_manager::execute_pass_mode_switching): new.
3586 (finish_optimization_passes): Convert to...
3587 (pass_manager::finish_optimization_passes): ...this.
3588 (finish_optimization_passes): Update for conversion of passes to
3590 (register_dump_files_1): Use has_gate since we cannot portably
3591 check a vtable entry against NULL.
3592 (dump_one_pass): Likewise.
3593 (ipa_write_summaries_2): Likewise.
3594 (ipa_write_optimization_summaries_1): Likewise.
3595 (ipa_read_summaries_1): Likewise.
3596 (ipa_read_optimization_summaries_1): Likewise.
3597 (execute_ipa_stmt_fixups): Likewise.
3598 (pass_manager::pass_manager): Rewrite pass-creation, invoking
3599 pass-creation functions rather than wiring up globals, and
3600 storing the results in fields of pass_manager generated using
3602 (pass_manager::dump_profile_report): Update for conversion of
3603 passes to C++ classes.
3604 (pass_manager::execute_ipa_summary_passes): Likewise.
3605 (execute_one_ipa_transform_pass): Likewise.
3606 (execute_one_pass): Use has_gate and has_execute since we cannot
3607 portably check a vtable entry against NULL.
3608 * pass_manager.h (pass_manager::finish_optimization_passes): New.
3609 (pass_manager): Use pass-instances.def to add fields for the
3610 various pass instances.
3611 * toplev.c (finalize): Update for move of
3612 finish_optimization_passes to a method of gcc::pass_manager.
3613 * toplev.h (finish_optimization_passes): Move to method of class
3615 * tree-pass.h (struct pass_data): New.
3616 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
3617 (opt_pass::gate): Convert to virtual function.
3618 (opt_pass::~opt_pass): New.
3619 (opt_pass::clone): New.
3620 (opt_pass::execute): Convert to virtual function.
3621 (opt_pass::opt_pass): New.
3622 (opt_pass::ctxt_): new.
3623 (gimple_opt_pass): Convert to subclass of opt_pass.
3624 (gimple_opt_pass::gimple_opt_pass): New.
3625 (rtl_opt_pass): Convert to subclass of opt_pass.
3626 (rtl_opt_pass::rtl_opt_pass): New.
3627 (ipa_opt_pass_d): Convert to subclass of opt_pass.
3628 (ipa_opt_pass_d::ipa_opt_pass_d): New.
3629 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
3630 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
3631 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
3632 invocation of pass_mode_switching to reflect this moving from a
3633 global to a member of gcc::pass_manager.
3634 (ix86_option_override): Rework how pass_insert_vzeroupper is
3635 added to the pass_manager to reflect autogenerated changes.
3636 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
3638 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
3640 PR rtl-optimization/57708
3641 * recog.c (peep2_find_free_register): Validate all regs in a
3644 2013-08-05 Jan Hubicka <jh@suse.cz>
3647 * cgraph.c (verify_cgraph_node): Accept local flags from other
3649 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
3650 (function_and_variable_visibility): Likewise.
3651 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
3653 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3655 * graph.c (init_graph_slim_pretty_print): Remove.
3656 (print_graph_cfg): Do not call it. Use local pretty printer.
3657 (start_graph_dump): Likewise.
3659 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3661 * gimple-pretty-print.c (buffer): Remove.
3662 (initialized): Likewise.
3663 (maybe_init_pretty_print): Likewise.
3664 (print_gimple_stmt): Do not call it. Use non-static local
3665 pretty_printer variable.
3666 (print_gimple_expr): Likewise.
3667 (print_gimple_seq): Likewise.
3668 (gimple_dump_bb): Likewise.
3670 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3672 * asan.c (asan_pp): Remove.
3673 (asan_pp_initialized): Likewise.
3674 (asan_pp_initialize): Likewise.
3675 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
3676 (asan_emit_stack_protection): Tidy. Use local pretty printer.
3677 (asan_add_global): Likewise.
3679 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
3681 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
3682 * diagnostic.h (diagnostic_flush_buffer): Adjust.
3683 * pretty-print.c (pp_formatted_text_data): Likewise.
3684 (pp_indent): Rename from pp_base_indent.
3685 (pp_format): Rename from pp_base_format.
3686 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
3687 (pp_format_verbatim): Rename from pp_base_format_verbatim.
3688 (pp_flush): Rename from pp_base_flush.
3689 (pp_set_line_maximum_length): Rename from
3690 pp_base_set_line_maximum_length.
3691 (pp_clear_output_area): Rename from pp_base_clear_output_area.
3692 (pp_set_prefix): Rename from pp_base_set_prefix.
3693 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
3694 (pp_emit_prefix): Rename from pp_base_emit_prefix.
3695 (pp_append_text): Rename from pp_base_append_text.
3696 (pp_formatted_text): Rename from pp_base_formatted_text.
3697 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
3698 (pp_remaining_character_count_for_line): Rename from
3699 pp_base_remaining_character_count_for_line.
3700 (pp_newline): Rename from pp_base_newline.
3701 (pp_character): Rename from pp_base_character.
3702 (pp_string): Rename from pp_base_string.
3703 (pp_maybe_space): Rename from pp_base_maybe_space.
3704 * asan.c (asan_pp_string): Adjust.
3705 (asan_emit_stack_protection): Likewise.
3706 (asan_add_global): Likewise.
3707 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
3708 * tree-mudflap.c (mf_varname_tree): Likewise.
3709 * tree-pretty-print.c (pp_tree_identifier): Rename from
3710 pp_base_tree_identifier.
3711 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
3712 Declare as function.
3714 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3716 * pretty-print.h (pp_bar_bar): New.
3717 (pp_ampersand_ampersand): Likewise.
3718 (pp_less_equal): Likewise.
3719 (pp_greater_equal): Likewise.
3720 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
3721 printer functions instead of pp_string or operators and punctuators.
3722 (dump_gimple_call): Likewise.
3723 (dump_gimple_omp_for): Likewise.
3724 (dump_gimple_transaction): Likewise.
3725 (dump_gimple_phi): Likewise.
3726 (pp_gimple_stmt_1): Likewise.
3727 * sched-vis.c (print_insn): Likewise.
3728 * tree-mudflap.c (mf_varname_tree): Likewise.
3729 * tree-pretty-print.c (dump_block_node): Likewise.
3730 (dump_generic_node): Likewise.
3732 2013-08-02 Jan Hubicka <jh@suse.cz>
3734 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
3736 * lto-streamer-out.c (tree_is_indexable): Results decls and
3737 parm decls are not indexable.
3738 (DFS_write_tree_body): Do not follow args and results.
3739 (hash_tree): Likewise.
3740 (output_functions): Rearrange so struct function is needed
3741 only when real body is output; be able to also ouptut abstract
3742 functions; output DECL_ARGUMENTS and DECL_RESULT.
3743 (lto_output): When not in WPA, ale store abstract functions.
3744 (write_symbol): Do not care about RESULT_DECL.
3745 (output_symbol_p): Handle correctly sbtract decls.
3746 * lto-streamer-in.c (input_function): Rearrange so struct
3747 function can be NULL at entry; allow streaming of
3748 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
3749 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
3750 sanity check during LTO.
3751 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
3752 RESULT_DECl and DECL_ARGUMENTS.
3753 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
3756 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3758 * pretty-print.h (pp_underscore): New.
3760 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
3761 printer functions instead of pp_character.
3762 (dump_binary_rhs): Likewise.
3763 (dump_ternary_rhs): Likewise.
3764 (dump_gimple_call_args): Likewise.
3765 (pp_points_to_solution): Likewise.
3766 (dump_gimple_call): Likewise.
3767 (dump_gimple_switch): Likewise.
3768 (dump_gimple_cond): Likewise.
3769 (dump_gimple_bind): Likewise.
3770 (dump_gimple_try): Likewise.
3771 (dump_gimple_omp_for): Likewise.
3772 (dump_gimple_omp_continue): Likewise.
3773 (dump_gimple_omp_single): Likewise.
3774 (dump_gimple_omp_sections): Likewise.
3775 (dump_gimple_omp_block): Likewise.
3776 (dump_gimple_omp_critical): Likewise.
3777 (dump_gimple_transaction): Likewise.
3778 (dump_gimple_asm): Likewise.
3779 (dump_gimple_phi): Likewise.
3780 (dump_gimple_omp_parallel): Likewise.
3781 (dump_gimple_omp_task): Likewise.
3782 (dump_gimple_omp_atomic_load): Likewise.
3783 (dump_gimple_omp_atomic_store): Likewise.
3784 (dump_gimple_mem_ops): Likewise.
3785 (pp_gimple_stmt_1): Likewise.
3786 (pp_cfg_jump): Likewise.
3787 (dump_implicit_edges): Likewise.
3788 (gimple_dump_bb_for_graph): Likewise.
3789 * graph.c (draw_cfg_node): Likewise.
3790 * langhooks.c (lhd_print_error_function): Likewise.
3791 * sched-vis.c (print_exp): Likewise.
3792 (print_value): Likewise.
3793 (print_pattern): Likewise.
3794 (print_insn): Likewise.
3795 (rtl_dump_bb_for_graph): Likewise.
3796 * tree-pretty-print.c (dump_function_declaration): Likewise.
3797 (dump_array_domain): Likewise.
3798 (dump_omp_clause): Likewise.
3799 (dump_location): Likewise.
3800 (dump_generic_node): Likewise.
3801 (print_struct_decl): Likewise.
3802 * diagnostic.c (diagnostic_show_locus): Use pp_space.
3804 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3806 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
3807 candidate table when replacing a candidate statement.
3808 (replace_rhs_if_not_dup): Likewise.
3809 (replace_one_candidate): Likewise.
3811 2013-08-02 Jan Hubicka <jh@suse.cz>
3812 Martin Liska <marxin.liska@gmail.com>
3814 * cgraphunit.c (add_new_function): Fix logic when adding from
3816 (assemble_thunk): Rename to ...
3817 (expand_thunk); .. this one; export; get it working with
3818 general functions; make produced gimple valid.
3819 * cgraph.h (expand_thunk): Declare.
3821 2013-08-02 Jan Hubicka <jh@suse.cz>
3823 * ipa-cp.c (gather_context_independent_values): Use
3824 ipa_get_param_move_cost.
3825 (get_replacement_map): Remove PARAM; move parameter folding
3827 (create_specialized_node): Update.
3828 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
3829 assert that we have gimple body; update move_cost.
3830 (count_formal_params): Assert that we have gimple body.
3831 (ipa_dump_param): New function.
3832 (ipa_alloc_node_params): Break out from ...
3833 (ipa_initialize_node_params): ... here.
3834 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
3835 (ipa_write_node_info): Stream move costs.
3836 (ipa_read_node_info): Read move costs.
3837 (ipa_update_after_lto_read): Do not recompute node params.
3838 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
3839 (ipa_get_param): Check we are not in WPA.
3840 (ipa_get_param_move_cost): New.
3841 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
3842 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
3843 parm numbers to be present.
3845 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3847 PR rtl-optimization/58048
3848 * lra-constraints.c (process_alt_operands): Don't check asm
3849 operand on register.
3851 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
3853 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
3854 the implied StoreLoad barrier for atomic operations if before.
3856 2013-08-02 Jan Hubicka <jh@suse.cz>
3857 Martin Liska <marxin.liska@gmail.com>
3859 * cgraph.c (cgraph_function_body_availability): Do not check
3861 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
3862 symtab_node_availability): Declare.
3863 * ipa.c (can_replace_by_local_alias): New.
3864 (function_and_variable_visibility): Use it.
3865 * symtab.c (symtab_for_node_and_aliases,
3866 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
3868 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3870 PR rtl-optimization/57963
3871 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
3872 (lra_constraints): Use them.
3874 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
3876 * config/arm/types.md (define_attr "type"): Add "load_acq"
3878 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
3880 (cortex_a53_store1): Likewise.
3882 2013-08-01 Jan Hubicka <jh@suse.cz>
3884 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
3885 partitions are not needed.
3887 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
3889 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
3890 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
3891 MAYBE_NON_Q_CLASS_P where appropriate.
3893 2013-08-01 Jan Hubicka <jh@suse.cz>
3895 * cgraph.h (release_function_body): Declare.
3896 * tree.c (free_lang_data_in_decl): Free, parameters and return values
3897 of unused delcarations.
3899 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3901 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3902 RTL form when subtracting a constant.
3904 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3906 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3907 Generate canonical plus rtx with negated immediate instead of minus
3909 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3911 2013-08-01 Jan Hubicka <jh@suse.cz>
3913 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3914 (cgraph_release_function_body): Likewise.
3915 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3916 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3917 to used_as_abstract_origin.
3918 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3919 symbols used as abstract origins.
3920 * cgraphunit.c (analyze_functions): Update.
3921 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3922 used_as_abstract_origin.
3923 * tree-inline.c (tree_function_versioning): Update
3924 used_as_abstract_origin; be ready for DECL_RESULT and
3925 DECL_ARGUMENTS to be NULL.
3927 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3928 for abstract functions.
3929 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3932 2013-08-01 Jan Hubicka <jh@suse.cz>
3934 * profile.c (compute_value_histograms): Fix thinko.
3936 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3938 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3939 aarch-common-protos.h to extra_headers.
3940 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3941 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3942 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3944 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3946 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3947 Include "../arm/types.md". Define "type" attribute for all patterns.
3948 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3951 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3953 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3954 to support power8 load fusion.
3955 (fusion_gpr_mem_load): Likewise.
3957 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3959 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3960 declarations for power8 load fusion.
3961 (emit_fusion_gpr_load): Likewise.
3963 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3964 tuning for power8, turn on fusion mode by default. Turn on sign
3965 extending fusion mode if normal fusion mode is on, and we are at
3967 (fusion_gpr_load_p): New function, return true if we can fuse an
3968 addis instruction with a dependent load to a GPR.
3969 (emit_fusion_gpr_load): Emit the instructions for power8 load
3972 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3973 (VSX load fusion peepholes): New peepholes to fuse together an
3974 addi instruction with a VSX load instruction.
3976 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
3977 peepholes to fuse an addis instruction with a load to a GPR base
3978 register. If we are supporting sign extending fusions, convert
3979 sign extending loads to zero extending loads and add an explicit
3982 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3984 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
3985 aarch-common-protos.h to extra_headers.
3986 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3987 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
3988 (arm_early_store_addr_dep): Likewise.
3989 (arm_no_early_alu_shift_dep): Likewise.
3990 (arm_no_early_alu_shift_value_dep): Likewise.
3991 (arm_no_early_mul_dep): Likewise.
3992 (arm_no_early_store_addr_dep): Likewise.
3993 (arm_mac_accumulator_is_mul_result): Likewise.
3994 (arm_mac_accumulator_is_result): Likewise.
3995 * config/arm/aarch-common.c: ... here. New file.
3996 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
3998 (arm_early_store_addr_dep): Likewise.
3999 (arm_no_early_alu_shift_dep): Likewise.
4000 (arm_no_early_alu_shift_value_dep): Likewise.
4001 (arm_no_early_mul_dep): Likewise.
4002 (arm_no_early_store_addr_dep): Likewise.
4003 (arm_mac_accumulator_is_mul_result): Likewise.
4004 (arm_mac_accumulator_is_result): Likewise.
4005 * config/arm/aarch-common-protos.h: ... here. New file.
4006 * config/arm/t-arm (aarch-common.o): Define.
4008 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4010 * config/arm/arm.md: Include new file "types.md".
4011 (define_attr "type"): Move from here to ...
4012 (define_attr "mul32"): Likewise.
4013 (define_attr "mul64"): Likewise.
4014 * config/arm/types.md: ... here. New file.
4016 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
4018 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
4019 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
4021 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4023 * gen-pass-instances.awk: Fix offset of substr().
4025 2013-07-31 David Malcolm <dmalcolm@redhat.com>
4027 * Makefile.in (pass-instances.def): New.
4028 (passes.o): Replace dependency on passes.def with one on
4031 * gen-pass-instances.awk: New.
4033 * passes.c (pass_manager::pass_manager): Use pass-instances.def
4034 rather than passes.def, updating local definition of NEXT_PASS
4035 macro to add an extra NUM parameter (currently unused).
4037 2013-07-30 David Malcolm <dmalcolm@redhat.com>
4039 * Makefile.in (PASS_MANAGER_H): New.
4040 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
4041 (passes.o): Likewise.
4042 (statistics.o): Likewise.
4043 (cgraphunit.o): Likewise.
4044 (context.o): Depend on PASS_MANAGER_H.
4046 * pass_manager.h: New.
4048 * cgraphunit.c (cgraph_add_new_function): Update for moves
4049 of globals to fields of pass_manager.
4050 (analyze_function): Likewise.
4051 (expand_function): Likewise.
4052 (ipa_passes): Likewise.
4053 (compile): Likewise.
4055 * context.c (context::context): New.
4056 * context.h (context::context): New.
4057 (context::get_passes): New.
4058 (context::passes_): New.
4060 * lto-cgraph.c (input_node): Update for moves of globals to
4061 fields of pass_manager.
4063 * passes.c (all_passes): Remove, in favor of a field of the
4064 same name within the new class pass_manager.
4065 (all_small_ipa_passes): Likewise.
4066 (all_lowering_passes): Likewise.
4067 (all_regular_ipa_passes): Likewise.
4068 (all_late_ipa_passes): Likewise.
4069 (all_lto_gen_passes): Likewise.
4070 (passes_by_id): Likewise.
4071 (passes_by_id_size): Likewise.
4072 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4073 the new class pass_manager.
4074 (set_pass_for_id): Convert to...
4075 (pass_manager::set_pass_for_id): ...method.
4076 (get_pass_for_id): Convert to...
4077 (pass_manager::get_pass_for_id): ...method.
4078 (register_one_dump_file): Move body of implementation into...
4079 (pass_manager::register_one_dump_file): ...here.
4080 (register_dump_files_1): Convert to...
4081 (pass_manager::register_dump_files_1): ...method.
4082 (register_dump_files): Convert to...
4083 (pass_manager::register_dump_files): ...method.
4084 (create_pass_tab): Update for moves of globals to fields of
4086 (dump_passes): Move body of implementation into...
4087 (pass_manager::dump_passes): ...here.
4088 (register_pass): Move body of implementation into...
4089 (pass_manager::register_pass): ...here.
4090 (init_optimization_passes): Convert into...
4091 (pass_manager::pass_manager): ...constructor for new
4092 pass_manager class, and initialize the pass_lists array.
4093 (check_profile_consistency): Update for moves of globals to
4094 fields of pass_manager.
4095 (dump_profile_report): Move body of implementation into...
4096 (pass_manager::dump_profile_report): ...here.
4097 (ipa_write_summaries_1): Update for moves of pass lists from
4098 being globals to fields of pass_manager.
4099 (ipa_write_optimization_summaries): Likewise.
4100 (ipa_read_summaries): Likewise.
4101 (ipa_read_optimization_summaries): Likewise.
4102 (execute_all_ipa_stmt_fixups): Likewise.
4104 * statistics.c (statistics_fini): Update for moves of globals to
4105 fields of pass_manager.
4107 * toplev.c (general_init): Replace call to
4108 init_optimization_passes with construction of the pass_manager
4111 * tree-pass.h (all_passes): Remove, in favor of a field of the
4112 same name within the new class pass_manager.
4113 (all_small_ipa_passes): Likewise.
4114 (all_lowering_passes): Likewise.
4115 (all_regular_ipa_passes): Likewise.
4116 (all_lto_gen_passes): Likewise.
4117 (all_late_ipa_passes): Likewise.
4118 (passes_by_id): Likewise.
4119 (passes_by_id_size): Likewise.
4120 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4121 the new class pass_manager.
4122 (get_pass_for_id): Remove.
4124 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4126 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
4129 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4131 * arm.md (mulhi3): New expand pattern.
4133 2013-07-30 Jan Hubicka <jh@suse.cz>
4134 Martin Liska <marxin.liska@gmail.com>
4136 * profile.c (compute_value_histograms): Do not ICE when
4137 there is mismatch only on some counters.
4139 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4141 PR rtl-optimization/57637
4142 * function.c (move_insn_for_shrink_wrap): Also check the
4143 GEN set of the LIVE problem for the liveness analysis
4144 if it exists, otherwise give up.
4146 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4148 PR tree-optimization/57993
4149 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
4150 replaced statement in the candidate table.
4151 (phi_add_costs): Return infinite cost when the hidden basis does
4152 not dominate all phis on which the candidate is dependent.
4153 (replace_one_candidate): Record replaced statement in the
4156 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
4158 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
4159 (ashlv2si3): New expander.
4160 (*ashlv2si3_i): New define_insn_and_split.
4161 * predicates.md (float_operation): Allow patterns with three
4164 PR rtl-optimization/58021
4165 * mode-switching.c (create_pre_exit): Always split off preceding
4166 insns if we are not at the basic block head.
4168 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
4170 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
4171 (UCLIBC_DYNAMIC_LINKER): New macro.
4172 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
4174 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
4175 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
4177 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
4178 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
4179 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
4180 for SF and DF modes. Use ieee_quad_format for TF mode.
4181 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
4182 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
4183 (mips_option_override): Handle `-mnan=legacy'.
4184 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
4185 `-mabs=2008' and `-mnan=2008'.
4186 (OPTION_DEFAULT_SPECS): Add "nan" default.
4187 (ASM_SPEC): Handle `-mnan='.
4188 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
4189 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
4190 comment accordingly.
4191 (neg<mode>2): Likewise.
4192 * config/mips/mips.opt (mabs, mnan): New options.
4193 * doc/install.texi (Configuration): Document `--with-nan=' option.
4194 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
4196 (MIPS Options): Document them.
4197 * config.gcc <mips*-*-*>: Handle `--with-nan='.
4198 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
4199 * configure: Regenerate.
4200 * config.in: Regenerate.
4202 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4204 * config/i386/i386.md (float post-reload splitters): Do not check
4205 for subregs of SSE registers.
4207 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4208 H.J. Lu <hongjiu.lu@intel.com>
4212 * config/i386/i386.md (post-reload splitter
4213 to avoid partial SSE reg dependency stalls): New pattern.
4215 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
4217 * config/s390/s390.md ("movcc"): Swap load and store instructions.
4219 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
4221 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
4222 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
4224 2013-07-26 Cary Coutant <ccoutant@google.com>
4226 * dwarf2out.c (die_checksum_ordered): Don't include template
4227 instantiations in signature.
4228 (is_template_parameter): New function.
4229 (is_template_instantiation): New function.
4230 (generate_skeleton_bottom_up): Don't include template instantiations
4232 (generate_skeleton): Likewise.
4233 (break_out_comdat_types): Move recursive call to break out nested
4235 (prune_unused_types_mark_generic_parms_dies): Call
4236 is_template_parameter.
4238 2013-07-26 Ian Bolton <ian.bolton@arm.com>
4240 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
4241 uses vector registers.
4242 * config/aarch64/iterators.md: Add attributes rtn and vas.
4244 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4245 Richard Earnshaw <richard.earnshaw@arm.com>
4247 * combine.c (simplify_comparison): Re-canonicalize operands
4249 * config/arm/arm.md (movcond_addsi): New splitter.
4251 2013-07-25 Sterling Augustine <saugustine@google.com>
4253 * dwarf2out.c (size_of_pubnames): Move code to...
4254 (include_pubname_in_output): ...here. New.
4255 (want_pubnames): Rearrange.
4256 (output_pubnames): Call include_pubname_in_output. Move assertion.
4258 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4260 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
4262 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4265 * doc/extend.texi: Remove obsolete builtins. Fix
4266 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
4268 2013-07-25 Jan Hubicka <jh@suse.cz>
4270 * cgraph.c (release_function_body): Break out from ...
4271 (cgraph_release_function_body): ... this one; also release DECL_RESULT
4273 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
4274 old_tree in the map.
4275 (create_specialized_node): Update.
4276 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
4278 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
4279 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
4280 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
4281 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
4284 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4286 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
4287 addsi3_carryin_alt2_<optab>): Correct output template.
4289 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4291 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
4292 Adjust for arm_restrict_it.
4293 Remove trailing whitespace.
4295 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
4297 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
4298 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
4300 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
4302 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
4304 PR rtl-optimization/57960
4305 * lra-constraints.c (process_alt_operands): Use the right mode
4306 when checking strict_low.
4308 2013-07-25 Jan Hubicka <jh@suse.cz>
4310 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
4311 * cgraph.c (cgraph_remove_node): Do not release function body
4312 when in cgraph streaming.
4313 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
4314 in other partitions are not considered reachable; fix handling of
4317 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4319 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
4321 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4326 * config/arm/arm.md ("*sibcall_insn): Replace use of
4327 Ss with US. Adjust output for v5 and v4t.
4328 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
4330 * config/arm/constraints.md ("Ss"): Rename to US.
4332 2013-07-25 Terry Guo <terry.guo@arm.com>
4334 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
4335 shift_add/shift_sub0/shift_sub1 RTXs.
4337 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
4338 Anton Blanchard <anton@au1.ibm.com>
4340 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
4341 (altivec_vpks<VI_char>ss): Likewise.
4342 (altivec_vpks<VI_char>us): Likewise.
4343 (altivec_vpku<VI_char>us): Likewise.
4344 (altivec_vpku<VI_char>um): Likewise.
4346 2013-07-24 David Malcolm <dmalcolm@redhat.com>
4348 Introduce context class.
4350 * Makefile.in (CONTEXT_H): New.
4351 (OBJS): Add context.o.
4352 (toplev.o): Add CONTEXT_H to dependencies.
4355 * toplev.c (general_init): Create the singleton gcc::context instance.
4361 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
4363 PR rtl-optimization/57968
4364 * mode-switching.c (create_pre_exit): Allow instructions that
4365 don't set a return register to need a non-exit mode.
4367 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4368 Anton Blanchard <anton@au1.ibm.com>
4370 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
4371 operands to vperm for little endian.
4372 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
4373 of lvsl to create the control mask for a vperm for little endian.
4375 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4376 Anton Blanchard <anton@au1.ibm.com>
4378 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
4379 two operands for little-endian.
4381 2013-07-23 Steve Ellcey <sellcey@mips.com>
4383 * config/mips/mips.c (mips_case_values_threshold): New.
4384 (TARGET_CASE_VALUES_THRESHOLD): Define.
4386 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4387 Anton Blanchard <anton@au1.ibm.com>
4389 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
4390 selection of field for vector splat in little endian mode.
4392 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4394 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
4395 expanders to rs6000.md.
4396 (ior<mode>3): Likewise.
4397 (and<mode>3): Likewise.
4398 (one_cmpl<mode>2): Likewise.
4399 (nor<mode>3): Likewise.
4400 (andc<mode>3): Likewise.
4401 (eqv<mode>3): Likewise.
4402 (nand<mode>3): Likewise.
4403 (orc<mode>3): Likewise.
4405 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
4408 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
4409 to split multi-word logical operations.
4410 (rs6000_split_logical_di): Likewise.
4411 (rs6000_split_logical): Likewise.
4413 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
4414 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
4415 and allow TImode operations in 32-bit.
4416 (vsx_and<mode>3_64bit): Likewise.
4417 (vsx_ior<mode>3_32bit): Likewise.
4418 (vsx_ior<mode>3_64bit): Likewise.
4419 (vsx_xor<mode>3_32bit): Likewise.
4420 (vsx_xor<mode>3_64bit): Likewise.
4421 (vsx_one_cmpl<mode>2_32bit): Likewise.
4422 (vsx_one_cmpl<mode>2_64bit): Likewise.
4423 (vsx_nor<mode>3_32bit): Likewise.
4424 (vsx_nor<mode>3_64bit): Likewise.
4425 (vsx_andc<mode>3_32bit): Likewise.
4426 (vsx_andc<mode>3_64bit): Likewise.
4427 (vsx_eqv<mode>3_32bit): Likewise.
4428 (vsx_eqv<mode>3_64bit): Likewise.
4429 (vsx_nand<mode>3_32bit): Likewise.
4430 (vsx_nand<mode>3_64bit): Likewise.
4431 (vsx_orc<mode>3_32bit): Likewise.
4432 (vsx_orc<mode>3_64bit): Likewise.
4434 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
4435 logical types in GPRs.
4437 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
4438 logical insns to rs6000.md, and allow TImode operations in
4440 (altivec_ior<mode>3): Likewise.
4441 (altivec_xor<mode>3): Likewise.
4442 (altivec_one_cmpl<mode>2): Likewise.
4443 (altivec_nor<mode>3): Likewise.
4444 (altivec_andc<mode>3): Likewise.
4446 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
4447 attributes for moving the 128-bit logical operations into
4449 (BOOL_REGS_OUTPUT): Likewise.
4450 (BOOL_REGS_OP1): Likewise.
4451 (BOOL_REGS_OP2): Likewise.
4452 (BOOL_REGS_UNARY): Likewise.
4453 (BOOL_REGS_AND_CR0): Likewise.
4454 (one_cmpl<mode>2): Add support for DI logical operations on
4455 32-bit, splitting the operations to 32-bit.
4459 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
4460 changes to combine the 32/64-bit code, allow logical operations on
4461 TI mode in 32-bit, and to use similar match_operator patterns like
4462 scalar mode uses. Combine the Altivec and VSX code for logical
4463 operations, and move it here.
4464 (ior<mode>3, 128-bit types): Likewise.
4465 (xor<mode>3, 128-bit types): Likewise.
4466 (one_cmpl<mode>3, 128-bit types): Likewise.
4467 (nor<mode>3, 128-bit types): Likewise.
4468 (andc<mode>3, 128-bit types): Likewise.
4469 (eqv<mode>3, 128-bit types): Likewise.
4470 (nand<mode>3, 128-bit types): Likewise.
4471 (orc<mode>3, 128-bit types): Likewise.
4472 (and<mode>3_internal): Likewise.
4473 (bool<mode>3_internal): Likewise.
4474 (boolc<mode>3_internal1): Likewise.
4475 (boolc<mode>3_internal2): Likewise.
4476 (boolcc<mode>3_internal1): Likewise.
4477 (boolcc<mode>3_internal2): Likewise.
4478 (eqv<mode>3_internal1): Likewise.
4479 (eqv<mode>3_internal2): Likewise.
4480 (one_cmpl1<mode>3_internal): Likewise.
4482 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
4484 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4485 Rename flag_stack_usage to flag_stack_usage_info.
4487 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
4489 * config/microblaze/sync.md: New file.
4490 * config/microblaze/microblaze.md: Include sync.md
4491 * config/microblaze/microblaze.c: Add print_operand 'y'.
4492 * config/microblaze/constraints.md: Add memory_contraint
4493 'Q' which is a single register.
4495 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
4497 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
4499 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
4501 * reload.c (find_reloads): Exit loop once we find this operand
4502 cannot be reloaded somehow for this alternative.
4504 * reload.c (find_reloads): Exit loop once we find a hard register.
4506 * rtlanal.c (computed_jump_p): Exit loop once we find label
4509 * i386.c (ix86_pad_returns): Exit loop after setting replace.
4511 * cfgloopmanip.c (remove_path): Exit loop after setting
4514 * gensupport.c (subst_dup): Avoid loop if code is not
4515 MATCH_DUP nor MATCH_OP_DUP.
4517 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
4519 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
4521 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4523 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
4524 true for SP_REGNUM if mode == ptr_mode.
4525 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
4526 with value R0_REGNUM + 31.
4528 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4530 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
4531 pad pointer-typed argument downward.
4533 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4535 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
4536 and __ILP32__ when the ILP32 model is in use.
4538 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4540 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
4541 (aarch64_load_symref_appropriately): In the case of
4542 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
4543 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
4544 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
4545 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
4546 if 'mode' doesn't equal to 'ptr_mode'.
4547 (aarch64_output_mi_thunk): Add an assertion on the alignment of
4548 'vcall_offset'; change to call aarch64_emit_move differently depending
4549 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
4550 to calculate the upper bound of 'vcall_offset'.
4551 (aarch64_cannot_force_const_mem): Change to also return true if
4553 (aarch64_legitimize_reload_address): In the case of large
4554 displacements, add new local variable 'xmode' and an assertion
4555 based on it; change to use 'xmode' to generate the new rtx and
4557 (aarch64_asm_trampoline_template): Change to generate the template
4558 differently depending on TARGET_ILP32 or not; change to use
4559 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
4560 (aarch64_trampoline_size): Removed.
4561 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
4562 and replace immediate literals with it. Change to use 'ptr_mode'
4563 instead of 'DImode' and call convert_memory_address if the mode
4564 of 'fnaddr' doesn't equal to 'ptr_mode'.
4565 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
4567 (aarch64_elf_asm_destructor): Likewise.
4568 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
4569 on TARGET_ILP32 instead of aarch64_trampoline_size.
4570 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
4571 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
4572 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
4573 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
4574 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
4575 (storewb_pair<GPI:mode>_<P:mode>): ... this.
4576 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
4577 depending on the value of 'mode'.
4578 (add_losym_<mode>): New.
4579 (ldr_got_small_<mode>): New, based on ldr_got_small.
4580 (ldr_got_small): Remove.
4581 (ldr_got_small_sidi): New.
4582 * config/aarch64/iterators.md (P): New.
4583 (PTR): Change to 'ptr_mode' in the condition.
4585 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4587 * config.gcc (aarch64*-*-*): Support --with-abi.
4588 (aarch64*-*-elf): Support --with-multilib-list.
4589 (aarch64*-*-linux*): Likewise.
4590 (supported_defaults): Add abi to aarch64*-*-*.
4591 * configure.ac: Mention AArch64 for --with-multilib-list.
4592 * configure: Re-generated.
4593 * config/aarch64/biarchilp32.h: New file.
4594 * config/aarch64/biarchlp64.h: New file.
4595 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
4597 (MULTILIB_DEFAULTS): Ditto.
4598 (DRIVER_SELF_SPECS): Ditto.
4599 (ASM_SPEC): Update to also substitute -mabi.
4600 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
4601 file whose name depends on -mabi= and -mbig-endian.
4602 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
4604 (POINTER_SIZE): New define.
4605 (POINTERS_EXTEND_UNSIGNED): Ditto.
4606 (enum aarch64_abi_type): New enumeration tag.
4607 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
4608 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
4609 (TARGET_ILP32): New define.
4610 * config/aarch64/aarch64.opt (mabi): New.
4612 (ilp32, lp64): New values for -mabi.
4613 * config/aarch64/t-aarch64 (comma): New define.
4614 (MULTILIB_OPTIONS): Ditto.
4615 (MULTILIB_DIRNAMES): Ditto.
4616 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
4617 * doc/invoke.texi: Document -mabi for AArch64.
4619 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
4621 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
4623 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4624 Anton Blanchard <anton@au1.ibm.com>
4626 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
4627 endianness when selecting field to splat.
4629 2013-07-22 Eric Christopher <echristo@gmail.com>
4631 * dwarf2out.c (die_odr_checksum): New function to use
4632 CHECKSUM_ macros and ULEB128 for DIE tag.
4633 (generate_type_signature): Use.
4635 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
4637 * config.gcc (sparc*-*-*): Accept leon3 processor.
4638 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
4639 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
4640 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
4641 * config/sparc/sparc.opt (enum processor_type): Add leon3.
4642 (mfix-ut699): Adjust comment.
4643 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
4644 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
4645 (CPP_CPU_SPEC): Likewise.
4646 (ASM_CPU_SPEC): Likewise.
4647 * config/sparc/sparc.c (leon3_cost): New constant.
4648 (sparc_option_override): Add leon3 support.
4649 (mem_ref): New function.
4650 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
4651 (sparc_do_work_around_errata): Look into the instruction in the delay
4652 slot and adjust accordingly. Add fix for the data cache nullify issues
4653 of the UT699. Change insertion position for the NOP.
4654 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
4655 (leon3_load): New reservation.
4656 (leon_store): Bump latency to 2.
4657 (grfpu): New automaton.
4658 (grfpu_alu): New unit.
4659 (grfpu_ds): Likewise.
4660 (leon_fp_alu): Adjust.
4661 (leon_fp_mult): Delete.
4662 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
4663 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
4664 * config/sparc/sparc.md (cpu): Add leon3.
4665 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
4667 (atomic_test_and_set): Likewise.
4670 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
4672 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
4673 default for R5900 targets.
4674 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
4675 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
4676 * config/mips/mips.c (mips_option_override): Report an error for
4677 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
4678 for -march=r5900 -mhard-float.
4680 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
4682 * df-problems.c (can_move_insns_across): Exit loop once we
4683 find a non-fixed, non-global register.
4685 * ipa-pure-const.c (propagate_nothrow): Exit loop after
4688 * omega.c (omega_eliminate_red): Break after setting red_found.
4689 (omega_problem_has_red_equations): Similarly after setting found.
4690 (omega_query_variable): Similarly after setting coupled.
4692 2013-07-22 Marek Polacek <polacek@redhat.com>
4694 * gimplify.c: Don't include gimple.h twice.
4696 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4698 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
4699 instead of TARGET_THUMB1.
4700 (Pz): New constraint.
4701 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
4703 (compare_negsi_si): Likewise.
4704 (compare_addsi2_op0): Likewise.
4705 (compare_addsi2_op1): Likewise.
4706 (addsi3_carryin_<optab>): Likewise.
4707 (addsi3_carryin_alt2_<optab>): Likewise.
4708 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
4709 for arm_restrict_it.
4710 (subsi3_carryin): Likewise.
4711 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
4712 (minmax_arithsi): Disable for arm_restrict_it.
4713 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
4714 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
4715 (satsi_<SAT:code>_shift): Likewise.
4716 (arm_shiftsi3): Add alternative for 16-bit encoding.
4717 (arm32_movhf): Disable for arm_restrict_it.
4718 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
4719 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
4721 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
4723 * config/arm/arm.md (attribute "insn"): Delete.
4724 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
4725 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
4726 (not_shiftsi): Update for attribute change.
4727 (not_shiftsi_compare0): Likewise.
4728 (not_shiftsi_compare0_scratch): Likewise.
4729 (arm_one_cmplsi2): Likewise.
4730 (thumb1_one_cmplsi2): Likewise.
4731 (notsi_compare0): Likewise.
4732 (notsi_compare0_scratch): Likewise.
4733 (thumb1_movdi_insn): Likewise.
4734 (arm_movsi_insn): Likewise.
4735 (movhi_insn_arch4): Likewise.
4736 (movhi_bytes): Likewise.
4737 (arm_movqi_insn): Likewise.
4738 (thumb1_movqi_insn): Likewise.
4739 (arm32_movhf): Likewise.
4740 (thumb1_movhf): Likewise.
4741 (arm_movsf_soft_insn): Likewise.
4742 (thumb1_movsf_insn): Likewise.
4743 (thumb_movdf_insn): Likewise.
4744 (movsicc_insn): Likewise.
4745 (movsfcc_soft_insn): Likewise.
4746 (and_scc): Likewise.
4747 (cond_move): Likewise.
4748 (if_move_not): Likewise.
4749 (if_not_move): Likewise.
4750 (if_shift_move): Likewise.
4751 (if_move_shift): Likewise.
4752 (if_shift_shift): Likewise.
4753 (if_not_arith): Likewise.
4754 (if_arith_not): Likewise.
4755 (cond_move_not): Likewise.
4756 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
4757 (neon_mov<mode>): Likewise.
4758 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
4759 (thumb2_movsi_vfp): Likewise.
4760 (movsf_vfp): Likewise.
4761 (thumb2_movsf_vfp): Likewise.
4762 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4764 (cortexa7_older_only): Likewise.
4765 (cortexa7_younger): Likewise.
4766 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
4767 (1020alu_shift_op): Likewise.
4768 (1020alu_shift_reg_op): Likewise.
4769 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
4770 (alu_shift_op): Likewise.
4771 (alu_shift_reg_op): Likewise.
4772 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
4773 (11_alu_shift_op): Likewise.
4774 (11_alu_shift_reg_op): Likewise.
4775 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
4776 (9_alu_shift_reg_op): Likewise.
4777 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4779 (cortex_a15_alu_shift): Likewise.
4780 (cortex_a15_alu_shift_reg): Likewise.
4781 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
4782 (cortex_a5_alu_shift): Likewise.
4783 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
4785 (cortex_a53_alu_shift): Likewise.
4786 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4788 (cortex_a7_alu_reg): Likewise.
4789 (cortex_a7_alu_shift): Likewise.
4790 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4791 (cortex_a8_alu_shift): Likewise.
4792 (cortex_a8_alu_shift_reg): Likewise.
4793 (cortex_a8_mov): Likewise.
4794 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4795 (cortex_a9_dp_shift): Likewise.
4796 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
4797 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
4798 (cortex_r4_mov): Likewise.
4799 (cortex_r4_alu_shift): Likewise.
4800 (cortex_r4_alu_shift_reg): Likewise.
4801 * config/arm/fa526.md (526_alu_op): Update for attribute change.
4802 (526_alu_shift_op): Likewise.
4803 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
4804 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
4805 (626te_alu_shift_op): Likewise.
4806 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
4807 (726te_alu_op): Likewise.
4808 (726te_alu_shift_op): Likewise.
4809 (726te_alu_shift_reg_op): Likewise.
4810 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
4811 (mp626_alu_shift_op): Likewise.
4812 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
4813 (pj4_alu_e1_conds): Likewise.
4814 (pj4_alu): Likewise.
4815 (pj4_alu_conds): Likewise.
4816 (pj4_shift): Likewise.
4817 (pj4_shift_conds): Likewise.
4818 (pj4_alu_shift): Likewise.
4819 (pj4_alu_shift_conds): Likewise.
4821 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4823 * config/arm/predicates.md (shiftable_operator_strict_it):
4825 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
4826 Disable cond_exec version for arm_restrict_it.
4827 (thumb2_smaxsi3): Convert to generate cond_exec.
4828 (thumb2_sminsi3): Likewise.
4829 (thumb32_umaxsi3): Likewise.
4830 (thumb2_uminsi3): Likewise.
4831 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
4832 (thumb2_neg_abssi2): Likewise.
4833 (thumb2_mov_scc): Add alternative for 16-bit encoding.
4834 (thumb2_movsicc_insn): Adjust alternatives.
4835 (thumb2_mov_negscc): Disable for arm_restrict_it.
4836 (thumb2_mov_negscc_strict_it): New pattern.
4837 (thumb2_mov_notscc_strict_it): New pattern.
4838 (thumb2_mov_notscc): Disable for arm_restrict_it.
4839 (thumb2_ior_scc): Likewise.
4840 (thumb2_ior_scc_strict_it): New pattern.
4841 (thumb2_cond_move): Adjust for arm_restrict_it.
4842 (thumb2_cond_arith): Disable for arm_restrict_it.
4843 (thumb2_cond_arith_strict_it): New pattern.
4844 (thumb2_cond_sub): Adjust for arm_restrict_it.
4845 (thumb2_movcond): Likewise.
4846 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
4847 (thumb2_zero_extendhisi2_v6): Likewise.
4848 (thumb2_zero_extendqisi2_v6): Likewise.
4849 (orsi_notsi_si): Likewise.
4850 (orsi_not_shiftsi_si): Likewise.
4852 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
4854 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
4855 instruction sequence is 1 byte shorter.
4857 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
4859 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
4860 it is not needed after split.
4862 2013-07-20 Iain Sandoe <iain@codesourcery.com>
4865 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
4866 second label for nonlocal goto receivers. Don't output pic base labels
4867 unless we're producing PIC; mark that action unreachable().
4868 (ix86_save_reg): If the function contains a nonlocal label, save the
4870 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
4871 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
4872 (update_pic_label_number_if_needed): New.
4873 (machopic_output_function_base_name): Adjust for nonlocal receiver
4875 (machopic_should_output_picbase_label): New.
4876 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
4877 (nonlocal_goto_receiver): New insn and split.
4879 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
4881 * config/aarch64/aarch64-builtins.c
4882 (aarch64_fold_builtin): Fold abs in all modes.
4883 * config/aarch64/aarch64-simd-builtins.def
4884 (abs): Enable for all modes.
4885 * config/aarch64/arm_neon.h
4886 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
4887 (vabs_f64): Add missing intrinsic.
4889 2013-07-19 Ian Bolton <ian.bolton@arm.com>
4891 * config/aarch64/arm_neon.h (vabs_s64): New function
4893 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
4896 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
4897 * config/avr/avr.md (adjust_len): Add `round'.
4898 * config/avr/avr-protos.h (avr_out_round): New prototype.
4899 (avr_out_plus): Add `out_label' argument.
4900 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4901 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4902 Handle the case where `insn' is just a pattern.
4903 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4904 (avr_out_round): New function.
4905 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4907 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4909 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4910 Add check for flag_stack_usage to handle -fstack-usage support
4912 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4914 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4915 interaction for new Power8 flags and VSX.
4917 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4920 * tree-inline.c (expand_call_inline): Emit errors during
4921 early_inlining only if optimization is not turned on.
4923 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4927 * passes.c (init_optimization_passes): Move the construction of
4928 the pass hierarchy into a new passes.def file.
4930 * Makefile.in (passes.o): Add dependency on passes.def.
4932 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4934 * passes.c (init_optimization_passes): Introduce macros for
4935 constructing the tree of passes (INSERT_PASSES_AFTER,
4936 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4937 TERMINATE_PASS_LIST).
4939 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4940 Wei Mi <wmi@google.com>
4942 PR rtl-optimization/57878
4943 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4945 (reload_pseudo_compare_func): Check nregs first for reload
4948 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4950 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4952 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4954 * read-rtl.c (validate_const_int): Once an invalid character is
4955 seen, quit the loop.
4957 * gengtype.c (write_roots): Similarly once we find the "deletable"
4958 or "if_marked" option.
4960 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4962 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4963 "xtab" and "sat". Move value "clz" from here to ...
4964 (attriubte "type"): ... here.
4965 (satsi_<SAT:code>): Delete "insn" attribute.
4966 (satsi_<SAT:code>_shift): Likewise.
4967 (arm_zero_extendqisi2addsi): Likewise.
4968 (arm_extendqisi2addsi): Likewise.
4969 (clzsi2): Update for attribute changes.
4970 (rbitsi2): Likewise.
4971 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4973 (arm_usatsihi): Likewise.
4974 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4976 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4978 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
4979 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
4980 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
4981 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
4982 in alphabetical order.
4983 (attribute "core_cycles"): Update for attribute changes.
4984 (arm_addsi3): Likewise.
4985 (addsi3_compare0): Likewise.
4986 (addsi3_compare0_scratch): Likewise.
4987 (addsi3_compare_op1): Likewise.
4988 (addsi3_compare_op2): Likewise.
4989 (compare_addsi2_op0): Likewise.
4990 (compare_addsi2_op1): Likewise.
4991 (addsi3_carryin_shift_<optab>): Likewise.
4992 (subsi3_carryin_shift): Likewise.
4993 (rsbsi3_carryin_shift): Likewise.
4994 (arm_subsi3_insn): Likewise.
4995 (subsi3_compare0): Likewise.
4996 (subsi3_compare): Likewise.
4997 (arm_andsi3_insn): Likewise.
4998 (thumb1_andsi3_insn): Likewise.
4999 (andsi3_compare0): Likewise.
5000 (andsi3_compare0_scratch): Likewise.
5001 (zeroextractsi_compare0_scratch
5002 (andsi_not_shiftsi_si): Likewise.
5003 (iorsi3_insn): Likewise.
5004 (iorsi3_compare0): Likewise.
5005 (iorsi3_compare0_scratch): Likewise.
5006 (arm_xorsi3): Likewise.
5007 (thumb1_xorsi3_insn): Likewise.
5008 (xorsi3_compare0): Likewise.
5009 (xorsi3_compare0_scratch): Likewise.
5010 (satsi_<SAT:code>_shift): Likewise.
5012 (arm_shiftsi3): Likewise.
5013 (shiftsi3_compare0): Likewise.
5014 (not_shiftsi): Likewise.
5015 (not_shiftsi_compare0): Likewise.
5016 (not_shiftsi_compare0_scratch): Likewise.
5017 (arm_one_cmplsi2): Likewise.
5018 (thumb_one_complsi2): Likewise.
5019 (notsi_compare0): Likewise.
5020 (notsi_compare0_scratch): Likewise.
5021 (thumb1_zero_extendhisi2): Likewise.
5022 (arm_zero_extendhisi2): Likewise.
5023 (arm_zero_extendhisi2_v6): Likewise.
5024 (arm_zero_extendhisi2addsi): Likewise.
5025 (thumb1_zero_extendqisi2): Likewise.
5026 (thumb1_zero_extendqisi2_v6): Likewise.
5027 (arm_zero_extendqisi2): Likewise.
5028 (arm_zero_extendqisi2_v6): Likewise.
5029 (arm_zero_extendqisi2addsi): Likewise.
5030 (thumb1_extendhisi2): Likewise.
5031 (arm_extendhisi2): Likewise.
5032 (arm_extendhisi2_v6): Likewise.
5033 (arm_extendqisi): Likewise.
5034 (arm_extendqisi_v6): Likewise.
5035 (arm_extendqisi2addsi): Likewise.
5036 (thumb1_extendqisi2): Likewise.
5037 (thumb1_movdi_insn): Likewise.
5038 (arm_movsi_insn): Likewise.
5039 (movsi_compare0): Likewise.
5040 (movhi_insn_arch4): Likewise.
5041 (movhi_bytes): Likewise.
5042 (arm_movqi_insn): Likewise.
5043 (thumb1_movqi_insn): Likewise.
5044 (arm32_movhf): Likewise.
5045 (thumb1_movhf): Likewise.
5046 (arm_movsf_soft_insn): Likewise.
5047 (thumb1_movsf_insn): Likewise.
5048 (movdf_soft_insn): Likewise.
5049 (thumb_movdf_insn): Likewise.
5050 (arm_cmpsi_insn): Likewise.
5051 (cmpsi_shiftsi): Likewise.
5052 (cmpsi_shiftsi_swp): Likewise.
5053 (arm_cmpsi_negshiftsi_si): Likewise.
5054 (movsicc_insn): Likewise.
5055 (movsfcc_soft_insn): Likewise.
5056 (arith_shiftsi): Likewise.
5057 (arith_shiftsi_compare0
5058 (arith_shiftsi_compare0_scratch
5059 (sub_shiftsi): Likewise.
5060 (sub_shiftsi_compare0
5061 (sub_shiftsi_compare0_scratch
5062 (and_scc): Likewise.
5063 (cond_move): Likewise.
5064 (if_plus_move): Likewise.
5065 (if_move_plus): Likewise.
5066 (if_move_not): Likewise.
5067 (if_not_move): Likewise.
5068 (if_shift_move): Likewise.
5069 (if_move_shift): Likewise.
5070 (if_shift_shift): Likewise.
5071 (if_not_arith): Likewise.
5072 (if_arith_not): Likewise.
5073 (cond_move_not): Likewise.
5074 (thumb1_ashlsi3): Set type attribute.
5075 (thumb1_ashrsi3): Likewise.
5076 (thumb1_lshrsi3): Likewise.
5077 (thumb1_rotrsi3): Likewise.
5078 (shiftsi3_compare0_scratch): Likewise.
5079 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
5080 (neon_mov<mode>): Likewise.
5081 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
5083 (thumb2_movsi_insn): Likewise.
5084 (thumb2_cmpsi_neg_shiftsi): Likewise.
5085 (thumb2_extendqisi_v6): Likewise.
5086 (thumb2_zero_extendhisi2_v6): Likewise.
5087 (thumb2_zero_extendqisi2_v6): Likewise.
5088 (thumb2_shiftsi3_short): Likewise.
5089 (thumb2_addsi3_compare0_scratch): Likewise.
5090 (orsi_not_shiftsi_si): Likewise.
5091 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5092 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
5094 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
5095 (1020alu_shift_op): Likewise.
5096 (1020alu_shift_reg_op): Likewise.
5097 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
5098 (alu_shift_op): Likewise.
5099 (alu_shift_reg_op): Likewise.
5100 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
5101 (11_alu_shift_op): Likewise.
5102 (11_alu_shift_reg_op): Likewise.
5103 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
5104 (9_alu_shift_reg_op): Likewise.
5105 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5107 (cortex_a15_alu_shift): Likewise.
5108 (cortex_a15_alu_shift_reg): Likewise.
5109 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
5111 (cortex_a5_alu_shift): Likewise.
5112 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
5114 (cortex_a53_alu_shift): Likewise.
5115 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5117 (cortex_a7_alu_reg): Likewise.
5118 (cortex_a7_alu_shift): Likewise.
5119 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
5121 (cortex_a8_alu_shift): Likewise.
5122 (cortex_a8_alu_shift_reg): Likewise.
5123 (cortex_a8_mov): Likewise.
5124 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
5125 (cortex_a9_dp_shift): Likewise.
5126 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
5128 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
5130 (cortex_r4_mov): Likewise.
5131 (cortex_r4_alu_shift): Likewise.
5132 (cortex_r4_alu_shift_reg): Likewise.
5133 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
5134 (526_alu_shift_op): Likewise.
5135 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
5136 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
5137 (626te_alu_shift_op): Likewise.
5138 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
5139 (726te_alu_op): Likewise.
5140 (726te_alu_shift_op): Likewise.
5141 (726te_alu_shift_reg_op): Likewise.
5142 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
5143 (mp626_alu_shift_op): Likewise.
5144 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
5145 (pj4_alu_e1_conds): Likewise.
5146 (pj4_alu): Likewise.
5147 (pj4_alu_conds): Likewise.
5148 (pj4_shift): Likewise.
5149 (pj4_shift_conds): Likewise.
5150 (pj4_alu_shift): Likewise.
5151 (pj4_alu_shift_conds): Likewise.
5152 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5154 (cortexa7_older_only): Likewise.
5155 (cortexa7_younger): Likewise.
5157 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5159 * ipa-pure-const.c (generate_summary): Rename to...
5160 (pure_const_generate_summary): ... this.
5162 2013-07-17 Iain Sandoe <iain@codesourcery.com>
5164 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
5166 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
5169 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
5172 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5174 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
5175 enabled without -march=zEC12.
5176 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
5179 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5181 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
5182 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
5183 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
5184 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
5185 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
5186 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
5187 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
5189 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
5190 and ISA_HAS_NMADD3_NMSUB3.
5191 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
5192 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
5193 (nmsub4<mode>, nmsub3<mode>): Likewise.
5194 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
5196 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5198 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
5199 TARGET_MIPS5400 checking.
5201 2013-07-16 Jakub Jelinek <jakub@redhat.com>
5202 Peter Bergner <bergner@vnet.ibm.com>
5204 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
5205 registers in the comment.
5206 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
5207 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
5208 rather than FIRST_PSEUDO_REGISTERS.
5210 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
5212 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
5213 enable extra ISA flags with TARGET_HTM.
5215 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5217 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
5220 2013-07-15 Cong Hou <congh@google.com>
5222 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
5223 in compare function for sorting.
5225 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
5227 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
5228 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
5229 * config/rs6000/rs6000.opt: Add -mhtm option.
5230 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
5231 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
5232 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
5233 __HTM__ if the HTM instructions are available.
5234 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
5235 htm_spr_reg_operand): New define_predicates.
5236 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
5237 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
5239 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
5240 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
5241 HTM builtin functions.
5242 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
5243 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
5244 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
5245 (rs6000_builtin_mask_calculate): Likewise.
5246 (rs6000_option_override_internal): Likewise.
5247 (bdesc_htm): Add new HTM builtin support.
5248 (htm_spr_num): New function.
5249 (htm_spr_regno): Likewise.
5250 (rs6000_htm_spr_icode): Likewise.
5251 (htm_expand_builtin): Likewise.
5252 (htm_init_builtins): Likewise.
5253 (rs6000_expand_builtin): Add support for HTM builtin functions.
5254 (rs6000_init_builtins): Likewise.
5255 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
5257 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
5258 (TARGET_HTM, MASK_HTM): Define macros.
5259 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
5260 (FIXED_REGISTERS): Likewise.
5261 (CALL_USED_REGISTERS): Likewise.
5262 (CALL_REALLY_USED_REGISTERS): Likewise.
5263 (REG_ALLOC_ORDER): Likewise.
5264 (enum reg_class): Likewise.
5265 (REG_CLASS_NAMES): Likewise.
5266 (REG_CLASS_CONTENTS): Likewise.
5267 (REGISTER_NAMES): Likewise.
5268 (ADDITIONAL_REGISTER_NAMES): Likewise.
5269 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
5270 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
5271 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
5272 * config/rs6000/htm.md: New file.
5273 * config/rs6000/htmintrin.h: New file.
5274 * config/rs6000/htmxlintrin.h: New file.
5276 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
5278 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5279 Define SYMBOL_TINY_GOT, update comment.
5280 * config/aarch64/aarch64.c
5281 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
5282 (aarch64_expand_mov_immediate): Likewise.
5283 (aarch64_print_operand): Likewise.
5284 (aarch64_classify_symbol): Likewise.
5285 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
5286 (ldr_got_tiny): Define.
5288 2013-07-13 Tobias Grosser <tobias@grosser.es>
5290 PR tree-optimization/54094
5291 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
5292 scheduling dimension for the parallelism check from the polyhedral
5293 information in the AST.
5294 * graphite-dependences.c (carries_deps): Do not assume the schedule is
5297 2013-07-13 Jason Merrill <jason@redhat.com>
5299 * print-tree.c (debug_vec_tree): Use debug_raw.
5300 (debug_raw (vec<tree, va_gc> &)): New.
5301 (debug_raw (vec<tree, va_gc> *)): New.
5302 * tree.h: Declare them.
5304 2013-07-13 Bin Cheng <bin.cheng@arm.com>
5306 * ifcvt.c (ifcvt_after_combine): New static variable.
5307 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
5309 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
5310 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
5311 rest_of_handle_if_after_reload): Pass new argument for if_convert.
5313 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
5315 * config/mips/mips.c (mips_expand_call): Remove empty statement.
5317 2013-07-12 Michael Matz <matz@suse.de>
5320 * convert.c (convert_to_real): Reject non-float inner types.
5322 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
5324 * config/aarch64/aarch64-protos.h
5325 (aarch64_simd_immediate_valid_for_move): Remove.
5326 * config/aarch64/aarch64.c (simd_immediate_info): New member.
5327 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
5329 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
5331 2013-07-11 Steve Ellcey <sellcey@mips.com>
5333 * config/mips/mips.c (mips_conditional_register_usage): Do not
5334 use t[0-7] registers in MIPS16 mode when optimizing for size.
5336 2013-07-11 Sriraman Tallam <tmsriram@google.com>
5338 * config/i386/i386.c (dispatch_function_versions): Fix array
5339 indexing of function_version_info to match actual_versions.
5341 2013-07-11 Teresa Johnson <tejohnson@google.com>
5343 * vec.h (struct va_gc): Move release out-of-line.
5344 (va_gc::release): Call ggc_free on released vec.
5346 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5348 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
5349 Require GOT register as additional operand in UNSPEC.
5350 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
5351 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
5352 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
5353 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
5354 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
5355 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
5356 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
5358 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
5361 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
5362 name seen by assembler/linker rather if available.
5364 2013-07-11 Andreas Schwab <schwab@suse.de>
5366 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
5368 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
5370 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
5372 2013-07-10 Joseph Myers <joseph@codesourcery.com>
5374 * doc/tm.texi.in: Move hook documentation to ....
5375 * target.def: ... here.
5377 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
5379 * doc/tm.texi: Regenerate.
5381 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
5384 * doc/invoke.texi: Document -Wconditionally-supported.
5386 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5389 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
5392 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5395 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
5396 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
5397 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
5398 Remove duplicate devices.
5399 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
5400 * config/avr/t-multilib: Regenerate.
5401 * config/avr/avr-tables.opt: Regenerate.
5402 * doc/avr-mmcu.texi: Regenerate.
5404 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5407 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
5409 2013-07-10 Graham Stott <graham.stott@btinternet.com>
5411 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
5412 the cost of MULT when optimizing for size.
5414 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5416 * config/cr16/cr16-protos.h: Don't include target.h.
5418 2013-07-09 Joseph Myers <joseph@codesourcery.com>
5420 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
5421 adjust register size for TDmode and TFmode for VSX registers.
5423 2013-07-08 Kai Tietz <ktietz@redhat.com>
5426 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
5427 hook_bool_const_tree_true.
5429 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5431 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
5432 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
5433 * config/s390/s390.md: Define FPR*_REGNUM constants.
5434 Fix FPR2_REGNUM constant (18 -> 17).
5435 ("*trunc<BFP:mode><DFP_ALL:mode>2")
5436 ("*trunc<DFP_ALL:mode><BFP:mode>2")
5437 ("trunc<BFP:mode><DFP_ALL:mode>2")
5438 ("trunc<DFP_ALL:mode><BFP:mode>2")
5439 ("*extend<BFP:mode><DFP_ALL:mode>2")
5440 ("*extend<DFP_ALL:mode><BFP:mode>2")
5441 ("extend<BFP:mode><DFP_ALL:mode>2")
5442 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
5445 2013-07-08 Graham Stott <graham.stott@btinternet.com>
5447 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
5449 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5451 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
5452 and cfun_fpr_bit_p to cfun_fpr_save_p.
5453 (s390_frame_area, s390_register_info, s390_frame_info)
5454 (s390_emit_prologue, s390_emit_epilogue)
5455 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
5457 * config/s390/s390.h: Define *_REGNUM macros for floating point
5460 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
5462 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
5464 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
5466 PR rtl-optimization/57786
5467 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
5468 and break out of the loop when it is set to false.
5470 2013-07-08 Jakub Jelinek <jakub@redhat.com>
5473 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
5474 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
5475 (const_int 63)) 0)).
5476 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
5477 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
5478 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
5480 PR rtl-optimization/57829
5481 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
5482 mask bits outside of mode are just sign-extension from mode to HWI.
5484 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5486 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
5487 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
5488 adjust_address instead of change_address to keep info about alignment.
5489 (emit_strmov): Remove.
5490 (emit_memmov): New function.
5491 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
5492 (expand_movmem_epilogue): Likewise and return updated rtx for
5494 (expand_constant_movmem_prologue): Likewise and return updated rtx for
5495 destination and source.
5496 (decide_alignment): Refactor, handle vector_loop.
5497 (ix86_expand_movmem): Likewise.
5498 (ix86_expand_setmem): Likewise.
5499 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
5501 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
5503 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
5504 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
5506 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
5508 * config/i386/sse.md (sse_movlhps): Change alternative 3
5509 of operand 2 to "m".
5511 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
5514 * config/i386/sse.md (iptr): New mode attribute.
5515 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
5516 (<sse>_vm<plusminus_insn><mode>3): Ditto.
5517 (<sse>_vmmul<mode>3): Ditto.
5518 (<sse>_vmdiv<mode>3): Ditto.
5519 (sse_vmrcpv4sf2): Ditto.
5520 (<sse>_vmsqrt<mode>2): Ditto.
5521 (sse_vmrsqrtv4sf2): Ditto.
5522 (<sse>_vm<code><mode>3): Ditto.
5523 (avx_vmcmp<mode>3): Ditto.
5524 (<sse>_vmmaskcmp<mode>3): Ditto.
5525 (<sse>_comi): Ditto.
5526 (<sse>_ucomi): Ditto.
5527 (*xop_vmfrcz_<mode>): Ditto.
5528 (*fmai_fmadd_<mode>): Ditto.
5529 (*fmai_fmsub_<mode>): Ditto.
5530 (*fmai_fnmadd_<mode>): Ditto.
5531 (*fmai_fnmsub_<mode>): Ditto.
5532 (*fma4i_vmfmadd_<mode>): Ditto.
5533 (*fma4i_vmfmsub_<mode>): Ditto.
5534 (*fma4i_vmfnmadd_<mode>): Ditto.
5535 (*fma4i_vmfnmsub_<mode>): Ditto.
5536 (*xop_vmfrcz_<mode>): Ditto.
5537 (sse_cvtps2pi): Ditto.
5538 (sse_cvttps2pi): Ditto.
5539 (sse_cvtss2si): Ditto.
5540 (sse_cvtss2si_2): Ditto.
5541 (sse_cvtss2siq_2): Ditto.
5542 (sse_cvttss2si): Ditto.
5543 (sse_cvttss2siq): Ditto.
5544 (sse_cvtsd2si): Ditto.
5545 (sse_cvtsd2si_2): Ditto.
5546 (sse_cvtsd2siq_2): Ditto.
5547 (sse_cvttsd2si): Ditto.
5548 (sse_cvttsd2siq): Ditto.
5549 (sse_cvtsd2ss): Ditto.
5550 (sse_cvtss2sd): Ditto.
5551 (avx2_pbroadcast<mode>): Ditto.
5552 (avx2_pbroadcast<mode>_1): Ditto.
5553 (*avx_vperm_broadcast_v4sf): Ditto.
5555 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
5556 (sse_movlhps): Ditto.
5557 (sse_storehps): Ditto.
5558 (sse_loadhps): Ditto.
5559 (sse_storelps): Ditto.
5560 (sse_loadlps): Ditto.
5561 (*vec_concatv4sf): Ditto.
5562 (*vec_interleave_highv2df): Ditto.
5563 (*vec_interleave_lowv2df): Ditto.
5564 (*vec_extractv2df_1_sse): Ditto.
5565 (*vec_extractv2df_0_sse): Ditto.
5566 (sse2_storelpd): Ditto.
5567 (sse2_loadlpd): Ditto.
5568 (sse2_movsd): Ditto.
5569 (*vec_concatv4si): Ditto.
5570 (vec_concatv2di): Ditto.
5572 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
5573 for Intel asm dialect.
5574 (mmx_punpcklwd): Ditto.
5575 (mmx_punpckldq): Ditto.
5577 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
5578 for intel assembler dialect.
5580 2013-07-06 Jakub Jelinek <jakub@redhat.com>
5583 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
5584 for BUILT_IN_C{LZ,LRSB}*.
5585 * tree.h (CASE_INT_FN): Add FN##IMAX case.
5586 * tree-vrp.c (extract_range_basic): Handle
5587 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
5588 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
5589 fall thru to code calling set_value*.
5590 * builtins.c (expand_builtin): Remove *IMAX cases.
5591 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
5592 if width is bigger than 2*HWI.
5594 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
5596 PR rtl-optimization/55342
5597 * lra-int.h (lra_subreg_reload_pseudos): New.
5598 * lra.c: Add undoing optional reloads to the block diagram.
5599 (lra_subreg_reload_pseudos): New.
5600 (lra_optional_reload_pseudos): Change comments.
5601 (lra): Init and clear lra_subreg_reload_pseudos. Clear
5602 lra_optional_reload_pseudos after undo transformations.
5603 * lra-assigns.c (pseudo_prefix_title): New.
5604 (lra_setup_reg_renumber): Use it.
5605 (spill_for): Ditto. Check subreg reload pseudos too.
5606 (assign_by_spills): Consider subreg reload pseudos too.
5607 * lra-constraints.c (simplify_operand_subreg): Use
5608 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
5609 (curr_insn_transform): Recognize and do optional reloads.
5610 (undo_optional_reloads): New.
5611 (lra_undo_inheritance): Call undo_optional_reloads.
5613 2013-07-05 Thomas Quinot <quinot@adacore.com>
5615 * tree-complex.c (expand_complex_operations_1): Fix typo.
5617 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
5619 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
5620 (tune_params): New member 'const vec_costs'.
5621 * config/aarch64/aarch64.c (generic_vector_cost): New.
5622 (generic_tunings): New member 'generic_vector_cost'.
5623 (aarch64_builtin_vectorization_cost): New.
5624 (aarch64_add_stmt_cost): New.
5625 (TARGET_VECTORIZE_ADD_STMT_COST): New.
5626 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
5628 2013-07-03 Jakub Jelinek <jakub@redhat.com>
5631 * config/i386/predicates.md (vsib_address_operand): Disallow
5632 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
5634 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
5637 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
5638 expand_builtin_setjmp_receiver.
5639 (expand_label): Adjust, call expand_builtin_setjmp_receiver
5640 with NULL for the label parameter.
5641 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
5642 the frame-pointer. Adjust comments.
5643 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
5644 only if LABEL is non-NULL.
5646 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
5648 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
5649 (ARM_ABI_AAPCS64): Ditto.
5651 (ARM_DEFAULT_ABI): Ditto.
5653 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
5655 * config/aarch64/aarch64-builtins.c
5656 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
5657 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
5659 * config/aarch64/aarch64-simd.md
5660 (aarch64_ld1<VALL:mode>): New.
5661 (aarch64_st1<VALL:mode>): Likewise.
5662 * config/aarch64/arm_neon.h
5663 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
5665 2013-07-02 Sriraman Tallam <tmsriram@google.com>
5667 * config/i386/i386.c (gate_insert_vzeroupper): Check if
5669 (ix86_option_override_internal):Turn on all -mavx target flags by
5670 default as they are dependent on AVX anyway.
5672 2013-07-02 Cary Coutant <ccoutant@google.com>
5674 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
5676 (loc_checksum_ordered): Likewise.
5677 (hash_loc_operands): Remove inline keyword.
5679 2013-07-02 Jakub Jelinek <jakub@redhat.com>
5681 PR tree-optimization/57741
5682 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
5683 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
5684 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
5685 Allow REAL_CST step_exprs if flag_associative_math.
5686 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
5688 2013-07-02 Ian Bolton <ian.bolton@arm.com>
5690 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
5692 2013-07-02 Ian Bolton <ian.bolton@arm.com>
5694 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
5696 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5698 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
5700 (iorsi3_insn): Likewise.
5701 (arm_xorsi3): Likewise.
5703 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
5705 * arm.md (attribute "wtype"): Delete. Move attribute values from here
5707 (attribute "type"): ... here, and prefix with "wmmx_".
5708 (attribute "core_cycles"): Update for attribute changes.
5709 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
5710 (tbcstv4hi): Likewise.
5711 (tbcstv2si): Likewise.
5712 (iwmmxt_iordi3): Likewise.
5713 (iwmmxt_xordi3): Likewise.
5714 (iwmmxt_anddi3): Likewise.
5715 (iwmmxt_nanddi3): Likewise.
5716 (iwmmxt_arm_movdi): Likewise.
5717 (iwmmxt_movsi_insn): Likewise.
5718 (mov<mode>_internal): Likewise.
5719 (and<mode>3_iwmmxt): Likewise.
5720 (ior<mode>3_iwmmxt): Likewise.
5721 (xor<mode>3_iwmmxt): Likewise.
5722 (add<mode>3_iwmmxt): Likewise.
5723 (ssaddv8qi3): Likewise.
5724 (ssaddv4hi3): Likewise.
5725 (ssaddv2si3): Likewise.
5726 (usaddv8qi3): Likewise.
5727 (usaddv4hi3): Likewise.
5728 (usaddv2si3): Likewise.
5729 (sub<mode>3_iwmmxt): Likewise.
5730 (sssubv8qi3): Likewise.
5731 (sssubv4hi3): Likewise.
5732 (sssubv2si3): Likewise.
5733 (ussubv8qi3): Likewise.
5734 (ussubv4hi3): Likewise.
5735 (ussubv2si3): Likewise.
5736 (mulv4hi3_iwmmxt): Likewise.
5737 (smulv4hi3_highpart): Likewise.
5738 (umulv4hi3_highpart): Likewise.
5739 (iwmmxt_wmacs): Likewise.
5740 (iwmmxt_wmacsz): Likewise.
5741 (iwmmxt_wmacu): Likewise.
5742 (iwmmxt_wmacuz): Likewise.
5743 (iwmmxt_clrdi): Likewise.
5744 (iwmmxt_clrv8qi): Likewise.
5745 (iwmmxt_clr4hi): Likewise.
5746 (iwmmxt_clr2si): Likewise.
5747 (iwmmxt_uavgrndv8qi3): Likewise.
5748 (iwmmxt_uavgrndv4hi3): Likewise.
5749 (iwmmxt_uavgv8qi3): Likewise.
5750 (iwmmxt_uavgv4hi3): Likewise.
5751 (iwmmxt_tinsrb): Likewise.
5752 (iwmmxt_tinsrh): Likewise.
5753 (iwmmxt_tinsrw): Likewise.
5754 (iwmmxt_textrmub): Likewise.
5755 (iwmmxt_textrmsb): Likewise.
5756 (iwmmxt_textrmuh): Likewise.
5757 (iwmmxt_textrmsh): Likewise.
5758 (iwmmxt_textrmw): Likewise.
5759 (iwmxxt_wshufh): Likewise.
5760 (eqv8qi3): Likewise.
5761 (eqv4hi3): Likewise.
5762 (eqv2si3): Likewise.
5763 (gtuv8qi3): Likewise.
5764 (gtuv4hi3): Likewise.
5765 (gtuv2si3): Likewise.
5766 (gtv8qi3): Likewise.
5767 (gtv4hi3): Likewise.
5768 (gtv2si3): Likewise.
5769 (smax<mode>3_iwmmxt): Likewise.
5770 (umax<mode>3_iwmmxt): Likewise.
5771 (smin<mode>3_iwmmxt): Likewise.
5772 (umin<mode>3_iwmmxt): Likewise.
5773 (iwmmxt_wpackhss): Likewise.
5774 (iwmmxt_wpackwss): Likewise.
5775 (iwmmxt_wpackdss): Likewise.
5776 (iwmmxt_wpackhus): Likewise.
5777 (iwmmxt_wpackwus): Likewise.
5778 (iwmmxt_wpackdus): Likewise.
5779 (iwmmxt_wunpckihb): Likewise.
5780 (iwmmxt_wunpckihh): Likewise.
5781 (iwmmxt_wunpckihw): Likewise.
5782 (iwmmxt_wunpckilb): Likewise.
5783 (iwmmxt_wunpckilh): Likewise.
5784 (iwmmxt_wunpckilw): Likewise.
5785 (iwmmxt_wunpckehub): Likewise.
5786 (iwmmxt_wunpckehuh): Likewise.
5787 (iwmmxt_wunpckehuw): Likewise.
5788 (iwmmxt_wunpckehsb): Likewise.
5789 (iwmmxt_wunpckehsh): Likewise.
5790 (iwmmxt_wunpckehsw): Likewise.
5791 (iwmmxt_wunpckelub): Likewise.
5792 (iwmmxt_wunpckeluh): Likewise.
5793 (iwmmxt_wunpckeluw): Likewise.
5794 (iwmmxt_wunpckelsb): Likewise.
5795 (iwmmxt_wunpckelsh): Likewise.
5796 (iwmmxt_wunpckelsw): Likewise.
5797 (ror<mode>3): Likewise.
5798 (ashr<mode>3_iwmmxt): Likewise.
5799 (lshr<mode>3_iwmmxt): Likewise.
5800 (ashl<mode>3_iwmmxt): Likewise.
5801 (ror<mode>3_di): Likewise.
5802 (ashr<mode>3_di): Likewise.
5803 (lshr<mode>3_di): Likewise.
5804 (ashl<mode>3_di): Likewise.
5805 (iwmmxt_wmadds): Likewise.
5806 (iwmmxt_wmaddu): Likewise.
5807 (iwmmxt_tmia): Likewise.
5808 (iwmmxt_tmiaph): Likewise.
5809 (iwmmxt_tmiabb): Likewise.
5810 (iwmmxt_tmiatb): Likewise.
5811 (iwmmxt_tmiabt): Likewise.
5812 (iwmmxt_tmiatt): Likewise.
5813 (iwmmxt_tmovmskb): Likewise.
5814 (iwmmxt_tmovmskh): Likewise.
5815 (iwmmxt_tmovmskw): Likewise.
5816 (iwmmxt_waccb): Likewise.
5817 (iwmmxt_wacch): Likewise.
5818 (iwmmxt_waccw): Likewise.
5819 (iwmmxt_waligni): Likewise.
5820 (iwmmxt_walignr): Likewise.
5821 (iwmmxt_walignr0): Likewise.
5822 (iwmmxt_walignr1): Likewise.
5823 (iwmmxt_walignr2): Likewise.
5824 (iwmmxt_walignr3): Likewise.
5825 (iwmmxt_wsadb): Likewise.
5826 (iwmmxt_wsadh): Likewise.
5827 (iwmmxt_wsadbz): Likewise.
5828 (iwmmxt_wsadhz): Likewise.
5829 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
5830 (iwmmxt_wabsdiffb): Likewise.
5831 (iwmmxt_wabsdiffh): Likewise.
5832 (iwmmxt_wabsdiffw): Likewise.
5833 (iwmmxt_waddsubhx): Likewise
5834 (iwmmxt_wsubaddhx): Likewise.
5835 (addc<mode>3): Likewise.
5836 (iwmmxt_avg4): Likewise.
5837 (iwmmxt_avg4r): Likewise.
5838 (iwmmxt_wmaddsx): Likewise.
5839 (iwmmxt_wmaddux): Likewise.
5840 (iwmmxt_wmaddsn): Likewise.
5841 (iwmmxt_wmaddun): Likewise.
5842 (iwmmxt_wmulwsm): Likewise.
5843 (iwmmxt_wmulwum): Likewise.
5844 (iwmmxt_wmulsmr): Likewise.
5845 (iwmmxt_wmulumr): Likewise.
5846 (iwmmxt_wmulwsmr): Likewise.
5847 (iwmmxt_wmulwumr): Likewise.
5848 (iwmmxt_wmulwl): Likewise.
5849 (iwmmxt_wqmulm): Likewise.
5850 (iwmmxt_wqmulwm): Likewise.
5851 (iwmmxt_wqmulmr): Likewise.
5852 (iwmmxt_wqmulwmr): Likewise.
5853 (iwmmxt_waddbhusm): Likewise.
5854 (iwmmxt_waddbhusl): Likewise.
5855 (iwmmxt_wqmiabb): Likewise.
5856 (iwmmxt_wqmiabt): Likewise.
5857 (iwmmxt_wqmiatb): Likewise.
5858 (iwmmxt_wqmiatt): Likewise.
5859 (iwmmxt_wqmiabbn): Likewise.
5860 (iwmmxt_wqmiabtn): Likewise.
5861 (iwmmxt_wqmiatbn): Likewise.
5862 (iwmmxt_wqmiattn): Likewise.
5863 (iwmmxt_wmiabb): Likewise.
5864 (iwmmxt_wmiabt): Likewise.
5865 (iwmmxt_wmiatb): Likewise.
5866 (iwmmxt_wmiatt): Likewise.
5867 (iwmmxt_wmiabbn): Likewise.
5868 (iwmmxt_wmiabtn): Likewise.
5869 (iwmmxt_wmiatbn): Likewise.
5870 (iwmmxt_wmiattn): Likewise.
5871 (iwmmxt_wmiawbb): Likewise.
5872 (iwmmxt_wmiawbt): Likewise.
5873 (iwmmxt_wmiawtb): Likewise.
5874 (iwmmxt_wmiawtt): Likewise.
5875 (iwmmxt_wmiawbbn): Likewise.
5876 (iwmmxt_wmiawbtn): Likewise.
5877 (iwmmxt_wmiawtbn): Likewise.
5878 (iwmmxt_wmiawttn): Likewise.
5879 (iwmmxt_wmerge): Likewise.
5880 (iwmmxt_tandc<mode>3): Likewise.
5881 (iwmmxt_torc<mode>3): Likewise.
5882 (iwmmxt_torvsc<mode>3): Likewise.
5883 (iwmmxt_textrc<mode>3): Likewise.
5884 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
5885 (wmmxt_pack): Likewise.
5886 (wmmxt_mult_c1): Likewise.
5887 (wmmxt_mult_c2): Likewise.
5888 (wmmxt_alu_c1): Likewise.
5889 (wmmxt_alu_c2): Likewise.
5890 (wmmxt_alu_c3): Likewise.
5891 (wmmxt_transfer_c1): Likewise.
5892 (wmmxt_transfer_c2): Likewise.
5893 (wmmxt_transfer_c3): Likewise.
5894 (marvell_f_iwmmxt_wstr): Likewise.
5895 (marvell_f_iwmmxt_wldr): Likewise.
5897 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5899 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5901 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5904 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5905 * lra-constraints.c (need_for_split_p): Check call used hard regs
5906 living through calls.
5908 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5909 call used regs for call insn.
5911 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5914 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5915 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5917 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5919 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5921 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5923 * lra-constraints.c (need_for_split_p): Check call used hard regs
5924 living through calls.
5926 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5929 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5930 to tie with any other modes. Eliminate Altivec vector mode tests,
5931 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5932 code, to return 0 if testing MODE2 for a condition, if we've
5933 already tested MODE1 for the same condition.
5935 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5937 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5940 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5942 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5943 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5945 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5947 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5949 (aarch64_symbolic_constant_p): Remove.
5950 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5951 static. Fix line length and white space.
5952 (aarch64_symbolic_constant_p): Remove.
5953 * config/aarch64/predicates.md (aarch64_valid_symref):
5954 Use aarch64_classify_symbol_expression.
5956 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5958 * config/arm/constraints.md (Ts): New constraint.
5959 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5961 (compare_scc): Use "Ts" constraint for operand 0.
5962 (ior_scc_scc): Likewise.
5963 (and_scc_scc): Likewise.
5964 (and_scc_scc_nodom): Likewise.
5965 (ior_scc_scc_cmp): Likewise for operand 7.
5966 (and_scc_scc_cmp): Likewise.
5967 * config/arm/thumb2.md (thumb2_movsi_insn):
5968 Add alternatives for 16-bit encodings.
5969 (thumb2_movhi_insn): Likewise.
5970 (thumb2_movsicc_insn): Likewise.
5971 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5972 (thumb2_negscc): Use "Ts" constraint.
5973 Move mvn instruction outside cond_exec block.
5974 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
5975 for 16-bit encodings.
5977 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5979 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
5981 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
5982 (mulsi3subsi): Likewise.
5983 (mulsidi3adddi): Likewise.
5984 (mulsidi3_v6): Likewise.
5985 (umulsidi3_v6): Likewise.
5986 (umulsidi3adddi_v6): Likewise.
5987 (smulsi3_highpart_v6): Likewise.
5988 (umulsi3_highpart_v6): Likewise.
5989 (mulhisi3tb): Likewise.
5990 (mulhisi3bt): Likewise.
5991 (mulhisi3tt): Likewise.
5992 (maddhisi4): Likewise.
5993 (maddhisi4tb): Likewise.
5994 (maddhisi4tt): Likewise.
5995 (maddhidi4): Likewise.
5996 (maddhidi4tb): Likewise.
5997 (maddhidi4tt): Likewise.
5998 (zeroextractsi_compare0_scratch): Likewise.
5999 (insv_zero): Likewise.
6000 (insv_t2): Likewise.
6001 (anddi_notzesidi_di): Likewise.
6002 (anddi_notsesidi_di): Likewise.
6003 (andsi_notsi_si): Likewise.
6004 (iordi_zesidi_di): Likewise.
6005 (xordi_zesidi_di): Likewise.
6006 (andsi_iorsi3_notsi): Likewise.
6008 (smax_m1): Likewise.
6010 (not_shiftsi): Likewise.
6011 (unaligned_loadsi): Likewise.
6012 (unaligned_loadhis): Likewise.
6013 (unaligned_loadhiu): Likewise.
6014 (unaligned_storesi): Likewise.
6015 (unaligned_storehi): Likewise.
6016 (extv_reg): Likewise.
6017 (extzv_t2): Likewise.
6019 (udivsi3): Likewise.
6020 (arm_zero_extendhisi2addsi): Likewise.
6021 (arm_zero_extendqisi2addsi): Likewise.
6022 (compareqi_eq0): Likewise.
6023 (arm_extendhisi2_v6): Likewise.
6024 (arm_extendqisi2addsi): Likewise.
6025 (arm_movt): Likewise.
6026 (thumb2_ldrd): Likewise.
6027 (thumb2_ldrd_base): Likewise.
6028 (thumb2_ldrd_base_neg): Likewise.
6029 (thumb2_strd): Likewise.
6030 (thumb2_strd_base): Likewise.
6031 (thumb2_strd_base_neg): Likewise.
6032 (arm_negsi2): Add alternative for 16-bit encoding.
6033 (arm_one_cmplsi2): Likewise.
6035 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6037 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
6038 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
6039 (movdfcc): Likewise.
6040 * config/arm/vfp.md (*thumb2_movsf_vfp):
6041 Disable predication for arm_restrict_it.
6042 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
6043 (*thumb2_movdfcc_vfp): Likewise.
6044 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
6045 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
6046 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
6047 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
6048 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
6049 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
6050 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
6051 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
6052 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
6053 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
6054 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
6055 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
6056 Disable predication for arm_restrict_it.
6058 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
6060 * config/i386/bmiintrin.h (_bextr_u32): New.
6061 (_bextr_u64): Ditto.
6063 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
6065 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
6066 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
6067 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
6068 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
6069 * config/mips/n32-elf.h: ...this new file.
6071 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
6074 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
6075 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
6077 2013-06-27 Catherine Moore <clm@codesourcery.com>
6079 * config/mips/mips-tables.opt: Regenerate.
6080 * config/mips/mips-cpus.def: Add m14ke and m14kec.
6081 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
6082 * doc/invoke.texi: Add -m14kc.
6084 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6087 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
6088 constraints of operand 1 and 2.
6091 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
6092 to match RTL canonicalization. Swap predicates and
6093 constraints of operand 1 and 2.
6095 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
6097 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
6098 Process OP_INOUT regs for splitting too.
6100 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6102 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
6103 decl before the loop, initialize to NULL.
6104 (vectorizable_load): Initialize ptr_incr to NULL.
6106 2013-06-27 Martin Jambor <mjambor@suse.cz>
6109 * ipa-ref.h (ipa_maybe_record_reference): Declare.
6110 * ipa-ref.c (ipa_maybe_record_reference): New function.
6111 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
6112 * ipa-cp.c (create_specialized_node): Record potential references from
6114 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
6116 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6118 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
6119 parameter 'mode' of type 'enum machine_mode mode'; change to pass
6120 'mode' to force_reg.
6121 (aarch64_add_offset): Update calls to aarch64_force_temporary.
6122 (aarch64_expand_mov_immediate): Likewise.
6124 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6126 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
6127 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
6129 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6131 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
6132 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6133 (struct machine_function): Add tbegin_p.
6134 (s390_canonicalize_comparison): Fold CC mode compares to
6135 conditional jump if possible.
6136 (s390_emit_jump): Return the emitted jump.
6137 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
6138 Handle CCRAWmode compares.
6139 (s390_option_override): Default to -mhtm if available.
6140 (s390_reg_clobbered_rtx): Handle floating point regs as well.
6141 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
6142 FPRs instead of df_regs_ever_live_p.
6143 (s390_optimize_nonescaping_tx): New function.
6144 (s390_init_frame_layout): Extend clobbered_regs array to cover
6146 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
6147 (s390_expand_tbegin): New function.
6148 (enum s390_builtin): New enum definition.
6149 (code_for_builtin): New array definition.
6150 (s390_init_builtins): New function.
6151 (s390_expand_builtin): New function.
6152 (TARGET_INIT_BUILTINS): Define.
6153 (TARGET_EXPAND_BUILTIN): Define.
6154 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
6155 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
6156 (s390_alc_comparison): Likewise.
6157 * config/s390/s390-modes.def: Add CCRAWmode.
6158 * config/s390/s390.h (processor_flags): Add PF_TX.
6159 (TARGET_CPU_HTM): Define macro.
6160 (TARGET_HTM): Define macro.
6161 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
6162 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
6163 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6164 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
6165 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
6166 (TBEGIN_MASK, TBEGINC_MASK): New constants.
6167 ("*cc_to_int"): Move up.
6168 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
6169 constants other than 0.
6170 ("*ccraw_to_int"): New insn and splitter definition.
6171 ("tbegin", "tbegin_nofloat", "tbegin_retry")
6172 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
6173 ("tx_assist"): New expander.
6174 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
6175 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
6176 * config/s390/s390.opt: Add -mhtm option.
6177 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
6178 * config/s390/htmxlintrin.h: New file.
6179 * config/s390/htmintrin.h: New file.
6180 * config/s390/s390intrin.h: New file.
6181 * doc/extend.texi: Document htm builtins.
6182 * config.gcc: Add the new header files to extra_headers.
6184 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
6186 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
6187 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
6189 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
6190 Pat Haugen <pthaugen@us.ibm.com>
6191 Peter Bergner <bergner@vnet.ibm.com>
6193 * config/rs6000/power8.md: New.
6194 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
6195 setting for power8 entry.
6196 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
6197 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
6198 test for Power4/Power5 only.
6199 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
6201 (force_new_group): Adjust comment.
6202 * config/rs6000/rs6000.md: Include power8.md.
6204 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
6206 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
6207 * config/arm/arm-protos.h (arm_max_conditional_execute): New
6209 (tune_params): Update comment.
6210 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
6211 (arm_max_conditional_execute): New function.
6212 (thumb2_final_prescan_insn): Use max_insn_skipped and
6213 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
6215 2013-06-25 Jakub Jelinek <jakub@redhat.com>
6217 PR tree-optimization/57705
6218 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
6219 SSA_NAME step, provided that it is not defined inside the loop.
6220 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
6221 (get_initial_def_for_induction): Handle SSA_NAME IV step.
6223 2013-06-25 Martin Jambor <mjambor@suse.cz>
6226 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
6227 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
6229 (ipa_note_param_call): Initialize member_ptr flag.
6230 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
6231 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
6232 (ipa_write_indirect_edge_info): Stream member_ptr flag.
6233 (ipa_read_indirect_edge_info): Likewise.
6235 2013-06-25 Richard Biener <rguenther@suse.de>
6238 * passes.c (init_optimization_passes): Move pass_fold_builtins
6239 and pass_dce earlier with -Og.
6241 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
6243 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
6244 <BIT_FIELD_REF>: Remove trailing TAB.
6245 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
6248 2013-06-24 Martin Jambor <mjambor@suse.cz>
6250 PR tree-optimization/57358
6251 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
6252 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
6253 (ipa_analyze_params_uses): Generate pessimistic info when true.
6255 2013-06-24 Martin Jambor <mjambor@suse.cz>
6257 PR tree-optimization/57539
6258 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
6259 global.inlined_to of the new node to it. All callers changed.
6260 * ipa-inline-transform.c (clone_inlined_nodes): New variable
6261 inlining_into, pass it to cgraph_clone_node.
6262 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
6263 ipa_free_edge_args_substructures.
6264 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
6265 rdesc linked list. Do not assert rdesc edges have inlined caller.
6266 Assert we have found an rdesc in the rdesc list.
6268 2013-06-24 Richard Biener <rguenther@suse.de>
6270 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
6271 (pointer_set_lookup): Declare.
6272 (class pointer_map): New template class implementing a
6273 generic pointer to T map.
6274 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
6275 pointer_map<T>::contains, pointer_map<T>::insert,
6276 pointer_map<T>::traverse): New functions.
6277 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
6278 (pointer_set_lookup): New function.
6279 (pointer_set_contains): Use pointer_set_lookup.
6280 (pointer_set_insert): Likewise.
6281 (insert_aux): Remove.
6282 (struct pointer_map_t): Embed a pointer_set_t.
6283 (pointer_map_create): Adjust.
6284 (pointer_map_destroy): Likewise.
6285 (pointer_map_contains): Likewise.
6286 (pointer_map_insert): Likewise.
6287 (pointer_map_traverse): Likewise.
6288 * tree-streamer.h (struct streamer_tree_cache_d): Use a
6289 pointer_map<unsigned> instead of a pointer_map_t.
6290 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
6291 (streamer_tree_cache_lookup): Likewise.
6292 (streamer_tree_cache_create): Likewise.
6293 (streamer_tree_cache_delete): Likewise.
6294 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
6295 pointer_map<unsigned> instead of a pointer_map_t.
6296 (lto_init_tree_ref_encoder): Adjust.
6297 (lto_destroy_tree_ref_encoder): Likewise.
6298 * lto-section-out.c (lto_output_decl_index): Likewise.
6299 (lto_record_function_out_decl_state): Likewise.
6300 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
6302 2013-06-24 Richard Biener <rguenther@suse.de>
6304 PR tree-optimization/57488
6305 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
6307 2013-06-24 Alan Modra <amodra@gmail.com>
6309 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
6310 (gen_easy_altivec_constant): Likewise.
6311 * config/rs6000/predicates.md (easy_vector_constant_add_self,
6312 easy_vector_constant_msb): Likewise.
6314 2013-06-23 Jakub Jelinek <jakub@redhat.com>
6317 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
6318 add missing return true.
6320 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
6323 * config/sh/predicates.md (general_extend_operand): Invoke
6324 general_movsrc_operand for memory operands.
6325 (general_movsrc_operand): Allow reg+reg addressing, do not use
6326 general_operand for memory operands.
6328 2013-06-23 Sriraman Tallam <tmsriram@google.com>
6330 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
6331 when current target options does not apply.
6332 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
6333 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
6334 * config/i386/bmiintrin.h: Pass appropriate target
6335 attributes to header.
6336 * config/i386/mmintrin.h: Ditto.
6337 * config/i386/nmmintrin.h: Ditto.
6338 * config/i386/avx2intrin.h: Ditto.
6339 * config/i386/fxsrintrin.h: Ditto.
6340 * config/i386/tbmintrin.h: Ditto.
6341 * config/i386/xsaveintrin.h: Ditto.
6342 * config/i386/f16cintrin.h: Ditto.
6343 * config/i386/xtestintrin.h: Ditto.
6344 * config/i386/xsaveoptintrin.h: Ditto.
6345 * config/i386/bmi2intrin.h: Ditto.
6346 * config/i386/lzcntintrin.h: Ditto.
6347 * config/i386/smmintrin.h: Ditto.
6348 * config/i386/wmmintrin.h: Ditto.
6349 * config/i386/x86intrin.h: Remove all header include guards.
6350 * config/i386/prfchwintrin.h: Ditto.
6351 * config/i386/pmmintrin.h: Ditto.
6352 * config/i386/tmmintrin.h: Ditto.
6353 * config/i386/xmmintrin.h: Ditto.
6354 * config/i386/popcntintrin.h: Ditto.
6355 * config/i386/rdseedintrin.h: Ditto.
6356 * config/i386/ammintrin.h: Ditto.
6357 * config/i386/emmintrin.h: Ditto.
6358 * config/i386/immintrin.h: Remove all header include guards.
6359 * config/i386/fma4intrin.h: Ditto.
6360 * config/i386/lwpintrin.h: Ditto.
6361 * config/i386/xopintrin.h: Ditto.
6362 * config/i386/ia32intrin.h: Ditto.
6363 * config/i386/avxintrin.h: Ditto.
6364 * config/i386/rtmintrin.h: Ditto.
6365 * config/i386/fmaintrin.h: Ditto.
6366 * config/i386/mm3dnow.h: Ditto.
6368 2013-06-22 Sriraman Tallam <tmsriram@google.com>
6370 * common/config/i386/i386-common.c: Handle LZCNT.
6372 2013-06-22 Andi Kleen <ak@linux.intel.com>
6374 * doc/extend.texi: Use __atomic_store_n instead of
6375 __atomic_store in HLE example.
6377 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
6379 * config/sh/sh.c: Remove <cstdlib> workaround.
6381 2013-06-21 Andi Kleen <ak@linux.intel.com>
6383 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
6385 2013-06-21 Andi Kleen <ak@linux.intel.com>
6387 * doc/extend.texi: Document that __atomic_clear and
6388 __atomic_test_and_set should only be used with bool.
6390 2013-06-20 Jan Hubicka <jh@suse.cz>
6392 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
6394 * tree.c (decls_same_for_odr): New function.
6395 (same_for_edr): New function.
6396 (types_same_for_odr): New function.
6397 (get_binfo_at_offset): Use it.
6398 * tree.h (types_same_for_odr): Declare.
6400 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
6401 Jason Merrill <jason@redhat.com>
6403 * system.h: Include <cstdlib> as well as <stdlib.h>.
6405 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
6408 * config/i386/i386.c (construct_container): Report error if
6409 long double is used with disabled x87 float returns.
6411 2013-06-20 Jan Hubicka <jh@suse.cz>
6413 * lto-cgraph.c (input_symtab): Do not set cgraph state.
6415 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
6417 PR rtl-optimization/57425
6418 PR rtl-optimization/57569
6419 * alias.c (write_dependence_p): Remove parameters mem_mode and
6420 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
6421 Changed all callers.
6422 (canon_anti_dependence): Get comments and semantics in sync.
6423 Add parameter mem_canonicalized. Changed all callers.
6424 * rtl.h (canon_anti_dependence): Update prototype.
6426 2013-06-20 Richard Biener <rguenther@suse.de>
6428 * data-streamer-in.c (streamer_read_uhwi): Optimize single
6429 byte case, inline streamer_read_uchar and defer section
6432 2013-06-20 Richard Biener <rguenther@suse.de>
6434 PR tree-optimization/57584
6435 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
6436 SSA names into the expanded expression that take part in
6437 abnormal coalescing.
6439 2013-06-19 Sharad Singhai <singhai@google.com>
6441 * gcov.c (print_usage): Handle new option.
6442 (process_args): Ditto.
6443 (get_gcov_intermediate_filename): New function.
6444 (output_intermediate_file): New function.
6445 (output_gcov_file): New function
6446 (generate_results): Handle new option.
6447 (release_function): Relase demangled name.
6448 (read_graph_file): Handle demangled name.
6449 (output_lines): Ditto.
6450 * doc/gcov.texi: Document gcov intermediate format.
6452 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
6455 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
6456 (lra_emit_add): Use the functions. Add comment about Y as an
6459 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
6462 * collect2.c (collect_atexit): New.
6463 (collect_exit): Delete.
6464 (main): Register collect_atexit with atexit.
6465 (collect_wait): Change collect_exit to exit.
6467 * collect2.h (collect_exit): Delete.
6468 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
6470 2013-06-19 Wei Mi <wmi@google.com>
6472 PR rtl-optimization/57518
6473 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
6474 if regno is used in paradoxical subreg.
6475 (update_equiv_regs): Check pdx_subregs[regno] before
6476 set a reg to be equivalent with a mem.
6478 2013-06-19 Matthias Klose <doko@ubuntu.com>
6481 * file-find.h (find_a_file): Add a mode parameter.
6482 * file-find.c (find_a_file): Likewise.
6483 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
6484 with X_OK for the executables.
6485 * collect2.c (main): Call find_a_file with X_OK.
6487 2013-06-19 Steve Ellcey <sellcey@mips.com>
6490 * config/mips/mips.md (casesi_internal_mips16_<mode>):
6491 Use NEXT_INSN instead of next_real_insn.
6493 2013-06-19 Jan Hubicka <jh@suse.cz>
6495 * cgraph.h (const_value_known_p): Replace by ...
6496 (ctor_for_folding): .. this one.
6497 * cgraphunit.c (process_function_and_variable_attributes): Use it.
6498 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
6499 * expr.c (expand_expr_real_1): Likewise.
6500 (string_constant): Likewise.
6501 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
6502 * ipa.c (process_references): Likewise.
6503 (symtab_remove_unreachable_nodes): Likewise.
6504 * ipa-inline-analysis.c (param_change_prob): Likewise.
6505 * gimple-fold.c (canonicalize_constructor_val): Likewise.
6506 (get_base_constructor): Likwise.
6507 * varpool.c (varpool_remove_node): Likewise.
6508 (varpool_remove_initializer): LIkewise.
6509 (dump_varpool_node): LIkwise.
6510 (const_value_known_p): Rewrite to ...
6511 (ctor_for_folding): ... this one.
6513 2013-06-19 Jakub Jelinek <jakub@redhat.com>
6516 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
6517 PERSONALITY in $PATH derived prefixes.
6519 2013-06-19 Jeff Law <law@redhat.com>
6521 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
6524 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
6525 (simplify_bitwise_binary): Use it to simpify certain binary ops on
6528 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
6530 * config/arm/vfp.md: Move VFP instruction classification documentation
6532 * config/arm/arm.md: ... here. Update instruction classification
6535 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
6537 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
6538 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
6539 pattern. Use more efficient sequences on ARMv5 and Thumb2.
6541 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
6544 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6545 with NEXT_INSN. Use tablejump_p to check for jump table data
6548 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
6551 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
6552 that now in C++ the value is correct per the C++ standards.
6554 2013-06-19 Richard Biener <rguenther@suse.de>
6556 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
6559 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6562 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6565 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6566 with next_active_insn.
6568 2013-06-18 Sriraman Tallam <tmsriram@google.com>
6570 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
6571 functions are inlined during failures to flag an error.
6572 * tree-inline.c (expand_call_inline): Allow the error to be flagged
6573 in early inline pass.
6575 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
6577 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
6580 2013-06-18 Julian Brown <julian@codesourcery.com>
6582 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
6583 Permit virtual register pre-reload if !strict.
6584 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
6586 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
6588 * config/arm/neon.md (movmisalign<mode>): Use
6589 neon_perm_struct_or_reg_operand instead of
6590 neon_struct_or_register_operand.
6591 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
6592 neon_permissive_struct_operand instead of neon_struct_operand.
6593 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
6594 neon_vector_mem_operand.
6595 * config/arm/predicates.md (neon_struct_operand): Adjust call to
6596 neon_vector_mem_operand.
6597 (neon_permissive_struct_operand): New.
6598 (neon_struct_or_register_operand): Rename to...
6599 (neon_perm_struct_or_reg_operand): This. Adjust call to
6600 neon_vector_mem_operand.
6602 2013-06-18 Richard Biener <rguenther@suse.de>
6604 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
6605 * lto-streamer.h: Include pointer-set.h.
6606 (struct lto_decl_slot): Remove.
6607 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
6608 Remove next_index entry.
6609 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
6610 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
6611 (lto_init_tree_ref_encoder): Adjust.
6612 (lto_destroy_tree_ref_encoder): Likewise.
6613 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
6614 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
6615 (lto_output_decl_index): Adjust.
6616 (lto_new_out_decl_state): Likewise.
6617 (lto_record_function_out_decl_state): Likewise.
6618 * lto-streamer-out.c (copy_function): Likewise.
6620 2013-06-18 Richard Biener <rguenther@suse.de>
6622 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
6623 * cgraphunit.c: Include cfgloop.h.
6624 (init_lowered_empty_function): Initialize the loop tree.
6625 (assemble_thunk): Insert new BBs into loops.
6627 2013-06-18 Richard Biener <rguenther@suse.de>
6629 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
6630 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
6631 the map from cache entry to cache index optional.
6632 (streamer_tree_cache_replace_tree): Adjust accordingly.
6633 (streamer_tree_cache_append): Likewise.
6634 (streamer_tree_cache_delete): Likewise.
6635 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
6636 streamer cache map from cache entry to cache index.
6637 * lto-streamer-out.c (create_output_block): Adjust.
6639 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
6641 * config/arm/arm.md (attribute "insn"): Move multiplication and
6642 division attributes to...
6643 (attribute "type"): ... here. Remove mult.
6644 (attribute "mul32"): New attribute.
6645 (attribute "mul64"): Add umaal.
6646 (*arm_mulsi3): Update attributes.
6647 (*arm_mulsi3_v6): Likewise.
6648 (*thumb_mulsi3): Likewise.
6649 (*thumb_mulsi3_v6): Likewise.
6650 (*mulsi3_compare0): Likewise.
6651 (*mulsi3_compare0_v6): Likewise.
6652 (*mulsi_compare0_scratch): Likewise.
6653 (*mulsi_compare0_scratch_v6): Likewise.
6654 (*mulsi3addsi): Likewise.
6655 (*mulsi3addsi_v6): Likewise.
6656 (*mulsi3addsi_compare0): Likewise.
6657 (*mulsi3addsi_compare0_v6): Likewise.
6658 (*mulsi3addsi_compare0_scratch): Likewise.
6659 (*mulsi3addsi_compare0_scratch_v6): Likewise.
6660 (*mulsi3subsi): Likewise.
6661 (*mulsidi3adddi): Likewise.
6662 (*mulsi3addsi_v6): Likewise.
6663 (*mulsidi3adddi_v6): Likewise.
6664 (*mulsidi3_nov6): Likewise.
6665 (*mulsidi3_v6): Likewise.
6666 (*umulsidi3_nov6): Likewise.
6667 (*umulsidi3_v6): Likewise.
6668 (*umulsidi3adddi): Likewise.
6669 (*umulsidi3adddi_v6): Likewise.
6670 (*smulsi3_highpart_nov6): Likewise.
6671 (*smulsi3_highpart_v6): Likewise.
6672 (*umulsi3_highpart_nov6): Likewise.
6673 (*umulsi3_highpart_v6): Likewise.
6674 (mulhisi3): Likewise.
6675 (*mulhisi3tb): Likewise.
6676 (*mulhisi3bt): Likewise.
6677 (*mulhisi3tt): Likewise.
6678 (maddhisi4): Likewise.
6679 (*maddhisi4tb): Likewise.
6680 (*maddhisi4tt): Likewise.
6681 (maddhidi4): Likewise.
6682 (*maddhidi4tb): Likewise.
6683 (*maddhidi4tt): Likewise.
6685 (udivsi3): Likewise.
6686 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
6687 (thumb2_mulsi_short_compare0): Likewise.
6688 (thumb2_mulsi_short_compare0_scratch): Likewise.
6689 * config/arm/arm1020e.md (1020mult1): Update attribute change.
6690 (1020mult2): Likewise.
6691 (1020mult3): Likewise.
6692 (1020mult4): Likewise.
6693 (1020mult5): Likewise.
6694 (1020mult6): Likewise.
6695 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
6697 (cortex_a15_mult64): Likewise.
6698 (cortex_a15_sdiv): Likewise.
6699 (cortex_a15_udiv): Likewise.
6700 * config/arm/arm1026ejs.md (mult1): Update attribute change.
6706 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
6707 (pj4_ir_div): Likewise.
6708 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
6709 (11_mult2): Likewise.
6710 (11_mult3): Likewise.
6711 (11_mult4): Likewise.
6712 (11_mult5): Likewise.
6713 (11_mult6): Likewise.
6714 (11_mult7): Likewise.
6715 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
6716 (cortex_a8_mla): Likewise.
6717 (cortex_a8_mull): Likewise.
6718 (cortex_a8_smulwy): Likewise.
6719 (cortex_a8_smlald): Likewise.
6720 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
6721 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
6722 (cortex_r4_mul_3): Likewise.
6723 (cortex_r4_mla_4): Likewise.
6724 (cortex_r4_mla_3): Likewise.
6725 (cortex_r4_smlald): Likewise.
6726 (cortex_r4_mull): Likewise.
6727 (cortex_r4_sdiv): Likewise.
6728 (cortex_r4_udiv): Likewise.
6729 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
6730 (cortex_a7_idiv): Likewise.
6731 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
6732 (9_mult2): Likewise.
6733 (9_mult3): Likewise.
6734 (9_mult4): Likewise.
6735 (9_mult5): Likewise.
6736 (9_mult6): Likewise.
6737 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
6738 (cortex_a53_sdiv): Likewise.
6739 (cortex_a53_udiv): Likewise.
6740 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
6741 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
6742 (mp626_mult2): Likewise.
6743 (mp626_mult3): Likewise.
6744 (mp626_mult4): Likewise.
6745 * config/arm/fa526.md (526_mult1): Update attribute change.
6746 (526_mult2): Likewise.
6747 * config/arm/arm-generic.md (mult): Update attribute change.
6748 (mult_ldsched_strongarm): Likewise.
6749 (mult_ldsched): Likewise.
6750 (multi_cycle): Likewise.
6751 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
6752 * config/arm/fa606te.md (606te_mult1): Update attribute change.
6753 (606te_mult2): Likewise.
6754 (606te_mult3): Likewise.
6755 (606te_mult4): Likewise.
6756 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
6757 (cortex_a9_mac16): Likewise.
6758 (cortex_a9_multiply): Likewise.
6759 (cortex_a9_mac): Likewise.
6760 (cortex_a9_multiply_long): Likewise.
6761 * config/arm/fa626te.md (626te_mult1): Update attribute change.
6762 (626te_mult2): Likewise.
6763 (626te_mult3): Likewise.
6764 (626te_mult4): Likewise.
6766 2013-06-18 Richard Biener <rguenther@suse.de>
6769 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
6771 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6774 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6775 with next_active_insn.
6777 2013-06-18 Alan Modra <amodra@gmail.com>
6779 * config/rs6000/rs6000.h (enum data_align): New.
6780 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
6781 (DATA_ABI_ALIGNMENT): Define.
6782 (CONSTANT_ALIGNMENT): Correct comment.
6783 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
6784 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
6786 2013-06-17 David Malcolm <dmalcolm@redhat.com>
6788 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
6789 ATTRIBUTE_UNUSED marking.
6791 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
6793 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
6794 alternative and update.
6795 (aarch64_dup_lanedi): Delete.
6796 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
6797 * config/aarch64/aarch64-simd-builtins.def: Update.
6799 2013-06-17 Richard Biener <rguenther@suse.de>
6801 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
6802 (lto_input_scc): Declare.
6803 (lto_input_tree_1): Likewise.
6804 (struct lto_stats_d): Add num_tree_bodies_output and
6805 num_pickle_refs_output.
6806 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
6807 (lto_read_tree_1): Split out from ...
6808 (lto_read_tree): ... this.
6809 (lto_input_scc): New function.
6810 (lto_input_tree_1): Split out from ...
6811 (lto_input_tree): ... this. Handle LTO_tree_scc.
6812 (lto_data_in_create): Create the streamer cache without hashes.
6813 * lto-streamer-out.c (create_output_block): Create the streamer
6814 cache with hashes when not doing WPA.
6815 (lto_write_tree_1): Split out from ...
6816 (lto_write_tree): ... this.
6817 (get_symbol_initial_value): New function.
6818 (lto_output_tree_1): Split out from ...
6819 (lto_output_tree): ... this. Write trees as series of SCCs
6820 using a DFS walk via DFS_write_tree.
6821 (struct sccs, struct scc_entry): New types.
6822 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
6823 (DFS_write_tree_body): New function.
6824 (DFS_write_tree): Likewise.
6825 (hash_tree): Likewise.
6826 (scc_entry_compare): Likewise.
6827 (hash_scc): Likewise.
6828 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
6829 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
6830 TREE_CHAIN as regular reference.
6831 (streamer_read_integer_cst): Remove.
6832 (streamer_get_pickled_tree): Adjust.
6833 * tree-streamer-out.c (streamer_write_chain): Disable streaming
6834 of DECL_EXTERNALs in BLOCK_VARS for now.
6835 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
6837 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
6838 Add hash value argument and record that if hashes are recorded
6840 (streamer_tree_cache_insert_1): Adjust.
6841 (streamer_tree_cache_insert): Likewise.
6842 (streamer_tree_cache_insert_at): Rename to ...
6843 (streamer_tree_cache_replace_tree): ... this and adjust.
6844 (streamer_tree_cache_append): Adjust.
6845 (record_common_node): Likewise.
6846 (streamer_tree_cache_create): Add argument whether to
6847 record hash values together with trees.
6848 (streamer_tree_cache_delete): Adjust.
6849 * tree-streamer.h (struct streamer_tree_cache_d): Add
6851 (streamer_read_integer_cst): Remove.
6852 (streamer_tree_cache_insert): Adjust.
6853 (streamer_tree_cache_append): Likewise.
6854 (streamer_tree_cache_insert_at): Rename to ...
6855 (streamer_tree_cache_replace_tree): ... this and adjust.
6856 (streamer_tree_cache_create): Add argument whether to record hashes.
6857 (streamer_tree_cache_get): Rename to ...
6858 (streamer_tree_cache_get_tree): ... this.
6859 (streamer_tree_cache_get_hash): New function.
6860 * tree.c (cache_integer_cst): New function.
6861 * tree.h (cache_integer_cst): Declare.
6862 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
6863 * lto-symtab.c (lto_varpool_replace_node): Only release
6864 DECL_INITIAL of non-prevailing decls.
6865 * varpool.c (varpool_remove_initializer): Do not release
6866 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
6868 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6870 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
6871 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
6872 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
6873 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
6874 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
6875 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
6876 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
6877 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
6878 instead of TARGET_64BIT.
6879 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
6880 Require ISA_HAS_<D>DIV.
6882 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
6884 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
6885 (mips*-*-linux*): Move default with_llsc setting to where other
6887 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
6889 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
6890 Likewise. Remove default with_tune setting. Move default float
6891 setting to its own block. Handle with_llsc in the same block as above.
6893 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
6895 PR rtl-optimization/57425
6896 PR rtl-optimization/57569
6897 * alias.c (write_dependence_p): Add new parameters mem_mode,
6898 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6899 Changed all callers.
6900 (canon_anti_dependence): New function.
6901 * cse.c (check_dependence): Use canon_anti_dependence.
6902 * cselib.c (cselib_invalidate_mem): Likewise.
6903 * rtl.h (canon_anti_dependence): Declare.
6905 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6907 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6908 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6909 ".set mips3" for 64-bit targets.
6911 2013-06-15 Dehao Chen <dehao@google.com>
6913 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6914 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6915 (gimple_check_call_args): Likewise.
6916 * value-prof.c (check_ic_target): Likewise.
6917 * ipa-inline.c (early_inliner): Likewise.
6918 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6919 * cgraph.c (cgraph_create_edge_1): Likewise.
6920 (cgraph_make_edge_direct): Likewise.
6922 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6925 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6926 rs6000_output_move_128bit to handle emitting quad memory
6927 operations. Set attribute length to 8 bytes.
6929 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6931 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6933 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6934 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6935 (aarch64_<su>mlsl<mode>): Likewise.
6937 2013-06-14 Mike Stump <mikestump@comcast.net>
6939 * Makefile.in (TARGET_H): Add insn-codes.h.
6941 2013-06-14 Alan Modra <amodra@gmail.com>
6945 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6946 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6947 bitfield expansion when EXPAND_MEMORY.
6948 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6950 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6952 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6953 test for clearing quad memory on 32-bit later.
6955 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6957 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6958 (fold_negate_expr): Likewise.
6959 (fold_real_zero_addition_p): Handle vectors.
6960 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6962 2013-06-14 Alan Modra <amodra@gmail.com>
6964 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6966 2013-06-13 Jan Hubicka <jh@suse.cz>
6968 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6969 Local comdats are not externally visible.
6970 * symtab.c (dump_symtab_base): Dump externally visible.
6971 (verify_symtab_base): Verify back links in the symtab hash.
6973 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6975 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
6976 CONVERT_EXPR as equal nodes.
6978 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6980 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
6982 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6984 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
6985 Generalize to complex and vector.
6986 * tree.c (build_all_ones_cst): New function.
6987 * tree.h (build_all_ones_cst): Declare it.
6989 2013-06-13 Alan Modra <amodra@gmail.com>
6991 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
6992 * config/rs6000/rs6000.md (signbittf2): New insn.
6993 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
6994 (abstf2_internal, cmptf_internal2): Likewise.
6995 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
6997 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6998 Pat Haugen <pthaugen@us.ibm.com>
6999 Peter Bergner <bergner@vnet.ibm.com>
7001 * config/rs6000/rs6000.c (emit_load_locked): Add support for
7002 power8 byte, half-word, and quad-word atomic instructions.
7003 (emit_store_conditional): Likewise.
7004 (rs6000_expand_atomic_compare_and_swap): Likewise.
7005 (rs6000_expand_atomic_op): Likewise.
7007 * config/rs6000/sync.md (larx): Add new modes for power8.
7009 (AINT): New mode iterator to include TImode as well as normal
7010 integer modes on power8.
7011 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
7012 that VSX registers are not considered. Use AINT mode iterator
7013 instead of INT1 to allow inclusion of quad word atomic operations
7015 (load_locked<mode>): Likewise.
7016 (store_conditional<mode>): Likewise.
7017 (atomic_compare_and_swap<mode>): Likewise.
7018 (atomic_exchange<mode>): Likewise.
7019 (atomic_nand<mode>): Likewise.
7020 (atomic_fetch_<fetchop_name><mode>): Likewise.
7021 (atomic_nand_fetch<mode>): Likewise.
7022 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
7024 (ATOMIC): On power8, add QImode, HImode modes.
7025 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
7026 modes that promote to SImode.
7027 (load_lockedti): Convert TImode arguments to PTImode, so that we
7028 get a guaranteed even/odd register pair.
7029 (load_lockedpti): Likewise.
7030 (store_conditionalti): Likewise.
7031 (store_conditionalpti): Likewise.
7033 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
7034 atomic load/store instructions.
7037 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
7039 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
7041 (insn_count): New attribute, with most cases extracted from...
7042 (length): ...here. Redefine most cases in terms of insn_count.
7043 (single_insn): Delete.
7044 (can_delay): Use insn_count to check for single instructions.
7045 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
7046 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
7047 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
7048 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
7049 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
7050 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
7051 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
7052 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
7053 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
7054 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
7055 rather than "length".
7056 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
7057 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
7058 Use "insn_count" rather than "length".
7059 * config/mips/mips-dsp.md
7060 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
7061 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
7064 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
7066 PR tree-optimization/57361
7067 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
7069 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
7071 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
7073 (aarch64_simd_combine<mode>): New instruction expansion.
7074 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
7076 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
7077 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
7079 2013-06-12 Jan Hubicka <jh@suse.cz>
7081 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
7082 decl has when in streaming stage.
7083 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
7084 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
7086 2013-06-12 Roland Stigge <stigge@antcom.de>
7089 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
7091 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7093 PR tree-optimization/57537
7094 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
7095 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
7097 2013-06-12 Richard Biener <rguenther@suse.de>
7099 * data-streamer.h (streamer_write_char_stream): CSE
7100 obs->current_pointer.
7101 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
7102 streamer_write_char_stream manually and optimize the resulting loop.
7103 (streamer_write_hwi_stream): Likewise.
7105 2013-06-12 Jan Hubicka <jh@suse.cz>
7107 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
7108 * cgraph.h (varpool_create_empty_node): Declare.
7109 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
7111 * symtab.c (symtab_unregister_node): Be lax about missin entries
7113 (symtab_get_node): Update comment.
7114 * varpool.c (varpool_create_empty_node): Break out from ...
7115 (varpool_node_for_decl): ... here.
7116 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
7118 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
7120 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
7121 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
7122 part. Use straight-line flow at the end.
7123 <COMPONENT_REF>: Remove superfluous else.
7124 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
7126 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7129 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
7130 target hook even for !TREE_PUBLIC decls. If no resolution info
7131 is available, return false for common and external decls.
7133 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7135 * config/rl78/constraints.md (U): New constraint.
7136 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
7139 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7142 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
7143 to allow returning address to AT_PLATFORM name.
7145 2013-06-11 Jan Hubicka <jh@suse.cz>
7147 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
7148 * cgraph.h (symtab_node_base): Add weakref flag.
7149 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
7150 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
7151 (output_weakrefs): Use weakref flag.
7152 * fold-const.c (simple_operand_p): Handle WEAK.
7153 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
7154 * ipa.c (varpool_externally_visible_p): Drop weakref.
7155 (function_and_variable_visibility): Update comment; fix weakref
7156 sanity checks; do not clear DECL_WEAK on them.
7157 * lto-cgraph.c (lto_output_node): update.
7158 (lto_output_varpool_node): Update.
7159 (input_overwrite_node): Update.
7160 (input_node): Update.
7161 (input_varpool_node): Update.
7162 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
7163 (lto_symtab_merge_symbols): Add sanity check.
7164 (lto_symtab_prevailing_decl): Do not special case weakrefs.
7165 * passes.c (rest_of_decl_compilation): Set static flag, too.
7166 * symtab.c (dump_symtab_base): Dump weakref.
7167 (verify_symtab_base): Sanity check weakrefs.
7168 (symtab_make_decl_local): Remove duplicated code.
7169 (symtab_alias_ultimate_target): Simplify.
7170 * varpool.c (varpool_create_variable_alias): Set weakref flag.
7172 2013-06-11 DJ Delorie <dj@redhat.com>
7174 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
7175 (rl78_unwind_word_mode): New.
7177 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7179 * final.c (debug_prefix_maps): Make static.
7181 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7183 * function.c (initial_trampoline): Remove stray copy.
7185 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
7187 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
7189 2013-06-11 Martin Jambor <mjambor@suse.cz>
7191 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
7192 within bounds at the beginning of the function.
7194 2013-06-11 Alan Modra <amodra@gmail.com>
7196 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
7198 (get_named_section): Don't NULL !DECL_P decl.
7200 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
7202 * doc/invoke.texi (core-avx2): Document.
7204 (atom): Updated with MOVBE.
7206 2013-06-11 Richard Biener <rguenther@suse.de>
7208 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
7210 2013-06-11 Anton Blanchard <anton@samba.org>
7212 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
7213 correct shift value in little-endian mode.
7215 2013-06-11 Jakub Jelinek <jakub@redhat.com>
7218 * varasm.c (get_variable_align): Move #endif to the right place.
7220 2013-06-10 Cary Coutant <ccoutant@google.com>
7222 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
7223 for hash so that hash table traversal order is deterministic.
7225 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
7226 Pat Haugen <pthaugen@us.ibm.com>
7227 Peter Bergner <bergner@vnet.ibm.com>
7229 * config/rs6000/vector.md (GPR move splitter): Do not split moves
7230 of vectors in GPRS if they are direct moves or quad word load or
7233 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
7235 (direct_move_p): Likewise.
7236 (quad_load_store_p): Likewise.
7238 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
7239 classes into bins based on the physical register type.
7240 (reg_class_to_reg_type): Likewise.
7241 (IS_STD_REG_TYPE): Likewise.
7242 (IS_FP_VECT_REG_TYPE): Likewise.
7243 (reload_fpr_gpr): Arrays to determine what insn to use if we can
7244 use direct move instructions.
7245 (reload_gpr_vsx): Likewise.
7246 (reload_vsx_gpr): Likewise.
7247 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
7248 information that is a simplification of register classes. Also
7249 precalculate direct move reload helpers.
7250 (direct_move_p): New function to return true if the operation can
7251 be done as a direct move instruciton.
7252 (quad_load_store_p): New function to return true if the operation
7253 is a quad memory operation.
7254 (rs6000_legitimize_address): If quad memory, only allow register
7255 indirect for TImode addresses.
7256 (rs6000_legitimate_address_p): Likewise.
7257 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
7258 (rs6000_reload_register_type): Likewise.
7259 (register_to_reg_type): Return register type.
7260 (rs6000_secondary_reload_simple_move): New helper function for
7261 secondary reload and secondary memory needed to identify anything
7262 that is a simple move, and does not need reloading.
7263 (rs6000_secondary_reload_direct_move): New helper function for
7264 secondary reload to identify cases that can be done with several
7265 instructions via the direct move instructions.
7266 (rs6000_secondary_reload_move): New helper function for secondary
7267 reload to identify moves between register types that can be done.
7268 (rs6000_secondary_reload): Add support for quad memory operations
7269 and for direct move.
7270 (rs6000_secondary_memory_needed): Likewise.
7271 (rs6000_debug_secondary_memory_needed): Change argument names.
7272 (rs6000_output_move_128bit): New function to return the move to
7273 use for 128-bit moves, including knowing about the various
7274 limitations of quad memory operations.
7276 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
7277 memory operations. call rs6000_output_move_128bit for the actual
7278 instruciton(s) to generate.
7279 (vsx_movti_64bit): Likewise.
7281 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
7282 (UNSPEC_P8V_MTVSRWZ): Likewise.
7283 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
7284 (UNSPEC_P8V_MTVSRD): Likewise.
7285 (UNSPEC_P8V_XXPERMDI): Likewise.
7286 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
7287 (UNSPEC_FUSION_GPR): Likewise.
7288 (FMOVE128_GPR): New iterator for direct move.
7289 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
7292 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
7293 loads and direct move instructions.
7294 (zero_extendsidi2_lfiwzx): Likewise.
7295 (extendsidi2_lfiwax): Likewise.
7296 (extendsidi2_nocell): Likewise.
7297 (floatsi<mode>2_lfiwax): Likewise.
7299 (floatunssi<mode>2_lfiwzx): Likewise.
7301 (fix_trunc<mode>_stfiwx): Likewise.
7302 (fixuns_trunc<mode>_stfiwx): Likewise.
7303 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
7304 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
7305 (parity<mode>2_cmpb): Set length/type attr.
7306 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
7307 for 'mr.' to fast_compare.
7308 (bpermd_<mode>): Change type attr to popcnt.
7309 (p8_fmrgow_<mode>): New insns for power8 direct move support.
7310 (p8_mtvsrwz_1): Likewise.
7311 (p8_mtvsrwz_2): Likewise.
7312 (reload_fpr_from_gpr<mode>): Likewise.
7313 (p8_mtvsrd_1): Likewise.
7314 (p8_mtvsrd_2): Likewise.
7315 (p8_xxpermdi_<mode>): Likewise.
7316 (reload_vsx_from_gpr<mode>): Likewise.
7317 (reload_vsx_from_gprsf): Likewise.
7318 (p8_mfvsrd_3_<mode>): LIkewise.
7319 (reload_gpr_from_vsx<mode>): Likewise.
7320 (reload_gpr_from_vsxsf): Likewise.
7321 (p8_mfvsrd_4_disf): Likewise.
7322 (multi-word GPR splits): Do not split direct moves or quad memory
7325 2013-06-10 David Malcolm <dmalcolm@redhat.com>
7327 * tree-into-ssa.c (interesting_blocks): Make static.
7329 2013-06-10 Jakub Jelinek <jakub@redhat.com>
7332 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
7333 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
7334 Use DATA_ABI_ALIGNMENT for that case instead if defined.
7335 (get_variable_align): New function.
7336 (get_variable_section, emit_bss, emit_common,
7337 assemble_variable_contents, place_block_symbol): Use
7338 get_variable_align instead of DECL_ALIGN.
7339 (assemble_noswitch_variable): Add align argument, use it
7340 instead of DECL_ALIGN.
7341 (assemble_variable): Adjust caller. Use get_variable_align
7342 instead of DECL_ALIGN.
7343 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
7345 (DATA_ABI_ALIGNMENT): Define.
7346 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
7347 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
7348 opt is false, only return the psABI mandated alignment increase.
7349 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
7350 (DATA_ABI_ALIGNMENT): ... this.
7351 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
7352 (DATA_ABI_ALIGNMENT): ... this.
7353 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
7354 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
7355 (DATA_ABI_ALIGNMENT): ... this.
7356 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
7357 * doc/tm.texi: Regenerated.
7359 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
7361 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
7362 cmp_code to construct REG_EQUAL note.
7364 2013-06-09 Jakub Jelinek <jakub@redhat.com>
7367 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
7368 that operands[2] doesn't overlap with operands[0].
7370 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
7371 Jan Hubicka <jh@suse.cz>
7373 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
7374 hack to mark symbols as used.
7376 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
7378 PR rtl-optimization/57559
7379 * lra-constraints.c (process_alt_operands): Don't discourage
7380 memory with known offset for offsetable memory constraint.
7381 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
7383 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
7385 * varasm.c (struct oc_local_state): Reorder fields.
7386 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
7387 and adjust accordingly.
7388 (output_constructor): Reorder initialization code and adjust call to
7389 output_constructor_bitfield.
7391 2013-06-07 Jan Hubicka <jh@suse.cz>
7393 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
7395 2013-06-07 David Malcolm <dmalcolm@redhat.com>
7397 * tree-object-size.c (unknown): Make const.
7399 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7401 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
7402 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
7403 for last alternative in the cpu_facility attribute.
7405 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7408 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
7409 (xordi3): Change operand 2 constraint to arm_xordi_operand.
7410 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
7411 * config/arm/constraints.md (Dg): New constraint.
7412 * config/arm/neon.md (xordi3_neon): Remove.
7413 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
7414 * config/arm/predicates.md (arm_xordi_operand): New predicate.
7416 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7418 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
7419 Clean up alternatives.
7421 2013-06-07 Alan Modra <amodra@gmail.com>
7423 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
7426 2013-06-07 Alan Modra <amodra@gmail.com>
7428 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
7430 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7432 * config/arm/constraints.md (Df): New constraint.
7433 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
7434 Correct length attribute for last two alternatives.
7436 2013-06-07 Alan Modra <amodra@gmail.com>
7438 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7439 override user -mfp-in-toc.
7440 (offsettable_ok_by_alignment): Consider just the current access
7441 rather than the whole object, unless BLKmode. Handle
7442 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
7443 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
7444 for -mcmodel=medium.
7445 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
7446 override user -mfp-in-toc or -msum-in-toc. Default to
7447 -mno-fp-in-toc for -mcmodel=medium.
7449 2013-06-06 DJ Delorie <dj@redhat.com>
7451 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
7452 TARGET_VALID_POINTER_MODE.
7454 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
7455 Pat Haugen <pthaugen@us.ibm.com>
7456 Peter Bergner <bergner@vnet.ibm.com>
7458 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7459 Document new power8 builtins.
7461 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
7462 condition code register, to allow 128-bit logical operations to be
7463 done in the VSX or GPR registers.
7464 (nor<mode>3): Use the canonical form for nor.
7465 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
7466 vclz*, and vpopcnt* vector instructions.
7467 (nand<mode>3): Likewise.
7468 (orc<mode>3): Likewise.
7469 (clz<mode>2): LIkewise.
7470 (popcount<mode>2): Likewise.
7472 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
7473 that only the GPRs are recognized.
7475 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7476 support for new power8 builtins.
7478 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
7480 (xscvdpspn): Likewise.
7486 (vpopcnt): Likewise.
7487 (vpopcntb): Likewise.
7488 (vpopcnth): Likewise.
7489 (vpopcntw): Likewise.
7490 (vpopcntd): Likewise.
7495 (eqv_v16qi3): Likewise.
7496 (eqv_v8hi3): Likewise.
7497 (eqv_v4si3): Likewise.
7498 (eqv_v2di3): Likewise.
7499 (eqv_v4sf3): Likewise.
7500 (eqv_v2df3): Likewise.
7502 (nand_v16qi3): Likewise.
7503 (nand_v8hi3): Likewise.
7504 (nand_v4si3): Likewise.
7505 (nand_v2di3): Likewise.
7506 (nand_v4sf3): Likewise.
7507 (nand_v2df3): Likewise.
7509 (orc_v16qi3): Likewise.
7510 (orc_v8hi3): Likewise.
7511 (orc_v4si3): Likewise.
7512 (orc_v2di3): Likewise.
7513 (orc_v4sf3): Likewise.
7514 (orc_v2df3): Likewise.
7516 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
7517 allow power8 quad mode in 64-bit.
7518 (rs6000_builtin_vectorized_function): Add support to vectorize
7519 ISA 2.07 count leading zeros, population count builtins.
7520 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
7521 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
7522 (builtin_function_type): Add vgbbd builtin function which takes an
7524 (altivec_expand_vec_perm_const): Add support for new power8 merge
7527 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
7528 that does not include TImdoe for use with 32-bit.
7529 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
7531 (UNSPEC_VSX_CVDPSPN): Likewise.
7532 (vsx_xscvdpspn): Likewise.
7533 (vsx_xscvspdpn): Likewise.
7534 (vsx_xscvdpspn_scalar): Likewise.
7535 (vsx_xscvspdpn_directmove): Likewise.
7536 (vsx_and<mode>3): Split logical operations into 32-bit and
7537 64-bit. Add support to do logical operations on TImode as well as
7538 VSX vector types. Allow logical operations to be done in either
7539 VSX registers or in general purpose registers in 64-bit mode. Add
7540 splitters if GPRs were used. For AND, add clobber of CCmode to
7541 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
7543 (vsx_and<mode>3_32bit): Likewise.
7544 (vsx_and<mode>3_64bit): Likewise.
7545 (vsx_ior<mode>3): Likewise.
7546 (vsx_ior<mode>3_32bit): Likewise.
7547 (vsx_ior<mode>3_64bit): Likewise.
7548 (vsx_xor<mode>3): Likewise.
7549 (vsx_xor<mode>3_32bit): Likewise.
7550 (vsx_xor<mode>3_64bit): Likewise.
7551 (vsx_one_cmpl<mode>2): Likewise.
7552 (vsx_one_cmpl<mode>2_32bit): Likewise.
7553 (vsx_one_cmpl<mode>2_64bit): Likewise.
7554 (vsx_nor<mode>3): Likewise.
7555 (vsx_nor<mode>3_32bit): Likewise.
7556 (vsx_nor<mode>3_64bit): Likewise.
7557 (vsx_andc<mode>3): Likewise.
7558 (vsx_andc<mode>3_32bit): Likewise.
7559 (vsx_andc<mode>3_64bit): Likewise.
7560 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
7561 and xxlorc instructions.
7562 (vsx_eqv<mode>3_64bit): Likewise.
7563 (vsx_nand<mode>3_32bit): Likewise.
7564 (vsx_nand<mode>3_64bit): Likewise.
7565 (vsx_orc<mode>3_32bit): Likewise.
7566 (vsx_orc<mode>3_64bit): Likewise.
7568 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
7570 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
7572 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
7573 (p8_vmrgow): Likewise.
7574 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
7575 GPRs to be split under VSX.
7576 (p8v_clz<mode>2): Add power8 count leading zero support.
7577 (p8v_popcount<mode>2): Add power8 population count support.
7578 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
7581 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
7584 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
7586 (vec_nand): Likewise.
7587 (vec_vclz): Likewise.
7588 (vec_vclzb): Likewise.
7589 (vec_vclzd): Likewise.
7590 (vec_vclzh): Likewise.
7591 (vec_vclzw): Likewise.
7592 (vec_vgbbd): Likewise.
7593 (vec_vmrgew): Likewise.
7594 (vec_vmrgow): Likewise.
7595 (vec_vpopcnt): Likewise.
7596 (vec_vpopcntb): Likewise.
7597 (vec_vpopcntd): Likewise.
7598 (vec_vpopcnth): Likewise.
7599 (vec_vpopcntw): Likewise.
7601 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7603 PR rtl-optimization/57468
7604 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
7607 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7609 PR rtl-optimization/57459
7610 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
7611 type when setting live regs.
7613 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7615 * config/s390/s390.opt (mlra): New option.
7616 * config/s390/s390.c (s390_decompose_address): Check displacement
7617 for all registers for LRA.
7618 (s390_secondary_reload): Don't used secondary reloads for LRA.
7619 (s390_lra_p): New function.
7620 (TARGET_LRA_P): Define.
7621 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
7622 of attribute cpu_facility to zarch for the last alternative.
7623 (*cmpmem_short): Ditto.
7625 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
7627 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
7628 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
7629 (arm_expand_prologue): Likewise.
7631 2013-06-06 Teresa Johnson <tejohnson@google.com>
7634 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
7635 as this is now done by redirect_edge_and_branch_force.
7636 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
7637 barriers, and fix interaction with splitting.
7638 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
7639 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
7640 reflect changes made in the routine.
7641 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
7642 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
7643 since this is called in cfglayout mode, and replace partition fixup
7644 with assert as that is now done by force_nonfallthru_and_redirect.
7645 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
7646 already be marked with region crossing note.
7647 (insert_section_boundary_note): Make non-static, gate on flag
7648 has_bb_partition, rewrite to also check for multiple partitions.
7649 (rest_of_handle_reorder_blocks): Remove call to
7650 insert_section_boundary_note, now done later during free_cfg.
7651 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
7652 * bb-reorder.h (insert_section_boundary_note): Declare.
7653 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
7654 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
7655 invoke insert_section_boundary_note.
7656 (try_redirect_by_replacing_jump): Remove unnecessary
7657 check for region crossing note.
7658 (fixup_partition_crossing): New function.
7659 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
7660 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
7661 in non-cfglayout mode.
7662 (force_nonfallthru_and_redirect): Fixup partition boundaries,
7663 remove old code that tried to do this. Emit barrier correctly
7664 when we are in cfglayout mode.
7665 (last_bb_in_partition): New function.
7666 (rtl_split_edge): Correctly fixup partition boundaries.
7667 (commit_one_edge_insertion): Remove old code that tried to
7668 fixup region crossing edge since this is now handled in
7669 split_block, and set up insertion point correctly since
7670 block may now end in a jump.
7671 (verify_hot_cold_block_grouping): Guard against checking when not in
7672 linearized RTL mode.
7673 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
7675 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
7676 rtl_verify_flow_info, so not called in cfglayout mode.
7677 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
7678 (fixup_reorder_chain): Remove old code that attempted to fixup region
7679 crossing note as this is now handled in force_nonfallthru_and_redirect.
7680 (duplicate_insn_chain): Don't duplicate switch section notes.
7681 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
7683 * basic-block.h (emit_barrier_after_bb): Declare.
7685 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7687 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
7688 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
7689 arm_usatsihi): Adjust alternatives for arm_restrict_it.
7691 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7693 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
7695 * config/arm/ldmstm.md: Regenerate.
7697 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7699 * config/arm/sync.md (atomic_loaddi_1):
7700 Disable predication for arm_restrict_it.
7701 (arm_load_exclusive<mode>): Likewise.
7702 (arm_load_exclusivesi): Likewise.
7703 (arm_load_exclusivedi): Likewise.
7704 (arm_load_acquire_exclusive<mode>): Likewise.
7705 (arm_load_acquire_exclusivesi): Likewise.
7706 (arm_load_acquire_exclusivedi): Likewise.
7707 (arm_store_exclusive<mode>): Likewise.
7708 (arm_store_exclusive<mode>): Likewise.
7709 (arm_store_release_exclusivedi): Likewise.
7710 (arm_store_release_exclusive<mode>): Likewise.
7712 2013-06-06 Richard Biener <rguenther@suse.de>
7714 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
7716 (lto_tag_is_tree_code_p): Adjust.
7717 (lto_tag_is_gimple_code_p): Likewise.
7718 (lto_gimple_code_to_tag): Likewise.
7719 (lto_tag_to_gimple_code): Likewise.
7720 (lto_tree_code_to_tag): Likewise.
7721 (lto_tag_to_tree_code): Likewise.
7722 * data-streamer.h (streamer_write_hwi_in_range): Use
7723 uhwi streaming to stream the normalized range.
7724 (streamer_read_hwi_in_range): Likewise.
7726 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7728 * config/arm/arm.md (enabled_for_depr_it): New attribute.
7729 (predicable_short_it): Likewise.
7730 (predicated): Likewise.
7731 (enabled): Handle above.
7732 (define_cond_exec): Set predicated attribute to yes.
7734 2013-06-05 Mike Stump <mikestump@comcast.net>
7736 * gdbinit.in (__FUNCTION__): Add.
7738 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
7740 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
7741 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
7743 2013-06-05 Jan Hubicka <jh@suse.cz>
7745 * varasm.c (mark_decl_referenced): Revert the removal until targets
7748 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
7750 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
7751 instead of mark_decl_referenced.
7753 2013-06-05 Jan Hubicka <jh@suse.cz>
7755 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
7756 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
7757 and symtab_used_from_object_file_p.
7758 (cgraph_make_node_local_1): Clear forced_by_abi.
7759 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
7760 * cgraph.h (symtab_node_base): Add forced_by_abi.
7761 (decide_is_variable_needed): Remove.
7762 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
7763 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
7764 (decide_is_symbol_needed): ... this one; handle symbols in general;
7765 always analyze virtuals; honnor forced_by_abi.
7766 (cgraph_finalize_function): Update.
7767 (varpool_finalize_decl): Update.
7768 (symbol_defined_and_needed): Remove.
7769 (analyze_functions): Update.
7770 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
7771 output_refs, input_overwrite_node): Handle forced_by_abi.
7772 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
7773 (address_taken_from_non_vtable_p): ... this one.
7774 (comdat_can_be_unshared_p_1): New function.
7775 (cgraph_comdat_can_be_unshared_p): Rename to ...
7776 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
7777 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
7778 (function_and_variable_visibility): Clear forced_by_abi as needed.
7779 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
7780 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
7781 * symtab.c (dump_symtab_base): Dump forced_by_abi.
7782 * varpool.c (decide_is_variable_needed): Remove.
7784 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7786 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
7787 (arm_option_override): Override arm_restrict_it where appropriate.
7788 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
7789 * config/arm/arm.opt (mrestrict-it): New command-line option.
7790 * doc/invoke.texi: Document -mrestrict-it.
7792 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7794 * tsan.c (tsan_atomic_table): Make const.
7796 2013-06-05 Richard Biener <rguenther@suse.de>
7798 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
7799 index associated with the tree we are supposed to replace.
7800 * tree-streamer-out.c (pack_ts_base_value_fields): Output
7801 TREE_ASM_WRITTEN as zero for everything but SSA names.
7803 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7805 * tree-ssa-structalias.c (call_stmt_vars): Make static.
7807 2013-06-04 Jan Hubicka <jh@suse.cz>
7809 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
7810 (input_node, input_varpool_node): Handle correctly external same
7812 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
7813 nodes at ltrans stage.
7815 2013-06-04 Jan Hubicka <jh@suse.cz>
7817 * ipa-inline.c (update_caller_keys): Fix availability test.
7818 (update_callee_keys): Likewise.
7819 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
7820 to follow ELF standard.
7822 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
7824 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
7825 (mips64r5900el-*-elf*): New configurations.
7826 * config/mips/mips-cpus.def (r5900): New processor.
7827 * config/mips/mips-tables.opt: Regenerate.
7828 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
7829 (mips_issue_rate): Handle PROCESSOR_R5900.
7830 (mips_reorg_process_insns): Force reorder mode for the R5900.
7831 * config/mips/mips.h (TARGET_MIPS5900): Define.
7832 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
7834 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
7836 * config/mips/mips.md (processor): Add r5900.
7837 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
7839 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7841 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
7842 into function to generate MOVI instruction.
7843 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
7844 (aarch64_preferred_simd_mode): Turn into wrapper.
7845 (aarch64_output_scalar_simd_mov_immediate): New function.
7846 * config/aarch64/aarch64-protos.h: Add prototype for above.
7848 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7850 * config/aarch64/aarch64.c (simd_immediate_info): Remove
7851 element_char member.
7852 (sizetochar): Return signed char.
7853 (aarch64_simd_valid_immediate): Remove elchar and other
7854 unnecessary variables.
7855 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
7856 Calculate element_char as required.
7857 * config/aarch64/aarch64-protos.h: Update and move prototype
7858 for aarch64_output_simd_mov_immediate.
7859 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
7862 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7864 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
7865 information completed by aarch64_simd_valid_immediate.
7866 (aarch64_legitimate_constant_p): Update arguments.
7867 (aarch64_simd_valid_immediate): Work with struct rather than many
7869 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
7870 (aarch64_simd_make_constant): Update arguments.
7871 (aarch64_output_simd_mov_immediate): Work with struct rather than
7872 many pointers. Output immediate directly rather than as operand.
7873 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7875 * config/aarch64/constraints.md (Dn): Update arguments.
7877 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7879 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
7881 (aarch64_simd_immediate_valid_for_move): Remove.
7882 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
7883 (aarch64_simd_make_constant): Update call.
7884 (aarch64_output_simd_mov_immediate): Update call.
7885 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7887 * config/aarch64/constraints.md (Dn): Update call.
7889 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7891 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
7892 return type to bool for prototype.
7893 (aarch64_legitimate_constant_p): Check for true instead of not -1.
7894 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
7895 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
7897 2013-06-04 Catherine Moore <clm@codesourcery.com>
7899 * config/mips/mips.opt (meva): New.
7900 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7901 (ASM_SPEC): Handle -meva.
7902 * doc/invoke.texi (meva): Document.
7904 2013-06-04 Alan Modra <amodra@gmail.com>
7906 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7909 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7911 * rtl.def: Add extra fourth optional field to define_cond_exec.
7912 * gensupport.c (process_one_cond_exec): Process attributes from
7914 * doc/md.texi: Document fourth field in define_cond_exec.
7916 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7918 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7919 out the processing order as in store_bit_field_1.
7921 2013-06-04 Jan Hubicka <jh@suse.cz>
7924 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7925 non-VAR_DECL node if it does not exist yet.
7927 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7929 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7930 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7931 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7932 target_cpu_default setting.
7934 2013-06-03 Teresa Johnson <tejohnson@google.com>
7936 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7937 default to -fopt-info=optimized instead of all.
7938 * doc/invoke.texi: Ditto.
7939 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7940 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7941 (execute_vect_slp): Emit BB vectorization success under
7942 MSG_OPTIMIZED_LOCATIONS.
7943 * tree-vect-slp.c (vect_slp_transform_bb): Change
7944 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7945 * tree-vect-loop.c (vect_transform_loop): Ditto.
7947 2013-06-03 Jason Merrill <jason@redhat.com>
7950 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7951 Use TARGET_EXPR for C++.
7953 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7955 PR rtl-optimization/57268
7956 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7957 if DEBUG_INSN_P (insn).
7960 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7962 PR rtl-optimization/57268
7963 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7964 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7966 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7968 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7969 (ix86_avoid_lea_for_addr): Likewise.
7970 (exact_dependency_1): Likewise.
7971 (ix86_adjust_cost): Likewise.
7972 (swap_top_of_ready_list): Fix formatting and !reload_completed check
7974 (do_reorder_for_imul): Fix typo, formatting and
7975 !reload_completed check removed.
7976 (ix86_sched_reorder): Fix typo and formatting.
7977 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
7980 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
7982 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
7984 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
7986 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
7987 <STRING_CST>: Likewise.
7988 <VECTOR_CST>: Likewise.
7990 2013-06-01 Janus Weil <janus@gcc.gnu.org>
7991 Mikael Morin <mikael@gcc.gnu.org>
7993 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
7994 * config.in: Regenerated.
7995 * configure: Regenerated.
7997 2013-06-01 Jan Hubicka <jh@suse.cz>
8000 * cgraphunit.c (compile): When weakref is not supported,
8001 set up transparent aliases before final output pass.
8002 * varasm.c (assemble_alias): Do not try to do it here.
8004 2013-06-01 Jan Hubicka <jh@suse.cz>
8007 * passes.c (for_per_function): Skip unanalyzed functions.
8009 2013-06-01 Jan Hubicka <jh@suse.cz>
8011 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
8012 (lto_symtab_merge_symbols_1): ... this one.
8013 (lto_symtab_merge_cgraph_nodes): Rename to ...
8014 (lto_symtab_merge_symbols): ... this one; simplify.
8015 * cgraph.c (same_body_aliases_done): Rename to ...
8016 (cpp_implicit_aliases_done): ... this one.
8017 (cgraph_create_function_alias): Update.
8018 (cgraph_same_body_alias): Update.
8019 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
8020 (verify_edge_corresponds_to_fndecl): Simplify.
8021 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
8022 (cgraph_node): Remove same_body_alias.
8023 (varpool_node): Remove alias_of and extra_name_alias.
8024 (same_body_aliases_done): Rename to ..
8025 (cpp_implicit_aliases_done): ... this one.
8026 (symtab_alias_ultimate_target): Add default parameter.
8027 (symtab_resolve_alias): New function.
8028 (fixup_same_cpp_alias_visibility): Declare.
8029 (cgraph_function_node): Add default parameter.
8030 (cgraph_node_asm_name): Likewise.
8031 (cgraph_function_or_thunk_node): Add default parameter; do
8032 not ICE when it is NULL.
8033 (varpool_variable_node): Likewise.
8034 * tree-emutls.c (create_emultls_var): Update.
8035 (ipa_lower_emutls): Update.
8036 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
8037 (cgraph_reset_node): Reset alias info.
8038 (cgraph_finalize_function): Update.
8039 (fixup_same_cpp_alias_visibility): Move to symtab.c.
8040 (analyze_function): Simplify.
8041 (cgraph_process_same_body_aliases): Simplify.
8042 (analyze_functions): Fixup same body aliases.
8043 (handle_alias_pairs): Simplify.
8044 (assemble_thunk): Update.
8045 (assemble_thunks_and_aliases): Update.
8046 (output_weakrefs): Rewrite.
8047 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
8048 (lto_output_varpool_node): Likewise.
8049 (compute_ltrans_boundary): Remve assert.
8050 (get_alias_symbol): New functoin.
8051 (input_node): Rewrite alias handling.
8052 (input_varpool_node): Likewise.
8053 * ipa-pure-const.c (propagate_pure_const): Fix formating.
8054 * ipa.c (process_references): Handle weakrefs correctly.
8055 (symtab_remove_unreachable_nodes): Likewise.
8056 * trans-mem.c (get_cg_data): Update.
8057 (ipa_tm_create_version_alias): Update.
8058 (ipa_tm_execute): Update.
8059 * symtab.c (dump_symtab_base): Dump aliases.
8060 (verify_symtab_base): Verify aliases.
8061 (symtab_node_availability): New function.
8062 (symtab_alias_ultimate_target): Simplify.
8063 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
8064 handle all the fixup cases.
8065 (symtab_resolve_alias): New function.
8066 * passes.c (ipa_write_summaries): Handle weakrefs.
8067 * varpool.c (varpool_analyze_node): Simplify.
8068 (assemble_aliases): Update.
8069 (varpool_create_variable_alias): Simplify.
8070 (varpool_extra_name_alias): Simplify.
8071 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
8072 (lto_symtab_merge_symbols): ... this one.
8074 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
8077 PR rtl-optimization/57268
8078 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8079 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8081 2013-06-01 Tobias Burnus <burnus@net-b.de>
8084 2013-05-31 Tobias Burnus <burnus@net-b.de>
8087 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
8090 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8092 PR rtl-optimization/57268
8093 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8094 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8096 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8098 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
8099 unordered comparison operators when -fno-trapping-math is in effect
8101 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
8102 and implement unordered comparison operators properly on the e500.
8104 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8106 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
8107 for constant scalar integers.
8108 (simplify_relational_operation_1): Likewise.
8110 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
8112 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
8113 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
8116 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8117 Igor Zamyatin <igor.zamyatin@intel.com>
8119 Silvermont (SLM) architecture performance tuning.
8120 * config/i386/i386.h (enum ix86_tune_indices): Add
8121 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
8122 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
8124 * config/i386/i386.c (initial_ix86_tune_features)
8125 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
8126 (ix86_lea_outperforms): Handle Silvermont tuning.
8127 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
8129 (ix86_use_lea_for_mov): Likewise.
8130 (ix86_avoid_lea_for_addr): Likewise.
8131 (ix86_lea_for_add_ok): Likewise.
8132 (exact_dependency_1): New function.
8133 (exact_store_load_dependency): Likewise.
8134 (ix86_adjust_cost): Handle Silvermont tuning.
8135 (do_reoder_for_imul): Likewise.
8136 (swap_top_of_ready_list): New function.
8137 (ix86_sched_reorder): Changed to handle Silvermont tuning.
8139 * config/i386/i386.md (peepholes that split memory operand in fp
8142 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8144 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
8145 Remove un-necessary braces.
8147 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8149 * config/aarch64/aarch64.c (aarch64_classify_symbol):
8150 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
8152 2013-05-31 Tobias Burnus <burnus@net-b.de>
8155 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
8157 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8160 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
8161 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
8162 * config/arm/neon.md (iordi3_neon): Remove.
8163 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
8164 * config/arm/predicates.md (imm_for_neon_logic_operand):
8165 Move to earlier in the file.
8166 (neon_logic_op2): Likewise.
8167 (arm_iordi_operand_neon): New predicate.
8169 2013-05-31 Richard Biener <rguenther@suse.de>
8171 PR tree-optimization/57478
8172 PR tree-optimization/57453
8173 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
8176 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8178 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
8179 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
8181 2013-05-30 Tobias Burnus <burnus@net-b.de>
8182 Thomas Koenig <tkoenig@gcc.gnu.org>
8185 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
8186 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
8188 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
8190 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
8192 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
8194 * target.def (register_usage_leveling_p): New hook.
8195 * targhooks.c (default_register_usage_leveling_p): New.
8196 * targhooks.h (default_register_usage_leveling_p): New prototype.
8197 * lra-assigns.c (register_usage_leveling_p): Use the hook.
8198 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
8199 * doc/tm.texi: Update.
8200 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
8202 2013-05-30 Ian Bolton <ian.bolton@arm.com>
8204 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
8205 (*insv_reg<mode>): New define_insn.
8207 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
8209 PR rtl-optimization/57439
8210 * postreload.c (move2add_valid_value_p): Check that we have
8211 a zero subreg_regno_offset when accessing the register in
8214 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8215 Igor Zamyatin <igor.zamyatin@intel.com>
8217 Silvermont (SLM) architecture pipeline model, tuning and
8219 * config.gcc: Add slm config options and target.
8221 * config/i386/slm.md: New.
8223 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
8225 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
8227 (ix86_target_macros_internal): Likewise.
8229 * gcc/config/i386/i386.c (slm_cost): New cost.
8230 (m_SLM): New macro flag.
8231 (initial_ix86_tune_features): Set m_SLM.
8232 (x86_accumulate_outgoing_args): Likewise.
8233 (x86_arch_always_fancy_math_387): Likewise.
8234 (processor_target_table): Add slm cost.
8235 (cpu_names): Add slm cpu name.
8236 (x86_option_override_internal): Set SLM ISA.
8237 (ix86_issue_rate): New case PROCESSOR_SLM.
8238 (ia32_multipass_dfa_lookahead): Likewise.
8239 (fold_builtin_cpu): Add slm.
8241 * config/i386/i386.h (TARGET_SLM): New target macro.
8242 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
8243 (processor_type): Add PROCESSOR_SLM.
8245 * config/i386/i386.md (cpu): Add new value "slm".
8246 (slm.md): Include slm.md.
8248 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
8249 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8251 * config/arm/arm-protos.h: Add and update function protos.
8252 * config/arm/arm.c (use_simple_return_p): New added.
8253 (thumb2_expand_return): Check simple_return flag.
8254 * config/arm/arm.md: Add simple_return and conditional simple_return.
8255 * config/arm/iterators.md: Add iterator for return and simple_return.
8257 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8259 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
8260 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
8261 (arm_emit_vfp_multi_reg_pop): Likewise.
8262 (thumb2_emit_ldrd_pop): Likewise.
8263 (arm_expand_epilogue): Add misc REG_CFA notes.
8264 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
8266 2013-05-29 Lawrence Crowl <crowl@google.com>
8268 * config/arm/t-arm: Update for below.
8270 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
8271 Change type to hash_table. Update dependent calls and types.
8273 * config/i386/t-cygming: Update for below.
8275 * config/i386/t-interix: Update for below.
8277 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
8278 Change type to hash_table. Update dependent calls and types.
8279 (i386_find_on_wrapper_list::wrappers): Likewise.
8281 * config/ia64/t-ia64: Update for below.
8283 * config/ia64/ia64.c (bundle_state_table):
8284 Change type to hash_table. Update dependent calls and types.
8286 * config/mips/mips.c (mips_reorg_process_insns::htab):
8287 Change type to hash_table. Update dependent calls and types.
8289 * config/sol2.c (solaris_comdat_htab):
8290 Change type to hash_table. Update dependent calls and types.
8292 * config/t-sol2: Update for above.
8294 2013-05-29 Teresa Johnson <tejohnson@google.com>
8296 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
8297 functions are not yet marked as defined.
8299 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
8300 Pat Haugen <pthaugen@us.ibm.com>
8301 Peter Bergner <bergner@vnet.ibm.com>
8303 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
8307 (vrotl<mode>3): Likewise.
8308 (vashl<mode>3): Likewise.
8309 (vlshr<mode>3): Likewise.
8310 (vashr<mode>3): Likewise.
8312 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8313 support for power8 V2DI builtins.
8315 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
8316 power8 V2DI builtins.
8317 (vupkhsw): Likewise.
8318 (vupklsw): Likewise.
8319 (vaddudm): Likewise.
8324 (vpkudum): Likewise.
8325 (vpksdss): Likewise.
8326 (vpkudus): Likewise.
8327 (vpksdus): Likewise.
8332 (vsubudm): Likewise.
8333 (vcmpequd): Likewise.
8334 (vcmpgtsd): Likewise.
8335 (vcmpgtud): Likewise.
8336 (vcmpequd_p): Likewise.
8337 (vcmpgtsd_p): Likewise.
8338 (vcmpgtud_p): Likewise.
8339 (vupkhsw): Likewise.
8340 (vupklsw): Likewise.
8341 (vaddudm): Likewise.
8346 (vpksdss): Likewise.
8347 (vpksdus): Likewise.
8348 (vpkudum): Likewise.
8349 (vpkudus): Likewise.
8354 (vsubudm): Likewise.
8356 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
8357 support for power8 V2DI instructions.
8359 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
8360 power8 V2DI instructions. Combine pack and unpack insns to use an
8361 iterator for each mode. Check whether a particular mode supports
8362 Altivec instructions instead of just checking TARGET_ALTIVEC.
8363 (UNSPEC_VPKUWUM): Likewise.
8364 (UNSPEC_VPKSHSS): Likewise.
8365 (UNSPEC_VPKSWSS): Likewise.
8366 (UNSPEC_VPKUHUS): Likewise.
8367 (UNSPEC_VPKSHUS): Likewise.
8368 (UNSPEC_VPKUWUS): Likewise.
8369 (UNSPEC_VPKSWUS): Likewise.
8370 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
8371 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
8372 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
8373 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
8374 (UNSPEC_VUPKHSB): Likewise.
8375 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
8376 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
8377 (UNSPEC_VUPKHSH): Likewise.
8378 (UNSPEC_VUPKLSB): Likewise.
8379 (UNSPEC_VUPKLSH): Likewise.
8381 (VI_char): Likewise.
8382 (VI_scalar): Likewise.
8383 (VI_unit): Likewise.
8385 (VP_small): Likewise.
8386 (VP_small_lc): Likewise.
8387 (VU_char): Likewise.
8388 (add<mode>3): Likewise.
8389 (altivec_vaddcuw): Likewise.
8390 (altivec_vaddu<VI_char>s): Likewise.
8391 (altivec_vadds<VI_char>s): Likewise.
8392 (sub<mode>3): Likewise.
8393 (altivec_vsubcuw): Likewise.
8394 (altivec_vsubu<VI_char>s): Likewise.
8395 (altivec_vsubs<VI_char>s): Likewise.
8396 (altivec_vavgs<VI_char>): Likewise.
8397 (altivec_vcmpbfp): Likewise.
8398 (altivec_eq<mode>): Likewise.
8399 (altivec_gt<mode>): Likewise.
8400 (altivec_gtu<mode>): Likewise.
8401 (umax<mode>3): Likewise.
8402 (smax<mode>3): Likewise.
8403 (umin<mode>3): Likewise.
8404 (smin<mode>3): Likewise.
8405 (altivec_vpkuhum): Likewise.
8406 (altivec_vpkuwum): Likewise.
8407 (altivec_vpkshss): Likewise.
8408 (altivec_vpkswss): Likewise.
8409 (altivec_vpkuhus): Likewise.
8410 (altivec_vpkshus): Likewise.
8411 (altivec_vpkuwus): Likewise.
8412 (altivec_vpkswus): Likewise.
8413 (altivec_vpks<VI_char>ss): Likewise.
8414 (altivec_vpks<VI_char>us): Likewise.
8415 (altivec_vpku<VI_char>us): Likewise.
8416 (altivec_vpku<VI_char>um): Likewise.
8417 (altivec_vrl<VI_char>): Likewise.
8418 (altivec_vsl<VI_char>): Likewise.
8419 (altivec_vsr<VI_char>): Likewise.
8420 (altivec_vsra<VI_char>): Likewise.
8421 (altivec_vsldoi_<mode>): Likewise.
8422 (altivec_vupkhsb): Likewise.
8423 (altivec_vupkhs<VU_char>): Likewise.
8424 (altivec_vupkls<VU_char>): Likewise.
8425 (altivec_vupkhsh): Likewise.
8426 (altivec_vupklsb): Likewise.
8427 (altivec_vupklsh): Likewise.
8428 (altivec_vcmpequ<VI_char>_p): Likewise.
8429 (altivec_vcmpgts<VI_char>_p): Likewise.
8430 (altivec_vcmpgtu<VI_char>_p): Likewise.
8431 (abs<mode>2): Likewise.
8432 (vec_unpacks_hi_v16qi): Likewise.
8433 (vec_unpacks_hi_v8hi): Likewise.
8434 (vec_unpacks_lo_v16qi): Likewise.
8435 (vec_unpacks_hi_<VP_small_lc>): Likewise.
8436 (vec_unpacks_lo_v8hi): Likewise.
8437 (vec_unpacks_lo_<VP_small_lc>): Likewise.
8438 (vec_pack_trunc_v8h): Likewise.
8439 (vec_pack_trunc_v4si): Likewise.
8440 (vec_pack_trunc_<mode>): Likewise.
8442 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
8444 (vec_vmaxsd): Likewise.
8445 (vec_vmaxud): Likewise.
8446 (vec_vminsd): Likewise.
8447 (vec_vminud): Likewise.
8448 (vec_vpksdss): Likewise.
8449 (vec_vpksdus): Likewise.
8450 (vec_vpkudum): Likewise.
8451 (vec_vpkudus): Likewise.
8452 (vec_vrld): Likewise.
8453 (vec_vsld): Likewise.
8454 (vec_vsrad): Likewise.
8455 (vec_vsrd): Likewise.
8456 (vec_vsubudm): Likewise.
8457 (vec_vupkhsw): Likewise.
8458 (vec_vupklsw): Likewise.
8460 2013-05-29 Jan Hubicka <jh@suse.cz>
8462 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
8463 flags; reorder rest of fields in more consistent way.
8464 (varpool_node): Remove analyzed, finalized and alias.
8465 (cgraph_ndoe): Likewise.
8466 (symtab_alias_ultimate_target): New function.
8467 (cgraph_function_node): Move offline.
8468 (cgraph_reset_node): Declare.
8469 (cgraph_comdat_can_be_unshared_p): Remove.
8470 (varpool_remove_initializer): Declare.
8471 (varpool_first_defined_variable, varpool_next_defined_variable
8472 cgraph_first_defined_function, cgraph_next_defined_function): Update.
8473 (cgraph_function_with_gimple_body_p): Update.
8474 (varpool_all_refs_explicit_p): Update.
8475 (symtab_alias_target): New function.
8476 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
8477 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
8478 (cgraph_function_or_thunk_node): Simplify using
8479 symtab_alias_ultimate_target.
8480 (varpool_variable_node): Likewise.
8481 * cgraph.c (cgraph_create_function_alias): Update.
8482 (cgraph_add_thunk): Update.
8483 (cgraph_remove_node): Update.
8484 (dump_cgraph_node): Do not dump removed flags.
8485 (cgraph_function_body_availability): Update.
8486 (cgraph_propagate_frequency): Update.
8487 (verify_cgraph_node): Check sanity of local flag.
8488 (cgraph_function_node): Move here from cgraph.h; revamp for
8489 cgraph_function_or_thunk_node.
8490 * lto-symtab.c (lto_varpool_replace_node): Update.
8491 (lto_symtab_resolve_can_prevail_p): Update.
8492 (lto_symtab_merge_cgraph_nodes): Update.
8493 * ipa-cp.c (determine_versionability, initialize_node_lattices,
8494 propagate_constants_accross_call, devirtualization_time_bonus,
8495 ipcp_propagate_stage): Update.
8496 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
8497 * ipa-inline-transform.c (clone_inlined_nodes,
8498 preserve_function_body_p): Update.
8499 * ipa-reference.c (propagate): Update.
8500 (write_node_summary_p): Update.
8501 * toplev.c (wrapup_global_declaration_2): Update.
8502 * cgraphunit.c (cgraph_analyze_function): Rename to ...
8503 (analyze_function) ... this one.
8504 (cgraph_process_new_functions): Update.
8505 (cgraph_reset_node): Export.
8506 (cgraph_finalize_function): Update.
8507 (cgraph_add_new_function): Update.
8508 (process_function_and_variable_attributes): Update.
8509 (varpool_finalize_decl): Update.
8510 (symbol_finalized): Remove.
8511 (symbol_finalized_and_needed): Rename to ...
8512 (symbol_defined_and_needed): ... update.
8513 (cgraph_analyze_functions): Update.
8514 (handle_alias_pairs): Update.
8515 (mark_functions_to_output): Update.
8516 (assemble_thunk): Update.
8517 (output_in_order): Update.
8518 (output_weakrefs): Update.
8519 (finalize_compilation_unit): Update.
8520 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
8521 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
8522 input_node, input_varpool_node): Update.
8523 * dbxout.c (dbxout_expand_expr): Update.
8524 * cgraphclones.c (cgraph_clone_node): Update.
8525 (cgraph_copy_node_for_versioning): Update.
8526 (cgraph_materialize_clone): Update.
8527 (cgraph_materialize_all_clones): Update.
8528 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
8529 propagate_pure_const, propagate_nothrow): Update.
8530 * lto-streamer-out.c (lto_output, write_symbol): Update.
8531 * ipa-utils.c (ipa_reverse_postorder): Update.
8532 * ipa-inline.c (can_inline_edge_p): Update.
8533 (update_caller_keys, ipa_inline): Update.
8534 * dwarf2out.c (reference_to_unused,
8535 premark_types_used_by_global_vars_helper): Update.
8536 * tree-eh.c (tree_could_trap_p): Update.
8537 * ipa-split.c (consider_split, execute_split_functions): Update.
8538 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
8539 has_addr_references_p): Update; move ahead in file for better
8541 (process_references): Simplify.
8542 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
8544 (cgraph_comdat_can_be_unshared_p): Make static.
8545 (cgraph_externally_visible_p): Update.
8546 (varpool_externally_visible_p): Update.
8547 (function_and_variable_visibility): Update.
8548 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
8549 ipa_tm_mark_force_output_node): Update.
8550 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
8551 estimate_edge_devirt_benefit, inline_generate_summary,
8552 inline_write_summary): Update.
8553 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
8554 * ipa-prop.c (ipa_compute_jump_functions): Update.
8555 (ipa_print_node_params, ipa_prop_read_section,
8556 ipa_update_after_lto_read, read_replacements_section): Update.
8557 * varasm.c (mark_decl_referenced): Update.
8558 (assemble_alias, dump_tm_clone_pairs): Update.
8559 * tree-inline.c (copy_bb): Update.
8560 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
8562 * symtab.c (dump_symtab_base): Print new flags.
8563 (verify_symtab_base): Verify new flags.
8564 (symtab_alias_ultimate_target): New function.
8565 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
8566 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
8568 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
8570 * i386.c (ix86_get_function_versions_dispatcher,
8571 ix86_generate_version_dispatcher_body): Update.
8572 (fold_builtin_cpu): Use varpool_add_new_variable.
8573 * varpool.c (varpool_remove_initializer): Break out from ...
8574 (varpool_remove_node): ... this one.
8575 (dump_varpool_node, varpool_node_for_asm,
8576 cgraph_variable_initializer_availability, varpool_analyze_node,
8577 varpool_assemble_decl, varpool_remove_unreferenced_decls,
8578 varpool_finalize_named_section_flags, varpool_create_variable_alias):
8581 2013-05-29 Jan Hubicka <jh@suse.cz>
8583 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
8585 2013-05-29 Easwaran Raman <eraman@google.com>
8587 PR tree-optimization/57442
8588 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
8589 when control exits the main loop.
8591 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
8593 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
8595 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
8596 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
8597 * rx/t-rx: Add rx100 under multi library matches option for nofpu
8600 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8602 PR tree-optimization/57441
8603 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
8604 Don't limit size of incr_vec to number of candidates.
8606 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
8608 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
8609 and mips16 directories.
8610 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
8611 (MULTILIB_DIRNAMES): Ditto.
8612 (MULTILIB_EXCEPTIONS): Add new exceptions.
8613 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
8614 (MULTILIB_DIRNAMES): Ditto.
8615 (MULTILIB_EXCEPTIONS): Add new exceptions.
8617 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8618 Marcus Shawcroft <marcus.shawcroft@arm.com>
8620 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
8621 SYMBOL_TINY_ABSOLUTE.
8622 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
8623 SYMBOL_TINY_ABSOLUTE.
8624 (aarch64_expand_mov_immediate): Likewise.
8625 (aarch64_classify_symbol): Likewise.
8626 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
8627 Permit SYMBOL_TINY_ABSOLUTE.
8628 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
8630 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8631 Marcus Shawcroft <marcus.shawcroft@arm.com>
8633 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
8634 Refactor if/switch. Replace gcc_assert with if.
8636 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8638 * config/i386/i386.c (initial_ix86_tune_features): Enable
8639 FP Reassociation for AMD bdver1 and bdver2.
8641 2013-05-29 Martin Jambor <mjambor@suse.cz>
8643 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
8644 and IMAGPART_EXPR do not occur within other handled_components.
8646 2013-05-29 Richard Biener <rguenther@suse.de>
8648 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
8649 access on whether the use is in the BB we currently try to
8651 (vect_bb_vectorization_profitable_p): Pass the BB we currently
8652 vectorize to vect_bb_slp_scalar_cost.
8654 2013-05-29 Richard Biener <rguenther@suse.de>
8656 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
8657 computing scalar cost offsetted by stmts that are kept live
8659 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
8660 for computation of scalar cost.
8662 2013-05-28 Steve Ellcey <sellcey@mips.com>
8664 * config/mips/mips-cpus.def (mips32r2): Change processor type.
8666 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
8668 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
8669 array notation built-in reduction functions.
8670 * doc/passes.texi (Passes): Added documentation about changes done
8672 * doc/invoke.texi (C Dialect Options): Added documentation about
8673 the -fcilkplus flag.
8674 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
8675 (BUILTINS_DEF): Depend on cilkplus.def.
8676 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
8677 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
8678 * cilkplus.def: New file.
8680 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
8682 PR rtl-optimization/57439
8683 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
8685 2013-05-28 Easwaran Raman <eraman@google.com>
8687 PR tree-optimization/57337
8688 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
8689 (find_insert_point): Correctly identify the insertion point
8690 when two statements with the same UID is compared.
8692 2013-05-28 Richard Biener <rguenther@suse.de>
8694 PR tree-optimization/56787
8695 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
8696 from the list of data references.
8697 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
8699 (vect_analyze_loop_operations): Likewise.
8700 (vect_transform_loop): Remove clobbers.
8702 2013-05-28 Martin Jambor <mjambor@suse.cz>
8704 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
8705 and REALPART_EXPRs have scalar type.
8707 2013-05-28 Richard Biener <rguenther@suse.de>
8709 PR tree-optimization/57411
8710 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
8712 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
8713 virtual operand propagation.
8715 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
8717 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
8718 destination register for bmasksi_vis.
8719 (vector_init_bshuffle): Likewise.
8720 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
8722 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
8724 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
8725 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
8726 mode if the instruction isn't available in the original mode.
8727 * config/sparc/sparc.opt (mfix-ut699): New option.
8728 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
8729 (divdf3): Turn into expander.
8730 (divdf3_nofix): New insn.
8731 (divdf3_fix): Likewise.
8732 (divsf3): Disable if -mfix-ut699.
8733 (sqrtdf2): Turn into expander.
8734 (sqrtdf2_nofix): New insn.
8735 (sqrtdf2_fix): Likewise.
8736 (sqrtsf2): Disable if -mfix-ut699.
8738 2013-05-27 Richard Biener <rguenther@suse.de>
8741 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
8742 block for the new loop.
8744 2013-05-27 Richard Biener <rguenther@suse.de>
8746 PR tree-optimization/57343
8747 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
8748 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
8749 (number_of_iterations_cond): Do not build the folded tree.
8751 2013-05-27 Richard Biener <rguenther@suse.de>
8755 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8756 OEP_CONSTANT_ADDRESS_OF retained.
8758 PR tree-optimization/57417
8759 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
8761 (set_ssa_val_to): Compare addresses using
8762 get_addr_base_and_unit_offset.
8764 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
8766 PR rtl-optimization/56833
8767 * postreload.c (move2add_record_mode): New function.
8768 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
8769 (move2add_use_add2_insn): Use move2add_record_sym_value.
8770 (move2add_use_add3_insn): Likewise.
8771 (reload_cse_move2add): Use move2add_valid_value_p and
8772 move2add_record_mode. Invalidate call-clobbered and REG_INC
8773 affected regs by setting reg_mode to VOIDmode.
8774 (move2add_note_store): Don't pretend the inside of a SUBREG is
8775 the actual destination. Invalidate single/leading registers by
8776 setting reg_mode to VOIDmode.
8777 Use move2add_record_sym_value, move2add_valid_value_p and
8778 move2add_record_mode.
8780 2013-05-27 Richard Biener <rguenther@suse.de>
8782 PR tree-optimization/57396
8783 * tree-affine.c (double_int_constant_multiple_p): Properly
8784 return false for val == 0 and div != 0.
8786 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8788 * config/mips/mips.h: Use #elif in preprocessor conditions.
8790 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8793 * config/mips/constraints.md (kl): New constraint.
8794 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
8795 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
8796 constraint for operand 0. Split after CSE for MIPS16. Emit a move
8798 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
8800 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8803 * config/mips/mips.c (mips_can_inline_p): New function.
8804 (TARGET_CAN_INLINE_P): Define.
8806 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
8808 * sched-int.h (ds_t, dw_t): Make unsigned int.
8809 Fix documentation that describes how all the ds_t bits are used.
8810 Reserve the last bit for delayed-branch scheduling.
8811 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
8812 (BITS_PER_DEP_WEAK): Fix definition and documentation.
8813 (gen_dep_weak_1): Remove prototype.
8814 * sched-deps.c (get_dep_weak_1): Make static.
8815 * target.def (speculate_insn, needs_block_p, gen_spec_check,
8816 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
8817 * doc/tm.texi: Regenerate.
8818 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
8820 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
8823 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
8825 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
8826 Sandra Loosemore <sandra@codesourcery.com>
8828 * config.gcc (powerpc-*): Allow native for with-cpu.
8830 2013-05-24 Jeff Law <law@redhat.com>
8832 PR tree-optimization/57124
8833 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
8834 conversion feeding a condition if the range has an overflow
8835 if -fstrict-overflow. Add warnings for when we do make the
8838 2013-05-24 Dehao Chen <dehao@google.com>
8840 * tree-cfg.c (locus_discrim_map): Fix the typo.
8841 (locus_discrim_hasher): Likewise.
8842 (locus_discrim_hasher::hash): Likewise.
8843 (locus_discrim_hasher::equal): Likewise.
8845 2013-05-24 Martin Jambor <mjambor@suse.cz>
8847 PR tree-optimization/57294
8848 * cgraph.h (ipa_record_stmt_references): Declare.
8849 * cgraphbuild.c (ipa_record_stmt_references): New function.
8850 (build_cgraph_edges): Use ipa_record_stmt_references.
8851 (rebuild_cgraph_edges): Likewise.
8852 (cgraph_rebuild_references): Likewise.
8853 * ipa-prop.c (ipa_modify_call_arguments): Discard references
8854 associated with the old statement and build references from the
8855 newly built statements.
8856 * ipa-ref.c (ipa_remove_stmt_references): New function.
8857 * ipa-ref.h (ipa_remove_stmt_references): Declare.
8859 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
8861 * lra-constraints.c (emit_spill_move): Use smaller mode for
8863 (check_and_process_move): Consider mem-reg moves for secondary
8865 (curr_insn_transform): Don't lose insns emitted before for
8866 secondary memory moves.
8867 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
8868 reg set up in the current insn.
8870 2013-05-24 Dehao Chen <dehao@google.com>
8872 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
8874 (locus_descrim_hasher::equal): Likewise.
8875 (build_gimple_cfg): New discriminator assignment algorithm.
8876 (make_edges): Likewise.
8877 (next_discriminator_for_locus): Likewise.
8878 (same_line_p): Likewise.
8879 (assign_discriminators): Likewise.
8880 (make_cond_expr_edges): Likewise.
8881 (make_gimple_switch_edges): Likewise.
8882 (make_goto_expr_edges): Likewise.
8883 (make_gimple_asm_edges): Likewise.
8885 2013-05-24 Ian Bolton <ian.bolton@arm.com>
8887 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
8888 X format specifier to only display bottom 16 bits.
8889 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
8890 immediate to match for operand 2, since it will be masked.
8892 2013-05-24 Richard Biener <rguenther@suse.de>
8894 PR tree-optimization/57287
8895 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
8896 all SSA names that occur in abnormal PHIs.
8898 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8901 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8902 that index is not negative.
8904 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8906 PR rtl-optimization/55177
8907 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8908 (simplify_byte_swapping_operation): New.
8909 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8910 (simplify_relational_operation_1): Deal with BSWAP.
8912 2013-05-23 Richard Henderson <rth@redhat.com>
8915 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8916 (ix86_reorg): Call it.
8918 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8921 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8922 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8923 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8925 2013-05-23 Christian Bruel <christian.bruel@st.com>
8928 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8930 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8931 Marcus Shawcroft <marcus.shawcroft@arm.com>
8933 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8934 * config/aarch64/constraints.md (Usa): Remove.
8935 * doc/md.texi (AArch64 Usa): Remove.
8937 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8938 Marcus Shawcroft <marcus.shawcroft@arm.com>
8940 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8941 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8942 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8943 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8945 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8947 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8948 instruction (AdvSIMD).
8949 * config/aarch64/aarch64-builtins.c
8950 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8951 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8953 2013-05-23 Martin Jambor <mjambor@suse.cz>
8956 * tree.h (contains_bitfld_component_ref_p): Declare.
8957 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8958 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8960 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8961 not access a bit-field. Assert all final offsets are byte-aligned.
8963 2013-05-23 Richard Biener <rguenther@suse.de>
8965 PR tree-optimization/57380
8966 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8967 least one invariant or re-used load.
8968 * passes.c (init_optimization_passes): Move pass_phiprop before
8971 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8973 * config/aarch64/aarch64-simd.md
8974 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
8976 2013-05-23 Richard Biener <rguenther@suse.de>
8979 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8980 OEP_CONSTANT_ADDRESS_OF retained.
8982 2013-05-23 Jakub Jelinek <jakub@redhat.com>
8985 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
8986 don't lower unit. Handle unit not being always BITS_PER_WORD.
8988 2013-05-23 Richard Biener <rguenther@suse.de>
8990 PR rtl-optimization/57341
8991 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
8992 instead of true_dependence.
8994 2013-05-22 David Malcolm <dmalcolm@redhat.com>
8996 * bb-reorder.c (branch_threshold): Make const.
8997 (exec_threshold): Ditto.
8999 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9000 Pat Haugen <pthaugen@us.ibm.com>
9001 Peter Bergner <bergner@vnet.ibm.com>
9003 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
9004 documentation for the power8 crypto builtins.
9006 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
9008 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
9009 macros for defining power8 builtin functions.
9010 (BU_P8V_AV_2): Likewise.
9011 (BU_P8V_AV_P): Likewise.
9012 (BU_P8V_VSX_1): Likewise.
9013 (BU_P8V_OVERLOAD_1): Likewise.
9014 (BU_P8V_OVERLOAD_2): Likewise.
9015 (BU_CRYPTO_1): Likewise.
9016 (BU_CRYPTO_2): Likewise.
9017 (BU_CRYPTO_3): Likewise.
9018 (BU_CRYPTO_OVERLOAD_1): Likewise.
9019 (BU_CRYPTO_OVERLOAD_2): Likewise.
9020 (XSCVSPDP): Fix typo, point to the correct instruction.
9021 (VCIPHER): Add power8 crypto builtins.
9022 (VCIPHERLAST): Likewise.
9023 (VNCIPHER): Likewise.
9024 (VNCIPHERLAST): Likewise.
9025 (VPMSUMB): Likewise.
9026 (VPMSUMH): Likewise.
9027 (VPMSUMW): Likewise.
9028 (VPERMXOR_V2DI): Likewise.
9029 (VPERMXOR_V4SI: Likewise.
9030 (VPERMXOR_V8HI: Likewise.
9031 (VPERMXOR_V16QI: Likewise.
9032 (VSHASIGMAW): Likewise.
9033 (VSHASIGMAD): Likewise.
9035 (VPERMXOR): Likewise.
9036 (VSHASIGMA): Likewise.
9038 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9039 __CRYPTO__ if the crypto instructions are available.
9040 (altivec_overloaded_builtins): Add support for overloaded power8
9043 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
9044 support for power8 crypto builtins.
9045 (builtin_function_type): Likewise.
9046 (altivec_init_builtins): Add support for builtins that take vector
9047 long long (V2DI) arguments.
9049 * config/rs6000/crypto.md: New file, define power8 crypto
9052 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9053 Pat Haugen <pthaugen@us.ibm.com>
9054 Peter Bergner <bergner@vnet.ibm.com>
9056 * doc/invoke.texi (Option Summary): Add power8 options.
9057 (RS/6000 and PowerPC Options): Likewise.
9059 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
9060 constraints.md instead of rs6000.h. Reorder w* constraints. Add
9061 wm, wn, wr documentation.
9063 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
9064 registers if direct move instructions are enabled.
9065 (wn): New constraint for no registers.
9066 (wq): New constraint for quad word even GPR registers.
9067 (wr): New constraint if 64-bit instructions are enabled.
9068 (wv): New constraint if power8 vector instructions are enabled.
9069 (wQ): New constraint for quad word memory locations.
9071 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
9072 constraint for 0..15 for crypto instructions.
9073 (gpc_reg_operand): If VSX allow registers in VSX registers as well
9074 as GPR and floating point registers.
9075 (int_reg_operand): New predicate to match only GPR registers.
9076 (base_reg_operand): New predicate to match base registers.
9077 (quad_int_reg_operand): New predicate to match even GPR registers
9078 for quad memory operations.
9079 (vsx_reg_or_cint_operand): New predicate to allow vector logical
9080 operations in both GPR and VSX registers.
9081 (quad_memory_operand): New predicate for quad memory operations.
9082 (reg_or_indexed_operand): New predicate for direct move support.
9084 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
9085 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
9086 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
9087 (POWERPC_MASKS): Add power8 options.
9088 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
9091 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9092 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
9094 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
9095 (-mpower8-fusion): New power8 options.
9096 (-mpower8-fusion-sign): Likewise.
9097 (-mpower8-vector): Likewise.
9098 (-mcrypto): Likewise.
9099 (-mdirect-move): Likewise.
9100 (-mquad-memory): Likewise.
9102 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
9104 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
9106 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
9107 (rs6000_debug_vector_unit): Add p8_vector.
9108 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
9109 definitions. Also print fusion state.
9110 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
9111 (rs6000_builtin_mask_calculate): Add power8 builtin support.
9112 (rs6000_option_override_internal): Add support for power8.
9113 (rs6000_common_init_builtins): Add debugging for skipped builtins
9115 (rs6000_adjust_cost): Add power8 support.
9116 (rs6000_issue_rate): Likewise.
9117 (insn_must_be_first_in_group): Likewise.
9118 (insn_must_be_last_in_group): Likewise.
9119 (force_new_group): Likewise.
9120 (rs6000_register_move_cost): Likewise.
9121 (rs6000_opt_masks): Likewise.
9123 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
9124 power8 capable assembler, default to power7 options.
9125 (TARGET_DIRECT_MOVE): Likewise.
9126 (TARGET_CRYPTO): Likewise.
9127 (TARGET_P8_VECTOR): Likewise.
9128 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
9129 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
9130 (VECTOR_MEM_P8_VECTOR_P): Likewise.
9131 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
9132 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
9133 (TARGET_XSCVDPSPN): Likewise.
9134 (TARGET_XSCVSPDPN): Likewsie.
9135 (TARGET_SYNC_HI_QI): Likewise.
9136 (TARGET_SYNC_TI): Likewise.
9137 (MASK_CRYPTO): Likewise.
9138 (MASK_DIRECT_MOVE): Likewise.
9139 (MASK_P8_FUSION): Likewise.
9140 (MASK_P8_VECTOR): Likewise.
9141 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
9142 temporary used by some of the direct move instructions to get two FP
9143 temporary registers does not force creation of a stack frame.
9144 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
9145 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
9146 that any VSX registers are tieable, even if they are also an
9147 Altivec vector mode.
9148 (r6000_reg_class_enum): Add wm, wr, wv constraints.
9149 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
9150 (RS6000_BTM_CRYPTO): Likewise.
9151 (RS6000_BTM_COMMON): Likewise.
9153 * config/rs6000/rs6000.md (cpu attribute): Add power8.
9154 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
9155 (enum rs6000_vector): Add power8 vector support.
9157 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9161 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
9162 (any_sibcall_could_use_r3): this and handle indirect calls.
9163 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
9165 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9167 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
9169 2013-05-22 Richard Biener <rguenther@suse.de>
9172 * profile.c (branch_prob): Do not split blocks that are
9173 abnormally receiving from ECF_RETURNS_TWICE functions.
9175 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9177 * recog.c (offsettable_address_addr_space_p): Fix calculation of
9178 address mode. Move pointer mode initialization to the same place.
9180 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9182 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
9183 while it has any effect.
9185 2013-05-21 Easwaran Raman <eraman@google.com>
9187 PR tree-optimization/57322
9188 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
9189 UID of the statement added to the BB to be 1.
9191 2013-05-21 Jakub Jelinek <jakub@redhat.com>
9193 PR tree-optimization/57331
9194 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
9195 of conversion from pointer type to integral type with integer.
9197 2013-05-21 Martin Jambor <mjambor@suse.cz>
9200 * ipa-prop.c (ipa_read_node_info): Process param_used and
9201 controlled_uses in the same order as when writing.
9203 2013-05-21 Magnus Granberg <baldrick@free.fr>
9206 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
9208 2013-05-21 Richard Biener <rguenther@suse.de>
9210 PR tree-optimization/57318
9211 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
9212 estimate stmts with side-effects as likely eliminated.
9214 2013-05-21 Richard Biener <rguenther@suse.de>
9216 PR tree-optimization/57330
9217 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
9218 preserve the call stmts fntype.
9220 2013-05-21 Richard Biener <rguenther@suse.de>
9222 PR tree-optimization/57303
9223 * tree-ssa-sink.c (statement_sink_location): Improve killing
9224 stmt detection and properly handle self-assignments.
9226 2013-05-21 Christian Bruel <christian.bruel@st.com>
9228 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
9229 spanning registers. LEAF_REG_REMAP is supported only for contiguous
9230 registers. Set register size out of the PARALLEL loop.
9232 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
9235 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
9236 (*fmasf4, *fmasf4_media): New insns.
9238 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9240 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
9241 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
9242 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
9243 (mips_idiv_insns): Update the comments to say that the returned
9244 instruction counts are in units of BASE_INSN_LENGTH.
9245 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
9246 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
9247 using 2 rather than 4 as the length of indirect MIPS16 and
9248 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
9249 length of a NOP. Don't divide MIPS16 lengths by 2.
9250 (mips16_split_long_branches): Assume a branch is long if the
9251 length is greater than 4 rather than 8.
9252 * config/mips/mips.md (length): Give MIPS16 lengths directly,
9253 rather than multiplying them by 2. Multiply instruction counts
9254 by BASE_INSN_LENGTH rather than 4.
9255 (*jump_mips16, tls_get_tp_mips16_<mode>)
9256 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
9258 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9260 * config/mips/mips.md (extended_mips16): Remove branch case.
9261 (length): Remove duplicated extended_mips16 test.
9263 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9265 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
9267 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
9269 * recog.h (Recog_data): Rename to...
9270 (recog_data_d): ...this.
9271 (recog_data): Update accordingly.
9272 * recog.c (recog_data): Likewise.
9273 * reload.c (save_recog_data): Likewise.
9274 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
9275 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
9277 2013-05-17 Julian Brown <julian@codesourcery.com>
9279 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
9280 found in a REG_EQUAL note, invalidate it.
9282 2013-05-17 Easwaran Raman <eraman@google.com>
9284 * tree-ssa-reassoc.c (find_insert_point): New function.
9285 (insert_stmt_after): Likewise.
9286 (get_def_stmt): Likewise.
9287 (ensure_ops_are_available): Likewise.
9288 (not_dominated_by): Likewise.
9289 (rewrite_expr_tree): Do not move statements beyond what is
9290 necessary. Remove call to swap_ops_for_binary_stmt...
9291 (reassociate_bb): ... and move it here.
9292 (build_and_add_sum): Assign UIDs for new statements.
9293 (linearize_expr): Likewise.
9294 (do_reassoc): Renumber gimple statement UIDs.
9296 2013-05-17 Jan Hubicka <jh@suse.cz>
9298 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
9300 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
9301 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
9303 (output_weakrefs): Update.
9305 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
9306 Martin Jambor <mjambor@suse.cz>
9309 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
9310 value that corresponds to the given aggval is found in values vector.
9312 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
9314 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
9315 sse, sse2, sse3, ssse3 and sse4a flags to options.
9317 2013-05-17 David Malcolm <dmalcolm@redhat.com>
9319 * gengtype-state.c: (s_expr_writer): New class, to handle
9320 prettifying of output layout of s-expressions.
9321 (state_writer): New class, to write out gtype.state.
9322 (state_written_type_count): Move this variable into member data of
9324 (s_expr_writer::s_expr_writer): New code: constructor for new class
9325 (state_writer::state_writer(): ditto
9326 (s_expr_writer::write_new_line): New function
9327 (s_expr_writer::write_any_indent): ditto
9328 (s_expr_writer::begin_s_expr): ditto
9329 (s_expr_writer::end_s_expr): ditto
9330 (write_state_fileloc): convert to method of state_writer...
9331 (state_writer:: write_state_fileloc): ...and use methods of
9332 s_expr_writer to write indentation into the gtype.state output file
9333 to visually represent the hierarchical structure of the list
9335 (write_state_fields): ditto, renaming to...
9336 (state_writer::write_state_fields)
9337 (write_state_a_string): ditto, renaming to...
9338 (state_writer::write_state_a_string)
9339 (write_state_string_option): ditto, renaming to...
9340 (state_writer::write_state_string_option)
9341 (write_state_type_option): ditto, renaming to...
9342 (state_writer::write_state_type_option)
9343 (write_state_nested_option): ditto, renaming to...
9344 (state_writer::write_state_nested_option)
9345 (write_state_option): ditto, renaming to...
9346 (state_writer::write_state_option)
9347 (write_state_options): ditto, renaming to...
9348 (state_writer::write_state_options)
9349 (write_state_lang_bitmap): ditto, renaming to...
9350 (state_writer::write_state_lang_bitmap)
9351 (write_state_version): ditto, renaming to...
9352 (state_writer::write_state_version)
9353 (write_state_scalar_type): ditto, renaming to...
9354 (state_writer::write_state_scalar_type)
9355 (write_state_string_type): ditto, renaming to...
9356 (state_writer::write_state_string_type)
9357 (write_state_undefined_type): ditto, renaming to...
9358 (state_writer::write_state_undefined_type)
9359 (write_state_struct_union_type): ditto, renaming to...
9360 (state_writer::write_state_struct_union_type)
9361 (write_state_struct_type): ditto, renaming to...
9362 (state_writer::write_state_struct_type)
9363 (write_state_user_struct_type): ditto, renaming to...
9364 (state_writer::write_state_user_struct_type)
9365 (write_state_lang_struct_type): ditto, renaming to...
9366 (state_writer::write_state_lang_struct_type)
9367 (write_state_param_struct_type): ditto, renaming to...
9368 (state_writer::write_state_param_struct_type)
9369 (write_state_pointer_type): ditto, renaming to...
9370 (state_writer::write_state_pointer_type)
9371 (write_state_array_type): ditto, renaming to...
9372 (state_writer::write_state_array_type)
9373 (write_state_gc_used): ditto, renaming to...
9374 (state_writer::write_state_gc_used)
9375 (write_state_common_type_content): ditto, renaming to...
9376 (state_writer::write_state_common_type_content)
9377 (write_state_type): ditto, renaming to...
9378 (state_writer::write_state_type)
9379 (write_state_pair_list): ditto, renaming to...
9380 (state_writer::write_state_pair_list)
9381 (write_state_pair): ditto, renaming to...
9382 (state_writer::write_state_pair)
9383 (write_state_typedefs): ditto, renaming to...
9384 (state_writer::write_state_typedefs)
9385 (write_state_structures): ditto, renaming to...
9386 (state_writer::write_state_structures)
9387 (write_state_param_structs): ditto, renaming to...
9388 (state_writer::write_state_param_structs)
9389 (write_state_variables): ditto, renaming to...
9390 (state_writer::write_state_variables)
9391 (write_state_srcdir): ditto, renaming to...
9392 (state_writer::write_state_srcdir)
9393 (write_state_files_list): ditto, renaming to...
9394 (state_writer::write_state_files_list)
9395 (write_state_languages): ditto, renaming to...
9396 (state_writer::write_state_languages)
9397 (write_state): create a state_writer instance and use it when
9398 writing out the state file
9400 2013-05-17 Mike Stump <mikestump@comcast.net>
9402 PR rtl-optimization/57304
9403 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
9404 accessing DF_REF_REAL_LOC.
9406 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9408 PR rtl-optimization/57281
9409 PR rtl-optimization/57300
9410 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
9411 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
9412 what the other splitter did if the registers are dead.
9414 2013-05-17 Richard Biener <rguenther@suse.de>
9416 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
9419 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9421 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
9424 2013-05-17 Marek Polacek <polacek@redhat.com>
9426 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
9427 length when doing non-zero store of storing '\0' to '\0'.
9429 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9431 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
9432 vect_external_def oprnd1 with loop_vinfo, try to emit
9433 optional cast, negation and and stmts on the loop preheader
9434 edge instead of into the pattern def seq.
9436 PR tree-optimization/57051
9437 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
9438 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
9440 2013-05-16 Nick Clifton <nickc@redhat.com>
9442 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
9443 (rl78_is_naked_func): New function.
9444 (rl78_expand_prologue): Skip prologue generation for naked functions.
9445 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
9446 * doc/extend.texi (naked): Add RL78 to the list of processors
9447 that supports this attribute.
9449 2013-05-16 Jeff Law <law@redhat.com>
9451 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
9453 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
9455 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
9456 cache parameters using detect_caches_amd also for CYRIX,
9457 NSC and TM2 signatures.
9459 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
9460 Dzianis Kahanovich <mahatma@eu.by>
9464 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9465 VIA/Centaur processors and determine their cache parameters
9466 using detect_caches_amd.
9468 2013-05-16 Teresa Johnson <tejohnson@google.com>
9470 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
9471 (rtl_verify_edges): New function.
9472 (rtl_verify_bb_insns): Ditto.
9473 (rtl_verify_bb_pointers): Ditto.
9474 (rtl_verify_bb_insn_chain): Ditto.
9475 (rtl_verify_fallthru): Ditto.
9476 (rtl_verify_bb_layout): Ditto.
9477 (rtl_verify_flow_info_1): Outline checks into new functions.
9478 (rtl_verify_flow_info): Ditto.
9480 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
9482 * cfghooks.c (copy_bbs): Add update_dominance argument.
9483 * cfghooks.h (copy_bbs): Update prototype.
9484 * tree-cfg.c (gimple_duplicate_sese_region):
9485 Add update_dominance argument.
9486 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
9487 * tree-ssa-loop-ch.c (copy_loop_headers): Update
9488 gimple_duplicate_sese_region call.
9489 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
9490 Update copy_bbs call.
9491 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
9492 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
9494 2013-05-16 Jakub Jelinek <jakub@redhat.com>
9496 * tree-vectorizer.h (NUM_PATTERNS): Increment.
9497 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
9498 vect_recog_rotate_pattern.
9499 (vect_recog_rotate_pattern): New function.
9501 2013-05-16 Jason Merrill <jason@redhat.com>
9503 * Makefile.in (LLINKER): New variable.
9504 (mostlyclean): Remove link mutex.
9505 * configure.ac: Handle --enable-link-mutex.
9506 * lock-and-run.sh: New script.
9508 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9511 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
9514 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9516 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
9518 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
9520 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
9521 * config/arm/arm.c (next_consecutive_mem): New function.
9522 (gen_movmem_ldrd_strd): Likewise.
9523 * config/arm/arm.md (movmemqi): Update condition and code.
9524 (unaligned_loaddi, unaligned_storedi): New patterns.
9526 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9528 * config.gcc: Obsolete *-*-solaris2.9*.
9529 * doc/install.texi (Specific, *-*-solaris2*): Document it.
9531 2013-05-16 Richard Biener <rguenther@suse.de>
9533 * passes.c (init_optimization_passes): Move pass_parallelize_loops
9534 earlier, after GRAPHITE transforms and IV canonicalization.
9536 2013-05-16 Jakub Jelinek <jakub@redhat.com>
9538 * omp-low.c (extract_omp_for_data): For collapsed loops,
9539 if at least one of the loops is known at compile time to
9540 iterate zero times, set count to 0.
9541 (expand_omp_regimplify_p): New function.
9542 (expand_omp_for_generic): For collapsed loops, if at least
9543 one of the loops isn't known to iterate at least once,
9544 add runtime check with setting count to 0.
9545 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
9546 For unsigned types if it isn't known at compile time that
9547 the loop will iterate at least once, add runtime check to bypass
9548 the whole loop if initial condition isn't true.
9550 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
9552 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
9554 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
9557 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
9558 transformations to avoid an infinite loop.
9560 2013-05-16 Marek Polacek <polacek@redhat.com>
9562 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
9564 2013-05-15 Leif Ekblad <leif@rdos.net>
9566 * config/i386/i386.c (ix86_decompose_address): Use
9567 DEFAULT_TLS_SEG_REG to access TLS segment register.
9568 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
9569 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
9570 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
9572 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
9575 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
9576 sibling calls to functions that would normally be lazily bound,
9577 unless $gp is call-clobbered.
9579 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
9581 * config/i386/i386.c (ix86_option_override_internal): Update
9582 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
9583 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
9584 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
9585 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
9587 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
9589 2013-05-15 Andreas Schwab <schwab@suse.de>
9591 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
9592 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
9595 2013-05-15 Teresa Johnson <tejohnson@google.com>
9597 * loop-unroll.c (report_unroll_peel): Check decision before
9598 emitting unroll/peel message.
9600 2013-05-15 Teresa Johnson <tejohnson@google.com>
9602 * function.h (has_bb_partition): New rtl_data flag.
9603 (bb_reorder_complete): Ditto.
9604 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
9605 instead of flag_reorder_blocks_and_partition.
9606 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
9607 with some enhancements.
9608 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
9609 * bb-reorder.c (connect_traces): Check for has_bb_partition
9610 instead of flag_reorder_blocks_and_partition.
9611 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
9612 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
9613 verify_hot_cold_block_grouping.
9614 (partition_hot_cold_basic_blocks): Set has_bb_partition.
9616 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9619 * config/arm/predicates.md (call_insn_operand): New predicate.
9620 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
9621 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
9622 if insn is not a tail call.
9623 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
9625 * config/arm/arm.h (enum reg_class): New caller save register class.
9626 (REG_CLASS_NAMES): Likewise.
9627 (REG_CLASS_CONTENTS): Likewise.
9628 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
9631 2013-05-15 Richard Biener <rguenther@suse.de>
9633 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
9634 of MSG_OPTIMIZED_LOCATIONS.
9635 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
9636 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
9638 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
9639 of MSG_OPTIMIZED_LOCATIONS.
9640 (execute_vect_slp): Likewise.
9641 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
9642 (vect_create_cond_for_alias_checks): Likewise.
9643 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
9644 (vect_recog_widen_mult_pattern): Likewise.
9645 (vect_recog_widen_sum_pattern): Likewise.
9646 (vect_recog_over_widening_pattern): Likewise.
9647 (vect_recog_widen_shift_pattern): Likewise.
9648 (vect_recog_vector_vector_shift_pattern): Likewise.
9649 (vect_recog_divmod_pattern): Likewise.
9650 (vect_recog_mixed_size_cond_pattern): Likewise.
9651 (vect_recog_bool_pattern): Likewise.
9652 (vect_pattern_recog_1): Likewise.
9654 2013-05-15 Martin Jambor <mjambor@suse.cz>
9656 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
9657 non-functions to builtin_unreachable.
9658 * ipa-inline-transform.c (inline_call): Do not assert estimates were
9659 correct when new direct edges were discovered.
9661 2013-05-15 Martin Jambor <mjambor@suse.cz>
9663 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
9664 header, print symbol order instead of node uid, print more information
9665 about indirect edge targets.
9666 (ipa_make_edge_direct_to_target): Print symbol order instead of node
9668 (ipa_make_edge_direct_to_target): Likewise.
9669 (remove_described_reference): Likewise.
9670 (propagate_controlled_uses): Likewise.
9671 (ipa_print_node_params): Also print symbol order.
9672 (ipcp_transform_function): Print symbol order instead of node uids.
9673 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
9674 (cgraph_get_create_real_symbol_node): Likewise.
9675 * ipa-cp.c (print_lattice): Likewise.
9676 (print_all_lattices): Likewise.
9677 (determine_versionability): Likewise.
9678 (initialize_node_lattices): Likewise.
9679 (estimate_local_effects): Likewise.
9680 (update_profiling_info): Likewise.
9681 (create_specialized_node): Likewise.
9682 (perhaps_add_new_callers): Likewise.
9683 (decide_about_value): Likewise.
9684 (decide_whether_version_node): Likewise.
9685 (identify_dead_nodes): Likewise.
9686 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
9687 (dump_inline_summary): Likewise.
9688 (estimate_node_size_and_time): Likewise.
9689 (inline_analyze_function): Likewise.
9690 * ipa-inline.c (report_inline_failed_reason): Likewise.
9691 (want_early_inline_function_p): Likewise.
9692 (edge_badness): Likewise.
9693 (update_edge_key): Likewise.
9694 (inline_small_functions): Likewise. Add dumping of order to two other
9696 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
9697 instead of node uids.
9698 (propagate_pure_const): Likewise.
9699 (propagate_pure_const): Likewise.
9700 * ipa-utils.c (dump_cgraph_node_set): Likewise.
9701 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
9702 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
9704 * tree-pretty-print.c (dump_function_header): Likewise.
9705 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
9706 Print symbol order instead of node uids.
9708 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9710 * config/s390/s390.c (s390_register_move_cost): Don't impose the
9711 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
9713 2013-05-15 Richard Biener <rguenther@suse.de>
9715 PR tree-optimization/57275
9716 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
9717 return value for fail to do runtime alias checks for gather loads.
9719 2013-05-15 Jan Hubicka <jh@suse.cz>
9723 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
9724 weakrefs are not external.
9725 (lto_symtab_merge_decls): Fix thinko when dealing with
9726 non-lto_symtab decls.
9727 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
9728 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
9729 * varpool.c (dump_varpool_node): Dump more flags.
9731 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9733 * config/i386/i386.c (processor_alias_table): Add instruction
9734 FSGSBASE for AMD bdver3 architecture.
9736 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9738 * tree.c (warn_deprecated_use): Print file:line using locus color.
9739 * diagnostic.c (diagnostic_report_current_module): Print file:line
9740 and file:line:column using locus color.
9742 2013-05-14 Mike Stump <mikestump@comcast.net>
9744 * gdbinit.in: Add __null.
9746 2013-05-14 Mike Stump <mikestump@comcast.net>
9748 * recog.h: Rename struct recog_data to Recog_data.
9749 * recog.c: Likewise.
9750 * reload.c (can_reload_into): Likewise.
9751 * config/picochip/picochip.c: Likewise.
9753 2013-05-14 Mike Stump <mikestump@comcast.net>
9755 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
9757 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
9759 * resource.h (struct resources): Remove unch_memory member.
9760 (CLEAR_RESOURCE): Don't clear unch_memory.
9761 * resource.c (mark_referenced_resources): Don't set it.
9762 (mark_set_resources): Likewise.
9763 (mark_target_live_regs): Don't clear it.
9764 (init_resource_info): Likewise.
9765 * reorg.c (resource_conflicts_p): Don't compare it.
9766 (redundant_insn): Don't set it.
9768 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
9770 * emit-rtl.c (next_label): Remove unused function.
9771 (skip_consecutive_labels, link_cc0_insns): Move to ...
9772 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
9773 only place where these functions are used, and make them static.
9775 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9777 * fold-const.c (fold_negate_expr): Handle vectors.
9778 (fold_truth_not_expr): Make it static.
9779 (fold_invert_truthvalue): New static function.
9780 (invert_truthvalue_loc): Handle vectors. Do not call
9781 fold_truth_not_expr directly.
9782 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
9783 <TRUTH_NOT_EXPR>: Do not cast to boolean.
9784 (fold_comparison): Handle vector constants.
9785 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
9786 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
9787 * tree.h (fold_truth_not_expr): Remove declaration.
9789 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
9791 * config/aarch64/aarch64-simd.md
9792 (aarch64_vcond_internal<mode>): Rename to...
9793 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
9794 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
9795 float modes. Clarify all iterator modes.
9796 (vcond<mode><mode>): Use new name for vcond expanders.
9797 (vcond<v_cmp_result><mode>): Likewise.
9798 (vcondu<mode><mode>: Likewise.
9799 * config/aarch64/iterators.md (VDQF_COND): New.
9801 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9804 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
9805 variable for the shift amount. Check that we shift by non-negative
9808 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
9811 * config/arm/arm.h (EPILOGUE_USES): Only return true
9812 for LR_REGNUM after epilogue_completed.
9814 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
9816 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
9819 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9822 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
9823 and Solaris 11+/x86 with gld.
9824 * configure: Regenerate.
9826 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9828 * expmed.c (expand_shift_1): Canonicalize rotates by
9829 constant bitsize / 2 to bitsize - 1.
9830 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
9831 case ROTATERT>: Likewise.
9834 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9836 * config/i386/i386.md (rotateinv): New code attr.
9837 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9838 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9839 roll $31, %eax, etc.
9841 2013-05-14 Richard Biener <rguenther@suse.de>
9844 * tree-eh.c (sink_clobbers): Give up for successors with
9845 multiple predecessors and no virtual uses.
9847 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
9849 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
9850 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
9852 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9855 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
9856 the case when both op0 and op1 have VOIDmode.
9858 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9860 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
9861 in multiply-accumulate mode.
9863 2013-05-13 Guozhi Wei <carrot@google.com>
9865 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
9867 2013-05-13 Kai Tietz <ktietz@redhat.com>
9870 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
9871 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
9872 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
9873 * config/i386/i386.c (ix86_option_override_internal): Likewise.
9874 (ix86_expand_prologue): Likewise.
9875 (ix86_expand_split_stack_prologue): Likewise.
9876 (legitimate_pic_address_disp_p): Likewise.
9877 (legitimize_pic_address): Likewise.
9878 (legitimize_tls_address): Likewise.
9879 (legitimize_pe_coff_symbol): Likewise.
9880 (output_pic_addr_const): Likewise.
9881 (construct_plt_address): Likewise.
9882 (ix86_expand_call): Likewise.
9883 (x86_output_mi_thunk): Likewise.
9884 (x86_function_profiler): Likewise.
9886 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
9888 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
9889 similar switch cases.
9890 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
9891 (aarch64_simd_mov_to_<mode>low): Delete.
9892 (aarch64_simd_mov_to_<mode>high): Delete.
9893 (move_lo_quad_<mode>): Add w<-r alternative.
9894 (aarch64_simd_move_hi_quad_<mode>): Likewise.
9895 (aarch64_simd_mov_from_*): Update type attribute.
9896 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
9899 2013-05-13 Jan Hubicka <jh@suse.cz>
9901 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9902 * config/i386/i386.c (ix86_compute_frame_layout,
9903 ix86_expand_epilogue, emit_i387_cw_initialization,
9904 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9905 ix86_local_alignment): Fix use of size/speed predicates.
9907 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9909 PR tree-optimization/45216
9910 PR tree-optimization/57157
9911 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9912 the (-Y) & (B - 1) variant if OP is |.
9913 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9914 return shifted. Use (-op1) & (prec - 1) as other_amount
9915 instead of prec - op1.
9917 2013-05-13 Martin Jambor <mjambor@suse.cz>
9920 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9921 (ipa_constant_data): New type.
9922 (ipa_jump_func): Use ipa_constant_data to hold information about
9923 constant jump functions.
9924 (ipa_get_jf_constant): Adjust to jump function type changes.
9925 (ipa_get_jf_constant_rdesc): New function.
9926 (ipa_param_descriptor): New field controlled_uses.
9927 (ipa_get_controlled_uses): New function.
9928 (ipa_set_controlled_uses): Likewise.
9929 * ipa-ref.h (ipa_find_reference): Declare.
9930 * ipa-prop.c (ipa_cst_ref_desc): New type.
9931 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9933 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9934 New parameter cs. Adjust all callers.
9935 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9936 (remove_described_reference): New function.
9937 (jfunc_rdesc_usable): Likewise.
9938 (try_make_edge_direct_simple_call): Decrement controlled use count,
9939 attempt to remove reference if it hits zero.
9940 (combine_controlled_uses_counters): New function.
9941 (propagate_controlled_uses): Likewise.
9942 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9943 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9944 (ipa_print_node_params): Print described use counter.
9945 (ipa_write_jump_function): Adjust to jump function type changes.
9946 (ipa_read_jump_function): New parameter CS, pass it to
9947 ipa_set_jf_constant. Adjust caller.
9948 (ipa_write_node_info): Stream controlled use count
9949 (ipa_read_node_info): Likewise.
9950 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9952 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9953 count. Remove cloning-added reference if it reaches zero.
9954 * ipa-ref.c (ipa_find_reference): New function.
9956 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9958 * config/i386/i386.c (processor_target_table): Modified default
9959 alignment values for AMD BD and BT architectures.
9961 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9963 * tree-vect-generic.c (uniform_vector_p): Move ...
9964 * tree.c (uniform_vector_p): ... here.
9965 * tree.h (uniform_vector_p): Declare it.
9966 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9969 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9971 PR tree-optimization/57230
9972 * tree-ssa-strlen.c (handle_char_store): Record length for
9973 array store from STRING_CST.
9975 PR tree-optimization/57230
9976 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
9979 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
9981 * config/epiphany/epiphany.c (epiphany_init): Check size of
9982 NUM_MODES_FOR_MODE_SWITCHING.
9983 (epiphany_expand_prologue):
9984 Remove CONFIG_REGNUM initial value handling code.
9985 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
9986 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
9987 (emit_set_fp_mode, epiphany_mode_after): Likewise.
9988 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
9989 Don't return 1 for FP_MODE_NONE.
9990 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
9991 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
9992 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
9993 * config/epiphany/epiphany.md (save_config): New pattern.
9995 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
9997 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
9999 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
10001 * config/i386/i386.md (memory): Handle sseishft1.
10002 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
10003 (*vec_extractv2di_1): Ditto.
10005 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
10007 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
10010 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
10012 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
10013 Add mthumb/march=armv7-a multilib.
10014 Add mthumb/march=armv7-r multilib.
10015 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
10017 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
10019 * config/v850/t-rtems: Add more multilibs.
10021 2013-05-10 Richard Biener <rguenther@suse.de>
10023 PR tree-optimization/57214
10024 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
10025 not propagate from SSA names that occur in abnormal PHI nodes.
10027 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
10029 * stor-layout.c (element_precision): New function.
10030 * machmode.h (element_precision): Declare it.
10031 * tree.c (build_minus_one_cst): New function.
10032 (element_precision): Likewise.
10033 * tree.h (build_minus_one_cst): Declare new function.
10034 (element_precision): Likewise.
10035 * fold-const.c (operand_equal_p): Use element_precision.
10036 (fold_binary_loc): Handle vector types.
10037 * convert.c (convert_to_integer): Use element_precision.
10038 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
10041 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
10043 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
10044 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
10045 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
10046 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10047 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
10048 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
10049 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
10050 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10051 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
10052 (Uuw8): New constraints.
10053 (Usb4): Move into alphabetical order.
10054 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
10055 (sd8_operand, ub8_operand, uw8_operand): New predicates.
10056 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
10057 previously unnamed patterns.
10058 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
10059 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
10060 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
10061 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
10062 of set_attr_alternative/if_then_else. Use extended_mips16 instead
10063 of specific lengths.
10065 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10067 * config/i386/i386.md (rotateinv): New code attr.
10068 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10069 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10070 roll $31, %eax, etc.
10072 PR tree-optimization/45216
10073 PR tree-optimization/57157
10074 * tree-ssa-forwprop.c (simplify_rotate): New function.
10075 (ssa_forward_propagate_and_combine): Call it.
10077 2013-05-10 Richard Biener <rguenther@suse.de>
10079 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
10080 disable peeling when we version for aliasing.
10081 (vector_alignment_reachable_p): Honor explicit user alignment.
10082 (vect_supportable_dr_alignment): Likewise.
10083 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
10084 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
10085 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
10086 then peeling to arrange for the cost-model check to come first.
10088 2013-05-10 Alan Modra <amodra@gmail.com>
10090 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
10091 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
10092 * configure: Regenerate.
10094 2013-05-10 Alan Modra <amodra@gmail.com>
10097 * varasm.c (default_elf_select_section): Move !DECL_P check..
10098 (get_named_section): ..to here before calling get_section_name.
10100 (default_section_type_flags): Add DECL_P check.
10101 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
10102 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
10104 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
10106 * config/epiphany/epiphany.c (epiphany_expand_prologue):
10107 When using gen_stack_adjust_str with a register offset, add a
10108 REG_FRAME_RELATED_EXPR note.
10110 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
10112 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
10113 (*vec_extractv4si_zext_mem): Ditto.
10114 (*vec_extractv2di): Add 0->x and x->x alternatives.
10115 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
10116 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
10118 2013-05-09 Jason Merrill <jason@redhat.com>
10120 N3639 C++1y VLA support
10121 * gimplify.c (gimplify_vla_decl): Don't touch an existing
10124 * tree.c (build_constructor_va): New.
10125 * tree.h: Declare it.
10127 2013-05-09 Martin Jambor <mjambor@suse.cz>
10130 * gimple-fold.c (canonicalize_constructor_val): Call
10131 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
10133 2013-05-09 Jan Hubicka <jh@suse.cz>
10134 Richard Biener <rguenther@suse.de>
10137 * symtab.c (symtab_make_decl_local): Do not add private names.
10139 2013-05-09 Jan Hubicka <jh@suse.cz>
10142 * symtab.c (insert_to_assembler_name_hash): Handle clones.
10143 (unlink_from_assembler_name_hash): Likewise.
10144 (symtab_prevail_in_asm_name_hash, symtab_register_node,
10145 symtab_unregister_node, symtab_initialize_asm_name_hash,
10146 change_decl_assembler_name): Update.
10148 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
10150 * config/aarch64/aarch64.md: New movtf split.
10151 (*movtf_aarch64): Update.
10152 (aarch64_movdi_tilow): Handle TF modes and rename to
10153 aarch64_movdi_<mode>low.
10154 (aarch64_movdi_tihigh): Handle TF modes and rename to
10155 aarch64_movdi_<mode>high
10156 (aarch64_movtihigh_di): Handle TF modes and rename to
10157 aarch64_mov<mode>high_di
10158 (aarch64_movtilow_di): Handle TF modes and rename to
10159 aarch64_mov<mode>low_di
10160 (aarch64_movtilow_tilow): Remove spurious whitespace.
10161 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
10163 (aarch64_print_operand): Update.
10165 2013-05-09 Alan Modra <amodra@gmail.com>
10167 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
10169 * configure: Regenerate.
10171 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10173 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
10174 splitter preparation statements.
10175 * config/i386/sse.md (*vec_extract* splitters): Ditto.
10176 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
10179 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10181 * gimple-ssa-strength-reduction.c (count_candidates): Change
10182 return value to int.
10183 (analyze_candidates_and_replace): Change type of length to int.
10185 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10187 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
10188 (*vec_extract<mode>): Use VI12_128 mode iterator.
10189 (*vec_extract<mode>_mem): Ditto.
10190 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
10193 2013-05-08 Diego Novillo <dnovillo@google.com>
10198 2012-08-17 Diego Novillo <dnovillo@google.com>
10201 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
10202 * config.in: Regenerate.
10203 * configure: Regenerate.
10204 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
10206 2013-05-08 Jan Hubicka <jh@suse.cz>
10209 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
10210 * cgraph.h (symtab_node_base): Add unique_name.
10211 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
10212 input_overwrite_node, input_varpool_node): Stream unique_name.
10213 * cgraphclones.c (cgraph_create_virtual_clone,
10214 cgraph_function_versioning): Set unique_name.
10215 * ipa.c (function_and_variable_visibility): Set unique_name.
10217 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10219 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
10220 (alloc_cand_and_find_basis): Restrict conditional candidate
10221 processing to CAND_MULTs.
10223 2013-05-08 Jan Hubicka <jh@suse.cz>
10226 lto-symtab.c (lto_symtab_symbol_p): New function.
10227 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
10228 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
10229 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
10230 Skip static symbols.
10232 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
10234 PR tree-optimization/57200
10235 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
10236 Only call inform if the preceding warning_at returns true.
10238 2013-05-07 Han Shen <shenhan@google.com>
10240 * cfgexpand.c (record_or_union_type_has_array_p): New function.
10241 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
10242 * common.opt (fstack-protector-strong): New option.
10243 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
10244 * doc/invoke.texi (Optimization Options): Document
10245 "-fstack-protector-strong".
10246 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
10248 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10250 * config/mips/mips.c (mips_machine_reorg2): Return 0.
10252 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
10254 * ira.c (update_equiv_regs): Add insn having equiv memory even if
10255 it is not lhs of the insn.
10256 (setup_reg_equiv): Remove insn having equiv memory which it is not
10258 * lra-constraints.c (process_address): Try to improve generation
10259 code for address base + disp.
10260 (lra_constraints): Make correct the code for checking insn setting
10261 up backward equivalence. Remove insn only if it is in the init
10263 * lra-eliminations.c (update_reg_eliminate): Change return value.
10264 (lra_eliminate): Use the result.
10266 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
10268 * config/i386/sse.md (ssescalarnummask): New mode attribute.
10269 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
10270 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
10271 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
10272 register target operands.
10273 (*vec_extractv8hi_sse2): New pattern.
10274 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
10275 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
10276 (*vec_extract<mode>_mem): New insn and split pattern.
10278 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
10280 * config/arm/arm.c (arm_asan_shadow_offset): New function.
10281 (TARGET_ASAN_SHADOW_OFFSET): Define.
10282 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
10283 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
10285 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10287 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
10288 (incr_vec_index): Return -1 if increment not found.
10289 (create_add_on_incoming_edge): Assert if increment not found.
10290 (record_increment): Limit number of increments recorded.
10291 (all_phi_incrs_profitable): Return false if an increment not found.
10292 (replace_profitable_candidates): Don't process increments that were
10294 (analyze_candidates_and_replace): Limit size of incr_vec.
10296 2013-05-07 Richard Biener <rguenther@suse.de>
10298 * calls.c (special_function_p): setjmp-like functions are leaf.
10299 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
10300 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
10302 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
10304 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
10305 (aarch64_simd_mov<mode>): New expander.
10306 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
10307 (aarch64_simd_mov_to_<mode>high): Likewise.
10308 (aarch64_simd_mov_from_<mode>low): Likewise.
10309 (aarch64_simd_mov_from_<mode>high): Likewise.
10310 (aarch64_dup_lane<mode>): Update.
10311 (aarch64_dup_lanedi): New instruction pattern.
10312 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
10313 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
10315 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10317 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
10318 (replace_mult_candidate): Remove unnecessary argument; remove
10319 unnecessary parameter from call to introduce_cast_before_cand.
10320 (replace_unconditional_candidate): Remove unnecessary parameter
10321 from call to replace_mult_candidate.
10322 (replace_conditional_candidate): Likewise.
10323 (insert_initializers): Use make_temp_ssa_name.
10324 (introduce_cast_before_cand): Remove unnecessary argument; use
10325 make_temp_ssa_name.
10326 (replace_one_candidate): Remove unnecessary argument; remove
10327 unnecessary parameter from calls to introduce_cast_before_cand.
10328 (replace_profitable_candidates): Remove unnecessary parameters
10329 from calls to replace_one_candidate.
10331 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10333 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
10334 phi def as possibly hiding a basis for a CAND_ADD whose operands
10335 have been commuted in the analysis.
10336 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
10338 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10340 * config/aarch64/aarch64.md
10341 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
10342 shift value between 0-4.
10344 2013-05-07 Richard Biener <rguenther@suse.de>
10346 * double-int.h (rshift): New overload.
10347 * double-int.c (rshift): New function.
10348 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
10349 (create_reference_ops_from_ref): Remove.
10350 (vn_reference_insert): Use shared ops for constructing the
10351 reference and copy it.
10353 2013-05-07 Richard Biener <rguenther@suse.de>
10355 PR middle-end/57190
10356 * tree-eh.c (sink_clobbers): Properly propagate
10357 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
10359 2013-05-07 Jakub Jelinek <jakub@redhat.com>
10361 PR tree-optimization/57149
10362 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
10363 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
10364 collect_phi_def_edges, execute_late_warn_uninitialized): Use
10365 uninit_undefined_value_p instead of ssa_undefined_value_p.
10368 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
10369 for modifier == EXPAND_INITIALIZER.
10371 2013-05-07 Anton Blanchard <anton@samba.org>
10373 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
10374 for powerpc64 little endian.
10375 * configure: Regenerate.
10377 2013-05-06 Graham Stott <grahams@btinternet.com>
10379 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
10380 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
10381 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
10382 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
10384 2013-05-06 Graham Stott <grahams@btinternet.com>
10386 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
10387 codes which allow non-lvalues.
10389 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
10391 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
10392 components are all 1s.
10393 (integer_minus_onep): New function.
10394 * tree.h (integer_minus_onep): Declare it.
10395 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
10396 integer_minus_onep instead of integer_all_onesp.
10398 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10401 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
10402 variations of these patterns.
10404 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
10406 * config/i386/i386.md (isa): Add x64_sse4 member.
10407 (enabled): Handle x64_sse4.
10408 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
10409 instruction for 64bit SSE4_1 targets. Update insn attributes.
10410 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
10411 instruction for SSE4_1 targets. Update insn attributes.
10412 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
10413 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
10414 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
10416 (*vec_extractv4si): Rename from *sse4_1_pextrd.
10417 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
10418 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
10420 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10423 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
10425 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
10427 * final.c (do_assembler_dialects): Don't handle curly braces and
10428 vertical bar escaped by % as dialect delimiters.
10429 (output_asm_insn): Print curly braces and vertical bar if escaped
10430 by % and ASSEMBLER_DIALECT defined.
10431 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
10432 * doc/tm.texi: Regenerated.
10434 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10436 * config/mips/mips.c: Include tree-pass.h.
10437 (mips_reorg): Split in pre- and post-dbr_schedule parts.
10438 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
10439 (pass_mips_machine_reorg2): New machine specific pass.
10440 (insert_pass_mips_machine_reorg2): New pass plugin definition.
10441 (mips_option_override): Register the new pass.
10442 * rtl.h (cleanup_barriers): Remove prototype.
10443 (dbr_schedule): Likewise.
10444 * jump.c (cleanup_barriers): Make static.
10445 * reorg.c (dbr_schedule): Likewise.
10447 2013-05-06 Richard Biener <rguenther@suse.de>
10449 PR tree-optimization/57185
10450 * tree-parloops.c (add_field_for_reduction): Handle anonymous
10451 SSA names properly.
10453 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
10456 * config/i386/i386.c (add_parameter_dependencies): Add dependence
10457 between "first_arg" and "insn", not "last" and "insn".
10459 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10461 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
10462 (find_candidates_in_block): Re-enable slsr_process_phi.
10463 (create_phi_basis): Fix double counting of candidate adjustment.
10465 2013-05-06 Richard Biener <rguenther@suse.de>
10467 PR middle-end/57147
10468 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
10469 the edge is also fallthru, preserve it and just clear the
10471 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
10472 also complex, preserve that and just clear the fallthru flag.
10473 * tree-inline.c (update_ssa_across_abnormal_edges): Also
10474 update virtual operands.
10476 2013-05-06 Alan Modra <amodra@gmail.com>
10478 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
10479 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
10480 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
10481 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
10482 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
10483 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
10485 2013-05-06 Alan Modra <amodra@gmail.com>
10487 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
10488 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
10489 (DEFAULT_ASM_ENDIAN): Define.
10490 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
10491 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
10492 Update -K PIC clause from sysv4.h.
10493 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
10494 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
10496 2013-05-06 Alan Modra <amodra@gmail.com>
10498 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
10499 twice for little-endian.
10500 (ashrdi3_no_power, ashrdi3): Support little-endian.
10502 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10505 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
10506 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
10508 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
10509 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
10510 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
10513 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
10514 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
10516 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
10517 * config/pa/pa.opt: Make mbig-switch a no-op.
10518 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
10519 (CASE_VECTOR_MODE): Always return SImode.
10520 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
10521 for the !TARGET_BIG_SWITCH case.
10522 * config/pa/pa-linux.h: Likewise.
10523 * config/pa/pa-openbsd.h: Likewise.
10524 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
10525 * config/pa/pa.md (short_jump): Remove define_insn.
10526 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
10527 (casesi0): Remove define_insn.
10528 (type): Remove btable_branch.
10529 (pa_combine_type): Likewise.
10530 (in_nullified_branch_delay): Likewise.
10531 (in_call_delay): Likewise.
10532 (define_delay): Likewise.
10533 (define_insn_reservation "Z3"): Likewise.
10534 (define_insn_reservation "Z4"): Likewise.
10535 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
10536 (pa_adjust_insn_length): Remove adjustment for btable branches.
10537 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
10540 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
10542 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
10543 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
10544 Add m->r,x alternatives.
10545 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
10546 splitters using SWI48x mode iterator.
10547 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
10548 TARGET_64BIT. Add m->x alternative.
10549 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
10550 Add o->x alternative. Enable for TARGET_SSE.
10551 (sse_storeq): Remove expander.
10552 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
10553 with memory input operand.
10554 (*vec_extractv2di_1 splitter): New.
10555 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
10556 * config/i386/i386.md (ssevecmodelower): New mode attribute.
10558 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
10560 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
10561 (INT_LOWPART): Delete.
10562 (extract_MB): Adjust.
10563 (extract_ME): Adjust.
10564 (print_operand): Adjust.
10566 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
10568 * config/rs6000/predicates.md (reg_or_add_cint_operand,
10569 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
10570 (reg_or_logical_cint_operand, easy_fp_constant,
10571 logical_const_operand): Delete "CONST_DOUBLE" case.
10572 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
10573 "HOST_BITS_PER_WIDE_INT == 64" test.
10574 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
10575 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
10576 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
10577 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
10579 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
10580 CONST_DOUBLE DImode/VOIDmode case.
10581 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
10582 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
10583 CONST_DOUBLE VOIDmode case.
10584 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
10585 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
10586 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
10587 Delete CONST_DOUBLE case.
10588 (splitters for mov FMOVE64 const_double): Delete
10589 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
10590 "HOST_BITS_PER_WIDE_INT >= 64" test.
10591 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
10593 (mov DI const_double): Delete.
10595 2013-05-04 Jakub Jelinek <jakub@redhat.com>
10597 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
10598 on op shows all bits zero in mode of a lowpart subreg, return zero.
10600 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
10603 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
10604 to save TFmode registers and DImode to save TImode registers for
10605 caller save operations.
10606 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
10607 mark being partially clobbered since they only use the first
10610 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
10611 and TDmode only use the upper 64-bits of each VSX register.
10613 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10615 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
10616 (find_candidates_in_block): Disable slsr_process_phi.
10618 2013-05-03 Guozhi Wei <carrot@google.com>
10620 * coverage.c (coverage_obj_init): Move the construction of gcov
10622 (build_init_ctor): ... here.
10624 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10626 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
10627 (slsr_cand_d): Redefine def_phi.
10628 (stride_status, phi_adjust_status, count_phis_status): New enums.
10629 (find_phi_def): New.
10630 (find_basis_for_base_expr): New.
10631 (find_basis_for_candidate): Handle hidden bases.
10632 (alloc_cand_and_find_basis): Handle phi candidates.
10633 (slsr_process_phi): New.
10634 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
10635 (create_mul_imm_cand): Likewise.
10636 (create_add_ssa_cand): Exclude phi base candidates.
10637 (create_add_imm_cand): Likewise.
10638 (slsr_process_cast): Likewise.
10639 (slsr_process_copy): Likewise.
10640 (find_candidates_in_block): Handle phi candidates.
10641 (dump_candidate): Likewise.
10642 (unconditional_cands): Delete.
10643 (unconditional_cands_with_known_stride_p): Delete.
10644 (phi_dependent_cand_p): New.
10645 (cand_increment): Handle phi-dependent candidates.
10646 (replace_dependent): Delete.
10647 (replace_mult_candidate): New.
10648 (replace_unconditional_candidate): New.
10649 (incr_vec_index): Move to avoid forward reference.
10650 (create_add_on_incoming_edge): New.
10651 (create_phi_basis): New.
10652 (replace_dependents): Delete.
10653 (replace_conditional_candidate): New.
10654 (phi_add_costs): New.
10655 (replace_uncond_cands_and_profitable_phis): New.
10656 (record_increment): Handle phi adjustments.
10657 (record_phi_increments): New.
10658 (record_increments): Handle phi adjustments.
10659 (phi_incr_cost): New.
10660 (lowest_cost_path): Handle phis.
10661 (total_savings): Likewise.
10662 (analyze_increments): Likewise.
10663 (ncd_with_phi): New.
10664 (ncd_of_cand_and_phis): New.
10665 (nearest_common_dominator_for_cands): Handle phi increments.
10666 (all_phi_incrs_profitable): New.
10667 (replace_profitable_candidates): Handle phi-dependent candidates.
10668 (analyze_candidates_and_replace): Likewise.
10670 2013-05-03 Teresa Johnson <tejohnson@google.com>
10673 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
10674 do not exceed REG_BR_PROB_BASE.
10676 2013-05-03 Jeff Law <law@redhat.com>
10678 PR tree-optimization/57144
10679 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
10680 operand of the condition will bit into the new type when eliminating
10681 a cast feeding a condition.
10683 2013-05-03 Jakub Jelinek <jakub@redhat.com>
10685 PR rtl-optimization/57130
10686 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
10687 of COMPARE as in_code to the recursive call if needed.
10689 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
10691 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
10692 (enabled): Handle new members.
10693 * config/i386/sse.md (*vec_concatv2si): Merge from
10694 *vec_concatv2si_sse2 and vec_concatv2si_sse.
10695 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
10697 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
10699 PR tree-optimization/57027
10700 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
10701 for fnms opportunity, check we got the prerequisite kind
10702 of tree / gimple before using accessor functions.
10704 2013-05-03 Richard Biener <rguenther@suse.de>
10706 * double-int.h (lshift): New overload without precision
10707 and arith argument.
10708 (operator *=, operator +=, operator -=): Move ...
10709 * double-int.c (operator *=, operator +=, operator -=): ... here
10710 and implement more efficiently.
10711 (mul_double_with_sign): Remove.
10712 (lshift_double): Adjust to take unsinged shift argument, push
10713 dispatching code to callers.
10714 (mul_double_wide_with_sign): Add early out for callers that
10715 are not interested in high parts or overflow.
10716 (lshift): New function.
10717 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
10718 dispatch code here.
10719 (lrotate, rrotate): Use logical shifts.
10720 * expr.c (get_inner_reference): Use lshift.
10721 * fixed-value.c (do_fixed_divide): Likewise.
10722 * tree-dfa.c (get_ref_base_and_extent): Likewise.
10723 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
10724 (indirect_refs_may_alias_p): Likewise.
10725 (stmt_kills_ref_p_1): Likewise.
10727 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
10729 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
10731 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
10733 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
10734 scalar form of FABD instruction.
10736 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
10738 * lra-constraints.c (process_alt_operands): Add checking alt
10739 number to choose the best alternative.
10741 2013-05-02 Richard Biener <rguenther@suse.de>
10743 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
10744 bitmap and its handling.
10745 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
10747 2013-05-02 Richard Biener <rguenther@suse.de>
10749 PR middle-end/57140
10750 * tree-inline.c (copy_loops): Properly handle removed loops.
10751 (copy_cfg_body): Mark destination loops for fixup if source
10752 loops needed fixup.
10754 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
10757 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
10758 generating simple_return for naked functions.
10760 2013-05-02 Martin Jambor <mjambor@suse.cz>
10762 PR middle-end/56988
10763 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
10764 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
10766 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
10767 ipa_agg_replacement_value structures.
10768 (known_aggs_to_agg_replacement_list): Likewise.
10769 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
10770 (read_agg_replacement_chain): Likewise.
10771 (ipcp_transform_function): Also check that by_ref flags match.
10773 2013-05-02 Richard Biener <rguenther@suse.de>
10775 * graphds.h (struct graph): Add obstack member.
10776 * graphds.c (new_graph): Initialize obstack and allocate
10778 (add_edge): Allocate edge from the obstack.
10779 (free_graph): Free the obstack instead of all edges and vertices.
10781 2013-05-02 Teresa Johnson <tejohnson@google.com>
10783 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
10785 * cfg.c (update_bb_profile_for_threading): Ditto.
10786 * tree-inline.c (copy_bb): Ditto.
10787 (copy_edges_for_bb): Ditto.
10788 (initialize_cfun): Ditto.
10789 (copy_cfg_body): Ditto.
10790 (expand_call_inline): Ditto.
10791 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
10792 (estimate_node_size_and_time): Ditto.
10793 (inline_merge_summary): Ditto.
10794 * cgraphclones.c (cgraph_clone_edge): Ditto.
10795 (cgraph_clone_node): Ditto.
10796 * sched-rgn.c (compute_dom_prob_ps): Ditto.
10797 (compute_trg_info): Ditto.
10799 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10801 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
10802 S reg when fp attribute set.
10803 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
10805 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10807 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
10809 (*and_one_cmplsi3_compare0_uxtw): Likewise.
10810 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
10811 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
10813 2013-05-02 Richard Biener <rguenther@suse.de>
10815 * tree-scalar-evolution.c (scev_info_hasher): Remove.
10816 (struct instantiate_cache_entry): New type.
10817 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
10818 (struct instantiate_cache_type): New type.
10819 (set_instantiated_value, get_instantiated_value): Remove.
10820 (get_instantiated_value_entry): New function.
10821 (instantiate_scev_name): Use the new cache and adjust.
10822 (instantiate_scev_poly): Adjust.
10823 (instantiate_scev_binary): Likewise.
10824 (instantiate_array_ref): Likewise.
10825 (instantiate_scev_convert): Likewise.
10826 (instantiate_scev_not): Likewise.
10827 (instantiate_scev_3): Likewise.
10828 (instantiate_scev_2): Likewise.
10829 (instantiate_scev_r): Likewise.
10830 (instantiate_scev): Likewise.
10831 (resolve_mixers): Likewise.
10833 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
10836 * lra-constraints.c (best_small_class_operands_num): Remove.
10837 (process_alt_operands): Remove small_class_operands_num. Take
10838 small classes operands into losers and only if the operand is not
10839 matched. Modify debugging output.
10840 (curr_insn_transform): Remove best_small_class_operands_num.
10843 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10845 * config/aarch64/aarch64-builtins.c
10846 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
10847 * config/aarch64/aarch64-simd-builtins.def
10848 (reduc_splus_): Add new modes.
10849 (reduc_uplus_): New.
10850 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
10851 (reduc_uplus_v4sf): Likewise.
10852 (reduc_splus_v4sf): Likewise.
10853 (aarch64_addv<mode>): Likewise.
10854 (reduc_uplus_<mode>): Likewise.
10855 (reduc_splus_<mode>): Likewise.
10856 (aarch64_addvv2di): Likewise.
10857 (reduc_uplus_v2di): Likewise.
10858 (reduc_splus_v2di): Likewise.
10859 (aarch64_addvv2si): Likewise.
10860 (reduc_uplus_v2si): Likewise.
10861 (reduc_splus_v2si): Likewise.
10862 (reduc_<sur>plus_<mode>): New.
10863 (reduc_<sur>plus_v2di): Likewise.
10864 (reduc_<sur>plus_v2si): Likewise.
10865 (reduc_<sur>plus_v4sf): Likewise.
10866 (aarch64_addpv4sf): Likewise.
10867 * config/aarch64/arm_neon.h
10868 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
10869 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
10870 add UNSPEC_SADDV, UNSPEC_UADDV.
10872 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
10874 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10876 * config/aarch64/arm_neon.h
10877 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
10879 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10881 * config/aarch64/aarch64-builtins
10882 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
10884 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10886 * config/aarch64/aarch64-simd-builtins.def
10887 (reduc_smax_): New.
10888 (reduc_smin_): Likewise.
10889 (reduc_umax_): Likewise.
10890 (reduc_umin_): Likewise.
10891 (reduc_smax_nan_): Likewise.
10892 (reduc_smin_nan_): Likewise.
10895 (smax): Update for V2SF, V4SF and V2DF modes.
10898 (smin_nan): Likewise.
10899 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10900 (<su><maxmin><mode>3): ...This, refactor.
10901 (s<maxmin><mode>3): New.
10902 (<maxmin_uns><mode>3): Likewise.
10903 (reduc_<maxmin_uns>_<mode>): Refactor.
10904 (reduc_<maxmin_uns>_v4sf): Likewise.
10905 (reduc_<maxmin_uns>_v2si): Likewise.
10906 (aarch64_<fmaxmin><mode>: Remove.
10907 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10909 (vmin<q>_f<32,64>): Likewise.
10910 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10912 (su): Add mappings for smax, smin, umax, umin.
10914 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10915 (FMAXMIN): Rename as...
10916 (FMAXMIN_UNS): ...This.
10918 (fmaxminv): Likewise.
10919 (fmaxmin): Likewise.
10921 (maxmin_uns_op): Likewise.
10923 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10925 * config/aarch64/arm_neon.h
10926 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10927 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10928 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10930 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10932 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10933 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10935 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10937 * config/aarch64/aarch64-simd.md
10938 (vcond<mode>_internal): Handle special cases for constant masks.
10939 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10940 (vcondu<mode><mode>): Likewise.
10941 (vcond<v_cmp_result><mode>): New.
10943 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10945 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10946 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10947 * config/aarch64/aarch64-simd-builtins.def
10948 (cmeq): Update to BUILTIN_VALLDI.
10953 * config/aarch64/arm_neon.h
10954 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10955 to builtins or C as appropriate.
10957 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10959 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10961 (cmhi): Rename to...
10963 * config/aarch64/aarch64-simd.md
10964 (simd_mode): Add SF.
10965 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10966 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10967 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10968 (cstore<mode>_neg): ...This.
10969 * config/aarch64/iterators.md
10971 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10972 (COMPARISONS): New.
10973 (UCOMPARISONS): Likewise.
10974 (optab): Add missing comparisons.
10980 (VCMP_S): Likewise.
10981 (VCMP_U): Likewise.
10982 (V_cmp_result): Add DF, SF modes.
10983 (v_cmp_result): Likewise.
10985 (vmtype): Likewise.
10986 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
10988 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
10990 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
10991 define_insn to define_insn_and_split.
10992 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
10993 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
10994 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
10995 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
10996 (thumb2_negscc): Likewise.
10998 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11000 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
11002 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11004 * config/arm/thumb2.md: Remove trailing whitespaces.
11006 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11008 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
11009 Use gen_int_mode rather than GEN_INT.
11011 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
11013 * value-prof.c (stream_in_histogram_value): Remove the strayed
11016 2013-04-30 Richard Biener <rguenther@suse.de>
11018 PR middle-end/57122
11019 * cfghooks.c (split_edge): Properly check for the loop latch edge.
11021 2013-04-30 Richard Biener <rguenther@suse.de>
11023 PR middle-end/57107
11024 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
11026 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
11028 PR rtl-optimization/56957
11029 PR rtl-optimization/57105
11030 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
11031 variable. Use just INSN_UID for determining whether an insn
11032 should be only disconnected from the insn stream.
11033 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
11035 2013-04-30 Jakub Jelinek <jakub@redhat.com>
11037 PR tree-optimization/57104
11038 * tsan.c (instrument_expr): Don't instrument accesses to
11039 DECL_HARD_REGISTER VAR_DECLs.
11041 2013-04-30 Richard Biener <rguenther@suse.de>
11043 * function.h (loops_for_fn): New inline function.
11044 (set_loops_for_fn): Likewise.
11045 * cfgloop.h (place_new_loop): Add struct function parameter.
11046 (get_loop): Likewise.
11047 (get_loops): Likewise.
11048 (number_of_loops): Likewise.
11049 (fel_next): Adjust.
11050 (fel_init): Likewise.
11051 * cfg.c (get_loop_copy): Adjust.
11052 * cfgloop.c (flow_loops_dump): Likewise.
11053 (record_loop_exits): Likewise.
11054 (verify_loop_structure): Likewise.
11055 * cfgloopanal.c (mark_irreducible_loops): Likewise.
11056 (estimate_reg_pressure_cost): Likewise.
11057 (mark_loop_exit_edges): Likewise.
11058 * cfgloopmanip.c (place_new_loop): Likewise.
11059 (add_loop): Likewise.
11060 (duplicate_loop): Likewise.
11061 * graph.c (draw_cfg_nodes): Likewise.
11062 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
11063 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
11064 (extract_affine_chrec): Likewise.
11065 (build_scop_iteration_domain): Likewise.
11066 * graphite.c (graphite_initialize): Likewise.
11067 * ira-build.c (create_loop_tree_nodes): Likewise.
11068 (more_one_region_p): Likewise.
11069 (rebuild_regno_allocno_maps): Likewise.
11070 (mark_loops_for_removal): Likewise.
11071 (mark_all_loops_for_removal): Likewise.
11072 (remove_unnecessary_regions): Likewise.
11073 (ira_build): Likewise.
11074 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
11075 * loop-init.c (fix_loop_structure): Likewise.
11076 (gate_rtl_move_loop_invariants): Likewise.
11077 (gate_rtl_unswitch): Likewise.
11078 (gate_rtl_unroll_and_peel_loops): Likewise.
11079 (rtl_doloop): Likewise.
11080 * lto-streamer-in.c (input_cfg): Likewise.
11081 * lto-streamer-out.c (output_cfg): Likewise.
11082 * modulo-sched.c (sms_schedule): Likewise.
11083 * predict.c (tree_estimate_probability): Likewise.
11084 (tree_estimate_probability_driver): Likewise.
11085 (estimate_loops): Likewise.
11086 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
11087 (move_sese_region_to_fn): Likewise.
11088 (debug_loop_num): Likewise.
11089 * tree-chrec.c (chrec_evaluate): Likewise.
11090 (hide_evolution_in_other_loops_than_loop): Likewise.
11091 (chrec_component_in_loop_num): Likewise.
11092 (reset_evolution_in_loop): Likewise.
11093 (evolution_function_is_invariant_rec_p): Likewise.
11094 * tree-if-conv.c (main_tree_if_conversion): Likewise.
11095 * tree-inline.c (copy_loops): Likewise.
11096 (copy_cfg_body): Likewise.
11097 (tree_function_versioning): Likewise.
11098 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
11099 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
11101 (add_to_evolution_1): Likewise.
11102 (scev_const_prop): Likewise.
11103 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
11104 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
11105 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
11106 (tree_ssa_lim_initialize): Likewise.
11107 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
11108 (verify_loop_closed_ssa): Likewise.
11109 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
11110 (tree_ssa_loop_im): Likewise.
11111 (tree_ssa_loop_unswitch): Likewise.
11112 (tree_vectorize): Likewise.
11113 (check_data_deps): Likewise.
11114 (tree_ssa_loop_ivcanon): Likewise.
11115 (tree_ssa_loop_bounds): Likewise.
11116 (tree_complete_unroll): Likewise.
11117 (tree_complete_unroll_inner): Likewise.
11118 (tree_parallelize_loops): Likewise.
11119 (tree_ssa_loop_prefetch): Likewise.
11120 (tree_ssa_loop_ivopts): Likewise.
11121 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
11122 * tree-vectorizer.c (vectorize_loops): Likewise.
11124 2013-04-29 Mike Frysinger <vapier@gentoo.org>
11126 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
11127 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
11128 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
11129 with EABI_LINK_SPEC.
11131 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11134 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
11137 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
11140 * lra-constraints.c (process_alt_operands): Discourage a bit more
11141 using memory for pseudos. Print cost dump for alternatives.
11142 Modify cost values for conflicts with early clobbers.
11143 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
11145 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11148 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
11150 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11152 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
11153 from/to S register.
11154 (movdi_aarch64): Support LDR/STR from/to D register.
11156 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11158 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
11159 or higher by default.
11161 2013-04-29 Richard Biener <rguenther@suse.de>
11163 PR middle-end/57075
11164 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
11165 even if not adding abnormal edges for calls that can make
11168 2013-04-29 Richard Biener <rguenther@suse.de>
11170 PR middle-end/57103
11171 * tree-cfg.c (move_stmt_op): Fix condition under which to update
11173 (move_stmt_r): Remove redundant checking.
11175 2013-04-29 Teresa Johnson <tejohnson@google.com>
11178 * basic-block.h (apply_scale): New function.
11179 (apply_probability): Use apply_scale.
11180 * gimple-streamer-in.c (input_bb): Ditto.
11181 * lto-streamer-in.c (input_cfg): Ditto.
11182 * lto-cgraph.c (merge_profile_summaries): Ditto.
11183 * tree-optimize.c (execute_fixup_cfg): Ditto.
11184 * tree-inline.c (copy_bb): Update comment to use apply_scale.
11185 (copy_edges_for_bb): Ditto.
11186 (copy_cfg_body): Ditto.
11188 2013-04-29 Tom de Vries <tom@codesourcery.com>
11190 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
11191 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
11192 (tail_merge_optimize): Handle current_loops == NULL.
11194 2013-04-26 Jeff Law <law@redhat.com>
11196 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
11197 (simplify_cond_using_ranges): Generalize code to simplify
11198 COND_EXPRs where one argument is a constant and the other
11199 is an SSA_NAME created by an integral type conversion.
11201 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11203 * config/arm/arm.md (store_minmaxsi): Use only when
11204 optimize_insn_for_size_p.
11206 2013-04-29 Christian Bruel <christian.bruel@st.com>
11209 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
11211 2013-04-29 Richard Biener <rguenther@suse.de>
11213 PR middle-end/57089
11214 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
11215 loop tree make sure to schedule a fixup for the child as well.
11216 (expand_omp_for_generic): Properly add loops.
11217 (expand_omp_for_static_nochunk): Likewise.
11218 (expand_omp_for_static_chunk): Likewise.
11219 (expand_omp_for): For the degenerate case fixup loops.
11220 (expand_omp_sections): Fix default bb placement in loops.
11221 (expand_omp_atomic_pipeline): Properly add loops.
11223 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11225 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
11227 2013-04-29 Tom de Vries <tom@codesourcery.com>
11229 * tree-ssa-tail-merge.c: Update header comment.
11231 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11233 * config/aarch64/arm_neon.h
11234 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
11235 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
11236 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
11237 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
11238 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
11239 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
11240 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
11241 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
11243 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11245 * config/aarch64/aarch64-simd.md
11246 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
11247 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
11248 fix_trunc, fixuns_trunc.
11249 (ftrunc<VDQF:mode>2): New.
11250 * config/aarch64/iterators.md (optab): Add fix, fixuns.
11251 (fix_trunc_optab): New.
11253 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11255 * config/aarch64/aarch64-builtins.c
11256 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
11257 iceilf, lround, iroundf.
11259 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11262 * config/i386/i386.h (enum ix86_tune_indices)
11263 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
11264 New, split from X86_TUNE_INTER_UNIT_MOVES.
11265 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
11266 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
11267 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
11268 (TARGET_INTER_UNIT_MOVES): Remove.
11269 * config/i386/i386.c (initial_ix86_tune_features): Update.
11270 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
11271 (ix86_expand_convert_uns_didf_sse): Use
11272 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11273 (ix86_expand_vector_init_one_nonzero): Ditto.
11274 (ix86_expand_vector_init_interleave): Ditto.
11275 (inline_secondary_memory_needed): Return true for moves from SSE class
11276 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
11277 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
11278 * config/i386/constraints.md (Yi, Ym): Depend on
11279 TARGET_INTER_UNIT_MOVES_TO_VEC.
11280 (Yj, Yn): New constraints.
11281 * config/i386/i386.md (*movdi_internal): Change constraints of
11282 operand 1 from Yi to Yj and from Ym to Yn.
11283 (*movsi_internal): Ditto.
11284 (*movdf_internal): Ditto.
11285 (*movsf_internal): Ditto.
11286 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
11287 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11288 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
11289 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
11290 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
11291 * config/i386/sse.md (movdi_to_sse): Ditto.
11292 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
11293 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
11294 TARGET_INTER_UNIT_MOVES.
11295 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
11296 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
11297 instead of TARGET_INTER_UNIT_MOVES.
11298 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
11299 operand 1 from Yi to Yj and from Ym to Yn.
11301 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11303 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
11304 (float_truncate_hi_): Likewise.
11305 (float_extend_lo_): Likewise.
11306 (float_truncate_lo_): Likewise.
11307 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
11308 (aarch64_float_extend_lo_v2df): Likewise.
11309 (vec_unpacks_hi_v4sf): Likewise.
11310 (aarch64_float_truncate_lo_v2sf): Likewise.
11311 (aarch64_float_truncate_hi_v4sf): Likewise.
11312 (vec_pack_trunc_v2df): Likewise.
11313 (vec_pack_trunc_df): Likewise.
11315 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11317 * config/aarch64/aarch64-builtins.c
11318 (aarch64_fold_builtin): Fold float conversions.
11319 * config/aarch64/aarch64-simd-builtins.def
11320 (floatv2si, floatv4si, floatv2di): New.
11321 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
11322 * config/aarch64/aarch64-simd.md
11323 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
11324 * config/aarch64/iterators.md (FLOATUORS): New.
11325 (optab): Add float, floatuns.
11326 (su_optab): Likewise.
11328 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11330 * config/aarch64/aarch64-builtins.c
11331 (aarch64_builtin_vectorized_function): Use new names for
11333 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
11334 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
11335 (fcvtzu): Split as...
11336 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
11337 (fcvtas): Split as...
11338 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
11339 (fcvtau): Split as...
11340 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
11341 (fcvtps): Split as...
11342 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
11343 (fcvtpu): Split as...
11344 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
11345 (fcvtms): Split as...
11346 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
11347 (fcvtmu): Split as...
11348 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
11349 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
11350 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
11351 (lfrintnusf, lfrintnudf): Likewise.
11352 * config/aarch64/aarch64-simd.md
11353 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
11355 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
11356 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
11357 (fcvt_pattern): Likewise.
11359 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11361 * config/aarch64/aarch64-simd.md
11362 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
11363 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
11365 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11367 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
11368 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
11369 (vrnd<a,m,n,p>_f32): Implement using builtins.
11370 (vrnd<i,x><q>_f<32, 64>): New.
11372 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11374 * config/aarch64/aarch64-builtins.c
11375 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
11376 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
11377 (frintz): Rename to...
11379 (frintp): Rename to...
11381 (frintm): Rename to...
11383 (frinti): Rename to...
11384 (nearbyint): ...this.
11385 (frintx): Rename to...
11387 (frinta): Rename to...
11389 * config/aarch64/aarch64-simd.md
11390 (aarch64_frint<frint_suffix><mode>): Delete.
11391 (<frint_pattern><mode>2): Convert to insn.
11392 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
11393 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
11394 (frint_pattern): Likewise.
11395 (frint_suffix): Likewise.
11397 2013-04-29 Richard Biener <rguenther@suse.de>
11399 PR tree-optimization/57081
11400 * loop-init.c: Include tree-flow.h.
11401 (loop_optimizer_finalize): Free number of iteration estimates.
11402 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
11404 2013-04-29 Jakub Jelinek <jakub@redhat.com>
11406 PR tree-optimization/57083
11407 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
11408 non-singleton shift count range, zero extend low_bound for uns case.
11410 * config/i386/predicates.md (general_vector_operand): New predicate.
11411 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
11412 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
11413 if they aren't nonimmediate operands. If their original values
11414 satisfy const_vector_equal_evenodd_p, don't shift them.
11415 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
11416 predicates. For the SSE4.1 case force operands[{1,2}] into registers
11417 if not nonimmediate_operand.
11418 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
11419 instead of register_operand.
11420 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
11422 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
11424 * stor-layout.c (finalize_size_functions): Allocate a structure and
11425 reset cfun before dumping the functions.
11427 2013-04-27 Jakub Jelinek <jakub@redhat.com>
11429 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
11432 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
11433 use xop_pmacsdqh if uns_p.
11434 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
11435 the immediate rotate count.
11437 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
11439 * rtl.h (struct rtx_def): Add comment for field jump.
11440 (LRA_SUBREG_P): New macro.
11441 * recog.c (register_operand): Check LRA_SUBREG_P.
11442 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11444 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11445 (lra_final_code_change): Skip subreg change for operators.
11446 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11447 if there are no operand changes.
11448 * lra-constraints.c (curr_insn_set): New.
11449 (match_reload): Set LRA_SUBREG_P.
11450 (emit_spill_move): Ditto.
11451 (check_and_process_move): Use curr_insn_set. Process only single
11452 set insns. Don't initialize sec_mem_p and change_p.
11453 (simplify_operand_subreg): Use LRA_SUBREG_P.
11454 (reg_in_class_p): New function.
11455 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11456 of #ifdef. Add code to remove cycling.
11457 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11458 non-null disp. Reload inner instead of disp when base and index
11459 are null. Try to put lo_sum into register.
11460 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11461 (check_and_process_move): Move code for move cost check to
11462 simple_move_p. Remove equiv_substitution.
11463 (simple_move_p): New function.
11464 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11465 curr_insn_set. Call check_and_process_move only for single set
11466 insns. Use the new function. Move call of check_and_process_move
11467 after operand equiv substitution and address process.
11469 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11472 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
11473 with nonlocal goto receivers or returns twice calls, ignore
11474 unininitialized values from abnormal edges to nl goto receiver
11475 or returns twice call.
11477 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11479 PR tree-optimization/57051
11480 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
11481 and VEC_RSHIFT_EXPR if shift count is a multiple of element
11484 2013-04-26 Richard Biener <rguenther@suse.de>
11486 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
11487 (expand_omp_taskreg): Likewise. Mark loops for fixup.
11488 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
11489 (fixup_loop_arrays_after_move): New function.
11490 (move_sese_region_to_fn): Properly outline the loop tree parts
11491 of the SESE region.
11493 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
11495 * config/i386/i386.md (type, unit): Fix long lines.
11497 2013-04-26 Richard Biener <rguenther@suse.de>
11499 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
11500 (lto-streamer-out.o): Likewise.
11501 * cfgloop.c (init_loops_structure): Export, add struct function
11502 argument and adjust.
11503 (flow_loops_find): Adjust.
11504 * cfgloop.h (enum loop_estimation): Add EST_LAST.
11505 (init_loops_structure): Declare.
11506 * lto-streamer-in.c: Include cfgloop.h.
11507 (input_cfg): Input the loop tree.
11508 * lto-streamer-out.c: Include cfgloop.h.
11509 (output_cfg): Output the loop tree.
11510 (output_struct_function_base): Do not drop PROP_loops.
11512 2013-03-26 Richard Biener <rguenther@suse.de>
11514 * tree-cfg.c (execute_build_cfg): Build the loop tree.
11515 (pass_build_cfg): Provide PROP_loops.
11516 (move_sese_region_to_fn): Remove loops that are outlined into fn
11518 * tree-inline.c: Include cfgloop.h.
11519 (initialize_cfun): Do not drop PROP_loops.
11520 (copy_loops): New function.
11521 (copy_cfg_body): Copy loop structure.
11522 (tree_function_versioning): Initialize destination loop tree.
11523 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
11524 (pass_parallelize_loops): Do IL verification.
11525 * loop-init.c (loop_optimizer_init): Fixup loops if required.
11526 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
11527 the CFG make sure we fixup loops as well.
11528 * tree-ssa-tail-merge.c: Include cfgloop.h.
11529 (replace_block_by): When merging loop latches mark loops for fixup.
11530 * lto-streamer-out.c (output_struct_function_base): Drop
11531 PROP_loops for now.
11532 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
11533 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
11534 * ipa-split.c: Include cfgloop.h.
11535 (split_function): Add the new return block to the loop tree root.
11536 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
11537 whether we have removed the forwarder block.
11538 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
11539 * cfgloop.h (place_new_loop): Declare.
11540 * cfgloopmanip.c (place_new_loop): Export.
11541 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
11542 (tree-switch-conversion.o): Likewise.
11543 (tree-complex.o): Likewise.
11544 (tree-inline.o): Likewise.
11545 (tree-ssa-tailmerge.o): Likewise.
11546 (ipa-split.o): Likewise.
11547 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
11548 (tree-ssa-copy.o): Likewise.
11549 * tree-switch-conversion.c: Include cfgloop.h
11550 (process_switch): If we emit a bit-test cascade, schedule loops
11552 * tree-complex.c: Include cfgloop.h.
11553 (expand_complex_div_wide): Properly add new basic-blocks to loops.
11554 * asan.c: Include cfgloop.h.
11555 (create_cond_insert_point): Properly add new basic-blocks to
11556 loops, schedule loop fixup.
11557 * cfgloop.c (verify_loop_structure): Check that looks are not
11559 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
11561 (expand_omp_for_generic): Likewise.
11562 (expand_omp_sections): Likewise.
11563 (expand_omp_atomic_pipeline): Schedule loops for fixup.
11564 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
11565 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
11566 is initialized, not when loops are present.
11567 * tree-parloops.c (parallelize_loops): Remove checking here.
11568 * passes.c (init_optimization_passes): Schedule a copy-propagation
11569 pass before complete unrolling of inner loops.
11571 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11573 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
11574 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
11575 (colorize_init): Add argument to _WIN32 version.
11576 * toplev.c: Include diagnostic-color.h.
11577 (process_options): Default to -fdiagnostics-color=auto if
11578 GCC_COLORS env var is in the environment.
11579 * common.opt (fdiagnostics-color=): Add Var and Init.
11580 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
11581 env var is in the environment, the default is auto rather than never.
11583 * diagnostic.h (file_name_as_prefix): Add context argument.
11584 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
11585 the string as locus.
11586 * langhooks.c (lhd_print_error_function): Adjust caller.
11588 2013-04-25 Lawrence Crowl <crowl@google.com>
11590 * var-tracking.c (shared_hash_def::htab):
11591 Change type to hash_table. Update dependent calls and types.
11593 2013-04-25 Lawrence Crowl <crowl@google.com>
11595 * Makefile.in: Update as needed below.
11597 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
11598 Move declaration to after the type's method definitons.
11600 * attribs.c (htab_t scoped_attributes::attribute_hash):
11601 Change type to hash_table. Update dependent calls and types.
11603 * bitmap.c (htab_t bitmap_desc_hash):
11604 Change type to hash_table. Update dependent calls and types.
11606 * cselib.c (htab_t cselib_hash_table):
11607 Change type to hash_table. Update dependent calls and types.
11609 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
11610 (hash_string_slot_node): Move implementation into lto-streamer.h
11611 struct string_slot_hasher.
11612 (eq_string_slot_node): Likewise.
11614 * data-streamer-out.c: Update output_block::string_hash_table
11615 dependent calls and types.
11617 * dwarf2cfi.c (htab_t trace_index):
11618 Change type to hash_table. Update dependent calls and types.
11620 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
11621 Change type to hash_table. Update dependent calls and types.
11622 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
11623 (htab_t optimize_external_refs::map): Likewise.
11624 (htab_t output_comp_unit::extern_map): Likewise.
11625 (htab_t output_comdat_type_unit::extern_map): Likewise.
11626 (htab_t output_macinfo::macinfo_htab): Likewise.
11627 (htab_t optimize_location_lists::htab): Likewise.
11628 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
11630 * except.c (htab_t ehspec_hash_type):
11631 Change type to hash_table. Update dependent calls and types.
11632 (assign_filter_values::ttypes): Likewise.
11633 (assign_filter_values::ehspec): Likewise.
11634 (sjlj_assign_call_site_values::ar_hash): Likewise.
11635 (convert_to_eh_region_ranges::ar_hash): Likewise.
11637 * gcse.c (htab_t pre_ldst_table):
11638 Change type to hash_table. Update dependent calls and types.
11640 * ggc-common.c (htab_t saving_htab):
11641 Change type to hash_table. Update dependent calls and types.
11642 (htab_t loc_hash): Likewise.
11643 (htab_t ptr_hash): Likewise.
11644 (call_count): Rename ggc_call_count.
11645 (call_alloc): Rename ggc_call_alloc.
11646 (loc_descriptor): Rename make_loc_descriptor.
11647 (add_statistics): Rename ggc_add_statistics.
11649 * ggc-common.c (saving_htab):
11650 Change type to hash_table. Update dependent calls and types.
11652 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
11653 (push_gimplify_context): Likewise.
11654 (pop_gimplify_context): Likewise.
11655 (struct gimple_temp_hash_elt): Added.
11656 (struct gimplify_hasher): Likewise.
11657 (struct gimplify_ctx.temp_htab):
11658 Change type to hash_table. Update dependent calls and types.
11660 * gimple-fold.c: Include gimplify-ctx.h.
11662 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
11663 Change type to hash_table. Update dependent calls and types.
11664 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
11665 avoid potential global name collision.
11667 * gimplify.c: Include gimplify-ctx.h.
11668 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
11669 (htab_t gimplify_ctx::temp_htab):
11670 Update dependent calls and types for new type hash_table.
11671 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
11672 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
11674 * gimplify-ctx.h: New.
11675 (struct gimple_temp_hash_elt): Move from gimplify.c.
11676 (class gimplify_hasher): New.
11677 (struct gimplify_ctx): Move from gimple.h.
11678 (htab_t gimplify_ctx::temp_htab):
11679 Change type to hash_table. Update dependent calls and types.
11681 * graphite-clast-to-gimple.c: Include graphite-htab.h.
11682 (htab_t ivs_params::newivs_index):
11683 Change type to hash_table. Update dependent calls and types.
11684 (htab_t ivs_params::params_index): Likewise.
11685 (htab_t print_generated_program::params_index): Likewise.
11686 (htab_t gloog::newivs_index): Likewise.
11687 (htab_t gloog::params_index): Likewise.
11689 * graphite.c: Include graphite-htab.h.
11690 4htab_t graphite_transform_loops::bb_pbb_mapping):
11691 Change type to hash_table. Update dependent calls and types.
11693 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
11694 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
11695 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
11697 * graphite-dependences.c: Include graphite-htab.h.
11698 (loop_is_parallel_p): Change hash table type of parameter.
11700 * graphite-htab.h: New.
11701 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
11702 (extern find_pbb_via_hash): Move from graphite-poly.h.
11703 (extern loop_is_parallel_p): Move from graphite-poly.h.
11704 (extern get_loop_body_pbbs): Move from graphite-poly.h.
11706 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
11707 (extern loop_is_parallel_p): Move to graphite-htab.h.
11708 (extern get_loop_body_pbbs): Move to graphite-htab.h.
11710 * haifa-sched.c (htab_t delay_htab):
11711 Change type to hash_table. Update dependent calls and types.
11712 (htab_t delay_htab_i2): Likewise.
11714 * ira-color.c (htab_t allocno_hard_regs_htab):
11715 Change type to hash_table. Update dependent calls and types.
11717 * ira-costs.c (htab_t cost_classes_htab):
11718 Change type to hash_table. Update dependent calls and types.
11720 * loop-invariant.c (htab_t merge_identical_invariants::eq):
11721 Change type to hash_table. Update dependent calls and types.
11723 * loop-iv.c (htab_t bivs):
11724 Change type to hash_table. Update dependent calls and types.
11726 * loop-unroll.c (htab_t opt_info::insns_to_split):
11727 Change type to hash_table. Update dependent calls and types.
11728 (htab_t opt_info::insns_with_var_to_expand): Likewise.
11730 * lto-streamer.h (struct string_slot): Move from data-streamer.h
11731 (struct string_slot_hasher): New.
11732 (htab_t output_block::string_hash_table):
11733 Change type to hash_table. Update dependent calls and types.
11735 * lto-streamer-in.c (freeing_string_slot_hasher): New.
11736 (htab_t file_name_hash_table):
11737 Change type to hash_table. Update dependent calls and types.
11739 * lto-streamer-out.c: Update output_block::string_hash_table dependent
11742 * lto-streamer.c (htab_t tree_htab):
11743 Change type to hash_table. Update dependent calls and types.
11745 * omp-low.c: Include gimplify-ctx.h.
11747 * passes.c (htab_t name_to_pass_map):
11748 Change type to hash_table. Update dependent calls and types.
11749 (pass_traverse): Rename to passes_pass_traverse.
11751 * plugin.c (htab_t event_tab):
11752 Change type to hash_table. Update dependent calls and types.
11754 * postreload-gcse.c (htab_t expr_table):
11755 Change type to hash_table. Update dependent calls and types.
11756 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
11758 * sese.c (debug_rename_map_1): Make extern.
11759 (htab_t copy_bb_and_scalar_dependences::rename_map):
11760 Change type to hash_table. Update dependent calls and types.
11762 * sese.h (extern debug_rename_map): Move to .c file.
11764 * store-motion.c (htab_t store_motion_mems_table):
11765 Change type to hash_table. Update dependent calls and types.
11767 * trans-mem.c (htab_t tm_new_mem_hash):
11768 Change type to hash_table. Update dependent calls and types.
11770 * tree-browser.c (htab_t TB_up_ht):
11771 Change type to hash_table. Update dependent calls and types.
11773 * tree-cfg.c (htab_t discriminator_per_locus):
11774 Change type to hash_table. Update dependent calls and types.
11776 * tree-complex.c: Include tree-hasher.h
11777 (htab_t complex_variable_components):
11778 Change type to hash_table. Update dependent calls and types.
11780 * tree-eh.c (htab_t finally_tree):
11781 Change type to hash_table. Update dependent calls and types.
11783 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
11784 struct int_tree_hasher.
11785 (extern int_tree_map_eq): Likewise.
11786 (uid_decl_map_hash): Removed.
11787 (extern decl_tree_map_eq): Likewise.
11789 * tree-hasher.h: New.
11790 (struct int_tree_hasher): New.
11791 (typedef int_tree_htab_type): New.
11793 * tree-inline.c: Include gimplify-ctx.h.
11795 * tree-mudflap.c: Include gimplify-ctx.h.
11797 * tree-parloops.c: Include tree-hasher.h.
11798 (htab_t eliminate_local_variables_stmt::decl_address):
11799 Change type to hash_table. Update dependent calls and types.
11800 (htab_t separate_decls_in_region::decl_copies): Likewise.
11802 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
11803 Change type to hash_table. Update dependent calls and types.
11805 * tree-sra.c (candidates):
11806 Change type to hash_table. Update dependent calls and types.
11808 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
11810 (int_tree_map_hash): Likewise.
11812 * tree-ssa-dom.c (htab_t avail_exprs):
11813 Change type to hash_table. Update dependent calls and types.
11815 * tree-ssa-live.c (var_map_base_init::tree_to_index):
11816 Change type to hash_table. Update dependent calls and types.
11818 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
11819 Change type to hash_table. Update dependent calls and types.
11821 * tree-ssa-phiopt.c (seen_ssa_names):
11822 Change type to hash_table. Update dependent calls and types.
11824 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
11825 Change type to hash_table. Update dependent calls and types.
11827 * tree-ssa-uncprop.c (equiv):
11828 Change type to hash_table. Update dependent calls and types.
11830 2013-04-25 Jakub Jelinek <jakub@redhat.com>
11832 PR rtl-optimization/57003
11833 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
11834 call note_stores with kill_clobbered_value callback again after
11835 killing regs_invalidated_by_call.
11837 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11839 * config/aarch64/aarch64-simd.md
11840 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
11841 (aarch64_simd_bsl<mode>): Likewise.
11842 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
11844 2013-04-25 Marek Polacek <polacek@redhat.com>
11846 PR tree-optimization/57066
11847 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
11849 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
11851 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
11853 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11855 * config/aarch64/aarch64-builtins.c
11856 (aarch64_fold_builtin): New.
11857 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
11858 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
11859 * config/aarch64/aarch64-simd-builtins.def (abs): New.
11860 * config/aarch64/arm_neon.h
11861 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
11863 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11864 Tejas Belagod <tejas.belagod@arm.com>
11866 * config/aarch64/aarch64-builtins.c
11867 (aarch64_gimple_fold_builtin): New.
11868 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
11869 * config/aarch64/aarch64-simd-builtins.def (addv): New.
11870 * config/aarch64/aarch64-simd.md (addpv4sf): New.
11871 (addvv4sf): Update.
11872 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
11874 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11876 * config/aarch64/aarch64.md
11877 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
11879 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11881 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
11882 (*ngcsi_uxtw): New pattern.
11884 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11885 Julian Brown <julian@codesourcery.com>
11887 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
11888 (TB_DREG): Add T_V4HF.
11889 (v4hf_UP): New macro.
11890 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11891 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11892 Handle initialisation of V4HF. Adjust initialisation of reinterpret
11894 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11895 (arm_vector_mode_supported_p): Handle V4HF.
11896 (arm_mangle_map): Handle V4HFmode.
11897 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11898 * config/arm/arm_neon_builtins.def: Add entries for
11899 vcvtv4hfv4sf, vcvtv4sfv4hf.
11900 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11901 (neon_vcvtv4hfv4sf): Likewise.
11902 * config/arm/neon-gen.ml: Handle half-precision floating point
11904 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11905 * config/arm/arm_neon.h: Regenerate.
11906 * config/arm/neon.ml (type elts): Add F16.
11907 (type vectype): Add T_float16x4, T_floatHF.
11908 (type vecmode): Add V4HF.
11909 (type features): Add Requires_FP_bit feature.
11910 (elt_width): Handle F16.
11911 (elt_class): Likewise.
11912 (elt_of_class_width): Likewise.
11913 (mode_of_elt): Refactor.
11914 (type_for_elt): Handle F16, fix error messages.
11915 (vectype_size): Handle T_float16x4.
11916 (vcvt_sh): New function.
11917 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11918 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11919 (string_of_mode): Handle V4HF.
11920 * doc/arm-neon-intrinsics.texi: Regenerate.
11922 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11924 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11925 format specifier in 'X' case.
11927 2013-04-25 Alan Modra <amodra@gmail.com>
11930 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11931 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11932 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11933 Repeat for many other rotate/shift and mask patterns using subregs.
11934 Name lshiftrt insns.
11935 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11936 on WORDS_BIG_ENDIAN.
11938 2013-04-25 Alan Modra <amodra@gmail.com>
11940 * config.gcc: Support little-endian powerpc-linux targets.
11941 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11942 (LINK_OS_LINUX_SPEC): Define.
11943 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11944 Preserve MASK_LITTLE_ENDIAN.
11945 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11946 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11947 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11948 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11949 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11950 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11951 Correct fp word order for little-endian. Don't shift toc entries
11952 smaller than a word for little-endian.
11953 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11954 (bswapdi2 splits): Correct low-part subreg for little-endian.
11955 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11956 low/high where such is correct only for be.
11957 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11958 little-endian for -mcall-aixdesc.
11960 2013-04-25 Alan Modra <amodra@gmail.com>
11962 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11963 replace_equiv_address_nv.
11965 2013-04-25 Alan Modra <amodra@gmail.com>
11967 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11969 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11972 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11973 * rtl.h (struct rtx_def): ...
11975 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11977 PR rtl-optimizations/57046
11978 * lra-constraints (split_reg): Set up lra_risky_transformations_p
11979 for multi-reg splits.
11981 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
11983 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
11985 2013-04-24 Sterling Augustine <saugustine@google.com>
11987 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
11988 (comp_dir_string, debug_str_dwo_section): New.
11989 (DEBUG_STR_DWO_SECTION): Rename to ...
11990 (DEBUG_DWO_STR_SECTION): ... this.
11991 (DEBUG_NORM_STR_SECTION): Delete.
11992 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
11993 (DEBUG_STR_DWO_SECTION_FLAGS): New.
11994 (find_AT_string): Move most logic to ...
11995 (find_AT_string_in_table): ... here. New.
11996 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
11997 add_skeleton_AT_string. Delete logic.
11998 (output_skeleton_debug_sections): Remove call to
11999 add_top_level_skeleton_die_attrs.
12000 (add_comp_dir_attribute): Move logic to comp_dir_string.
12001 (dwarf2out_init): Initialize debug_str_dwo_section.
12002 (output_indirect_string): Call find_string_form.
12003 (output_indirect_strings): Rewrite.
12004 (prune_unused_types): Empty skeleton_debug_str_hash.
12005 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
12006 (dwarf2out_finish): Call output_indirect_strings.
12008 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12010 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
12012 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12014 * rtl.h (struct rtx_def): Add comment for field jump.
12015 (LRA_SUBREG_P): New macro.
12016 * recog.c (register_operand): Check LRA_SUBREG_P.
12017 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12019 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12020 (lra_final_code_change): Skip subreg change for operators.
12021 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12022 if there are no operand changes.
12023 * lra-constraints.c (curr_insn_set): New.
12024 (match_reload): Set LRA_SUBREG_P.
12025 (emit_spill_move): Ditto.
12026 (check_and_process_move): Use curr_insn_set. Process only single
12027 set insns. Don't initialize sec_mem_p and change_p.
12028 (simplify_operand_subreg): Use LRA_SUBREG_P.
12029 (reg_in_class_p): New function.
12030 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12031 of #ifdef. Add code to remove cycling.
12032 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12033 non-null disp. Reload inner instead of disp when base and index
12034 are null. Try to put lo_sum into register.
12035 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12036 (check_and_process_move): Move code for move cost check to
12037 simple_move_p. Remove equiv_substitution.
12038 (simple_move_p): New function.
12039 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12040 curr_insn_set. Call check_and_process_move only for single set
12041 insns. Use the new function. Move call of check_and_process_move
12042 after operand equiv substitution and address process.
12044 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
12046 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
12047 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
12048 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
12050 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12052 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
12054 2013-04-24 Marek Polacek <polacek@redhat.com>
12056 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
12057 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
12058 (select_loops_exit_conditions): Likewise.
12059 (number_of_iterations_for_all_loops): Likewise.
12060 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
12061 (scev_analysis): Likewise.
12063 2013-04-02 Catherine Moore <clm@codesourcery.com>
12064 Chao-ying Fu <fu@mips.com>
12066 * config/mips/micromips.md (jraddiusp): New pattern.
12067 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
12068 instruction if possible.
12070 2013-04-24 Alan Modra <amodra@gmail.com>
12072 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
12074 2013-04-24 Julian Brown <julian@codesourcery.com>
12075 Chung-Lin Tang <cltang@codesourcery.com>
12077 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
12078 dependency behavior in enumeration type DIE generation. Add TODO note
12079 to comments about future DW_FORM_sdata/udata re-work of related code.
12081 2013-04-23 Lawrence Crowl <crowl@google.com>
12083 * Makefile.in: Update as needed below.
12085 * hash-table.h (class hash_table):
12086 Correct many methods with parameter types compare_type to the correct
12087 value_type. (Correct code was unlikely to notice the change.)
12088 (hash_table::elements_with_deleted) New.
12089 (class hashtable::iterator): New.
12090 (hashtable::begin()): New.
12091 (hashtable::end()): New.
12092 (FOR_EACH_HASH_TABLE_ELEMENT): New.
12094 * statistics.c (statistics_hashes):
12095 Change type to hash_table. Update dependent calls and types.
12097 * tree-into-ssa.c (var_infos):
12098 Change type to hash_table. Update dependent calls and types.
12100 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
12101 Change type to hash_table. Update dependent calls and types.
12103 * tree-ssa-loop-im.c (struct mem_ref.refs):
12104 Change type to hash_table. Update dependent calls and types.
12106 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
12107 Change type to hash_table. Update dependent calls and types.
12109 * tree-ssa-sccvn.c (vn_tables_s::nary):
12110 Change type to hash_table. Update dependent calls and types.
12111 (vn_tables_s::phis): Likewise.
12112 (vn_tables_s::references): Likewise.
12114 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
12115 (vn_reference_eq): Update parameter and return types.
12117 * tree-ssa-structalias.c (pointer_equiv_class_table):
12118 Change type to hash_table. Update dependent calls and types.
12119 (location_equiv_class_table): Likewise.
12121 * tree-vect-data-refs.c: Consequential changes for making
12122 peeling a hash_table.
12124 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
12125 (destroy_loop_vec_info): Dependent hash_table update.
12127 * tree-vectorizer.h (peeling_htab):
12128 Change type to hash_table. Update dependent calls and types.
12130 2013-04-23 Shiva Chen <shiva0217@gmail.com>
12132 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
12133 to check the register content is equal or not.
12134 * lra-constraints.c (match_reload): Use lra_assign_reg_val
12135 to assign register content record.
12136 * lra-eliminations.c (update_reg_eliminate): Use
12137 lra_update_reg_val_offset to update register content offset.
12138 * lra-int.h (struct lra_reg): Add offset member.
12139 (lra_reg_val_equal_p): New static inline function.
12140 (lra_update_reg_val_offset): New static inline function.
12141 (lra_assign_reg_val): New static inline function.
12142 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
12143 to assign register content record.
12144 (initialize_lra_reg_info_element): Initial offset to zero.
12146 2013-04-23 Catherine Moore <clm@codesourcery.com>
12148 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
12149 operands. Record compression.
12151 2013-04-23 Xinliang David Li <davidxl@google.com>
12153 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
12155 2013-04-23 Richard Biener <rguenther@suse.de>
12157 PR middle-end/57036
12158 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
12159 parameter, only add abnormal goto edges from the copied body
12160 if the call could perform abnormal gotos.
12161 (copy_cfg_body): Adjust.
12163 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
12165 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
12167 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
12169 * coretypes.h (gimple_stmt_iterator): Add struct to make
12172 2013-04-23 Richard Biener <rguenther@suse.de>
12174 PR tree-optimization/57026
12175 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
12176 from SSA names occuring in abnormal PHI nodes.
12178 2013-04-22 Andi Kleen <ak@linux.intel.com>
12180 * lto/lto.c (print_lto_report_1): Fix LTO report names.
12182 2013-04-22 Andi Kleen <ak@linux.intel.com>
12184 * lto/lto.c (print_lto_report_1): Declare early.
12185 (read_cgraph_and_symbols): Call print_lto_report_1 early.
12187 2013-04-22 Andi Kleen <ak@linux.intel.com>
12189 * common.opt (-flto-report-wpa): Add.
12190 * doc/invoke.texi (-flto-report-wpa): Add.
12191 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
12194 2013-04-22 Xinliang David Li <davidxl@google.com>
12196 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
12197 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
12198 * Makefile.in: New dependency
12200 David Daney <ddaney.cavm@gmail.com>
12202 * configure.ac (gcc_cv_as_micromips_support): Use the
12203 --fatal-warnings option.
12204 * configure: Regenerate.
12206 2013-04-22 Marek Polacek <polacek@redhat.com>
12209 * tsan.c (instrument_expr): Don't instrument expression
12210 in case its size is zero.
12212 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
12216 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
12218 * config/alpha/alpha.c (TARGET_LRA_P): New define.
12220 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12222 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
12223 (gimple_stmt_iterator): New typedef.
12224 * gimple.h (gimple_stmt_iterator): Rename to...
12225 (gimple_stmt_iterator_d): ... This.
12226 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
12227 trees be valid for GIMPLE and GENERIC.
12228 (TARGET_GIMPLE_FOLD_BUILTIN): New.
12229 * gimple-fold.c (gimple_fold_call): Call target hook
12230 gimple_fold_builtin.
12231 * hooks.c (hook_bool_gsiptr_false): New.
12232 * hooks.h (hook_bool_gsiptr_false): New.
12233 * target.def (fold_stmt): New.
12234 * doc/tm.texi: Regenerate.
12236 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
12239 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
12240 a set sp if no stack realignment.
12242 2013-04-22 Nick Clifton <nickc@redhat.com>
12244 * config.gcc (tilegx-linux): Extend extra_objs rather than
12246 (tilepro-linux): Likewise.
12248 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12250 * config/aarch64/aarch64-builtins.c
12252 (CF0, CF1, CF2, CF3, CF4, CF10): New.
12253 (VAR<1-12>): Add MAP parameter.
12254 (BUILTIN_*): Likewise.
12255 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
12256 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
12257 (aarch64_ushl_n<mode>): Likewise.
12258 (aarch64_sshr_n<mode>): Likewise.
12259 (aarch64_ushr_n<mode>): Likewise.
12260 (aarch64_<maxmin><mode>): Likewise.
12261 (aarch64_sqrt<mode>): Likewise.
12262 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
12263 (vshr<q>_n_*): Likewise.
12265 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12267 * config/aarch64/aarch64-builtins.c
12268 (aarch64_simd_builtin_type_mode): Handle SF types.
12270 (BUILTIN_GPF): Define.
12271 (aarch64_init_simd_builtins): Handle SF types.
12272 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
12273 (frecps): Likewise.
12274 (frecpx): Likewise.
12275 * config/aarch64/aarch64-simd.md
12276 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
12277 (aarch64_frecpe<mode>): New.
12278 (aarch64_frecps<mode>): Likewise.
12279 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
12280 (v8type): Add frecp<esx>.
12281 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
12282 (aarch64_frecps<mode>): Likewise.
12283 * config/aarch64/iterators.md (FRECP): New.
12284 (frecp_suffix): Likewise.
12285 * config/aarch64/arm_neon.h
12286 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
12288 2013-04-22 Christian Bruel <christian.bruel@st.com>
12291 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
12292 (REG_CLASS_NAMES): Idem.
12293 (REG_CLASS_CONTENTS): Idem.
12294 (REGCLASS_HAS_FP_REG): Idem.
12295 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
12296 (sh_conditional_register_usage): Idem.
12298 2013-04-21 Jeff Law <law@redhat.com>
12300 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
12301 (ssa_forward_propagate_and_combine): Use it.
12303 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12305 * lra.c: Update the flow chart diagram.
12307 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12309 PR rtl-optimization/56847
12310 * lra-constraints.c (process_alt_operands): Discourage alternative
12311 with non-matche doffsettable memory constraint fro memory with
12314 2013-04-19 Richard Biener <rguenther@suse.de>
12316 PR tree-optimization/56982
12317 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
12319 * gimplify.c (gimplify_call_expr): Notice special calls.
12320 (gimplify_modify_expr): Likewise.
12321 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
12322 abnormal control flow receivers.
12323 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
12324 in the same way as cfun->has_nonlocal_labels.
12325 (gimple_purge_dead_abnormal_call_edges): Likewise.
12326 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
12327 receivers start a basic-block.
12329 2013-04-19 Richard Biener <rguenther@suse.de>
12331 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
12333 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
12334 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
12335 (SLP_TREE_LOAD_PERMUTATION): Add.
12336 (vect_transform_slp_perm_load): Adjust prototype.
12337 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
12338 (vect_free_slp_instance): Likewise.
12339 (vect_create_new_slp_node): Likewise.
12340 (vect_supported_slp_permutation_p): Remove.
12341 (vect_slp_rearrange_stmts): Adjust.
12342 (vect_supported_load_permutation_p): Likewise. Inline
12343 vect_supported_slp_permutation_p here.
12344 (vect_analyze_slp_instance): Compute load permutations per
12345 slp node instead of per instance.
12346 (vect_get_slp_defs): Adjust.
12347 (vect_transform_slp_perm_load): Likewise.
12348 (vect_schedule_slp_instance): Remove redundant code.
12349 (vect_schedule_slp): Remove hack for PR56270, add it ...
12350 * tree-vect-stmts.c (vectorizable_load): ... here, do not
12351 CSE loads for SLP. Adjust.
12353 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
12355 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
12356 spelling in two comments.
12358 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
12361 * config/arm/arm.c (load_multiple_sequence): Require SP
12362 as base register for loads if SP is in the register list.
12364 2013-04-19 Martin Jambor <mjambor@suse.cz>
12366 PR tree-optimization/56718
12367 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
12368 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
12369 and made public. Adjusted all callers.
12370 (ipa_intraprocedural_devirtualization): New function.
12371 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
12372 (ipa_intraprocedural_devirtualization): Likewise.
12373 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
12375 2013-04-19 Richard Biener <rguenther@suse.de>
12377 PR tree-optimization/57000
12378 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
12380 2013-04-19 Terry Guo <terry.guo@arm.com>
12382 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
12384 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
12385 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
12386 (cortex_m4_fmacs): Use new reservations.
12387 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
12389 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
12391 PR rtl-optimization/56999
12392 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
12394 (lra_coalesce): Remove split_origin_bitmap and related code.
12395 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
12396 ranges if necessary.
12398 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
12400 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
12402 (ix86_expand_call): Remove clobbered_registers array and use
12403 x86_64_ms_sysv_extra_clobbered_registers instead.
12404 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
12406 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
12408 * config/i386/i386.md (*call_rex64_ms_sysv): Use
12409 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
12410 (*call_value_rex64_ms_sysv): Ditto.
12412 2013-04-18 Cary Coutant <ccoutant@google.com>
12414 * dwarf2out.c (output_pubnames): Check die_perennial_p of
12415 parent instead of die_mark.
12417 2013-04-18 Diego Novillo <dnovillo@google.com>
12419 * gimple.c (create_gimple_tmp): New.
12420 (get_expr_type): New.
12421 (build_assign): New.
12422 (build_type_cast): New.
12423 * gimple.h (enum ssa_mode): Define.
12424 (gimple_seq_set_location): New.
12425 * asan.c (build_check_stmt): Change some gimple_build_* calls
12426 to use build_assign and build_type_cast.
12428 2013-04-18 Richard Biener <rguenther@suse.de>
12430 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
12431 handle negative step. Remove redundant checks.
12432 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
12433 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
12434 for negative step and grouped loads fail to vectorize.
12436 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
12438 * emit-rtl.c (reset_insn_used_flags): New function.
12439 (reset_all_used_flags): Use it.
12440 (verify_insn_sharing): New function.
12441 (verify_rtl_sharing): Fix verification for SEQUENCEs.
12443 2013-04-18 Jakub Jelinek <jakub@redhat.com>
12445 PR tree-optimization/56984
12446 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
12447 and (x >> M) >= N don't register any assertion if N << M is the
12450 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
12452 * lower-subreg.c (resolve_simple_move): If called self-recursive,
12453 do not delete_insn insns that have not yet been emitted, only
12454 unlink them with remove_insn.
12455 * df-scan.c (df_insn_delete): Revert r197492.
12457 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
12459 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
12460 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
12462 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
12464 * config/arm/arm.md (movsicc_insn): Convert define_insn into
12465 define_insn_and_split.
12466 (and_scc,ior_scc,negscc): Likewise.
12467 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
12469 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
12471 * config/arm/arm.c (use_return_insn): Return 0 for targets that
12472 can benefit from using a sequence of LDRD instructions in epilogue
12473 instead of a single LDM instruction.
12475 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
12478 * doc/extend.texi: Fix typo.
12480 2013-04-17 Richard Biener <rguenther@suse.de>
12482 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
12483 (vect_build_slp_tree): ... here.
12484 (vect_build_slp_tree_1): Compute which stmts of the SLP group
12485 match. Remove special-casing of mismatched complex loads.
12486 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
12487 re-try the match with swapped commutative operands.
12488 (vect_supported_load_permutation_p): Remove special-casing of
12489 mismatched complex loads.
12490 (vect_analyze_slp_instance): Adjust.
12492 2013-04-17 Richard Biener <rguenther@suse.de>
12494 PR rtl-optimization/56921
12495 * cfgloop.h (struct loop): Add simple_loop_desc member.
12496 (struct niter_desc): Mark with GTY(()).
12497 (simple_loop_desc): Do not use aux field but simple_loop_desc.
12498 * loop-iv.c (get_simple_loop_desc): Likewise.
12499 (free_simple_loop_desc): Likewise.
12502 2013-04-16 Richard Biener <rguenther@suse.de>
12504 PR rtl-optimization/56921
12505 * loop-init.c (pass_rtl_move_loop_invariants): Add
12506 TODO_do_not_ggc_collect to todo_flags_finish.
12507 (pass_rtl_unswitch): Same.
12508 (pass_rtl_unroll_and_peel_loops): Same.
12509 (pass_rtl_doloop): Same.
12511 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
12513 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
12514 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
12515 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
12517 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
12518 * tree-streamer.c (record_common_node): Adjust reference in comment.
12520 2013-04-17 Terry Guo <terry.guo@arm.com>
12522 * config/arm/cortex-m4.md: Add a new bypass.
12524 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12526 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
12528 (*subs_<optab><mode>_multp2): New pattern.
12529 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
12530 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
12532 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12534 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
12535 (*subs_mul_imm_<mode>): New pattern.
12537 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
12540 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
12541 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
12542 (vsx_movti_32bit): Change j->wa to O->wa.
12544 2013-04-16 Richard Biener <rguenther@suse.de>
12546 PR rtl-optimization/56921
12547 * loop-init.c (pass_rtl_move_loop_invariants): Add
12548 TODO_do_not_ggc_collect to todo_flags_finish.
12549 (pass_rtl_unswitch): Same.
12550 (pass_rtl_unroll_and_peel_loops): Same.
12551 (pass_rtl_doloop): Same.
12553 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
12555 * config/arm/arm.c (emit_multi_reg_push): New declaration
12556 for an existing function.
12557 (arm_emit_strd_push): New function.
12558 (arm_expand_prologue): Used here.
12559 (arm_emit_ldrd_pop): New function.
12560 (arm_expand_epilogue): Used here.
12561 (arm_get_frame_offsets): Update condition.
12562 (arm_emit_multi_reg_pop): Add a special case for load of a single
12563 register with writeback.
12565 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
12567 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
12570 2013-04-16 Richard Biener <rguenther@suse.de>
12572 PR tree-optimization/56756
12573 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
12574 (first_mem_ref_loc): New.
12575 (execute_sm): Place the load temporarily before a previous
12576 access instead of in the latch edge to ensure its SSA dependencies
12577 are defined at points dominating the load.
12579 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
12581 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
12582 correct fix by moving header and footer insn to the footer of
12583 the merged basic block. Clear BB_END of the merged-away block.
12585 PR middle-end/43631
12586 * emit-rtl.c (make_note_raw): New function.
12587 (link_insn_into_chain): New static inline function.
12588 (add_insn): Use it.
12589 (add_insn_before, add_insn_after): Factor insn chain linking code...
12590 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
12591 using link_insn_into_chain.
12592 (note_outside_basic_block_p): New helper function for emit_note_after
12593 and emit_note_before.
12594 (emit_note_after): Use nobb variant of add_insn_after if the note
12595 should not be contained in a basic block.
12596 (emit_note_before): Use nobb variant of add_insn_before if the note
12597 should not be contained in a basic block.
12598 (emit_note_copy): Use make_note_raw.
12599 (emit_note): Likewise.
12600 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
12601 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
12602 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
12603 the moved barrier the tail of the basic block it follows.
12604 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
12606 2013-04-15 Jakub Jelinek <jakub@redhat.com>
12608 PR tree-optimization/56962
12609 * gimple-ssa-strength-reduction.c (record_increment): Only set
12610 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
12611 either rhs1 or rhs2 is equal to c->base_expr.
12613 2013-04-15 Richard Biener <rguenther@suse.de>
12615 PR tree-optimization/56933
12616 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
12618 (GROUP_READ_WRITE_DEPENDENCE): Remove.
12619 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
12620 * tree-vect-data-refs.c (vect_analyze_group_access): Move
12621 dependence check ...
12622 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
12624 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
12625 GROUP_READ_WRITE_DEPENDENCE.
12627 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12629 * emit-rtl.c (reset_all_used_flags): New function.
12630 (verify_rtl_sharing): Call reset_all_used_flags before and after
12631 performing the checks.
12633 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12635 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
12636 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
12637 * config/arm/constraints.md (De): New constraint.
12638 * config/arm/neon.md (anddi3_neon): Delete.
12639 (neon_vand<mode>): Expand to standard anddi3 pattern.
12640 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
12641 Move earlier in the file.
12642 (neon_inv_logic_op2): Likewise.
12643 (arm_anddi_operand_neon): New predicate.
12645 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12647 * configure.ac (gcc_cv_ld_as_needed): Set
12648 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
12649 Use -z ignore, -z record on *-*-solaris2*.
12650 (HAVE_LD_AS_NEEDED): Update comment.
12651 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
12652 * configure: Regenerate.
12653 * config.in: Regenerate.
12654 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
12655 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
12656 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
12657 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
12658 equivalents. Fix markup.
12659 * doc/tm.texi: Regenerate.
12661 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
12663 * config/i386/i386.opt: New option mstack-protector-guard=.
12664 * config/i386/i386-opts.h: Add enum stack_protector_guard.
12665 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
12666 TARGET_SSP_TLS_GUARD.
12667 * config/i386/i386.c (ix86_option_override_internal): Set
12668 ix86_stack_protector_guard.
12669 * config/i386/i386.md (stack_protect_set): Enable for
12670 TARGET_SSP_TLS_GUARD only.
12671 (stack_protect_set_<mode>): Ditto.
12672 (stack_protect_test): Ditto.
12673 (stack_protect_test_<mode>): Ditto.
12674 * doc/invoke.texi (i386 Option): Document.
12676 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
12679 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
12680 (S_MODES): Set H_MODE bit.
12681 (SF_MODES): Set only S_MODE and SF_MODE bits.
12682 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
12683 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
12684 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
12685 <MODE_FLOAT>: Likewise.
12687 2013-04-15 Joey Ye <joey.ye@arm.com>
12689 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
12691 2013-04-15 Joey Ye <joey.ye@arm.com>
12693 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
12695 (thumb_far_jump_used_p): Count instruction size and set
12698 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
12700 * reorg.c (fill_simple_delay_slots): Reindent block of code.
12701 * resource.c (mark_target_live_regs): Reformat conditional block.
12703 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
12705 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
12706 notes, they are emitted only just before final.
12707 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
12709 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
12711 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
12712 * cfgrtl.c (delete_insn): Call it here instead.
12713 * lra-spills.c (lra_final_code_change): Use delete_insn.
12714 * haifa-sched.c (sched_remove_insn): Likewise.
12715 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
12716 returning to the nop pool.
12717 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
12718 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
12720 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
12722 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
12723 * doc/tm.texi: Regenerated.
12725 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
12727 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
12730 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
12732 * df-core.c (df_find_def): Compare register numbers.
12733 (df_find_use): Likewise.
12735 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
12738 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
12739 lra_in_progress for return.
12741 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
12743 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
12744 define_insn into define_insn_and_split and emit movsicc patterns.
12746 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
12748 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
12750 2013-04-12 Richard Biener <rguenther@suse.de>
12752 * tree-pass.h (TODO_do_not_ggc_collect): New.
12753 * passes.c (execute_one_ipa_transform_pass): Honor
12754 TODO_do_not_ggc_collect.
12755 (execute_one_pass): Likewise.
12758 2013-04-10 Richard Biener <rguenther@suse.de>
12760 * passes.c (init_optimization_passes): Remove reload pass.
12761 * ira.c (do_reload): Merge into ...
12763 (rest_of_handle_reload): Remove.
12764 (pass_reload): Likewise.
12765 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12766 of reload for vzeroupper pass placement.
12768 2013-04-12 Jakub Jelinek <jakub@redhat.com>
12770 PR tree-optimization/56918
12771 PR tree-optimization/56920
12772 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
12773 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
12774 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
12775 use wide_mul_with_sign method.
12777 2013-04-12 Richard Biener <rguenther@suse.de>
12779 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
12780 not be considered a gimple constant.
12782 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
12784 * fold-const.c (const_binop): Handle vector shifts by a scalar.
12785 (fold_binary_loc): Call const_binop also for mixed vector-scalar
12788 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
12789 Jakub Jelinek <jakub@redhat.com>
12791 * opts.c: Include diagnostic-color.h.
12792 (common_handle_option): Handle OPT_fdiagnostics_color_.
12793 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
12794 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
12795 (diagnostic-color.o): New.
12796 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
12797 (diagnostic_color_rule): New enum.
12798 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
12799 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
12800 the location string.
12801 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
12802 either NULL, or color kind.
12803 * diagnostic-color.c: New file.
12804 * diagnostic-color.h: New file.
12805 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
12807 * doc/invoke.texi (-fdiagnostics-color): Document.
12808 * pretty-print.h (pp_show_color): Define.
12809 (struct pretty_print_info): Add show_color field.
12810 * diagnostic.c: Include diagnostic-color.h.
12811 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
12812 macros. Colorize error:, warning: etc. strings and also the location
12814 (diagnostic_show_locus): Colorize the caret line.
12815 * pretty-print.c: Include diagnostic-color.h.
12816 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
12817 inside of %< %> quotes or quoted through q format modifier.
12819 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12821 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
12823 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12825 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
12826 code in CC_NZ mode.
12827 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
12830 2013-04-11 Marek Polacek <polacek@redhat.com>
12832 PR tree-optimization/48184
12833 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
12835 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
12837 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
12838 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
12839 (skip_simple_arithmetic): Tidy up.
12840 * tree.h (skip_simple_constant_arithmetic): Declare.
12842 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12844 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
12846 2013-04-11 Richard Biener <rguenther@suse.de>
12848 * tree-vect-loop.c (get_initial_def_for_induction): Properly
12849 generate vector constants.
12851 2013-04-11 Richard Biener <rguenther@suse.de>
12853 PR tree-optimization/56878
12854 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
12855 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
12857 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
12858 Prefer to align the DR with the most invariant base address.
12860 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12862 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
12865 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
12867 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
12868 floating-point vector comparisons against 0.
12870 2013-04-11 Jakub Jelinek <jakub@redhat.com>
12872 PR tree-optimization/56899
12873 * fold-const.c (extract_muldiv_1): Apply distributive law
12874 only if TYPE_OVERFLOW_WRAPS (ctype).
12876 2013-04-11 Bin Cheng <bin.cheng@arm.com>
12879 * ira-costs.c (scan_one_insn): Check whether the source rtx of
12880 loading has side effect.
12882 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
12884 * config/sparc/sparc.c: Include tree-pass.h.
12885 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
12886 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
12887 head of file. Change return type. Split off gate function.
12888 (sparc_gate_work_around_errata): New function.
12889 (pass_work_around_errata): New pass definition.
12890 (insert_pass_work_around_errata) New pass insert definition to
12891 insert pass_work_around_errata just after delayed-branch scheduling.
12892 (sparc_option_override): Insert the pass.
12893 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
12895 2013-04-10 David S. Miller <davem@davemloft.net>
12897 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12898 or -mcpu=hypersparc.
12900 * target.def (cstore_mode): New hook.
12901 * target.h: Include insn-codes.h
12902 * targhooks.c: Likewise.
12903 (default_cstore_mode): New function.
12904 * targhooks.h: Declare it.
12905 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12906 * doc/tm.texi: Rebuild.
12907 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12908 target hook, rather than inspecting the insn_data.
12909 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12910 (TARGET_CSTORE_MODE): Redefine.
12911 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12913 * config/sparc/predicates.md (cstore_result_operand): New special
12915 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12916 Use it for operand 0.
12917 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12918 (*snesi_special): Likewise.
12919 (*snesi_zero): Likewise.
12920 (*seqsi_zero): Likewise.
12921 (*sltu_insn): Likewise.
12922 (*sgeu_insn): Likewise.
12923 (*seqdi_special): Make operand 0 and comparison operation be of
12925 (*snedi_special): Likewise.
12926 (*snedi_special_vis3): Likewise.
12927 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12928 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12929 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12930 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12931 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12932 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12933 (*sltu_extend_sp64): Likewise.
12934 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12935 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12936 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12937 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12938 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12940 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12942 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12943 (aarch64_start_file): Use the new function.
12945 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12946 Jason Merrill <jason@redhat.com>
12948 * common.opt: Add -gdwarf.
12949 * opts.c (common_handle_option): Handle it.
12950 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12952 2013-04-10 Richard Biener <rguenther@suse.de>
12954 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12955 (execute_one_ipa_transform_pass): But unconditionally here.
12956 (execute_one_pass): And here.
12957 (init_optimization_passes): Remove reload pass.
12958 * tree-pass.h (TODO_ggc_collect): Remove.
12959 (pass_reload): Likewise.
12960 * ira.c (do_reload): Merge into ...
12962 (rest_of_handle_reload): Remove.
12963 (pass_reload): Likewise.
12964 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12965 of reload for vzeroupper pass placement.
12966 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12967 and todo_flags_finish of all passes.
12969 2013-04-10 Richard Biener <rguenther@suse.de>
12971 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12972 first_const_oprnd field, rename first_def_type to first_op_type.
12973 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
12974 (vect_get_and_check_slp_defs): Always use the type of the
12975 operand. Allow mixed vect_external_def, vect_constant_def types.
12976 (vect_get_constant_vectors): Handle mixed vect_external_def,
12977 vect_constant_def types.
12979 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
12981 PR tree-optimization/55524
12982 * tree-ssa-math-opts.c
12983 (convert_mult_to_fma): Don't use an fms construct
12984 when we don't have an fms operation, but fnma, and it looks
12985 likely that we'll be able to use the latter.
12987 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12989 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
12991 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
12992 inline fail caused by overwritable functions.
12994 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
12996 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
12997 unnecessary bits in the constant power of two case.
12999 2013-04-10 Richard Biener <rguenther@suse.de>
13001 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
13002 broken code swapping operands.
13003 (vect_build_slp_tree): Do not compute load permutations here.
13004 (vect_analyze_slp_instance): Compute load permutations here,
13005 after building the SLP tree.
13007 2013-04-09 Christian Bruel <christian.bruel@st.com>
13009 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
13010 of next/prev_real_insn.
13012 2013-04-09 Jan Hubicka <jh@suse.cz>
13014 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13015 Drop aliased parameter.
13016 (function_and_variable_visibility): Do not handle alias pairs.
13017 * cgraph.c (varpool_externally_visible_p): Update prototype.
13018 * varpool.c (varpool_add_new_variable): Update.
13020 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13022 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
13024 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
13026 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
13028 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
13030 2013-04-09 Marek Polacek <polacek@redhat.com>
13032 PR tree-optimization/48762
13033 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
13035 2013-04-09 Richard Biener <rguenther@suse.de>
13037 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
13039 (vect_build_slp_tree): Likewise.
13040 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
13041 calculating the cost of a SLP instance.
13042 (vect_analyze_slp_instance): Use it from here, after building
13045 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13047 PR middle-end/56883
13048 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
13049 expand_omp_for_static_chunk): Use simple_p = true in
13050 force_gimple_operand_gsi calls when assigning to addressable decls.
13052 2013-04-09 Jeff Law <law@redhat.com>
13054 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
13055 when the boolean was created by converting a wider object which
13056 had a boolean range.
13058 2013-04-09 Richard Biener <rguenther@suse.de>
13060 * tree-vectorizer.h (slp_void_p): Remove.
13061 (slp_tree): Typedef before _slp_tree declaration.
13062 (struct _slp_tree): Use a vector of slp_tree as children.
13063 (vect_get_place_in_interleaving_chain): Remove.
13064 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
13066 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
13068 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
13069 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
13070 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
13071 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
13072 Use slp_node instead of slp_void_p and adjust.
13074 2013-04-09 Richard Biener <rguenther@suse.de>
13076 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
13077 work that is not necessary.
13079 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13081 PR tree-optimization/56854
13082 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
13083 forward into clobber stmts if it would change MEM_REF lhs into
13086 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
13088 * tree.c (type_hash_lookup, type_hash_add): Make static.
13089 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
13091 2013-04-09 Richard Biener <rguenther@suse.de>
13093 * tree.h (unsave_expr_now): Remove.
13094 * tree-inline.c (mark_local_for_remap_r): Remove.
13095 (unsave_expr_1): Likewise.
13096 (unsave_r): Likewise.
13097 (unsave_expr_now): Likewise.
13098 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
13099 (propagate_tree_value): Likewise.
13101 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
13103 * doc/rtl.texi (sequence): Rewrite documentation to match the
13104 current use of SEQUENCE rtl objects.
13105 * rtl.def (SEQUENCE): Likewise.
13107 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
13108 Update documentation.
13109 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
13110 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
13112 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
13114 2013-04-08 Teresa Johnson <tejohnson@google.com>
13116 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
13117 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
13119 (estimate_edge_size_and_time): Add comment to suggest using rounding
13121 (estimate_node_size_and_time): Ditto.
13122 (remap_edge_change_prob): Use helper rounding divide methods.
13123 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
13124 (gimple_mod_pow2_value_transform): Ditto.
13125 (gimple_mod_subtract_transform): Ditto.
13126 (gimple_ic_transform): Ditto.
13127 (gimple_stringops_transform): Ditto.
13128 * stmt.c (conditional_probability): Ditto.
13129 (emit_case_dispatch_table): Ditto.
13130 * lto-cgraph.c (merge_profile_summaries): Ditto.
13131 * tree-optimize.c (execute_fixup_cfg): Ditto.
13132 * cfgcleanup.c (try_forward_edges): Ditto.
13133 * cfgloopmanip.c (scale_loop_profile): Ditto.
13135 (duplicate_loop_to_header_edge): Ditto.
13136 (lv_adjust_loop_entry_edge): Ditto.
13137 * tree-vect-loop.c (vect_transform_loop): Ditto.
13138 * profile.c (compute_branch_probabilities): Ditto.
13139 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
13140 * lto-streamer-in.c (input_cfg): Ditto.
13141 * gimple-streamer-in.c (input_bb): Ditto.
13142 * ipa-cp.c (update_profiling_info): Ditto.
13143 (update_specialized_profile): Ditto.
13144 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
13145 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
13147 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13148 (compute_trg_info): Ditto.
13149 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
13150 (purge_dead_edges): Ditto.
13151 * loop-unswitch.c (unswitch_loop): Ditto.
13152 * cgraphclones.c (cgraph_clone_edge): Ditto.
13153 (cgraph_clone_node): Ditto.
13154 * tree-inline.c (copy_bb): Ditto.
13155 (copy_edges_for_bb): Ditto.
13156 (initialize_cfun): Ditto.
13157 (copy_cfg_body): Ditto.
13158 (expand_call_inline): Ditto.
13160 2013-04-08 Kai Tietz <ktietz@redhat.com>
13162 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
13163 TARGET_CYGWIN64 by TARGET_64BIT.
13165 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13167 * config/epiphany/epiphany.md (GPR_1): New constant.
13168 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
13169 * config/epiphany/epiphany.c (gen_compare_reg):
13170 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
13171 is already in place.
13172 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
13173 Don't require being called during rtl expansion; If y operlaps r0,
13175 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
13176 (epiphany_expand_epilogue): Likewise.
13178 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
13179 Don't use CC_FPmode for ORDERED / UNORDERED.
13180 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
13182 * config/epiphany/constraints.md (CnL): New constraint.
13183 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
13184 * config/epiphany/predicates.md (add_operand): Allow 1024.
13186 * config/epiphany/epiphany.md (logical_op): New code iterator.
13187 (op_mnc): New code attribute.
13188 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
13189 (mov_f+1, mov_f+2): New peephole2 patterns.
13191 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
13192 (cstoresi4): Also allow re-use of zero result when doing a NE
13193 comparison to a non-zero operand.
13194 Use (clobber (scratch)) for first insn if the gpr output is not needed.
13196 * config/epiphany/epiphany.md (<insn_opname>v2si3):
13197 Use gen_addsi3_i / gen_subsi3_i.
13199 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13203 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
13204 contain anything but clobbers, at most one __builtin_stack_restore,
13205 optionally debug stmts and final resx, and if it has at least one
13206 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
13207 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
13208 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
13209 which isn't defaut definition, remove them.
13210 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
13211 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
13212 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
13213 with MEM_REF LHS with SSA_NAME address.
13215 2013-04-08 Jeff Law <law@redhat.com>
13217 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
13219 2013-04-08 Richard Biener <rguenther@suse.de>
13221 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
13223 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
13224 determined vector type.
13225 (vect_analyze_data_refs): Likewise.
13226 (vect_get_new_vect_var): Adjust.
13227 (vect_create_destination_var): Preserve SSA name versions.
13228 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
13229 not dump anything here.
13231 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13233 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
13234 Add member lr_slot_known.
13235 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
13237 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
13238 Remove code that sets lr_slot_offset according to what a previous
13239 version of epiphany_emit_save_restore used to do.
13240 (epiphany_emit_save_restore): When doing an lr save or restore,
13241 set/verify lr_slot_known and lr_slot_offset.
13243 2013-04-08 Xinyu Qi <xyqi@marvell.com>
13246 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
13249 2013-04-08 Richard Biener <rguenther@suse.de>
13251 * alias.c (find_base_term): Fix thinko in previous change.
13253 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13255 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
13256 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
13257 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
13258 if possible to compute val.
13259 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
13260 For QImode integers don't require anything about precision. Use
13261 const_with_all_bytes_same to find out if the constant doesn't have
13262 repeated bytes in it.
13264 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13266 * config/s390/s390.c (s390_expand_insv): Only accept insertions
13269 2013-04-08 Marek Polacek <polacek@redhat.com>
13271 PR rtl-optimization/48182
13272 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
13275 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13278 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
13279 nuses, make sure we have a label.
13281 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13284 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
13285 (rs6000_emit_swdiv_low_precision): Remove.
13286 (rs6000_emit_swdiv): Rewrite to handle between one and four
13287 iterations of Newton-Raphson generally; modify required number of
13288 iterations for some cases.
13289 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
13291 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
13293 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
13294 set-but-unused variable.
13296 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
13297 basic blocks of released function bodies garbage-collectable.
13299 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
13300 (struct rtl_opt_pass): Add TODO_df_finish.
13302 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
13304 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13306 * config/arm/constraints.md (q): New constraint.
13307 * config/arm/ldrdstrd.md: New file.
13308 * config/arm/arm.md (ldrdstrd.md) New include.
13309 (arm_movdi): Use "q" instead of "r" constraint
13310 for double-word memory access.
13311 (movdf_soft_insn): Likewise.
13312 * config/arm/vfp.md (movdi_vfp): Likewise.
13313 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
13314 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
13315 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
13316 (mem_ok_for_ldrd_strd): Likewise.
13317 (output_move_double): Update assertion.
13319 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13321 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
13323 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13325 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
13326 define_insn_and_split.
13327 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
13329 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13331 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
13332 define_insn_and_split.
13333 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
13334 (shiftsi3_compare): New pattern.
13335 (rrx): New pattern.
13336 * config/arm/unspecs.md (UNSPEC_RRX): New.
13338 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13340 * config/arm/arm.md (negdi_extendsidi): New pattern.
13341 (negdi_zero_extendsidi): Likewise.
13343 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13345 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
13346 define_insn_and_split.
13347 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
13348 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
13350 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13352 * config/arm/arm.md (arm_subdi3): Convert define_insn into
13353 define_insn_and_split.
13354 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
13355 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
13357 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13359 * config/arm/arm.md (subsi3_carryin): New pattern.
13360 (subsi3_carryin_const): Likewise.
13361 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
13362 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
13364 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13366 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
13368 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13370 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
13371 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
13373 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13375 * config/arm/arm.c (arm_expand_builtin): Change fcode
13376 type to unsigned int.
13378 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13380 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
13382 2013-04-04 Ian Lance Taylor <iant@google.com>
13384 * doc/standards.texi (Standards): The Go frontend supports the Go 1
13387 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
13389 PR middle-end/56729
13390 * df-scan.c (df_insn_delete): Disable failing assert.
13392 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13394 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
13395 New function prototype.
13396 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
13397 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
13398 (arm_builtin_vectorized_function): New function.
13400 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13402 * config/arm/arm_neon_builtins.def: New file.
13403 * config/arm/arm.c (neon_builtin_data): Move contents to
13404 arm_neon_builtins.def.
13405 (enum arm_builtins): Include neon builtin definitions.
13406 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
13407 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
13409 2013-04-04 Marek Polacek <polacek@redhat.com>
13411 PR tree-optimization/48186
13412 * predict.c (maybe_hot_frequency_p): Return false if
13413 HOT_BB_FREQUENCY_FRACTION is 0.
13414 (cgraph_maybe_hot_edge_p): Likewise.
13416 2013-04-04 Richard Biener <rguenther@suse.de>
13418 PR tree-optimization/56826
13419 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
13422 2013-04-04 Richard Biener <rguenther@suse.de>
13424 PR tree-optimization/56213
13425 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
13426 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
13428 2013-04-04 Richard Biener <rguenther@suse.de>
13430 PR tree-optimization/56837
13431 * tree-loop-distribution.c (classify_partition): For non-zero
13432 values require that the value has the same precision as its
13433 mode to be useful as memset value.
13435 2013-04-03 Nick Clifton <nickc@redhat.com>
13437 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
13438 (fmssf4): Use fmsf.s on E3V5 architectures.
13439 (fnmasf4): Use fnmaf.s on E3V5 architectures.
13440 (fnmssf4): Use fnmsf.s on E3V5 architectures.
13442 2013-04-03 Jeff Law <law@redhat.com>
13444 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
13445 (lra-eliminations.o): Likewise.
13447 2013-04-03 Teresa Johnson <tejohnson@google.com>
13449 * gcov-io.c (compute_working_sets): Moved most of body of old
13450 compute_working_sets here from profile.c.
13451 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
13452 (gcov_working_set_t): Moved typedef here from basic-block.h
13453 (compute_working_set): Declare.
13454 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
13455 (get_working_sets): Renamed from compute_working_set,
13456 replace most of body with call to new compute_working_sets.
13457 (get_exec_counts): Replace call to compute_working_sets
13458 to get_working_sets.
13459 * profile.h (get_working_sets): Renamed from compute_working_set.
13460 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
13461 to get_working_sets.
13462 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
13463 * gcov-dump.c (dump_working_sets): New function.
13465 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
13467 * hwint.c (sext_hwi, zext_hwi): New functions.
13468 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
13469 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
13470 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
13471 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
13472 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
13473 (sext_hwi, zext_hwi): New functions.
13475 2013-04-03 Jeff Law <law@redhat.com>
13477 PR tree-optimization/56799
13478 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
13479 back test for widening conversion erroneously dropped in prior change.
13481 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13484 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
13485 instead of next_real_insn.
13487 2013-04-03 Marek Polacek <polacek@redhat.com>
13490 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
13492 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13495 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
13497 (thumb1_output_casesi): Likewise.
13498 (thumb2_output_casesi): Likewise.
13500 2013-04-03 Richard Biener <rguenther@suse.de>
13502 PR tree-optimization/56817
13503 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
13505 (tree_unroll_loops_completely_1): ... new function to manually
13506 walk the loop tree, properly defering outer loops of unrolled
13507 loops to later iterations.
13509 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13511 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
13512 (vectorizable_load): Likewise.
13513 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
13514 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
13516 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13518 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
13521 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13523 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
13525 2013-04-03 Bin Cheng <bin.cheng@arm.com>
13527 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
13529 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13531 PR tree-optimization/56790
13532 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
13535 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13537 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
13539 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
13540 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
13543 2013-04-03 Jakub Jelinek <jakub@redhat.com>
13546 * tree.h (force_folding_builtin_constant_p): New decl.
13547 * builtins.c (force_folding_builtin_constant_p): New variable.
13548 (fold_builtin_constant_p): Fold immediately also if
13549 force_folding_builtin_constant_p.
13551 2013-04-03 Richard Biener <rguenther@suse.de>
13553 PR tree-optimization/56812
13554 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
13555 DRs of the same interleaving chain are independent.
13557 2013-04-02 Jason Merrill <jason@redhat.com>
13559 * gdbinit.in (pbb): Use debug fn.
13561 2013-04-02 Lawrence Crowl <crowl@google.com>
13563 * sese.h (struct ivtype_map_elt_s): Remove unused.
13564 (extern debug_ivtype_map): Remove unused.
13565 (extern eq_ivtype_map_elts): Remove unused.
13566 * sese.c (debug_ivtype_map): Removed unused.
13567 (debug_ivtype_map_1): Removed unused.
13568 (debug_ivtype_elt): Remove unused.
13569 (eq_ivtype_map_elts): Remove unused.
13572 2013-04-02 Kai Tietz <ktietz@redhat.com>
13575 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
13576 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
13577 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
13579 (legitimize_pe_coff_symbol): Likewise.
13580 (is_imported_p): New helper-function.
13581 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
13582 for Windows x64 targets.
13583 (ix86_expand_prologue): Optimize for pe-coff targets.
13584 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
13585 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
13586 medium/large code-model.
13587 (legitimize_pic_address): Likewise.
13588 (legitimize_tls_address): Likewise.
13589 (ix86_expand_call): Likewise.
13590 (x86_output_mi_thunk): Likewise.
13591 (get_dllimport_decl): Add new beimport argument.
13592 (construct_plt_address): Don't assert for x64 pe-coff targets.
13593 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
13595 (SYMBOL_FLAG_STUBVAR): New macro.
13596 (SYMBOL_REF_STUBVAR_P): Likewise.
13597 * config/i386/winnt.c (stub_list): New structure.
13598 (stub_head): New local variable.
13599 (i386_pe_record_stub): New function.
13600 (i386_pe_file_end): Emit refptr-stubs.
13602 2013-04-02 Jakub Jelinek <jakub@redhat.com>
13604 PR rtl-optimization/56745
13605 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
13606 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
13609 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
13610 and both of them are MEM_REFs, just compare first argument for
13611 equality and attempt to deal even with differing offsets.
13614 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
13615 of gimple_clobber_p to be MEM_REF.
13616 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
13617 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
13618 after gimplification.
13619 * asan.c (get_mem_ref_of_assignment): Don't instrument
13620 gimple_clobber_p stmts.
13621 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
13622 gimple_clobber_p stmt if they have MEM_REF lhs and
13623 are dead because of another gimple_clobber_p stmt.
13624 * tree-ssa-live.c (clear_unused_block_pointer): Treat
13625 gimple_clobber_p stmts like debug stmts.
13626 (remove_unused_locals): Remove clobbers with MEM_REF lhs
13627 that refer to unused VAR_DECLs or uninitialized values.
13628 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
13629 gimple_clobber_p stmts if they refer to removed parameters.
13630 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
13633 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
13635 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
13636 using SWI48 mode attribute.
13638 2013-04-02 Wei Mi <wmi@google.com>
13640 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
13641 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
13642 *<rotate_insn><mode>3_mask in i386.md.
13644 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
13646 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
13648 2013-04-02 Richard Biener <rguenther@suse.de>
13650 PR tree-optimization/56778
13651 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
13652 Runtime alias tests are not supported for gather loads.
13653 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
13654 stmts referenced from SSA operands before updating SSA form.
13656 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
13657 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13659 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
13660 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
13661 * config/arm/cortex-a53.md: New file.
13662 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
13663 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
13664 * config/arm/arm.c (arm_issue_rate): Likewise.
13665 * config/arm/arm-tune.md: Regenerate
13666 * config/arm/arm-tables.opt: Regenerate.
13667 * config/arm/arm-cores.def: Add cortex-a53.
13669 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
13671 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
13674 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
13676 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
13677 scalar load/store operations using B/H registers.
13678 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
13680 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
13682 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
13684 * config/aarch64/aarch64.c
13685 (aarch64_simd_scalar_immediate_valid_for_move): New.
13686 * config/aarch64/aarch64-protos.h
13687 (aarch64_simd_scalar_immediate_valid_for_move): New.
13688 * config/aarch64/constraints.md (Dh, Dq): New.
13689 * config/aarch64/iterators.md (hq): New.
13691 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
13693 * reorg.c (get_branch_condition): Deal with conditional returns.
13694 (fill_simple_delay_slots): Remove dead code dealing with jumps.
13696 2013-04-01 Wei Mi <wmi@google.com>
13698 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
13699 Truncate operand 2 using %b asm operand modifier.
13700 (*<shift_insn><mode>3_mask): Ditto.
13701 (*<rotate_insn><mode>3_mask): Ditto.
13703 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
13705 PR middle-end/56798
13706 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
13708 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
13710 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
13712 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
13714 2013-03-30 Lawrence Crowl <crowl@google.com>
13716 * dse.c (clear_alias_sets): Remove never set.
13717 (disqualified_clear_alias_sets): Remove never set.
13718 (clear_alias_mode_pool): Remove never set.
13719 (dse_step0): Remove condition that is never true.
13720 (canon_address): Remove condition that is never true.
13721 (dse_step7): Remove condition that is never true.
13722 (rest_of_handle_dse): Remove condition that is never true.
13723 (rest_of_handle_dse::did_global): Remove never read from above.
13724 (dse_step2_spill): Remove never called from above.
13725 (dse_step5_spill): Remove never called from above.
13727 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
13729 * doc/md.texi (Standard Names) <casesi>: Update documentation for
13730 JUMP_TABLE_DATA changes.
13731 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
13732 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
13733 (Insns) <jump_table_data>: New entry.
13734 * doc/tm.texi: Regenerate.
13736 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
13738 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
13739 for table jump at the end of a basic block using tablejump_p.
13740 * targhooks.c (default_invalid_within_doloop): Likewise.
13741 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
13742 target hook implementation that is identical to the default hook.
13743 (rs6000_invalid_within_doloop): Remove.
13745 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
13746 unused variable from tablejump_p call.
13748 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
13749 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
13750 (INSN_DELETED_P): Likewise.
13751 (emit_jump_table_data): New prototype.
13752 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
13753 after 4th as unused.
13754 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
13755 * sched-vis.c (print_insn): Likewise.
13756 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
13757 insn for compatibility with back ends that use next_active_insn to
13758 identify jump table data.
13759 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
13760 (remove_insn): Likewise.
13761 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
13763 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
13764 (emit_jump_table_data): New function.
13766 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
13767 basic block, a JUMP_TABLE_DATA never is.
13768 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
13769 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
13770 off from code handling real insns.
13771 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
13772 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
13773 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
13774 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
13775 is not a NONDEBUG_INSN_P.
13776 * ira-costs.c (scan_one_insn): Likewise.
13777 * jump.c (mark_all_labels): Likewise.
13778 (mark_jump_label_1): Likewise.
13779 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13780 * lra.c (get_insn_freq): Expect all insns reaching here to be in
13782 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
13783 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
13784 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
13785 JUMP_TABLE_DATA_P insns.
13786 (calculate_elim_costs_all_insns): Likewise.
13787 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
13788 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
13789 (delete_output_reload): Code style fixups.
13790 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
13791 insn flags on this non-insn.
13792 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
13793 as scheduling barriers, for pre-change compatibility.
13794 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
13795 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
13797 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
13798 redundant JUMP_TABLE_DATA_P test.
13799 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
13800 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13801 (frv_for_each_packet): Likewise.
13802 * config/i386/i386.c (min_insn_size): Likewise.
13803 (ix86_avoid_jump_mispredicts): Likewise.
13804 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13805 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
13806 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13807 (mips16_insn_length): Robustify.
13808 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
13809 (mips16_split_long_branches): Likewise.
13810 * config/pa/pa.c (pa_combine_instructions): Likewise.
13811 * config/rs6000/rs6000.c (get_next_active_insn): Treat
13812 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
13813 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
13814 as contributing to pool range lengths.
13815 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
13816 Remove redundant JUMP_TABLE_DATA_P test.
13817 (sh_loop_align): Likewise.
13818 (split_branches): Likewise.
13819 (sh_insn_length_adjustment): Likewise.
13820 * config/spu/spu.c (get_branch_target): Likewise.
13822 2013-03-29 Jan Hubicka <jh@suse.cz>
13824 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
13825 gcov streaming; stream hot bb threshold to ltrans.
13826 * predict.c (get_hot_bb_threshold): Break out from ....
13827 (maybe_hot_count_p): ... here.
13828 (set_hot_bb_threshold): New function.
13829 * lto-section-in.c (lto_section_name): Add profile.
13830 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
13831 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
13832 and data-streamer.h
13833 (histogram_entry): New structure.
13834 (histogram, histogram_pool): New global vars.
13835 (histogram_hash): New structure.
13836 (histogram_hash::hash): New method.
13837 (histogram_hash::equal): Likewise.
13838 (account_time_size): New function.
13839 (cmp_counts): New function.
13840 (dump_histogram): New function.
13841 (ipa_profile_generate_summary): New function.
13842 (ipa_profile_write_summary): New function.
13843 (ipa_profile_read_summary): New function.
13844 (ipa_profile): Decide on threshold.
13845 (pass_ipa_profile): Add ipa_profile_write_summary and
13846 ipa_profile_read_summary.
13847 * Makefile.in (ipa.o): Update dependencies.
13848 * lto-streamer.h (LTO_section_ipa_profile): New section.
13850 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
13852 * tree.h (VAR_P): New.
13854 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
13857 * doc/invoke.texi ([-fwhole-program]): Fix typo.
13859 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
13861 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
13862 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13863 (control_flow_insn_p): Likewise.
13864 * cfgrtl.c (duplicate_insn_chain): Likewise.
13865 * final.c (get_attr_length_1): Likewise.
13866 (shorten_branches): Likewise.
13867 (final_scan_insn): Likewise.
13868 * function.c (instantiate_virtual_regs): Likewise.
13869 * gcse.c (insert_insn_end_basic_block): Likewise.
13870 * ira-costs.c (scan_one_insn): Likewise.
13871 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13872 * lra.c (check_rtl): Likewise.
13873 * reload1.c (elimination_costs_in_insn): Likewise.
13874 * reorg.c (follow_jumps): Likewise.
13876 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
13877 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13878 (thumb_far_jump_used_p): Likewise.
13879 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
13880 (workaround_speculation): Likewise.
13881 (add_sched_insns_for_speculation): Likewise.
13882 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
13883 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13884 (frv_for_each_packet): Likewise.
13885 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
13886 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
13887 (final_emit_insn_group_barriers): Likewise.
13888 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13889 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13890 (mips16_insn_length): Likewise.
13891 * config/pa/pa.c (pa_reorg): Likewise.
13892 (pa_combine_instructions): Likewise.
13893 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
13894 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
13895 (sh_reorg): Likewise.
13896 (split_branches): Likewise.
13897 * config/spu/spu.c (get_branch_target): Likewise.
13899 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13902 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13904 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13905 Fix declaration name.
13907 2013-03-28 Lawrence Crowl <crowl@google.com>
13909 * graphds.h (struct graph.indicies): Remove unused.
13910 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13911 (SCOP_ORIGINAL_PDDRS): Remove unused.
13912 * sese.h (extern insert_loop_close_phis): Removed unused.
13913 (extern insert_guard_phis): Removed unused.
13914 (extern ivtype_map_elt_info): Removed unused.
13915 (new_ivtype_map_elt): Removed unused.
13916 * sese.c (ivtype_map_elt_info): Removed unused.
13918 2013-03-28 Lawrence Crowl <crowl@google.com>
13920 * Makefile.in: Add several missing include dependences.
13922 (test-dump.o): New. This object is not added to any executable,
13923 but is present for ad-hoc testing.
13925 (debug (const bitmap_head_def &)): New.
13926 (debug (const bitmap_head_def *)): New.
13928 (extern debug (const bitmap_head_def &)): New.
13929 (extern debug (const bitmap_head_def *)): New.
13931 (debug (edge_def &)): New.
13932 (debug (edge_def *)): New.
13934 (debug (basic_block_def &)): New.
13935 (debug (basic_block_def *)): New.
13937 (dump_node (const_tree, int, FILE *)): Correct source file.
13939 (debug (die_struct &)): New.
13940 (debug (die_struct *)): New.
13942 (extern debug (die_struct &)): New.
13943 (extern debug (die_struct *)): New.
13944 * gimple-pretty-print.c
13945 (debug (gimple_statement_d &)): New.
13946 (debug (gimple_statement_d *)): New.
13947 * gimple-pretty-print.h
13948 (extern debug (gimple_statement_d &)): New.
13949 (extern debug (gimple_statement_d *)): New.
13951 (debug (ira_allocno_copy &)): New.
13952 (debug (ira_allocno_copy *)): New.
13953 (debug (ira_allocno &)): New.
13954 (debug (ira_allocno *)): New.
13956 (extern debug (ira_allocno_copy &)): New.
13957 (extern debug (ira_allocno_copy *)): New.
13958 (extern debug (ira_allocno &)): New.
13959 (extern debug (ira_allocno *)): New.
13961 (debug (live_range &)): New.
13962 (debug (live_range *)): New.
13964 (debug (lra_live_range &)): New.
13965 (debug (lra_live_range *)): New.
13967 (debug (lra_live_range &)): New.
13968 (debug (lra_live_range *)): New.
13970 (debug (omega_pb_d &)): New.
13971 (debug (omega_pb_d *)): New.
13973 (extern debug (omega_pb_d &)): New.
13974 (extern debug (omega_pb_d *)): New.
13976 (debug (const rtx_def &)): New.
13977 (debug (const rtx_def *)): New.
13979 (debug_tree (tree): Move within file.
13980 (debug_raw (const tree_node &)): New.
13981 (debug_raw (const tree_node *)): New.
13982 (dump_tree_via_hooks (const tree_node *, int)): New.
13983 (debug (const tree_node &)): New.
13984 (debug (const tree_node *)): New.
13985 (debug_verbose (const tree_node &)): New.
13986 (debug_verbose (const tree_node *)): New.
13987 (debug_head (const tree_node &)): New.
13988 (debug_head (const tree_node *)): New.
13989 (debug_body (const tree_node &)): New.
13990 (debug_body (const tree_node *)): New.
13991 (debug_vec_tree (tree): Move and reimplement in terms of dump.
13992 (debug (vec<tree, va_gc> &)): New.
13993 (debug (vec<tree, va_gc> *)): New.
13995 (extern debug (const rtx_def &)): New.
13996 (extern debug (const rtx_def *)): New.
13998 (debug_raw (simple_bitmap_def &)): New.
13999 (debug_raw (simple_bitmap_def *)): New.
14000 (debug (simple_bitmap_def &)): New.
14001 (debug (simple_bitmap_def *)): New.
14003 (extern debug (simple_bitmap_def &)): New.
14004 (extern debug (simple_bitmap_def *)): New.
14005 (extern debug_raw (simple_bitmap_def &)): New.
14006 (extern debug_raw (simple_bitmap_def *)): New.
14008 (debug (vinsn_def &)): New.
14009 (debug (vinsn_def *)): New.
14010 (debug_verbose (vinsn_def &)): New.
14011 (debug_verbose (vinsn_def *)): New.
14012 (debug (expr_def &)): New.
14013 (debug (expr_def *)): New.
14014 (debug_verbose (expr_def &)): New.
14015 (debug_verbose (expr_def *)): New.
14016 (debug (vec<rtx> &)): New.
14017 (debug (vec<rtx> *)): New.
14019 (extern debug (vinsn_def &)): New.
14020 (extern debug (vinsn_def *)): New.
14021 (extern debug_verbose (vinsn_def &)): New.
14022 (extern debug_verbose (vinsn_def *)): New.
14023 (extern debug (expr_def &)): New.
14024 (extern debug (expr_def *)): New.
14025 (extern debug_verbose (expr_def &)): New.
14026 (extern debug_verbose (expr_def *)): New.
14027 (extern debug (vec<rtx> &)): New.
14028 (extern debug (vec<rtx> *)): New.
14030 (_list_iter_cond_expr): Make inline instead of static.
14032 (debug (sreal &)): New.
14033 (debug (sreal *)): New.
14035 (extern debug (sreal &)): New.
14036 (extern debug (sreal *)): New.
14038 (extern debug_raw (const tree_node &)): New.
14039 (extern debug_raw (const tree_node *)): New.
14040 (extern debug (const tree_node &)): New.
14041 (extern debug (const tree_node *)): New.
14042 (extern debug_verbose (const tree_node &)): New.
14043 (extern debug_verbose (const tree_node *)): New.
14044 (extern debug_head (const tree_node &)): New.
14045 (extern debug_head (const tree_node *)): New.
14046 (extern debug_body (const tree_node &)): New.
14047 (extern debug_body (const tree_node *)): New.
14048 (extern debug (vec<tree, va_gc> &)): New.
14049 (extern debug (vec<tree, va_gc> *)): New.
14051 (debug (struct loop &)): New.
14052 (debug (struct loop *)): New.
14053 (debug_verbose (struct loop &)): New.
14054 (debug_verbose (struct loop *)): New.
14055 * tree-dump.c: Add header dependence.
14057 (extern debug (struct loop &)): New.
14058 (extern debug (struct loop *)): New.
14059 (extern debug_verbose (struct loop &)): New.
14060 (extern debug_verbose (struct loop *)): New.
14062 (debug (data_reference &)): New.
14063 (debug (data_reference *)): New.
14064 (debug (vec<data_reference_p> &)): New.
14065 (debug (vec<data_reference_p> *)): New.
14066 (debug (vec<ddr_p> &)): New.
14067 (debug (vec<ddr_p> *)): New.
14069 (extern debug (data_reference &)): New.
14070 (extern debug (data_reference *)): New.
14071 (extern debug (vec<data_reference_p> &)): New.
14072 (extern debug (vec<data_reference_p> *)): New.
14073 (extern debug (vec<ddr_p> &)): New.
14074 (extern debug (vec<ddr_p> *)): New.
14076 (debug (pt_solution &)): New.
14077 (debug (pt_solution *)): New.
14079 (extern debug (pt_solution &)): New.
14080 (extern debug (pt_solution *)): New.
14082 (debug (_var_map &)): New.
14083 (debug (_var_map *)): New.
14084 (debug (tree_live_info_d &)): New.
14085 (debug (tree_live_info_d *)): New.
14087 (extern debug (_var_map &)): New.
14088 (extern debug (_var_map *)): New.
14089 (extern debug (tree_live_info_d &)): New.
14090 (extern debug (tree_live_info_d *)): New.
14092 2013-03-28 Jan Hubicka <jh@suse.cz>
14094 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
14096 2013-03-28 Ian Bolton <ian.bolton@arm.com>
14098 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
14099 record only when desired or required.
14101 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
14103 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
14104 *vec_extractv2di_1_rex64. Use x64 isa attribute.
14106 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14108 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
14109 (*andsi3_compare0_uxtw): New pattern.
14110 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
14111 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
14113 2013-03-28 Jan Hubicka <jh@suse.cz>
14115 * data-streamer-in.c (streamer_read_gcov_count): New function.
14116 * gimple-streamer-out.c: Include value-prof.h.
14117 (output_gimple_stmt): Output histogram.
14118 (output_bb): Use streamer_write_gcov_count.
14119 * value-prof.c: Include data-streamer.h
14120 (dump_histogram_value): Add HIST_TYPE_MAX.
14121 (stream_out_histogram_value): New function.
14122 (stream_in_histogram_value): New function.
14123 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
14124 (stream_out_histogram_value, stream_in_histogram_value): Declare.
14125 * data-streamer-out.c (streamer_write_gcov_count): New function.
14126 (streamer_write_gcov_count_stream): New function.
14127 * lto-cgraph.c (lto_output_edge): Update counter streaming.
14128 (lto_output_node): Likewise.
14129 (input_node, input_edge): Likewise.
14130 * lto-streamer-out.c (output_cfg): Update streaming.
14131 * lto-streamer-in.c (input_cfg): Likewise.
14132 * data-streamer.h (streamer_write_gcov_count,
14133 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
14134 * gimple-streamer-in.c: Include value-prof.h
14135 (input_gimple_stmt): Input histograms.
14136 (input_bb): Update profile streaming.
14138 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14140 * genmodes.c (emit_max_int): New function.
14141 (emit_insn_modes_h): Added call to emit_max_function.
14142 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
14144 * machmode.def: Fixed comment.
14146 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14148 * combine.c (try_combine): Removed useless assert.
14149 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
14151 2013-03-28 Marek Polacek <polacek@redhat.com>
14152 Richard Biener <rguenther@suse.de>
14154 PR tree-optimization/56695
14155 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
14156 build signed result of a vector comparison.
14157 * tree-cfg.c (verify_gimple_comparison): Check that a result
14158 of a vector comparison has signed type.
14160 2013-03-28 Richard Biener <rguenther@suse.de>
14162 PR tree-optimization/37021
14163 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
14164 do not restrict gaps between groups.
14165 * tree-vect-stmts.c (vectorizable_load): Properly account for
14166 a gap between groups.
14168 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
14170 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
14171 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
14174 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
14176 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
14177 * doc/extend.texi (Named Address Spaces): Ditto.
14178 (Variable Attributes): Ditto.
14180 2013-03-27 Kai Tietz <ktietz@redhat.com>
14182 * config.build: Add support for cygwin x64 target.
14183 * config.gcc: Likewise.
14184 * config.host: Likewise.
14185 * configure.ac: Likewise
14186 * configure: Regenerated.
14188 2013-03-27 Kai Tietz <ktietz@redhat.com>
14190 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
14191 * config/i386/t-cygwin-w64: New file.
14192 * config/i386/cygwin-w64.h: New file.
14193 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
14194 and add support for x64-cygwin target.
14195 (CPP_SPEC): Likewise.
14196 (CXX_WRAP_SPEC_LIST): Undefine before define.
14197 (LIBGCJ_SONAME): Use 15 as version.
14199 2013-03-27 Richard Biener <rguenther@suse.de>
14201 PR tree-optimization/56716
14202 * tree-ssa-structalias.c (perform_var_substitution): Adjust
14203 dumping for ref nodes.
14205 2013-03-27 Martin Jambor <mjambor@suse.cz>
14207 PR tree-optimization/55334
14208 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
14209 restricted pointers to arrays.
14211 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
14213 * Makefile.in (.SUFFIXES): Add .cc.
14214 (.c.o): Apply same recipe for implicit rule .cc.o.
14216 2013-03-27 Richard Biener <rguenther@suse.de>
14218 PR tree-optimization/37021
14219 * tree-vect-data-refs.c (vect_check_strided_load): Allow
14220 REALPART/IMAGPART_EXPRs around the supported refs.
14221 * tree-ssa-structalias.c (find_func_aliases): Assume that
14222 floating-point values are not used to transfer pointers.
14224 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
14226 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
14227 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
14228 * gcc/doc/tm.texi: Regenerate.
14229 * gcc/targhooks.h (default_has_ifunc_p): New.
14230 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
14231 * gcc/config/linux-protos.h: New file.
14232 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
14233 this hook for linux which disables support of indirect functions in
14235 * gcc/config/linux-android.c: New file.
14236 * gcc/config/t-linux-android.c: Ditto.
14237 * gcc/config.gcc: Added new object file linux-android.o.
14238 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
14239 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
14240 * gcc/varasm.c (do_assemble_alias): Likewise.
14241 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
14242 doesn't support indirect functions.
14243 * configure: Regenerate.
14245 2013-03-27 Bin Cheng <bin.cheng@arm.com>
14248 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
14249 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
14252 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14254 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
14256 2013-03-27 Terry Guo <terry.guo@arm.com>
14258 * config/arm/arm-cores.def: Added core cortex-r7.
14259 * config/arm/arm-tune.md: Regenerated.
14260 * config/arm/arm-tables.opt: Regenerated.
14261 * doc/invoke.texi: Added entry for core cortex-r7.
14263 2013-03-27 Walter Lee <walt@tilera.com>
14265 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
14266 double-decrement of next_scratch_regno.
14268 2013-03-27 Walter Lee <walt@tilera.com>
14270 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
14272 (insn_v1mulus): Ditto.
14273 (insn_v2muls): Ditto.
14275 2013-03-27 Walter Lee <walt@tilera.com>
14277 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
14278 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
14280 2013-03-27 Walter Lee <walt@tilera.com>
14282 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
14283 (*sibcall_value): Ditto.
14285 2013-03-27 Walter Lee <walt@tilera.com>
14287 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
14288 (insn_mnz_v8qi): ... this ...
14289 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
14291 (insn_v<n>mnz): Replaced by ...
14292 (insn_v1mnz): ... this ...
14293 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
14295 (insn_mz_<mode>): Replaced by ...
14296 (insn_mz_v8qi): ... this ...
14297 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
14299 (insn_v<n>mz): Replaced by ...
14300 (insn_v1mz): ... this ...
14301 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
14304 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
14306 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
14308 2013-03-26 Roland McGrath <mcgrathr@google.com>
14310 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
14311 than fprintf with a non-constant, non-format string.
14313 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14315 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
14316 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
14317 operand 0 predicate.
14318 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
14319 attribute. Use general_x64nomem_operand as operand 1 predicate.
14320 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
14321 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14322 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
14323 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14324 (mov<mode>_insv_1): Remove expander. Merge insn with
14325 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
14326 Use general_x64nomem_operand as operand 1 predicate.
14327 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
14328 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
14329 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14330 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
14331 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
14332 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
14333 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14334 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
14335 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
14336 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
14337 (general_x64nomem_operand): Ditto.
14339 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
14341 * config/rtems.opt: Add -pthread option.
14343 2013-03-26 Richard Biener <rguenther@suse.de>
14345 * alias.c (find_base_term): Avoid redundant and not used recursion.
14346 (base_alias_check): Get the initial base term from the caller.
14347 (true_dependence_1): Compute and pass base terms to base_alias_check.
14348 (write_dependence_p): Likewise.
14349 (may_alias_p): Likewise.
14351 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
14353 * config/aarch64/aarch64.c (aarch64_classify_address): Support
14354 PC-relative load in SI modes and above only.
14356 2013-03-26 Xinyu Qi <xyqi@marvell.com>
14358 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
14359 * config/arm/iwmmxt.md (WCGR0): Update.
14360 (WCGR1, WCGR2, WCGR3): Likewise.
14362 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14364 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
14365 Use x64 and nox64 isa attributes.
14367 2013-03-26 Richard Biener <rguenther@suse.de>
14369 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
14370 alignment computations and rely on get_object_alignment_1
14371 for the !TYPE_P case.
14372 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
14374 2013-03-26 Walter Lee <walt@tilera.com>
14376 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
14377 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
14379 2013-03-25 Jeff Law <law@redhat.com>
14381 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
14382 check for INTEGRAL_TYPE_P that was missing due to checking in
14383 wrong version of prior patch.
14385 2013-03-25 Walter Lee <walt@tilera.com>
14387 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
14388 TILEGX_INSN_SHUFFLEBYTES1.
14389 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
14391 (tilegx_builtins): Ditto.
14392 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
14394 2013-03-25 Walter Lee <walt@tilera.com>
14396 * config/tilegx/tilegx.md (floatsisf2): New pattern.
14397 (floatunssisf2): New pattern.
14398 (floatsidf2): New pattern.
14399 (floatunssidf2): New pattern.
14401 2013-03-25 Walter Lee <walt@tilera.com>
14403 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
14404 tests for constraint J, K, N, P.
14406 2013-03-25 Walter Lee <walt@tilera.com>
14408 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
14409 Use indirect/pcrel encoding.
14410 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
14413 2013-03-25 Steve Ellcey <sellcey@mips.com>
14415 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
14416 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
14417 * config/mips/mips.c (mips_option_override): Set IMADD default.
14418 * config/mips/mips.h (PTF_AVOID_IMADD): New.
14419 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
14420 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
14421 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
14422 * doc/invoke.texi (-mimadd/-mno-imadd): New.
14424 2013-03-25 Jeff Law <law@redhat.com>
14426 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
14427 slightly to avoid creating and folding useless trees. Simplify
14428 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
14430 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
14432 * config/i386/i386.md (*zero_extendsidi2): Merge with
14433 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
14434 * config/i386/predicates.md (x86_64_zext_operand): Rename from
14435 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
14436 targets. Clarify comment.
14438 2013-03-25 Martin Jambor <mjambor@suse.cz>
14440 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
14441 pass-through jump functions differently.
14442 (ipa_read_jump_function): Likewise. Also use setter functions to set
14445 2013-03-25 Martin Jambor <mjambor@suse.cz>
14447 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
14448 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
14450 (ipa_get_indirect_edge_target): New function.
14451 (devirtualization_time_bonus): New parameter known_aggs, pass it to
14452 ipa_get_indirect_edge_target. Update all callers.
14453 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
14454 ipa_get_indirect_edge_target_1 instead of calling
14455 ipa_get_indirect_edge_target.
14456 (create_specialized_node): Pass aggvlas to
14457 ipcp_discover_new_direct_edges.
14459 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14461 * config/arm/arm.md (f_sels, f_seld): New types.
14462 (*cmov<mode>): New pattern.
14463 * config/arm/predicates.md (arm_vsel_comparison_operator): New
14466 2013-03-25 Kai Tietz <ktietz@redhat.com>
14468 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
14469 POSIX-printf for mingw-hosted builds.
14471 2013-03-25 Richard Biener <rguenther@suse.de>
14473 PR middle-end/56694
14474 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
14475 must-not-throw stmt location.
14477 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14479 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
14480 Emit load-acquire versions when acq is true.
14481 (arm_emit_store_exclusive): Add rel parameter.
14482 Emit store-release versions when rel is true.
14483 (arm_split_compare_and_swap): Use acquire-release instructions
14485 of barriers when appropriate.
14486 (arm_split_atomic_op): Likewise.
14487 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
14488 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
14489 (VUNSPEC_SLX): Likewise.
14490 (VUNSPEC_LDA): Likewise.
14491 (VUNSPEC_STL): Likewise.
14492 * config/arm/sync.md (atomic_load<mode>): New pattern.
14493 (atomic_store<mode>): Likewise.
14494 (arm_load_acquire_exclusive<mode>): Likewise.
14495 (arm_load_acquire_exclusivesi): Likewise.
14496 (arm_load_acquire_exclusivedi): Likewise.
14497 (arm_store_release_exclusive<mode>): Likewise.
14499 2013-03-25 Catherine Moore <clm@codesourcery.com>
14501 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
14502 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
14503 * config/mip/predicates.md (lwsp_swsp_operand,
14504 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
14505 sb16_operand, db4_operand, db7_operand, ib3_operand,
14506 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
14507 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
14508 andi16_operand): New predicates.
14509 * config/mips/mips.md (compression): New attribute.
14510 (enabled): New attribute.
14511 (length): Consider compression in computing length.
14512 (shift_compression): New code attribute.
14513 (*add<mode>3): New operands. Record compression.
14514 (sub<mode>3): Likewise.
14515 (one_cmpl<mode>2): Likewise.
14516 (*and<mode>3): Likewise.
14517 (*ior<mode>3): Likewise.
14518 (unnamed pattern for xor): Likewise.
14519 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
14520 (*<optab><mode>3): Likewise.
14521 (*mov<mode>_internal: Likewise.
14522 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
14523 (mips_unsigned_immediate_p): New.
14524 (umips_lwsp_swsp_address_p): New.
14525 (m16_based_address_p): New.
14526 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
14527 (mips_unsigned_immediate_p): New prototype.
14528 (lwsp_swsp_address_p): New prototype.
14529 (m16_based_address_p): New prototype.
14530 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
14531 (mips_signed_immediate_p): New function.
14532 (m16_based_address_p): New function.
14533 (lwsp_swsp_address_p): New function.
14534 (mips_print_operand_punctuation): Recognize short delay slot insns
14535 for microMIPS.add<mode>3"
14537 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14540 * config/arm/iterators.md (v_cmp_result): New mode attribute.
14541 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
14543 2013-03-25 Richard Biener <rguenther@suse.de>
14545 PR tree-optimization/56689
14546 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
14549 2013-03-25 Richard Biener <rguenther@suse.de>
14551 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
14553 (memory_references): Likewise.
14554 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
14555 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
14556 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
14557 (gather_mem_refs_in_loops): Fold into ...
14558 (analyze_memory_references): ... this. Move initialization
14559 to tree_ssa_lim_initialize.
14560 (fill_always_executed_in): Rename to ...
14561 (fill_always_executed_in_1): ... this.
14562 (fill_always_executed_in): Move contains_call computation to
14563 this new function from ...
14564 (tree_ssa_lim_initialize): ... here.
14565 (tree_ssa_lim): Call fill_always_executed_in.
14567 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
14569 * postreload.c (reload_combine): Fix code detecting returns.
14571 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
14573 * function.c (emit_use_return_register_into_block): On cc0 targets,
14574 do not emit the sequence between cc0 setter and user.
14576 2013-03-25 Kai Tietz <ktietz@redhat.com>
14578 * config/i386/predicates.md (local_symbolic_operand): Interpret
14579 dll-imported symbols as none-local.
14581 2013-03-25 Richard Biener <rguenther@suse.de>
14583 * tree-ssa-loop-im.c (struct depend): Remove.
14584 (struct lim_aux_data): Make depends a vec of gimples.
14585 (free_lim_aux_data): Adjust.
14586 (add_dependency): Likewise.
14587 (set_level): Likewise.
14589 2013-03-25 Richard Biener <rguenther@suse.de>
14591 PR middle-end/56434
14592 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
14593 the pointer returned by calls with ECF_MALLOC set.
14595 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
14597 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
14599 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
14601 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
14602 using MMXMODE mode iterator.
14603 (*move<mode>_internal): Merge with *movv2sf_internal and
14604 *movv2sf_internal_rex64 using MMXMODE mode iterator.
14606 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
14608 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
14609 (record_last_mem_set_info): Likewise.
14611 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
14612 of XNEWVEC followed by memset.
14613 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
14615 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
14617 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
14618 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
14619 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
14620 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
14621 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
14622 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
14623 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
14624 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
14625 BARRIER_P instead of GET_CODE.
14627 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14629 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
14630 inaccuracy in the probing code.
14632 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
14633 (ctrapdi4): Likewise.
14635 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14637 * calls.c (expand_call): Add missing guard to code handling return
14638 of non-BLKmode structures in MSB.
14639 * function.c (expand_function_end): Likewise.
14641 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14643 * combine.c (try_combine): Adjust comment. Do not add the set of
14644 insn #0 if the destination indirectly is set or dies in insn #2.
14645 Tidy up code to distribute a new note.
14647 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
14649 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
14650 also for alternatives 16 and 17.
14652 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
14654 * config/i386/sse.md (*mov<mode>_internal): Merge with
14655 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
14656 Emit insn template depending on type attribute. Use
14657 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14658 movd instead of movq mnemonic for interunit moves. Rewrite mode
14659 attribute calculation. Remove unit attribute calculation.
14660 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
14661 Set prefix_data16 attribute for DImode ssemov types.
14662 Use Ym instead of y for SSE-MMX conversion alternatives.
14663 Reorder operand constraints.
14665 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
14667 * df.h (df_insn_delete): Adjust prototype.
14668 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
14669 and let it decide whether mark the basic block dirty.
14670 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
14671 * df-scan.c (df_insn_info_delete): New helper function, split
14672 off from df_insn_delete.
14673 (df_scan_free_bb_info): Use it.
14674 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
14676 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
14677 that the insn is actually an insn and it has a non-NULL basic block.
14678 Do not mark basic block dirty if only deleting a DEBUG_INSN.
14680 2013-03-22 Richard Biener <rguenther@suse.de>
14682 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
14684 (mem_ref_alloc): Do not allocate them.
14685 (refs_independent_p): Do not query or maintain a cache.
14687 2013-03-22 Richard Biener <rguenther@suse.de>
14689 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
14690 (gather_mem_refs_in_loops): Do not compute it.
14691 (analyze_memory_references): Do not allocate it.
14692 (tree_ssa_lim_finalize): Do not free it.
14693 (for_all_locs_in_loop): Do not query all_refs_in_loop.
14695 2013-03-22 Richard Biener <rguenther@suse.de>
14697 * is-a.h (as_a): Use gcc_checking_assert.
14699 2013-03-22 Ian Bolton <ian.bolton@arm.com>
14701 * config/aarch64/aarch64.c (aarch64_print_operand): New
14702 format specifier for printing a constant in hex.
14703 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
14704 format specifier for printing second operand.
14706 2013-03-22 Richard Biener <rguenther@suse.de>
14708 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
14710 (gather_mem_refs_in_loops): Perform store accumulation here.
14711 (create_vop_ref_mapping_loop): Remove.
14712 (create_vop_ref_mapping): Likewise.
14713 (analyze_memory_references): Initialize refs_stored_in_loop.
14714 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
14715 (record_indep_loop): Remove.
14716 (record_dep_loop): New function.
14717 (ref_indep_loop_p_1): Adjust to only walk over references
14718 in the loop, not its subloops.
14719 (ref_indep_loop_p): Rename to ...
14720 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
14721 maintaining a more fine-grained cache.
14722 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
14723 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
14725 2013-03-22 Richard Biener <rguenther@suse.de>
14727 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
14728 (struct mem_ref): Make accesses_in_loop a vec of a vec of
14729 aggregate mem_ref_loc.
14730 (free_mem_ref_locs): Inline into ...
14731 (memref_free): ... this and adjust.
14732 (mem_ref_alloc): Adjust.
14733 (mem_ref_locs_alloc): Remove.
14734 (record_mem_ref_loc): Adjust.
14735 (get_all_locs_in_loop): Rewrite into ...
14736 (for_all_locs_in_loop): ... this iterator.
14737 (rewrite_mem_ref_loc): New functor.
14738 (rewrite_mem_refs): Use for_all_locs_in_loop.
14739 (sm_set_flag_if_changed): New functor.
14740 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
14741 (ref_always_accessed): New functor.
14742 (ref_always_accessed_p): Use for_all_locs_in_loop.
14744 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14746 * tree-pass.h (PROP_gimple_lvec): New.
14747 * passes.c (dump_properties): Handle PROP_gimple_lvec.
14748 (init_optimization_passes): Move pass_lower_vector.
14749 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
14751 (pass_lower_vector): Provide PROP_gimple_lvec.
14752 (pass_lower_vector_ssa): Likewise.
14753 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
14755 2013-03-21 Mark Wielaard <mjw@redhat.com>
14757 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
14759 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14761 * config/i386/i386.md (*movdi_internal): Disparage slightly
14762 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
14763 conversion alternatives.
14765 2013-03-21 Jakub Jelinek <jakub@redhat.com>
14767 PR middle-end/48087
14768 * diagnostic.def (DK_WERROR): New kind.
14769 * diagnostic.h (werrorcount): Define.
14770 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
14771 promoted to DK_ERROR, increment DK_WERROR counter instead of
14773 * toplev.c (toplev_main): Call print_ignored_options even if
14774 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
14775 even if just werrorcount is non-zero.
14778 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
14780 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
14781 (string_cst_pool_decl): New function.
14782 (optimize_one_addr_into_implicit_ptr): New function.
14783 (resolve_addr_in_expr): Optimize DWARF location expression
14784 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
14785 which doesn't live in memory, but has DW_AT_location or
14786 DW_AT_const_value, or refers to a string literal, into
14787 DW_OP_GNU_implicit_pointer.
14788 (optimize_location_into_implicit_ptr): New function.
14789 (resolve_addr): If removing DW_AT_location of a variable because
14790 it was DW_OP_addr of address of the variable, but the variable doesn't
14791 live in memory, try to emit const value attribute for the initializer.
14793 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14795 * tree.h (VECTOR_TYPE_P): New macro.
14796 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
14797 TYPE_MODE): Use it.
14798 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
14799 VEC_COND_EXPR cannot be lvalues.
14800 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
14802 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14804 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
14805 Restrict the transformation to equal modes.
14807 2013-03-21 Richard Biener <rguenther@suse.de>
14809 PR tree-optimization/39326
14810 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
14811 (MEM_ANALYZABLE): Adjust.
14812 (record_mem_ref_loc): Move bitmap ops ...
14813 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
14814 unanalyzable refs, do not record locations for it.
14815 (analyze_memory_references): Allocate ref zero as shared
14817 (refs_independent_p): Do not test for unanalyzed mems here.
14818 (ref_indep_loop_p_1): Special-case disambiguation against
14819 the unanalyzed ref.
14820 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
14822 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
14824 * config/arm/arm-protos.h (tune_params): Add
14825 prefer_neon_for_64bits field.
14826 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
14827 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
14828 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
14829 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
14830 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
14831 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
14832 (arm_option_override): Handle -mneon-for-64bits new option.
14833 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
14834 (prefer_neon_for_64bits): Declare new variable.
14835 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
14836 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
14837 (arch_enabled): Handle new arch types. Remove support for onlya8
14839 (one_cmpldi2): Use new arch names.
14840 (zero_extend<mode>di2, extend<mode>di2): Ditto.
14841 * config/arm/arm.opt (mneon-for-64bits): Add option.
14842 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
14843 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
14844 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
14846 * doc/invoke.texi (-mneon-for-64bits): Document.
14848 2013-03-21 Richard Biener <rguenther@suse.de>
14850 PR tree-optimization/39326
14851 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
14852 (sort_bbs_in_loop_postorder_cmp): New function.
14853 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
14855 2013-03-21 Richard Biener <rguenther@suse.de>
14857 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
14858 (vect_insert_into_interleaving_chain): Likewise.
14859 (vect_drs_dependent_in_basic_block): Inline ...
14860 (vect_slp_analyze_data_ref_dependence): ... here. New function,
14862 (vect_analyze_data_ref_dependence): ... here. Simplify.
14863 (vect_check_interleaving): Simplify.
14864 (vect_analyze_data_ref_dependences): Likewise. Split out ...
14865 (vect_slp_analyze_data_ref_dependences): ... this new function.
14866 (dr_group_sort_cmp): New function.
14867 (vect_analyze_data_ref_accesses): Compute data-reference groups
14868 here instead of in vect_analyze_data_ref_dependence. Use
14869 a more efficient algorithm.
14870 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
14871 vect_slp_analyze_data_ref_dependences. Call
14872 vect_analyze_data_ref_accesses earlier.
14873 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
14874 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
14875 (vect_slp_analyze_data_ref_dependences): New prototype.
14877 2013-03-21 Richard Biener <rguenther@suse.de>
14879 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
14880 ref is stored in the loop.
14881 (find_refs_for_sm): Walk only over all stores.
14882 (store_motion_loop): Allocate from lim_bitmap_obstack.
14883 (store_motion): Likewise.
14885 2013-03-21 Richard Biener <rguenther@suse.de>
14887 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
14888 Update virtual SSA form.
14890 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14892 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
14893 * configure: Regenerate.
14894 * config.in: Regenerate.
14895 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
14896 if !HAVE_LD_EH_FRAME_CIEV3.
14898 2013-03-21 Richard Biener <rguenther@suse.de>
14900 * tree-cfg.c (verify_expr_no_block): New function.
14901 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14902 nor DECL_VALUE_EXPR have locations with associated blocks.
14903 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14904 (clear_unused_block_pointer): Remove code dealing with
14905 blocks in DECL_DEBUG_EXPR locations.
14907 2013-03-21 Richard Biener <rguenther@suse.de>
14909 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14910 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14911 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14912 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14913 instead of DECL_DEBUG_EXPR_IS_FROM.
14914 * gimplify.c (gimplify_modify_expr): Likewise.
14915 * tree-cfg.c (verify_expr_location_1): Likewise.
14916 * tree-complex.c (create_one_component_var): Likewise.
14917 * tree-sra.c (create_access_replacement): Likewise.
14918 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14919 (clear_unused_block_pointer): Likewise.
14920 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14921 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14922 * var-tracking.c (var_debug_decl): Likewise.
14923 (track_expr_p): Likewise.
14924 * tree-inline.c (add_local_variables): Likewise. Set
14925 DECL_HAS_DEBUG_EXPR_P after copying it.
14926 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14927 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14929 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14932 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14933 * configure: Regenerate.
14934 * config.in: Regenerate.
14935 * config/i386/i386.md (*movdf_internal): Use
14936 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14937 movd instead of movq mnemonic for interunit moves.
14938 (*movdi_internal): Ditto.
14940 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14942 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14943 (abd<mode>_3): New pattern.
14944 (aba<mode>_3): New pattern.
14945 (fabd<mode>_3): New pattern.
14947 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14949 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14950 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14951 occurrence of REGISTER_PREFIX as its empty string.
14953 2013-03-20 Jeff Law <law@redhat.com>
14955 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14956 addititional equivalences for equality comparisons between an SSA_NAME
14957 and a constant where the SSA_NAME was set from a widening conversion.
14959 2013-03-20 Walter Lee <walt@tilera.com>
14961 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14963 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14965 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14966 depending on type attribute.
14967 (*movti_internal): Ditto.
14968 (*movtf_internal): Ditto.
14969 (*movxf_internal): Ditto.
14970 (*movdf_internal): Ditto.
14971 (*movsf_internal): Ditto.
14973 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14975 * config/i386/i386.md (*movti_internal): Set prefix attribute to
14976 maybe_vex for sselog1 and ssemov types.
14977 (*movdi_internal): Reorder operand constraints.
14978 (*movsi_internal): Ditto. Set prefix attribute to
14979 maybe_vex for sselog1 and ssemov types.
14980 (*movtf_internal): Set prefix attribute to maybe_vex
14981 for sselog1 and ssemov types.
14982 (*movdf_internal): Ditto. Set prefix_data16 attribute for
14983 DImode ssemov types. Reorder operand constraints.
14984 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
14985 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
14986 attribute for SImode ssemov types. Reorder operand constraints.
14988 2013-03-20 Martin Jambor <mjambor@suse.cz>
14990 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
14991 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
14993 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
14995 * config/rs6000/predicates.md (indexed_address, update_address_mem
14996 update_indexed_address_mem): New predicates.
14997 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
14998 attribute for load/store instructions.
14999 * config/rs6000/dfp.md (movsd_store): Likewise.
15000 (movsd_load): Likewise.
15001 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
15002 (unnamed HI->DI extend define_insn): Likewise.
15003 (unnamed SI->DI extend define_insn): Likewise.
15004 (unnamed QI->SI extend define_insn): Likewise.
15005 (unnamed QI->HI extend define_insn): Likewise.
15006 (unnamed HI->SI extend define_insn): Likewise.
15007 (unnamed HI->SI extend define_insn): Likewise.
15008 (extendsfdf2_fpr): Likewise.
15009 (movsi_internal1): Likewise.
15010 (movsi_internal1_single): Likewise.
15011 (movhi_internal): Likewise.
15012 (movqi_internal): Likewise.
15013 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
15014 attribute for load/store instructions.
15015 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
15017 (mov<mode>_softfloat): Likewise.
15018 (mov<mode>_hardfloat32): Likewise.
15019 (mov<mode>_hardfloat64): Likewise.
15020 (mov<mode>_softfloat64): Likewise.
15021 (movdi_internal32): Likewise.
15022 (movdi_internal64): Likewise.
15023 (probe_stack_<mode>): Likewise.
15025 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
15027 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
15028 floating point, and decimal floating point to reload iterator.
15030 * config/rs6000/constraints.md (wl constraint): New constraints to
15031 return FLOAT_REGS if certain options are used to reduce the number
15032 of separate patterns that exist in the file.
15033 (wx constraint): Likewise.
15034 (wz constraint): Likewise.
15036 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15037 -mdebug=reg, print wg, wl, wx, and wz constraints.
15038 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15039 Initialize the reload functions for 64-bit binary/decimal floating
15041 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15042 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15043 create the buffer on the stack to overcome not having a 32-bit
15045 (rs6000_emit_move): Likewise.
15046 (rs6000_secondary_memory_needed_rtx): Likewise.
15047 (rs6000_alloc_sdmode_stack_slot): Likewise.
15048 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15049 via xxlxor, just like DFmode 0.0.
15051 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
15052 define as 1 if we are running on a power7 or newer.
15053 (enum r6000_reg_class_enum): Add new constraints.
15055 * config/rs6000/dfp.md (movsd): Delete, combine with binary
15056 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15057 with other moves by using conditional constraits (wg). Use LFIWZX
15058 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
15059 (movsd splitter): Likewise.
15060 (movsd_hardfloat): Likewise.
15061 (movsd_softfloat): Likewise.
15063 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
15064 binary and decimal floating point moves.
15065 (fmove_ok): New attributes to combine binary and decimal floating
15066 point moves, and to combine power6x (mfpgpr) moves along normal
15068 (real_value_to_target): Likewise.
15069 (f32_lr): Likewise.
15070 (f32_lm): Likewise.
15071 (f32_li): Likewise.
15072 (f32_sr): Likewise.
15073 (f32_sm): Likewise.
15074 (f32_si): Likewise.
15075 (movsf): Combine binary and decimal floating point moves. Combine
15076 power6x (mfpgpr) moves with other moves by using conditional
15077 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
15078 (mov<mode> for SFmode/SDmode); Likewise.
15079 (SFmode/SDmode splitters): Likewise.
15080 (movsf_hardfloat): Likewise.
15081 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
15082 (movsf_softfloat): Likewise.
15083 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
15085 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
15086 wx and wz constraints.
15088 * config/rs6000/constraints.md (wg constraint): New constraint to
15089 return FLOAT_REGS if -mmfpgpr (power6x) was used.
15091 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
15094 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15095 -mdebug=reg, print wg, wl, wx, and wz constraints.
15096 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15097 Initialize the reload functions for 64-bit binary/decimal floating
15099 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15100 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15101 create the buffer on the stack to overcome not having a 32-bit
15103 (rs6000_emit_move): Likewise.
15104 (rs6000_secondary_memory_needed_rtx): Likewise.
15105 (rs6000_alloc_sdmode_stack_slot): Likewise.
15106 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15107 via xxlxor, just like DFmode 0.0.
15109 * config/rs6000/dfp.md (movdd): Delete, combine with binary
15110 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15111 with other moves by using conditional constraits (wg). Use LFIWZX
15112 and STFIWX for loading SDmode on power7.
15113 (movdd splitters): Likewise.
15114 (movdd_hardfloat32): Likewise.
15115 (movdd_softfloat32): Likewise.
15116 (movdd_hardfloat64_mfpgpr): Likewise.
15117 (movdd_hardfloat64): Likewise.
15118 (movdd_softfloat64): Likewise.
15120 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
15121 64-bit binary and decimal floating point moves.
15122 (FMOVE64X): Likewise.
15123 (movdf): Combine 64-bit binary and decimal floating point moves.
15124 Combine power6x (mfpgpr) moves with other moves by using
15125 conditional constraits (wg).
15126 (mov<mode> for DFmode/DDmode): Likewise.
15127 (DFmode/DDmode splitters): Likewise.
15128 (movdf_hardfloat32): Likewise.
15129 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
15130 (movdf_softfloat32): Likewise.
15131 (movdf_hardfloat64_mfpgpr): Likewise.
15132 (movdf_hardfloat64): Likewise.
15133 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
15134 (movdf_softfloat64): Likewise.
15135 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
15136 (reload_<mode>_load): Move to later in the file so they aren't in
15137 the middle of the floating point move insns.
15138 (reload_<mode>_store): Likewise.
15140 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
15143 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
15144 constraint if -mdebug=reg.
15145 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
15146 Enable using dd reload support if needed.
15148 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
15149 binary and decimal floating point moves in rs6000.md.
15150 (movtd_internal): Likewise.
15152 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
15153 decimal floating point moves.
15155 (movtf_internal): Likewise.
15156 (mov<mode>_internal, TDmode/TFmode): Likewise.
15157 (movtf_softfloat): Likewise.
15158 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
15160 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
15161 movdi_internal64, using wg constraint for move direct operations.
15162 (movdi_internal64): Likewise.
15164 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
15165 MODES_TIEABLE_P for selected modes. Print the numerical value of
15166 the various virtual registers. Use GPR/FPR first/last values,
15167 instead of hard coding the register numbers. Print which modes
15168 have reload functions registered.
15169 (rs6000_option_override_internal): If -mdebug=reg, trace the options
15170 settings before/after setting cpu, target and subtarget settings.
15171 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
15172 and for secondary reload failures in rs6000_secondary_reload_inner.
15173 (rs6000_secondary_reload_fail): Likewise.
15174 (rs6000_secondary_reload_inner): Likewise.
15176 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
15177 macros for first/last GPR and FPR registers.
15178 (LAST_GPR_REGNO): Likewise.
15179 (FIRST_FPR_REGNO): Likewise.
15180 (LAST_FPR_REGNO): Likewise.
15182 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
15183 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
15184 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
15185 (vcond<mode><mode>): Likewise.
15186 (vcondu<mode><mode>): Likewise.
15187 (vector_gtu<mode>): Likewise.
15188 (vector_gte<mode>): Likewise.
15189 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
15190 to prevent the compiler from converting DImode operations to TImode.
15191 (ior<mode>3): Likewise.
15192 (and<mode>3): Likewise.
15193 (one_cmpl<mode>2): Likewise.
15194 (nor<mode>3): Likewise.
15195 (andc<mode>3): Likewise.
15197 * config/rs6000/constraints.md (wt constraint): New constraint
15198 that returns VSX_REGS if TImode is allowed in VSX registers.
15200 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
15201 constant under VSX.
15203 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
15204 similar to TImode, but it is restricted to being in the GPRs.
15206 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
15207 TImode to occupy a single VSX register.
15209 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
15210 -mvsx-timode for power7/power8.
15211 (power7 cpu): Likewise.
15212 (power8 cpu): Likewise.
15214 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
15215 sure that TFmode/TDmode take up two registers if they are ever
15216 allowed in the upper VSX registers.
15217 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
15219 (rs6000_init_hard_regno_mode_ok): Likewise.
15220 (rs6000_debug_reg_global): Add debugging for PTImode and wt
15221 constraint. Print if LRA is turned on.
15222 (rs6000_option_override_internal): Give an error if -mvsx-timode
15223 and VSX is not enabled.
15224 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
15225 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
15226 to reg+offset addressing. Use PTImode when checking offset
15227 addresses for validity.
15228 (reg_offset_addressing_ok_p): Likewise.
15229 (rs6000_legitimate_offset_address_p): Likewise.
15230 (rs6000_legitimize_address): Likewise.
15231 (rs6000_legitimize_reload_address): Likewise.
15232 (rs6000_legitimate_address_p): Likewise.
15233 (rs6000_eliminate_indexed_memrefs): Likewise.
15234 (rs6000_emit_move): Likewise.
15235 (rs6000_secondary_reload): Likewise.
15236 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
15237 reloads to fpr registers to continue to use reg+offset addressing,
15238 but 64-bit reloads to altivec registers need reg+reg addressing.
15239 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
15240 it. Treat LO_SUM like a PLUS operation.
15241 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
15242 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
15243 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
15244 registers to share a register with a smaller sized type, since VSX
15245 puts scalars in the upper 64-bits.
15246 (print_operand): Add support for PTImode.
15247 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
15248 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
15249 registers, but don't have arithmetic support.
15250 (rs6000_memory_move_cost): Add test for VSX.
15251 (rs6000_opt_masks): Add -mvsx-timode.
15253 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
15256 (VSr): Use wt constraint for TImode.
15257 (VSv): Drop TImode support.
15258 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
15259 (vsx_movti_64bit): Likewise.
15260 (vsx_movti_32bit): Likewise.
15261 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
15262 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
15263 one '?' on the appropriate output constraint. Do not allow TImode
15264 logical operations on 32-bit systems.
15265 (vsx_ior<mode>3): Likewise.
15266 (vsx_xor<mode>3): Likewise.
15267 (vsx_one_cmpl<mode>2): Likewise.
15268 (vsx_nor<mode>3): Likewise.
15269 (vsx_andc<mode>3): Likewise.
15270 (vsx_concat_<mode>): Likewise.
15271 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
15273 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
15274 OPTION_MASK_VSX_TIMODE.
15275 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
15276 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
15278 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
15279 (TI2 iterator): New iterator for TImode, PTImode.
15280 (wd mode attribute): Add values for vector types.
15281 (movti_string): Replace TI move operations with operations for TImode
15282 and PTImode. Add support for TImode being allowed in VSX registers.
15283 (mov<mode>_string, TImode/PTImode): Likewise.
15284 (movti_ppc64): Likewise.
15285 (mov<mode>_ppc64, TImode/PTImode): Likewise.
15286 (TI mode splitters): Likewise.
15288 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
15291 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
15293 PR tree-optimization/56355
15294 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
15295 Also handle integers with undefined overflow.
15297 2013-03-20 Catherine Moore <clm@codesourcery.com>
15298 Maciej W. Rozycki <macro@codesourcery.com>
15299 Tom de Vries <tom@codesourcery.com>
15300 Nathan Sidwell <nathan@codesourcery.com>
15301 Iain Sandoe <iain@codesourcery.com>
15302 Nathan Froyd <froydnj@codesourcery.com>
15303 Chao-ying Fu <fu@mips.com>
15305 * doc/extend.texi: (micromips, nomicromips, nocompression):
15306 Document new function attributes.
15307 * doc/invoke.texi (minterlink-compressed, mmicromips,
15308 m14k, m14ke, m14kec): Document new options.
15309 (minterlink-mips16): Update documentation.
15310 * doc/md.texi (ZC, ZD): Document new constraints.
15311 * configure.ac (gcc_cv_as_micromips): Check if linker
15312 supports the .set micromips directive.
15313 * configure: Regenerate.
15314 * config.in: Regenerate.
15315 * config/mips/mips-tables.opt: Regenerate.
15316 * config/mips/micromips.md: New file.
15317 * constraints.md (ZC, ZD): New constraints.
15318 * config/mips/predicates.md (movep_src_register): New predicate.
15319 (movep_src_operand): New predicate.
15320 (non_volatile_mem_operand): New predicate.
15321 * config/mips/mips.md (multimem): New type.
15322 (length): Differentiate between 17-bit and 18-bit branch offsets.
15323 (MOVEP1, MOVEP2): New mode iterator.
15324 (mov_<load>l): Use ZC constraint.
15325 (mov_<load>r): Likewise.
15326 (mov_<store>l): Likewise.
15327 (mov_<store>r): Likewise.
15328 (*branch_equality<mode>_inverted): Add microMIPS support.
15329 (*branch_equality<mode>): Likewise.
15330 (*jump_absolute): Likewise.
15331 (indirect_jump_<mode>): Likewise.
15332 (tablejump_<mode>): Likewise.
15333 (<optab>_internal): Likewise.
15334 (sibcall_internal): Likewise.
15335 (sibcall_value_internal): Likewise.
15336 (prefetch): Use constraint ZD.
15337 * config/mips/mips.opt (minterlink-compressed): New option.
15338 (minterlink-mips16): Now an alias for minterlink-compressed.
15339 (mmicromips): New option.
15340 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15341 (compare_and_swap_12): Likewise.
15342 (sync_add<mode>): Likewise.
15343 (sync_<optab>_12): Likewise.
15344 (sync_old_<optab>_12): Likewise.
15345 (sync_new_<optab>_12): Likewise.
15346 (sync_nand_12): Likewise.
15347 (sync_old_nand_12): Likewise.
15348 (sync_new_nand_12): Likewise.
15349 (sync_sub<mode>): Likewise.
15350 (sync_old_add<mode>): Likewise.
15351 (sync_old_sub<mode>): Likewise.
15352 (sync_new_add<mode>): Likewise.
15353 (sync_new_sub<mode>): Likewise.
15354 (sync_<optab><mode>): Likewise.
15355 (sync_old_<optab><mode>): Likewise.
15356 (sync_new_<optab><mode>): Likewise.
15357 (sync_nand<mode>): Likewise.
15358 (sync_old_nand<mode>): Likewise.
15359 (sync_new_nand<mode>): Likewise.
15360 (sync_lock_test_and_set<mode>): Likewise.
15361 (test_and_set_12): Likewise.
15362 (atomic_compare_and_swap<mode>): Likewise.
15363 (atomic_exchange<mode>_llsc): Likewise.
15364 (atomic_fetch_add<mode>_llsc): Likewise.
15365 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15366 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15367 (umips_save_restore_pattern_p): Likewise.
15368 (umips_load_store_pair_p): Likewise.
15369 (umips_output_load_store_pair): Likewise.
15370 (umips_movep_target_p): Likewise.
15371 (umips_12bit_offset_address_p): Likewise.
15372 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15373 (mips_base_mips16): Rename this...
15374 (mips_base_compression_flags): ...to this. Update all uses.
15375 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15376 (mips_mips16_decl_p): Delete.
15377 (mips_nomips16_decl_p): Delete.
15378 (mips_get_compress_on_flags): New function.
15379 (mips_get_compress_off_flags): New function.
15380 (mips_get_compress_mode): New function.
15381 (mips_get_compress_on_name): New function.
15382 (mips_get_compress_off_name): New function.
15383 (mips_insert_attributes): Support multiple compression types.
15384 (mips_merge_decl_attributes): Likewise.
15385 (umips_12bit_offset_address_p): New function.
15386 (mips_start_function_definition): Emit .set micromips directive.
15387 (mips_call_may_need_jalx_p): New function.
15388 (mips_function_ok_for_sibcall): Add microMIPS support.
15389 (mips_print_operand_punctuation): Support short delay slots and
15391 (umips_swm_mask, umips_swm_encoding): New.
15392 (umips_build_save_restore): New function.
15393 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15394 (was_mips16_p): Remove.
15395 (old_compression_mode): New.
15396 (mips_set_compression_mode): New function.
15397 (mips_set_current_function): Add microMIPS support.
15398 (mips_option_override): Likewise.
15399 (umips_save_restore_pattern_p): New function.
15400 (umips_output_save_restore): New function.
15401 (umips_load_store_pair_p_1): New function.
15402 (umips_load_store_pair_p): New function.
15403 (umips_output_load_store_pair_1): New function.
15404 (umips_output_load_store_pair): New function.
15405 (umips_movep_target_p) New function.
15406 (mips_prepare_pch_save): Add microMIPS support.
15407 * config/mips/mips.h (TARGET_COMPRESSION): New.
15408 (TARGET_CPU_CPP_BUILTINS): Update macro
15409 to use new compression flags and to support microMIPS.
15410 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15411 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15412 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15413 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15414 (ASM_SPEC): Support mmicromips and mno-micromips.
15415 (M16STORE_REG_P): New macro.
15416 (MIPS_CALL): Support TARGET_MICROMIPS.
15417 (MICROMIPS_J): New macro.
15418 (mips_base_mips16): Rename this...
15419 (mips_base_compression_flags): ...to this.
15420 (UMIPS_12BIT_OFFSET_P): New macro.
15421 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15422 (MULTILIB_DIRNAMES): Likewise.
15423 2013-03-20 Richard Biener <rguenther@suse.de>
15425 PR tree-optimization/56661
15426 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
15427 the result does not have to be distinct.
15429 2013-03-20 Richard Biener <rguenther@suse.de>
15431 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
15434 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15435 Steven Bosscher <steven@gcc.gnu.org>
15437 PR rtl-optimization/56605
15438 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
15440 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15443 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
15444 that require movd instead of movq.
15446 2013-03-20 Richard Biener <rguenther@suse.de>
15448 * tree-ssa-structalias.c (struct variable_info): Add pointer
15449 to the first field of an aggregate with sub-vars. Make
15450 this and the pointer to the next subfield its ID.
15451 (vi_next): New function.
15452 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
15453 storedanything_id, integer_id): Increment by one.
15454 (new_var_info, get_call_vi, lookup_call_clobber_vi,
15455 get_call_clobber_vi): Adjust.
15456 (solution_set_expand): Simplify and speedup.
15457 (solution_set_add): Inline into ...
15458 (set_union_with_increment): ... this. Adjust accordingly.
15459 (do_sd_constraint): Likewise.
15460 (do_ds_constraint): Likewise.
15461 (do_complex_constraint): Simplify.
15462 (build_pred_graph): Adjust.
15463 (solve_graph): Likewise. Simplify and speedup.
15464 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
15465 get_constraint_for_component_ref, get_constraint_for_1,
15466 first_vi_for_offset, first_or_preceding_vi_for_offset,
15467 create_function_info_for, create_variable_info_for_1,
15468 create_variable_info_for, intra_create_variable_infos): Adjust.
15469 (init_base_vars): Push NULL for ID zero.
15470 (compute_points_to_sets): Adjust.
15472 2013-03-20 Richard Biener <rguenther@suse.de>
15474 * cfgloop.c (verify_loop_structure): Streamline and avoid
15475 ICEing on corrupt loop tree.
15476 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
15479 2013-03-20 Richard Biener <rguenther@suse.de>
15481 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
15482 check whether an SSA update is needed.
15484 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
15486 * config/mips/constraints.md (T): Rename to...
15490 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
15491 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
15493 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15495 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
15496 (*subsi3_carryin_uxtw): Likewise.
15498 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15500 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
15501 (*rorsi3_insn_uxtw): Likewise.
15503 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15505 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
15506 (*extrsi5_insn_uxtw): Likewise.
15508 2013-03-19 Richard Biener <rguenther@suse.de>
15510 PR tree-optimization/56273
15511 * passes.c (init_optimization_passes): Move second VRP after DOM.
15513 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
15515 * config/i386/i386.md (*movti_internal): Merge from
15516 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
15517 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
15518 nox64 isa attributes.
15520 2013-03-18 Richard Biener <rguenther@suse.de>
15522 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
15524 (merge_node_constraints): Likewise.
15525 (build_succ_graph): Likewise.
15526 (valid_graph_edge): Inline into single caller.
15527 (unify_nodes): Likewise. Use bitmap_set_bit return value
15529 (scc_visit): Fix formatting and variable use.
15530 (do_sd_constraint): Use gcc_checking_assert.
15531 (do_ds_constraint): Likewise.
15532 (do_complex_constraint): Likewise.
15533 (condense_visit): Likewise. Cleanup.
15534 (dump_pred_graph): New function.
15535 (perform_var_substitution): Dump the pred-graph before
15536 variable substitution.
15537 (find_equivalent_node): Use gcc_checking_assert.
15538 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
15540 2013-03-18 Richard Biener <rguenther@suse.de>
15542 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
15543 Remove cond_expr_stmt_list argument and do not gimplify the
15545 (vect_loop_versioning): Adjust.
15546 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
15547 Cleanup to use less temporaries.
15548 (vect_create_data_ref_ptr): Cleanup.
15550 2013-03-18 Jakub Jelinek <jakub@redhat.com>
15552 PR tree-optimization/56635
15553 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
15554 require types_compatible_p types.
15556 2013-03-18 Nick Clifton <nickc@redhat.com>
15558 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
15559 spurious backslash.
15561 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
15562 Add missing line to comment describing function.
15564 2013-03-18 Richard Biener <rguenther@suse.de>
15566 PR tree-optimization/56210
15567 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
15568 Handle string / character search functions.
15569 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
15571 2013-03-18 Richard Biener <rguenther@suse.de>
15573 PR middle-end/56483
15574 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
15575 and implement properly.
15576 * gimple.h (gimple_cond_single_var_p): Remove.
15578 2013-03-18 Richard Biener <rguenther@suse.de>
15580 * tree-data-ref.h (find_data_references_in_loop): Declare.
15581 * tree-data-ref.c (get_references_in_stmt): Use a stack
15582 vector pre-allocated in the callers.
15583 (find_data_references_in_stmt): Adjust.
15584 (graphite_find_data_references_in_stmt): Likewise.
15585 (create_rdg_vertices): Likewise.
15586 (find_data_references_in_loop): Export.
15587 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
15588 Compute dependences here...
15589 (vect_analyze_data_refs): ...not here. When we encounter
15590 a non-vectorizable data reference in basic-block vectorization
15591 truncate the data reference vector. Do not bother to
15592 fixup data-dependence information for gather loads.
15593 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
15594 of data references, as reported.
15596 2013-03-18 Richard Biener <rguenther@suse.de>
15598 PR tree-optimization/3713
15599 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
15600 has_constants and expr.
15601 (stmt_has_constants): Properly valueize SSA names when deciding
15602 whether the stmt has constants.
15604 2013-03-18 Richard Biener <rguenther@suse.de>
15606 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
15607 whole function when there is nothing to do.
15608 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
15609 * tree-vectorizer.c (vectorize_loops): Update virtual and
15610 loop-closed SSA once.
15611 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
15613 2013-03-18 Richard Biener <rguenther@suse.de>
15615 PR middle-end/56113
15616 * domwalk.c (bb_postorder): New global static.
15617 (cmp_bb_postorder): New function.
15618 (walk_dominator_tree): Replace scheme imposing an order for
15619 visiting dominator sons by one sorting them at the time they
15620 are pushed on the stack.
15622 2013-03-18 Richard Biener <rguenther@suse.de>
15624 PR tree-optimization/39326
15625 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
15626 (struct mem_ref): Replace mem member with ao_ref typed member.
15627 (MEM_ANALYZABLE): Adjust.
15628 (memref_eq): Likewise.
15629 (mem_ref_alloc): Likewise.
15630 (gather_mem_refs_stmt): Likewise.
15631 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
15632 (execute_sm_if_changed_flag_set): Adjust.
15633 (execute_sm): Likewise.
15634 (ref_always_accessed_p): Likewise.
15635 (refs_independent_p): Likewise.
15636 (can_sm_ref_p): Likewise.
15638 2013-03-18 Jakub Jelinek <jakub@redhat.com>
15641 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
15642 return 1 even for !unsignedp.
15644 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15646 * config/i386/i386.md (isa): Add x64 and nox64.
15647 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
15648 (*pushtf): Enable *roF alternative for x64 isa only.
15649 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
15650 mode attribute of integer alternatives to DImode for TARGET_64BIT.
15651 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
15652 (*movtf_internal): Merge from *movtf_internal_rex64 and
15653 *movtf_internal_sse. Use x64 and nox64 isa attributes.
15654 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
15655 nox64 isa attributes.
15656 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
15657 nox64 isa attributes.
15658 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
15660 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15662 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15664 2013-03-17 Jakub Jelinek <jakub@redhat.com>
15667 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
15668 class names. Remove trailing comma after "ALL_REGS".
15670 2013-03-16 Jan Hubicka <jh@suse.cz>
15672 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
15673 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
15674 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
15675 of cgraph_get_create_node.
15676 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
15678 2013-03-16 Jason Merrill <jason@redhat.com>
15681 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
15682 with DW_AT_default_value.
15684 2013-03-16 Jakub Jelinek <jakub@redhat.com>
15686 * BASE-VER: Set to 4.9.0.
15688 2013-03-14 Andi Kleen <ak@linux.intel.com>
15691 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
15692 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
15693 Document _x* TSX intrinsics.
15695 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
15696 David Holsgrove <david.holsgrove@xilinx.com>
15698 * configure.ac: Add MicroBlaze TLS support detection.
15699 * configure: Regenerate.
15700 * config/microblaze/microblaze-protos.h
15701 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
15702 symbol_mentioned_p, label_mentioned_p): Add prototypes.
15703 * config/microblaze/microblaze.c (microblaze_address_type): Add
15704 ADDRESS_TLS and tls_reloc address types.
15705 (microblaze_address_info): Add tls_reloc.
15706 (TARGET_HAVE_TLS): Define.
15707 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
15708 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
15709 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
15710 load_tls_operand, microblaze_call_tls_get_addr,
15711 microblaze_legitimize_tls_address): New functions.
15712 (microblaze_classify_unspec): Handle UNSPEC_TLS.
15713 (get_base_reg): Use microblaze_tls_symbol_p.
15714 (microblaze_classify_address): Handle TLS.
15715 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
15716 label_mentioned_p and microblaze_tls_referenced_p.
15717 (microblaze_legitimize_address): Handle TLS.
15718 (microblaze_address_insns): Handle ADDRESS_TLS.
15719 (pic_address_needs_scratch): Handle TLS.
15720 (print_operand_address): Handle TLS.
15721 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
15722 (microblaze_expand_move): Handle TLS.
15723 (microblaze_legitimate_constant_p): Check
15724 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
15725 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
15726 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
15727 (PIC_OFFSET_TABLE_REGNUM): Set.
15728 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
15729 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
15730 (addsi3, movsi_internal2, movdf_internal): Update constraints
15731 * config/microblaze/predicates.md (arith_plus_operand): Define
15732 (move_operand): Redefine as move_src_operand,
15733 check microblaze_tls_referenced_p.
15735 2013-03-14 Ian Bolton <ian.bolton@arm.com>
15737 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
15738 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
15740 2013-03-14 Ian Bolton <ian.bolton@arm.com>
15742 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
15745 2013-03-14 Jakub Jelinek <jakub@redhat.com>
15747 PR tree-optimization/53265
15748 * common.opt (Waggressive-loop-optimizations): New option.
15749 * tree-ssa-loop-niter.c: Include tree-pass.h.
15750 (do_warn_aggressive_loop_optimizations): New function.
15751 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
15752 if number_of_latch_executions returned constant.
15753 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
15754 early. If number_of_latch_executions returned constant, set
15755 nb_iterations_upper_bound back to it.
15756 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
15758 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
15759 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
15761 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
15762 (MULTILIB_OSDIRNAMES): Set.
15763 * genmultilib: If defaultosdirname doesn't start with :: , set
15764 defaultosdirname2 instead, clear it and emit two . multilib_raw
15765 entries instead of just one.
15767 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
15769 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15770 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
15771 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15772 (SUBTARGET_OVERRIDE_OPTIONS): New.
15774 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
15777 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
15778 (musermode): Convert to Var(TARGET_USERMODE).
15779 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
15780 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
15781 * config/sh/sh.c (sh_option_override): Use
15782 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
15783 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
15785 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
15787 (udivsi3_i4_single, divsi3_i4_single): Use
15788 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
15790 2013-03-13 Dave Korn <dave.korn.cygwin@....>
15792 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
15795 2013-03-13 Richard Biener <rguenther@suse.de>
15797 PR tree-optimization/56608
15798 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
15799 calls when vectorizing basic-blocks.
15801 2013-03-13 Jakub Jelinek <jakub@redhat.com>
15804 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
15807 2013-03-12 Jakub Jelinek <jakub@redhat.com>
15809 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
15811 2013-03-11 Jan Hubicka <jh@suse.cz>
15814 * lto-streamer-out.c (output_symbol_p): Skip references from
15815 constructors of external variables.
15817 2013-03-11 Jan Hubicka <jh@suse.cz>
15819 PR middle-end/56571
15820 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
15822 * emit-rtl.c (verify_rtx_sharing): Likewise.
15823 (copy_insn_1): Likewise.
15824 * rtl.c (copy_rtx): Likewise.
15826 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
15829 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
15830 output_operand_lossage message.
15832 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
15835 * arm.c (shift_op): Validate RTL pattern on the fly.
15836 (arm_print_operand, case 'S'): Don't use shift_operator to validate
15839 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15842 * config/pa/pa.md (call_value): Check for calls to powf and direct to
15843 new call patterns that clobber %fr12.
15844 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
15845 split and postreload patterns.
15846 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
15847 registers %fr12 and %fr12R as call used.
15849 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
15851 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
15852 (canon_address, record_store, replace_read, check_mem_read_rtx,
15853 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
15854 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
15855 rest_of_handle_dse): Likewise.
15857 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
15859 PR middle-end/56524
15860 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
15862 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
15863 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
15864 (save_optabs_if_changed): Replace with...
15865 (init_tree_optimization_optabs): ...this.
15866 * optabs.c (save_optabs_if_changed): Rename to...
15867 (init_tree_optimization_optabs): ...this. Take the optimization node
15868 as argument. Do nothing if the base optabs are already correct.
15869 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
15870 to recompute optabs.
15871 * function.h (function): Remove optabs field.
15872 * function.c (invoke_set_current_function_hook): Call
15873 init_tree_optimization_optabs. Use the result to initialize
15876 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
15878 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
15879 if GTMA_HAS_NO_INSTRUMENTATION.
15880 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
15881 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
15882 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
15883 * gimple-pretty-print.c (dump_gimple_transaction): Handle
15884 GTMA_HAS_NO_INSTRUMENTATION.
15886 2013-03-08 Jakub Jelinek <jakub@redhat.com>
15888 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
15891 2013-03-08 Marek Polacek <polacek@redhat.com>
15892 Jakub Jelinek <jakub@redhat.com>
15894 PR tree-optimization/56478
15895 * predict.c (is_comparison_with_loop_invariant_p): Change the
15896 type of loop_step to tree.
15897 (predict_loops): Adjust.
15898 (predict_iv_comparison): Perform the computations on double_ints.
15900 2013-03-08 Richard Biener <rguenther@suse.de>
15902 PR tree-optimization/56570
15903 * tree-cfg.c (verify_expr_location_1): Verify locations for
15905 * tree-sra.c (create_access_replacement): Strip locations
15906 from DECL_DEBUG_EXPRs.
15908 2013-03-08 Richard Biener <rguenther@suse.de>
15910 * tree-inline.c (expand_call_inline): Do not associate
15911 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15912 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15914 2013-03-08 Richard Biener <rguenther@suse.de>
15916 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15917 or block changes with -Og. Fix for location / block encoding
15918 changes and PHI arguments with locations.
15920 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15922 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15924 (struct output_info): Likewise.
15925 (register_overhead): Remove bad gcc_assert.
15926 (bitmap_find_bit): If there is only a single bitmap element, do not
15927 count a miss as a search.
15928 (print_statistics): Update for counter type changes.
15929 (dump_bitmap_statistics): Likewise. Print headers such that they
15930 are properly lined up with the printed counters.
15932 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15934 PR tree-optimization/56559
15935 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15936 check that it has only a single use.
15938 2013-03-07 Richard Biener <rguenther@suse.de>
15940 * doc/invoke.texi (fwhole-program): Discourage use in combination
15943 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15945 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15947 PR tree-optimization/56539
15948 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15949 instead of GSI_CONTINUE_LINKING as last argument to
15950 force_gimple_operand_gsi. Adjust function comment.
15952 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15955 PR middle-end/56548
15956 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15957 promoted mode, convert the result back to the original mode.
15959 2013-03-06 Richard Biener <rguenther@suse.de>
15961 PR middle-end/56294
15962 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15963 (insert_updated_phi_nodes_compare_uids): New function.
15964 (update_ssa): Sort symbols_to_rename after UID before
15965 traversing it to insert PHI nodes.
15967 2013-03-06 Richard Biener <rguenther@suse.de>
15969 PR middle-end/50494
15970 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15971 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
15974 2013-02-13 Richard Biener <rguenther@suse.de>
15977 * varasm.c (output_constant_def_1): Get the decl representing
15978 the constant as argument.
15979 (output_constant_def): Wrap output_constant_def_1.
15980 (make_decl_rtl): Use output_constant_def_1 with the decl
15981 representing the constant.
15982 (build_constant_desc): Optionally re-use a decl already
15983 representing the constant.
15984 (tree_output_constant_def): Adjust.
15986 2013-03-06 Joey Ye <joey.ye@arm.com>
15989 * gcc.c (convert_white_space): New function.
15990 (main): Handles white space in function name.
15992 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
15995 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
15996 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
15997 to SH_DIV_CALL_TABLE for TARGET_SH2.
15998 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
16000 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
16001 call-table options.
16003 2013-03-05 Sterling Augustine <saugustine@google.com>
16004 Cary Coutant <ccoutant@google.com>
16007 * dwarf2out.c (resolve_addr): Don't call
16008 remove_loc_list_addr_table_entries a second time for the same
16011 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16014 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
16015 (avoid_complex_debug_insns): New function.
16016 (expand_debug_locations): Call it.
16018 PR rtl-optimization/56484
16019 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
16020 lifetimes of hard registers on small register class machines.
16022 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16024 * config/microblaze/microblaze-protos.h: Rename
16025 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
16026 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
16028 (microblaze_fast_interrupt_function_p): New function.
16029 (microblaze_is_interrupt_handler): Rename to
16030 microblaze_is_interrupt_variant and add fast_interrupt check.
16031 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
16032 (save_restore_insns): Likewise.
16033 (compute_frame_size): Likewise.
16034 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
16035 (microblaze_globalize_label): Likewise.
16036 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
16037 * config/microblaze/microblaze.md: Use wrapper
16038 microblaze_is_interrupt_variant.
16040 2013-03-05 Kai Tietz <ktietz@redhat.com>
16042 * sdbout.c (sdbout_one_type): Switch to current function's section
16043 supporting cold/hot.
16045 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16047 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
16050 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16052 PR middle-end/56461
16053 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
16054 if VALGRIND_GET_VBITS is defined, temporarily make object
16055 memory all defined, and restore previous valgrind addressability
16056 and definability afterwards. Free this_object at the end.
16058 PR middle-end/56461
16059 * lra.c (lra): Call lra_clear_live_ranges if live_p,
16060 right before calling lra_create_live_ranges, also call it
16061 when clearing live_p. Only call lra_clear_live_ranges
16062 at the end if live_p.
16064 PR middle-end/56461
16065 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
16067 2013-03-05 Richard Biener <rguenther@suse.de>
16069 PR tree-optimization/56521
16070 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
16073 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
16076 * except.h (remove_unreachable_eh_regions): New prototype.
16077 * except.c (remove_eh_handler_splicer): New function, split out
16078 of remove_eh_handler.
16079 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
16080 warning about running it on many EH regions one at a time.
16081 (remove_unreachable_eh_regions_worker): New function, walk the
16082 EH tree in depth-first order and remove non-marked regions.
16083 (remove_unreachable_eh_regions): New function.
16084 * tree-eh.c (mark_reachable_handlers): New function, split out
16085 from remove_unreachable_handlers.
16086 (remove_unreachable_handlers): Use mark_reachable_handlers and
16087 remove_unreachable_eh_regions.
16088 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
16089 and remove_unreachable_eh_regions.
16091 2013-03-05 Richard Biener <rguenther@suse.de>
16093 PR middle-end/56525
16094 * loop-init.c (fix_loop_structure): Remove loops in two stages,
16095 not freeing them until the end.
16097 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16099 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
16101 2013-03-05 Richard Biener <rguenther@suse.de>
16103 PR tree-optimization/56270
16104 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
16105 of loads after scheduling an SLP instance.
16107 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16109 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
16111 (check_gcc_parallelize): Run guality.exp as a separate job from
16112 vect.exp with unsorted.exp and $(dg_target_exps) separately from
16113 struct-layout-1.exp with stackalign.exp.
16115 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
16117 PR middle-end/56461
16118 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
16119 load_index sbitmap even if some bit in it isn't set.
16121 PR middle-end/56461
16122 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
16123 (discover_iteration_bound_by_body_walk): Change queues to
16124 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
16125 spelling in comment. Call safe_push on queues[bound_index] directly.
16126 Release queues[queue_index] in every iteration unconditionally.
16127 Release bounds vector.
16129 PR middle-end/56461
16130 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
16131 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
16132 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
16135 2013-03-05 Richard Biener <rguenther@suse.de>
16138 * tree-inline.c (remap_blocks_to_null): New function.
16139 (expand_call_inline): When expanding a call stmt without
16140 an associated block inline remap all callee blocks to NULL.
16142 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16144 PR rtl-optimization/56494
16145 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
16146 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
16147 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
16149 PR middle-end/56461
16150 * sel-sched-ir.c (free_sched_pools): Release
16151 succs_info_pool.stack[succs_info_pool.max_top] vectors too
16152 if succs_info_pool.max_top isn't -1.
16155 * opts.c (opts_obstack, opts_concat): Moved to...
16156 * opts-common.c (opts_obstack, opts_concat): ... here.
16158 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16160 PR middle-end/56461
16161 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
16163 2013-03-04 Martin Jambor <mjambor@suse.cz>
16165 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
16166 all appropriate places.
16168 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
16170 PR tree-optimization/56424
16171 * ipa-split.c (split_function): Do not set the RSO flag if result is
16172 not by reference and its type is a register type.
16174 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
16176 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
16177 (microblaze_legitimate_pic_operand): Likewise
16178 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
16179 new function microblaze_legitimate_pic_operand
16180 * config/microblaze/microblaze-protos.h
16181 (microblaze_legitimate_pic_operand): Declare.
16183 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16185 * config/microblaze/predicates.md (call_insn_simple_operand):
16186 New predicate for supported rtx code types.
16187 * config/microblaze/microblaze.md (call_internal1): Use
16188 call_insn_simple_operand predicate.
16190 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16192 PR middle-end/56461
16193 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
16194 partitions.ordered_remove.
16196 PR middle-end/56461
16197 * tree-vect-stmts.c (vectorizable_conversion): Don't call
16198 vec_oprnds0.create (1) for modifier == NONE.
16200 PR middle-end/56461
16201 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
16202 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
16203 vec_oprnds1 right before pushing anything to it for
16206 PR middle-end/56461
16207 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
16208 set nbbs to 0 instead of having separate code path.
16209 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
16210 instead of false as last argument if returning NULL.
16212 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
16214 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
16215 the attribute is now called "target" instead of "option".
16216 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
16217 * doc/tm.texi.in (Target Attributes): Likewise document the correct
16218 attribute/pragma name for TARGET_OPTION_VALID_P and
16219 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
16220 * doc/tm.texi: Regenerated.
16222 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
16224 * config/microblaze/microblaze.c:
16225 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
16226 * config/microblaze/microblaze.h: Add -mxl-reorder to
16228 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
16229 instructions emitted if TARGET_REORDER.
16230 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
16231 or 0 for -m/-mno case, but initialises as 2 to detect default use case
16234 2013-03-01 Xinliang David Li <davidxl@google.com>
16236 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
16239 2013-03-01 Jakub Jelinek <jakub@redhat.com>
16241 PR middle-end/56461
16242 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
16243 vector even when returning true. Fix up function comment formatting.
16245 PR middle-end/56461
16246 * ira-build.c (ira_loop_nodes_count): New variable.
16247 (create_loop_tree_nodes): Initialize it.
16248 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
16250 PR middle-end/56461
16251 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
16252 method on dr_chain and result_chain.
16253 * tree-vect-stmts.c (vectorizable_store): Only call
16254 result_chain.create if j == 0.
16256 PR middle-end/56461
16257 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
16258 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
16259 before overwriting it.
16261 2013-03-01 Tobias Burnus <burnus@net-b.de>
16263 * doc/extended.texi (C Extensions): Change order in @menu
16265 (Other MIPS Built-in Functions): Move last MIPS entry before
16266 "picoChip Built-in Functions".
16267 (SH Built-in Functions): Move after RX Built-in Functions.
16268 * doc/gcc.texi (Introduction): Change order in @menu
16270 * doc/md.texi (Constraints): Ditto.
16271 * gty.texi (Type Information): Ditto.
16272 (User-provided marking routines for template types): Make
16274 * doc/invoke.texi (AArch64 Options): Move before
16275 "Adapteva Epiphany Options".
16277 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
16278 Jakub Jelinek <jakub@redhat.com>
16281 * asan.c (gate_asan): Lookup no_sanitize_address instead of
16282 no_address_safety_analysis attribute.
16283 * doc/extend.texi (no_address_safety_attribute): Rename to
16284 no_sanitize_address attribute, mention no_address_safety_analysis
16285 attribute as deprecated alias.
16287 2013-02-28 Jakub Jelinek <jakub@redhat.com>
16289 PR middle-end/56461
16290 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
16291 type to vec<vec<tree> > *.
16292 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
16293 to be vec<tree> instead of vec<tree> *, set vec_defs
16294 to vNULL and call vec_defs.create (number_of_vects), adjust other
16296 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
16297 vectorizable_condition): Adjust vect_get_slp_defs callers.
16299 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16301 * config/aarch64/aarch64.c
16302 (aarch64_float_const_representable): Remove unused variable.
16304 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16306 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
16308 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16310 * config/aarch64/aarch64-builtins.c
16311 (aarch64_init_simd_builtins): Make static.
16313 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16315 * config/aarch64/aarch64.c
16316 (aarch64_simd_make_constant): Make static.
16318 2013-02-28 Martin Jambor <mjambor@suse.cz>
16320 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
16321 with no initialization to the RHS of debug statements.
16323 2013-02-28 Martin Jambor <mjambor@suse.cz>
16325 PR tree-optimization/56294
16326 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
16328 (get_access_replacement): Do not call create_access_replacement.
16329 Assert a replacement exists.
16330 (get_repl_default_def_ssa_name): Create the replacement declaration
16333 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16335 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
16336 final_end_function.
16338 2013-02-28 Marek Polacek <polacek@redhat.com>
16340 PR rtl-optimization/56466
16341 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
16342 if we're changing a loop.
16343 (peel_loops_completely): Likewise.
16345 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
16348 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
16350 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16353 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
16354 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
16355 INTX_FTYPE_FX, FX_FTYPE_INTX.
16356 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
16358 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16360 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
16361 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
16362 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
16363 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
16364 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
16365 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
16366 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
16367 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
16368 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
16369 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
16370 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
16371 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
16372 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
16373 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
16374 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
16375 (avrxmega6): Increase max flash segments from 5 to 6.
16376 * config/avr/t-multilib: Regenerate.
16377 * config/avr/avr-tables.opt: Regenerate.
16378 * doc/avr-mmcu.texi: Regenerate.
16380 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16382 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
16383 (avr_device_to_arch): Rename to avr_device_to_ld.
16384 (avr_device_to_as): New prototype.
16385 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
16386 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
16387 * config/avr/driver-avr.c (avr_device_to_as): New.
16388 (avr_device_to_arch): Rename to avr_device_to_ld.
16390 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16392 PR middle-end/56461
16393 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
16394 method on dr_chain and result_chain.
16396 PR middle-end/56461
16397 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
16398 pointer_set_destroy on not_executed_last_iteration.
16400 PR middle-end/56461
16401 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
16403 PR middle-end/56461
16404 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
16405 FOR_EACH_DEFINED_FUNCTION when freeing state.
16407 PR middle-end/56461
16408 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
16410 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
16413 PR middle-end/56461
16414 * ipa-cp.c (decide_whether_version_node): Call vec_free on
16415 known_aggs[i].items and release known_aggs vector.
16417 PR middle-end/56461
16418 * ipa-reference.c (propagate): Free node_info even for alias nodes.
16420 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16422 * config/microblaze/microblaze.c (microblaze_emit_compare):
16423 Use xor for EQ/NE comparisions.
16424 * config/microblaze/microblaze.md (cstoresf4): Add constraints
16425 (cbranchsf4): Adjust operator to comparison_operator.
16427 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16429 PR middle-end/56461
16430 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
16432 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
16433 vec_safe_push, always update *slot.
16434 (redirect_edge_var_map_clear): Use vec_free.
16435 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
16436 (free_var_map_entry): Use vec_free.
16437 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
16438 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
16440 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
16442 PR middle-end/45472
16443 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
16444 when the may_trap_p bit of the exprs being merged differs.
16445 Reorder tests for speculativeness in the logical and operator.
16447 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16449 * incpath.c (add_standard_paths): Use reconcat instead of concat
16450 where appropriate and avoid leaking memory.
16452 * opts.h: Include obstack.h.
16453 (opts_concat): New prototype.
16454 (opts_obstack): New declaration.
16455 * opts.c (opts_concat): New function.
16456 (opts_obstack): New variable.
16457 (init_options_struct): Call gcc_init_obstack on opts_obstack.
16458 (finish_options): Use opts_concat instead of concat
16459 and XOBNEWVEC instead of XNEWVEC.
16460 * opts-common.c (generate_canonical_option, decode_cmdline_option,
16461 generate_option): Likewise.
16462 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
16463 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
16466 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
16467 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
16469 2013-02-26 Jakub Jelinek <jakub@redhat.com>
16471 PR middle-end/56461
16472 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
16474 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
16476 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
16477 (arm_block_move_unaligned_straight): Likewise.
16478 (arm_adjust_block_mem): Likewise.
16480 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
16483 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
16484 temp, cond and label.
16485 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
16488 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
16489 * config/c6x/c6x.h (dbx_register_map): Update declaration.
16492 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
16493 of prologue/epilogue functions.
16496 * config/tilegx/tilegx.c (tilegx_expand_prologue):
16497 Remove unused variable cfa_offset.
16498 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
16501 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
16502 type promotion to unsigned.
16505 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
16506 for HOST_WIDE_INT of 32 bit / same size as int.
16507 (arm_block_move_unaligned_straight): Likewise.
16508 (arm_adjust_block_mem): Likewise.
16511 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
16514 2013-02-26 Marek Polacek <polacek@redhat.com>
16516 PR tree-optimization/56426
16517 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
16519 2013-02-26 Richard Biener <rguenther@suse.de>
16522 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
16523 unused variable loops.
16525 2013-02-26 Jakub Jelinek <jakub@redhat.com>
16527 PR tree-optimization/56448
16528 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
16529 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
16530 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
16531 later operands of the references, or even first operand for
16532 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
16534 PR tree-optimization/56443
16535 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
16536 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
16537 to type_for_mode langhook.
16539 2013-02-25 Matt Turner <mattst88@gmail.com>
16541 * doc/invoke.texi: Document r4700.
16543 2013-02-25 Richard Biener <rguenther@suse.de>
16545 PR tree-optimization/56175
16546 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
16548 (simplify_bitwise_binary): ... here. Also guard the conversion
16549 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
16551 2013-02-25 Catherine Moore <clm@codesourcery.com>
16554 2013-02-24 Catherine Moore <clm@codesourcery.com>
16555 Maciej W. Rozycki <macro@codesourcery.com>
16556 Tom de Vries <tom@codesourcery.com>
16557 Nathan Sidwell <nathan@codesourcery.com>
16558 Iain Sandoe <iain@codesourcery.com>
16559 Nathan Froyd <froydnj@codesourcery.com>
16560 Chao-ying Fu <fu@mips.com>
16562 * doc/extend.texi: (micromips, nomicromips, nocompression):
16563 Document new function attributes.
16564 * doc/invoke.texi (minterlink-compressed, mmicromips,
16565 m14k, m14ke, m14kec): Document new options.
16566 (minterlink-mips16): Update documentation.
16567 * doc/md.texi (ZC, ZD): Document new constraints.
16568 * configure.ac (gcc_cv_as_micromips): Check if linker
16569 supports the .set micromips directive.
16570 * configure: Regenerate.
16571 * config.in: Regenerate.
16572 * config/mips/mips-tables.opt: Regenerate.
16573 * config/mips/micromips.md: New file.
16574 * constraints.md (ZC, AD): New constraints.
16575 * config/mips/predicates.md (movep_src_register): New predicate.
16576 (movep_src_operand): New predicate.
16577 (non_volatile_mem_operand): New predicate.
16578 * config/mips/mips.md (multimem): New type.
16579 (length): Differentiate between 17-bit and 18-bit branch offsets.
16580 (MOVEP1, MOVEP2): New mode iterator.
16581 (mov_<load>l): Use ZC constraint.
16582 (mov_<load>r): Likewise.
16583 (mov_<store>l): Likewise.
16584 (mov_<store>r): Likewise.
16585 (*branch_equality<mode>_inverted): Add microMIPS support.
16586 (*branch_equality<mode>): Likewise.
16587 (*jump_absolute): Likewise.
16588 (indirect_jump_<mode>): Likewise.
16589 (tablejump_<mode>): Likewise.
16590 (<optab>_internal): Likewise.
16591 (sibcall_internal): Likewise.
16592 (sibcall_value_internal): Likewise.
16593 (prefetch): Use constraint ZD.
16594 * config/mips/mips.opt (minterlink-compressed): New option.
16595 (minterlink-mips16): Now an alias for minterlink-compressed.
16596 (mmicromips): New option.
16597 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16598 (compare_and_swap_12): Likewise.
16599 (sync_add<mode>): Likewise.
16600 (sync_<optab>_12): Likewise.
16601 (sync_old_<optab>_12): Likewise.
16602 (sync_new_<optab>_12): Likewise.
16603 (sync_nand_12): Likewise.
16604 (sync_old_nand_12): Likewise.
16605 (sync_new_nand_12): Likewise.
16606 (sync_sub<mode>): Likewise.
16607 (sync_old_add<mode>): Likewise.
16608 (sync_old_sub<mode>): Likewise.
16609 (sync_new_add<mode>): Likewise.
16610 (sync_new_sub<mode>): Likewise.
16611 (sync_<optab><mode>): Likewise.
16612 (sync_old_<optab><mode>): Likewise.
16613 (sync_new_<optab><mode>): Likewise.
16614 (sync_nand<mode>): Likewise.
16615 (sync_old_nand<mode>): Likewise.
16616 (sync_new_nand<mode>): Likewise.
16617 (sync_lock_test_and_set<mode>): Likewise.
16618 (test_and_set_12): Likewise.
16619 (atomic_compare_and_swap<mode>): Likewise.
16620 (atomic_exchange<mode>_llsc): Likewise.
16621 (atomic_fetch_add<mode>_llsc): Likewise.
16622 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16623 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16624 (umips_save_restore_pattern_p): Likewise.
16625 (umips_load_store_pair_p): Likewise.
16626 (umips_output_load_store_pair): Likewise.
16627 (umips_movep_target_p): Likewise.
16628 (umips_12bit_offset_address_p): Likewise.
16629 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16630 (mips_base_mips16): Rename this...
16631 (mips_base_compression_flags): ...to this. Update all uses.
16632 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16633 (mips_mips16_decl_p): Delete.
16634 (mips_nomips16_decl_p): Delete.
16635 (mips_get_compress_on_flags): New function.
16636 (mips_get_compress_off_flags): New function.
16637 (mips_get_compress_mode): New function.
16638 (mips_get_compress_on_name): New function.
16639 (mips_get_compress_off_name): New function.
16640 (mips_insert_attributes): Support multiple compression types.
16641 (mips_merge_decl_attributes): Likewise.
16642 (umips_12bit_offset_address_p): New function.
16643 (mips_start_function_definition): Emit .set micromips directive.
16644 (mips_call_may_need_jalx_p): New function.
16645 (mips_function_ok_for_sibcall): Add microMIPS support.
16646 (mips_print_operand_punctuation): Support short delay slots and
16648 (umips_swm_mask, umips_swm_encoding): New.
16649 (umips_build_save_restore): New function.
16650 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16651 (was_mips16_p): Remove.
16652 (old_compression_mode): New.
16653 (mips_set_compression_mode): New function.
16654 (mips_set_current_function): Add microMIPS support.
16655 (mips_option_override): Likewise.
16656 (umips_save_restore_pattern_p): New function.
16657 (umips_output_save_restore): New function.
16658 (umips_load_store_pair_p_1): New function.
16659 (umips_load_store_pair_p): New function.
16660 (umips_output_load_store_pair_1): New function.
16661 (umips_output_load_store_pair): New function.
16662 (umips_movep_target_p) New function.
16663 (mips_prepare_pch_save): Add microMIPS support.
16664 * config/mips/mips.h (TARGET_COMPRESSION): New.
16665 (TARGET_CPU_CPP_BUILTINS): Update macro
16666 to use new compression flags and to support microMIPS.
16667 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16668 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16669 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16670 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16671 (ASM_SPEC): Support mmicromips and mno-micromips.
16672 (M16STORE_REG_P): New macro.
16673 (MIPS_CALL): Support TARGET_MICROMIPS.
16674 (MICROMIPS_J): New macro.
16675 (mips_base_mips16): Rename this...
16676 (mips_base_compression_flags): ...to this.
16677 (UMIPS_12BIT_OFFSET_P): New macro.
16678 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16679 (MULTILIB_DIRNAMES): Likewise.
16681 2013-02-25 Tom de Vries <tom@codesourcery.com>
16683 PR rtl-optimization/56131
16684 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
16685 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
16686 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
16688 2013-02-25 Tobias Burnus <burnus@net-b.de>
16690 * doc/invoke.texi (-fsanitize=): Move from optimization
16691 to debugging options.
16693 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
16695 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
16697 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
16698 Alexander Monakov <amonakov@ispras.ru>
16700 PR middle-end/56077
16701 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
16702 flush pending lists also on non-jumps. Adjust comment.
16704 2013-02-24 Catherine Moore <clm@codesourcery.com>
16705 Maciej W. Rozycki <macro@codesourcery.com>
16706 Tom de Vries <tom@codesourcery.com>
16707 Nathan Sidwell <nathan@codesourcery.com>
16708 Iain Sandoe <iain@codesourcery.com>
16709 Nathan Froyd <froydnj@codesourcery.com>
16710 Chao-ying Fu <fu@mips.com>
16712 * doc/extend.texi: (micromips, nomicromips, nocompression):
16713 Document new function attributes.
16714 * doc/invoke.texi (minterlink-compressed, mmicromips,
16715 m14k, m14ke, m14kec): Document new options.
16716 (minterlink-mips16): Update documentation.
16717 * doc/md.texi (ZC, ZD): Document new constraints.
16718 * configure.ac (gcc_cv_as_micromips): Check if linker
16719 supports the .set micromips directive.
16720 * configure: Regenerate.
16721 * config.in: Regenerate.
16722 * config/mips/mips-tables.opt: Regenerate.
16723 * config/mips/micromips.md: New file.
16724 * constraints.md (ZC, AD): New constraints.
16725 * config/mips/predicates.md (movep_src_register): New predicate.
16726 (movep_src_operand): New predicate.
16727 (non_volatile_mem_operand): New predicate.
16728 * config/mips/mips.md (multimem): New type.
16729 (length): Differentiate between 17-bit and 18-bit branch offsets.
16730 (MOVEP1, MOVEP2): New mode iterator.
16731 (mov_<load>l): Use ZC constraint.
16732 (mov_<load>r): Likewise.
16733 (mov_<store>l): Likewise.
16734 (mov_<store>r): Likewise.
16735 (*branch_equality<mode>_inverted): Add microMIPS support.
16736 (*branch_equality<mode>): Likewise.
16737 (*jump_absolute): Likewise.
16738 (indirect_jump_<mode>): Likewise.
16739 (tablejump_<mode>): Likewise.
16740 (<optab>_internal): Likewise.
16741 (sibcall_internal): Likewise.
16742 (sibcall_value_internal): Likewise.
16743 (prefetch): Use constraint ZD.
16744 * config/mips/mips.opt (minterlink-compressed): New option.
16745 (minterlink-mips16): Now an alias for minterlink-compressed.
16746 (mmicromips): New option.
16747 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16748 (compare_and_swap_12): Likewise.
16749 (sync_add<mode>): Likewise.
16750 (sync_<optab>_12): Likewise.
16751 (sync_old_<optab>_12): Likewise.
16752 (sync_new_<optab>_12): Likewise.
16753 (sync_nand_12): Likewise.
16754 (sync_old_nand_12): Likewise.
16755 (sync_new_nand_12): Likewise.
16756 (sync_sub<mode>): Likewise.
16757 (sync_old_add<mode>): Likewise.
16758 (sync_old_sub<mode>): Likewise.
16759 (sync_new_add<mode>): Likewise.
16760 (sync_new_sub<mode>): Likewise.
16761 (sync_<optab><mode>): Likewise.
16762 (sync_old_<optab><mode>): Likewise.
16763 (sync_new_<optab><mode>): Likewise.
16764 (sync_nand<mode>): Likewise.
16765 (sync_old_nand<mode>): Likewise.
16766 (sync_new_nand<mode>): Likewise.
16767 (sync_lock_test_and_set<mode>): Likewise.
16768 (test_and_set_12): Likewise.
16769 (atomic_compare_and_swap<mode>): Likewise.
16770 (atomic_exchange<mode>_llsc): Likewise.
16771 (atomic_fetch_add<mode>_llsc): Likewise.
16772 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16773 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16774 (umips_save_restore_pattern_p): Likewise.
16775 (umips_load_store_pair_p): Likewise.
16776 (umips_output_load_store_pair): Likewise.
16777 (umips_movep_target_p): Likewise.
16778 (umips_12bit_offset_address_p): Likewise.
16779 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16780 (mips_base_mips16): Rename this...
16781 (mips_base_compression_flags): ...to this. Update all uses.
16782 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16783 (mips_mips16_decl_p): Delete.
16784 (mips_nomips16_decl_p): Delete.
16785 (mips_get_compress_on_flags): New function.
16786 (mips_get_compress_off_flags): New function.
16787 (mips_get_compress_mode): New function.
16788 (mips_get_compress_on_name): New function.
16789 (mips_get_compress_off_name): New function.
16790 (mips_insert_attributes): Support multiple compression types.
16791 (mips_merge_decl_attributes): Likewise.
16792 (umips_12bit_offset_address_p): New function.
16793 (mips_start_function_definition): Emit .set micromips directive.
16794 (mips_call_may_need_jalx_p): New function.
16795 (mips_function_ok_for_sibcall): Add microMIPS support.
16796 (mips_print_operand_punctuation): Support short delay slots and
16798 (umips_swm_mask, umips_swm_encoding): New.
16799 (umips_build_save_restore): New function.
16800 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16801 (was_mips16_p): Remove.
16802 (old_compression_mode): New.
16803 (mips_set_compression_mode): New function.
16804 (mips_set_current_function): Add microMIPS support.
16805 (mips_option_override): Likewise.
16806 (umips_save_restore_pattern_p): New function.
16807 (umips_output_save_restore): New function.
16808 (umips_load_store_pair_p_1): New function.
16809 (umips_load_store_pair_p): New function.
16810 (umips_output_load_store_pair_1): New function.
16811 (umips_output_load_store_pair): New function.
16812 (umips_movep_target_p) New function.
16813 (mips_prepare_pch_save): Add microMIPS support.
16814 * config/mips/mips.h (TARGET_COMPRESSION): New.
16815 (TARGET_CPU_CPP_BUILTINS): Update macro
16816 to use new compression flags and to support microMIPS.
16817 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16818 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16819 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16820 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16821 (ASM_SPEC): Support mmicromips and mno-micromips.
16822 (M16STORE_REG_P): New macro.
16823 (MIPS_CALL): Support TARGET_MICROMIPS.
16824 (MICROMIPS_J): New macro.
16825 (mips_base_mips16): Rename this...
16826 (mips_base_compression_flags): ...to this.
16827 (UMIPS_12BIT_OFFSET_P): New macro.
16828 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16829 (MULTILIB_DIRNAMES): Likewise.
16831 2013-02-24 Jakub Jelinek <jakub@redhat.com>
16834 * target-globals.c (save_target_globals): For init_reg_sets and
16835 target_reinit remporarily set this_fn_optabs to this_target_optabs.
16837 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
16839 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
16840 * config/aarch64/t-aarch64
16841 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
16843 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
16845 PR inline-asm/56148
16846 * lra-constraints.c (process_alt_operands): Reload operand
16847 conflicting with earlier clobber only if no more other conflicting
16850 2013-02-22 Jakub Jelinek <jakub@redhat.com>
16853 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
16854 if not linking a shared library.
16856 2013-02-22 Seth LaForge <sethml@google.com>
16858 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
16860 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
16862 * config/arm/arm.md (split for extendsidi): Update condition.
16863 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
16864 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
16865 (qhs_zextenddi_cstr): Likewise.
16867 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16869 PR middle-end/56420
16870 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
16871 avoid signed wrapping.
16872 (expand_mult): Handle properly multiplication by
16873 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
16874 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
16875 in the compiler if coeff is HOST_WIDE_INT_MIN.
16876 (expand_divmod): Don't make ext_op1 static, change it's type to
16877 uhwi. Avoid undefined behavior in -INTVAL (op1).
16879 PR rtl-optimization/50339
16880 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
16882 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
16883 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
16884 into splitting_ashiftrt field.
16885 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
16887 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
16890 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
16892 PR middle-end/56108
16893 * trans-mem.c (execute_tm_mark): Do not expand transactions that
16894 are sure to go irrevocable.
16896 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
16898 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16899 scalars are valid operands.
16901 2013-02-21 Martin Jambor <mjambor@suse.cz>
16903 PR tree-optimization/56310
16904 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16905 only matching indices and non-negative final offsets.
16906 (intersect_aggregates_with_edge): Pass src_idx to
16907 agg_replacements_to_vector. Pass src_idx insstead of index to
16908 intersect_with_agg_replacements.
16910 2013-02-21 Martin Jambor <mjambor@suse.cz>
16912 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16913 instead of hard-wired defaults.
16915 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16917 * doc/invoke.texi (MIPS Options): Update documentation of the
16918 floating-point multiply-accumulate instruction restrictions.
16920 2013-02-21 Kostya Serebryany <kcc@google.com>
16922 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16923 asan_shadow_offset on x86_64 linux.
16925 2013-02-21 Richard Biener <rguenther@suse.de>
16927 PR tree-optimization/56415
16929 2013-02-11 Richard Biener <rguenther@suse.de>
16931 PR tree-optimization/56273
16932 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16935 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16938 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16941 PR inline-asm/56405
16942 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16943 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16945 2013-02-20 Jan Hubicka <jh@suse.cz>
16947 PR tree-optimization/56265
16948 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16949 when target is referenced for first time.
16951 2013-02-20 Richard Biener <rguenther@suse.de>
16953 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16954 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16955 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16956 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16957 not return anything.
16958 (rename_ssa_copies): Do not remove unused locals.
16959 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16960 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16961 * passes.c (execute_function_todo): Do not schedule unused locals
16962 removal if cleanup_tree_cfg did something.
16963 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16964 about the number of removed locals.
16966 2013-02-20 Richard Biener <rguenther@suse.de>
16968 PR tree-optimization/56398
16969 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16971 2013-02-20 Martin Jambor <mjambor@suse.cz>
16973 PR tree-optimization/55334
16974 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
16975 restricted pointers to arrays.
16977 2013-02-20 Richard Biener <rguenther@suse.de>
16978 Jakub Jelinek <jakub@redhat.com>
16980 PR tree-optimization/56396
16981 * tree-ssa-ccp.c (n_const_val): New static variable.
16982 (get_value): Return NULL for SSA names we don't have a lattice
16984 (ccp_initialize): Initialize n_const_val.
16985 * tree-ssa-copy.c (n_copy_of): New static variable.
16986 (init_copy_prop): Initialize n_copy_of.
16987 (get_value): Return NULL_TREE for SSA names we don't have a
16990 2013-02-20 Martin Jambor <mjambor@suse.cz>
16992 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
16994 2013-02-20 Richard Biener <rguenther@suse.de>
16996 * genpreds.c (write_lookup_constraint): Do not compare first
16997 letter of the constraint again.
16999 2013-02-20 Richard Biener <rguenther@suse.de>
17001 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
17003 (get_use_iv_cost): Terminate hashtable walk when coming across
17006 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
17008 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
17009 reassociation for avx2 targets.
17011 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17013 * config/microblaze/microblaze.c: microblaze_has_clz = 0
17014 Add version check for v8.10.a to enable microblaze_has_clz
17015 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
17016 version and TARGET_PATTERN_COMPARE check
17017 * config/microblaze/microblaze.md: New clzsi2 instruction
17019 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17021 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
17022 function before branching.
17024 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17026 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
17028 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
17030 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17032 PR middle-end/55889
17033 * sel-sched.c: Include ira.h.
17034 (implicit_clobber_conflict_p): New function.
17035 (moveup_expr): Use it.
17036 * Makefile.in (sel-sched.o): Depend on ira.h.
17038 2013-02-19 Richard Biener <rguenther@suse.de>
17040 PR tree-optimization/56384
17041 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
17042 (vn_hash_type): Split out from ...
17043 (vn_hash_constant_with_type): ... here.
17044 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
17045 (vn_phi_eq): Compare types from vn_phi_s structure.
17046 (vn_phi_lookup): Populate vn_phi_s type.
17047 (vn_phi_insert): Likewise.
17049 2013-02-19 Jakub Jelinek <jakub@redhat.com>
17051 PR tree-optimization/56350
17052 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
17053 if haven't found reduction or nested cycle operand, rather than
17054 asserting we must find it.
17056 PR tree-optimization/56381
17057 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
17060 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
17061 Jakub Jelinek <jakub@redhat.com>
17064 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
17065 (swap_optab_enable): Same.
17066 (init_all_optabs): Use argument instead of global.
17067 * tree.h (struct tree_optimization_option): New field target_optabs.
17068 * expr.h (init_all_optabs): Add argument to prototype.
17069 (TREE_OPTIMIZATION_OPTABS): New.
17070 (save_optabs_if_changed): Protoize.
17071 * optabs.h: Declare this_fn_optabs.
17072 * optabs.c (save_optabs_if_changed): New.
17073 Declare this_fn_optabs.
17074 (init_optabs): Add argument to init_all_optabs() call.
17075 * function.c (invoke_set_current_function_hook): Handle per
17077 * function.h (struct function): New field optabs.
17078 * config/mips/mips.c (mips_set_mips16_mode): Handle when
17079 optimization_current_node has changed.
17080 * target-globals.h (save_target_globals_default_opts): Protoize.
17081 * target-globals.c (save_target_globals_default_opts): New.
17083 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17086 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
17087 registers %fr12 and %fr12R as call used.
17090 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
17091 and HImode, require all displacements to be an integer multiple of
17093 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
17094 only allow QImode and HImode when reload is in progress and strict is
17095 true. Likewise for symbolic addresses. Use base14_operand to check
17096 displacements in REG+BASE addresses.
17098 2013-02-18 Richard Biener <rguenther@suse.de>
17100 PR tree-optimization/56366
17101 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17102 handle sign-conversion of outer-loop initial induction value.
17104 2013-02-18 Richard Biener <rguenther@suse.de>
17106 PR middle-end/56349
17107 * cfghooks.c (merge_blocks): If we merge a latch into another
17108 block adjust references to it.
17109 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
17110 (verify_loop_structure): Verify that a recorded latch is in fact
17113 2013-02-18 Richard Biener <rguenther@suse.de>
17115 PR tree-optimization/56321
17116 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
17117 order SSA name release and virtual operand unlinking.
17119 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17121 * config/microblaze/microblaze.md (save_stack_block): Define.
17122 (restore_stack_block): Likewise.
17124 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17126 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
17127 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
17128 * config/microblaze/microblaze.c (microblaze_option_override):
17129 Bail out early for PIC modes when target does not support PIC.
17131 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17133 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
17134 Replace with a microblaze version.
17135 (microblaze_trampoline_init): Adapt for microblaze.
17136 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
17139 2013-02-16 Jakub Jelinek <jakub@redhat.com>
17140 Dodji Seketeli <dodji@redhat.com>
17143 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
17144 (instrument_mem_region_access): Do not forget to always put
17145 instrumentation of the of 'base' and 'base + len' in a "if (len !=
17146 0) statement, even for cases where either 'base' or 'base + len'
17147 are not instrumented -- because they have been previously
17148 instrumented. Simplify the logic by putting all the statements
17149 instrument 'base + len' inside a sequence, and then insert that
17150 sequence right before the current insertion point. Then, to
17151 instrument 'base + len', just get an iterator on that statement.
17152 And do not forget to update the pointer to iterator the function
17153 received as argument.
17155 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
17157 PR rtl-optimization/56348
17158 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
17160 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
17162 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
17163 (clean_graph_dump_file): Pass base to start_graph_dump.
17165 2013-02-14 Richard Henderson <rth@redhat.com>
17168 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
17170 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
17172 * collect2-aix.h: Define F_LOADONLY.
17174 2013-02-14 Richard Biener <rguenther@suse.de>
17177 * varasm.c (output_constant_def_1): Get the decl representing
17178 the constant as argument.
17179 (output_constant_def): Wrap output_constant_def_1.
17180 (make_decl_rtl): Use output_constant_def_1 with the decl
17181 representing the constant.
17182 (build_constant_desc): Optionally re-use a decl already
17183 representing the constant.
17184 (tree_output_constant_def): Adjust.
17186 2013-02-14 Dodji Seketeli <dodji@redhat.com>
17189 * asan.c (instrument_builtin_call): Really put the length of the
17190 second source argument into src1_len.
17192 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17194 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
17195 argument. If it is false, don't create edge from then_bb to
17197 (insert_if_then_before_iter): Pass true to it.
17198 (build_check_stmt): Pass false to it.
17199 (transform_statements): Flush hash table only on extended basic
17200 block boundaries, rather than at the beginning of every bb.
17201 Don't flush hash table on nonfreeing_call_p calls.
17202 * tree-flow.h (nonfreeing_call_p): New prototype.
17203 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
17205 2013-02-13 David S. Miller <davem@davemloft.net>
17207 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
17209 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
17212 * ira.c (max_regno_before_ira): Move from ...
17214 (fix_reg_equiv_init): Use max_regno_before_ira instead of
17217 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17219 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
17221 2013-02-13 Richard Biener <rguenther@suse.de>
17224 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
17225 globals in MEM_REFs.
17227 2013-02-13 Richard Biener <rguenther@suse.de>
17229 * loop-init.c (loop_optimizer_init): Clear loop state when
17230 re-initializing preserved loops.
17231 * loop-unswitch.c (unswitch_single_loop): Return whether
17232 we unswitched the loop. Do not verify loop state here.
17233 (unswitch_loops): When we unswitched a loop discover new loops.
17235 2013-02-13 Kostya Serebryany <kcc@google.com>
17237 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
17239 * sanitizer.def: Rename __asan_init to __asan_init_v1.
17241 2013-02-12 Dodji Seketeli <dodji@redhat.com>
17243 Avoid instrumenting duplicated memory access in the same basic block
17244 * Makefile.in (asan.o): Add new dependency on hash-table.h
17245 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
17246 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
17247 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
17248 (free_mem_ref_resources, has_mem_ref_been_instrumented)
17249 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
17250 (get_mem_ref_of_assignment): New functions.
17251 (get_mem_refs_of_builtin_call): Extract from
17252 instrument_builtin_call and tweak a little bit to make it fit with
17254 (instrument_builtin_call): Use the new
17255 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
17256 of is_gimple_builtin_call.
17257 (instrument_derefs, instrument_mem_region_access): Insert the
17258 instrumented memory reference into the hash table.
17259 (maybe_instrument_assignment): Renamed instrument_assignment into
17260 this, and change it to advance the iterator when instrumentation
17261 actually happened and return true in that case. This makes it
17262 homogeneous with maybe_instrument_assignment, and thus give a
17263 chance to callers to be more 'regular'.
17264 (transform_statements): Clear the memory reference hash table
17265 whenever we enter a new BB, when we cross a function call, or when
17266 we are done transforming statements. Use
17267 maybe_instrument_assignment instead of instrumentation. No more
17268 need to special case maybe_instrument_assignment and advance the
17269 iterator after calling it; it's now handled just like
17270 maybe_instrument_call. Update comment.
17272 2013-02-13 Richard Biener <rguenther@suse.de>
17274 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
17275 Fix loop discovery code.
17277 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
17279 PR inline-asm/56148
17280 * lra-constraints.c (process_alt_operands): Match early clobber
17281 operand with itself. Check conflicts with earlyclobber only if
17282 the operand is not reloaded. Prefer to reload conflicting operand
17283 if earlyclobber and matching operands are the same.
17285 2013-02-12 Richard Biener <rguenther@suse.de>
17288 * lto-streamer-out.c (write_symbol): Do not output symbols
17289 for hard register variables.
17291 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
17294 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
17295 (umulsidi3_insn, mulsidi3_insn): New insns.
17297 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
17299 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
17300 (struct tune_params): Add vec_costs field.
17301 * config/arm/arm.c (arm_builtin_vectorization_cost)
17302 (arm_add_stmt_cost): New functions.
17303 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
17304 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
17305 (arm_default_vec_cost): New struct of type cpu_vec_costs.
17306 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
17307 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
17308 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
17309 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
17311 2013-02-12 Richard Biener <rguenther@suse.de>
17314 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
17315 decls again if possible.
17317 2013-02-12 Richard Biener <rguenther@suse.de>
17319 PR middle-end/56288
17320 * tree-ssa.c (verify_ssa_name): Fix check, move
17321 SSA_NAME_IN_FREE_LIST check up.
17323 2013-02-12 Jakub Jelinek <jakub@redhat.com>
17324 Steven Bosscher <steven@gcc.gnu.org>
17326 PR rtl-optimization/56151
17327 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
17328 equal to op0 or op1, and last_insn pattern is CODE operation
17329 with MEM dest and one of the operands matches that MEM.
17331 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
17333 * doc/extend.texi: Document Function Multiversioning and "default"
17334 parameter string to target attribute.
17335 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
17336 target attribute parameter is "default".
17337 (ix86_compare_version_priority): Remove checks for target attribute.
17338 (ix86_mangle_function_version_assembler_name): Change error to sorry.
17339 Remove check for target attribute equal to NULL. Add assert.
17340 (ix86_generate_version_dispatcher_body): Change error to sorry.
17342 2013-02-11 Iain Sandoe <iain@codesourcery.com>
17343 Jack Howarth <howarth@bromo.med.uc.edu>
17344 Patrick Marlier <patrick.marlier@gmail.com>
17347 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
17348 define ENDFILE_SPEC as TM_DESTRUCTOR.
17349 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
17351 2013-02-11 Alexander Potapenko <glider@google.com>
17352 Jack Howarth <howarth@bromo.med.uc.edu>
17353 Jakub Jelinek <jakub@redhat.com>
17356 * config/darwin.c (cdtor_record): Rename ctor_record.
17357 (sort_cdtor_records): Rename sort_ctor_records.
17358 (finalize_dtors): New routine to sort destructors by
17359 priority before use in assemble_integer.
17360 (machopic_asm_out_destructor): Use finalize_dtors if needed.
17362 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
17364 PR rtl-optimization/56275
17365 * simplify-rtx.c (avoid_constant_pool_reference): Check that
17366 offset is non-negative and less than cmode size before
17367 calling simplify_subreg.
17369 2013-02-11 Richard Biener <rguenther@suse.de>
17371 PR tree-optimization/56264
17372 * cfgloop.h (fix_loop_structure): Adjust prototype.
17373 * loop-init.c (fix_loop_structure): Return the number of
17374 newly discovered loops.
17375 * tree-cfgcleanup.c (repair_loop_structures): When new loops
17376 are discovered, do a full loop-closed SSA rewrite.
17378 2013-02-11 Richard Biener <rguenther@suse.de>
17380 PR tree-optimization/56273
17381 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
17383 (check_array_ref): Fix missing newline in dumps.
17384 (search_for_addr_array): Likewise.
17386 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
17388 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
17390 2013-02-09 Jakub Jelinek <jakub@redhat.com>
17393 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
17395 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
17397 PR rtl-optimization/56246
17398 * lra-constraints.c (simplify_operand_subreg): Try to reuse
17400 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
17401 constraints are satisfied.
17403 2013-02-08 Jeff Law <law@redhat.com>
17406 * emit-rtl.c (reg_is_parm_p): New function.
17407 * regs.h (reg_is_parm_p): New prototype.
17408 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
17409 callee-clobbered registers.
17411 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
17414 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
17415 If there is no implicit builtin declaration, just return NULL.
17417 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17419 * config/i386/sse.md (FMAMODEM): New mode iterator.
17420 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
17421 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
17423 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17425 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
17426 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
17427 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
17429 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17431 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
17432 (microblaze*-*-elf): Likewise.
17433 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
17435 * config/microblaze/microblaze-c.c: Add builtin defines for
17436 _LITTLE_ENDIAN and _BIG_ENDIAN.
17437 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
17438 add to TARGET_DEFAULT flags.
17439 Expand ASM_SPEC and LINK_SPEC.
17440 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
17441 * config/microblaze/microblaze.md: Update extendsidi2 and
17442 movdi_internal instructions to use low-order / high-order reg
17444 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
17445 options and inversemask / mask of LITTLE_ENDIAN.
17446 * config/microblaze/t-microblaze: Expand multilib options to
17447 include mlittle-endian (le) and update exceptions patterns.
17449 2013-02-08 Jakub Jelinek <jakub@redhat.com>
17451 PR rtl-optimization/56195
17452 * lra-constraints.c (get_reload_reg): Don't reuse regs
17453 if they have smaller mode than requested, if they have
17454 wider mode than requested, try to return a SUBREG.
17456 PR tree-optimization/56250
17457 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
17458 if type is unsigned and code isn't MULT_EXPR.
17460 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
17462 PR tree-optimization/56064
17463 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
17464 bits according to mode.
17465 * fixed-value.h (fixed_from_double_int)
17466 (const_fixed_from_double_int): Adjust comments.
17468 2013-02-08 Richard Biener <rguenther@suse.de>
17471 * lto-streamer.h (struct data_in): Remove current_file, current_line
17472 and current_col members.
17473 * lto-streamer-out.c (lto_output_location): Stream changed bits
17474 en-block for efficiency.
17475 * lto-streamer-in.c (clear_line_info): Remove.
17476 (lto_input_location): Cache current file, line and column
17477 globally via local statics. Read changed bits en-block.
17478 (input_function): Do not call clear_line_info.
17479 (lto_read_body): Likewise.
17480 (lto_input_toplevel_asms): Likewise.
17482 2013-02-08 Michael Matz <matz@suse.de>
17484 PR tree-optimization/52448
17485 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
17486 (nt_call_phase): New static.
17487 (add_or_mark_expr): Only mark accesses with newer phase than any
17489 (nonfreeing_call_p): New.
17490 (nt_init_block): Update nt_call_phase, mark blocks as visited.
17491 (nt_fini_block): Keep blocks marked as visited.
17492 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
17494 2013-02-08 Richard Biener <rguenther@suse.de>
17496 * ira.c (ira): Free broken dominator information.
17498 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17500 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
17502 2013-02-08 Marek Polacek <polacek@redhat.com>
17504 * cfgloop.c (verify_loop_structure): Add more checking of headers.
17506 2013-02-08 Richard Biener <rguenther@suse.de>
17508 PR middle-end/56181
17509 * cfgloop.h (flow_loops_find): Adjust.
17510 (bb_loop_header_p): Declare.
17511 * cfgloop.c (bb_loop_header_p): New function split out from ...
17512 (flow_loops_find): ... here. Adjust function signature,
17513 support incremental loop structure update.
17514 (verify_loop_structure): Cleanup. Verify a loop is a loop.
17515 * cfgloopmanip.c (fix_loop_structure): Move ...
17516 * loop-init.c (fix_loop_structure): ... here.
17517 (apply_loop_flags): Split out from ...
17518 (loop_optimizer_init): ... here.
17519 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
17520 in incremental mode, only remove dead loops here.
17522 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
17525 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
17526 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
17527 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
17528 (*round<mode>3.libgcc): New insns for fixed-modes.
17529 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
17530 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
17531 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
17532 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
17533 implementations. Define to __builtin_avr_absFX,
17534 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
17535 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
17536 __builtin_avr_countlsFX, respectively.
17537 * config/avr/avr-c.c (target.h): Include it.
17538 (enum avr_builtin_id): New enum.
17539 (avr_resolve_overloaded_builtin): New static function.
17540 (avr_register_target_pragmas): Use it to set
17541 targetm.resolve_overloaded_builtin.
17542 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
17543 tree nodes used by DEF_BUILTIN.
17544 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
17545 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
17546 <AVR_BUILTIN_xxBITS>: Same.
17548 2013-02-08 Richard Biener <rguenther@suse.de>
17550 * cfgloop.c (verify_loop_structure): Properly handle
17551 a loop exiting to another loop header.
17552 * ira-int.h (ira_loops): Remove.
17553 * ira.c (ira_loops): Remove.
17554 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
17555 (do_reload): Use loop_optimizer_finalize.
17556 * ira-build.c (create_loop_tree_nodes): Use get_loops and
17557 number_of_loops to access the loop tree.
17558 (more_one_region_p): Likewise.
17559 (finish_loop_tree_nodes): Likewise.
17560 (rebuild_regno_allocno_maps): Likewise.
17561 (mark_loops_for_removal): Likewise.
17562 (mark_all_loops_for_removal): Likewise.
17563 (remove_unnecessary_regions): Likewise.
17564 (ira_build): Likewise.
17565 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
17567 2013-02-08 Richard Biener <rguenther@suse.de>
17569 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
17570 * ipa-pure-const.c (analyze_function): Avoid calling
17571 mark_irreducible_loops twice.
17572 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
17574 2013-02-07 David S. Miller <davem@davemloft.net>
17576 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
17578 * var-tracking.c (vt_add_function_parameter): Test the presence of
17579 HAVE_window_save properly and do not remap argument registers when
17580 we have a leaf function.
17582 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
17585 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
17587 * config/i386/i386.c (ix86_print_operand): Ditto.
17589 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
17591 * lra-constraints.c (process_alt_operands): Fix recently added comment.
17593 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
17595 PR rtl-optimization/56225
17596 * lra-constraints.c (process_alt_operands): Check that reload hard
17597 reg can hold value for strict_low_part.
17599 2013-02-07 Jakub Jelinek <jakub@redhat.com>
17602 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
17603 dwarf2out_end_function.
17604 (in_first_function_p, maybe_at_text_label_p,
17605 first_loclabel_num_not_at_text_label): New variables.
17606 (dwarf2out_var_location): In the first function find out
17607 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
17608 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
17611 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
17613 PR rtl-optimization/56178
17614 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
17615 SUBREG of a register. Tidy up related block of code.
17616 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
17617 note if the source is a register or a SUBREG of a register.
17619 2013-02-07 Jakub Jelinek <jakub@redhat.com>
17622 * config/rs6000/rs6000.md (ptrm): New mode attr.
17623 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
17624 call_value_indirect_aix<pttrsize>,
17625 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
17628 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17630 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
17631 if -bnortl. Convert to strcmp and strncmp.
17633 2013-02-07 Alan Modra <amodra@gmail.com>
17636 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
17637 addresses won't wrap when offsetting.
17638 (rs6000_secondary_reload): Provide secondary reloads needed for
17639 wrapping LO_SUM addresses.
17641 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
17643 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
17644 MACH, just __MACH__.
17646 2013-02-06 Richard Biener <rguenther@suse.de>
17648 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
17649 instead of calling fix_loop_structure.
17651 2013-02-06 Jakub Jelinek <jakub@redhat.com>
17653 PR middle-end/56217
17654 * omp-low.c (use_pointer_for_field): Return false if
17655 lower_send_shared_vars doesn't generate any copy-out code.
17657 2013-02-06 Tom de Vries <tom@codesourcery.com>
17659 PR rtl-optimization/56131
17660 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
17661 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
17662 of the label is NULL. Add comment.
17664 2013-02-05 Jakub Jelinek <jakub@redhat.com>
17666 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
17669 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
17670 (STATIC_LIBTSAN_LIBS): Likewise.
17671 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
17672 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
17673 is defined, don't add anything else beyond that.
17674 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
17675 (LINK_COMMAND_SPEC): Use them.
17677 PR tree-optimization/56205
17678 * tree-stdarg.c (check_all_va_list_escapes): Return true if
17679 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
17680 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
17682 2013-02-05 Richard Biener <rguenther@suse.de>
17684 PR tree-optimization/53342
17685 PR tree-optimization/53185
17686 * tree-vectorizer.h (vect_check_strided_load): Remove.
17687 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
17688 not disallow peeling for vectorized strided loads.
17689 (vect_check_strided_load): Make static and simplify.
17690 (vect_analyze_data_refs): Adjust.
17691 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
17692 correctly when vectorizing strided loads.
17694 2013-02-05 Richard Biener <rguenther@suse.de>
17696 * doc/install.texi: Refer to ISL, not PPL.
17698 2013-02-05 Jan Hubicka <jh@suse.cz>
17700 PR tree-optimization/55789
17701 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
17703 2013-02-05 Jan Hubicka <jh@suse.cz>
17705 PR tree-optimization/55789
17706 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
17707 the dead call anyway.
17709 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
17712 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
17714 2013-02-04 Alexander Potapenko <glider@google.com>
17715 Jack Howarth <howarth@bromo.med.uc.edu>
17716 Jakub Jelinek <jakub@redhat.com>
17719 * config/darwin.c (sort_ctor_records): Stabilized qsort
17720 on constructor priority by using original position.
17721 (finalize_ctors): New routine to sort constructors by
17722 priority before use in assemble_integer.
17723 (machopic_asm_out_constructor): Use finalize_ctors if needed.
17725 2013-02-04 Jakub Jelinek <jakub@redhat.com>
17728 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
17729 about visibility on artificial decls.
17730 * config/sol2.c (solaris_assemble_visibility): Likewise.
17732 2013-02-04 Kai Tietz <ktietz@redhat.com>
17735 * config/i386/i386.c (function_value_ms_64): Add additional valtype
17736 argument and improve checking of return-argument types for 16-byte
17738 (ix86_function_value_1): Add additional valtype argument on call
17739 of function_value_64.
17740 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
17741 handling infunction_value_64 function.
17743 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
17745 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
17747 2013-02-04 Richard Biener <rguenther@suse.de>
17749 PR tree-optimization/56188
17750 * tree-ssa-structalias.c (label_visit): Consider case with
17751 initially non-empty points-to set.
17752 (perform_var_substitution): Dump node mapping and clean up.
17754 2013-02-04 Richard Guenther <rguenther@suse.de>
17757 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
17758 node prevail as last resort.
17759 (lto_symtab_merge_decls): Remove guard on LTRANS here.
17760 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
17762 2013-02-04 Richard Biener <rguenther@suse.de>
17764 PR tree-optimization/56113
17765 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
17767 (equiv_class_lookup_or_add): ... this.
17768 (label_visit): Adjust and fix error in previous patch.
17769 (perform_var_substitution): Adjust.
17771 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
17773 * config/sh/divtab.c: Fix formatting and comments throughout the file.
17774 * config/sh/sh4-300.md: Likewise.
17775 * config/sh/sh4a.md: Likewise.
17776 * config/sh/constraints.md: Likewise.
17777 * config/sh/sh.md: Likewise.
17778 * config/sh/netbsd-elf.h: Likewise.
17779 * config/sh/predicates.md: Likewise.
17780 * config/sh/sh-protos.h: Likewise.
17781 * config/sh/ushmedia.h: Likewise.
17782 * config/sh/linux.h: Likewise.
17783 * config/sh/sh.c: Likewise.
17784 * config/sh/superh.h: Likewise.
17785 * config/sh/elf.h: Likewise.
17786 * config/sh/sh4.md: Likewise.
17787 * config/sh/sh.h: Likewise.
17789 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17791 * config/pa/constraints.md: Adjust unused letters. Change "T"
17792 constraint to match_test floating_point_store_memory_operand().
17793 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
17794 (base14_operand): New.
17795 (floating_point_store_memory_operand): New.
17796 (integer_store_memory_operand): Revise to use base14_operand and
17797 reg_plus_base_memory_operand.
17798 (move_dest_operand): Allow symbolic_memory_operands.
17799 (symbolic_memory_operand): Check for LO_SOM.
17800 (symbolic_operand): Change default case to break.
17801 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
17802 CONST_DOUBLE values to be reloaded by putting them into memory when
17803 the destination is a floating point register.
17804 (movdf): Remove code to handle CONST_DOUBLE.
17806 (reload_indf_r1): New.
17807 (reload_insf_r1): New.
17808 Consistently use "Q" and "T" constraints with integer and floating
17809 point move instructions, respectively.
17810 (movdi): Remove FAIL.
17811 Change predicate for source operand unamed DImode move from
17812 general_operand to move_src_operand.
17813 (umulsidi3): Change predicate for destination operand to
17815 Likewise for similar unamed patterns.
17816 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
17817 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
17818 (hppa_legitimize_address): Simplify mask calculation.
17819 (pa_emit_move_sequence): Revised handling of secondary reloads from
17820 REG+D addresses for floating point loads and stores. Directly handle
17821 loading CONST0_RTX (mode) to a floating point register.
17822 (pa_secondary_reload): Handle reloading DF and SFmode constant values
17823 to floating point registers. Don't restrict secondary reloads to
17824 floating point registers to integer modes. Revise some comments and
17826 (TARGET_LEGITIMATE_ADDRESS_P): Define.
17827 (pa_legitimate_address_p): New.
17828 (pa_legitimize_reload_address): New.
17829 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
17830 (STRICT_REG_OK_FOR_BASE_P): New.
17831 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
17832 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
17834 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
17835 Andrew Dixie <andrewd@gentrack.com>
17837 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
17840 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
17842 * expmed.c (extract_bit_field_1): Pass the full width of the
17843 structure to get_best_reg_extraction_insn.
17845 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17848 * configure.ac (use_cxa_atexit): Add AIX.
17849 * configure: Regenerate.
17851 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
17853 2013-02-01 Jakub Jelinek <jakub@redhat.com>
17856 * final.c (need_profile_function): New variable.
17857 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
17858 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
17859 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
17860 notes, targetm.asm_out.function_prologue doesn't emit anything,
17861 HAVE_prologue and profiler should be emitted before prologue,
17862 set need_profile_function instead of emitting it.
17863 (final_scan_insn): If need_profile_function, emit
17864 profile_function on the first NOTE_INSN_BASIC_BLOCK or
17865 NOTE_INSN_FUNCTION_BEG note.
17867 2013-02-01 Richard Henderson <rth@redhat.com>
17869 * config/rs6000/rs6000.md (smulditi3): New.
17872 * config/alpha/alpha.md (umulditi3): New.
17874 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17876 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
17877 (ASM_OUTPUT_ALIGNED_LOCAL): New.
17879 2013-02-01 Richard Biener <rguenther@suse.de>
17881 PR tree-optimization/56113
17882 * tree-ssa-structalias.c (label_visit): Reduce work for
17883 single-predecessor nodes.
17885 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
17887 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
17888 range isn't testing for zero.
17890 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
17892 PR middle-end/56113
17893 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
17895 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
17896 Nick Clifton <nickc@redhat.com>
17898 * config/v850/constraints.md (Q): Define as a memory constraint.
17899 * config/v850/predicates.md (label_ref_operand): New predicate.
17900 (e3v5_shift_operand): New predicate.
17901 (ior_operator): New predicate.
17902 * config/v850/t-v850: Add e3v5 multilib.
17903 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17904 (v850_gen_movdi): Prototype.
17905 * config/v850/v850.c: Add support for e3v5 architecture.
17906 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17908 (construct_save_jarl): Add e3v5 long JARL support.
17909 (v850_adjust_insn_length): New function. Adjust length of call
17910 insns when using e3v5 instructions.
17911 (v850_gen_movdi): New function: Generate instructions to move a
17913 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17914 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17915 (TARGET_USE_FPU): Enable for e3v5.
17916 (CONST_OK_FOR_W): New macro.
17917 (ADJUST_INSN_LENGTH): Define.
17918 * config/v850/v850.md (UNSPEC_LOOP): Define.
17919 (attr cpu): Add v850e3v5.
17920 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17921 (movdi): New pattern.
17922 (movdi_internal): New pattern.
17923 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17924 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17925 (cstoresf4): Likewise.
17926 (cstoredf4): Likewise.
17927 (insv): New pattern.
17928 (rotlso3_a): New pattern.
17929 (rotlsi3_b): New pattern
17930 (rotlsi3_v850e3v5): New pattern.
17931 (doloop_begin): New pattern.
17932 (fix_loop_counter): New pattern.
17933 (doloop_end): New pattern.
17934 (branch_normal): Add e3v5 long branch support.
17935 (branch_invert): Likewise.
17936 (branch_z_normal): Likewise.
17937 (branch_z_invert): Likewise.
17938 (branch_nz_normal): Likewise.
17939 (branch_nz_invert): Likewise.
17940 (call_internal_short): Add e3v5 register-indirect JARL support.
17941 (call_internal_long): Likewise.
17942 (call_value_internal_short): Likewise.
17943 (call_value_internal_long): Likewise.
17944 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17945 (mloop): New option.
17946 * config.gcc: Add support for configuring v840e3v5 target.
17947 * doc/invoke.texi: Document new v850 specific command line options.
17949 2013-01-31 Paul Koning <ni1d@arrl.net>
17953 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17954 children if parent is a class.
17955 (prune_unused_types_prune): Don't add DW_AT_declaration.
17957 2013-01-31 Richard Biener <rguenther@suse.de>
17959 PR tree-optimization/56157
17960 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17961 match up operand with SLP child.
17963 2013-01-31 Jason Merrill <jason@redhat.com>
17966 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17967 parameters the first time.
17968 (gen_scheduled_generic_parms_dies): Check completeness here.
17970 2013-01-31 Richard Biener <rguenther@suse.de>
17972 PR middle-end/53073
17973 * common.opt (faggressive-loop-optimizations): New flag,
17974 enabled by default.
17975 * doc/invoke.texi (faggressive-loop-optimizations): Document.
17976 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
17977 infer_loop_bounds_from_undefined by it.
17979 2013-01-31 Richard Biener <rguenther@suse.de>
17981 PR tree-optimization/56150
17982 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
17983 visit virtual operands.
17984 (find_uses_to_rename_bb): Likewise.
17986 2013-01-31 Richard Biener <rguenther@suse.de>
17988 PR tree-optimization/56150
17989 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
17990 mixed store non-store stmts.
17992 2013-01-30 Jakub Jelinek <jakub@redhat.com>
17995 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
17996 LIBASAN_EARLY_SPEC is defined.
17997 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
17998 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
18000 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
18003 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
18004 invalid args instead of ICEing on it.
18005 (ix86_valid_target_attribute_tree): Return error_mark_node if
18006 ix86_valid_target_attribute_inner_p failed.
18007 (ix86_valid_target_attribute_p): Return false only if
18008 ix86_valid_target_attribute_tree returned error_mark_node. Allow
18009 target("default") attribute.
18010 (sorted_attr_string): Change argument from const char * to tree,
18011 merge in all target attribute arguments rather than just one.
18012 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
18013 instead of free. Avoid using strcat.
18014 (ix86_mangle_function_version_assembler_name): Mangle
18015 target("default") as if no target attribute is present. Adjust
18016 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
18017 instead of xmalloc and XDELETEVEC instead of free.
18018 (ix86_function_versions): Don't return true if one of the decls
18019 doesn't have target attribute. If they don't and one of the decls
18020 is DECL_FUNCTION_VERSIONED, report an error. Adjust
18021 sorted_attr_string caller. Use XDELETEVEC instead of free.
18022 (ix86_supports_function_versions): Remove.
18023 (make_name): Fix up formatting.
18024 (make_dispatcher_decl): Remove resolver_name and its initialization.
18025 Avoid leaking memory.
18026 (is_function_default_version): Return true if there is
18027 target("default") attribute rather than no target attribute at all.
18028 (make_resolver_func): Avoid leaking memory.
18029 (ix86_generate_version_dispatcher_body): Likewise.
18030 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
18031 * target.def (supports_function_versions): Remove.
18032 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
18033 * doc/tm.texi: Regenerated.
18035 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
18037 PR rtl-optimization/56144
18038 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
18039 for values with side effects.
18041 2013-01-30 Richard Biener <rguenther@suse.de>
18043 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
18044 (sparseset_pop): Likewise.
18045 * cfganal.c (compute_idf): Likewise. Increase work-stack size
18046 to be able to use quick_push in the worker loop.
18048 2013-01-30 Marek Polacek <polacek@redhat.com>
18050 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
18052 2013-01-30 Richard Biener <rguenther@suse.de>
18055 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
18057 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
18059 PR tree-optimization/56064
18060 * fixed-value.c (fixed_from_double_int): New function.
18061 * fixed-value.h (fixed_from_double_int): New prototype.
18062 (const_fixed_from_double_int): New static inline function.
18063 * fold-const.c (native_interpret_fixed): New static function.
18064 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
18065 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
18066 (native_encode_fixed): New static function.
18067 (native_encode_expr) <FIXED_CST>: Use it.
18068 (native_interpret_int): Move double_int worker code to...
18069 * double-int.c (double_int::from_buffer): ...this new static method.
18070 * double-int.h (double_int::from_buffer): Prototype it.
18072 2013-01-30 Richard Biener <rguenther@suse.de>
18074 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
18075 New pointer-map and obstack.
18076 (init_alias_vars): Allocate pointer-map and obstack.
18077 (delete_points_to_sets): Free them.
18078 (find_what_var_points_to): Cache result.
18079 (find_what_p_points_to): Adjust for changed interface of
18080 find_what_var_points_to.
18081 (compute_points_to_sets): Likewise.
18082 (ipa_pta_execute): Likewise.
18084 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18086 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
18087 * configure: Regenerate.
18088 * config.in: Regenerate.
18089 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
18090 #nobits/#progbits if supported.
18092 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
18095 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
18096 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
18097 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
18099 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18101 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
18102 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
18104 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18106 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
18107 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
18109 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18111 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
18113 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
18114 * config/arm/cortex-a7.md: New bypasses using
18115 arm_mac_accumulator_is_result.
18117 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18119 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
18120 (cortex_a7_neon_mla): Likewise.
18121 (cortex_a7_fpfmad): New reservation.
18122 (cortex_a7_fpmacs): Use ffmas and update required units.
18123 (cortex_a7_fpmuld): Update required units and latency.
18124 (cortex_a7_fpmacd): Likewise.
18125 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
18126 (cortex_a7_neon). Likewise.
18127 (bypass) Update participating units.
18129 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18131 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
18132 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
18134 * config/arm/vfp11.md (vfp_farith): Use ffmas.
18135 (vfp_fmul): Use ffmad.
18136 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
18137 (cortex_r4_fmacd): Use ffmad.
18138 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
18139 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
18140 (cortex_a9_fmacd): Use ffmad.
18141 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
18142 (cortex_a8_vfp_macd): Use ffmad.
18143 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
18144 (cortex_a5_fpmacd): Use ffmad.
18145 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
18146 (cortex_a15_vfp_macd): Use ffmad.
18147 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
18149 2013-01-29 Jason Merrill <jason@redhat.com>
18152 * varasm.c (default_assemble_visibility): Don't warn about
18153 visibility on artificial decls.
18155 2013-01-29 Richard Biener <rguenther@suse.de>
18157 PR tree-optimization/56113
18158 * tree-ssa-structalias.c (equiv_class_lookup): Also return
18160 (label_visit): Free duplicate bitmaps and record the leader instead.
18161 (perform_var_substitution): Adjust.
18163 2013-01-29 Richard Biener <rguenther@suse.de>
18165 PR tree-optimization/55270
18166 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
18167 the CFG, schedule loops for fixup.
18169 2013-01-29 Nick Clifton <nickc@redhat.com>
18171 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
18174 2013-01-28 Leif Ekblad <leif@rdos.net>
18176 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
18177 * config/i386/i386.h (TARGET_RDOS): New macro.
18178 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
18179 * config/i386/i386.c (ix86_option_override_internal): For 64bit
18180 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
18181 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
18182 DEFAULT_LARGE_SECTION_THRESHOLD.
18183 * config/i386/i386.md (R14_REG, R15_REG): New constants.
18184 * config/i386/rdos.h: New file.
18185 * config/i386/rdos64.h: New file.
18187 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
18190 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
18193 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18195 PR rtl-optimization/56117
18196 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
18197 call cselib_lookup_from_insn on the MEM before calling
18198 add_insn_mem_dependence.
18200 2013-01-28 Richard Biener <rguenther@suse.de>
18202 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
18203 to a stmt that didn't have one.
18204 (copy_phis_for_bb): Likewise for PHI arguments.
18205 (copy_debug_stmt): Likewise for debug stmts.
18207 2013-01-28 Richard Biener <rguenther@suse.de>
18209 PR tree-optimization/56034
18210 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
18211 (partition_builtin_p): Adjust.
18212 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
18213 it is the last partition.
18214 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
18215 up the vertex for the definition.
18216 (classify_partition): Classify whether a partition is a
18217 PKIND_REDUCTION, thus has uses outside of the loop.
18218 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
18219 Merge all PKIND_REDUCTION partitions into the last partition.
18220 (tree_loop_distribution): Seed partitions from reductions as well.
18222 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18224 PR tree-optimization/56125
18225 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
18226 pow(x,c) into sqrt(x) * powi(x, n/2) or
18227 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
18228 optimizing for size.
18229 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
18230 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
18233 PR tree-optimization/56094
18234 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
18235 to UNKNOWN_LOCATION while gimplifying expr.
18237 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
18240 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
18241 operand 0 in movabs insn template for -masm=intel asm alternative.
18242 (*movabs<mode>_2): Ditto for operand 1.
18244 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
18247 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
18250 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18252 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
18255 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18257 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
18258 Undef to avoid warning.
18260 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
18262 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
18263 * configure: Regenerate.
18265 2013-01-25 Jakub Jelinek <jakub@redhat.com>
18267 PR tree-optimization/56098
18268 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
18269 for stmts with volatile ops.
18270 (cond_store_replacement): Don't optimize if assign has volatile ops.
18271 (cond_if_else_store_replacement_1): Don't optimize if either
18272 then_assign or else_assign have volatile ops.
18273 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
18276 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18278 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
18280 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18282 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
18283 missing ':' in asm example.
18285 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
18287 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
18288 entries into lane and laneq entries.
18289 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
18290 Remove AdvSIMD scalar modes.
18291 (aarch64_sq<r>dmulh_laneq<mode>): New.
18292 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
18294 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
18295 builtin implementations to relfect changes in RTL in aarch64-simd.md.
18296 * config/aarch64/iterators.md (VCOND): New.
18299 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18302 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
18303 Add NULL LIBNAME argument to existing definitions.
18304 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
18305 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
18306 * config/avr/avr.c (DEF_BUILTIN): Same.
18307 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
18308 (avr_expand_builtin): Expand to a vanilla call if a libgcc
18309 implementation is available (DECL_ASSEMBLER_NAME is set).
18310 (avr_fold_absfx): New static function.
18311 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
18312 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
18313 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
18314 AVR_BUILTIN_ABSLLK.
18315 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
18316 (abshk, absk, abslk, absllk): Provide as static inline functions.
18318 2013-01-25 Marek Polacek <polacek@redhat.com>
18320 PR tree-optimization/56035
18321 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
18323 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
18325 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
18326 (*movtf_internal_rex64): Add (!o,C) alternative
18327 (*movxf_internal_rex64): Ditto.
18328 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
18330 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
18332 * doc/invoke.texi: fix typo.
18333 * doc/objc.texi: fix typo.
18335 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
18337 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
18338 for the first two alternatives.
18340 2013-01-24 Diego Novillo <dnovillo@google.com>
18342 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
18343 (ggc-zone.o): Remove.
18344 * configure.ac: Remove option --with-gc.
18345 * configure: Re-generate.
18346 * doc/install.texi: Remove documentation for --with-gc.
18347 * gengtype.c (write_enum_defn): Remove. Update all users.
18348 (write_Types_process_field): Remove generation of gt_e_* argument.
18349 (output_type_enum): Remove. Update all users.
18350 (write_enum_defn): Remove. Update all users.
18351 (enum alloc_zone): Remove. Update all users.
18352 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
18353 * ggc-common.c (ggc_splay_alloc): Remove first argument.
18354 Update all callers.
18355 (struct ptr_data): Remove field TYPE. Update all users.
18356 (gt_pch_note_object): Remove argument TYPE. Update all users.
18357 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
18359 * ggc-none.c (ggc_alloc_typed_stat): Remove.
18360 (struct alloc_zone): Remove.
18361 (ggc_internal_alloc_zone_stat): Remove.
18362 (ggc_internal_cleared_alloc_zone_stat): Remove.
18363 * ggc-page.c (ggc_alloc_typed_stat): Remove.
18364 (ggc_pch_count_object): Remove last argument. Update all users.
18365 (ggc_pch_alloc_object): Remove last argument. Update all users.
18366 (struct alloc_zone): Remove.
18367 * ggc-zone.c: Remove.
18368 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
18369 (struct alloc_zone): Remove.
18370 (ggc_alloc_typed_stat): Remove.
18371 (ggc_alloc_typed): Remove.
18372 (ggc_splay_alloc): Remove first argument.
18373 (rtl_zone): Remove. Update all users.
18374 (tree_zone): Remove. Update all users.
18375 (tree_id_zone): Remove. Update all users.
18376 (ggc_internal_zone_alloc_stat): Remove. Update all users.
18377 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
18378 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
18379 * tree-ssanames.c: Remove references to zone allocator in comments.
18381 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
18383 * config/avr/avr.c (avr_out_fract): Make register numbers that
18384 might be outside of source operand signed.
18386 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
18388 * config/i386/constraints.md (Yf): New constraint.
18389 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
18390 of f constraint to conditionaly disable x87 register preferences.
18391 (*movdf_internal): Ditto.
18392 (*movsf_internal): Ditto.
18394 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
18396 PR inline-asm/55934
18397 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
18398 that have operands with impossible constraints.
18399 Add a FIXME for a speed-up opportunity.
18400 * lra-constraints.c (process_alt_operands): Verify that a class
18401 selected from constraints on asms is valid for the operand mode.
18402 (curr_insn_transform): Remove incorrect comment.
18404 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
18406 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
18407 TOC operand is a valid symbol ref in the constant pool.
18409 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18411 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
18413 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
18416 * config/avr/stdfix.h: New file.
18417 * t-avr (stdfix-gcc.h): New rule to build it.
18418 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
18420 2013-01-23 Kostya Serebryany <kcc@google.com>
18422 * config/darwin.h: remove dependency on
18423 CoreFoundation (asan on Mac OS).
18425 2013-01-23 Jakub Jelinek <jakub@redhat.com>
18428 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
18429 instead of cmpdi_operand for first comparison operand.
18430 Don't assert that comparison operands aren't both constants.
18432 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
18434 * doc/install.texi (Downloading the Source): Update references to
18435 downloading separate components.
18437 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
18439 * doc/extend.texi (__int128): Improve grammar.
18441 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
18444 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
18445 alternative to (o,r).
18446 (*movdi_internal_rex64): Remove (!o,n) alternative.
18447 (DImode immediate->memory splitter): Remove.
18448 (DImode immediate->memory peephole2): Remove.
18449 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
18450 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
18451 alternative to (!o,*r).
18452 (*movtf_internal_sse): New pattern.
18453 (*movxf_internal_rex64): New pattern.
18454 (*movxf_internal): Disable for TARGET_64BIT.
18455 (*movdf_internal_rex64): Remove (!o,F) alternative.
18457 2013-01-22 Jakub Jelinek <jakub@redhat.com>
18459 PR middle-end/56074
18460 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
18461 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
18462 * tree-vect-loop-manip.c (find_loop_location): Also ignore
18463 stmt locations where LOCATION_LOCUS of the stmt location is
18464 UNKNOWN_LOCATION or BUILTINS_LOCATION.
18467 * config/i386/i386.md (UNSPEC_STOS): New.
18468 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
18469 *strsetqi_1): Add UNSPEC_STOS.
18471 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
18474 * doc/invoke.texi: Remove left over -Wsynth example.
18476 2013-01-21 Jakub Jelinek <jakub@redhat.com>
18478 PR tree-optimization/56051
18479 * fold-const.c (fold_binary_loc): Don't fold
18480 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
18481 a narrowing conversion, or widening conversion from signed
18484 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
18486 PR rtl-optimization/56023
18487 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
18488 dependent on debug instruction.
18490 2013-01-21 Martin Jambor <mjambor@suse.cz>
18492 PR middle-end/56022
18493 * function.c (allocate_struct_function): Call
18494 invoke_set_current_function_hook earlier.
18496 2013-01-21 Jakub Jelinek <jakub@redhat.com>
18498 * reload1.c (init_reload): Only initialize reload_obstack
18499 during the first call.
18501 2013-01-21 Marek Polacek <polacek@redhat.com>
18503 * cfgloop.c (verify_loop_structure): Fix up grammar.
18505 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
18507 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
18508 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
18510 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18513 * config/arm/marvell-pj4.md: Update copyright year.
18514 Fix up use of alu to alu_reg and simple_alu_imm.
18516 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
18518 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
18520 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
18523 * lra-constraints.c (curr_insn_transform): Don't reuse original
18524 insn for secondary memory move when memory mode should be different.
18526 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18528 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
18529 atomic_storedi_1): New patterns.
18531 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18533 btver2 pipeline descriptions.
18534 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
18536 * config/i386/i386.md (btver2_decode): New type attributes.
18537 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
18539 * config/i386/btver2.md: New file describing btver2 pipelines.
18541 2013-01-19 Andrew Pinski <apinski@cavium.com>
18543 PR tree-optimization/52631
18544 * tree-ssa-sccvn (visit_use): Before looking up the original
18545 statement, try looking up the simplified expression.
18547 2013-01-19 Anthony Green <green@moxielogic.com>
18549 * config/moxie/moxie.c (moxie_expand_prologue): Set
18550 current_function_static_stack_size.
18552 2013-01-18 Jakub Jelinek <jakub@redhat.com>
18554 PR tree-optimization/56029
18555 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
18556 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
18558 2013-01-18 Sharad Singhai <singhai@google.com>
18560 PR tree-optimization/55995
18561 * dumpfile.c (dump_loc): Print location only if available.
18562 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
18564 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
18567 * lra-constraints.c (curr_insn_transform): Reuse original insn for
18568 secondary memory move.
18569 (inherit_reload_reg): Use rclass instead of cl for
18570 check_secondary_memory_needed_p.
18572 2013-01-18 Jakub Jelinek <jakub@redhat.com>
18574 PR middle-end/56015
18575 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
18576 the case where writing real complex part of target modifies op1.
18578 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
18580 * config/aarch64/aarch64-simd.md
18581 (aarch64_vcond_internal<mode>): Handle unordered cases.
18582 * config/aarch64/iterators.md (v_cmp_result): New.
18584 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
18585 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18587 * config/arm/marvell-pj4.md: New file.
18588 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
18589 * config/arm/arm.md (generic_sched): Add marvell_pj4.
18590 (generic_vfp): Likewise.
18591 * config/arm/arm-cores.def: Add marvell-pj4.
18592 * config/arm/arm-tune.md: Regenerate.
18593 * config/arm/arm-tables.opt: Regenerate.
18594 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
18595 * doc/invoke.texi: Document marvell-pj4.
18597 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
18599 * config/aarch64/arm_neon.h: Map scalar types to standard types.
18601 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18606 * var-tracking.c (negative_power_of_two_p): New.
18607 (global_get_addr_cache, local_get_addr_cache): New.
18608 (get_addr_from_global_cache, get_addr_from_local_cache): New.
18609 (vt_canonicalize_addr): Rewrite using the above. Adjust the
18611 (vt_stack_offset_p): Remove.
18612 (vt_canon_true_dep): Always canonicalize loc's address.
18613 (clobber_overlapping_mems): Make sure we have a MEM.
18614 (local_get_addr_clear_given_value): New.
18615 (val_reset): Clear local cached entries.
18616 (compute_bb_dataflow): Create and release the local cache.
18617 Disable duplicate MEMs clobbering.
18618 (emit_notes_in_bb): Clobber MEMs likewise.
18619 (vt_emit_notes): Create and release the local cache.
18620 (vt_initialize, vt_finalize): Create and release the global
18621 cache, respectively.
18622 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
18624 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18626 PR libmudflap/53359
18627 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
18628 not found in the symtab.
18630 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18633 PR rtl-optimization/55547
18634 PR rtl-optimization/53827
18637 * alias.c (offset_overlap_p): New, factored out of...
18638 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
18639 the conservative special case for symbolic constants. Don't
18640 adjust zero sizes on alignment.
18642 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
18644 PR rtl-optimization/52573
18645 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
18646 REG_UNUSED for the same register.
18648 2013-01-17 Richard Biener <rguenther@suse.de>
18649 Marek Polacek <polacek@redhat.com>
18651 PR rtl-optimization/55833
18652 * loop-unswitch.c (unswitch_loops): Move loop verification...
18653 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
18654 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
18655 Set it to true when we're removing a loop from hierarchy tree in
18656 an irreducible region.
18657 (fix_bb_placements): Adjust caller.
18658 (fix_loop_placements): Likewise.
18660 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
18662 * config/avr/builtins.def (DEF_BUILTIN): Factor out
18663 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
18664 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
18665 Remove ID. Adjust comments.
18666 * config/avr/avr-c.c (avr_builtin_name): Remove.
18667 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
18668 * config/avr/avr.c (avr_tolower): New static function.
18669 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
18670 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
18671 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
18674 2013-01-17 Jan Hubicka <jh@suse.cz>
18676 PR tree-optimization/55273
18677 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
18679 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
18682 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
18683 store through atomic_store<mode>_1.
18684 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
18686 2013-01-17 Martin Jambor <mjambor@suse.cz>
18688 PR tree-optimizations/55264
18689 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
18690 for virtual methods.
18691 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
18692 virtual methods before inlining is over.
18693 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
18695 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
18698 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
18700 PR rtl-optimization/56005
18701 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
18702 pending reads for prefetch.
18704 2013-01-16 Ian Bolton <ian.bolton@arm.com>
18706 * config/aarch64/aarch64.md
18707 (*cstoresi_neg_uxtw): New pattern.
18708 (*cmovsi_insn_uxtw): New pattern.
18709 (*<optab>si3_uxtw): New pattern.
18710 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
18711 (*<optab>si3_insn_uxtw): New pattern.
18712 (*bswapsi2_uxtw): New pattern.
18714 2013-01-16 Richard Biener <rguenther@suse.de>
18716 * tree-inline.c (tree_function_versioning): Remove set but
18717 never used variable.
18719 2013-01-16 Richard Biener <rguenther@suse.de>
18721 PR tree-optimization/55964
18722 * tree-flow.h (rename_variables_in_loop): Remove.
18723 (rename_variables_in_bb): Likewise.
18724 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
18725 (copy_loop_before): Adjust and delete update-ssa status.
18726 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
18727 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
18728 (rename_variables_in_loop): Remove.
18729 (slpeel_update_phis_for_duplicate_loop): Likewise.
18730 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
18731 use available cfg machinery instead of duplicating it.
18732 Update PHI nodes and perform poor-mans SSA update here.
18733 (slpeel_tree_peel_loop_to_edge): Adjust.
18735 2013-01-16 Richard Biener <rguenther@suse.de>
18737 PR tree-optimization/54767
18738 PR tree-optimization/53465
18739 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
18740 (vrp_visit_phi_node): For PHI arguments coming via backedges
18741 drop all symbolical range information.
18742 (execute_vrp): Compute backedges.
18744 2013-01-16 Richard Biener <rguenther@suse.de>
18746 * doc/install.texi: Update CLooG and ISL requirements to
18749 2013-01-16 Christian Bruel <christian.bruel@st.com>
18752 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
18753 (broken_move): Handle UNSPECV_SP_SWITCH_B.
18754 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
18756 2013-01-16 DJ Delorie <dj@redhat.com>
18758 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
18759 (UNSPECV_SP_SWITCH_E): New.
18760 (sp_switch_1): Change to an unspec.
18761 (sp_switch_2): Change to an unspec. Don't use post-inc when we
18764 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
18766 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
18767 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
18768 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
18769 (expand_mem_thread_fence): Ditto.
18770 (expand_mem_signal_fence): Ditto.
18771 (expand_atomic_load): Ditto.
18772 (expand_atomic_store): Ditto.
18774 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
18776 PR rtl-optimization/55547
18777 PR rtl-optimization/53827
18780 * alias.c (memrefs_conflict_p): Set sizes to negative after
18783 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18786 * function.c (thread_prologue_and_epilogue_insns): Always
18787 add crtl->drap_reg to set_up_by_prologue.set, even if
18788 stack_realign_drap is false.
18790 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18792 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
18793 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
18794 *call): Fix indention.
18796 2013-01-15 Tom de Vries <tom@codesourcery.com>
18799 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
18802 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
18804 PR rtl-optimization/55153
18805 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
18807 2013-01-15 Martin Jambor <mjambor@suse.cz>
18809 PR tree-optimization/55920
18810 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
18811 accesses as grp_to_be_debug_replaced.
18813 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18815 PR tree-optimization/55920
18816 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
18817 there is non-useless type conversion needed from debug rhs to lhs,
18818 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
18820 2013-01-15 Joseph Myers <joseph@codesourcery.com>
18821 Mikael Pettersson <mikpe@it.uu.se>
18824 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
18826 (ASM_OUTPUT_CASE_LABEL): Remove.
18827 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
18828 * final.c (shorten_branches): Update alignment of labels before
18829 jump tables if CASE_VECTOR_SHORTEN_MODE.
18831 2013-01-15 Richard Biener <rguenther@suse.de>
18834 * system.h: Do not include gmp.h for building host tools.
18836 2013-01-15 Richard Biener <rguenther@suse.de>
18838 PR middle-end/55882
18839 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
18840 account for bitpos when computing alignment.
18842 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
18844 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
18845 (ix86_target_macros_internal): Likewise.
18847 * config/i386/i386.c (m_CORE2I7): Removed.
18848 (m_CORE_HASWELL): New macro.
18849 (m_CORE_ALL): Likewise.
18850 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
18851 (initial_ix86_arch_features): Likewise.
18852 (processor_target_table): Initializations for Core avx2.
18853 (cpu_names): New names "core-avx2".
18854 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
18855 PROCESSOR_CORE_HASWELL.
18856 (ix86_issue_rate): New case.
18857 (ia32_multipass_dfa_lookahead): Likewise.
18858 (ix86_sched_init_global): Likewise.
18860 * config/i386/i386.h (TARGET_HASWELL): New macro.
18861 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
18862 (processor_type): New PROCESSOR_HASWELL.
18864 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18866 PR tree-optimization/55955
18867 * tree-vect-loop.c (vectorizable_reduction): Give up early on
18868 *SHIFT_EXPR and *ROTATE_EXPR codes.
18870 PR tree-optimization/48766
18871 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
18872 -ftrapv disable -fwrapv.
18874 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18877 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
18878 etc. to 1 and not to __flash.
18879 Use LL suffix for __INT24_MAX__ with -mint8.
18880 Use ULL suffix for __UINT24_MAX__ with -mint8.
18882 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18884 * config/avr/avr-arch.h
18885 (struct base_arch_s): Use typedef avr_arch_t instead.
18886 (struct arch_info_s): Use typedef avr_arch_info_t instead.
18887 (struct mcu_type_s): Use typedef avr_mcu_t instead.
18888 * config/avr/avr.c: Same.
18889 * config/avr/avr-devices.c: Same.
18890 * config/avr/driver-avr.c: Same.
18891 * config/avr/gen-avr-mmcu-texi.c: Same.
18892 * config/avr/avr-mcus.def: Adjust comment.
18894 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
18896 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
18897 * config/aarch64/iterators.md (VALLDI): New.
18899 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18900 Andi Kleen <ak@linux.intel.com>
18903 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18904 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18907 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18909 * config/avr/avr-stdint.h: Remove trailing blanks.
18910 * config/avr/avr-log.h: Same.
18911 * config/avr/avr-arch.h: Same.
18912 * config/avr/avr-devices.c: Same.
18913 * config/avr/avr-dimode.md: Same.
18914 * config/avr/predicates.md: Same.
18915 * config/avr/avr-c.c: Same. And fix typo.
18917 * config/avr/avr-protos.h: Same. And:
18918 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18919 (init_cumulative_args): Rename to avr_init_cumulative_args.
18920 (expand_prologue): Rename to avr_expand_prologue.
18921 (expand_epilogue): Rename to avr_expand_epilogue.
18922 (adjust_insn_length): Rename to avr_adjust_insn_length.
18923 (notice_update_cc): Rename to avr_notice_update_cc.
18924 (final_prescan_insn): Rename to avr_final_prescan_insn.
18925 * config/avr/avr.c: Same.
18926 * config/avr/avr.h: Same.
18927 * config/avr/avr.md: Remove trailing blanks.
18928 (prologue): Use avr_expand_prologue.
18929 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18931 2013-01-14 Richard Biener <rguenther@suse.de>
18933 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18934 verify_location, collect_subblocks): New functions.
18935 (verify_gimple_in_cfg): Verify that locations only reference
18936 BLOCKs in the functions BLOCK tree.
18938 2013-01-14 Richard Biener <rguenther@suse.de>
18940 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18942 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18944 (insert_out_of_ssa_copy_on_edge): Likewise.
18945 (rewrite_close_phi_out_of_ssa): Likewise.
18946 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18948 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18949 propagated constants.
18950 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18953 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18955 * config/avr/avr-modes.def: Add GPL copyright notice.
18957 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18959 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18960 MEMMODEL_MASK to determine memory model.
18961 (atomic_store<mode>): Ditto from operands[2].
18962 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18964 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18967 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18968 (fold_gimple_assign): Don't call unshare_expr here.
18969 (fold_ctor_reference): Call unshare_expr.
18971 2013-01-13 Terry Guo <terry.guo@arm.com>
18973 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
18974 * doc/fragments.texi: Document MULTILIB_REUSE.
18975 * gcc.c (multilib_reuse): New internal spec.
18976 (set_multilib_dir): Also search multilib from multilib_reuse.
18977 * genmultilib (tmpmultilib3): Refactor code.
18978 (tmpmultilib4): Ditto.
18979 (multilib_reuse): New multilib argument.
18981 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
18983 * Makefile.in: Update copyright.
18985 2013-01-12 Tom de Vries <tom@codesourcery.com>
18987 PR middle-end/55890
18988 * calls.c (expand_call): Check if arg_nr is valid.
18990 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18992 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
18993 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
18994 documentation. Add missing '__' in front of
18995 __builtin_ia32_packssdw256.
18997 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19000 * config/s390/s390.c (s390_preferred_reload_class): Do not return
19001 NO_REGS for larl operands.
19002 (s390_reload_larl_operand): Use s390_load_address instead of
19005 2013-01-11 Richard Biener <rguenther@suse.de>
19007 * tree-cfg.c (verify_node_sharing_1): Split out from ...
19008 (verify_node_sharing): ... here.
19009 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
19011 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
19013 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
19014 Substitute TREECHECKING.
19015 * configure: Regenerate.
19016 * Makefile.in (TREECHECKING): New.
19018 2013-01-11 Richard Guenther <rguenther@suse.de>
19020 PR tree-optimization/44061
19021 * tree-vrp.c (extract_range_basic): Compute zero as
19022 value-range for __builtin_constant_p of function parameters.
19024 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
19026 Update copyright years.
19028 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
19030 PR rtl-optimization/55672
19031 * lra-eliminations.c (mark_not_eliminable): Permit addition with
19032 const to be eliminable.
19034 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
19036 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
19037 * configure: Regenerate.
19039 2013-01-10 Richard Biener <rguenther@suse.de>
19041 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
19043 2013-01-10 Richard Biener <rguenther@suse.de>
19046 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
19047 locations for virtual PHI arguments.
19048 (rewrite_update_phi_arguments): Likewise.
19050 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
19052 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
19055 2013-01-10 Jakub Jelinek <jakub@redhat.com>
19057 PR tree-optimization/55921
19058 * tree-complex.c (expand_complex_asm): New function.
19059 (expand_complex_operations_1): Call it for GIMPLE_ASM.
19061 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19064 * config/s390/s390.c (s390_symref_operand_p)
19065 (s390_loadrelative_operand_p): Merge the two functions.
19066 (s390_check_qrst_address, print_operand_address): Add parameters
19067 to s390_loadrelative_operand_p invokation.
19068 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
19069 (s390_reload_larl_operand, s390_secondary_reload): Use
19070 s390_loadrelative_operand_p instead of s390_symref_operand_p.
19071 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
19073 2013-01-09 Mike Stump <mikestump@comcast.net>
19075 * dse.c (record_store): Remove unnecessary assert.
19077 2013-01-09 Jan Hubicka <jh@suse.cz>
19079 PR tree-optimization/55569
19080 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
19081 * cfgloop.h (scale_loop_profile): Likewise.
19083 2013-01-09 Jan Hubicka <jh@suse.cz>
19086 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
19088 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
19090 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
19092 PR middle-end/55114
19093 * expr.h (maybe_emit_group_store): Declare.
19094 * expr.c (maybe_emit_group_store): New function.
19095 * builtins.c (expand_builtin_int_roundingfn): Call it.
19096 (expand_builtin_int_roundingfn_2): Likewise.
19098 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
19100 PR rtl-optimization/55829
19101 * lra-constraints.c (match_reload): Add code for absent output.
19102 (curr_insn_transform): Add code for reloads of matched inputs
19105 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
19107 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
19108 attribute of movddup insn to DF.
19109 (*vec_interleave_lowv2df): Ditto.
19110 (vec_dupv2df): Ditto.
19112 2013-01-09 Jan Hubicka <jh@suse.cz>
19114 PR tree-optimiation/55875
19115 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
19116 EVERY_ITERATION parameter.
19117 (number_of_iterations_exit): Check if exit is executed every iteration.
19118 (idx_infer_loop_bounds): Similarly here.
19119 (n_of_executions_at_most): Simplify
19120 to only test for cases where statement is dominated by the
19121 particular bound; handle correctly the "postdominance" test.
19122 (scev_probably_wraps_p): Use max loop iterations info
19123 as a global bound first.
19125 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
19126 Nick Clifton <nickc@redhat.com>
19128 * config/v850/v850.md (cbranchsf4): New pattern.
19129 (cstoresf4): New pattern.
19130 (cbranchdf4): New pattern.
19131 (cstoredf4): New pattern.
19132 (movsicc): Disallow floating point comparisons.
19133 (cmpsf_le_insn): Fix order of operators.
19134 (cmpsf_lt_insn): Likewise.
19135 (cmpsf_eq_insn): Likewise.
19136 (cmpdf_le_insn): Likewise.
19137 (cmpdf_lt_insn): Likewise.
19138 (cmpdf_eq_insn): Likewise.
19139 (cmpsf_ge_insn): Use LE comparison.
19140 (cmpdf_ge_insn): Likewise.
19141 (cmpsf_gt_insn): Use LT comparison.
19142 (cmpdf_gt_insn): Likewise.
19143 (cmpsf_ne_insn): Delete pattern.
19144 (cmpdf_ne_insn): Delete pattern.
19145 * config/v850/v850.c (v850_gen_float_compare): Use
19146 gen_cmpdf_eq_insn for NE comparison.
19147 (v850_float_z_comparison_operator)
19148 (v850_float_nz_comparison_operator): Move from here ...
19149 * config/v850/predicates.md: ... to here. Move GT and GE
19150 comparisons into v850_float_z_comparison_operator.
19151 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
19153 (v850_float_nz_comparison_operator): Likewise.
19155 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19157 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
19158 with calls to gen_insvsi/gen_insvdi.
19160 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19162 * config/i386/i386.c (initial_ix86_tune_features): Set up
19163 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
19165 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
19166 Jakub Jelinek <jakub@redhat.com>
19168 PR tree-optimization/48189
19169 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
19170 If nitercst is 0, don't predict the exit edge.
19172 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19174 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
19175 in asm_fprintf with reg_names.
19176 (aarch64_print_operand_address): Likewise.
19177 (aarch64_return_addr): Likewise.
19178 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
19180 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19182 * config/pa/pa.h (VAL_U6_BITS_P): Define.
19183 (INT_U6_BITS): Likewise.
19184 * config/pa/predicates.md (uint6_operand): New predicate.
19185 (shift5_operand, shift6_operand): Likewise.
19186 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
19188 (lshrdi3): Use shift6_operand.
19189 (shrpsi4, shrpdi4): New insn patterns.
19190 (extzv): Delete expander.
19191 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
19192 predicates in unamed zero extract patterns. Tighten common constraint.
19193 (extv): Delete expander.
19194 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
19195 predicates in unamed sign extract patterns. Tighten common constraint.
19196 (insv): Delete expander.
19197 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
19198 predicates in unamed insert patterns. Tighten common constraint.
19199 Change uint32_operand predicate to uint6_operand predicate in unamed
19200 DImode pattern to insert constant values of type 1...1xxxx.
19202 2013-01-04 Jan Hubicka <jh@suse.cz>
19204 PR tree-optimization/55823
19205 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
19208 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19209 Uros Bizjak <ubizjak@gmail.com>
19211 PR rtl-optimization/55845
19212 * df-problems.c (can_move_insns_across): Stop scanning at
19213 volatile_insn_p source instruction or give up if
19214 across_from .. across_to range contains any volatile_insn_p
19217 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19219 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
19220 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
19222 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
19223 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
19225 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19228 * asan.c (asan_clear_shadow): New function.
19229 (asan_emit_stack_protection): Use it.
19231 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19233 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
19234 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
19235 with tab instead of space.
19237 2013-01-08 Nick Clifton <nickc@redhat.com>
19239 * config/rl78/rl78.c (rl78_expand_prologue): Always select
19240 register bank 0 at the start of an interrupt handler.
19241 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
19244 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19246 * config/aarch64/aarch64-simd.md
19247 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
19248 (aarch64_simd_bsl): Likewise.
19249 (aarch64_vcond_internal<mode>): Likewise.
19250 (vcond<mode><mode>): Likewise.
19251 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
19252 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
19254 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19256 * config/aarch64/aarch64-builtins.c
19257 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
19259 2013-01-08 Martin Jambor <mjambor@suse.cz>
19262 * tree-sra.c (analyze_access_subtree): Return true also after
19263 potentially creating a debug-only replacement.
19265 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19267 PR middle-end/55890
19268 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
19270 PR tree-optimization/54120
19271 * tree-vrp.c (range_fits_type_p): Don't allow
19272 src_precision < precision from signed vr to unsigned_p
19273 if vr->min or vr->max is negative.
19274 (simplify_float_conversion_using_ranges): Test can_float_p
19275 against CODE_FOR_nothing.
19277 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19278 Richard Biener <rguenther@suse.de>
19280 PR middle-end/55851
19281 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
19282 types instead of just INTEGER_TYPE types.
19284 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
19286 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
19289 2013-01-07 Steve Ellcey <sellcey@mips.com>
19292 * config/mips/mips.opt: Change mad to mmad to match documentation.
19294 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19297 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
19298 .progmemx.data now.
19300 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19303 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
19304 (avr_addrspace_t): Add .section_name field.
19305 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
19307 (avr_addrspace): Same. Initialize .section_name. Remove last
19308 NULL entry. Put __memx into .progmemx.data.
19309 (progmem_section_prefix): Remove.
19310 (avr_asm_init_sections): No need to initialize progmem_section.
19311 (avr_asm_named_section): Use avr_addrspace[].section_name to get
19312 section name prefix.
19313 (avr_asm_select_section): Ditto. And use get_unnamed_section to
19314 retrieve the progmem section.
19315 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
19316 boundary to run over avr_addrspace[].
19317 (avr_register_target_pragmas): Ditto.
19319 2013-01-06 Jakub Jelinek <jakub@redhat.com>
19321 * varasm.c (output_constant_def_contents): For asan_protect_global
19322 protected strings, adjust DECL_ALIGN if needed, before testing for
19324 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
19325 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
19327 (output_object_block): For asan protected decls, emit asan padding
19328 after their contents.
19329 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
19330 (asan_finish_file): Test it here instead.
19332 2013-01-07 Nick Clifton <nickc@redhat.com>
19333 Matthias Klose <doko@debian.org>
19334 Doug Kwan <dougkwan@google.com>
19335 H.J. Lu <hongjiu.lu@intel.com>
19338 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
19340 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
19342 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
19344 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
19346 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
19348 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19351 * doc/install.texi (Cross-Compiler-Specific Options): Document
19354 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
19356 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
19357 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
19358 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
19359 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
19360 vqmovun_high_s64): Fix source operand number and update copyright.
19362 2013-01-07 Richard Biener <rguenther@suse.de>
19364 PR middle-end/55890
19365 * gimple.h (gimple_call_builtin_p): New overload.
19366 * gimple.c (validate_call): New function.
19367 (gimple_call_builtin_p): Likewise.
19368 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19369 Use gimple_call_builtin_p.
19370 (find_func_clobbers): Likewise.
19371 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
19372 (strlen_optimize_stmt): Likewise.
19374 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
19376 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
19377 (vld1q_dup_*): Likewise.
19378 (vld1_*): Likewise.
19379 (vld1q_*): Likewise.
19380 (vld1_lane_*): Likewise.
19381 (vld1q_lane_*): Likewise.
19383 2013-01-07 Richard Biener <rguenther@suse.de>
19385 * lto-streamer.h (LTO_minor_version): Bump to 2.
19387 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
19389 * config/aarch64/aarch64-protos.h
19390 (aarch64_const_double_zero_rtx_p): Rename to...
19391 (aarch64_float_const_zero_rtx_p): ...this.
19392 (aarch64_float_const_representable_p): New.
19393 (aarch64_output_simd_mov_immediate): Likewise.
19394 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
19395 move immediate case.
19396 * config/aarch64/aarch64.c
19397 (aarch64_const_double_zero_rtx_p): Rename to...
19398 (aarch64_float_const_zero_rtx_p): ...this.
19399 (aarch64_print_operand): Allow printing of new constants.
19400 (aarch64_valid_floating_const): New.
19401 (aarch64_legitimate_constant_p): Check for valid floating-point
19403 (aarch64_simd_valid_immediate): Likewise.
19404 (aarch64_vect_float_const_representable_p): New.
19405 (aarch64_float_const_representable_p): Likewise.
19406 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
19407 (aarch64_output_simd_mov_immediate): New.
19408 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
19409 (*movdf_aarch64): Likewise.
19410 * config/aarch64/constraints.md (Ufc): New.
19411 (Y): call aarch64_float_const_zero_rtx.
19412 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
19414 2013-01-07 Richard Biener <rguenther@suse.de>
19416 PR tree-optimization/55888
19417 PR tree-optimization/55862
19418 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
19419 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
19420 not if it is contained therein.
19422 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19424 * config/avr/t-avr: Typo.
19426 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19429 * config/avr/t-avr: Don't automatically rebuild
19430 $(srcdir)/config/avr/t-multilib
19431 $(srcdir)/config/avr/avr-tables.opt
19432 $(srcdir)/doc/avr-mmcu.texi
19433 (avr-mcus): New phony target to build them on request.
19434 (s-avr-mlib, s-avr-mmcu-texi): Remove.
19435 * avr/avr-mcus.def: Adjust comments.
19437 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
19439 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
19441 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
19443 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
19445 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
19447 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
19449 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
19451 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
19452 to generate profiling.
19453 * config/rs6000/aix64.h (LIB_SPEC): Same.
19455 2013-01-04 Andrew Pinski <apinski@cavium.com>
19457 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
19459 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
19461 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
19463 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
19465 (ix86_expand_move): Ditto.
19466 (ix86_zero_extend_to_Pmode): Ditto.
19467 (ix86_expand_call): Ditto.
19468 (ix86_expand_special_args_builtin): Ditto.
19469 (ix86_expand_builtin): Ditto.
19471 2013-01-04 Richard Biener <rguenther@suse.de>
19473 PR tree-optimization/55862
19474 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
19475 translating them through PHI nodes.
19477 2013-01-04 Martin Jambor <mjambor@suse.cz>
19479 PR tree-optimization/55755
19480 * tree-sra.c (sra_modify_assign): Do not check that an access has no
19481 children when trying to avoid producing a VIEW_CONVERT_EXPR.
19483 2013-01-04 Marek Polacek <polacek@redhat.com>
19485 PR middle-end/55859
19486 * opts.c (default_options_optimization): Clarify error message.
19488 2013-01-04 Richard Biener <rguenther@suse.de>
19490 PR middle-end/55863
19491 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
19494 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19497 * config/pa/pa.md (movsi): Revert previous change.
19498 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
19501 2013-01-03 Richard Henderson <rth@redhat.com>
19503 * config/i386/i386.c (ix86_expand_move): Always assign to op1
19504 after eliminating TLS symbols.
19506 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
19509 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
19510 * graphite-poly.c (debug_gmp_value): Likewise.
19512 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
19515 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
19516 selected code model, define __code_mode_small__, __code_model_medium__,
19517 __code_model_large__, __code_model_32__ or __code_model_kernel__.
19518 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
19519 xchg temporary register with %k. Declare temporary register as
19521 [__x86_64__]: For medium and large code models, preserve %rbx register.
19523 2013-01-03 Richard Biener <rguenther@suse.de>
19525 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
19526 (dump_subscript): Adjust.
19527 (finalize_ddr_dependent): Do not dump redundant info.
19528 (analyze_siv_subscript): Adjust.
19529 (subscript_dependence_tester): Likewise.
19530 (compute_affine_dependence): Likewise.
19532 2013-01-03 Richard Biener <rguenther@suse.de>
19535 2013-01-03 Richard Biener <rguenther@suse.de>
19537 PR tree-optimization/55857
19538 * tree-vect-stmts.c (vectorizable_load): Do not setup
19539 re-alignment for invariant loads.
19541 2013-01-02 Richard Biener <rguenther@suse.de>
19543 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
19544 invariant load do not generate a vector load from the scalar location.
19546 2013-01-03 Richard Biener <rguenther@suse.de>
19548 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
19549 for not vectorizing.
19550 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
19551 not build INDIRECT_REFs, call get_name once only.
19552 (vect_create_data_ref_ptr): Likewise. Dump base object kind
19553 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
19555 2013-01-03 Richard Biener <rguenther@suse.de>
19557 PR tree-optimization/55857
19558 * tree-vect-stmts.c (vectorizable_load): Do not setup
19559 re-alignment for invariant loads.
19561 2013-01-03 Richard Biener <rguenther@suse.de>
19564 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
19565 prefer a built-in decl.
19567 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19569 * gcc.c (process_command): Update copyright notice dates.
19570 * gcov.c (print_version): Likewise.
19571 * gcov-dump.c (print_version): Likewise.
19573 PR rtl-optimization/55838
19574 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
19575 iv0.step, iv1.step and step.
19577 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19578 Marc Glisse <marc.glisse@inria.fr>
19580 PR tree-optimization/55832
19581 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
19582 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
19583 integer_{one,zero}_node.
19585 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19588 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
19589 * var-tracking.c (reverse_op): Don't add reverse ops to
19590 VALUEs that have already
19591 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
19593 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
19595 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
19597 2013-01-02 Teresa Johnson <tejohnson@google.com>
19599 * dumpfile.c (dump_loc): Print filename with location.
19600 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
19601 new location_t parameter to emit complete unroll message with
19602 new dump framework.
19603 (canonicalize_loop_induction_variables): Compute loops location
19604 and pass to try_unroll_loop_completely.
19605 * loop-unroll.c (report_unroll_peel): New function.
19606 (peel_loops_completely): Use new dump format with location
19607 for main dumpfile message, and invoke report_unroll_peel on success.
19608 (decide_unrolling_and_peeling): Ditto.
19609 (decide_peel_once_rolling): Remove old dumpfile message subsumed
19610 by report_unroll_peel.
19611 (decide_peel_completely): Ditto.
19612 (decide_unroll_constant_iterations): Ditto.
19613 (decide_unroll_runtime_iterations): Ditto.
19614 (decide_peel_simple): Ditto.
19615 (decide_unroll_stupid): Ditto.
19616 * cfgloop.c (get_loop_location): New function.
19617 * cfgloop.h (get_loop_location): Declare.
19619 2013-01-02 Sriraman Tallam <tmsriram@google.com>
19621 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
19624 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19626 PR middle-end/55198
19627 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
19628 BLKmode objects when EXPAND_MEMORY is specified.
19630 2013-01-02 Sriraman Tallam <tmsriram@google.com>
19632 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
19634 (fold_builtin_cpu): Do not share cpu model decls across statements.
19636 2013-01-02 Jason Merrill <jason@redhat.com>
19639 * tree.c (build_array_type_1): Revert earlier change.
19641 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
19643 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
19645 * config/aarch64/aarch64-tune.md: Re-generate.
19647 2013-01-02 Richard Biener <rguenther@suse.de>
19649 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
19650 invariant load do not generate a vector load from the scalar location.
19652 2013-01-02 Richard Biener <rguenther@suse.de>
19655 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
19656 * configure: Regenerate.
19658 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
19660 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
19661 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
19662 (expand_builtin_int_roundingfn_2): Keep the original target around
19663 for the fallback case.
19665 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
19667 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
19668 to be clear for sign changes.
19670 2013-01-01 Jan Hubicka <jh@suse.cz>
19672 * ipa-inline-analysis.c: Fix formatting.
19674 2013-01-01 Jakub Jelinek <jakub@redhat.com>
19676 PR tree-optimization/55831
19677 * tree-vect-loop.c (get_initial_def_for_induction): Use
19678 gsi_after_labels instead of gsi_start_bb.
19680 Copyright (C) 2013 Free Software Foundation, Inc.
19682 Copying and distribution of this file, with or without modification,
19683 are permitted in any medium without royalty provided the copyright
19684 notice and this notice are preserved.